Contents:
Sam's Repair Briefs - Complete: 1 to 99
Copyright (c) 1994, 1995, 1996, 1997, 1998
All Rights Reserved
Reproduction of this document in whole or in part is permitted if both of the following conditions are satisfied:
This series presents case studies of selected repair problems from my archives. (Archives sounds more impressive than scribbled notes and schematics, doesn't it?). I hope that these summaries are useful for those of you are interested in tackling your own repairs. These will NOT be of the form: "replace C418 when vertical size is reduced on HyperTech TV chassis HT312". Rather, they will document specific but common problems with TVs, VCRs, CDs, computer peripherals, etc. The symptoms, testing, diagnostic procedure, repair procedure, and comments will be included so that you can learn from my approach (and my screwups). If you want specific solutions to well known or repeat problems with your RCA or Zenith, then the repair professionals who frequent this group will be in a better position to help since they work on many of the same line of equipment on a regular basis. However, if you want to develop a general diagnostic approach, then this series may provide some tips and insight based on theory and experience. Not all of these are exciting cast-of-thousands repairs. Many tend to be mundain but address common problems with consumer electronic equipment. You will no doubt recognize some as directly pertaining to some appliance or electronic device that you had repaired (or junked) in the past. There will probably even be some of those 'dogs' that we all hate - the problems that never seem to go away. In any case, these will all be based on my true experiences with minimal embellishment.
Note: These articles are archived at: * http://www.repairfaq.org/ under the heading: Assorted Repair Briefs. These are maintained by Filip Gieszczykiewicz. Repair Guides are also available at this URL for devices and problems similar to those covered in the Repair Briefs(c) series. ************************************************************************ 1 * Daisy Wheel Printer - Carriage gets stuck * 2 * Panasonic PV1461 VCR with Dead Power Supply * 3 * USR Data Modem Won't Dial * 4 * RCA 25" TV Blows Fuse * 5 * Pioneer CD Changer will not Recognize Disc * 6 * Dead Power One 400 W Multiple Output Switcher * 7 * Goldstar CR820U TV with no Color * 8 * Panasonic PV3720 VCR with Erratic Horizontal Video Noise * 9 * Original ATT Touch Tone Phone will not Dial * 10 * Pioneer PD5100 CD Player Trashed * 11 * Yamaha R8 Receiver with Tuning Problem * 12 * Sylvania TV with no Horizontal Sync * 13 * Dead Microwave * 14 * Tandy Color TV/Monitor with hum bars * 15 * Realistic CD Player Randomly Shuts Off * 16 * Sony CDU31A CD ROM Drive Failure * 17 * Panasoic Color TV Shakes and Pulsates * 18 * Dead Zenith color TV * 19 * Goldstar VCR caught in infinite eject loop * 20 * Nintendo Game Console will not Reset * 21 * Sharp VC7864U VCR Erratic * 22 * Magnavox Console TV Deflection Problems * 23 * Hewlett Packard AN/USM281A Oscilloscope with Multiple Problems * 24 * Sony Servolock Turntable Erratic * 25 * Kenwood KX-55C Cassette Deck with dead Transport * 26 * Kenwood CD Changer with erratic behavior * 27 * Panasonic PV 4820 VCR with multiple problems * 28 * JVC HR-D860U VCR mangled tape * 29 * Panasonic VCR with video noise, audio hum, and erratic operation * 30 * JVC 6 Disc CD Changer gets stuck * 31 * Mitsubishi HS-318UR VCR Dropped * 32 * Toshiba T1000XE PA8706 AC Adapter * 33 * Mac Plus with no Video * 34 * RCA EFR485 (CTC111) Color TV Part 1: Messed up Colors * 35 * Panasonic PV1414 Closed Circuit TV Camera with no Video * 36 * Heathkit IM2202 Digital Multimeter - Dead * 37 * RCA EFR485 (CTC111) Color TV Part 2: No Picture * 38 * SYLVANIA RXX170-WA01 Color TV Clicks - No Power * 39 * Sony Compact Stereo Model HP-179-P - FM Dead * 40 * Mitsubishi HS-U53 - Bad Tracking * 41 * Mitsubishi FA2100-CW FAX Machine will not Send * 42 * Mitsubishi HS3381UR VCR Tracking Bands * 43 * Panasonic PV2812 VCR - no Play or Record * 44 * JVC HRD-550U VCR - Tracking Problems * 45 * SEARS (Goldstar) VCR Part 1: Broken Cassette Loader * 46 * Beckman Model 310B Digital Multimeter with Random Display * 47 * Quasar TT4259WW Color TV - Dead * 48 * SEARS (Goldstar) VCR Part 2: Shutdown in Play and Record * 49 * HP DeskJet Professional Printer - Part 1: Dead * 50 * Panasonic PV1545 HiFi VCR - Eats Tapes * 51 * HP DeskJet Professional Printer - Part 2: Missing Lines * 52 * Aztech CDA-268-01A CDROM Drive - Drawer Continuously Closing * 53 * Craftsman Electric Drill - Worn Motor Bearing * 54 * NEC CDR-260 Double Speed CDROM Drive - Intermittent * 55 * RCA FPR560ER Color TV - Erratic Behavior * 56 * Toshiba M4200 VCR - Weak Intermittent Audio * 57 * HP DeskJet Professional Printer - Part 3: Print Fades Out * 58 * Magnavox FD2000-SL01 CD Player - Dead * 59 * GE 13AC1504W Color TV - Dead (with other problems) * 60 * Canon FaxPhone 80 - More or Less Dead * 61 * Sony D14 Portable Compact Disc Player - Smoked * 62 * Zenith System 3 Color TV - Intermittent Blue Fog * 63 * Sony Boombox CD Player - Erratic Shutoff * 64 * Book Tape Player - Missing Channels 3 and 4 * 65 * Aiwa CSD-707 Boombox CD Player - Doesn't Recognize CDs * 66 * Sony KV-13TR-20 Color TV - Erratic Reception * 67 * Magnavox Phonograph - Record Changer Problems * 68 * Clocks, Clock Radio, and Cordless Phone - Smoked * 69 * GE Portable Color TV - Dropped * 70 * Nutone RF Wireless Chime - No Response to Button * 71 * Tandy Color TV/Monitor - Intermittent Shutdown * 72 * Braun Quartz Clock - No Movement * 73 * Sony D88 Portable CD Player Part 1 - Does Not Recognize Discs * 74 * Sony D88 Portable CD Player Part 2 - Erratic Audio Noise * 75 * Panasonic CT-714 Color TV - Dead * 76 * Emerson VGA Monitor - Dead Doggy * 77 * Sony KV-2675R Color TV - Will Not Power Up * 78 * Magnavox 31 Inch TV - Dropped and Fried * 79 * ConairPhone Desk Phone - Almost Dead * 80 * Sharp 13KM15 Color TV - Dead * 81 * Sony KV-19TR20 Color TV - No Reception * 82 * Emerson CGA Color Monitor - Dead * 83 * Tatung CM1495 Multisync Color Monitor - Dead * 84 * Sony Portable CD Player - Dead * 85 * Lambda LUS-8-12 Switching Supply - Fried * 86 * Fisher Power Amp with Blown Channel * 87 * Mac Plus and Original Apple 20 MB Hard Drive - Multiple Problems * 88 * HP DeskJet 500C Color Ink-Jet printer - No Printing * 89 * Craftsman Eager 1 Lawn Mower - Rod Disaster * 90 * Harman-Kardon Model 520 Stereo Receiver - Multiple Problems * 91 * Multifunction Desk Phone - Erratic Operation * 92 * Aiwa Sports 250 Boombox - Dropped * 93 * Kenmore Microwave - No Heat and Dead Buttons on Touchpad * 94 * Hoover Vacuum Cleaner Doesn't Pick Up * 95 * Subaru Auto AM/FM/cassette - Intermittent Reception * 96 * GE Frost-Free Refrigerator - Reduced Cooling * 97 * Panasonic VCR Power Supply - Comedy of Errors * 98 * Kenmore Window Fan - No Breeze * 99 * JVC 20 Inch Color TV - Flat Line * ************************************************************************
Patient: Panasonic KX-P3131 Daisy Wheel PC printer.
Symptoms: After some random amount of printing, the carriage would get
stuck and just twitch instead of moving across the platen.
Testing: Using its internal test function, I let it print out until the
problem ocurred.
Once the problem ocurred, the carriage would just kind of twitch back
and forth. Grabbing it, it was obvious that the stepper motor was providing
no effective power. Since this is usually indicative of a missing phase
to the stepper motor, I immediately suspected either:
1. Bad connection - cable or soldering.
2. Bad driver.
3. Bad motor.
Since it was intermittent with no relationship to time, heat, position, etc.,
this is most likely (1).
I then removed the cover to gain access to the circuit board and motor
connections. It was necessary to defeat the cover interlock to get
it to come one. Now, to get it to screw up again.
I left it printing out the ASCII character set and got a byte to eat.
When I came back it was busy gouging a hole in the paper. So now for
the critical test: Will pressing on the stepper motor connector cause
a change? The answer is --- Yes! The carriage started moving again meaning
that it is likely a bad connection.
Of course, to gain access to the underside of the circuit board required
removing a zillion screws and the entire mechanical assembly. But once
this was accomplished - immediate gratification. There were obvious
bad solder joints around several pins of the stepper motor connector.
I resoldered these and few others that were suspicious and inspected the
rest of the board. If only I could remember which screws went where!
Apparently, the continuous vibration of the assembly eventually caused
the connections to fail. This is not likely a heat related problem though
it could be just plain bad quality control.
Once reassembled, I left it happily printing out page after page of ACSII
characters. Then, just to be sure, I connected an old laptop and printed
a few pages of Repair Notes.
Comments: This is one of those dream problems since their solution is so
obvious and so definitive. There is no doubt that the cure will last.
Unfortunately, the tough 'dogs' are the ones you lose hair over.
Patient: Garage Sale Panasonic PV1461 VCR (Asked $10, offered and accepted $5).
Symptoms: The power supply had been identified as being bad by the seller.
Testing: I plugged it in - nothing as expected. This was to verify that
all functions and display were dead.
I removed cover and found that the main fuse had been removed. Hmmm, this
usually means catastrophic power supply problem. This is the typical Panasonic
switching power supply. Well, I have completely rebuilt these, so no biggie.
(Of course, there was one I blew up, but that is another story).
The hardest part was removing the power supply. It is buried underneath
the bottom circuit board necessitating the removal of this board and the
front panel. Do you think they design these things this way to discourage
tampering? There was plastic in specific places to prevent removal from
the top even though it would have been trivial to design for easy removal.
I disconnected power supply from VCR. Fortunately, this was just a connector.
I tested across switchmode power transistor with ohmmeter - dead short C-B-E.
I looked in ECG for 2SC3890 - ECG379 with that infamous # indicating an
electrical but not mechanical match.. Since I had some BU406s, I looked
this number up in ECG guessing that the BU406 would be have similar ratings
and be usable in a small switcher - guess what, ECG379. The difference is
that the 2SC3890 is totally plastic while the BU406 is a metal tab TO220. OK,
so I cut out a bit of mica to serve as an insulator and used a nylon screw.
This is temporary as I intend to get the proper replacement (2SC3890 - $2.15
from MCM Electronics).
I also checked continuity from the main filter cap to the C and E leads
of the transistor to rule out a blown fusable emitter resistor. I checked
other semiconductors as well - all fine as far as my VOM was concerned.
Fortunately, the only casualty seems to have been the transistor and the
fuse was fast enough to prevent any damage due to its shorting.
To power up the supply, I initially used a Variac with a 25 W light bulb
in series with the line. Note that since I do not have any 1.6 A
fuses, the fuse is shorted. The light bulb will provide the current limiting
for now. I use my dual outlet widget box plugging in the supply to one
outlet and a lamp with the 25W light bulb into the other (the outlets are
wired in series for exactly this sort of application). This whole rig
was plugged into an isolation trasformer for safety.
I then identified the primary output and connected my VOM to this. It would
be in the range 5-15 V, probably 12 V based on the filter capacitors (16 V).
All set? Crank up power. Output comes up to about 13 V at around 50 VAC
in. Light Bulb hardly flickered. If it had not stopped in the 12 V range,
this would either indicate that there was a problem in the regulation or
that a load was needed. Had this happened, I would have put a 22 ohm 5W
resistor on this output and retested. Cranking up the Variac to full
voltage causes no noticeable change to output.
OK, connect VCR. Lights, VCR, power! Nothing. Light bulb is now
glowing but there is no indication of life from front panel and no action
from motors.
Hmmmm.
Examining the nameplate, the expected power consumption is 29 W which is
way more then one can get through a 25 W light bulb. (Expect to be able
to draw maybe half of the bulb's ratings). So I go hunting for a 60 W
light bulb.
This time the motors twitch and the front panel comes alive.
Inserting a tape and holding my breath - tape starts to load but
then aborts with poweroff. Go for it! The Variac had been set at
about 90 VAC, so I crank it up to 120. Now everything works. I go
dig up a TV and verify that the basic functions are ok. Doesn't appear
to even need much cleaning. Even the idler tire appears to be in good
condition.
Add a 2SC3890 and box of 1.6 A fuses to my next MCM order.
Comments: this is the sort of repair that might not pay for a professional
shop to undertake. The time to disassemble the VCR, identify the problem,
replace the transistor and fuse, and verify correct operation could be
excessive, or at least has the potential to be excessive. If it turned
out to be more than a transistor and fuse or was beyond repair, they might
have to eat the cost in time and materials. In addition, there is no real
way to guarantee that other marginal components won't cause the problem to
repeat a week or month in the future. Upgrade/repair kits are available for
these supplies and would probably represent the lowest risk investment for
a permenant repair. No doubt, the previous owner had taken it in for repair
and been quoted a rediculous price to replace the entire power supply module.
Patient: USR 14.4 K baud external data modem.
Symptoms: Modem works in most respects but when it goes to dial, the tones
are superimposed on the dialtone which never goes away.
Testing: I plugged it into an old laptop kept for the specific purpose of
testing of random external peripherals. Indeed, the AT commands
worked just fine but tone dialing did not work. Interestingly,
pulse dialing would work 90% of the time but a connection was
never completely established.
At this point, I put a scope on each side of the 600 ohm coupling transformer.
A normal 2 or 3 volt p-p signal was present on the logic side of the
transformer during dialing. However, a much attenuated signal was present
on the phone line side - probably .2 V p-p or so. I probably could have
used a set of crystal headphones to just listen to the relative amplitudes
of the logic and phone line side signals instead of a scope. (Magnetic
headphones would have too low an input impedance.)
On a hunch, I did try replacing the transformer with one from my junkbox
but as expected, this produced no change.
The phone line is a nasty place for electronic components - 90 V ringing
signal, lightning strikes, pickup of EMPs (from nuclear bombs), etc.
The first place to look for fried components is therefore the phone line
circuitry. For a modem and no schematics, with the possible exception
of the power supply, it is also probably the only place where there is
any real chance of finding a problem.
Since there are a manageable number of discrete parts that connect the
phone line to the transformer, an ohmmeter check was in order. Unfortunately,
many of them were really itty bitty surface mount parts. After a couple
of go-arounds, this proved to bear fruit as an SO23 device marked on the
circuit board as a diode turned out to be shorted though I seem to have
missed it on my first pass. Carefully unsoldering the almost microscopic
part confirmed that it had turned into a dead short.
Note: markings for these devices were not always complete. Nonetheless,
basic ohmmeter checks could be made with enough confidence to tentatively
eliminate all except the single shorted diode.
Removing the diode and retesting proved that dialout was possible and that
normal communications at 14.4 K baud was normal. The markings on the diode
did not permit me to identify whether it was a simple diode or a zener.
I still do not know what function the diode actually served.
I replaced it with a 1N4148. The modem has been tested to confirm that
it is not damaged by the ringing voltage. Since the application does not
require dial-in access, I do not know if there is still a problem with
this mode.
Comments: Where a problem can be narrowed down to a small section of
circuitry, ohmmeter tests can prove successful in identifying parts
that have failed shorted or open. This may be the only option when
confronted with a device for which obtaining a schematic would be
difficult or impossible - or not worth the effort. While we might
consider a modem to be a throwaway item these days, the time need to
do basic testing of the phone line side components is minimal and,
as in this case, may be all that is needed.
Patient: RCA 26" TV, Chassis CTC131A/M. Garage sale acquisition ($5).
Previous owner explained that it failed to work after power
was restored following a power outage due to a thunderstorm.
They were more than happy to have me remove it from their garage.
Symptoms: Dead. Located fuse - blown due to short (not overload). Blackened
inside glass. Note: appearance of blown fuses is very significant.
Testing: Replaced fuse. This was in hindsight a mistake without using
a series light bulb to limit the current. The new fuse did not
blow but the lights dimmed momentarily. Apparently a fusable
resistor had sacrificed itself to protect the fuse.
Without schematics, I decided to trace the AC line circuits. These were
partially buried inside a metal box with some big fat resistors on a
separate board.
Measuring across the power resistors revealed that one 2.7 ohm resistor
was now open. That explained the new failure. Some further searching
located a TO3 power transistor inside the metal shield. Measuring C-E came
up 0 ohms. This was not the horizontal output transistor but was rather
in the power supply - a flyback switching supply separate from the deflection
circuits.
The transistor was marked with an RCA part number. My handy dandy ECG
Semiconductor Master Substitution guide listed it as a typical 350/800 volt
switching transistor. I did not have an exact replacement but figured
that a horizontal output transistor would probably work at least temporarily
in this application, so I used my favorite BU208A in its place.
As a temporary substitute for the fusable 2.7 ohm resistor, I put in
a regular power resistor with the understanding that a safe replacement
would be installed before the TV was buttoned up.
Now for the test. This would be classic use for a series light bulb and/or
variac for the initial power on. However, for whatever reasons, I did not
bother - and lucked out in this case. The TV came on just fine. The only
adjustments needed were to focus and horizontal position - totally unrelated
to the original problem.
The recommended ECG replacement for the chopper transistor was an ECG385,
a 350 V 10 A switching transistor. I elected to put in an ECG386 to give
myself a little extra margin. The ECG386 is rated at 500 V and 20 A.
Other specifications were similar or superior.
Comments: Since most modern consumer electronics are powered all the time
even if the power switch is off, it is always a good idea to unplug everything
during a lightning storm or if a blackout occurs. A nearby lightning
strike can easily impose huge transients on the AC line. When power is
restored following a failure, the initial power-on may not be clean including
mini-brownouts, spikes, multiple cycles, etc. These are all hard on
switching supply based equipment. Now, I fully realize that few of ua
actually follow this advise. (Of course, other screwups can result
in similar damage. I once was given a bag of dead stuff from a friend
of a friend who had been doing a little wiring in his house. Somehow
he managed to connect 240 V to a 120 V circuit - only for a second....)
This is one example where a failure with the most catastrophic impact
on performance (it was dead, after all) has among a very simple solution
(transistor, resistor, fuse). I much prefer these to the 'color noise
on channels in the UHF band' variety.
Patient: Pioneer PDM601 6 Disc CD Changer rejects all CDs.
Symptoms: Disc is loaded from cassette, disc spins at various speeds
for about 5 seconds, then it gives up, unloads this disc and
loads next until there are no more.
Testing: Testing consisted of removing cover and observing behavior.
Everything appears to happen normally except that disc directory
is never read.
This is one of those problems that has an obvious cause and solution once
experienced, diagnosed, and repaired the first time. I will outline my
approach the first time I came across this one describing this
particular case history from the perspective of the novice wannabe
repair expert.
Since the disc is loaded and spins, it is likely that the laser and focus
servo are functioning (though perhaps no guarantee, but I hope to get to that
in another episode).
At first, I suspected (incorrectly) that an adjustment was needed, so
I did what I always now warn against - turning any of the internal controls.
I thought that they had been returned to the original settings but was
not positive.
Pioneer CD players usually include a test function - a button on the main
circuit board marked 'TEST'. Normally, test functions are invoked either
by simply depressing the button or by holding it depressed when power is
turned on. In this case I discovered that if this button were depressed
when the unit was switched on, the display would change and certain front
panel buttons would now function controlling the servo circuits directly.
(Note that the approximate duration of the previous paragraph was about
1 year as I put the thing aside unable to make any headway during the
first go-around.)
After some experimentation with the front panel controls, I came to the
conclusion that:
* STOP turns all servos OFF.
* TRACK FWD enable FOCUS servo (and loads disc 1 in changer).
* PLAY enables SPINDLE servo.
* PAUSE enables TRACKING servo (and PLAY mode).
* MANUAL SEARCH FWD or REV to move Laser head.
Playing with these for awhile revealed that it was possible to move
the laser pickup, engage the focus servo, and make the disc rotate, Further
experimentation confirmed that the internal controls were more or less
in their correct positions - tracking offset could make the pickup slew from
one end to the other, for example.
All this was leading nowhere until I accidentally happened to engage
all servos in the middle of a CD - and the track display suddenly appeared.
I reached for my headphones and confirmed that the CD was playing, just fine.
A little further investigation allowed me to determine that the CD would
play fine from about the middle to the end but would get progressively
noisier when moving toward the starting track and would be totally
unplayable at the very start.
Now, what could depend so fundamentally on track position? Well, there
are two possibilities: spindle motor speed and PLL frequency. A little
careful tweaking of PLL center frequency had little effect.
One thing I noticed was that one of the servo driver ICs was running
quite hot. This should have raised a red flag but again, this was
the first time I had seen this problem. I also observed that putting
a heatsink on the IC and blowing on this would permit the disc to play
error free much closer to its start.
So now, you are saying, "what a moron, everyone knows that Pioneer CD
spindle motors are crap". The confusing thing here was that the spindle
motor was not dead, just marginal. So, all basic observations came
up negative.
Anyway, back to the saga. Suspecting the driver IC, I obtained a replacement
from MCM Electronics and swapped it. No change. Measuring motor voltage
showed a maximum of 1.7 V or so at any time including startup. Since the
driver is known now to be good and power was confirmed to be stable, I
started to suspect the motor.
Disconnecting the spindle motor and cycling the player revealed that the driver
was putting almost 10 V on the motor terminals but the motor was no doubt
partially shorted and dropping this to less than 2 V with the consequential
high power dissipation in the driver.
Now for the long shot. While the player was attempting to spin up
and read the disc directory, I gave the motor a squirt of degreaser
through its ventilation holes. The motor took off - went totally
overspeed. Power off. Wait for degreaser to evaporate. Try again.
Now, the directory came up the first time even though my internal controls
were still no doubt not perfectly tweaked. All functions worked perfectly.
For the first time is about 2 years, the player was producing music (without
the help of TEST mode!
I then performed a normal alignment of the internal controls (but this
is another story).
Measuring the motor voltage now showed greater than 5 V at spinup and a
range of 2-.5 V between start and end of disc.
Comments: At this point the proper course of action would be to replace
the spindle motor. However, since this is my CD player and replacing spindle
motors is sometimes a pain, I will just keep an eye on performance.
A pretty good indication of the motor's state is the time to spinup. If
this deteriorates again I will be forced to replace the motor. For now,
it continues to be satisfactory.
The initial confusion here was due to the fact that the motor was not
totally dead, just weak enough to cause a problem with the inner tracks
and more importantly, the directory. This is one of those cases where
the old style turntable with a bad weak motor would have been much easier to
troubleshoot,
Patient: Power One multiple output 400 W switching power supply.
Symptoms: Totally dead. No schematics available.
Testing: Applying power produces no output. Only observation is that
lights flicker indicating the input filter capacitors are
charging so this eliminates a blown line fuse as a possibility.
Unfortunately, this unit is not what I would call a 'simple switcher'.
In addition to the main switchmode transistor, there are 2 other power
transistors, a uA723 IC regulator, countless discrete transistors, resistors,
capacitors, some components I cannot identify. This is all on the primary
side of the transformer. Tracing the circuit is virtually out of the
question due to its complexity. The only good news is that I have
several identical units so I can compare readings between the bad one and a
working unit. Powering with a Variac produces similar lack of any output.
This is not a case of the outputs shutting down - there is simply no startup.
First check: power to main filter caps, continuity of thermal protector,
power to uA723, power to switchmode power transistor. These all are fine.
Next, checked components around input including power transistors, large
resistors (suspecting a startup problem), capacitors, etc. All ok.
Finally, about to give up, I decide to just test resistance across a more
or less random selection of components. Everything is identical until
I put my meter across a 6.8 M ohm resistor. On the good unit, this measures
above 1 M ohm. On the bad unit, it measures about 40 K ohms. I
unsolder components around this resistor until I located a 2N4124 transistor
that makes a difference when removed. Testing on the x1 range of my VOM
it tests fine but on the x1K scale, there is significant leakage in the
reverse biased junctions. Comparing with a nearby 2N4124, this is definitely
not the normal characteristics of a 2N4124. The 2N4124 is a general purpose
transistor so I replaced it with a handy 2N3904.
Powering with a Variac, the supply now comes up fine. I have no idea
of the function of the bad transistor.
Comments: This might be called a blind repair. Like bad connections, the failure
mechanism and function of the bad part will probably never be known. It
is not known whether the transistor was marginal to begin with and its
characteristics just drifted over time, or whether it went bad.
The basic assumptions which permit this technique to work at all
are that for a sudden change in behavior in a system with mostly discrete
parts, one of these parts has changed its resistance enough that an ohmmeter
check has a chance to find it and that the circuitry is interconnected
enough that checking a relatively small subset of node combinations has
a good chance of locating the bad part.
Patient: Goldstar model CR820U 19" color TV. Paid $1 at garage sale. If
I can repair it cheaply, it will be given to someone in need free
or for a nominal charge.
Symptoms: Basically operational except absolutely no color - not even any
color noise between channels. Color and Tint knobs have no
effect. In addition, brightness control has very limited range -
brightness slightly too high. Several other minor problems
including cracked tuner knob, dirty tuner, broken antenna wires.
Testing: This involved tuning local strong channels, adjusting fine tuning,
tweaking AGC, etc. Under no circumstances was there any hint
of color in the picture or between channels.
Fortunately, such symptoms narrow down the possible area of investigation
to the chroma decoder circuitry. Searching the circuit board for a likely
subsystem, I found that the set uses a TA7608P chroma chip. Fortunately,
this chip is listed in my ECG Semiconductor Master Replacement Guide with
a cross to ECG1532. Naturally, I suspect the IC at this point but know better
than to just go out and find a replacement - I have been burned in this way
with an RCA TV color problem - maybe I will say more on that in another
Repair Brief. While the ECG does have a pinout, this does not really provide
enough information to probe the circuit.
Off to the library to obtain the SAM's Photofact for this set. $0.75 poorer,
I copied the complete schematic and another interesting page - a chart showing
the resistances to ground for all pins on all of the intergrated circuits
used in the TV. One item I forgot to look for was the block diagram that
may have been included of the TA7608P chip. Oh well.
One thing I did try once armed with the SAM's was to attempt to tweak the
subbrightness control. Even this had very little effect. OK, on to the
fun stuff.
First test: confirm that the resistances of the circuit match those printed
on the resistance chart. The chart specifies a DMM that applies less than
.1 V on the ohms scales (to prevent forward biasing of any semiconductor
junctions). Hopefully, my DMM satisfies this requirement. First step:
make sure the main filter capacitor is discharged before making any resistance
measurements. Done. Unfortunately. these tests do not reveal anything amiss.
Tests of this type are not guaranteed to find any problems. However, there
is a fair chance that a shorted or open part would show up as a bad reading
in the near that part. In my case, if there were any bad parts,
the circuit topology prevented this simple resistance test from detecting
them.
Now for the live tests. I don't have a color bar generator so I just
have to hope that a broadcast channel will provide a signal that is close
enough.
In the interest of safety, all the following tests are made with the set
powered off of my isolation transformer.
Voltage measurements were inconclusive. Although some where off by 20%,
this could be due to my non standard input signal. There did not appear
to be any particular pattern.
Next, I used a scope to look at the testpoints for which SAM's supplied
waveforms. Again, these will be different than the ones using a standard
colorbar signal but the overall appearance should enable me to determine
if a particular output is dead or the amplitude is way off.
All the waveforms looked reasonable except one - the output of what I
deduce to be a gated chroma amplifier. Its output is almost dead. The
color reference oscillator (3.58... MHz) looks fine as do the chroma input
and color burst gating pulses. All the supply voltages and decoupling
pins look fine as well. This doesn't look good for the chip. The signal
seems to be getting in but the chroma amp would seem to be dead. Fortunately,
I could not locate an inexpensive replacement from my usual sources. The
TA7608P is probably obsolete. Even the ECG1532 is not available from
MCM Electronics or Dalbani. I did not get to the point of trying ECG
directly. You will see why I say 'fortunately' in a moment.
When confronted with a situation of this type, I usually try some experiments.
What would happen if I apply the chroma signal directly to the point in the
circuit that is the output of that dead chroma amp? I take a 10K resistor
and jump the input of the chip to the chroma amp output (and input to
the chroma demodulator.) Now, I have colored stripes on the screen indicating
that the chroma demodulator circuit is probably functioning. (Without the
color burst to phase lock, I could not hope for anything more). So, it
still looks like that bad amplifier. Well, one last desperate effort....
I use my sam's patented magic spit. This has served in on numerous occasions
mostly when locating clock noise, marginal timing, or glitches in high
speed digital systems but hey, the world is really analog anyhow. I am
not joking. Those who have done serious debugging know exactly what I
am talking about. Moisten your finger. Run it up and down the pins on a
suspect device. If something changes you have either (1) found a
particularly critical or high impedance but normally behaving circuit
(for example the frequency determining LC network of an oscillator) or
(2) something that is open or on the edge.
Magic spit to the rescue: running my finger (one hand in my pocket, isolation
transformer, etc.) over the two rows of pins on the chroma chip proved to
yield immediate results. Around the low number pins, the color suddenly
appeared grossly overloaded but with some indication of correlation with
the picture. This is the first time I have seen any indication of picture
related color. Hum.... OK, now to narrow it down. I take an insulated wire,
and strip both ends. Holding one bare end with my fingers, I touch each
of the pins in turn of the chip. Touching pin 6 has the most dramatic
effect producing the distorted colors. What is connected to pin 6? According
to the schematic, it is a 2.2 M ohm resistor to pin 3 and a decoupling
capacitor to the power supply. It is unlikely that the capacitor would
fail in such a way as to cause this behavior. So it must be the resistor.
Rummaging around in my resistor cabinet I come up with a 2.2 M ohm resistor.
and tack it across pins 3 and 6 (with power off!).
Now I get distorted color but this looks a lot more like a color TV than
what I had before. Time for a better antenna. That helped a little but
the picture is way too dark. Well, I did fiddle with the sub-brightness
control back when it had minimal effect. Locating the sub-brightness
control again, it now functions as expected having more than a sufficient
range. When adjusted to produce a picture of normal brightness (with the
user control mid-range), the color appears normal. Some additional fiddling
with the Color and Tint controls yields a fine looking color picture.
Removing and testing the original 2.2 M ohm resistor confirms an open
circuit. Soldering in my replacement completes the repair.
Now to clean the tuner, wipe down the case, repair the knob and antenna....
Total cost $1.77 including original purchase price (I threw in a hypothetical
$.02 for the resistor).
As I am writing this, I am watching the Goldstar TV - my trusty RCA
has just died after 14 years with no picture (raster and sound ok - that
will be another Repair Brief).
Comments: It is always tempting to suspect the expensive or unavailable
part first. Very often, as in this case, this proves to be erroneous.
Had the TA7608P been readily available at reasonable cost, I would have
probably replaced it only to find no change in behavior. This would,
however, have saved time.
Be warned that the 'sam's magic spit' approach must be used with caution.
You must understand the safety implications of touching *any* live circuit
especially with moistened fingers. I use an isolation transformer for
debugging. However, even with this precaution, I would think twice
before doing this on a live chassis (the Goldstar signal circuits are
isolated from the power line).
Patient: Panasonic PV3720 3 head VCR.
Symptoms: Band of what looked like tracking noise would come and go
depending on tape being played, speed of tape, whether at
start or end, etc. The noise was confined to the top 1/3
of the picture. Its height could vary from just a couple of
video lines to a band occupying 20 % of the screen.
Testing: Several tapes were played initially. Problem would be nearly
absent with some but severe on others. It was generally worse
with EP recorded tapes compared to SP tapes.
I generally do not like problems of this type because one of the more
likely possibilities is of a worn video head. This is one of the classic
symptoms yet it could have a number of other causes. The approach must
be to eliminate as best as possible the alternative causes until the risk
of purchasing a new video is minimized.
Alternate #1: dirty heads. Head cleaning with a wet cleaning tape
followed by a manual cleaning had little effect.
Alternate #2: tape path alignment. Visual inspection of the tape
movement showed nothing out of the ordinary. Tape motion was very
smooth and uniform with no wiggling, wavering, or wondering. All tape
guides were properly positioned, perfectly vertical (where appropriate)
and the tape appeared to be riding at the correct height on the video head
cylinder.
Alternate #3: backtension. Insufficient backtension could result in
similar problems. Inspection seemed to indicate that backtension was
normal. Manually increasing backtension by gently pressing the backtension
level to the left made a slight improvement. Increasing the spring tension
did the same. However, these were not dramatic effects and backtension
is not a critical setting to obtain a clean picture (though it is important
to be accurate to minimize head wear and clogging).
Alternate #4: roller guide height. Although visual inspection of the
of the tape path alignment proved negative I decided to confirm roller
guide height by careful adjustment of the supply side roller guide -
carefully noting its original position. (Problems at the top of the
picture would be related to the supply side roller guide.) Optimal
position for both EP and SP was at the original setting.
This left the video head as a likely candidate and at this point based
on the age of the machine, a new video head cylinder (MCM Electronics,
ERH433) was ordered and installed. Success! There was no doubt about
the improvement. The noise bards completely disappeared and the normal
backtension provided more then adequate head-tape contact.
Comments: Subtle problems that eventually point to the video heads are
among the more difficult to diagnose with enough confidence to risk ordering
an expensive video head and find out that the problem was elsewhere.
This was a case of video head wear (as opposed to a mechanical or
electronic failure of the heads). The chance of having an identical
video head available to swap - which is the best test - is quite small,
especially for a 3 head type. If this were a 4 head machine, some meaningful
comparisons could be made during playback since a different set of
heads is often used depending on tape speed and mode.
Thus, unless there is visible damage to the video heads or something
like an open winding that could be revealed by simple testing, it
comes down to eliminating as best as possible the alternatives until
only the head remains a likely possibility.
Patient: Classic ATT Touch Tone Wall Phone. 'Battlewagon' usually considered
indestructible.
Symptoms: Everything worked fine except dialing. For some buttons, dial
tone would not go away. For others, tones would be accepted but
would be erratic and incorrect digits. Certain tones sounded
weak or single frequency.
Testing: All buttons were tested. It was found that the problem was not
even consistent as some buttons would not work all the time.
While the internal wiring of one of these old phones is intimidating,
the basic tone dialing circuitry is an amazing example of simplicity.
About the only things that fail yet still permit some tone generation
are the pot core coils that determine tone frequency. Therefore, this is
the first thing to check.
Sure enough, the core that deals with rows has split where the two halfs
are joined. This seems to be a common problem due to both the age and brittle
cement used on some revs of this model phone, and probably, as a result
of rough treatment when hanging up the handset.
These cores must be aligned before being glued back together. In addition,
there is an adjustment plug which may need to be tweaked. I align by ear as
follows: Put a known good tone dialing phone and the bad phone on the same
phone line. Depending on which core is bad, depress either an entire (same)
row or column of buttons on both phones. (Adhesive tape is handy to
hold down the buttons unless you have four hands.) By depressing the entire
set of buttons, you are disabling the other tone generator so you hear a pure
tone. Without turning the fine adjustment plug (assuming it was not disturbed;
if it was, set it mid-range or the same as the one in the other core), rotate
the loose core top until a zero beat is obtained. As your rotate the core, you
will hear the tone change. As it approaches the correct setting, you will
hear the tones beat against each other. When you are set correctly, the
pitches will be equal and the beat frequency will go to zero. Mark the
position of the core with a pen or pencil and then glue with Epoxy or
other general purpose adhesive (around the outside - not on the mating
surfaces as this will affect the tone frequencies). After the glue sets,
confirm and adjust the plug core if needed. These cores use a strange
triangular core tool - I made mine by filing down an aluminum roofing nail
(do not use a ferrous material).
Comments: Those classic ATT touch tone phones are virtually indestructible.
However, broken cores (or actually, just broken joints on the cores) are
common but easily repaired once you know what to look for. Setting the tones
by referencing a known good phone seems to be a very reliable technique
as the zero beat permits an adjustment to better than .1%. Note that if
the reference phone is a more modern (and flimsy digital one), then pushing
multiple buttons may not work as it does with the old analog models. Setting
the frequency using the normal dual tones will work - it is just not as easy..
Patient: Pioneer PD5100 single disc CD player in really sad shape. Garage
sale acquisition paid $2. Probably paid too much. This episode
is more aptly called a restoration rather than a repair. Perhaps
a feature length saga. The challenge was irresistible.
Symptoms: Tray wasn't even on track, just sitting inside; Flapper ripped
off of mountings, electronic condition unknown. It is obvious
that the owner had attempted something - it would be generous
to call it a repair - and was unable or did not bother to get it
back together.
Testing: Not applicable at this point. With loose parts removed, power
was applied to determine if there was any hope at all. At least
the front panel came alive and pressing Eject resulted in the
tray loading motor spinning.
In order to attempt to play a disc, the controller needs to think that the
tray is closed. It will then go through its startup cycle. In the case
of this player, there is a limit switch - somewhere. Rather than trying
to locate it, I decide to put the tray back on its tracks. This is
easy but there is still something wrong as it jams when the Eject button
is pressed. So be it, leave that for later. At least the limit switch
will be activated. Rummaging around in the pile of lonely parts removed
from the carcass, I locate the clamper cover with the magnet. I pop in
a garbage CD, put the clamper cover on top (make a mental not to press
Eject under any circumstances as the tray, disc, cover, and anything else that
is not screwed down would probably fly across the room) and press the
power button. Some success - the disc spins and the directory is correctly
displayed. The display came up rather quickly indicating that most of the
optics and servos cannot be far out of alignment. This is quite remarkable!
With mounting anticipation, I connect the audio outputs to my amplifier
and press play. The disc spins and makes repeated attempts to start
playing at track 1 but it is obvious that something is terribly wrong.
Attempting to play other tracks results in similar behavior.
The pickup appears to actually move to the general vicinity of the
correct track but is unable to locate and lock onto the time/track that
is selected. Pioneer CD players perform a very audible search to home in
on the correct disc location; there was no evidence of this search.
I next attempt careful adjustment of the servo controls. Note that I do
not expect this to help the problem based on how quickly the directory
was displayed. However, the tracking could still be off and with care,
there should be little risk of making things worse. Who knows what controls
the owner touched in a misguided hope of performing a miracle. First, I marked
the *exact* position of each control with a felt tip pen. This will get me
back to the supposedly good positions no matter what. The only controls
that would likely have an effect are those related to tracking. Careful
tweaking of tracking balance, tracking offset, and tracking gain have
no detectable effect. I put them back in their original position and
verify that the player still recognizes the disc. So far so good?
At least the moron who butchered this thing does not seem to have
touched the electronic adjustments.
At this point what do we know? Well, we know that all of the major
components of the optical deck work including the laser, photodiode
array, fine focus and tracking voice coil actuators, and spindle motor.
These are all needed to read the disc index. The spindle motor, a
common problem in Pioneer CD players is fine as its toughest task is
at disc startup where the speed is greatest. Since the disc index is
located at the very inner extent of the disc, we do not know if the
sled servo (coarse tracking) is working correctly, only that it is
doing something - it resets to the inner track if manually moved away
and it does move to the approximate position of the selected track.
Well, Pioneer CDs have a TEST mode. Where is the button? I hunt all over
for the little button and am about to give up when there it is! Hidden by
the cables to the front panel.
OK, press TEST while switching on power. Now I have control of the
servos. A little experimentation confirms that focus and spindle rotation
seem to be functional: (of course, we knew that, right?) With no disc in
place, the focus search routine is initiated by pressing TRACK FWD. The
disc will only spin if focus lock is achieved and this is confirmed with
a disc in place. So far nothing new. I am able to move the pickup
back and forth on its tracks by pressing SEARCH FWD and REV.
However, when entering the correct sequence to play at an arbitrary
point on the disc, weird things happen. If I use the SEARCH FWD and REV
buttons to move the pickup to a particular spot on the disc, press
TRACK FWD to close the focus servo, PLAY to start spindle rotation, and
then PAUSE to actually start playing, the track and time info is only
displayed for an instant. Then, the pickup seems to move toward and bump
against the inner limit. Sometimes, a couple of times are displayed in
rapid succession which are not sequential as they should be. In fact, they
nearly always are far apart and the second is usually a lower time than
the first. Then the display is blank.
Hum, I don't have a schematic so this could be the end of the line. But,
I do see one chip on the circuit board that is getting unusually hot and
I know from past experience that it is a servo driver - TA8410K. I have
absolutely no idea if it is related in any way to the problem or really,
for that matter, what the problem is. I only know that (1) it has only
10 pins and is easy to replace, (2) I have a replacement in my parts
box, and (3) it is getting hot (which may or may not be a fault since
I know these type of chips to run at least warm).
Getting to the bottom of the circuit board proves to be a bit harder
than anticipated requiring removal of most of the snap type connectors.
I guess these are cheaper than real connectors for Pioneer but a pain
for servicing (cables are terminated in tinned wires and placed in the
connector housing, then a cover is pressed down to lock them in place).
I manage to only mangle one of these (cosmetic damage only).
Replacement goes smoothly. Getting all the connectors back in place
is loads of fun but the effort is worthwhile! Now, the disc plays
on the first attempt. There are still some tracking problems but
this is a distinct improvement. In all honesty, I am not sure that the
chip made the difference - it could have been a bad connections at one
of the connectors. The new chip runs warm, perhaps not quite as hot
as the old one, I am really not positive. I put a heatsink on it in any
case (as I always do with these chips - just for insurance.
Next I tackle the mechanical restoration. First step: get the tray
to move smoothly. Without going into terrible detail, the tray consists
of two parts whose relative motion raises and lowers the disc. There
appears to be something missing which controls when this raising and lowering
takes place as the disc is lowered even before the tray moves into the
machine. Sometimes there is a ball that controls this and a little
examination reveals a grease trail where such a ball could have been.
A corresponding hole in the tray bottom confirms this. I didn't notice
any such ball in the parts pile but it could have easily been lost (I later
found it near a corner of my workbench) but for now, I located a similar
sized steel ball in my steel ball collection. With the ball in place,
the tray now moves smoothly in and out and the disc is raised and lowered
at the proper time.
Now for the clamper.
This is a much sorrier affair as the clamper is mounted to the deck
sheetmetal with a couple of plastic standoffs that have been totally
snapped off at their bases. First I try simply glueing them but this
does not appear to be solid enough. In addition to the glue, I am able
to clamp one down with a metal scrap that I carefully shape and screw
down. For the other, I made a splint using a screw through a drilled
hole into a neighboring strut. Now the clamper moves up and down at the
proper time but the cover disk with the magnet seems to hit the tray.
The part that seems to help out has totally disappeared so I take a brass
rod and mount it in its place. Even without the rest of the mechanism,
this seems to work fine. This rod, wrapped with electrical tape to prevent
damage to the disc, prevents the disc from flopping around too much.
Disc loads; disc unloads; all is well.
I then went through the electrical servo adjustment procedure as outlined
in the CD Player Notes, final tweaking by maximizing the amplitude and
stability of the 'eye' pattern. I made the mistake of attempting to touch
the 'tangential adjustment' (at least that is what I think it is - without
a proper alignment disc, this appears to be very difficult) and spent
some frantic minutes until I was able to restore it to its original
position. Beethoven's Ninth Symphony comes in handy as it runs almost
to the edge of the disc (74 minutes) necessary to access the tangential
adjustment. I even risked careful adjustments of the LD - laser power just
to determine that it was not at the limit of its power. It was not.
I am fairly confident at this point that the adjustments are pretty much
where they should be - and they are very close to their original position.
Now the CD player works fairly well though it does not seem to have
as much disc defect tolerance as I would expect. I do not know if there is
still a fault either optical, mechanical, or electronic as all tests
that I can perform without service info seem satisfactory. Considering
what the player went through, this has still been a rewarding experience.
Comments: I consider this to be more of a learning experience than a repair.
At the outset, I did not expect to be able to get nearly as far as I did.
It was fun as such things go.
While I am in favor of home repairs, this is an example of a situation where
whoever attempted the repair of a problem due most likely to the bad servo
driver IC, totally destroyed any possibility of a professional even going
beyond looking at the unit and stating: "Yup, that was a CD player once upon
a time long long ago. To whom should I send the flowers?"
Patient: Yamaha R8 stereo receiver, about 4 years old at the time of the
failure.
Symptoms: Fm reception is totally dead. Station numbers change erratically,
not possible to save presets. Some AM stations work but most do not.
This happened without warning - turned it on one day and it was sick.
Testing: This involved methodically checking to see what functions are
operational. Incrementing and decrementing of FM station frequencies
is not operational in seek mode, only in manual. There is not
reception on any FM station frequencies. Incrementing or
decrementing AM station numbers across certain boundaries (I forget
the exact locations) causes a sudden jump of 800 KHz and may actually
jump to an illegal station frequency. Various other modes are non
functional including saving of memory presets. Even the hard reset
does not store the factory presets.
I purchased the service manual for this unit - a nice piece of documentation
and very reasonable - about $12. However, this is an example of modern
technology where even schematics, pin descriptions of the various LSI
chips, parts lists, etc. are not really adequate when so much depends
on firmware (in 3 microcontrollers) which is not provided. It turned out
to be difficult to even determine where each function is centered.
Some electrical tests that were performed:
Power supply voltages were verified.
Waveforms were checked on frequency synthesizer chip (LC7210).
Function of PLL charge pump was verified in both AM and FM. Output
(VCO control voltage) was consistent with frequency display when
reception was possible but not at other times. However, this could not
be a problem with the charge pump, only the digital control.
Intermediate 4 bit busses were checked for stuck-at faults - there were none.
The first real clue is that since even some manual tuning functions are
faulty, this is probably a digital fault. Presumably in manual, the
station display is driven by the microcontroller that drives the
synthesizer chip rather than being returned by that chip after a station
search. Even in this mode (for AM), there is the issue of the 800 KHz jump.
This is not approximate but exact and probably due to a stuck bit
representing the 800 place value. The question then became: where
was the bad bit? It is not on one of the intermediate busses as these
were tested.
Could it be in the tuning microcontroller? Maybe, but then I would expect
other functions controlled by this chip to be faulty (like mode setting,
etc.) This is not the case. Could it be in the frequency synthesizer
chip? Probably as only station tuning functions are defective. Could it
be elsewhere? There do not appear to be any other busses or digital
control lines that could cause the set of problems that are present.
However, not confident enough of the diagnosis of the faulty LC7210
synthesizer to spend the $25 or so that Yamaha would probably charge and
not finding this part in any of my normal mail order sources, I set the
receiver aside for a while. I dig out my garage sale NAD for use in
the meantime.
A couple of MCM catalog editions later - what's this? LC7210 - $6. I will
spring for that. Next MCM order arrives, solder in a socket as I always
do where possible. Replacement chip cures all problems!
With 20/20 hindsight, it is almost possible to identify the place
inside the LC7210 where the 800s bit bus fault occurs based on the
symptoms and the rudimentary block diagram provided in the service
manual.
Comments: Although not evident from the description above, this was a
frustrating experience even with the service manual because there really
was not enough information present to make the logical inferences needed
to come to a definitive conclusion as to the defective part. Modern
consumer electronics include more and more microcontrollers where the
intelligence is buried in firmware and not the hardware itself. Without
firmware listings, a microprocessor is just a black box even with pins
listings and internal block diagrams. It would be nice if the service
manual would at least provide better indications of which functions is located
where - identifying the functions of each of the components. (It would also
be nice if they were written or at least edited by Americans (in the case
of a manual destined for the U.S. market). Some of the translations are,
well, a bit strange.
Patient: Sylvania Color Television, about 20 years old.
Symptoms: Suddenly, the picture lost *all* horizontal hold. There was no
evidence of any kind of attempt at lock in. I do not know whether
this happened at power-on or while in use.
Testing: With strong signal, it was determined that horizontal hold had
no effect. It is as though the H sync is not making it to the
lock circuitry. Adjusting horizontal hold makes picture move
across screen. Angle of sides of picture changes but there is
no lock - even incorrect - at any setting.
Using my isolation transformer, I prepare to scope the relevant signals.
I obtain the SAM's for the set. I check for the sync signal at the input
to IC400 (I think). It is there. This should be a snap - bad IC! Well,
that is exactly what happens - a careless slip of the scope probe and not
only a snap, but a crackle and a pop - and now I have no video, no HV, no
deflection - nothing.
OK, so what started out as a simple signal problem is now a major (at least
cost and pride wise) power supply problem.
Checking the first TO3 transistor I can locate - short - one dead
transistor. This is the power supply series chopper.
Checking the horizontal output transistor (HOT) with an
ohmmeter - short - second dead transistor.
After removing transistor, I check for rectified line voltage at the
input to the chopper - nothing. Tracing this back I soon locate an
open fusable resistor.
So, whatever I touched probably caused the HOT to fail (forced on for
too great a time can blow the HOT as a single shot event). The shorted
HOT probably then took out the chopper transistor.
This is not fun. It is not likely to be inexpensive either. It does seem
that no other parts have been sacrificed. Fusable resistors and driver
transistors seem ok in so far as my meter is concerned. I still assume
that the original problem was caused by a faulty IC400 but this point it is
impossible to confirm this since the set id dead-dead.
Damage:
Chopper transistor - $10, horizontal output transistor - $6. IC400 - $15,
fusable resistor - $1.
After replacing the components, making sure to use mica insulators
and silicone heat sink compound for the transistors, the set comes
back alive. Sync is fine. A little touchup of the video background
and gain controls (unrelated to the sync problem) and we are done. Ouch.
Comments: The lessons learned here came at a cost - but mostly to my
pride. Cascade failures are all too easy to induce through carelessness.
Power supply circuits are not forgiving. One would think that probing
the sync signal would not be able to kill anything. However, the design
of power supply and deflection systems share some common characteristics.
One of these is that a single instance of an improper drive waveform
can blow the switching transistor as a single shot event - excessive
current or excessive flyback voltage. This is a matter of exceeding the safe
operating area of the transistor.
What you learn: if possible, make all connections to your test equipment
with power off. Insulate all but the last mm of your probe so that any
slip cannot cause a short. Work methodically, think things through, don't
be over-eager, don't take shortcuts.
Patient: Microwave oven (don't recall the brand). Symptoms: Totally dead - no front panel display or anything else. Testing: Plugged unit into live outlet confirms description of problem. First step: remove cover. Second step: confirm that HV capacitor is discharged. Although the unit has been unplugged for several days, it never hurts to be careful. Discharge with high value high wattage resistor (well insulated) and confirm with HV voltmeter. WARNING WARNING WARNING etc. Microwave ovens are probably the most dangerous piece of consumer electronic equipment in terms of potential for electrocution while being repaired. Much more so than TVs, for example. * There is up to 5,000 V negative with respect to the chassis while powered and stored in the capacitor when shut down. The energy stored in this capacitor is roughly ten times that stored in the CRT second anode of a TV or monitor. A malfunctioning oven that does not blow a fuse is potentially even more of a hazard since the capacitor cannot discharge into the magnetron load after shutdown. * The available current exceeds 1 A. Depending on the transformer, this may flow until the fuse blows or circuit breaker trips. * An isolation transformer (assuming you had one large enough) or GFCI will not protect against HV shock since the HV is already on the secondary side of the HV transformer and, as noted, the HV return is the chassis. * A circuit breaker or fuse is not sensitive enough and too slow to provide any protection for you. You might think at first that the possibility of microwave exposure is the principal danger. However, unless the door seal, oven chamber, or waveguide or its seals are damaged, there is no way for microwave radiation to escape. This still means that proper precautions should be taken - don't operate the oven for longer than needed with the cover off and don't stand too close. Third step: test fuse. Open. Since these have ceramic bodies, it is not easy to determine if the fuse died due to an overload or a short by visual examination. A microwave oven can blow a fuse for several possible reasons. Some of these are: 1. Shorted Magnetron. 2. Shorted capacitor. 3. Shorted rectifier. 4. Defective door interlock switches. 5. Bad Triac (possibly causes transformer core saturation due to operating as half wave rectifier or SCR). 6. Bad transformer (shorted turns cause overload). Fortunately, all but the last two are generally easy to identify using just an ohmmeter. Sometimes, an intermittent or short that occurs only at full voltage will prove elusive, however. Some quick checks reveal that the capacitor is a dead short. When replacing a microwave oven capacitor, it is important to get a fairly close match for the capacitance. The uF rating of the capacitor affects the microwave power output. Note that the 'working volts' rating on a microwave oven capacitor is not the same as on common capacitors found in other electronic equipment. It is not the maximum voltage permitted across the capacitor but closer to the VRMS rating of the HV transformer. And of course, before you start pulling wires off (1) mark down where they go and (2) discharge/check for voltage on the cap one more time. Replacing the capacitor with one from MCM Electronics brings the oven back to life. Comments: It is highly likely that the capacitor failed due to a defect in manufacture rather than some other underlying problem in the circuitry. When one thinks about how a capacitor is constructed - rolled up layers of foil and dielectric - it is amazing that capacitors do not fail that often. Any nick, thin spot, etc. represents a point of excess stress and can fail as in this case after considerable use - resulting in a short circuit, dead oven, and unhappy chef.
Patient: Tandy color TV/monitor found curbside. Unit appears to be in
good physical condition. Remote is missing. Well, you cannot
have everything.
Symptoms: When first turned on, TV appears to function normally. Why
was it tossed? Well, after 30 seconds or so, a pair of hum bars
begins to appear in the picture gradually getting worse until
horizontal width and sync are affected.
Testing: Using a Variac, there is a point below normal line voltage where
set operates perfectly. OK, so I will keep a Variac attached
to the unit!
After removing the cover, the first thing to suspect is the main filter
capacitor. If this should dry up and lose some of its value, these would
be the exact symptoms. Jumpering (with power off) of a known good capacitor
I keep for this purpose doesn't change anything. But what is this? A
discolored resistor catches my eye. Maybe it is changing value as it heats
and causing these symptoms. I wait a reasonable time for the set to cool
and measure the resistor - 360 ohms. OK, replace with new one. Expecting
this to cure the problem I am disappointed when there is absolutely no
change.
Off to the library for the SAMs Photofact. Darn - SAMs does not have a
service folder for this model. Nor for any similar models that I can
determine (the librarian was very cooperative).
Well, the problem seems to be heat related. I get out my trusty can
of cold spray. After going through nearly the entier can, it would seem
that there is only one part that has an effect on the hum bars when it is
chilled. It is the SCR that is part of the power supply regulator. Rather
than simply obtaining a replacement, I decide to trace the circuit to
determine, if possible, the possible cause of the problem figuring at this
point that the SCR is simply sensitive to heat.
During normal operation, an IC drives the gate of the SCR but what is this??
Until the secondary supplies kick in and provide power to the IC, the SCR
is driven by - you guessed it - the mysterious resistor. The other end of the
resistor goes to the raw DC on the main filter capacitor. Now that is
odd....Since I do not believe much in coincidences, I now start rethinking
the significance of this. Maybe that resistor is not quite what it appears
to be.
First, I remove it and see what happens: nothing. Power on, power off,
nothing.
Next, I momentarily touch the resistor to the circuit pad - the set comes
alive. Then I remove it. The set remains alive. And, after several
minutes, no hum bars. Hum....
I then try increasingly larger values of resistance until turn on is not
reliable - 15 K seems marginal, so I will go with 8.2 K ohms = that is
over 25 times what I measured! No wonder there were hum bars indicating
regulation problems - that low value resistor was totally overwhelming
the poor IC in driving the SCR.
The set is used daily and has been operating without further problems
for over 5 years since reviving it. It works great with a $10 universal
remote control.
How did the resistor get damaged in the first place? I have no idea - maybe
its wattage was slightly underrated and it just finally decided to poop out.
I have no way of knowing what the original value was supposed to be or even,
for that matter, the wattage.
Comments: resistors can and do change value, sometimes, as in this case,
quite dramatically. Without a schematic, there is no easy way to determine
when and if this has happened - and what the original value should have been.
However, any discoloration, burn, or scorch marks should arouse suspicion.
With 20/20 hindsight, these signs may indicate the presence of carbon -
a fairly low resistance substance and thus reduced resistance is likely.
And yet again, a $.02 part brings a complex renders a complex piece
of equipment inoperative.
Patient: Realistic portable CD player, about 4 years old.
Symptoms: Player will shut off at totally random times or sometimes will
not recognize the disc. There is a clicking associated with
the problem - probably focus search failing.
Testing: Attempting to play various CDs to completion provided no
indication that the particular CD or power adapter made the
slightest difference. There were some false leads with respect
to the latter but these turned out to be strictly coincidence.
The lens was inspected and cleaned anyhow with no change.
The first hint of the source of the fault came as a result of an
observation that pressing on the cover would sometimes either cause
the player to stop in the middle of a disc or allow it to recognize and
begin playing a disc when it would not otherwise cooperate.
Perhaps, the interlock switch was not being pressed in far enough. So,
rather than open the unit (I really don't like messing with the insides
of portable CDs if I can help it - you will see why in a few moments),
I glued a bit of plastic to the post that pokes the switch.
This seemed to help. For a few weeks, the problems had for the most
part gone away and the owner was a happy camper. Not surprisingly,
this fix was only temporary.
Since the quick fix had some effect, it is very likely that I am on
the right track. I will have to open it and deal with the switch
face-to-face.
This is not too bad except that it is necessary to remove the main
circuit board to access the switch which is mounted on a little board
of its own. Four screws (large enough to actually see without a
microscope) to get the bottom off, another couple to remove
the main board. One more and I can remove or at least extend the
switch circuit board far enough to inspect its solder connections and
get at the switch.
The solder doesn't look too bad but there might still be hairline
cracks that are not readily visible. A little reflow and they
should be fine. (Problems with solder joints here are not related to
heat as in a TV or monitor but rather due to the mechanical stress that
is applied to the switch every time the lid is closed.)
Now for the switch. It appears that the cover of the switch can be
snapped off relatively easily. The contacts appear somewhat gummy
so I clean these and pop the cover back on.
Tests with an ohmmeter now show the switch action to be solid. Wiggling
the switch lever and/or the entire switch has no effect.
Great, put it back together and I am done.
After replacing the switch board, main board, and bottom cover - the test.
Fanfare please!
Nothing. The player is dead as a door nail. It now will not even focus
and gives up almost immediately.
Off come the screws. Almost immediately, it is obvious what has happened.
In replacing the main board, I accidentally squashed one of the printed
cables linking the optical pickup and main board, partially severing
the cable. In fact, 2 of the 4 conductors are cut. This is the focus
and tracking drive cable so it is pretty important. What a pain!
Fortunately, luck is on my side with respect to the location of
the break - it is at a non-flexing part of the cable. Therefore,
repairing the cable should not be that difficult since once the conductors
are connected electrically, they can be coated with a sealer and flexing
will not be a problem.
To repair a cable of this type, I have two options: I can attempt to
jumper the break with some fine strands of wire or I can go point-to-point
from the circuit board to the destination on the optical pickup. However,
the latter connections are nearly hidden and would be difficult to solder.
I opt for the first. Using an Xacto knife, I carefully scrape the
orange mylar coating from both sides of the break. Then with #30
wire, I carefully solder across the break for each of the conductors.
A spring clothespin holds the wire in place during the soldering.
The entire affair is then coated with some clear sealer to reinforce
it mechanically and provide insulation. It isn't pretty, but it
will work fine. For added protection, I add a layer of plastic electrical
tape.
Now, finally, reassembling the unit keeping cable routing firmly in
mind, there should be no problem.
And, as expected, the player comes back to life and is rock solid with
respect to playing and recognizing discs. The oops should have no effect
on the expected longevity of the player.
Comments: we all can point to those minor disasters where we have overlooked
something where we should have been more careful. Whenever reassembling
anything, it is imperative that lead dress (ok, fancy term for how the
cables are routed) is kept firmly in the forefront of your mind. It seems
that with more and more miniaturization, this is an increasingly
important and at times, frustrating consideration. First of all,
it is very tempting to say when disassembling the unit 'this is obvious,
no need to write it down'. Bad move. Often, it appears much less
obvious when putting everything back in its place. I have never quite
figured out how they do it during manufacturing - correctly most of the time.
Ignoring cable routing can lead, as in this case, to severed wires. It can
also result in shorting between wires or between wires and sharp metal
brackets or shields. Broken wires can usually be repaired if they can
be located. Shorted signals can result in additional hard-to-locate
collateral damage which can really turn your hair gray.
What is even scarier is that with line connected electronics or appliances like
vacuum cleaners and even toasters - this can lead to electrically live
parts accessible to the user. Sometimes, the plastic insulation on typical
internal wiring will not fail immediately but will cold flow and cause
problems later. So, one should always make every effort to assure that
no wiring is being pinched and for metal cased appliances, check that the
case is not electrically live - has a high resistance (usually
infinite, but at least a few M ohms to both wires of the AC line
with any on/off switches in both positions) after the repair is completed.
For non-heating appliances or electronics, a little electrical tape goes
a long way. For heating appliances you really need to make sure that bare
wires are routed far from any exposed metal of the case taking into account
as well any motion that may occur during normal operation or due to being
knocked about or dropped.
Patient: Sony CDU33A CDROM Drive given to me supposedly brand new but broken.
Symptoms: Drive behaves the same as a similar working drive until it
is accessed. Then, there is no response by any DOS or
Windows software. No CDs are recognized, always get the message:
Abort, Fail, Retry?
Testing: I keep an old (well, what other type are there?) 286 PC clone
system around for the primary purpose of testing peripherals.
Installing the drive and software confirms the reported
behavior. I was given two similar drives. The other one was
reported as being intermittent but seems to work fine in my
test system. This one was indeed dead.
Since it is impossible to observe the behavior of the pickup and, in
particular, the lens with the cover on, the first step is to get at
the guts.
Fortunately, the CDU33A is quite simple to disassemble.
There are only two major components: the Printed Wiring Board (PWB) where all
the active electronics are located and the Optical Deck including laser,
optics, and pickup worm drive mechanism.
The other parts include the upper plastic casting and metal shroud,
solenoid latch assembly, right and left guide rails, drawer assembly,
and front bezel, two springs, bottom plate, 6 screws.
There are only two electrical connectors inside: one flat printed cable
linking the PWB and optical deck and a two pin connector supplying power
to the eject solenoid. This is in pleasant contrast to some other CDROM
drives I have seen with a half dozen or more small connectors spread
all over the PWB making removal and testing very difficult and risky.
After about 10 minutes, I have the drive apart and can now reassemble the
major components on the bench outside the case to observe behavior.
I prop up the circuit board and reconnect the flexible cable - noting the
orientation marks. I can now run the drive with full visibility of
the mechanism and optics. With a CD in place, there is no danger
from the laser beam. I make sure the PWB cannot short to anything and
that the whole affair cannot tip over.
Having set up this contraption (you would have to see it to appreciate
appreciate this terminology), I am ready to continue testing.
Naturally, it now works perfectly.
No amount of abuse seems to phase it - wiggling cables, flexing the
circuit board, trying multiple CDs, all fail to reproduce the
original problems. Could it be the case? I can think of no reason
why it should make a difference? Is there anything else different?
I don't think so. Perhaps the sled was jammed somehow and disassembling
the drive fixed it. Who knows.
After reassembly, the drive continues to function perfectly.
Comments: How many times has someone brought you a 'broken' device which
has magically started working again on your bench. It certainly cannot
hurt your reputation. Admittedly, here, I had to actually do an exploratory
before rejuvenation to convince it that I meant business.
It has now been almost a year and the drive continues to function. I can
only guess that the cable may have been poorly seated or had some
dirt stuck in the contacts. Until it fails again, there isn't much more
to try. Unfortunately, the saying: "if it ain't broke, don't fix it" now
applies. I have no idea if the drive will ever again fail within its normal
life expectancy, but in the meantime, where did I put my Win95 CD?
(No comments, please, about choice of OS).
Patient: Late model Panasonic 17" color TV
Symptoms: Width slightly reduced. Slight evidence of 60 Hz hum bar,
brightness pulsating, raster shaking, somewhat channel dependent.
Testing: All of these symptoms were easily reproduced on the bench. The
60 Hz hum bar is the giveaway indicating a low voltage power
supply problem.
Rather than operating the TV off a Variac to confirm lack of regulation,
I decide to just try the most likely solution - a replacement main filter
capacitor. With power off and making sure the main filter capacitor is
discharged, I use a pair of clip leads to jumper my test cap across its
terminals.
The set now works perfectly.
Removing the old capacitor (not easy as the rivlets really do make nice
heat sinks), testing with my trusty Radio Shack DMM on its capacitance scale
reveals that the value has dropped by over 85% - pretty amazing that the
set worked at all!
One highly overpriced replacement filter capacitor (I used a local
distributor instead of my favorite mail order sources) and the deed is done.
No disasters on this one!
Comments: This capacitor was mounted next to a large heat sink - possibly
the power regulator. When replacing electrolytics, we often ignore one
very important specification - the temperature rating. Either the original
capacitor was defective or it was not rated for the thermal conditions inside
a compact TV. The TV was not that old - maybe 3 or 4 years at most. We all
can point to equipment we own that is still working after 20 or 30 years
going strong on the original filter capacitors.
Patient: Zenith 19" color TV, about 5 years old. This TV is owned by
a neighbor of mine who has small kids - more on this later.
Symptoms: Dead as a door nail. Only evidence that it is connected to
the line is a momentary flicker of lights when TV is turned on
indicating that the main filter capacitor is being charged.
Testing: This set has a pull-type on-off switch. There were no blown
fuses. Checking with a voltmeter shows 150 V on the main
filter capacitor with the switch in the on position. Ditto
for the collector of the HOT.
This would seem to indicate that there is a problem with the startup
drive to the Horizontal Output Transistor (HOT).
Off to the library for the SAMs....
Many Zenith TVs use a simple multivibrator to generate a startup
signal to the horizontal driver transistor until the flyback
can generate the secondary voltages needed to operate the deflection
ICs. Once these voltages are present, the startup circuit is disabled.
Indeed, such a design is used for this TV.
Checking with a scope (powering the TV through my isolation transformer)
at the base of the HOT shows no drive signal.
Tracing back, there is no signal at the driver transistor or from the
output of the startup circuit. One of the two transistors in the startup
multivibrator is bad.
I do not have a suitable replacement - it is a high voltage low current
Zenith part similar to an MPSA43 - 200V. I will need to obtain one, or
better yet, two to replace both transistors in the multivibrator.
To confirm that the rest of the TV is operational, I use a common
technique to 'jump start' a TV where the startup circuit is defective.
This is to inject a signal of around 15-16 KHz directly into the
base of the HOT to substitute for the startup circuit.
With the TV turned on, momentarily touching the output of a pulse generator
set for 15 KHz and a couple of volts amplitude to the HOT base brings the TV
to life. Everything appears normal except that the TV does not start on its
own. Somehow, I don't think my neighbor would approve of this solution.
(Also, I am not giving up my pulse generator!).
Caution: jump starting a TV like this is risky. In addition to the dangers of
mucking with a live TV, injecting a signal with improper characteristics
into the HOT can destroy it and possible a lot of other circuitry - instantly.
For example, a single cycle with too long an ON time can blow the HOT from
overcurrent while driven on or overvoltage during flyback. Use this approach
with care.
Replacement of the multivibrator transistors with the exact Zenith parts
completes the repair successfully.
Comments: Examining the schematic of the startup circuit reveals that it
appears to be designed to fail - especially with kids about. While
the transistors are rated at 200 V (they are running on the 150 B+
from the line power supply), the transistor power rating is only .6 W.
Even though they are running in a switching mode, I believe that
repeated on/off cycles can stress these to the breaking point. Something
was mentioned about my neighbor's kids turning the TV on and off repeatedly.
I have not duplicated this experiment but suspect that such treatment
at least may contribute to premature failure. Fortunately, in this case, it
was only in the startup circuit.
Power-on is a stressful time for many types of equipment due to inrush
current, transient voltage, so many things changing quickly, etc.
In addition, designers may not study and characterize the behavior
during startup with the same amount of care that they presumably
(we hope) do for steady-state operation.
Patient: Samsung VCR, Model VR 2610.
Symptoms: Inserting a tape works fine - it plays, it records, it FFs, it
REWs. However, attempting to eject a cassette results in an
infinite loop - the VCR grabs the tape back just before it
pokes out of the slot. Sometimes, the tape can be grabbed in
time but usually the cassette does not exit far enough.
Testing: Symptoms confirmed. With the top off, it is easy to catch the
tape but I don't suppose this would be an acceptable solution.
In addition, the cassette carriages seems king of sloppy - loose
for want of a better term. This would indicate a mechanical
problem with the cassette basket - the mechanism which moves
the cassette into position inside the VCR.
First step: a close examination of the basket mechanism. Nothing obvious - no
broken parts visible.
Next step: attempt to remove the basket. With most VCRs, this is a simple
matter of 4 screws and perhaps a connector. Not here. There are 4 screws,
but once the screws are removed, only one side wants to come loose. The
left side, with most of the gears and whatsits, is firmly fixed to the
base of the tape deck. No doubt, there are critical timing relationships
that might be disturbed once removed. It stays for now.
Perhaps, removing the bottom cover will reveal something. 8 screws later,
bottom cover off. What's this? A spring!! So now, we know that something
is indeed broken and most likely in the basket somewhere. This sort of spring
is not the type to have just popped off - it is a close wound coil spring
with hooks at each end. And, guess what, there is also a tiny bit of white
nylon which was probably the tab onto which the spring was hooked at one end.
A close examination of the visible portions of the basket above and below
deck finally turns up something now that I know generally what I am looking
for. Thankfully, it is accessible and I hopefully don't need to pursue
removing the basket which almost certainly would not be a fun thing to do.
The spring is supposed to be connecting two gear-type wheels in the EJECT
mechanism. With the spring sprung, these were free to rotate when they should
not have and their free play was sufficient to cause the EJECT operation
to screw up.
So, how to repair? There is no good way to glue nylon and even if there
was, the tab is so small that it would be impossible to provide a strong
enough bond to withstand the spring force. Replacement of the part with
the broken tab is a possibility though again not a pleasent one - it would
require removing the basket. Of course, replacing the entire basket is
another unpleasent and expensive options. Installing a metal post in place
of the tab is also a possibility - one that I do not really want to
contemplate.
Well, it appears as though there is nothing particularly critical about
the spring placement. Is there an alternative location to connect the
end with the broken tab? Yes, it would appear that it will be sufficient
to hook it around another large wire spring. However, then it is probably
stretched too long, so I make a link out of a piece of a paper clip and
this seems to be about right. (Paper clips, bailing wire, scotch tape
and chewing gum (well maybe not chewing gum) are among my favorite things).
Getting all this in place under spring tension between the edge of the case
and the basket plastic frame proves a bit of challenge - requring a dental
picks, needlenose pliers, patience, and few carefully chosen four letter
words - but I prevail. The EJECT operation now works perfectly.
While not pretty, I believe the newly designed spring attachment will be
much more robust than the original. I should write to Samsung!
Comments: This is a another case of poor design - there can be no other way
of describing it. The spring is rather large (you can visualize it, can't
you?) and the tab much too small. Another .0001 cent of plastic and it
would outlast the rest of the VCR. There was absolutely no excuse as
there is plenty of space to enlarge and reinforce the tab.
Patient: Nintendo original game system.
Symptoms: Power light blinks indicating that it is not able to run the
program contained in the game cartridge.
Testing: Tried multiple cartridges without success.
The most common problem with these units is a worn or dirty system unit
game cartridge connector. In this case, the red power/status light will
continue to flash even after the RESET button is pressed with a
game cartridge in place. Replacements are available for about $9
from the usual sources (MCM Electronics, etc.)
First, I try another game cartridge - the one that is not working may
just have dirty contacts or may be defective. This does not work.
So I need to get inside. Fortunately, unlike some other consumer stuff,
this is quite easy. Six screws underneath followed by about a dozen
to remove the metal shield and circuit board so the connector can be removed
and inspected.
Before removing the connector from the circuit board edge, I give the system
another chance to redeem itself. With the latching mechanism removed, it is
possible to press the cartridge down somewhat lower than normal increasing
the chances for good contact. Indeed when this is done, it is possible
to occasionally get a good reset and game startup on the TV. This
certainly confirms the original suspicion.
Now, can I revive the original connector or must it be replaced? There are
three kinds of problems that generally occur with these connectors:
1. Wear of the contacts to the game cartridge. Although full pressure
is not applied until the cartridge is latched, there is still wear
every time a game is inserted or removed. This take its toll.
It is often possible to use a dental pick or a bent paper clip (one
of my favorite tools!) to slightly spread the spring contacts so that
they grip the edge contacts of the game cartridge more tightly. On a
high mileage unit, however, they may be worn through to the point of
actually breaking in half resulting in replacement as the only option.
2. Corrosion at the connector to the circuit board. Cleaning with
a pencil eraser or at most some very fine sandpaper (600-1000 grit)
will usually restore to as-new condition unless some really corrosive
agent was at involved.
3. Kid grime on game cartridge edge connector transferred to Nintendo
cartridge connector. Favorites: Coke, sugar candy, and ice cream.
So, the first step is cleaning of both sets of connector contacts and
the main circuit board edge finger. You may need to use a variety of
solvents to completely remove all crud. Water may work better on sugar
syrup than normal contact cleaner or alcohol. For the edge finger, a
pencil eraser very useful.
Don't neglect the game cartridge connectors. These generally do not wear
but may collect all kinds of strange stuff. Rather than fight with the
security screws that you may find holding the case together, I usually
simply use a Qtip with water, contact cleaner, or alcohol - or one after
the other - to clean these contacts. Again, very fine sandpaper may be
needed in extreme cases.
Even if these procedures only make a slight improvement - you can press down
on the cartridge and the machine will respond to the RESET button - you have
confirmed that the connector is indeed the problem. In many cases, just
cleaning will result in reliable operation for a long time to come.
In the case of this particular system, all three problems were present.
However, for the time being at least, the system has responded well to
treatment.
Comments: While the original Nintendo game machine is a couple of generations
out of date, many are still in use. And, hey, young kids usually don't care.
OK, you don't have to admit to being the one who cannot resist just a couple
rounds of 'Super Mario III'!
Old Nintendos can usually be picked up for $5-20 at garage sales sometimes
complete with a selection of games, sometimes bare. The games go for
$1-$5 depending on the barganing skills of the kid selling the stuff.
However, a bad connector is almost a sure bet with a secondhand system.
Consider that most electronic connectors are typically rated in terms
of hundreds of insertion/removal cycles. A Nintendo machine must endure
thousands of not necessarily gentle cycles over its lifetime. The connector
was not designed for that. Furthermore, you are likely to find all kinds
of muck inside, mostly unidentified, and often difficult to remove.
Nonetheless, these things are remarkably robust, electronic failures
are infrequent, and they can usually be revived without much difficulty.
Patient: Sharp Model VC7864U VCR in generally good condition, garage
sale acquisition, $1.
Symptoms: Former owner complained about difficulty in ejecting.
Testing: Tried playing multiple cassettes (not all at once!). For the
most part, the VCR behaved normally. Maybe just a bit sluggish
loading but no other obvious problems. Why did he dump it?
I did my usual cleaning - rubber parts did not look to bad, leave them
for now. Even the idler tire appears to be in decent condition. I will
use the VCR and see if any problems appear.
The first sign of trouble appears once when attempting to use REVIEW
mode - the VCR abruptly stopped and attempted to unload the tape. The
loading motor was spinning but nothing was happening (I think it was turning
in the wrong direction and the belt was slipping - I am not sure). Oh boy,
time to leave the cover off. Manually giving the motor shaft (fortunately
it is accessible from above the deck) a couple of turns convinces the VCR
to complete a correct unload cycle.
Well, this sounds like the classic 'if it is an erratic Sharp VCR, the mode
switch must be dirty or bad' problem.
(2 years pass as I am in no mood to bother with this repair at the moment.)
OK, now I have a need for a reasonably decent VCR to replace my cousin's
Mitsubishi HS328U which is finally dying. So, I dig the Sharp out of the
closet and see about its condition. Now, it doesn't even want to play a
tape at all. Well, I know I have to deal with the mode switch, so first
things first.
The mode switch on this model is sandwiched between the loading gears
and a mounting plate - all parts of what I will call the 'loading gear
assembly'. To access the mode switch, this entire unit needs to be
removed and partially disassembled. The gears operate the roller guide
loading mechanism, and a couple of cam operated levers which are conveniently
hidden when it is removed or reinstalled. It is driven by the loading
motor via a couple of idler gears.
Timing marks: In the unloaded position, there is a hole in one gear
that appears to line up with a slot. So, with the roller guides retracted
(and the gears which operate this linkage have timing marks which also
line up), this hole should be centered in the slot. Fine. This appears
to be the only critical relationship with respect to removing the loading
gear assembly.
I unsolder the 4 connections to the mode switch, remove 3 screws, and -
sproing! What was that? OK, one or both cams still had a lever with
spring pressure applied. Hopefully, it will be possible to extend these
these when replacement time comes along.
With the loading gear assembly removed, it is still not possible to access
the mode switch. Now to disassemble it. There are two fancy cam gears which
obviously must be timed correctly - in one position there appear to be
an arrow and triangular hole which line up. I add a couple of marks of
my own for good measure with a felt tip pen. A simple split washer
holds the gear I need to remove onto its shaft. (Note: these split
washers are not designed to be reused but with care in removal, they
can usually be replaced without any long term problems. Of course,
a professional would have an assortment of replacement sizes handy.)
Removing the gear carefully, there don't appear to be any flat washers
or spacers to worry about.
Once the gear is removed - making a note as to which side is up though
this is pretty obvious - the mode switch is exposed. Squeezing the
center of the split shaft enables the cover to be popped off and the
interior appears. I almost lost the springy wiper as it is not fixed to the
plastic cover but popped free when first removed. A frantic search was
needed to locate it on the floor. The wiper fingers and encoder contact
traces seem to be in good condition but whatever was used as a lubricant
is a little gummy and might be the problem. A simple cleaning seems to
take care of that. I also bend the wiper fingers a bit to increase the
contact force very slightly.
Now, to get everything back together. First, the wiper is replaced and
the mode switch cover is snapped back in place. Free rotation is confirmed.
Then, the gear that was removed is returned to its shaft along with a
cam follower lever that was under it. The split washer is replaced.
To install the entire loading gear assembly means that the original
gear timing relationships must be re-established. In addition, care must
be taken to make sure those two cam follower levers I mentioned previously
are properly positioned. This takes a bit of work but eventually, I
am convinced that everything works as it should. The screws are tightened
and then free movement of all the parts is confirmed by manually cycling the
loading mechanism. The 4 mode switch connections are then resoldered.
Now for the test. Since this was not a hard failure to begin with, there
is no guarantee that any problems will be detected.
The tape seems to load correctly but then the VCR unloads and shuts down.
What is wrong? It would appear that the takeup reel is not turning. Hum,
probably that rubber wasn't as great as I had assumed a couple of years back.
I now do a more complete cleaning and, in particular, remove the idler tire
and inspect it. It appears to be ok but as a test, I turn it inside-out.
Now, everything works as expected. Testing with a cassette cheater (shell),
there appears to be adequate takeup torque. I clean the idler tire again and
reinstall it in the normal configuration. All modes appear functional even
when testing with a full takeup reel - requiring the most takeup torque.
I will order new rubber anyhow and replace it at a convenient time or if
problems reappear.
This VCR now appears to operate reliably and consistently. I have seen
no evidence of the original erratic behavior. Only time will tell for sure.
Comments: I cannot overemphasize the importance of making careful notes as
well as adding timing marks of your own when removing any parts of a VCR which
could conceivably have critical timing relationships. Not doing this can
really mess up your day. Err on the side of excess - it won't cost you
anything.
Sharp VCRs seem to be particularly prone to mode switch problems: Of the
3 Sharp VCRs under my control, 3 of them have developed dirty mode switches
resulting in a variety of erratic symptoms including, as noted, going into
the wrong mode as well as aborting the tape loading operation for no good
reason.
It would seem that a VCR design using an optical mode switch instead of one
with sliding contacts would be much more reliable at only modest additional
cost. After all, VCRs already use a number of optical sensors and cheap
computer mice use optical encoders not very different in design from a mode
switch. At least, it would be nice if mode switches were readily accessible.
Some are visible as soon as the bottom cover is removed. Others require
substantial disassembly with associated risks of incorrect reassembly resulting
in mechanical timing problems or even damage when the unit is cycled.
Patient: Approximately 10 year old BIG Magnavox color console (when
furniture meant something).
Symptoms: Horizontal deflection jittery, possible vertical collapse, arcing
flyback - all in one set! This info from Dave whose friend
owns the set. Dave is a tech at work who is now doing more
software than hardware (not necessarily by choice).
Testing: I did not actually see the original problems, nor did I have
access to the entire set as Dave came in one morning with the
guts of this set under his arm (more like both arms). We actually
attempted to power it without the yoke or CRT but there was
absolutely no evidence of anything. Surprise surprise.
Since the original description of the problems is somewhat incomplete,
a visual inspection is made and the HOT is tested for shorts just to be
sure. There were none. However, the visual inspection did confirm that
the flyback had a narrow but rather long (maybe a couple of inches) crack
in its housing, There was no conclusive evidence of arcing but this is
one area where the original symptoms were fairly definitive as the owner
stated that there was arcing around the flyback. (He probably knew just
enough to be dangerous, but hopefully has not done anything we will regret.)
This would explain the jittery horizontal but what about the vertical problems?
Were there really vertical problems. I never did get a good answer to this
question - at least not until later.
While it is likely that the flyback could be patched up at least temporarily,
it was decided to order a new one. The owner was willing to spend up to
$150 to repair the set - I have no idea why. No match from places like
MCM Electronics - must go directly to Magnavox (Philips, actually). $71,
ouch. Admittedly, this is one of the spiffiest flybacks I have seen lately
(at least since that A-line Zenith with the cool ribbed plastic coil form). It
has a detachable CRT anode wire - wire and suction cup sold separately!
Well, for $71, you cannot expect everything.
Although we have agreed to order the flyback, I decide to test the old one
anyhow, so next day I bring in my flyback testing widget (12 V chopper, see
document on flyback testing). This is the first time Dave has seen this
tester and Ed (our chief digital design engineer) is also curious but stands
at a safe distance, having a great deal of respect for a few puny KV. Ed
always stands at a safe distance when anything higher than 5 V is involved!
First step - locate the HV return. In this case, it is obvious because (1)
a separate bare wire is brought out to a pin and (2) this wire is connected
to no other pins on the flyback. (With a built in HV rectifier, it is
not possible to use a normal DMM to locate this wire.)
Next step - wrap a ten turn coil around the core of the flyback and connect
this to the chopper.
Apply power - a nice healthy arc can be drawn from the HV lead of the flyback
to the return connection, current draw on power supply is low. Flyback is
quite functional. This does not test for breakdown at full voltage but does
rule out hard shorted turns. (Ed can be overheard mumbling something about
sticking with 5 V logic.)
Result is fine by me, owner wants new flyback and this one will make
a great HV supply for a plasma globe or something - someday.
So, we pile the chassis and all its attachments onto a table in the corner
of the testing lab awaiting our shiny new flyback (minus the red wire, some
assembly required). It looks kind of pathetic there but no one else dares
go anywhere near let alone touch it after an off-hand comment about charged
capacitors!
Approximately 5 days later, our new flyback arrives and is soldered into place.
Next morning: I see Dave pulling up in his Chevy wagon. Guess what is
in the back? The entire huge, heavy TV, belly down. Oops. We quickly find
a place for it somewhat out of the way in a back room.
Apparently, there is no arcing and the horizontal deflection is stable.
But, there is absolutely no vertical at all - flat lined. OK, so the rumors
about vertical collapse were not exaggerated.
A little more visual inspection reveals a couple of interesting observations.
First, all the deflection circuitry - both horizontal and vertical - is
clustered in a small area near the flyback. In addition, the crack in the
crack in the original flyback is adjacent to some of the *vertical* output
circuitry. So, perhaps, the arcing was making its way to something in the
vertical deflection. What kind of output chip is it? Ah, my favorite - a
TDA3654. Fortunately, I have a bunch of them to keep one of my tough dogs
fed. So, I am well prepared if need be.
A quick measurement of power to the TDA3654 reveals that there is none.
Maybe this won't be so bad after all. Tracing back with an ohmmeter and
what do I find? An open fusable resistor! And, in exactly the right place
to be killing power to the chip. Could it be this easy? Actually - yes
in this case. I install a normal 1 ohm 1/4 watt resistor (only for testing).
I also use the ohmmeter to confirm that the rectifiers in the vicinity are
healthy. We are set!
Power! At first there is nothing on the screen but then snow gradually
appears - and it is full screen. There is no antenna. Of course, reception
inside our building is nearly non-existent due to all the computer RF
interference and steel beam construction. However, we quickly locate
a pair of rabbit ears (or maybe it was just a couple of feet of hookup
wire) and tune one of the few channels that is viewable at all - which happens
to be broadcasting the morning cartoons. But that is just fine. Everything
appears normal and I remind Dave to replace that resistor with a proper
flameproof variety.
Dave cannot believe it.
Ed is nowhere to be found.
Everyone loves the cartoons.
Comments: The mechanism for the vertical failure still remains obscure.
Apparently, the arc caused a momentary but not fatal short circuit in
some part in the vertical output circuitry which blew the resistor.
We always hear how sensitive ICs are to static - here we have 25KV of raw
power discharging nearby with apparently no permanent damage except to a
25 cent resistor.
Patient: Hewlett Packard AN/USM281A Oscilloscope. This is the militarized
version of the HP180 lab scope. PL1186 dual channel 50 Mhz vertical
plug-in, PL1187 delayed sweep timebase plug-in.
Symptoms: Horizontal position shifted almost off the screen; delayed sweep
and B timebase inoperative. Alternate triggering erratic at low
intensities(??).
Testing: How does one test a scope? Well, put it through its paces with
reasonable input signals - a 10 MHz clock oscillator provides a
nice test signal. Maybe another scope would be handy?
Prologue: (You can tell right off that this will be a feature length saga.)
I bought this scope at a garage sale. Now, understand, garage and tag sales
around the Philadelphia area where I live are usually of the "Aunt Minnie's
old silver plate" variety. Electronic equipment is usually limited to comatose
VCRs and color TVs that play in B/W (not that I complain about this sort of
stuff for the right price - as little as possible). However, one little
ad catches my eye: one item amongst all the bric-a-brac is 'test equipment'.
BTW, I never go to flea markets with any serious intention of buying anything.
It is clear where their stuff generally originates. All the junk I turn
down at garage sales ends up with hugely inflated prices at flea markets!
I get to THE sale relatively early (I am not quite the garage sale addict
type who gets up at 5 AM to be first in line). All that is visible are
a couple of pathetic old signal generators - one audio, the other RF. Well,
$10 for an RF signal generator isn't too bad. I could probably have
bargained him down to $5 but first the all important question: Anything else?
(Not that I expected any sort of affirmative response given the assortment
of hat boxes, deflated basketballs, and old Christmas decorations.) However,
surprise surprise! "There is one other item." So he crawls under a table
and drags out an HP AN/USM281A - a real oscilloscope! "Well, I have this,
um, oscilloscope. It is solid state, dual channel, 50 Mhz, etc." Now,
I am paying really close attention (but of course, not wanting to show it).
The only oscilloscopes I had seen at garage sales until this time (beside
my $3 Tek 321, but that is another story) are usually the really beaten up
Eico variety). He is actually doing a pretty fair sell job. So, how much
are you asking? "I would like to get $100 for it." Very interesting. Can I
try it? "Sure." So, he props 'my' scope on top of a rickety old bar stool
(I would have been quite upset if the thing had gone crashing to the floor
but still didn't want to act interested enough to suggest he find a more
stable spot.) I figure that if it appears to work at all, $100 is a good
price even if I have do some repair and calibration. I fiddle with the
controls, also noting that it comes with two nice looking 1X/10X probes.
Suddenly, the scope really cooperates - it must really want a new home
being so lonely stuck in the back of that garage. The trace scoots off to
the right of the screen. None of the front panel controls have enough range
to bring it back. I mumble: I cannot get the trace back. He says "Oh, um,
uh..." Before he can get too far, how about $50. "Sure, ok." I didn't do
enough testing to find out that the delayed sweep was also dead. For that
matter, even with the 10 turn delay time pot in plain view, the existence of
a delayed sweep mode did not register. I only found that out later. No
amount of fiddling would produce any difference between the A timebase and
Mixed A+B. The B timebase was totally dead.
So, how to go about tackling this? I have no service manual, no schematics,
and looking at circuit boards, the semiconductors are HP house numbers.
I didn't have an ECG manual. Fortunately, the component side of the
circuit boards are readily accessible. However, tracing the wiring is a
real treat with HP's love affair with multicolored striped bundled wiring
harnesses. I could try to buy a manual (this was somewhat before the days
of sci.electronics.repair). Nah, that would be cheating (and probably
expensive). I did try our local HP sales rep when we were looking into
their logic analyzers but he did the usual salesperson thing and lost interest
once we signed on the dotted line.
So, I had to repair it the old fashioned way - ohmmeter, circuit tracing,
seat of the pants, etc.
Objective #1: find the horizontal position problem. Even if the delayed
sweep remains broken, a dual trace 50 MHz scope is very useful. Fortunately,
this problem appears solid now (and not intermittent as was the case
previously) so ohmmeter tests of the horizontal sweep board components
should be possible. Tracing back from the deflection plate connections to
the CRT finally results in a difference between apparently symmetric sections
of the output stage. An 11 K ohm power resistor tests open! Well, that
wasn't too bad. It doesn't appear as though there is any cause other than age.
Replacing the resistor restores full range control of horizontal position.
About 1/2 hour to find. Not too shabby.
At this point, I considered the operation a success and put the other problems
aside. Partially, this was because what I had was usable and partially because
I did not expect the delayed sweep/B timebase problem to be nearly as easy
to solve.
(Actual elapsed time: about 2 years. Well you know how I suggest 'sleeping on
a problem' when you cannot solve it immediately!).
Objective #2: Fix the B timebase. The only help I have with this is the
fact that the A and B timebase circuitry is quite similar so some comparisons
of resistance measurements will be possible. However, for some reason,
my first pass over the components with an ohmmeter does not turn up anything
obvious.
Using another scope (a Tek 564), I poke around the B timebase circuitry
a bit to see if I can locate any interesting signals. I suppose my definition
of an 'interesting signal' is one that is doing anything - not flatlined.
What's this? A ramp waveform - it must be derived from the A timebase
as it maintains a fixed relationship with the A timebase output but goes into
the B timebase circuitry. My guess would be that this was be used to provide
a gating pulse of variable width controlled by the delayed time (10 turn) pot.
Maybe following this signal will lead to something. Sure enough, it goes
to what must be a comparator circuit since the pot also connects nearby.
And - what is that? A charred resistor! Now, why didn't I see that before?
Fortunately, there is another identical circuit across the board and
that resistor is readable - 100 ohms. Now we are cooking (hopefully not
literally). With great expectations - switch on. Still no action from the
Mixed A+B or B timebase. As far as I can tell, not a thing has changed.
The B timebase is still as dead as a door nail. The good news is that the
resistor is not getting hot, so that is encouraging - probably.
Next, I continue on with my search for shorted or open parts with an ohmmeter.
This time, I will be more systematic hopefully not missing anything. Finally,
results! Another blown part. This time, it is a transistor. From the
remaining good junction I can at least tell it is supposed to be PNP.
Measuring the voltages across C-E, it appears as though a 2N2907 will be of
adequate ratings (at least voltage wise). Hopefully. Soldering in the new
transistor from the top of the board is not fun - but more fun than attempting
to remove the board entirely. Power: Darn, still no action.
Continuing with the ohmmeter finally turns up a shorted diode but what type?
Again, duplication comes to the rescue. It is a 54 V zener. I use a pair
of 25 V or so zeners temporarily to substitute for this unusual diode. Now,
finally, the B timebase is responding to treatment! Having done what I always
caution against - turning adjustments without marking them - the behavior
is a bit - strange, but the two relevant controls (don't ask me at this point
what they did) could be set for proper operation of the Mixed A+B as well as
B timebase.
How could 3 parts in apparently not directly connected circuits be blown?
I have no idea except to suspect that the previous owner may have attempted
to repair the B timebase and never stumbled upon the actual cause which
was most likely the bad zener but managed to blow other parts in the process.
Objective #3: Determine why turning the intensity way down causes the
alternate sweep mode to get stuck on one trace. This one is a minor
annoyance only and really doesn't affect the utility of the scope in any
way. It is just not quite perfect. I finally locate a couple of adjustments
in the vertical plug-in that seem to have an effect on the rise time of the
start of the sweep or something - they seem to modify the strange behavior
as well as affecting the linearity of the left 1 cm or so of the screen for
high sweep speeds. I finally find a compromise position that seems to be
satisfactory.
Wow, a whole bunch of simple problems but now it appears to be in pretty
good condition. A couple of drops of oil for the delay time control,
clean the switches and controls, and a piece of Plexiglas to protect the CRT
and it is done.
Comments: What do they say? "Real engineers don't need no f***ing manuals".
Or is that programmers? Well, whatever. A service manual would sure have
made the task a lot easier. But where is the sport in repairing something
with actual accurate documentation?!
The AN/USM281A is still a good solid easy to use oscilloscope for general
design and debugging. I don't know if this one ever saw duty in the Navy
but I have seen this model going for about $250-300 from test equipment
outfits like Tucker. In my tests, the scope will display a viewable waveform
and lock at frequencies greater than 80 MHz even though the specs are for 50
MHz vertical bandwidth. The mainframe itself, I think, is rated for 100 MHz.
It takes standard HP180 series plug-ins so if I ever come across any of those
at another garage sale.... (about the same time pigs learn to fly!)
Patient: Jeff's Sony turntable. Two speed, quartz locked, nice - if you
care about such things anymore.
Symptoms: Platter turns but 'lock' light flickers and speed is slow and
uneven (even to the unaided eye and without listening, it is
obviously struggling).
Testing: Both speeds (33-1/3 and 45) have similar problems. The selector
switches appear to be clean and solid. By gently touching the
spinning platter, it is found that there is also essentially
no torque - it stops very easily.
It was quite obvious that the servo system was having difficulty reliably
locking - as evidenced by the flickering 'lock' indicator and lack of
torque. In addition, the platter did not appear to want to start up
reliably at some rotational positions.
So, what does it use as a reference? Remove the platter! This requires
popping a cosmetic cover and large E-clip. Aha, what is this?
A little pickup near the edge that looks sort of like a tape head.
However, unlike an audio or digital tape head, it has a series of offset
laminations about 1 mm in thickness. Not visible but inferred is a magnetic
stripe pattern on the inner surface of the platter which is in close proximity
with the pickup when the platter is installed. Better keep magnets far
away. That pattern was put on with a special jig at the factory - there
would be no way to reconstruct it if some, say, accident were to take place.
As a long shot, I attempt to adjust the pickup closer to the platter surface.
Perhaps the magnetic pattern has weakened or something else has drifted.
Ouch, now it is rubbing. A little further back. Ah, now it is clear.
Not too surprisingly, there is no change. Symptoms are identical.
Now what? Unfortunately, there is no way to get at the circuitry - under
the platter - when the platter is installed. There is no way to excite
the magnetic pickup with the platter removed. Not true! After storing
the platter at a safe distance, a simple magnet will generate a 20 mV
signal out of the pickup. This is probably greater than the normal
signal level.
The circuit is pretty simple - couple of transistors and other stuff
being fed from the pickup for feedback and hall effect sensors for
motor control. The motor is a brushless DC 4 pole type with the
commutation control external and on the same board as the servo lock PLL.
Maybe a little signal tracing is in order. Using my magnet, I can
see the feedback signal making its way through to what I assume is
part of the PLL - probably the phase detector.
In retrospect, even suspecting the PLL's feedback signal was probably
not valid. The key is the dependence on rotational position during
startup.
So, what about the motor signals? There appear to be two outputs from the
motor (in addition to the coil driving signals). What do these look
like? The first (from H1 - Hall sensor 1) flips between +5 V or so and
ground as the motor rotor is rotated through a complete revolution (Wow,
Peter Piper picked a pack of...., sorry, just practicing.) However,
H2 seems to be mostly dead - stuck at an intermediate voltage and only
varying by a fraction of a volt. Well, that could certainly be a problem.
(If this had not turned up anything, I would have gone on the determine if
the driving signals were correct).
Unfortunately, the actual sensors appear to be under the rotor. First the
rotor and then the entire motor circuit board has to be removed to access
H2. There is a slight risk that removing the powerful rotor magnet from
its pole pieces will cause it to lose some magnetization (this can happen when
removing the rotors of high energy servo motors, for example, if a dummy
rotor core is not slipped in as the rotor is removed). However, I risk it
and the rotor magnet seems to be just as strong as ever when replaced. I
stick it on a piece of soft steel as a keeper to be sure there no possibility
of it weakening while it is off the turntable.
The tricky part is accessing the bottom of the motor circuit board without
ripping up the connections to the (fixed) servo board. The actual Hall
effect sensor is in a four pin package about 2 mm on a side. There is at
least a red dot to indicate pin 1. I use an ohmmeter to double check that
the device itself is bad by measuring directly on the package pins. Confirmed,
there is are very noticeably different resistance readings across the pins
of the good and suspect parts.
I sent Jeff to buy the part (my 'customers' have to do something in return
for free repairs and Jeff can convince anybody of anything).
Fortunately, Sony Parts was happy to sell this $4 (wow!) part to someone
who is not a licensed Sony repair center. They were very helpful in fact,
looking up the part number for H2 in their parts catalog.
Reassembly is straightforward though I had no idea if the alignment of
the sensor was adequate until testing.
Now, at least the waveforms from H1 and H2 are of nearly the same shape
and about 90 degrees out of phase as they should be. Replacing the
platter and the lock light is now solid with decent torque from the motor.
The platter will start quickly and reliably from any initial position.
Next day: Jeff comes in complaining that the turntable is vibrating!
Indeed, during part of the rotation, there is a distinct vibration
that can be felt from the base. What could this be? Best guess is that
the pickup needed to repositioned. And, that was the problem. Apparently,
my initial adjustment in trying to locate the original problem had resulted
in the pickup being positioned further rather than closer from the platter.
Some careful incremental movement and the vibrations disappear.
Comments: The elegant simplicity of this turntable is impressive. Too bad
that turntables aren't really popular anymore. Well, maybe not bad. I
cannot complain about the benefits of CDs though other may disagree.
Fortunately, this turned out to be a motor problem and not an actual
fault in the PLL which would probably have necessitated a schematic.
I am not sure if Jeff ever used the turntable after it was repaired (about
4 years ago) as I cannot imagine him wanting to deal with vinyl. Maybe
that is why there has been no return call!
Patient: Tag sale Kenwood KX-55C cassette deck marked $2-, as-is, 'dead
transport'. I paid a whole $1. Otherwise, it seemed in good
condition. Well, at least they were honest.
Symptoms: Front panel indicators are alive and the logic seems to repond
to the PLAY and STOP as there is the sound of a solenoid but
neither the capstan nor takeup reel shows any indication of even
attempting to move. There is no sound of a whirring motor.
Testing: With the cover off, nothing appears locked up - everything turns
freely and the belts seem to be in reasonable condition. The motor
connections are readily found. Voltage to the motor is about 10 V
in all modes so in all likelihood, the problem lies inside the motor
casing. The only connections to the motor are the two power leads.
There is no external tachometer or other apparent feedback used for
speed control.
So far, this seems pretty straightforward. The motor bracket is fastened
to the transport with 2 screws, a bit awkward to reach but I finally
prevail. The motor itself is contained in an second metal case to
provide shielding. To get inside requires prying off the cover at the
non-shaft end with a screwdriver. So far so good.
Various techniques are used to perform speed regulation in a cassette
tape deck. However, for a totally internal regulator, there are usually
only two possibilities: mechanical governor and voltage regulator.
In this case, it turned out to be the latter - a little circuit board
contained several components including a power device which was shorted.
I was not able to identify or cross reference this part so I decided
to make my own regulator.
However, first I needed to determine (1) if there was anything else
wrong with the deck and (2) the proper voltage for the motor.
The only think I know so far is that the voltage to the motor is less
than 10 V.
There is no strobe disk to watch under fluorescent lighting.
Adjusting tape speed:
Make a recording of a single tone on a tape recorder you trust - one
with accurate speed.
My Heathkit audio signal generator and Yamaha cassette deck will do.
I use a variable DC power supply to drive the motor after soldering
wires to the motor terminals and remounting the motor in the deck. Since
I will not be attempting to squash my regulator inside the case, I should
not need to touch the motor again. I hope.
Apply power. Adjust the voltage to the motor to about 5 V and hit
PLAY: It turns! And, there is sound from my amp. Uh Oh, what happened?
The transport went into stop mode. In a VCR, I would suspect an idler
tire or belt. What was that? Did someone say belt? Right, the belt has
popped off of the tape counter. Maybe the tape counter has a contact on
it which is used to sense that the takeup reel is turning.
Now, it stays on but the pitch is way low.
Then, I adjust the speed while listening to this same source simultaneously
with the tape being played back on the unit to be adjusted. As the speed
is adjusted, the pitch changes. As it approaches the correct setting,
the tones beat against each other. When it is set correctly, the pitches
will be equal and the beat frequency will go to zero.
Even if you are tone deaf, it is easy to adjust the pitch accuracy
to better than 1/10 of a semitone using this method.
However, for this initial test, accuracy is not needed as long as the
approximate voltage is determined.
The result: A little over 6 V. Excellent, any vanilla flavor IC regulator
will do that from a 10 V input. The is quite low.
A scrap of circuit board, an LM317 (on heatsink, though this was probably
not essential at the 100 mA or so current required by the motor), a couple
of caps and resistors later, and Presto! a regulator. Speed adjust is done
with a 50 turn pot. (Well, it was handy.) I do not attempt to stuff this
into the motor case but screw it to the power transformer with an insulating
standoff.
I now repeat the speed adjustment - more carefully this time.
One thing I do note is that after a few minutes of continuous operation,
the motor speed and thus pitch changes by a detectable but not unacceptable
amount (still only a fraction of a semitone) and then stabilizes. This is
not due to the voltage regulator as the voltage is rock solid. I assume
that the windings of the motor heat slightly and increase in resistance.
Well, no feedback for speed control, what can be expected?
Regulator running cool, repair completed.
(5 or so years later:)
While listening to a tape, motor grinds to a halt. So what is up now?
Take off the cover. Hard to believe I had all the screws installed!
Checking with a voltmeter, there is only a few tenths of a volt across
the motor but normal 10 V at the input to the regulator circuit. With the
motor disconnected, the output of the regulator is exactly correct - 6.3
volts. So the regulator is most likely fine. What about the motor. Testing
with an ohmmeter reveals that the resistance varies between about .5 and 2
ohms - much less than I would expect. So, there is a problem with the motor.
It is extremely unlikely that the entire motor shorted out.
This is the classic partially shorted Mabuchi-type motor scenario. I am
not going to let a little motor get the better of me! I first try spraying
it out with some degreaser - no dice, no change. I will just have to
take it apart.
The end-plate is held on with a couple of bent metal fingers, pry them
out. However, I cannot just pull the end off as it will rip the delicate
metal brushes in the process. There are two access holes - I usually call
them ventilation holes but I bet their real purpose is to permit a brush
spreader tool (pretty impressive name, probably deserves a trademark) to be
used to allow the brushes to clear the lip on the commutator. My brush
spreader tool(tm) is - as usual - a bent paper clip.
With the end removed, I can now safely pop the entire armature out of the
motor case after removing the pulley (plastic, easy press fit).
(Note, this would be risky with precision servo motors using using high
strength rare earth magnets as the very act of removing the armature could
lead to instant partial demagnetization of the magnets. This is generally
not a problem with these cheap PM motors.)
The commutator looks a bit dirty but the real problem seems to be some
metal particles bridging the segments - probably dislodged by the
continuous rubbing action of the brushes. This is easily cleaned and
the gaps between the segments are cleared with a pointy dental pick.
Now, measuring across any pair of segments results in a 20-22 ohm reading,
much more reasonable.
Reassembly is uneventful and we are operational once again.
Comments: I can never resist attempting to repair a normally non-repairable
part like a motor. Sometimes, they turn out not to be repairable or I loose
a critical tiny tiny part to an unexplored corner of may basement but that
is the exception. In this case, it was quite easy. The motor did not actually
appear worn - the commutator and brushes were relatively smooth and undamaged.
Therefore, a replacement would probably not be substantially longer lived
than my restored motor.
Patient: Kenwood carrousel 5 disc CD changer. This is an 'everything in a
drawer' type design.
Symptoms: Sometimes, it will recognize the disc and start playing. Most of
the time, it will give up and go on to the next slot.
Testing: Several discs were tried without any conclusive differences
in behavior. If a disc is successfully recognized, it will play
without problems until about the midpoint and then the player
will abort and move on to the next. Attempting to manually
search forward past track 6 (on one particular disc being
used for tests) will result in an abort as well.
Unfortunately, once a disc is rejected, the silly thing remembers that
the spot (1 of 5) is empty so without cycling power, it is not possible
to even attempt to play the same disc twice in succession.
However, this is a minor irritation.
The first problem was getting the case open. While the screws appeared to
be ordinary Philips head screws, it required almost sitting on the
screwdriver to persuade them to break free. Hope he never expects to
have any warranty work done. The heads are barely recognizable!
Once open, the disc in the play slot is clearly visible. When the player
fails to recognize a disc, it doesn't spin at all. Ha! a spindle motor
problem. (Rubbing my hands!)
But first, check out startup. With suitable contortions, it is just
possible to make out the edge of the lens with no disc in place. Pressing
play results in the expected behavior of the lens moving up and down several
times before giving up (with no CD present).
There is no chance of being able to see the lens with a disc in place.
Forget it.
Well, maybe it is the spindle motor. I put a disc in and press play. Disc
moves into position, pause, give up and move on.
I'll show it!
Cycle power. Press play. Disc moves into position. And -- now: give
the disc a little twist. Yes! now the CD picks up speed and is recognized
and starts playing.
Hummmm.
Sounds more and more like a spindle motor. Perhaps a dead spot or
partially shorted winding. Or maybe a weak driver. Is this consistent?
For the most part, it is. Sometimes, the disc will be recognized and will
play on its own but most of the time, a little help is needed. Once it
starts, a dead spot would not matter and the hardest time for a driver is
when the motor is starting.
However, what about the problem of not being able to play to the end
of a disc once play is successful? I confirm that this is still indeed
the case. Yes, it is impossible to play past track 6 on this particular
disc. Trying another one, there is a also a time location beyond which
play aborts.
What does this mean? It now doesn't sound like the motor as the toughest
job for the spindle motor is at the start of a disc.
However, what else could it be?
So, I roll up my sleeves and prepare to do battle with the spindle motor,
at least testing it. This is not going to be fun. The spindle motor
is located underneath the pullout drawer and there doesn't appear ti be
any easy access. Well, every long journey begins with a single step....
As I am removing the bar with the clamper above the optical deck, something
catches my eye. What was that? This is the first unobstructed view of the
lens. That doesn't look quite right. Normally, the lens of a CD player or
other optical drive is shiny with a bluish tinge. This one is fuzzy purple.
That can't be right. Well, cleaning is easy enough not believing (but
hoping) that a dirty lens could be the problem all along.
One Q-tip later and, yes, all the former problems disappear. Discs
are recognized reliably, play flawlessly and to the end, and search
works fine for the entire disc.
How could a dirty lens prevent the spindle from starting but then
play apparently normally until a certain track? Well, I never
actually did confirm that focus was established in the situations where
the player gave up. But to have problems with focus but then be able
to play substantially free of problems. Well, no one ever said that
understanding was always guaranteed.
When discussing the results with the owner, he confirmed that the house
was somewhat dusty but I know that no one in his family smokes and the
player was not in a kitchen.
Oh well, at least I did not have to take the whole bloody thing (the owner is
an Englishman) to pieces.
Comments: Why am I reporting on a simple case of a CD player that
required a lens cleaning? Well, to point out that many problems really
do have simple solutions even if the initial symptoms may point elsewhere.
In a portable, cleaning the lens would have been done as a matter of
course due to the easy access. Cleaning the lens in this changer meant
removing the cover as well as the bar holding the clamper. And, some of
the symptoms did not really point to lens cleaning.
Patient: Panasonic PV-4820 VCR is basically comatose. Owner says her son
tried to change the 'fuse' to fix the problem!!???
Symptoms: Front panel display is alive but not much else. There is no
response to any buttons or the remote control.
Testing: Tried all the front panel buttons and remote control, unplugging,
etc. Nothing.
The fuse in the power supply is fine. What the @#$% is she talking about?
Once the top and bottom covers are removed, a thorough visual inspection is
performed. One intersting observation is that the capstan seems to be
binding - the rotor of the direct drive capstan motor is rubbing against
the bottom support bracket. Surely, this cannot be the cause of all the
problems, could it? A little more thought and examination of how the capstan
is mounted reveals the likely answer: there is a white plastic 'knob' with
a cross-slot poking through the main circuit board on top of the VCR. This
knob adjusts the height of the capstan bearing. No doubt the so called 'fuse'
that was attempted to be replaced was this white knob which does kind of
resemble a 3AG type fuse holder. OK, mumbling something about no user
serviceable parts inside, I adjusted the capstan height so that it is clear
of the bracket and spins freely. I make a note to check for any problems
with the capstan servo after the main problem is found since as expected,
there is no change in behavior with the capstan free to spin..
Now, back to the main event. Since there are multiple major system problems -
no response to any buttons qualifies - a power supply problem is indicated.
This VCR uses the famous (or infamous) Panasonic switching power supply
(though I did not know of its fame at the time). I locate the power supply
connector and start making measurements. I do not have a pinout or schematic
for this particular power supply but there just happens to be a similar
model Panasonic in the main conference room. How, I wonder if anyone will miss
it for a little while? Any important customers today?
One thing that is fairly obvious: there is no evidence of the typical
5-6 volts that most VCRs use for their logic. The closest voltage is
3.5 V - on a couple of pins. Checking on the power supply connector of
the other VCR, these pins correspond to the 5.1 V outputs. The other
voltages are a bit high but this is not surprising since there is little
load with nothing else working and the regulation is probably faulty.
Rather than working on the power supply in-place I decide to remove it to
the comfort of my workbench. This is for three reasons. First, if for
some reason, the outputs should skyrocket, I will not have the rest of the
VCR to fry. Next, moving the power supply around will probably break the
connections due to fatigue anyhow (it is soldered). Finally, it is simply
more convenient. Only 9 or so wires are involved, so no big deal.
Five minutes later....
Now, for a switching supply, I probably need a load at least on the regulated
output. I locate a 15 ohm power resistor, that should be about right for a
+5 V low current (probably under 1 amp) logic supply.
Powering up using a Variac feeding my isolation transformer, the power supply
is still outputting +3.5 V on the +5.1 V output. Good, at least it is not an
excessive load problem on the +5.1 V outputs (which would indicate a problem
elsewhere in the VCR. What about the other outputs? Well, they all test
somewhat high (based on the measurements from the other VCR). Thus, an
overload on any of those outputs is unlikely as well.
So, what could be wrong? Next step: draw out the circuit. This isn't a
complicated power supply, no controller chip, just several transistors,
resistors, diodes, an optoisolator, etc.
A half hour or so later, I have a not quite complete schematic. There
could still be errors, but it is mostly there and sufficiently detailed
to identify potentially bad parts.
Since this power supply is basically working - nothing smoked or blew up -
this limits the possible problems considerably.
Testing components in the regulator feedback circuits does not reveal anything.
What about the optocoupler? With power off, I tack a wire across the input
terminals (low voltage side). Now, if the optocoupler were leaky, this
should make no difference. Bringing up the AC on my Variac, the +5.1 V output
quickly reaches +5 - probably at an input of around 50 VAC - and shows no
sign of stopping there. This is the expected behavior if the optoisolator
were good. Cross that off the list. This also confirms that there is no
excess load on the +5.1 V supply holding it down.
What about the filter components for the +5.1 V? Jumping a good electrolytic
across the input capacitor (C16) of the Pi filter has no effect. So both
the capacitors are likely good as the first should be sufficient to hold
regulation. What else is there? What about the zener, D11? No dice, it and
its buddy, the only other zener, D15, test good.
What else is there? Hey, what is this capacitor, C21? It seems to be
in series between the +5.1 V output and the feedback circuitry? Since it is
in series, leakage is important. I need to swap it or test for leakage
out-of-circuit. I always opt to swap given a choice.
Ah ha! Success. Replacing with a good 1 uF 50 V capacitor results in
precisely 5.1 V. The capacitor, C21, seems to bypass the zener when the
output is changing - it in effect limits the rise of the +5.1 V output.
Thus, if it turns leaky and passes current, not only will the +5.1 V
output rise slower, it will be prevented from reaching its full specified
value. The poor regulator does not know the difference between a rising
output and a leaky capacitor.
Next day, replacing the power supply in the VCR restores full operation.
There appear to be no problems with the capstan. All modes are fully
functional. No other none user serviceable parts appear to have been
touched.
I write a little note explaining what 'No User Serviceable Parts' really means.
Comments: Knowing what I do now, checking C21 followed by C16 and C17,
and then the AC line filter capacitor, C4, would be the first thing to do
whenever one of these or similar Panasonic power supplies shows up with
incorrect output voltages. What this type of failure indicates is that
components like capacitors really are like peas in a pod. When from the
same manufacturer and lot number they are all very similar with respect to
failure modes like drying up and losing capacitance or turning leaky when
used in similar environments.
Most troubleshooting on the professional level is a matter of keeping good
notes and/or having a database of previous repair tech tips. While on the
subject, how about those solder connections in RCA/GE chassis CTC175/76/77
and all the others for that matter?...
As far as that 'fuse'. I must admit, had I known nothing about VCRs, the
white knob does look an awful lot like a fuse holder. To a homeowner, it
represents an irresistible temptation. Fortunately, he stopped short of
attempting to remove the 'fuse' entirely which would have no doubt resulted
in serious damage. And, no adjustment screws had been tightened! Thank you
for small miracles.
Patient: JVC HR-D860U HiFi VCR, about 2-3 years old.
Symptoms: According to the owner, the VCR gobbled up a tape and her
husband 'removed' the cassette but admitted he really wasn't
sure of what he was doing(!!).
Testing: In situations like this, a careful inspection of the mechanism
should be the first step. For JVCs of this era, parts tend to
fall off of the roller guide assemblies. In some cases, attempting
to play a tape can cause expensive damage - a loose roller guide
can swing up and smash the video heads.
Sure enough, the supply side roller guide assembly was loose on its
track and upon removing the bottom cover, the infamous brass pin fell
onto the workbench. One drop of Epoxy takes care of the roller guide
repair.
So what else is wrong? Was any damage caused by removing the tape
or playing a tape with the faulty roller guide assembly?
The VCR appears to be behaving just fine, thank you. At least it played
the one test tape fine that I tried.
Return to owner. Try it.
Next day: It's back: Doesn't work on SLP or pause modes. Uh Oh, that sounds
like a smashed video head. Sure enough, a close visual inspection reveals
that one of the video head chips is broken. Most likely, this resulted from
the roller guide flopping around and/or during 'removal' of the stuck
cassette. Unfortunately, repair price wise, this is a 6 head VCR. Lowest
head price from my source: $71. Ouch. I tell the owner. "That isn't too
bad, go ahead". OK, I guess she remembers what it cost new.
One week later: installation is straightforward as the connections are
all to a circuit board with no flying leads. Pop the new head on and
everything seems to be fine but - what's this - flag waving? Was that
there before? I sort of remembered some on my cheapo monitor with the
old head (the part that was still good, didn't think much about it at the
time).
So, I go and try the VCR on the Panasonic TV in the main conference
room. A little better, still some flag waving. However, everything
else seems normal - back tension in particular. It could just be
my tapes. If recorded on a VCR with significantly different back tension
and/or video heads that are not perfectly aligned on the drum (not a
user or tech adjustment), flag waving is possible and does not indicate
anything is wrong - just different. Many TVs will have a suitable fast
vertical sync response or a special channel with this characteristic so
not generally a problem.
I returned the VCR, concerned that it would still not be quite right. But,
next day, everyone was happy and even with my description, had no idea what
I was talking about. Plays just like new. Flag waving, what's that?
Comments: this is one of those situations where the recommendation: "always
confirm what does and does not work before touching a piece of equipment"
applies fully. Unfortunately, with the broken roller guide, this was not
quite possible at the outset. However, I should have paid more attention
to the behavior once the roller guide was repaired and before replacing
the video head drum.
Interestingly, this was one of my first contacts with the JVC parts shedding
VCRs but in the next couple months, came across two more JVCs - one with
a dropped brass pin and the other with a loose plastic hinge pin in the
roller guide positioning linkage. Then, I 'fixed' a coworker's JVC over
the phone (425 miles away) by his description of the symptoms: "Sure,
you lost a brass pin. Pull the bottom off....".
VCRs really are like peas in a pod.
Patient: Panasonic VCR with several problems. Owner's complaint is video
noise - sometimes the video disappears to be replaced by noise.
Symptoms: Indeed, video playback is poor. Tracking does not help.
At times, the video totally disappears as though the heads are
clogged. I also note that there seems to be a hum in the audio
at times and there is some erratic operation when switching modes.
Testing: I tested with multiple tapes. Thoroughly cleaned the video heads
using a cleaning tape, then cleaning sticks, then my thumb.
In no case was clear video totally restored and the heads seemed
to get clogged quite quickly. There is definitely hum in the
audio. Occasionally, when entering play, the VCR will abort and
unload.
This may be a case of multiple independent problems.
The hum leads me to suspect the power supply. This VCR uses a power
transformer, rectifier, filter capacitor prior to linear regulators.
Fortunately, unlike some Panasonic VCRs, the power supply in this one
is easily removed and may be tested easily. Putting a scope on the
main filter capacitor reveals that there is quite a lot of ripple - and
that it is also somewhat erratic in amplitude. This could indeed account
for the hum and erratic behavior. Some simple calculation show that
the capacitance must be reduced by 75%. I locate a capacitor that is
still perhaps only 2/3 of the labeled value and jerryrig it in place.
The ripple is now much much reduced and the hum is gone. Time will tell
whether the erratic operations is cured.
The owner was probably not even aware of these problems!
Now for the video problems - which are unchanged.
All symptoms point to a bad head. It is quite likely this machine
has seen significant use. The owner is apartment bound and spends a lot
of time in front of TV.
Indications of a bad video head include:
* Any visible damage to the ferrite chips.
No visible damage in this case.
* Excessive video snow which cannot be eliminated by the tracking controls.
An image where more or less good video alternates with snow at a 30 Hz rate
means that one of the 2 heads in a pair is either dirty or bad.
Definitely present at times.
* Excessive video snow or no picture on some playback speeds (SP, LP, EP,
still) since different sets of heads (in 4 head or more) machines are often
used for different speeds. If this is due to wear, then it would probably
gradually deteriorate and not happen suddenly.
Yes as described above. This is a two head machine so all speeds are
affected.
* Inability of certain internal adjustments such as backtension to eliminate
erratic tracking problems may indicate a worn video head. Horizontal
bands of video noise may come and go at various places in the picture
depending on what speed is being used or the playback location on the tape
(beginning, middle, end). These may come and go in a periodic cycle.
No significant response to backtension. All guideposts seems to be
locked solidly in correct position.
* Need to frequently clean the video heads even if you are only using new
good quality (name brand) tapes. Video heads are normally self cleaning but
very worn heads can tend to collect tape oxide resulting in a noisy, snowy,
or totally missing picture.
Absolutely - even after a manual cleaning (or using the Mark 1 Thumb),
bad video seems to return quite quickly.
* You have just been playing a rental, damaged, or spiced tape and you
notice any of the above symptoms.
She mostly plays old tapes. However, as noted there is no visible damage.
Nonetheless, old worn heads may be more prone to clogging.
Given that most of the qualifications for bad video heads are met, I
go and order a new upper cylinder from MCM Electronics. It is about $30.
This does indeed eliminate the video snow.
There is no further evidence of the hum or erratic behavior.
Comments: This VCR gets heavy use. I had cleaned it a few months before
the present episode and noted at that time that video playback was not
that great. But at the time, it didn't seem poor enough to warrant the
expense of head replacement. However, now it was definitely unusable.
The gradual and progressive degradation is a classic symptom of worn
video heads.
Patient: JVC 6 disc cartridge type CD changer used by Chinese restaurant
for background music.
Symptoms: Music gets stuck on outer tracks of some discs. Sections start
repeating, stuttering, etc.
Testing: All CDs exhibit the same problems at approximately the same
time into the disc.
History: This is the infamous 'Chinese Restaurant' CD player. I believe
the problems started when after several years, they decided to change or
add a disc. Yes, believe it or not, we had been subjected to the same
music for as long as we had been visiting this restaurant - at least a
couple of years. We were regulars. On this occasion, we were just getting
through the soup (these are the full course lunch specials) when, what was
that? The music got stuck for an instant. Being into this stuff, my ears
are tuned to typical CD problems. Must have been my imagination. Then,
halfway through the main course (Shrimp with Lobster sauce), there it was
again. This time, it really got stuck, repeating the same measure for
several minutes before anyone noticed and shut it off. Never being able
to refuse a challenge, when settling the check, I inquire: so music broken?
"Yes, not work." OK, I fix. So we walk out of there with this CD player
under my arm and something to do for the afternoon.
Problems of this sort are almost always mechanical with simple causes and
easy solutions. Attempting to play a disc - now it was totally incapable
of even recognizing the index - resulted in the motor that moves the optical
pickup just spinning its wheels. The worm gear was not moving and the
pickup remained stuck about 3/4 of the way to the outside of the disc. Giving
it a little help, resulted in resetting to the inner track limit and then
successfully beginning to play the disc.
The most difficult part about this affair was getting to the worm gear as the
optical deck is mounted upside-down and to remove it requires partially
disassembling the changer mechanism. However, once this is done, the problem
is quite obvious: Chinese grease! The last 1/4 portion of the worm gear is
gummed up. As noted, the same 6 discs had been playing for as long as I
can remember and none of these ever got to this outer track location. In
addition to cleaning and oiling the worm gear, I decide to order a new belt
which couples the motor to the worm gear as the slipping has probably weakened
it. However, I will return the player to them in the meantime as it should
work for a few weeks at least.
Well, that was about 3 years ago. The same new set of discs is still playing
and I still have that replacement belt sitting in my desk drawer. We did get
a free meal for the 3 of us out of the deal.
Comments: I know that many people are very careful to pamper their
equipment. However, unless the various mechanisms are fully exercised
to their limits at least occasionally - be it the pickup on a CD player
or the suspension on an automobile - dirt, grime, and corrosion can set
in and result in expensive repairs later on. At least that is my theory.
Despite my general low opinion of many JVC products, I have to admit that
this CD changer has been very reliable. It has probably been running 8-10
hours a day, 6 days a week, for over 5 years requiring only this one minor
cleaning and lube job during the entire time. That is not a bad record!
Patient: Mitsubishi HS-318UR VCR apparently fell off the TV where it was
perched from a height of about 4 feet. Owner had taken it to
a repair shop for an estimate - it was recommended he get a new
VCR. So, this was basically given to me. Well, loads of thanks!
He also had some broken pieces in a little plastic bag - a corner
of the front panel and a bit of a circuit board.
Symptoms: Totally dead, no display, no other signs of life. (Yes, I
know, you are not supposed to plug in a dropped VCR until
a thorough visual inspection is made inside and out.)
Testing: Not applicable.
This is an irresistible challenge. Aside from the chipped corner of the
front panel, there doesn't appear to be any exterior damage. But, the
piece of circuit board is ominous as it is apparent that there are numerous
traces broken - and probably more inside.
Overall, VCRs are quite tough. However, falling in just the wrong
way can do substantial and possibly not immediately visible damage.
I have heard of someone fighting off a would-be mugger with a VCR but
this too is not recommended practice!
In this case, the shop's estimate was way beyond what the VCR was worth
and the owner was perfectly happy to upgrade to a newer model. I did make
sure to get the remote control from him just in case a miracle happened -
or the remote was the only way left to access the major functions.
After removing the top and bottom covers and front panel, the extent of the
damage became evident. The good news was that the main board appeared to
be intact and the mechanism itself has no obvious damage though this will
not be fully confirmed until the electronic problems are addressed. At least,
there are no obviously broken or bent parts - and no loose screws or other
parts fall out. Nothing has popped loose.
I cycle the cassette loading and tape loading mechanism manually by turning
the appropriate motor shafts. Everything appears to be free and seems to
operate properly.
Inspecting for broken electronic parts does not reveal anything. At least
the display is undamaged. However, the corner of the front panel display
board is broken off - this is the piece in the plastic bag - thank you for
saving it! In addition, cracks have propagated in a couple of directions
cutting even more traces. Yuk. And, these are really, really fine traces.
Double yuk. And, this is a double sided board. Triple yuk. Why couldn't
the display have cracked in two - put it out of its misery.
There is always a slight risk that the initial impact has already fried
electronic parts as a result of a momentary short or from broken circuit
traces and there will still be problems even after repairing the visible
damage and/or replacing the broken components. Or, for that matter, that
my initial power test fried something else. Well, nothing ventured, nothing
gained.
Optimistically, at least 25 - possibly considerably more - 10 mil width
traces are cut. Some of the breaks are only visible with a strong
magnifying glass but as they say, a break is a break is a break.
First, I repair the board physically with some quick setting Epoxy. It
almost looks like a circuit board again. There is one little piece that
is missing - and some traces go smack through its center. Well, you cannot
have everything! Since, this is going to be a feature length restoration,
I decide to remove the display board from the VCR - unfortunately, it is
soldered in, 40 or 50 pins worth. At least, this is straightforward, just
tedious.
With the board removed, it will at least be easier to get at both sides
as needed. There is no particularly good place to start - they are all
equally bad. Well, every journey begins with a single first step....
Where I can locate suitable end points, I solder fine wire (insulated
#30 wirewrap wire) between these terminals. However, in many cases, it
is simply not possible to trace far enough to locate the ends. In these
cases, I use bare #30 wire directly across the break. A few I take
a short cut and just bridge with solder - this will bite me in the end.
I attempt to double check each connection after soldering for correct
wiring and that there are no shorts before proceeding to the next.
Unfortunately, one small piece of circuit board is completely missing
and traces pass through this piece. I guess as best as possible the
routing of these traces and hope. It seems to be part of the tuner memory
circuit.
After what seems to be an eternity - it is probably on the order of 5
hours - I am prepared for the smoke test. This requires reattaching
the connections to the mainboard and I will thus loose the mobility
of having the display board separate. Well, so be it. This isn't even
fun!
I plug it in! Well, no smoke at least. The display shows blinking 12:00.
Progress! Still, ready to pull the plug, I hit the power button. What
do you know - that even works. And there is some sound from the transport
initializing.
Now to try the various functions. Let's see, no response from clock set,
channel up, tuning controls. It is obvious that there are still a bunch
of problems. Other anomalies: the Quick Record indicator is stuck on.
What's this? The display just went dark. Oops, there is is back again.
Pressing on the board seems to have an effect. There is still an
intermittent somewhere.
Well, this is a start. Back to work....
Solder, solder, solder. After finding a half dozen more broken traces
that I missed on the first go around, try again. Now, the clock can
actually be set (well, except for the minutes). Quick record is still
stuck. There goes the display off again. So the intermittent is still
intermittent. Ouch! Why is the controller chip so hot? Pull the plug.
Checking over the connections, nothing seems amiss but in the rats nest
that the back of the board has become, who can say?
Following some traces that run under the display, I locate and repair
a couple more that I had missed before. I rewire those few traces
that I had just bridged with solder. I never do figure out exactly
how the traces running through that missing bit of board should be wired.
This is as good as it will ever get.
For some reason, the Quick Record light is now out, I have no idea why.
The display at least seems solid. The controller chip is cool again.
Also have no idea why. Now to try a tape....
Success of sorts. The tape will play but the sound is muddy. CUE, REV,
FF, and REW even work. It is not possible to reset the tape counter and
it is also not possible to set the channel memory. Previously set
channels can be tuned so that, at least, is not really a problem. And,
who needs the stink'n tape counter anyhow.
It is not possible to adjust the A/C head azimith enough to get truly
good sound. Only later, do I discover that the tape I was using for
testing had been recorded on a VCR with incorrectly set azimith so this
was probably a non-problem.
Just when I thought success was in hand, all of a sudden, after playing
for about 10 minutes, the tape suddenly speeds up to about 10X normal
with mickey mouse sound and loses sync. This is not CUE mode but yet
another problem. The STOP button works but EJECT overshoots the correct
stopping position as the main motor is obviously operating at increased
speed in all modes. Resetting requires unplugging the VCR for a few minutes
and manually rotating the EJECT gear a couple of turns. This is going to
be a peculiar way to watch a movie! Something is obviously heating up,
possibly causing a failure of one or more of the power supply voltages.
After spending about 15 hours on this VCR, I am in no mood to (1) give up
after being so close or (2) do any more serious work. Is there an
alternative? Yes, let me cheat. Maybe simply providing a bit more cooling
to the power supply will keep it happy. So, I dig up a little blower removed
from a defunct ultrasonic humidifier and prop it behind the VCR. Can this
possibly work? The answer is - yes. Apparently, the added cooling is just
enough to prevent the thermal runaway problem. Now, it is actually possible
to play or record a 2 hour movie (first experiment: Star Trek IV). With
this setup, the VCR will play or record indefinitely without spontaneously
going into warp drive mode.
Comments: If you take a piece of equipment that has been dropped into a
repair shop, the estimate you get may make the national debt look like pocket
change in comparison. Attempting to repair a VCR or any other complex device
that has been dropped is a very uncertain challenge - and since time is
money for a professional, spending an unknown amount of time on a single
repair is very risky. There is no harm is getting an estimate (though
many shops charge for just agreeing that what you are holding is - a VCR!)
I suspect that the final problem - the thermal runaway is not related to
the repairs I made on the display board but due, perhaps, to a heatsink
that was loosened that I did not notice. Since the small amount of
additional cooling provided by my blower was adequate to keep the VCR
happy, it could even have been a pre-existing condition that just did not
show up in a cooler location.
While a lot of effort was expended on this single restoration, for me it
was well worth it. This was my first VCR. Despite the less than total
success, the satisfaction was considerable. While the HS-318UR has long
since been retired I still hesitate to cannibalize it for parts figuring
that it has some historical significance.
Patient: Toshiba PA8706 power adapter (Supply #1) for a colleague's T1000XE
laptop. I can hear everyone saying "right, one of those really
state-of-the-art computers". However, the power adapters for newer
models are quite similar so any lessons learned are still valuable.
A second identical power pack will be pressed into service for the
loan of some of its organs.
Symptoms: When the bad supply is connected to a laptop with a known good
battery pack, the red power LED flashes continuously. No amount
of reinserting the battery or plugging/unplugging the adapter's
DC plug clears the fault. (These contortions are sometimes needed
when starting with a totally dead battery and are sort of normal.)
Testing: A voltmeter reveals that the output, 12 VDC, is cycling at the
same rate as the LED. This occurs whether connected to the
laptop or not. Comparing this with a good adapter shows that even
without a load, the output should be a steady 12 V. The other
power pack works fine with this laptop as well. Therefore the
fault is definitely in the power pack.
These power packs are actually universal (90-240 VAC or DC) input switching
power supplies which makes for a much more exciting adventure than a simple
transformer!
The first challenge was getting inside. There were obviously snaps around
the sides but it took a bit of detective work (read: the @#$% thing would not
split in two) to determine that there was a concealed Torx screw (fortunately
not of the security variety) under the manufacturer's label. Once this was
removed (by puncturing the decal rather than pealing it back), the case came
apart easily. At least it wasn't sealed and potted in Epoxy.
So, what do we have? It looks like a fairly typical small switcher - line
filter, bridge rectifier, filter capacitor (big!), switching transistor
and rectifier diodes in TO220 cases mounted to a full length heatsink,
C-L-C 'pi' output filter. Part types for the power semiconductors cannot
be determined immediately because of silicone rubber heat sink boots that
entirely conceal the devices.
Uh oh, what's this? A conformal coated hybrid controller. I better hope it
isn't bad as there is **no** way of determining what is inside. Obtaining
a replacement would no doubt be impossible as well. Naturally, this is the
first component that I suspect - wrongly, of course.
Well, what does a cycling output mean? Usually, this is an indication of
an overvoltage or overcurrent condition forcing the controller to continuously
shut down the supply and restart. A common cause for an overcurrent fault
are shorted components on the secondary side of the supply. Overvoltage may
result from defective regulator feedback components. The controller may
simply be bad and misinterpreting the overcurrent and overvoltage sense inputs.
What does the output look like? On my analog VOM which has a fairly fast
response, it appears as though it is cycling between 0 and a full 12 V (not
something in between that would indicate a severe overload or short. Running
the supply on a Variac reveals that the frequency of the cycle decreases as
the input voltage is reduced but that the voltage swing remains the same.
What is likely: the main switchmode transistor is probably good but there
may be shorted devices in the secondary side of the supply resulting in
overcurrent, a defective regulator resulting in overvoltage, or a faulty
sense circuit incorrectly shutting the supply down.
So, as always, I test all of the semiconductors that I can identify with
an ohmmeter. Of course, I cannot do this on the hybrid circuit so there
is an uncertainty that I might have missed something. However, everything
I can test checks out fine.
What is next? Capacitors, resistors. All of these seem fine. I substitute
some small electrolytics just to be sure. No change. Removing the output
filter capacitors one at a time also results in no meaningful change. (Maybe
the cycling rate increased - I did not check.)
At this point, I am temporarily out of ideas and the supply gets shelved
for a few months. There is a spare but no one is eager to give it up
(or maybe I just don't want to ask) for parts swapping - which is what I
really want to do. Guess which part I want to swap?
A garage sale comes to the rescue as I find an identical power pack marked
$5 but it is late in the sale (after all, how many people collect these
things!) and I get it for $2.50. Whether it works or not, this is a handy
source of parts - as long as it does not have the identical problem.
Supply #2: I plug it in and measure its output - nothing. Humm. A dead
power pack could mean anything but as with any AC adapter, the first thing
to check is the wire. And, sure enough, a little squeezing and bending
of the DC plug end results in momentary contact and 12 V out. No doubt,
the original owner spent the $60 or so for a replacement - not having
read the FAQ on AC adapter repair.
Now, with a little effort, I should have a working power pack - and a source
of known good parts as well. It turns out not to be such a simple repair
as I have to carefully slit the rubber plug cover and then fish the good
wire ends through that to reattach them to the plug itself. As they say, not
difficult, just tedious. I have been there before.
Back to Supply #1. Now, I start swapping like crazy. The hybrid controller
is first, of course. As you guessed, no change. Next is the transformer,
then the switchmode transistor (a MOSFET), capacitors. No change, no change,
no change.
Finally, the one part that I should have suspected all along - even if my
VOM said it was good - the 12 V rectifier. Swapping this part - a TO220
with a pair of high efficiency (or something) fast recovery diodes did the
trick. I know, you are saying: bozo, even in your FAQ, you say to never
fully trust an ohmmeter test. Well, it finally did bite me.
Comments: Most of the time, silicon semiconductors stay broken once they
fail but not always. In this case, for whatever reason, the diodes would
work like - well diodes - for long enough to provide full output voltage
but then one or both of them would turn into short circuits or at least
develop high leakage. Thus, reducing the input voltage would lengthen
this cycle as it would take longer for full reverse voltage to develop
on the diodes as a result of the filter capacitors charging.
With switching power supplies it is always tempting to blame the controller,
transformer, and switchmode transistor - because they are expensive, difficult
to obtain, or both. However, for a supply that shows some signs of life,
these components are rarely the cause of the problem.
Swapping parts between a working and dead unit - or between channels of a
stereo amplifier or dual channel oscilloscope - is always an ideal way
of eliminating selected parts from suspicion. It is a definitive test.
There is a slight risk (maybe not so slight with power supplies and
power amplifiers) that putting a good part into a circuit with some other
fault will blow the good part. But, this risk can generally be minimized
by using a series light bulb along with a Variac to limit current to the
critical semiconductors.
Patient: A colleague's Apple Mac Plus computer. Another state-of-the-art
piece of technology. But, hey, these things are still useful.
Symptoms: Power-on results in nothing on the screen and the familiar friendly
'bong' is missing as well. Just some unidentified electrical noise
indicating that it is getting power.
Testing: Nothing with the case on. It is obviously not operating as a
computer and does not go through the normal boot sequence or even
get to the point of asking for a diskette.
This is the 'classic' all-in-one style Mac, physically almost identical to
the original MacIntosh computer and the Mac 512K. The much newer Mac Classic
is constructed along similar lines. All require the special 'extra long
Mac wrench' (1 foot long Torx) to get at the two screws under the 'handle'.
Not having one of these (or actually, having lost mine), I make a suitable
substitute by grinding down the end of a triangular file handle. It works
quite well, thank you.
There are 5 Torx screws: 2 lower rear, one inside the backup battery
compartment, and 2 under the handle (they are not all quite identical
depending on if they screw into plastic or metal). Once the screws are
out, a case splitting tool is supposed to be used but a wide straight
blade screwdriver or other metal strip can be used **carefully** to
separate the two sections of the case without damage.
With the case removed, I start by examining the large Molex connectors for
damage and cold solder joints - there are none apparent (at this time).
How to identify voltages? Well, the floppy is pretty standard, so at least
+5 and +12 should be easily found. OK, +5 is low and +12 is very low.
Something is probably shorted. How about some ohmmeter checks?
Since this is probably a monitor problem, the natural place to start is
the horizontal deflection circuit. The small B/W monitor used in the
Mac Plus runs off of +12 on the isolated side of the main switchmode
power supply so at least this will not be a shocking experience.
Hmmm, HOT (BU406) is a dead short. It also appears to have seriously
overheated as the circuit board at its legs is somewhat blackened. The
most likely cause of overheating followed by failure is a bad flyback.
Nothing else in the vicinity appears bad based on the ohmmeter tests.
How does one test a flyback transformer?
Three ways:
1. Compare resistance readings of windings with a known good flyback. If
these match, substitute it. I really like this one and for a short test,
it is low risk - the chance of ruining a good flyback on a low voltage
deflection circuit is quite small. It is also a definitive test.
2. Perform a 'ring test'. Pulse a winding in parallel with a capacitor
and see how quickly the oscillations die down. A shorted turn or two
will totally kill the Q of the flyback.
3. Drive it with a known good chopper - this is my fun flyback tester - and
confirm that it does not load the drive circuit significantly and that it
puts out some high voltage.
See the document: "Testing of Flyback (LOPT) Transformers" for further info.
Happily, (1) is an option as I have a good Mac Plus just sitting around
doing nothing (you don't act surprised!). There appear to be several
variations on the basic flyback model - 157-0042B/B but they appear to
be interchangeable. Unfortunately, the ohmmeter comparison is inconclusive
and cannot test the HV winding in any case.
So, I decide to swap the flybacks - it is only 8 or 10 solder connections,
so quick and painless. I rummage around in my junk box (you probably get
the idea that my junk box is as big as a rail car) and locate a BU406D. The
built in damper diode should not hurt anything (there is an external one
in the Mac already).
Installing the transistor on its heatsink and soldering in the flyback takes
all of 5 minutes. Now for the smoke test. There is probably no need (and
no easy way anyhow) to use a series light bulb or Variac to limit the current.
The Mac switchmode power supply is current limited and unlikely to be able to
kill the HOT in the few seconds it will take to verify that the screen comes
up. I will know almost immediately if the voltages are more normal as the
friendly bong should sound almost as soon as the power switch is flipped.
And, indeed it does. A few seconds later, I am greeted with the inquiring
Mac Icon and inserting a boot disk results in the smiling Mac Icon and
a proper boot sequence. The picture appears to be bright, well focused,
and stable. A few tweaks to position and size, and it is perfect.
Now to order a flyback. $14.99 from MCM Electronics. I leave the BU406D
in place permanently.
Three years later: it's back!
New set of symptoms: single vertical line. This is a classic (no pun intended)
problem with Mac monitors - bad connections to the deflection yoke. And,
sure enough, whacking it brings back the horizontal, though it is somewhat
erratic.
The cause is obvious: one pin of the Molex yoke connector has staged a
Three Mile Island. It is totally melted and the normally white plastic
is blackened and brittle.
Rather than replacing the connector, I elect to just do a bypass and solder
a jumper from the remaining wire stump directly to its destination on the
monitor circuit board. Who ever removes yoke connectors anyhow? The remaining
three pins and sockets (1 horizontal and 2 vertical) appear to be in perfect
condition so I leave them alone.
I like easy problems for a change of pace!
Comments: Old Macs, like JVC VCRs that lose parts and Sony TVs with bad
tuner shield soldering, have a particular set of problems that are nearly
universal. Run a Mac Plus long enough and you **will** see a bad Molex
connector or dead flyback.
The flybacks usually fail either by shorted turns in the primary or secondary.
Cracked cores or HV rectifier failures are also possible but not common).
When it is a primary short, I believe there is a good chance the HOT will
blow if not caught immediately. A secondary short may just increase the
load enough to cause overheating but not immediate failure. Sometimes, there
will still be enough high voltage generated to have some light on the screen
but the picture, if any, will be highly distorted. (I have a few Mac flybacks
at this point and it seems to be a toss up as to which set of windings failed.
I like it when the primary goes bad because I can then perform a flyback
primary windingectomy and have a very nice core and high voltage winding
for constructing neat and nifty high voltage invertors. Generating more than
12,000 VDC from 12 VDC is quite easy and reworked Mac flybacks used in this
manner seem to be virtually indestructible. See the document: "Various Schematics and Diagrams" for a couple of complete designs using previously
owned, low mileage flybacks.
Patient: My color television - the one I actually bought new (what a concept)
at a discount store (after the yoke died on my Zenith B/W 21" tube
set from the 1960s). The RCA was about 4 years old at the time of
this failure.
Symptoms: At first, I just thought the camera work on the show I was watching
was a little strange, maybe the lighting at dusk or something like
that. But when the commercials also looked strange, it was obvious
that the TV was the problem. There were no yellows or blues! All
colors were in shades of red, orange, and cyan. This makes bananas
look particularly icky.
Testing: I tried different channels, adjusting the user controls, etc. No
change. Black and white programs were normal and turning the color
control all the way down resulted in a normal B/W picture.
Therefore, the CRT and its drive are fine, no missing colors. This
is a case of messed up colors - and there can be only one small
area of the TV's circuitry that can be responsible - the chroma
decoder. I suppose, a short between two of the primary color
signals - say blue and green - could result in a somewhat similar
symptom but it was clear by rotating the tint control that this
simple explanation did not hold up.
For a while, I just watched my weird TV since until I could get the Sams,
there would not be much I could do. After a couple days, it didn't seem as
strange but nonetheless, I didn't want to have to explain the situation to
anyone who visited. So, off to the library. This is a good excuse to copy
nearly the complete Sams folder for the CTC111 chassis even though a
relatively small area is involved (in this problem).
A single IC performs all the chroma functions including generation of the
3.58 MHz reference, gating of the color burst, I and Q decoding, and generation
of the R, G, and B drive to the CRT socket board. A disconcertingly large
number of discrete components surround this chip.
There are basically three signal inputs: luminance (B/W video), chrominance
(color information), and a color burst gating pulse.
Since the B/W picture is normal, the luminance input must be fine.
There is no likely scenario where a fault in a prior subsystem (i.e.,
tuner or IF) could mess up the chrominance in any way that would explain
the symptoms. The gating pulse matches the Sams waveform.
What about voltage measurements? These seem to be pretty close though at
least one appears to be a misprint in the Sams (I never do quite figure
out if it is or not, it did not have anything to do with the problem but
my control settings might not have been the same).
What else is there? The output R, G, and B drives we know are messed up
but do they appear funny in any way? Nope, they are all unique and the
amplitudes are similar - thus no apparent shorts between them or to anywhere
else.
The chip could be bad. Yes, I should have known better but this was over
10 years ago and ICs in TVs were relatively new. Therefore, I obtain a
replacement - $15, not too bad considering how many legs it has! I carefully
unsolder the old one and install a socket (as I always do in these situations
as damage to the cheaply made circuit boards is likely if it needs to be
changed again or the replacement is bad).
With great expectations, the new chip is plugged in, the TV is turned on and-
That looks even worse than before! Not only are the colors more messed up
but the contrast seems to be off as well. So, either the replacement is bad
(yeh, right) or the chip is not the problem. While bad parts are possible,
the symptoms looks suspiciously similar, if not quite identical. Therefore,
I conclude that it must be something else. $15 wasted? No, I learned a lot:
If the chip is good and the signals to it look good, there can only be one
set of alternatives - a bad discrete component or solder connection in the
vicinity of the chip.
Well, they all look like parts. No smoke has leaked out that I can see.
First, to determine which parts are likely to affect only the color. Based
on the controls (color and tint) and identifying parts that would have an
effect on phase or frequency response narrows it down quite a bit. There
are some resistors, capacitors, and a couple of small inductors.
A methodical test of resistances between pairs of nodes finally turns up
something. There is an inductor in what looks like a filter between two
pins on the chip which measures open! That would do it. It is marked
6 on the schematic and that is probably the normal resistance - it certainly
should not test open. On the parts list, I finally find it - a 39 uH
inductor. Just as a quick test, I take a 1 M resistor and wind a very
precise random number of turns of #30 magnet wire around it. The 1M is
irrelevant but acts as a convenient form for the coil.
Now, finally, a change in the color. Nowhere near correct but this is the
first time that anything approaching reasonable colors have appeared since
the start of this affair. Still not quite right. I get a another 1M resistor
body and wind a whole bunch more turns onto it. Now, that is a lot better.
Actually, probably good enough but I have access to an inductance bridge at
work so, resistor and #30 wire in hand, I finally manage to come up with a
fair approximation to a39 uH inductor. Sure, you are saying, just go buy
one. You know that is not my style!
Comments: There is no good excuse for an inductor in a low power circuit to
fail. The only explanation can be that one of the connections to the outside
leads was not made properly or the fine wire of the coil had been nicked
and finally just broke. I attempted to repair it but unfortunately lost
the wire inside the potting compound (it was something like #40 wire - very
very thin). Just bad quality control, not an induced failure.
Although it cost me $15 to replace the (good) chip, in the end, this was
probably well worth it as it definitively (well, almost) eliminated a large
unknown from consideration. The TV worked well for the next 12 years and
only recently developed the 'No Picture' problem to be dealt with in a
subsequent Repair Brief.
Patient: A friend's brother's closed circuit TV camera from a building
he rented (or something similarly convoluted).
Symptoms: No video. Output will sync the small monitor that came with it but
absolutely no signs of any response to light. You could set off a
magnesium flare 2 inches in front of it - lens or no lens - and
there would be no indication of any change to the video.
Testing: There are no user adjustments beyond focus (there isn't even an
iris on the lens). The camera was connected to a working monitor.
Nothing, zippo, zilch, nadda.
This is a fairly typical 1/2" vidicon NTSC B/W video camera. 16 mm, f1.6
lens with focus ring, 24 VAC power. Output is baseband (RS170) video from
a BNC jack. Apparently, it had been working but then at some point just quit.
No one knows if it was a sudden or gradual failure.
First, some vidicon theory:
A vidicon is one of a number of similar camera pickup tubes based on a non-
storage photoconductive target. An electron gun, not to dissimilar from the one
in a TV or monitor CRT generates a beam of electrons. A number of electrodes
shape and accelerate this beam toward the target. A pair of magnetic
deflection coils provide horizontal and vertical scanning exactly as in
a TV. A focus coil is used rather than focus electrodes (as are common in
modern CRTs).
The photoconductive target is high resistance when dark but low resistance
when illuminated. When the image from the lens is focused onto the target
coating, the light and dark areas of the image results in low and high
resistance areas on the target. As the electron beam scans, the target
current is a direct analog of the brightness variations of the image. A
very high impedance amplifier then boosts this signal to the level required
for RS170 video (roughly .7 V p-p into 75 ohms) which is combined with the
composite sync from the master timing generator.
Anyway, back to the story...
My friend Bill has quite a decently equipped lab related to his business
(special analog integrated circuits) but when it comes to something a bit
out of the ordinary, he is less than, shall we say, well prepared. To test
this camera on the bench requires a 24 VAC supply for its power. You would
think that locating a suitable transformer between a basement and garage full
of junk (sorry Bill, that is what a lot of it is!) would be a piece of cake.
Nope. We had to jerry-rig a couple of adjustable AC supplies in series to get
something approximating 24 V. Oops, phase backwards. Reverse and try again.
Finally, 24 VAC, and not too much smoke (though there was some - probably
a dead cockroach or something bigger). :-)
From the outset, Bill is insisting that the tube is bad. How? "Oh, the heater
will be gone". OK, where is the multimeter?
We use his 25 year old Lafayette VOM with the bent probes (one of which is
missing the banana plug at the multimeter end so the wire is just stuffed
into the hole). It is probably pins 3 and 4 of the vidicon. Yep, seems
low resistance to me. When the camera is powered, in the dark, it is indeed
just possible to make out the faint orange glow of the heater inside the tube
but this was not obvious at first. The voltage is probably correct, something
like 5 or 6 V DC.
Being the more conservative type, I suggest trying to trace out some of the
circuit to determine what's what.
Bill, on the other hand wants to start poking around with the scope. OK,
fine. At least, many of the connector pins are labeled with signal names.
There are even various voltage test points on the circuit boards.
Well, first, let use check the marked voltages. Using the state-of-the-art
multimeter, every marked voltage seems to be dead on. (Now, don't get me
wrong. I also have an identical VOM which I still use quite a bit. However,
it is in somewhat less battle weary.)
Unfortunately, all the marked voltages are power supply outputs - not pins
on the vidicon tube which is what is really needed (as you will see later).
OK, what next. "Let's check deflection." Sure, humor him. These are clearly
marked on the yoke connector. On the scope, the vertical looks like a decent
sawtooth but the horizontal is a really short pulse. I, figuring the
inductance of the horizontal coil suggest that this might be normal. Bill
doesn't buy into this. I also suggest that even if the horizontal deflection
were dead, there would be some light sensitivity - the response is real time
with respect to wherever the beam is hitting. As long as it is somewhere on
the target, there should be some response to light. The horizontal deflection
coils are AC coupled - capacitor is good - so the beam could not be fully
off target. Bill is still fixated on a deflection problem. Oh well, humor
him some more.
At this point, it is late, I want to do this in my more methodical way, so
I bundle the whole thing into a bag and take it home on the back of my 10
speed.
Now, I do not have the nice working setup that Bill does, but I do have parts.
You want parts, I have parts. In particular, I have no trouble locating a
24 V transformer and a line cord! So that problem, at least does not exist
at my place.
Since Bill won't sleep until this deflection thing can be put to bed also,
I decide to trace the circuit to see if the pulsed waveform is entirely
consistent with how it would appear to operate. I still believe it is
and it is irrelevant anyway.... The circuit is very conventional, not too
unlike a horizontal output circuit in any TV or monitor except that there
is only an inductor in place of the flyback. Checking the input signal,
it is quite clear that the circuit is doing what it should.
Then I get a phone call. Hi Bill. So we are discussing this result someone
mentions that it would be easy with a current probe to determine if the
deflection waveforms were correct. Current probe! Hello? Did someone
mention the words 'current probe' by an chance? Of course. I have a
Tek high frequency clamp on current probe - cost me a whole 10 cents at a
garage sale. The last time I used it, Bill and I were trying to measure
the instantaneous current pulse in an electronic flash unit (about 60-100 A).
Why didn't we think of this earlier (Bill has one as well but I bet he paid
more for his! Then again, with Bill, you can never quite tell).
This takes all of five minutes - the wires are clearly marked. And, what do
you know, a perfect current sawtooth! Hi Bill, guess what? Horizontal current
is a perfect sawtooth.
Next, I decide to attempt to figure out what each of the signals connected
to the vidicon do. There are 7 pins, the internal electrode arrangement
is clearly visible, and the names are printed on the vidicon socket board:
Pin 1: G1 ; Grid 1 - sets the beam current (like brightness, 0 to -80V).
Pin 2: K ; Cathode - electron emitter. (0 V, 20 V for blanking)
Pin 3: H+ ; Filament, around 5-6 V DC.
Pin 4: H- ;
Pin 5: G2 ; Grid 2 - this is very close to the cathode. (400 V) (1)
Pin 6: G3 : Grid 3 - this runs most of the length of the tube. (300 V) (2)
Pin 7: G4 : Grid 4 - this is a short ring near the target. (400 V)
1. This is suspicious as I would expect it to be somewhere in between G1
and G3 in value. However, based on the circuit, this probably is correct
as it is fed through a resistor (not voltage divider) directly from the
400 V supply.
2. This voltage seems to be the feedback for the power supply regulator.
____ _____________________
| \_____ ____
| ___
| | _______K
| |< H+,H-
| |_______
| _____ ____ G1
| ____ _____________________/ G2
| G4 G3
|
Target (Video signal out on external contact ring). Biased at around +10 V.
Touching or even going near the target ring at the front of the vidicon
results in a hum signal in the video output. Therefore, the video amplifier
chain is almost certainly working correctly.
Taking measurements while adjusting the Beam and Target pots seems to
indicate that they have the expected effect. There is just no beam current!
What else? Maybe Bill is right - the tube is shot.
I start rechecking the measurements - probe slips, zap - what was that?
Did I see a picture for a second or two? What did I touch? Figuring a
cold solder joint, I start prodding in the vicinity. Nothing. OK,
recreate the incident. The G2 test point and another pad are very
close. If I momentarily touch the two, there is a tiny spark and a faint
but recognizable picture appears for a few seconds and then fades out.
This is entirely repeatable. What is that other pad? It turns out to be
G1 - the beam control grid. Surprise, surprise!
At first I thought that dragging down the G3 voltage or the +300 V rail
was having the effect but, of course, it turned out to be pulling up on
the G1 voltage - making it more positive with respect to the cathode. In
fact, shorting G1 to its buddy next door on the same connector, +H, results
in a very nice picture. (Recall that +H is around +5 or +6 V).
For now, I will see if I can get it working consistently and worry about the
explanation later. I grab a 1 M resistor and solder it between +300 V
and G1. Now, it is possible to adjust Beam (and all the other controls I
messed up) to get a quite decent picture. The resistor provides enough
positive current into the BEAM pot so that the G1 voltage can easily be
brought up to 30 V - well beyond what is needed.
As far as I can tell, the video is fine, the settings are stable. So,
we have a case of a device that works in way it should not - which usually
means that the 'fix' is masking some other fault. Without schematics with
voltage levels or a spec sheet for the vidicon, there is not much more I
can do. Bill will no doubt want the thing back anyhow and it will be
hard to refuse in the interest of further research.
Comments: There are times when total comprehension is not required. It
would be a simple matter to substitute a known good vidicon - if I had one.
There is only so much time worth spending on something like a B/W CCTV
camera. Based on on my testing, it works reliably. I have had it on
for several hours without any change in behavior.
Almost certainly, there is still a fault - perhaps Bill was correct all
along and the vidicon is indeed bad in some way - not the way he thought,
but bad is bad except when it works! Or, maybe one of the voltages is
indeed far enough off to be killing the video. Maybe it will drift
further out of spec over time and the 'fix' will now longer be enough.
For now, it works, move on. This camera will never be used in any
mission-critical application, probably chip inspection, so if it dies,
Bill will just have to unpack the Mark-1 magnifying eyeballs from storage.
Patient: Heathkit 3-1/2 digit bench/portable digital multimeter from garage
sale junk box. Apparently, this kit was never entirely completed
as the cover was still in its protective plastic wrap.
Symptoms: No display or any other signs of life. No smoke either so at
least that is a positive sign.
Testing: Switching through all ranges and modes results in similar behavior
as in no signs of life.
Here is a situation that is not quite typical of a repair: a device which is
in a totally unknown state. Was it constructed properly? Were all the parts
good? Did some prior attempt to make it work blow something out? In many
ways it is much better to be working on a device that was functional and
then blew its top than to be dealing with something that never worked
properly and whose history is not available.
Well, with every long journey, start with a single step...
I don't intend to spend a lot of time on this. If initial tests do not
produce any revelations, shelve it. However, I do have the complete
instruction manual and schematics so this will not be a totally blind
challenge.
Since there is no action of any kind - not even a hint of a display - either
the power supply is bad or the controller - an MK6013 - is bad. Hope for
the power supply.
This meter is designed to be used on the bench or as a portable. It uses
4 C-size NiCd cells trickle charged when plugged into its detachable power
cord. Checking on each of these indicates that they are charging. After
a few minutes, there is greater than 5 V across the battery which should be
adequate for normal operation. Still nothing. Unplugging, shows that the
cells (despite some crystallization around the ends) do hold a charge.
The power supply runs off of the NiCd battery and uses a DC-DC convertor
to generate all the voltages required for the circuitry except the +5
which is obtained from the battery through a pass transistor. The neon 7
segment displays require +/- 170 V and the analog circuitry uses +/-12 V.
Listening carefully, it is just possible to hear the whine of this DC-DC
convertor so at least it is starting up. Are the voltage outputs correct?
Well, the +170 testpoint measures about +70, the -170 V testpoint and +12 are
also low and -12. Humm, -12 is nearly non-existent.
Putting a scope on one of the inputs to the convertor transformer shows a
highly asymmetric waveform. This is a totally symmetric power oscillator.
Therefore, such a skewed waveform probably indicates excess load on one
of the outputs - probably on just one diode of the rectifier for that output.
Before I start mucking around in the power supply, I unplug the MK6013
chip and set it aside in conductive foam. I don't really want to blow
it out (assuming it is still good) through some power supply screwup.
All other parts are common and readily available. I probably would never
get around to searching for a replacement for the MK6013, however.
First step: disconnect all windings of the transformer that are not
needed to run the convertor oscillator. Fortunately, this is easily
accomplished without removing the transformer. Suck the solder, make
sure the pins are clear of the pads. Power! OK, now the waveform is
a nice squarewave (much higher frequency and no longer audible). This
proves that (1) the transformer is likely fine and (2) it is an excessive
secondary load problem.
Which output?
First, I reconnect the high voltage (+/-170 V) winding. Still a nice
squarewave. The +/- 170 V testpoints measure about 180 V and there is
just a hint of a glow from part of the display. Without logic drive,
this could be correct.
So, the problem is in the low voltage portion of the power supply or
somewhere else powered from the low voltage. Good news and bad news.
This narrows down the problem - to most of the circuitry in the entire unit!
However, the fault is likely in the power supply.
Next, reconnect the transformer winding for the +/-12. Now, the bad
waveform is back. The asymmetry of the input waveform almost certainly
points to a bad rectifier as this is the only component that could
produce a load current depending on the polarity of the input. Once
rectified, excess load should affect both polarities of the input waveform
equally and thus not result in the asymmetric waveform.
First, is it the positive or negative 12 V that is the problem? Pulling
one transistor easily confirms that the +12 V output is not the problem
I then unsolder each of the two rectifiers for the +12 V and nothing changes.
Now for the -12. Unsoldering D212 makes no difference but removing
D214 restores the symmetry of the waveform. Even without this diode in
place, all output voltages are now well within expected tolerances. A
1N4007 takes the place of its dead buddy.
Now for the acid test: replace the controller chip. And - Yes! - we
have a display. It even looks somewhat correct. Obviously, the meter
has not been calibrated but it does produce the expected results for
DC V and Ohms. The + indicator and a couple of segments are out but some
quick jiggling and tapping confirms bad connections. The bad segments are
due to loose socket pins and easily remedied with a pair of needlenose pliers.
I never do discover the actual cause of the bad + indicator but the problem
went away after cycling the AC/DC switch several times.
Some quick calibration shows that it is in pretty good condition. There
may possibly still be some minor accuracy problems with a couple of the
ranges but in general - the @$#% thing works!
A bad diode! A stupid, 5 cent, bad 1N4002 diode! Can you imagine the
frustration of the original owner who, after meticulously constructing
the kit - following every instruction to the letter, double checking and
triple checking - is unable to make it work because of a simple power
supply problem? Keep in mind that this is not a modern DMM where the
major parts consist of: case, selector, IC, LCD. This is 1975 technology
with many many discrete parts on two 4" x 6" circuit boards. The effort
probably represented several dozen hours labor with a hot soldering iron
and the extensive use of choice 4 letter words.
Comments: A Heathkit IM2202 Multimeter is nothing to write home about by
today's standards but is still useful nontheless. The significance here,
if there is any, is in the type of situation in which one is attempting to
make a newly constructed unit operational. There are subtle differences in
the diagnostic procedures required due to the unknowns involved. In this
particular situation, there was no way of knowing how much damage might have
been done through prior unsuccessful troubleshooting or whether - as it turned
out - some original parts were bad as well. This was obviously not the first
electronics kit constructed by this person - there were a number of successful
examples at the same sale. Therefore, I did have some confidence that the
basic construction was solid and accurate and there would be some minor
problem involved - which as it turned out, was correct reasoning.
Patient: My color television - the one I actually bought new (what a concept)
at a discount store (after the yoke died on my Zenith B/W 21" tube
set of many years. This is the second of a 2 part Repair Briefs.
Symptoms: I turned it on and - sound but no picture. No light visible on
the screen. That is kind of strange. Except for the color problem
10 years ago, this has been a reliable set. (No tuner shield solder
problems, knock on plastic!)
Testing: No user controls have any effect. The screen remains black.
There is no flash of light when powering it off either. There
is static on the screen however, so I suspect that HV is fine.
Nothing is shutting down on its own either.
Since there is sound and HV static, I assume the horizontal deflection is
running. As a simple test, I turn up the SCREEN control. This results in
a raster. Therefore the power supplies, horizontal, vertical, high voltage,
tuner, and sound are all in fine shape. Just no #$%^ picture. Well, that
narrows the area of search considerably.
This will require the Sams and a scope to trace - sure, no problem. I still
have the photocopied Sams from dealing with the previous problem and my scope
is just dying to chomp on something interesting.
The only difficulty is that the bottom of the main circuit board is not
accessible when mounted in the cabinet. However, maybe RCA thought of
this - there are two offset screw holes which (whether by design or not,
I do not know) permit the circuit board to be moved about 8 inches to the
rear - and the cabling is even long enough. Now, with the TV on its side,
everything I need to get to is out in the open.
With my isolation transformer - always a prerequisite - I can conveniently
examine various points in the video chain. However, some testpoints are
concealed under a shield soldered at multiple points. A slight annoyance -
the soldering on this shield is really really good.
Just when I have settled in to try to determine where the video signal is
getting lost, the picture pops on. Actually, it kind of dribbles on - first
a weak rainbow pattern and then a second later the good picture. Could this
be a bad connection? Perhaps, but no amount of wiggling, jiggling, whacking,
or cursing, has the slightest effect. Cycling power also does not induce
the problem. It is now as though the TV is not broken.
Hmmm. sounds like an electrolytic cap or thermal problem. (BTW, Bill - from
Repair Brief 35: "Panasonic PV1414 Closed Circuit TV Camera with no Video" is
sure that the problem is a bad electrolytic cap and wants to destroy the
ozone layer with a can of CFC based freeze spray to locate it.)
Let it cool for an hour.....
One hour later - picture comes on instantly!
Let it cool for a day.......
One day later - picture comes on instantly!
Leave it alone for a week................
Finally - no picture - for about a minute, then rainbow, then solid picture.
So, how does one go about troubleshooting a problem of this sort? Give me
a problem that takes a minute or hour to show itself any day but one that
is broken for only a minute or two? Not fun.
Well, perhaps even when it is working, there will be some voltage or
waveform that isn't quite right. Maybe a visual examination will turn
up some potentially bad electrolytic capacitors. Maybe one has lost some
of its value and a capacitance check will reveal the culprit. Right,
keep dreaming.
First step: narrow down the search area. At this point, I have no idea
exactly where the video signal is getting lost - not even which subsystem!
I do know that sync is stable - examining the raster shows that the lines
are correctly interlaced when a proper channel is tuned. So, this puts
it after the sync separator. Since this is fed from the video at the
output of the video IF, I know at least that it is functioning to this
point. That leaves the entire luminance channel - joy - as well as the
chroma/luminance IC (the one I tried replacing in Part 1 of this story,
U700).
I spend some time down one blind alley - the chroma input to U700.
Suspecting this signal, I confirm that it is indeed always present
even when there is no picture. A simple test of shorting this signal
to ground at the chip input would have revealed that the most its absence
would do is result in a B/W picture. One dead end.
Here is a rough block diagram of the video chain:
+--------------+
+-------| Lum Peak/Amp |-------+
| +--------------+ |
| C v U700
TP-A +-----------+ +------------+ B +-----------------+ RGB to
Video IF ---+---| Comb Proc |----| Chroma Buf |---| Lum/Chroma Proc |--- CRT
| +-----------+ +------------+ +-----------------+ Drive
To Sync Sep. D ^
+------------+ |
From Sync Sep ----| Burst Gate |--------+
+------------+ |
|
+----------------+ |
From Vertical Deflection ----| Vertical Blank |------+
+----------------+
I know that the video signal is fine at TP-A at all times. I also know now
that TP-B is fine and does not kill the video in any case should it disappear.
Next, I identify the pins on the Lum/Chroma Proc chip (U700) that can
possibly kill the picture totally. It appears as though there are
only five: 4 of these are controls or power supply voltages (constant DC
levels) and the other one is the luminance input, TP-C.
One week, later, preparing for battle I am set to measure the 4 static
voltages as soon as power is applied. The scope is hanging on TP-C
as well to determine if that signal is present at all times. A signal here
would eliminate all the stuff in the 'Lum Peak/Amp' box from consideration
(and there is considerable stuff in this box - I have greatly simplified
the block diagram).
Charge!, or, err, power!
Voltages: Measure, measure, measure, measure! These all check out within
5 percent or so.
TP-C signal: Yep, it comes up immediately and appears to be valid video and
of proper amplitude.
There is no detectable change in any of these when the picture appears
30 seconds later.
That's just great. All five possibilities seem to be fine. Well, when
the plausible are eliminated, check the implausible.
First, however, I breakdown and decide to swap U700 with my spare in a
week when I apply power again. As expected (but hoped against), the screen
remains black for the expected minute. When the picture appears, the color
balance is somewhat off (recall, that in our last episode, there were more
severe problems with this chip but that was when the actual fault had not
been located). However, adjustment of the CRT drives fixes this. For all
intents and purposes, the replacement U700 performs identically to the
original.
So, there must be some other input to U700 that is messed up.
Well, there is another possible candidate. I had not paid serious attention
to it before but it is all that is left - the signal derived from 'Burst Gate'
and 'Vertical Blank'. I initially discounted this since there was a signal
present even when the picture was blank - but was it the correct signal?
Brief checks previously had shown it to healthy - but that was when there
was a picture.
One way to find out: short it to ground and see what happens. And - yes -
this kills the picture. What is also significant, is that when the short
is removed, the picture comes back in a very similar way to when the 'problem'
goes away - some initial rainbow effects before it stabilizes.
So, next power cycle (1 week) I am intent on catching this signal in the act!
This signal should look like a 2 volt pedestal during horizontal blanking
with a narrow 5 volt pulse riding on top. This is called the 'Sandcastle'.
I am now thoroughly familiar with what the correct shape should be.
___
|\ ^
|| |
__||_ 7V
________| |_________ _v_
1 Week passes......
Scope, are you ready? "At your service".
Look at that! The pedestal is there but the narrow pulse is missing.
Finally a clue. Some quick voltage measurements of the 'Vertical Blank'
circuitry shows that it is impossible for it to be at fault as the pedestal
would not appear normal. Therefore, it must be the 'Burst Gate'.
This is a fairly simple circuit - a transistor buffer is driven into saturation
by the inverse blanking signal. An RC delays its turnoff at which point
the positive going pulse at the collector is coupled via a C-L-R network
via a diode to U700. The following is somewhat simplified:
12V
|
/
\ R744
/
\ C710 CR700
+---||---+---|>|---+---/\/\--- TP-D
_____ __ R742 |/ | |
|_| ----/\/\---+---| Q701 C /
_|_ |\ C L703 \ R741
C709 --- | C /
_|_ _|_ _|_ _|_
- - - -
Next step: to determine if this transistor is working correctly. This time,
scope probes on input and output. The verdict: input and output are both
stable instantly at power-on. Having confirmed this, I immediately kill
power. Maybe these quick checks will permit whatever is being cured in
a minute to remain sick. Now what?
What about the other side of C710. Nothing. How can that be?
Coupling capacitor? Quickly bridging another one doesn't result in any
change. Kill power. I cannot afford the bench space, need to get this
wrapped up!
On a hunch, I check the resistance across L703. 20 K ohms, Huh?
No way, it is marked 21 ohms on the schematic. Not another dud
inductor! (Recall that this was the problem in Part I - messed up
colors).
I pull L703 - it measures open. How could an open inductor result in
a flat-line output? One might think this would increase the output.
Well, with the inductor removed or open, the diode charges the capacitor
as it would in a power supply and results in the diode output being clamped
to zero volts. It actually makes sense.
First I try winding an inductor but even with as many turns of #32 wire
as I can fit on a 1/2 watt resistor body, the value must be way too low
as the output is still dead.
Rummaging around in my inductor drawer, I locate one that looks kind of
similar. Unmarked, but of approximately the right size and construction
(I really don't have a great stock of inductors.) The original is marked
471 which I assume to be 470 uH but I didn't photocopy the page of the Sams
Photofact with inductor ratings, unfortunately. The results with my unmarked
replacement are mixed - the picture is there (well, that really doesn't
prove anything) but it seems a little bright. This signal looks fine,
maybe the bright picture is my imagination. OK, try a different larger one.
This even measures the same on my ohmmeter as the original should (like
that means anything). However, the picture now appears normal.
Since the poor lonely inductor sitting on my bench still measures infinity
ohms, I am confident that this is indeed the problem. Exactly what
mechanism results in a delayed start inductor is not quite clear. It is not
heat as there is no time for any thermal effects and the power dissipation
in the inductor is about as close to zero as one can imagine. Is it simply
the voltage pulses appearing across some kind of marginal semiconductor-like
junction formed by corrosion between the coil wire and the leads that
eventually results in good contact?
Attempting to 'disassemble' the broken inductor simply results in broken
inductor pieces everywhere so I will never really know for sure what happened -
until another one of the half dozen or so similar inductors in the TV decides
to do its open circuit thing.
Maybe someday I will actually order the correct replacements for both of the
dud inductors - this new unmarked one and the homemade inductor that fixed
the previous messed up color problem. For now, it seems to be fine and the
TV shows won't be any better with the proper replacement parts anyhow.
Comments: No doubt next time (sure, if there is a next time) I will test all
the similar inductors first! What could possibly lead to a batch of unreliable
inductors is another one of those mysteries of the universe. After all,
an inductor is just a coil of wire soldered to a couple of leads. There is no
thermal or mechanical shock and the circuit is very low power in any case.
The TV is not in a damp location or subject to any other kind of abuse that
I know of though based on the appearance of the innards of the inductor,
some type of deterioration may have taken place.
How should one diagnose a problem of this type? In hindsight, I guess
testing components in the vicinity of the Lum/Chroma Proc after waiting
a week and with power off would make sense. But, some of the elapsed time
was required to localize the problem to that chip. However, once it was
clear that one of the signals to U700 was messed up, the ohmmeter checks
would have greatly reduced the additional required debug time. Would
freeze spray have worked? Perhaps - if anyone had thought to hit the
inductor. There were no electrolytic capacitors anywhere in the circuitry
around U700.
Patient: Garage sale TV. Waited for price to be marked: $5. Paid $3.
Then asked if there was remote - they wanted to charge me $30 for
that! Just kidding.
Symptoms: Pressing front panel power button results in a click but no
indication of deflection or anything else. By listening carefully,
it is just possible to determine that the 'on' and 'off' clicks
are not quite the same sound. Therefore, the system controller
thinks it is cycling power - this is not a shutdown problem. There
is no response of any kind to any buttons on the remote control.
Testing: With the cover off, the first check is to determine if the horizontal
output transistor (HOT) is shorted - it is not. Applying power
(using my isolation transformer), the main filter capacitor shows
the expected +150 V from the bridge rectified 115 VAC. Shutting
power off causes this voltage to decay over the course of about
10 seconds - therefore, the power relay, its control, and main
filter capacitor are probably good. Checking voltage at the HOT
shows that it is present (though high - I suspect this is simply
due to the fact that there is no substantial load with the set not
running). Thus, this is a startup problem and not a catastrophic
failure of the HOT, flyback, or some other expensive part. More
importantly, startup problems usually have a definite cause and are
not likely to reoccur.
The next step is to obtain the Sams Photofact for this set. A trip to the
library is in order. Until then, what about the remote control? Is it bad
or is there something else in the TV itself that is not responding to the
remote?
A quick check with my IR Tester reveals that the remote is not putting out
any IR signal for any button presses. I double-check the batteries - they
are good and will operate other remote controls. Opening the remote control
is not as bad as some - one screw and some snaps along the sides. Nothing
looks immediately amiss - no broken parts. But this is the component side
of the circuit board. Depressing some catches around the edge allows it to
be removed. Hey, what is this? Along one edge, a trace has been lifted off
and broken by one of the catches. Did I do that? I thought I was being
careful enough. Well, it needs to be jumpered and soldered in any case.
That does it! The remote now puts out a strong IR signal - and operates the
TV (to the extent that this is possible - at least it clicks in exactly the
same way). I still have no idea of how the trace got broken or if I did it
though this would imply an impossible sequence of events as I did not open the
remote until *after* it did not work. But it did definitely fix the remote
problem. For good measure, I also put a dab of glue on the ceramic resonator
which is flapping in the breeze.
Now, back to the main event. Checking the Sams index is disappointing - this
model is not listed in the (1995) manual. However, a model with a very similar
number is. Well, asking for the folder doesn't cost anything. Examining
the schematics of this folder (for a RXX168-WA01) shows that there are
enough similarities to justify copying costs. In fact, it appears to be
virtually identical (I cannot actually locate any differences of any
consequence). Generally, on the first round, I copy the schematics and
any resistance charts and chip descriptions (block diagrams). Later, if
I need specific information like part numbers, I will copy the appropriate
additional pages.
Since there is voltage on the HOT, there is almost certainly a problem
with the startup drive. Checking the base of the HOT shows that there is
no drive present when power is turned on.
Checking the base of the horizontal driver transistor shows that there is
nothing there either.
How is startup drive derived?
Various models of TVs and monitors use different techniques: multivibrator,
deflection chip powered from standby supply, high value resistor supplies
current to horizontal driver or HOT directly, etc.
For this TV, a single chip is used for sync and deflection, a TDA8305 - TV
Signal Processor. However, after examining every pin on this chip, it is
apparent that there is no DC power available until after the set starts up
and the flyback supply is running. The supply voltages for the chip are all
13 V - rectified, filtered, and regulated off of a winding on the flyback.
This leaves only one possibility - there is a high value resistor between
the 130 V B+ and the base of the horizontal driver transistor.
The collector of the horizontal driver transistor has the expected +130 V B+
so the transistor is not turning on or is open (not likely).
First, I unsolder the base of the horizontal driver transistor so I can check
the junctions with an ohmmeter - it appears good. Better make sure to
reconnect that before applying power or else there will be 130 V on the IC!
Next, I test across R502 - startup resistor. It injects current into the
base of the horizontal driver transistor. This applies a pulse to the
HOT to initiate the flyback supply. And, what do you know! The resistor
tests open! At first I thought I was just using the wrong scale on my VOM.
However, it is indeed open - might as well be a pair of disconnected wires.
It is a 47 K ohm 1/2 W film resistor. A quick power calculation shows that
130 V across a 47 K ohm resistor is about .3 W - close enough to suspect that
simple heat dissipation and thermal cycling over time resulted in an isolated
failure. In fact, there is no possible mechanism that would result in a
forced failure of this resistor. Even if the 130 V supply was running high,
the maximum dissipation is just about 1/2 W.
A replacement - with a 1 W resistor this time - restores normal operation.
Not a bad set for $3 (+$2.40 in copying costs and a resistor).
Comments: These are the types of repairs that pay the rent for repair shops.
A quickly identified cause with an equally quick and definitive cure. A return
problem related to this repair is extremely unlikely. Realistically, a 15
minute repair from start to finish is possible.
This Sylvania TV is also nice to work on. Everything is on one circuit board,
components are clearly marked, and it is possible to access everything without
disconnecting more than the degauss and speaker connections (not needed for
testing in any case). Too bad that a 2 cent resistor was specified with a
marginal power rating. All too frequently, startup problems in TVs, monitors,
and other switchmode power supplies, are caused by open high value resistors.
Patient: Sony compact stereo system - AM/FM radio, turntable, cassette (at
least it is not an 8-track!!!). This is one of my charity repairs.
(actually not charity in this case but low income people who would
otherwise not get the stuff repaired at all.)
Symptoms: AM works fine but there is only low level static from FM. General
whacking and prodding produces no reaction.
Testing: Exercised all buttons and controls, no change. An antenna
seems to make little difference.
Getting to the **top** of the tuner circuit board is not too bad - just
remove the turntable, uh, record changer (you know what they are, right?).
Four screws and disconnect the motor and audio connectors.
However, even before that, I have to vacuum the thing! Why? Apparently,
in packing it up (in a pillowcase, no less), some powdered laundry detergent
or something similar was spilled into the thing. Yes, the adventures I
have! Even my shop vac has difficulty getting every last particle out but
for now, it will have to do.
With the turntable removed and set aside, the tuner and preamp board is
readily accessible.
The FM tuning capacitor and front end are easily identified. Touching
various points in this circuitry results in a large change in output but
no selectivity. There appears to be a strong station which peaks around
102 MHz but it is obvious there is no heterodyning action as there is
absolutely no selectivity. However, the front end is sensitive to my
touch and therefore it is likely that the following stages are good.
I was just about to give up and head for the library and possible Sams
when I decided to poke around a little more. My some quirk, I happened to
press on a coil in the front end box and instantly good reception returned -
as long as I kept my finger on that coil! Another bad connection? A wooden
stick worked just as well as my finger (I was sure relieved since I didn't
look forward to spending the rest of my existence holding onto that coil).
So, probably a bad connection. When operating, reception seems to be
absolutely perfect. Most probably the local oscillator was at fault.
Now, the fun begins.
Getting to the bottom of the circuit board is a whole lot more difficult
than getting to the top.
First, the back panel needs to be removed. Then several screws holding the
circuit board and a ground strap that needs to be unsoldered from a shield.
Now it may be moved, but is still attached to the front panel via the dial
cord. And you know how much I love to restring dial cords! For a while,
I thought the cord would stay put even when pivoting the circuit board up
enough to inspect and resolder but no such luck. With a thoroughly
disheartening 'pling' the dial cord popped off of its pulleys. Well, at
least the board is totally free and the bottom can be accessed.
I resolder all the pads in the vicinity of the coil since may of the
connections looked questionable (though there were really none that I
could say definitively were bad).
After flipping the board back over, reception now appears to be solid.
Pressing, prodding, and pounding have no effect. Will it be permanent?
I surely hope so!
Replacing the dial cord turns out to be easier than I had expected (unlike
that time where I melted it with my soldering iron!) The only question was
how many turns on the tuning knob shaft but there is obviously only one
possibility (3) based on the total length.
Checking out the rest of the unit:
The tape deck appears to be functioning, Fortunately, none of the soap
powder made its way into the tape mechanism.
The turntable, however, is a sad story. Soap powder everywhere. I remove
the platter to find - more soap powder everywhere. It is necessary to
disassemble, clean, and regrease the main platter bearing and changer gears.
Only then do I notice that the counterbalance weight for the tone arm is
missing and the stylus is bent and would need to be replaced. Checking
with the owner, it turns out that she doesn't play records (I can see why)
and therefore, as much as it pains me, I will not bother to restore the
turntable at this time (and probably never). It would be simple - I would
have to come up with a suitable substitute for the weight and order a stylus.
Finally, I check out the dial indicators. One is burned out and another is
intermittent (broken filament). The stereo light is burned out as well.
These will come from MCM. At first I thought that substituting light bulbs
I had in my stock would suffice but the cost of the proper replacements
was so low that it wouldn't be worth it and mine would not work as well.
Comments: The world it seems is full of these 25 year old compact stereo
systems. While audiophiles and snobs may look down their noses at them,
the majority of people are perfectly happy with their performance (or lack
thereof). Repairing one of these may not have the glamor of working on
a laserdisc or projection TV but this lack may be more than made up for
by the flood of enthusiastic appreciation from the owner - especially if
they would otherwise not have been able to afford a repair.
Problems with compact stereos are very often low tech - lubrication, dirt,
bad connections, dried up capacitors, burnt out light bulbs. A schematic
is rarely needed and access to the interior is usually straightforward (though
perhaps not as easy as one would expect as in this case).
Patient: My Mitsubishi HS-U53 HiFi VCR purchased new in 1990 (yes, hard
to believe I actually bought something new!). Pampered, used
mostly to tape Startrek and NOVA episodes. Not abused at all.
How can it do this to me?
Symptoms: Tape is accepted but attempting to play or even display a still
results in a garbage picture - tracking is way off. In addition,
play results in no tape movement and eject follows as if the VCR
finds the content distasteful.
Testing: I tried several tapes but it was obvious that something was very
wrong.
I enjoy the challenge of reviving a garage sale acquisition purchased for $1
but I really do not like having to fix something for which I paid real money!
Nonetheless, I have no choice. Actually, this VCR has always been a bit
finicky about how a tape is inserted. If pushed into the cassette slot
too slowly, it might grab the tape and spit it out. If the tape was
accepted, there was a chance that it would not seat correctly and never
advance. Therefore, I wasn't entirely surprised by the failure.
Despite all my advise about most problems being simple and mechanical, I was
naturally thinking that **my** problem would be something esoteric!
Once the cover is off, the extent of the problem can be seen. This VCR
uses a 'rapid access transport' which means that the tape is immediately
wrapped around the video heads in the fully loaded position except that the
pinch roller may not be fully engaged. Starting from this position is
indeed quite fast - less than a second to enter play or any other mode
except super fast rewind which unloads the tape back into the cassette.
The tape is in very light contact with the spinning drum but with no
tension applied - which winds down on its own after a few minutes to
conserve motor life (wear of the video heads would be negligible in this
state).
Anyhow, back to the story. Checking the tape position after loading
reveals that the pinch roller is not moving properly over and down as
it should. Further close examination reveals the reason: there should be
two wings on a spiral plastic cam gear which guides the pinch roller
movement. One of these has apparently broken off allowing it to lose
contact with its mate. (I did find the broken-off piece - I really like
to account for all parts so they cannot jam something at a later time.)
Fortunately, MCM Electronics has a repair kit for about $12 which includes
the entire assembly with a new pinch roller as well as one belt. I guess
this is a common problem. Inspecting the pinch roller reveals some fine
cracks (somewhat unusual for a pinch roller of this age) so a replacement is
worthwhile.
To be safe, I scribe timing marks on the 2 or 3 gears that might be
disturbed during the swap. This is always a wise precaution if no
service manual is available. As it turned out, this was not needed
as the new pinch roller assembly dropped right in without moving
anything.
Comments: Why does this sort of thing happen? Originally I thought that
Mitsubishi saved 2 cents on the plastic reinforcements and thus timed it
to fail after 4 years. Perhaps this is the case. However, the consensus
now appears to be that the lubrication of the shaft on which the pinch
roller assembly slides gums up or dries out resulting in binding and
subsequent breakage.
In either case, the end result is an annoying repair.
I did make sure to liberally lubricate the sliding shaft but fully expect the
problem to reoccur in a few years. Now, if I only would do the types of
preventive maintenance I recommend to others perhaps this could be avoided!
Patient: Mitsubishi FA2100-CW facsimile machine. Old but perfectly
usable technology. It should send and receive faxes. What
more could you want?
Symptoms: Receive seems to work properly. Attempting to send a fax results
in continual redials and the destination phone is never reached.
Testing: Initially, problem was not known. So I had to set up WINFAX on
my PC to provide a suitable source/sink at the other end.
See below. Once I unjammed the paper roll, receive seems to be
fine. Attempting to send a FAX, however, results in repeated
attempts to dial without success. Local copy function works
correctly.
This fax machine was apparently acquired by my 'customer' as a result
of the down-sizing of a local hospital. There is a user's guide but no
installation or setup info. Well, better assume all the DIP switches
are set correctly. Write them down just in case I should accidentally
move one without realizing it (I actually did not do this and was lucky.
However, it is a good idea).
At first I thought that the fax machine was never connecting to the phone
line and I tried to identify which of the 4 internal relays was supposed
to do this. Using a multimeter, the one whose coil was being activated
was easily determined but there was nothing measurable across its contacts.
Monitoring the phone like with a DMM showed, however, that it was picking
up - the 50 V on-hook voltage drops to something like 5 V when dialing.
Further testing by monitoring the phone line both with the fax machine's
phone and a separate extension reveals that the dial tone never goes away
and the tones seem to be somewhat weak though apparently of the correct
DTMF frequencies. The fax machine's built in phone does dial correctly,
however. Interesting.
Time for some inspection. Disassembly is actually fairly easy - 2 screws to
remove the bottom panel, 5 screws to remove the circuit board and a heat sink.
Only complication is the roughly 10 cables that need to be disconnected to
free the logic board from its wiring harness.
Oops, what is this? The backup battery is hanging by one lead. Apparently,
this had been replaced relatively recently (I really have no way of knowing)
and whoever did it was too lazy to access the bottom of the circuit board.
Therefore, they attempted to solder in the new pack by heating the leads from
the top instead of the pads. It may have worked long enough to close up the
unit but was certainly not a long term repair. The other lead is somewhat
loose as well.
While I don't expect this to fix the main problem, it is dealt with first.
Additional wires need to be solder to the flimsy tabs on the battery pack
since they break off as soon as I attempt to reinstall it in the circuit
board holes. Some silicone sealer is now used to secure the pack to the
board.
I reinstall the board with all 10 of its cables to confirm that **the**
problem still exists and that it will now hold its time and other memory
settings.
Now the machine is continuously trying to eject the input copy. What is
going on? After some minor panic and confirmation that all connectors are
indeed installed correctly, I realize that I am trying to run it on its
side and there are some interlock switches that are gravity dependent.
Setting the unit upright makes it a whole lot happier.
It now appears to hold its settings but as expected, will still not dial
for fax transmission.
Out comes the board again.
So, what next? Tracing the circuit might be possible - it is a double sided
board - but what about my usual approach - test all the semiconductors and
capacitors in the area of the phone line circuitry. This has proven successful
on more than one occasion (see Repair Brief #2 - USR Data Modem Won't Dial",
for example).
And, guess what? I found a little black diode that tested *open*. Now, that
is unusual. What type? The part number is not listed in any of my databooks
and cross references. Well, if it walks and talks like a 1N400x type, that is
what I will try. The symbol on the circuit board is just a diode so I assume
it is not something fancy like a zener or Tranzorb surge protector. So I
put in one of my lonely looking 1N4007s - I dare you to blow! Hopefully,
this wasn't supposed to be some kind of high speed or high efficiency type.
All other components that I can reasonably identify test out at least
in-circuit with my DMM.
Reinstalling the logic board with all 10 cables is now pretty quick. I have
the technique perfected!
So, now for the test.
Mode: Manual Dial. Telephone number: My PC's modem number.
Uh oh, behavior is the same. I am just about to scream or beat something
over the CRT when I pick up my extension phone and hear: "Please touch 1
for sales, touch 2 for service.....". Huh? My PC does not have a voice
mail system. It dialed a wrong number!!! Actually, I told it to dial a
wrong number. After all, how often would I need to dial *my* PC? I didn't
even remember my own phone number correctly.
OK, this will work. After transposing a couple of digits, I get it to
dial the correct number but now WINFAX is not picking up. Not surprising
as I never configured it for receive.
Finally, I am able to send a fax to my PC.
The customer will need to re-enter the other stored info like rapid dial
numbers, their fax number, header and footers, etc., but it would seem that
the fax machine is now functional.
Now where did I put the fax number of that pizza place?
Comments: This is yet another example of a simple problem bringing down
sophisticated technology. I would have had no chance of successfully
troubleshooting anything in the fax logic. However, phone line circuitry
problems are quite common - especially after the storms we had the previous
week. (I have a VCR coming in which failed during the storm as well).
Without the circuit diagram, I have no way of knowing if this was, indeed,
storm related damage but the probably is quite high that it was.
Patient: A former colleague's Mitsubishi HS3381UR VCR. (He now works in
California and we are near Philadelphia. However, he visits
frequently. It worked fine before the move (except for the
occasional spasms of B/W only playback but that is another story),
he says.
Symptoms: Tape plays but the bottom of the picture is snow and the remainder
of the picture has an appearance similar to CUE (forward search)
mode - multiple tracking noise bands are visible. However, the
tape is clearly playing at normal speed as the audio is is normal.
Testing: Confirmed with several tapes. Same symptoms. CUE and REV result
in similarly confused pictures.
The picture is broken up. The top two thirds has the same exact appearance as
it would in search mode (CUE or REV) - 3 or 4 bands of noise with good
video in between. The bottom third is snow. However, the sound is normal
(but no HiFi sound) so the speed is normal and this is strictly a tracking
problem. Once you have seen this set of symptoms a couple of times, the
diagnosis takes about 3 milliseconds.
We did try to help diagnose the problem long distance. Unfortunately,
his description of the picture was not clear enough or the email connection
was bad - or something. This turned out to be hopeless. His interest in
repairing things usually ends with changing batteries in the remote control
anyhow. I am just glad he didn't cause additional problems by using sandpaper
on the video heads or attempting to straighten the tilted guideposts with a
pair of pliers!
So, next time in town, the VCR is dumped in our hardware lab....
It rattles! And, this is not even a JVC!
Upon removing the cover, the problem is obvious: the right side
tilted guidepost has fallen out. It was probably fine as long as the
VCR was in an upright position. However, the baboons who packed up his
apartment probably had it sitting upside-down for the move so the guidepost
worked loose. This is a 5 minute fix using 5 minute Epoxy using just the
smallest amount - a dab on either side - then pressed fully into place
and rotated if possible to distribute the adhesive (you don't want a
continuous coating as trapped air in the cavity may cause the post to pop
out before the Epoxy cures.) Tape path alignment should not be needed.
Comments: It seems as though older Mitsubishis are virtually guaranteed to
have at least one loose guidepost. I have seen 3 or 4 in rapid succession.
These often don't cause problems until the VCR is moved but the symptoms
are quite obvious once you have seen a couple (even if you don't shake
the VCR!).
Mitsubishis, JVCs, who else?
Patient: Panasonic PV2812 VCR used for a lot of time shifting. The owner
is apparently quite eager - desperate would be a better term - to
have it back.
Symptoms: Play results in snow and shutdown in a few seconds. Record also
results in shutdown. FF and REW work normally.
Testing: Multiple tapes result in identical behavior - just snow in play
and then shutdown.
This is one of those VCRs with almost no controls on the front panel.
Therefore, testing almost anything beyond play, record, and channel selection
requires a remote control. I didn't get the remote control so I had to use
one from another older Panasonic VCR. Some of the buttons behaved a bit,
well, strangely. At least the basic functions seem to be standard.
With the top off, the problem is quickly identified. This model uses a
tape transport that loads the tape half-way as soon as a cassette is
inserted. A plastic lever called a 'sector gear' drives a metal lever
with a vertical post called a 'control arm' which is supposed to pull
the tape out of the cassette. (Thanks to Frank Fendley (Studio Sound
Service, frank.fendley@datacom.iglou.com) for providing the correct
terminology for these parts and for identifying the specific part numbers
so I could obtain the necessary replacement).
The first observation is that the tape does not move after play or record
are pressed and the microcontroller is shutting down due to lack of takeup
reel rotation.
However, the underlying cause is not a common idler or belt but failure of
the control arm to pull the tape beyond the pinch roller. Thus, in
play, the pinch roller comes down and presses against the capstan but
there is no tape there! Without the rotating capstan to pull the tape
through, it just sits there. Without control pulses, I guess the blue
screen circuitry kicks in and there is no picture. Eventually, the lack
of takeup reel rotation results in a shutdown to the stop position.
So, what can prevent the control arm from doing its thing? If this
were a late model Sony, it would be dried up lubrication but in this case
the arm swings quite freely.
However, it appears to get hung up just short of the pinch roller. How does
this work? A spiral multifunction cam gear that looks somewhat like a
miniature washing machine agitator (supposedly called a 'Pressure Roller
Lift Cam' by Philips who apparently invented the technique) is used to:
1. rotate the mode switch via gear teeth around its bottom edge,
2. move the control arm back and forth using a cam follower lever called
a sector gear whose end presses against a cam affair near the bottom.
3. move the pinch roller up and down and press it against the capstan
using a stirrup type cam follower which starts at the top and follows
the spiral.
The mode switch appears to be fine and properly timed - in the tape half
loaded position, the timing marks line up precisely and in the tape-out
position, its timing mark lines up with a notch on the plastic case. One
can only assume that there was some intelligence at work here!
The pinch roller appears to do exactly what it should coming down and
then pressing against the capstan - which does rotate properly during
the time it is supposed to be playing.
Therefore, we are left with the control arm itself. This is driven by a
cam on the outside of the main gear but not directly - there is another
lever - the sector gear - which actually contacts the cam. This drives
the control arm via a set of 6 gear teeth. As a result of the way the two
levers are pinned, a slight change in cam position results in a large
movement of the control arm.
Except that it is getting hung up on something.
Unfortunately, it is almost impossible to view the goings on with everything
in place. However, I really don't want to take anything off that (1) might
require retiming or (2) have circuit boards hanging by multiple cables (as
would be required for this series of Panasonic and clone VCRs).
What appears to be happening is that the actual control arm (which is
made of metal) is getting hung up. The sector gear is made of plastic
and - guess what! - has a crack!!! It isn't a break and not even much of
a crack but will need replacement.
However, is the crack the cause or the result of the current problems?
It is just a small crack so I assume that the overall behavior will not
be as grossly affected as it seems to be. Could the crack have been caused
by some other problem?
Mechanically, there really isn't much to this. The only way it could not
work would be for the control arm to have changed vertical height somehow
or bent, or for there to be some obstruction or gummed up lubrication
preventing full movement.
It seems as though a projection on the metal control arm itself is
supposed to fit into and clear the cam as it pivots. Imagine: Cam moves
lever #1 (the sector gear); lever #1 meshes with the gear teeth of lever #2
(the metal one with the control arm post); and a projection on lever #2
then fits into a cavity on the same cam gear. Thus, this cannot be
a timing or electronic problem as the mechanics are totally self contained
and only depend on the relationships of the two levers. The gear teeth
that mesh them are timed properly - the little timing marks align. But the
projection just stubbornly keeps hanging against the cam.
I did remove the metal lever completely to check for cracks or bending - there
was no evidence of either of these conditions. I also tried raising and
lowering it a carefully recorded amount - no change except that when raised,
it hit the pinch roller before getting stuck so it was obviously not
too low to begin with.
Finally, I decide to remove the pinch roller assembly - this seems pretty
harmless as it is fully up and everything should operate fine with my
cassette cheater, pinch roller or not. This gives a clearer view of
what is going on.
It appears as though there is considerable stress on the plastic lever as
the crack widens significantly during part of its travel. However, upon
unloading, the pinch roller does move fully as it should. Interesting.
Careful observation now reveals that the projection must be entering a
cavity in the cam which is not visible and the cause is now likely to be
that crack in the plastic lever.
When loading, there is enough stress to deform the lever and due to the
mechanical amplification of the lever system, shifts the point of entry
of the projection just enough for it to miss the cavity every time.
However, during unload, there is less stress and thus it pops in place
for at least part of the way.
With a little less spring torque on the control arm, operation seems to
be reliable even with the bad plastic part. Thus, the crack was the cause
and not the effect - which is comforting as I was not looking forward to
a convoluted failure!
Once the replacement arrived and was installed, I confirmed the timing (I just
avoided ruining my whole day as the agitator thing tried to jump up but
I calmed it down), and then lubricated some key points as well. The VCR
was then whisked back to its expectant owner.
Comments: Once again, a very simple common problem like a cracked plastic
part can result in symptoms that are confusing. This would have been a lot
easier if there was better access to the mechanism. Unfortunately, this
series of Panasonic VCRs puts the main circuit board on top and while not
actually covering the cam gear/pinch roller/loading arm certainly gets in
the way. I don't like removing it because there are a bunch of cables that
never seem to go back just the way they should - at least not without some
persuasion.
Unlike the Mitsubishi broken cam follower problem (Repair Brief 40: Mitsubishi
HS-U53 - Bad Tracking) which bears some similarity, there doesn't appear
to be any real cause for this failure except just normal wear and tear
and perhaps too flimsy construction of the plastic sector gear. The design
with a projection that must enter a recess seems like some sort of interlock
to guarantee that the tape is fully beyond the pinch roller before it comes
down and would possible crunch it. In all fairness, there must have been
some merit to this design. At no time was there any damage to the tape
despite such a fundamental failure of the tape loading mechanism.
However, once a crack develops, it is likely to get worse as the stress
is a lot greater on the part when that projection cannot enter its cavity
in the cam gear.
Patient: A colleague's JVC HRD-550U 2 head VCR. I was actually expecting
this job about 4 months ago but being his third VCR, apparently
wasn't an acute emergency case.
Symptoms: Since this is a JVC, I don't even bother to plug it in before
taking off the top cover and jigging the guide posts.
Testing: Jiggling the roller guide assemblies reveals that the right hand one
is loose. Removing the bottom cover produces the missing brass pin.
I cycle the mechanism half way and pull the plug to gain access to the roller
guide assemblies. It is important to take care at this point to prevent the
relatively unconstrained roller guide from flopping up and hitting the spinning
video heads - a relatively expensive lesson.
So far so good, a drop of Epoxy and that guide post is better than new.
Pop in a tape and...No picture. This was supposed to be straightforward. :-(
Both roller guides are seating properly but - wait - guess what? A missing
tilted guide post!
Two dropped parts in one JVC at the same time! Fabulous.
Sure enough, the post is sitting quietly on the workbench minding its own
business not caring one wit for being absent from work.
Break out the Epoxy once again. At least this post stays put once pressed
into place (unlike the Mitsubishis that have the same problem but the cavity
in which the posts are inserted are so closely machined that they trap air
and the posts keep wanting to pop back out.)
Try #3: Ouch! The tape is grabbed on the wrong side of the roller guide.
After carefully extracting the cassette, it is obvious that the roller guides
are not fully retracting into the cassette. Thus, when the cassette drops,
sometimes the tape is in front of the one of the posts. How can that be?
In addition, sometimes the tape would not load, whirring motors, and it
would give up and shut down.
The repairs look fine - brass post snug against the shoulder and nothing
to catch on anyhow. What about the other one.
Now some history of this machine. It dropped the plastic pin on the opposite
side roller guide linkage a couple of years ago. I replaced this and
reinforced it with a tiny screw. The screw is quite secure but it appears
to be this roller guide that is getting hung up, but on what?
At first, I thought that a plastic projection on the underside was hitting
a sliding widget (great name, huh?) which it seemed to not quite miss as
it should. When in the fully unloaded position, the projection on this
widget keeps the roller guides in the retracted position. Apparently, it
is doing it job too well and preventing it from retracting in the first place!
Carefully bending the plastic which allows the roller guide to pop home
results in correct loading but failed retraction the next time EJECT
is pushed.
Great! I have a VCR that can be loaded at the factory with one movie
and will work fine as long as the cassette is not ejected!
So, is there a timing problem - something off by one tooth - or something
else? The only thing unusual about that roller guide assembly is my
reinforcing screw. Remove it!
And, sure enough, now loading and unloading is flawless. Apparently,
the head of the screw added just enough height above the roller guide
assembly to catch on the metal of the cassette basket in the down position.
Lesson: when you reinforce the roller guide hinge, use a very thin headed
screw.
I have no idea why this didn't show up after the first repair or maybe it
has been a problem all along and the owner never realized it. Some careful
filing leaving just enough head slot for a jeweler's screwdriver blade and
everything appears to be happy.
Comments: This single VCR has has now had 3 of the 6 possible common
dropped parts - drop off. In a professional service situation, it would
really be prudent to head off the inevitable and reinforce or glue the
others. However, they seem as secure as when brand new so I leave them
alone. Perhaps, I just need material for a future Repair Brief!
Patient: Well worn SEARS Model 580.53471750. This is a Goldstar chassis.
Symptoms: Cassette platform is loose indicating a broken part. The owner
apparently had someone extract the cassette - cover screws were
missing (again - this is not the first such instance).
Testing: Cassette jams as soon as it is inserted. Fortunately, no additional
damage is done.
Well, removing the cover is easy at least. Now, is the broken basket
assembly the cause or the effect?
To remove the cassette basket assembly requires taking off the front panel
(3 screws), video head cover sheetmetal (3 screws), and then 6 screws to
actually unfasten the unit itself. The complete assembly can then be unplugged
and removed to the convenience of my workbench.
And, hey, what do you know? There is a transverse shaft which keeps the
two sides in sync - it drives the left-hand side from a motor and gear reducer
on the right-hand side. The little right-hand gear is - missing! Not just
fallen off but gone. Apparently, whoever extracted the cassette did a little
'clean-up'. No, it didn't fall off, it is **gone**.
Thus, I need one gear. Better make that two gears - its left-hand mate appears
to have a fine crack just waiting to spread.
Frank Fendley of Studio Sound Service identified the part numbers - at first
thinking I wanted the large drive gear which has a couple of projections and
a spring. This mistake, however, got me to looking at that part and noticed
that a plastic post had developed a crack and *was* in the process of
breaking, so add one of those to the list.
Hopefully, there is nothing else wrong with the VCR that caused the missing
gear to crack in half. However, without the basket in place, it is kind of
difficult to be sure. I probably could have nursed it through the cycle
even with only the one gear in place but I have confidence that this is
indeed the main problem.
The parts arrive 2 days later. The two small gears fit perfectly but the
large gear is apparently a slightly different revision and I need to ream
out a section of the hole to accommodate a shoulder on the mounting shaft.
I, of course, violate my Rule #1 - mark everything before removal and a
couple of email messages back and forth are needed to get the timing adjusted
properly.
After a little lubrication with plastic safe grease and reassembly, normal
functions appear to be restored.
When I returned the VCR to its owner, she commented: "So, that little plastic
piece was the problem?". Thanks. No user serviceable parts....
Comments: The lesson of this story (aside from not messing up the timing)
is to always check related components - mechanical as well as electrical -
for possible stress or pending failures. This applies equally to capacitors
in a power supply and plastic gears in the cassette loader.
Patient: Beckman portable DMM Model 310B I acquired because - you guessed it -
it was broken.
Symptoms: Most ranges result in a random display alternating between large
arbitrary positive and negative numbers. There do appear to be
differences depending on mode - DC V vs. Ohms but not in any
decipherable way.
Testing: Changing battery, applying various inputs, whacking it, using the
DMM to hammer nails - no change. Though it might have been my
imagination, the character of the screwups may have changed after
pounding some 20 penny spikes!
I had taken a look at this multimeter to attempt to repair it a couple of
years ago. At that time, I completely (so I thought) disassembled it to
locate any standard chips hoping that if I could get pinouts, I could
determine what was wrong. There were none except for a couple CMOS CD parts -
one in the clock generator and the other in the piezo buzzer circuit.
Everything else was custom and what was worse was that I could not even locate
the main A/D and display chip! It had to be there somewhere. since the 2 or
3 custom 16 pin chips I could find were not large enough and not in the
proper place to perform these key functions (they turned out to be precision
resistor networks - I think).
So, for two years, this thing sat on top of Jim's filing cabinet gathering
dust.
Finally, with nothing better to do (right, I can hear you saying: "sure,
likely story") I set out to solve the mystery.
After pulling off the LCD panel, cleaning its contacts, and reinstalling - with
no change in behavior - I noticed a 'wart' on the backside of the PCB board
behind the display. What's that? It has no business being there unless it
were hiding something. This 'wart' was a couple of molded plastic pieces
that with 20-20 hindsight was obviously clamping something in place.
Prying off the first plastic piece revealed - a second one.
Prying off the second plastic piece revealed - a little circuit card with
a blob of Epoxy - chip-on-board - in the center and 40 or so gold pads
on the long edges. Ah ha! The A/D and display chip at last. Of course,
a lot of help this will be if I need a replacement - no markings of any kind.
Connections were via those aligned rubber strips LCD watch manufacturers
are so fond of - and which I always suspect for bad connections.
So, I clean the chip-board contacts and reinstall. No significant change.
OK, maybe a little deeper. Next, I remove the chip-board, its mounting, and
the two rubber connector strips and clean everything with cotton swabs and
alcohol.
Now, finally, a change. The display looks to be half way normal. In fact,
at that point I thought it was entirely fixed. Checking with a handy .1 %
resistor that just happened to fall out of the sky, the ohms scale seems
to be dead-on. Hurray, but premature.
At this point, I go about cleaning up the case thinking that this was pretty
easy.
Then, I tried it on a 1.5 V Alkaline battery and got - 3.43 V. Huh? Did
I pick up an Lithium by accident? Nope.
My AC line measures around 250 V according to the meter. Since none of
my VCRs have melted down, perhaps, there is still a problem.
Well, maybe there are still some bad connections. So I go through the
entire exercise of disassembling the chip-board once again and double check
all pads. Reassembling results in no change and jiggling, pressing, and
other wise fondling the display/IC area doesn't alter anything.
What next? How can ohms be perfect and all other ranges (I also tested
DC current and Diode check) be off by more than a factor of 2?
As noted, this thing has a chip-on-board IC - unmarked - for the A/D and
display driver and custom Beckman parts for key resistor networks. Thus, it
is virtually impossible to determine any circuit details (not to mention the
32 pole rotary selector which cannot be used with the unit disassembled for
testing).
There is one pot - full scale adjust - which has about a 10 % range end-to-end
and affects everything but resistance. Perhaps, resistance is calculated
by performing a comparison with a known resistance. Then, they would both
be off by the same amount. Maybe I could bugger this pot to drop it down
but I would really like to determine the root cause.
Modifying the adjustment pot circuit turns out to be impossible since it
is connected to the above mentioned unidentified resistor networks and any
attempt would be a shot in the dark.
So, now it works fine as long as I multiply all my (non-resistance) readings
by about .4. I thought I was able to 'fix' DC volts by installing a 1.3 M ohm
resistor across a point found by trial and error. However, this did not
help AC volts or any of the current scales. And, apparently it messed up
the offset as + and - readings seem to differ by an unacceptable amount.
For awhile (like a overnight), I just let the problem bounce around.
Suspecting that the reference voltage was incorrect - low resulting in
the higher than correct readings, I searched for anything that might be
a voltage reference. There were a bunch of transistor-like things but one
stood out since while it had three legs, two of these were shorted together.
Jumpering across the unconnected pins with a high value resistor resulted in
the readings increasing. Measuring across it with a working multimeter read
.54 V. Disconnecting the part entirely resulted in readings that were now
**low** by a factor of about 2.5. Since the present part at .54 V makes it
read too high by a factor of just over 2, a reference of 1.2 V should be
about right. I knew I had saved that bag of 1.2 V zeners for a reason!
Sure enough, putting a 1.2 V zener in place of the unidentified part results
in nearly perfect accuracy on all ranges I tested - DCV, ACV, DC mA and Diode
Check. No doubt it could be tweaked now if I had some voltage reference that
I trusted. And, Ohms is still fine (actually probably better since the higher
reference will increase resolution).
So, I still need to identify the reference part: HCC 8069 J010?
It is a TO-92 but doesn't show up in my ECG or SK databooks.
After posting symptoms to sci.electronics.repair, I did get some useful
comments:
(From: Kevin Carney (carneyke@watson.ibm.com)).
"I've repaired many of the 300 series over the past ten years or so. The
problem described sound like one of the finger contacts driven by the cam
wheel is staying closed or not making contact. If this meter has these
type of contacts it may be your problem. Solving it requires patience taking
apart bending contacts putting together and testing. Good luck, Kevin."
As noted, it does not appear to have been bad contacts in this case though
this info is certainly worthwhile to keep in mind for the future. The unit
had very little use and except for the one contact I bent putting it back
together :-( (and subsequently had to unbend) the selector switches (all 32
sets of contacts) appear to be in pretty good condition.
Then, one Netter who prefers to remain anonymous replied:
"I have the part in stock. I will snail mail it to you if you send me your
address. The Beckman part number listed in the manual is FG3000-231-102.
It is a 1.2 Volt Zener in a TO-92 package. The markings on the part are
LT1004 CZ-1.2."
So I guessed correctly. My 1.2 V zener is marked LM305, BZ1-2, K002. I do
not know if it is equivalent to the specified part so I accept this generous
offer.
Comments: Test equipment is the stuff one depends on to troubleshoot other
broken equipment. Therefore, it is disheartening in some ways to have to
repair a DMM. To find multiple problems on a lightly used instrument is
also disappointing. I suppose it is possible that my prior poking around
searching for the random display problem caused the zener to fail but what
an unlikely part to damage! I am not entirely sure the bad connections
explanation is entirely correct or at least that it is a permanent repair
but I will keep my eye on that.
Now all I need is a voltage standard to set its calibration.
Patient: Bill's brother-in-law's 17 inch color TV.
Symptoms: Only the channel numbers work. In all other respects, it is a
paper weight - and a big one at that.
Testing: No change when run up on a Variac. It remains dead.
So, Bill gives me a call at 7:30 in the evening: "I have this TV, about 13
inches. Only the channel numbers work, nothing else. Can I drop it off?"
No, Bill, if you want to come and help to troubleshoot it, that is fine, pick
a time. Otherwise, hold on to it. You can clutter up your basement. "OK,
I will see if I can get dad to watch Bobby". Any excuse to get away from the
kid. Five minutes later: "How about a half hour?" Sure, fine.
About 45 minutes later (30 minutes of Bill time), he comes lugging this
huge TV down the basement stairs. At first, I thought it was at least
19 inches but measuring the CRT, 17 is about right. Well, at least it
*should* be easy to work on - lots of empty space.
Plug it in and just as expected, only the channel numbers work. No picture,
no sound, no static, no deflection whine. The thing has push button controls
so I guess the standby power is ok. Time to check the power supply and HOT.
Eight screws later, the back comes off cleanly. There is a nice 3" x 4" parts
layout diagram, still legible, so that is a help.
Check the fuses. One is black. Bill, which is the fuse closest to the back?
"Uh, looks like the B+".
Check the horizontal output transistor (HOT) with my trusty 25 year old
Lafayette VOM. Case (collector) to other pins: 0 ohms. OK, bad HOT.
What else?
Unfortunately, they saved a couple cents and soldered the transistor rather
than using a socket. Bill, write this down: The yellow wire is on the right.
"Done."
Removing the bad transistor is easy though we manage to lose one of the
lockwashers. Well, as long as it is not on the circuit board somewhere.
I go into my HOT drawer and pull out a 2SD871. Nah, that is too good a part
for testing *this* set, how about a BU208A? I won't be too unhappy if one of
these blows. (Note: I have no intention of leaving either in place once
the set works. I will order the proper replacement, a 2SD950.)
Installation is equally smooth with the yellow wire on the right. These is
also a little insulating sleeve which I swear went on the yellow wire pin
but Bill thinks went on the other. Oh well, I will put an insulator on both
later on. Replace the fuse.
I go get the Variac and series light bulb widget and plug these in to my
isolation transformer rig.
Starting with a 100 W light bulb.
Bring up the voltage until about 100 V - channel indicator lit - push ON
button. Light bulb flashes brightly and set shuts off.
OK, Bill, unplug the degauss coil from the mainboard.
"What? Huh? Where is it?"
Locate the coil and work backwards. Here, this connector....
Try again. Similar result. Up voltage to 120. Now, it stays on, not
to full brightness but more than I would expect if the set were working
properly.
There is still no picture, no sound, no static on the screen. I think I
can hear the deflection whine, however. The HOT is holding.
Try a larger light bulb!
I put in a 150 W outdoor flood. Try again. Turn Variac to 130 V (or whatever
its maximum is).
There is still no picture, no sound, no static on the screen. I definitely
can hear the deflection whine. The HOT is still holding.
Then, Bill says: "Shut it off. Shut if off. I see smoke...."
Sure enough, there is a wisp of smoke coming from the vicinity of the flyback.
It stops too quickly to determine the source. There is a small electrolytic
there as well and the smoke could be originating elsewhere. The odor could
be burning plastic or burning electrolytic.
Now for the real smoke test! I go get a foot square piece of Plexiglas
that was once part of a homemade HV capacitor to use as a shield if something
should decide to blow up.
Power it up again. After about 20 seconds, the smoke appears but it is
still not possible to determine if it is the cap, flyback, or something
else.
Well, change the cap. It is a 1 uF, 160 V electrolytic. I will show it!
I put in a .8 uF, 200 V polyester type. No change - still smokes. This
eliminates the capacitor. Bill is sure it is coming from the flyback at
this point anyhow.
Bill is getting ichy as he knows his dad won't be able to take much more of
Bobby. So I suggest that we pull the flyback and I will test it later. The
big pins are a pain but attempting to discharge the CRT HV under the suction
cup thing yields nothing as expected.
Bill is about to leave. Hey, how about buttoning this baby up so it is not
cluttering up my workbench. "OK," Just a couple of screws, put the other
stuff in a plastic bag. "I think the flyback is bad."
Next day:
I get out my flyback tester widget - the 12 V chopper - and hook it up.
Ten turns of wire around the core connected to its output. If the flyback
is good, this should excite it to produce 8 or 10 KV with only a small
load on the power supply.
I locate the HV return on the flyback by turning up the juice just enough
so it is oscillating - barely. Measuring on the 5000 V scale of my Simpson
260 shows a couple KV between the CRT HV connector and only one pin on the
base of the flyback. That must be it. The others show zero volts.
Now, turning up the input power to my normal 'full' results in a nice 1/4
inch arc between the HV output and the return. Is the flyback good? Just
then, an arc develops between the return pin and its neighbor. What is this?
Kill power. Did I pick the wrong return pin?
I drag out the TV chassis to confirm that the pin I selected was indeed
the HV return. It was. The other pin is a winding for one of the auxiliary
outputs of the flyback.
So, what is going on? A quick check with the multimeter solves the mystery:
All resistance measurements are reasonable except one - and a fundamental one
at that. I measure 1.76 K ohms between the CRT HV connector and three pins
on the base of the flyback - and one of those is the pin that arced.
Therefore, this flyback is history. Apparently, internal breakdown between
the output of the HV rectifier or multiplier and the low voltage windings
resulted in destruction of the original HOT and blown fuse. This is a pretty
spectacular failure mode!
I call Bill on the phone: flyback is bad as expected. Short, CRT HV to pins
on base. "I already told him it probably wasn't worth fixing".
The flyback is definitely bad but could it have damaged other circuitry
in the TV? After all, until the HOT popped, that HV was arcing internally
to the low voltage windings. I suspect that it did not cause any damage,
however. Despite the fact that the HV output is on the CRT capacitance and
could really zap something, it is likely that the short didn't develop
suddenly but over at least enough time for the capacitor to discharge
harmlessly. The actually current available from the flyback HV output
is quite small and the low voltage flyback windings have almost no
resistance to ground.
We will probably never know since Bill talked the 'customer' out of spending
the money for the expected repair parts. The HOT would probably be around $5
and the flyback is $26 from Dalbani. Add shipping and this will likely come
to over $35 for a 13 year old TV. And, there would still be a chance that
it would not work at all without additional parts or have long term
reliability problems.
So, if anyone has a good TLF14617F flyback they want to sell cheap....
Otherwise, the set goes to the great TV spare parts graveyard in the sky (or
actually in the attic of my garage).
Comments: I promised that not all these stories would have entirely
happy endings. This is a case where determining the extent of the
damage and cutting losses made the most sense. If I come across a
replacement flyback (sure, like they are all standardized!) then it
will be worth the effort to see if that is all that is needed.
Patient: Well worn SEARS Model 580.53471750. This is the continuing saga
of the VCR described in Repair Brief #45: "SEARS (Goldstar) VCR
Part 1: Broken Cassette Loader".
Symptoms: VCR will enter Play and Record successfully but then shut down in
anywhere from 10 seconds to 10 minutes.
Testing: I was able to confirm these symptoms and it didn't take long, about
2 minutes the first time, 10 seconds the next. I tried several
tapes without any apparent change.
It was only two weeks since returning this unit to the 'customer' after
repairing the cassette loader. These are classic symptoms of dead rubber
but I knew the rubber parts were fine having replaced them not so long ago.
And, observing the behavior as it shut down, there was never a problem with
spilled tape or a weak takeup reel. This is confirmed by testing the takeup
reel torque using a cassette cheater (shell).
The reel rotation sensors would be the next natural suspects. What about them?
Well, there is only one, for the takeup reel. The tape counter does increment
but on closer examination, there does appear to be some weirdness. Instead
of counting 0000, 0001, 0002, 0003, etc. It occasionally skips counts. So
the sequence might be: 0000, 0001, 0003, 0005, 0006.... I don't know how
the microcontroller determines that the tape is moving but it might just
test the least significant bit periodically. If the counter is skipping
counts - say only doing odd numbers for a while - this could end up not
changing for too long.
So, now I start playing with the takeup reel with the VCR in STOP mode.
Fortunately, the counter is active and I can simply twirl the takeup spindle
to my heart's content. It doesn't take long to realize that IT is behaving
strangely. If I go very slowly, there will be times when the counter display
will free-run, counting rapidly and continuously. This is, in fact, most
likely what is happening - the reel is turning slowly enough that as it
passes through these 'bad' areas, the counter skips a count or two. Why?
Putting a scope on the sensor signal doesn't reveal anything amiss - it
changes smoothly from low to high and back again. However, when it is
approximately in the middle of its range, the counter does its free-run
thing.
I trace the circuit to a buffer on a little circuit board tucked in a
very inconvenient spot needing to extricate it from several connectors.
Examining this circuit shows that it is, well, just a buffer. No hysteresis
or debounce. Its output is changing as expected. Tracing the output reveals
that it is going to a large multilegged creature - the main system controller
chip. So, if that does the debounce, I am kind of out of luck as I am not
going to invest in a new microcontroller.
Well, I will show it! I build a little widget board with a single transistor
(2N3904) and a couple resistors as an additional buffer and put this in series
with the original signal. I then add a feedback resistor from the output of
both buffers to the input of the first one. This adds just enough hysteresis
to prevent the circuit from even likely lingering in the bad area.
At first, this works like a champ but then the sensor seemed to be losing
sensitivity and its output refused to go low! Could a dying sensor be the
entire problem? No, I don't think so. That free-running count problem would
still exist even if the transitions were sharper - it might be unlikely during
normal tape movement but could still happen if the tape stopped at a just the
wrong spot.
Removing the sensor under the takeup reel requires popping the split washer
and pulling the spindle - taking care not to lose the washers under the
spindle. It could be worse.
The sensor is a roughly 3/16 inch diameter affair with an LED and photodiode
pointing at a four quadrant aluminum reflecting pattern on the bottom of the
spindle. Everything is clean and undamaged so perhaps the LED or photodiode
is in the process of failing. It doesn't really matter which is bad as failure
of either would render the device useless.
What to do? My usual places like MCM Electronics do not list a replacement
and I really wouldn't have a minimum order anyhow. I check with Frank
Fendley - he has it - but I cannot justify the $5 S&H for a $2 part.
I will make one from the guts of an optoisolator! I at least want to do
this to further test my theory (though at this point I am nearly certain
that the sensor is the **final** problem).
I totally destroy 2 optoisolators in the process but finally extract both
an LED and photodiode intact. Fortunately, my friend Bill has bags of these
left over from his switchmode power supply design days.
The circuit board is marked so it is easy determine which leads are the LED
and which are the photodiode. It isn't pretty but with a bit of filing and
other manipulation, using the correct chants, etc., it finally is positioned
to have enough sensitivity to activate the counter. Unfortunately, the top
of the LED rubs on the bottom of the spindle during rewind slightly scraping
the reflecting pattern but not enough to affect anything. A bit more filing
and a key incantation and it seems to be solid. The response does not
quite seem to be as sensitive as I would like but operation appears to be
consistent, reliable, and repeatable from one end of a T120 tape to the other.
In fact, I ran the VCR for about 20 hours in Play and Rec, end-to-end of
tape, and at SP and EP speeds without any apparent problems. I know I can
obtain the replacement part but for now my kludge is just fine, thank you.
The @#$% VCR will probably now die in a couple months in some other
interesting and creative manner.
Comments: This is my VCR from hell. It seems to have problems with multiple
unrelated failures at nearly the same time. I have already had to replace
the video heads (due probably to just plain wear and tear) and also suspect
a bad connection in the video circuitry as well. Sometimes, I have simply
needed to clean the video heads. In addition, the 'customer' insists on
using old, worn, and damaged tapes.
Assuming that the cause of the aborted Play and Record was a weakening sensor
and/or bad debounce circuit inside the microcontroller, there is no way of
tying these in with the prior failure of the gear on the tape loader mechanism.
If it were just out of warranty I would credit Goldstar with timing multiple
parts to fail simultaneously but it is at least 8 years old at this point.
Patient: Garage sale HP DeskJet Professional ink jet printer in pretty good
physical condition except see below. $5*.
Symptoms: Switching power on results in absolutely no indication of life either
from the front panel LEDs or motor movement.
Testing: The output of the power brick is marked and I was able to verify
that it was approximately correct - 20 VAC. Since it is a simple
transformer, it is probably good. Attempts to evoke a response from
the printer without opening the case were all dismal failures.
You probably know me better than to believe I paid a whole $5 for a broken
printer. Well, I did get a PC power supply and a bunch of old hard drives
and controllers as well. As of this writing, the power supply and at least
one of the hard drives are known good.
Unfortunately, I neglected to ask what the original problem was. But from
his attitude, the printer must have died totally. Otherwise, he would have
likely mentioned that it was probably a minor problem (and attempted to
charge more than $5).
Anyhow, the printer is in a box sitting on the ground and somewhat waterlogged.
I don't know if the stuff was out all night and it rained or what.
First test, as noted, was of the AC adapter which is apparently just a
transformer. Its output read a bit higher than the listed value that is
typical of an unloaded wall transformer.
After drying the case off, I plugged the printer in and turned it on.
Nothing. No sign of movement from any of the motors, no lights on the
display, and no response to any buttons.
OK, time to get inside. Some HP engineer probably won an award for the
mechanical design of this thing. (I seem to recall reading about this
product line in a past HP report.) Indeed, it is easy to get apart and
the subsystems are readily accessible. The top of the case is held in
place with 4 snaps. There are only 5 modules inside: power supply,
logic board, print head driver board, printer mechanism, button and display
panel. Except for the fact that HP seems to used weird size philips head
screws, the entire thing can be disassembled to the module level in about
10 minutes. However, I would have preferred screws to hold the top of the
case in place rather than snaps as it is necessary to deal with the snaps
every time the top is removed - which during troubleshooting is quite
frequently as the paper tray is part of the top cover.
Back to the story. All I have done to this point is remove the top of the
case. However, the next time I switch it on, the print head whomps to the
left end of its travel and seems to be trying to fling itself off into space.
Power off. Let's try that once again. Power on: Whomp-whir! Power off.
Then, I notice the puddle of water on the logic board!
First, I try to deal with in-place by mopping up what I can get to. Then,
I start up the old air compressor and use that to blow and dry the water.
Well, it isn't really up to the job so I start up the 3 horse shop vac on
blow. However, I cannot get under the print mechanism where the bulk of the
logic board is located.
Removing the print mechanism requires only disconnecting of 3 connectors
and a ground strap (one of those darn HP philips head screws!).
Now, it is easy to get to all parts of the logic board. I also remove the
EPROM and dry out under it to be sure.
I replace the print mechanism.
Ready? Power on: Nothing. OK, we try that once again. I said power on:
some lights on the display. There is no response from the buttons and no
motor movement. Several more power cycles results in somewhat random lights
but no other action.
OK, I will take out the logic board so I can inspect the underside for
water. This requires removing 7 or 8 screws and the power connector.
The board is fine. I also reseat the power connectors.
Reassembling - ready? Power on: Now, I get lights AND the print head
seems to be doing something reasonable, like the reset sequence. Back
and forth, prime, and then the On-Line light comes on. Its pitiful brain
thinks that everything is fine and ready to print. Is it?
Along with the printer came about 15 sheets of somewhat water damaged (but
now dry) paper. This should be good enough to test it even if every third
sheet jams and tears.
Now, how to do a self test? Fortunately, the complete HP users' manual
was part of the deal. "Hold the FONT button while turning on power".
OK, no problem.
Now, the printer goes through all the expected motions of initialization
and then proceeds to load a sheet of paper and....
Nothing. Actually, as far as it is concerned, the self test worked fine.
But, there is no ink on the paper! Well, the cartridge could be empty
but it doesn't feel that way. Based on my past experience, it is nearly
full. The nozzles could be clogged but sucking on the business end of
the print head (yuck) results in some ink drops appearing.
At this point since I do not have another cartridge to try, I decide to
strip it down completely once again and do a thorough inspection. Somehow,
I don't think the water was the cause of the original problem. All connectors
are reseated (including those to the print head which are the flexible printed
variety).
Everything looks fine except for the solder around the DC power connector
output pins of the power supply. There may be a hint of cracks around
several of them. Well, my soldering iron makes short work of those!
Put it all back together once again. No change. There is always a chance
that operating the printer with the waterlogged logic board may have damaged
something - always fear the worst, right? Therefore, I decide to do a little
exploratory probing of the print head driver board since everything seems to
be fine except there is no action from the ink jet nozzles.
A bit of ink jet theory: There are something like 50 thin film heating
elements inside the micromachined chamber at the business end of the print
head. These heaters are pulsed at precisely the right time by the logic
to vaporize the ink in direct contact with them and expel a drop of ink
toward the paper. If a heater is bad or a nozzle is clogged, there will
be a missing line (out of the 50 possible lines) on the paper. Priming
is supposed to assure that the nozzles are clear, loaded, and ready to go.
First I test for power - there is both +5 V and +20 V. I assume the +5 is
for the logic and +20 for the print head nozzle heater elements. There
are something like 50 nozzles and I have no intention of testing them all
but it is fairly easy to determine that the print head receives +20 for
power and the nozzle driver pull low to turn them on. Probing with a scope
while the printer is supposed to be doing the self test confirms that there
are pulses at a representative sample of the nozzle wires.
The only thing left to do is try a new cartridge. But, that has to wait
until I can 'borrow' one from the office.
Two days later.
None of the cartridges we use at the office are exactly the same part number
but they do have the same array of gold pads and the same size print head
itself. I take one of each of the two types we have (one is the super-high
capacity type - all right!.
I first try the cartridge that looks exactly the same as the one that came
with the printer (though the part number is different).
And, what do you know! The self test is nearly perfect. There is one missing
line. This could be a bad driver (hope not) or defective cartridge (yeh,
right!).
Next, I try the PaintJet 'high capacity' cartridge and this also works but
now there are 2 missing lines. :-(
Going back and forth, they are consistent. I am not sure if one of the
two missing lines are the same on both cartridges. Could something be
marginal or is the priming not working? However, all other nozzles seem
to be rock solid. Reseating the connectors to the print head makes no
difference. If I knew which drivers were involved I could look at the
signals but it will be difficult to trace the circuitry from the driver
board to the actual nozzle.
Comments: I suspect the original problem resulting in the dead printer to
have been a cold solder joint on the DC power connector which I repaired.
I don't really think that the nozzle problem was caused by the water since
the print head driver board was never wet. Since the data connection to the
print head driver board is a 20 pin cable, this must be a common bus and thus
it is unlikely that any failure on the main logic board could manifest itself
as a single or pair of bad nozzles. Stay tuned.
Patient: Garage sale oldie but goodie. Asked $10, paid $5 ("Sure, I just
want it out of here"). I did suggest to the owner that he could
probably fix it himself for under $3 but I guess he hadn't read
the FAQ, or it didn't match his decor, or something. I did ask
about the remote but that had long since walked.
Symptoms: I was told by the owner that he had to open the VCR to extract the
tape. I asked: "You didn't do anything that I will be cursing you
out for, did you?". The answer was 'no' but he did not sound all
that convincing....
Testing: No take-up reel movement in FF, CUE, or Play modes even with a
cassette cheater and no load on the take-up reel. Not that I am
terribly surprised.
If this had not been a HiFi VCR, I would probably have passed as I have
enough Panasonics of that era. However, I also know that those machines
can be kept going with minimal effort and investment almost indefinitely.
The first step is virtually automatic: replace the idler tire.
Removing the idler on this VCR can be accomplished without any disassembly
beyond taking off the top cover - but this was easily done as the owner
never replaced the @#$% screws. Grrrrr.
Pop off the split washer (careful - don't loose it. Yes, I know, you are
not supposed to reuse these....).
Since I don't happen to have the proper size tire in my inventory, I first try
to turn the old one inside-out but this little bugger refuses to cooperate.
I finally find one that can be stretched to fit until I obtain a replacement
(actually get one from my secondary inventory at work).
FF and REW now work fine. Play, however, results in a picture which is
nearly total snow with a distorted picture showing through and no HiFi
sound. I use my Mark 1 thumb on the heads - no change. I then clean the
video heads using cleaning sticks and alcohol - no change. Then, I realize
that I have not tried to adjust the tracking control and on some machines,
if tracking is way off, the picture will not just have noise but will be
totally unwatchable. Sure enough, the picture and HiFi sound now come in
clearly.
CUE (forward search) works fine. REV (reverse search) results in snow and
then shutdown. Then, I realize that I have not replaced the split washer
on the idler. I don't know if this was the cause, but replacing that washer
results in REV working fine after this. Maybe, the idler assembly was
creeping up on its shaft.
For good measure, I also check the two belts under the deck. They are in
good condition. Therefore, I just clean the belts and pulleys. I also
give the top side a general cleaning - capstan, pinch roller, roller guides,
fixed guide posts, full erase head, and A/C head stack. The only detectable
oxide/dirt buildup was on the capstan and pinch roller and even this wasn't
that significant.
Recording seems to work as well. I leave it recording for several hours to
be sure that there are no thermal problems.
Comments: This VCR is about 11 years old but in excellent condition. Older
Panasonic VCRs were built very solidly with a lot of metal in the transport
and reliable electronics. As noted, little tends to go wrong - rubber parts,
power supply capacitors. Thus, keeping them going requires minimal effort.
While these machines don't have spectacular features by today's standards,
they will outlast a bucket load of the cheaply constructed junk that passes
for many modern VCRs.
Patient: Garage sale HP DeskJet Professional ink-jet printer in pretty good
physical condition. Original problem was described in "Repair
Brief #49 - HP DeskJet Professional Printer - Part 1: Dead."
Symptoms: Ink-jet nozzles #45 and #47 do not work resulting in a pair of white
lines every 1/8" or so on the printout.
Testing: A variety of new and used print cartridges produce similar results
except that in some cases, only 1 line is missing (??).
As you will recall, this printer was dead (and somewhat waterlogged) when
I picked it up for $5 at a garage sale. Drying it out and soldering a few
suspicious connections resulted in everything working except for 2 nozzles
on the print head - #45 and #47. Well, 24/25ths of it works, what do you
want for $5?? :-)
Interestingly, when using a new PaintJet cartridge, only 1 nozzle appeared to
be bad...
The electronics for the DeskJet is divided between two circuit boards - the
main logic board and a smaller print head driver board. Due to this way of
partitioning I did not think that the water I found on the logic board could
result in damage in such a way that only 1 or 2 nozzles were affected.
(From: Paul Grohe (grohe@galaxy.nsc.com) in reply to Repair Brief #49).
"I concur. If just *one* jet is not firing, then it is on the
driver/flex-cable/connector/cartridge side. All the nozzle
decoding is done on the driver board, so the 20 pin interconnect
cable is okay. The DC (well..really 20VAC) power connector does
take some abuse in normal service, this could have aggravated the
cold joint."
Given this behavior, there are several possibilities:
1. Bad driver(s) - stuck off resulting in no heating and no ink.
2. Bad driver(s) - stuck on burning out print head heater(s).
3. Bad connections to print head - flex cable, connector, or contacts.
(From: Paul Grohe (grohe@galaxy.nsc.com)).
"I assume you have cleaned the contacts (with a Q-tip, on both cartridge and
socket). Use a magnifying glass and check *each* of the gold 'bump' contacts.
Repeated cartridge swapping, or improper insertion, can cause a crack to form
around the base of the 'bump' and the pad (or the pad and the trace). The
'bumps' can also be 'flattened' by cartridges that were forced in at too
much of an angle. There should also be some 'give' or 'sponginess' to the
contact area to assure even contact with the cartridge."
I have done this inspection - everything looked ok (at least as best I can
without removing the flex cable).
I removed the driver board and gave it a thorough visual inspection. As
expected, the soldering was perfect. Leave it to HP.
(1) would be difficult to find until I had a complete wiring diagram of
the driver board, flex cable, and print head. Therefore, I defer on this
until I have exhausted other possibilities.
(2) would be rather disappointing as this would mean that I have already
blown two new print cartridges ($20+ each) in testing (and that the problem
was probably in one of the 40 pin HP ASICs). At first I was *sure* this
was the case (of course, always fear the worst!) as testing between certain
contacts on the print cartridges resulted in unexpected readings. Most of
the resistances were around 32 ohms except for 1 which was open (blah) and
another which was high (51 ohms).
Then I examined the pads under a magnifying lens and found that the open
contact was indeed - open and not even connected to anything by design. The
51 ohm reading was too consistent - all 3 of my cartridges (including the
original that came with the printer) measured nearly the same value. If
this were due to a partial burnout, it would be an amazing coincidence.
The shorted driver theory was finally put to rest when I took one of the
cartridges into the office and confirmed that it functioned properly in a
working printer.
(From: Paul Grohe (grohe@galaxy.nsc.com)).
"Check for broken/bad traces in the flex-cable that goes from the driver board
to the cartridge. Ohm out the cable between the supply commons and the
individual driver lines (at the PCB) with the cartridge in place. I think
the jet resistance was about 50 ohms (It's been a while). There were four
separate jet sections (commons). All four commons were tied to the +20V
supply through four separate (12 ohm?) series current limiting resistors.
The driver outputs seemed to be grounded emitter, open collector (w/clamp
diode?). The jets themselves are driven individually and are not multiplexed."
This (3) seemed like the next step. I drew a diagram of the two 28 pin
connectors on the driver board. Then I identified the 4 common lines (1 for
each 12 or 13 nozzle heaters). These went to some common 10 ohm resistors to
the +20 V power supply. I then measured between the nearest common and each
pin to the print head. When I was on a heater fed by that common, the
resistance would be about 32 ohms. When to one fed from a different common,
it would be around 52 ohms (32 + 2 * 10). In every case but 3, these made
sense:
* One of these measured high even though it seemed to be no different than
the others. I finally concluded (or rationalized) that this was the funny
high resistance element in the cartridge.
* The other two tested open on the cartridge that had 2 defective nozzles but
one of these was the proper 32 ohms on the PaintJet cartridge that had only
one bad nozzle. Interesting.
Time to disassemble the flex cable connecting the driver board to the print
head.
The retainers on the flex cable at the connector-end just snap into the open
position and the cable comes free. Four screws release the print head cover
and the cable then pops free of its indexing posts.
Under the contact area is a rubber pad with a little bump for each contact.
All the bumps seem to be in good condition and minding their own business.
A careful inspection of the flex cable shows that it is in virtually perfect
condition as well with no cracks or wear at any locations including the 56
contact points.
However, the indexing holes seem to be a bit deformed. Could the cable
have worked out of position slightly resulting in poor contact to a couple
of pads? Wishful thinking, maybe....
While it is it out, I confirm connections between the two open pins and the
corresponding contact pads - finally correctly locating them. Following
the many fine traces is a bit tricky to say the least.
Reassembly is equally straightforward. I take care to center the flex cable
on its indexing posts. Is it my imagination or are the contact pads now
sitting flatter and more uniformly than before?
On a hunch - I really did not expect anything to change - I check the two
open pins. And, what do you know? They now measure 32 ohms. At this point,
I am now, of course, absolutely confident that the printer will work!
The first few lines out of the self test are disappointing as it appears
as though there are now many marginal nozzles. However, guess what? The
top line of text which previously listed the numbers of the bad nozzles
is the normal printout - ID E. Before it was: 45 47 ID E. The printer's
pathetic brain thinks all the nozzles are working fine. Then, half way
into the second page, the print died out totally. OK, maybe that cartridge
needs to be cleaned and primed. It has been bounced around, turned
upside-down, and otherwise abused. A little blowing into the vent hole
(drip, drip - love that dreaded black finger disease) and a wipe - and we
try again. Now it is perfect - no missing lines, no smudges, run, drips,
or errors. I have a working printer!
Comments: Once again, that actual problems turn out to be exceedingly simple:
bad connections in both cases. As noted previously, the HP DeskJet series
in general is a well engineered design with only a half dozen basic components.
While my printer is one of the oldest, the fundamental design has not really
changed dramatically in the last several years as evidenced by the fact that
print cartridges for some much more modern printers work just fine in this
old machine. The print quality with a new cartridge is nearly laser-quality.
Yes, HP seems to come out with a new, faster, cheaper, color. etc. printer
every few weeks. But, looking inside newer printers shows that their basic
design and construction is quite similar.
Patient: Aztech CDA-268-01A double speed CDROM drive in excellent physical
condition.
Symptoms: Drawer is constantly trying to close. Shutting off power and
pulling the drawer out half-way results in drawer closing and
motor spinning its wheels.
Testing: A universal CDROM interface card (Panasonic, Sony, Mitsumi)
was used to confirm that this was not a case of there just not
being an interface card present. See the text below.
This drive was sent to me as non-working having been removed from a PC after
it failed with the symptoms described above. An IDE CDROM was purchased to
replace it so it was not known whether the interface (probably a sound card)
was at fault.
I first began testing drive simply connected to a spare PC power supply (with
an auto headlight to provide the required minimum load for the supply).
It's behavior was consistent - the poor little drawer motor just continuously
tried to close the drawer regardless of pushing the any buttons. The Busy
light remained solidly on. The motor was getting quite toasty in the process.
Note: this drive has two front panel buttons that serve multiple functions.
to open and close the drawer and start and advance tracks for audio play
without software. While I was not sure of their exact function at this
time, neither button evoked a response of any kind.
So, I wonder if my magic spit will have any effect? Some careful trials on
the interface connector revealed a pair of pins where pressing with exactly
the *correct* concentration of magic spit(tm) would cause the motor to quit
and the Busy light to go out momentarily.
Hmmm. Maybe the interface card is required after all. There might be nothing
wrong with the drive. The original interface may have been bad. Since the
replacement used IDE, this would have gone undetermined.
What type of interface is it? Originally, I thought I was actually told IDE
but a close examination of the pin configuration showed this to be impossible
as the grounds were in the wrong place. It is a 40 pin connector, so SCSI
is ruled out. Therefore, it must be one of the proprietary interfaces:
Panasonic, Sony, or Mitsumi.
Since I am not willing to cannibalize one of my working PCs to obtain an
interface, this waits until the next time I am in the office. A co-worker
(OK, actually, the boss management weeny) who insisted on having a PC built
for him but rarely uses it and never has to the best of my knowledge used
the CDROM won't notice if the card disappears for a few days: "What, your
CDROM drive doesn't respond anymore? it must have gone bad from too much
loneliness.... I will get to it as soon as possible." HeHeHe. :-)
The interface card is a universal type - a jumper will select between the
three types. It is obvious from the pin configuration that the Mitsumi
interface is the only one that makes any sense with one complete row of
ground pins.
With the interface installed on my state-of-the-art DTK 286 mainboard, the
behavior is unchanged. The drawer still insists on attempting to close
forever.... With the cable installed, the magic spit no longer has any
effect so the pins involved are probably being driven from the interface.
Thus, the drive is really and truly bad. I assume that what was happening
was that my magic spit(tm) was triggering a master reset and thus disabling
all operations. This is encouraging in some ways - the drive is not completely
hosed.
What next? Checking the drawer switch is fairly easy and it is definitely
good - zero resistance at either end of its travel.
To get at most of the circuitry on the board requires removing a sheet metal
shield - 3 solder connections, no big deal.
I expect that if the drawer switch signal goes directly to an LSI chip or
ASIC, this drive becomes spare parts as such a multilegged creature is not
likely to be easily obtained, inexpensive, or replaced without a great deal
of cursing even with proper soldering equipment.
However, on this drive, the drawer switch signal goes to a 74LS244 octal buffer
in a surface mount package (joy!) and a pullup resistor. Checking at the
buffer input, the signal swings very nicely between 0 and 5 V.
Checking at the output we have - stuck around 2 V. Is the buffer bad? That
is what I thought at first but then I realized that:
* The enable to the buffer is not simply tied to ground (on) but is driven
from another chip.
* The output is connected to a data bus and not a dedicated chip input.
Therefore, I am seeing the average bus activity of that bit and not a
solid level.
Using a fine soldering iron and dental pick to lift the leg, I disconnect the
output pin of the 'LS244. Now, this output tests low regardless of the drawer
switch signal level - either it is bad and open or it is tri-stated. However,
momentarily grounding the enable pin to the 'LS244 allows the output to swing
the full 5 V following the drawer switch signal and this confirms that the
the chip is actually good.
The enable signal looks sick - it is floating at around 1.5 V. How is it
generated? Would you believe, a surface mount 74LS00??? Most of the circuitry
on this board is in a few large custom chips and yet there are still a few
jelly beans as well! Grumpf, there is no accounting for designers' tastes.
Better for my chances of repair. Interestingly, there are absolutely no
adjustments of any kind on the mainboard - probably a digital servo system.
Checking the LS00 shows that the gate in question has its inputs on pins
13 and 12 and its output on pin 11. Pin 13 has a nice periodic pulse on it
but pin 12 looks as dead as pin 11 - not a solid logic low but close enough
to ground that the output (it is a NAND after all) should be high and it is
not. Lifting the output pin makes no difference so it is not being loaded
down. Power and ground connections to the chip seem solid. Checking the
other 3 gates of the chip show that they are equally screwed up with outputs
that are in never-never land and do not change significantly (though there is
some slight response) when the inputs switch. My suspicion is an internal
power problem with the chip, not that this helps us.
_ 1/8 74LS244
| \
Door Switch o----------------| >o----X-------o Data bus
|_/ (Break at X and
13 _____ o Enable pulse data bus)
___|___|___|__ o---------| \ 11 |
12| NAND |o---+ Dead - about 1.5 V
1 _____ +---|_____/
+---| \ 3 | NANDs are sections of
| 2| NAND |o----+ 74LS00 in SMT package
Ground o--+---|_____/ Dead - about 1 V
To see if the relevant output of the 'LS244 (the bus line) actually has an
effect, I connect a fine insulated wire to it while it is still disconnected
from the 'LS244 and power the drive. Now, momentarily touching this to ground
will sometimes get the drawer to stop and even sometimes the CD will try to
spin up (accelerating past Warp 10 and never actually succeeding in reading
the directory (I presume since it never stabilized). I expect this is due
to the fact that I am overriding whatever else is supposed to be on that
bus line and it is getting really confused. Poor thing. :-(
At first I thought a neighboring 74LS04 (hard to believe, huh?) was also bad
but the designer morons at Aztech or Mitsumi or whoever actually did this
board did not tie many of the unused inputs to valid levels so they were
just floating. (With LS TTL, unused gates should be forced to a constant
output by tying their inputs to ground or through a pullup to Vcc, whichever
is appropriate for the logic.) So that was a false alarm.
In fact, only 2 of the gates on the LS00 are used (1 of these as an inverter,
see the diagram!). Only *1* inverter on the LS04 even connected to anything.
What a waste!
My initial thought was to wire up the unused LS00 gates to substitute for the
bad ones but these, too, were deader than the proverbial dodo. I thought
about wiring some unused sections of other jellybean chips (there was a mostly
unused 74LS08 across country) to create the needed logic. But sanity prevailed
and I decided to try to locate a replacement. I finally found an old PC I/O
card (that might even have been bad) with a 74LS00 in the same SMT package.
Fortunately, at the office, we have a PACE rework station so removing the
old LS00 and its replacement goes smoothly using a tip that heats all of
the pins simultaneously. After cleaning up the pads, I use a super-fine
tip soldering iron to tack pins 11 and 4 and then solder the rest of them
without too much difficulty.
Now for the test: Although I did not take the rest of the drive, I should
be able to cause the Busy light to go out if I can convince the drive that
the drawer is actually opened. And, sure enough, momentarily grounding the
drawer switch signal results in the Busy light going out. Furthermore, the
behavior is slightly different on every other push of the Open/Close button
indicating that it is actually trying to read the disc when it thinks the
drawer is closed. Of course, with no optical pickup attached, this might
be rather difficult!
The initial functional test will be made with an audio CD since that should
be good enough to confirm basic operation. Apply power: drawer closes and
stops, Busy light goes out. OK, we are cooking (no smoke!). Insert disc,
press the button. Drawer closes, discs starts spinning, sort of. Well,
actually it starts spinning in the wrong direction (counterclockwise). Then,
apparently it thinks better of it and accelerates clockwise. While it does
not go ballistic, it is certainly spinning faster than the 1X speed. Of
course, I really don't know what it should be doing not having an instruction
manual! A little experimentation with the buttons and then it seems to be
spinning more slowly. Maybe I hit the combination to play the audio CD.
Time for the headphones.
And, sure enough, track 1 is playing, a bit scratchy, but nonetheless, there
is music! A little more experimentation with the buttons reveals that the
left hand button advances the tracks and will cycle back to the beginning
once the last track is played. The right button pauses the play with one push
and ejects with a second push. Of course, if I could interpret the icons on
on the bezel, I might have been able to figure this out without trial and
error!
The scratchy sound is a little disconcerting knowing that data readout is
more critical. Furthermore, it seems to have problems accessing the outer
tracks. Stopping the CD results in similar strange rotation before it finally
decides to open the drawer.
Well, the CD I am using has seen better days, being a casualty of a couple
other Repair Brief CD experiences. So, I get a good CD. Unfortunately, now
my Vivaldi is not even recognized! Grrrrr.
I clean the lens.... No change.
There is one other possibility that doesn't involve thousands of lives and
millions of dollars - I have not yet replaced that sheet metal shield.
Perhaps there is some interference between the electronics and the pickup.
Sure enough, after soldering the shield into place, *all** detectable
problems vanish - even the peculiar wrong-way rotation.
Now for the computer. Rather than putting together my state-of-art DTK 286
system again, I will install the CDROM in a real PC (OK, well a real PC from
about 5 years ago).
Hardware installation is a snap. What about a driver? Since it has a
Mitsumi interface, I assumed it would work with a Mitsumi driver. Wrong.
After finally finding the correct I/O address (as my interface card is not
marked well - just jumpers for Bits 1-4 but what position is Bit 1 and is
a jumper a 1 or a 0?), it will still not initialize properly though it does
find the card - the light on the drive flashes. No combination of IRQ or
DMA makes any difference.
Internet to the rescue! A Lycos search turns out to be totally worthless not
even being able to quickly locate the Aztech homepage. Did you know that
there are 4 or 5 companies with Aztech in their name?
However, using Yahoo, an entry of "Aztech CDROM" results in the third or forth
entry being a link to the Aztech Utility/Driver FTP directory at:
http://www.aztech.com.sg/c&t/ftp/cdrom.htm
Sure enough, there is an entry for my CDA268-01A, atcd268.zip, Install disk
V1.35, 53,561 KB.
Download, unzip, pop into the PC's A drive and 5 minutes later I have a
working CDROM. No runs, no drips, and no errors. I was tempted to install
Win95 but I figured that wouldn't work really well on a 4 MB 386!
Comments: Despite what I had to say about the Aztech engineers, this drive
seems to be a very clean design. There are virtually no discrete parts and
no adjustments of any kind on the mainboard. The layout is fairly wide open
on a double sided printed circuit board. There is also an LSI chip on the
optical pickup itself - perhaps the focus/tracking servos and front-end
decoding.
I expected this CDROM to either have a trivial problem like a bad connection
or bad drawer switch or a fried custom irreplaceable chip. I would never have
anticipated that a jelly bean 74LS00 would up and die. But that it is what it
did as there is no evidence of any kind of trauma, spike, or spill. One
assumes that 20 year old technology will be reliable. This is yet another
example where the initial expectations and fears can be totally unfounded.
Patient: Garage sale electric drills. I picked up two of these (similar
models) at different times).
Symptoms: Running the drill at anything near full speed results in a spine
tingling squeal.
Testing: Putting a drop of oil in the rear bearing will quiet it down for
a few minutes but this is not a long term cure.
This is a classic case of cost cutting (or how much the Marketing department
controls the Company) resulting in early failure. Simple bronze bushings are
used at both ends of the motor shaft. At the gear-box end, this is acceptable
as this is enclosed and shielded against contaminants. However, at the handle
end, all kinds of stuff can find its way into the motor and bearing. In
particular, when using the drill with a sanding disk, fine powder easily
infiltrates the motor absorbing lubrication. (Please, no comments about using
the proper tool for the job. The fact of the matter is that electric drills do
get used in this type of service.) The result is a dry bearing which rapidly
wears if not attended to. It is hard to ignore as the result is a spine
tingling squeal whenever the drill is running.
How to deal with it? I could probably have purchased a replacement bronze
bushing from Sears Parts or used the good one from the other end of one of
the drills to fix the other. However, what is really needed is a double
sealed ball bearing instead. The seal is the important part though at the
speed at which the motor runs, a ball bearing isn't a bad idea in any case.
I have upgraded a couple of these drills to ball bearings. The substitution
is straightforward requiring disassembly of the drill - removing of the
front gear reducer and then one side of the case. At this point, the old
sleeve bearing is easily freed from its mounting (just the plastic of the
case) and pulled from the shaft. The shaft is likely undamaged unless you
attempted to continue running the drill even after going deaf.
The drills I upgraded had bearings that were 7/8" OD, 5/16" thick, and with
a 5/16" ID center hole. The old ones were worn by almost 1/32" oversize
for the center hole but the motor shaft was undamaged. I found suitable
replacement double sealed ball bearings in my junk box but I would assume
that they are fairly standard - possibly even available from Sears Parts as
I bet they are used in the next model up.
If the gear reducer needs to come apart to access the motor, take note of
any spacer washers or other small parts so you can get them back in exactly
the correct locations. Work in a clean area to avoid contaminating the
grease packing.
The bearing should be a press fit onto the shaft. Very light sanding of
the shaft with 600 grit sandpaper may be needed - just enough so that the
new bearing can be pressed on. Or, gently tap the center race with hammer
(protected with a block of wood). Make sure that the bearing is snug when
mounted so that the outer race cannot rotate - use layers of thin heat
resistant plastic if needed to assure a tight fit (the old sleeve bearing
was keyed but your new ball bearing probably won't have this feature).
These drills now run as smoothly as Sears much more expensive models. Of
course, the chuck will probably fall off at any moment...
Comments: I had to do a similar upgrade to a cheap shop vac which had basically
the same problem - the top bronze bushing had lost its lubrication resulting
in accelerated wear and failure. Unfortunately, access to this was quite
a bit more difficult requiring almost total disassembly of the motor/blower
unit. These sorts of failures are common with inexpensive hand-vacs as well.
This isn't rocket science but a simple modification like this can significantly
extend the life of a tool that would otherwise be discarded. Of course, if it
were designed properly in the first place, such upgrades would not be needed.
Patient: A friend of a coworker's CDROM Drive inherited from his company
when it broke (of course). Got that?
Symptoms: The thing was sitting on the edge of Dave's desk when I showed up
in the morning. All the pieces were there, some assembly required.
I could not really get a coherent description of the problem - I
thought it had something to do with the drawer opening... Sound
familiar? (Recall "Repair Brief #52 - Aztech CDA-268-01A CDROM
Drive - Drawer Continuously Closing").
Testing: I had to reassemble it first! Then I suggested I take it home
since I really didn't have time to deal with it at work.
First, I give it a careful examination but find nothing. I then reassembled
it to the point where power can be safely applied. I did not bother with
the interface, just the power. For these occasions, I have an old switching
supply conveniently mounted under my PC desk - and one that doesn't even
require one of those quaint headlamp minimum loads! At first the drive
appeared hesitant to spin the CD but then after a couple of minutes, seems
to be fine - the 2X and Busy lights flash on momentarily as the disc spins
at what I assume to be 2X speed. It seemed to behave all night as I would
every so often push the Eject button to open and close the drawer.
At this point, I figured the problem had gone away or was solved or something.
Suspecting bad connections at the flex cable connector, I disassembled the
unit once again and cleaned and reseated these. I cleaned the lens as
well while it was accessible Testing showed it to still be fine at least
in so far as the basic reading of the disc identification was concerned.
Email to Dave: The CDROM seems to work, I will try it when I come in next.
Next time at the office, I installed the drive in a PC that we had been using
for various CDROM and CDR projects which was conveniently in pieces.... It was
running Win95, so I assumed that the CD would be found automatically. You
know, Plug 'n Pray!
The drive has what I assumed to be an IDE interface - the missing pin was
in the correct location. I finally found the info on jumper setting where
one might actually expect to find it - on the label!
However, trying all reasonable combinations of jumpers (Slave, 8/16 bit) with
the harddisk set for Master on a single IDE controller did not result in any
recognition by the New Hardware Wizard. The first time I tried it - with
no jumpers resulting in the CDROM being set for Master - I did see harddisk
errors and the Busy light on the CDROM drive flashed as the harddisk was
being accessed.
Then, I noticed that it did not seem to be initializing properly, again - the
red and orange LEDs did not flash on when I inserted a disc. Hurrrumph.
"Hey Dave, well I thought it was fixed but now it is dead again."
Back home and with the top cover off, I watch as it is supposed to be spinning
up. Now the disc will not turn at all! However, if I give the spindle help,
it will seem to try for a little while but never get to the proper speed. Ah
ha! Can we spell - are you ready: Spindle Motor?
Disassembling the drive further to access the bottom of the optical deck
reveals a cheap brush type permanent magnet motor - essentially the same as
those used on our favorite Pioneer CD players and changers. We know what a
joy they are!
Testing with an ohmmeter results in readings between 12 ohms and 0 as the
motor shaft is slowly rotated. This is one sick motor.
First, I tried spraying it with tuner and control cleaner through the
ventilation/brush access holes in the rear. This did not seem to make
any noticeable change.
I can use a power supply to attempt to clear the short. First, I unsoldered
the red wire so that there is no chance of blowing any of the circuitry
when external power is applied.
First, I got my genuine Heath variable power supply but this proved incapable
of spinning it as the current limit kicked in at 700 mA. Into a dead short,
this was not enough and the voltage stubbornly remained near zero.
So, I got my trusty "destructo-proof give it all you can take" variable
power supply. Now we are in business!
Turning it up to about 10 V allows the motor to spin at high speed hopefully
flinging the metal whiskers or other crud off of the commutator.
That did it! Now, the reading varies between about 15 and 20 ohms which
seems much more reasonable. I resolder the red wire.
This will work now!
Sure enough, spinup is once again consistent. At least now, I am sure
it will remain that way for a little while at least.
So what about trying it on a PC at home. Now, which PC to use.... Paul Grohe
suggested the 'blue one' when I asked him if he knew of a proper software
driver for the NEC. OK, where is the blue paint? :-)
After giving up on my version of Netscape trying to download a .EXE file, I
finally used ftp to get NEC-IDE.EXE, a self extracting .ZIP file containing
README.TXT and NEC_IDE.SYS which is a universal driver for NEC IDE CDROM
drives under DOS/WIN31. (Win95 has these drivers built in but my newly
painted 'blue' PC is only a 4 MB 33 MHz 386.)
After following the instructions on modifications to the CONFIG.SYS and
AUTOEXEC.BAT files, I am ready. The jumpers are set for Slave and 8 bit.
Long pause.... "No drives found - initialization aborted". Boot completes
Grrrr, maybe it is 16 bit...
Long pause.... "No drives found - initialization aborted". Boot completes.
Double Grrrr, let me remove the Slave jumper, at least that will confirm
if the drive is recognizing IDE commands as it should conflict with the
harddisk and cause harddisk errors or cause the boot to hang with "HDD
Controller Failure".
Long pause.... "No drives found - initialization aborted". Boot completes.
Wait, that messed things up when I tried it at work (before I figured out
that the Slave jumper was needed at all). What else could be wrong....??
Maybe, power needs to be cycled to reset the drive.
Slave, 8 bit:
Long pause.... "No drives found - initialization aborted".
One more time - Slave, 16 bit:
Short pause, Busy light flashes, "One drive found....." Success! The drive
is now accessible under DOS and Windows.
Semi-documented software is soooo much fun! :-)
Email to Dave: The drive works, same spindle motor problem as Pioneer changer.
I cleared it for now but changing motor will be a slight pain....
I never did notice any problem with the drawer.
Comments: It somewhat amazes me that many CDROM drives still use the same
cheap brush type permanent magnet motors as Pioneer and many other CD players.
When driven at more than 1X speeds especially, these are prone to metal
particle migration or metal whisker formation or whatever - more commonly
known as crud buildup. Thus, the same problems that are so common with many
brands of CD players are destined to appear on CDROM drives - and the same, at
least temporary - cures are effective. The use of a brushless spindle
motor - even common in some basic CD players - would totally eliminate the
possibility of this failure mode. Even some cheap Walkmen tape players use
brushless drive motors. Perhaps, these drives are *not supposed* to last
that long by design, good excuse to upgrade.
Patient: RCA FPR560ER (CTC130C chassis) 19 inch color television. This is
a nice TV but it has been driving Bill up a wall.... (You will
recall my friend Bill from Repair Briefs #35 and #47.)
Symptoms: The TV will power itself on/off and/or lose channel lock for random
amounts of times at random times.
Testing: Whacking did not seem to have any definitive effect - but this TV
is quite solidly constructed so it is not clear that the whacking
reached the relevant components.
I got involved with this after Bill had initially attempted to repair the TV
for what he thought was a simple case of a blown horizontal output transistor
(HOT). He had replaced it with something from Radio Shack that they called
an HOT :-). The transistor ran extremely hot and lasted about an hour. Now
we know why Radio Shack called it a HOT! (I don't even think Radio Shack
lists higher power transistors in its current catalog - I wonder why!). Next,
he replaced it with a TIP552 which has decent specs. An additional heat sink
was added as well (although we know this should not be needed with a properly
functioning deflection circuit). Nonetheless, this seems to be holding up.
However, Bill then noticed the onset of the erratic behavior and believed
that the blown HOT somehow caused the set to start acting weird.
I suspected that the erratic power cycling caused the original HOT to fail.
Thus, this was a symptom of another problem, not the cause.
The behavior took on a variety of forms. Among them were the following:
* Sometimes the screen would go black for a few seconds with or without the
channel numbers incrementing at random and then come back as though nothing
happened.
* Sometimes the TV would actually appear to click off for a few seconds or
longer.
* Sometimes in the middle of the night the TV would just decide to turn itself
on at full volume.
Needless to say, this is no way for a self respecting TV to act!
Bill and I decide to spend a little time on it. Bill drags out a rickety
typewriter (you remember them, right?) table to put it on. Will this hold,
I ask? "Sure. If it collapses, we won't need to repair it." OK, keeping
my distance. It holds. However, the TV operates fine for the few minutes
we have allocated.
Bill (who used to design switching power supplies) is attempting to come up
with some explanation involving a problem with the power supply or control
logic.
We, of course, know better. From early on, I was working on the hypothesis
that bad connections were involved but where were they? No amount of wacking
seemed to evoke a response. And, the TV never would perform on demand. We
even dragged it over to my place and I would run it whenever I was around but
it never screwed up - until s couple of weeks after it was lugged back to
Bill's house. I kept telling him that it wasn't fixed. But, Bill is the kind
of guy who figures if it doesn't screw up immediately, the problem probably
went away. Yeh, right.
For about a year, he was using the TV on a switched outlet. He would shut
it off totally when not in use and live with the erratic behavior when it was
on. During the course of an evening, it probably would only misbehave once
or twice.
At least the HOT seems to be surviving.....
Finally, it became too much and it was relegated a corner of the basement
where it sat for another 6 months or so. Every so often he would mention
the TV but I wasn't really that eager to deal with it.
Then, while on a cleanup crusade, "Sam, take this thing. If you cannot fix
it, trash it. I will go get a 27 inch set on sale."
So, we lug it back to my work bench. Even though it is only a 19 inch set,
it is quite heavy and somewhat bulky. Since the only way I can get it off
my bench will be to fix it, I start the search....
Unlike some more modern sets that can best be described as a CRT with a wart
(the main board), this has one circuit board for deflection and video, one for
the tuner, another for the A/V inputs, power supply, etc.
Checking each in turn (I didn't start with the tuner hoping - more like wishful
thinking - that the problem would be elsewhere since the tuner required more
disassembly):
* Power supply - all soldering looks good.
* Deflection and video - removing this requires disconnecting some cables
but once this is done, it comes out enough to examine in detail. Everything
looks fine.
* A/V - this is unlikely to be the cause so I do not attempt to get to the
underside which would be a pain.
* Tuner - to get at the bottom of the board requires unsoldering a metal cover
at several points. My Weller soldering iron is not up to this but a 140 W
soldering gun comes to the rescue.
Almost immediately, some suspect pins come into view. However, they are not
obviously bad. First, I try to measure for resistance changes as I flex the
board and/or press on each pin. This turns out to be impossible as there
is no convenient place to hang clip leads and my DMM's response is too slow
anyhow. The only way to find out is to now prod each one with power on.
There are three main blocks of large pins. Viewed in the normal orientation,
these are at the upper left, middle top, and upper right. Starting at the
upper left results in immediate gratification: tapping lightly with my
solder sucker causes the set to power on and off in a manner very similar to
the original symptoms. There is definitely a hairline crack around at least
one of these pins. Well, there is only one way of knowing for sure. I
resolder all five of the pins on this connector. Now, it is rock solid.
Hitting it with a (insulated) hammer has no effect.
I also solder all the other similar size connector pins on the board.
Now, to come up with suitable substitutes for half the cabinet screws which
Bill lost....
Comments: While this is not the infamous CTC175/176/177 tuner solder problem,
these older RCAs seem to have a less severe case of the same disease. I have
seen others with cracked solder connections often in similar locations.
Unfortunately, the only way to be sure of the repair is to let Bill start
using it again. One can never be quite sure of a repair to an intermittent
problem that happens so infrequently.
Patient: Garage sale Toshiba M4200, $13. Owner mentioned audio problems
suggesting that it needed a new audio head. I love it when the
owner offers a diagnosis - they are almost invariably wrong.
Symptoms: Indeed, audio is somewhat weak but not as bad as I was led to
believe. I expected nothing. When pressing START, the audio
seems to fade in rather than coming on at correct volume every
time.
Testing: Playing various tapes at multiple speeds exhibit similar symptoms
to a greater or lesser extent.
First, I tried cleaning the tape path. This seemed to make a slight
improvement but not much. Adjusting the azimuth setting on the A/C helped
a bit more (but I found out later that it was the VCR on which the recordings
were made that was misaligned.)
As a more extensive test, I let it record for a couple of hours. The beginning
of this tape played back somewhat weak but was otherwise quite listenable.
However, about halfway through, the audio dropped out totally over a span
of a few seconds. After this, only silence.
As another test, I play a tape (Battle of the Bulge, if you care) and after
about an hour and a half, the audio starts fading out to be replaced first
by silence and then a sort of whine/buzz.
Interestingly, the video is rock solid and perfect under all conditions.
Could this be an electronic problem? Perhaps, the audio driver/preamp chip
or its power supply is failing due to heat? We always tend to suspect the
hard-to-locate or expensive problems first.
A careful examination of the tape path reveals the problem but at first not
the cause: the tape is moving up on the A/C head so that the audio track is
no longer aligned with the record/play head. The control track is wide
enough that enough overlap is still present and there are no servo problems.
I can only surmise that alignment is such that a combination of the audio
and its guard band are in contact with the audio head. Very gently pressing
on the edge of the tape restores full audio volume but eventually it wanders
off again.
The tape movement is also not mirror smooth - it is rippling between the
A/C head and right roller guide. However, no tape edge damage seems to
result.
So, what is the cause?
At first, I thought that there was some actual mechanical alignment problem.
It could not be with the roller guides as the video is perfect (this is a
non-HiFi VCR so there is no issue of HiFi head alignment). Roller guide
tilt, fixed guide post vertical height or alignment, even A/C head tilt,
can affect this.
Careful experimentation with the adjustment of the A/C head doesn't result
in any noticeable improvement.
Could one of the fixed guide posts have shifted? Not likely, they are screwed
down tight and locked.
What about the capstan and pinch roller? I had cleaned them but could the
capstan have shifted position changing it angle or something like that?
Again, not likely - the 3 screws are secure and this VCR did not fall off
of a 10 story building as far as I know.
This leaves the pinch roller.
The pinch roller seems to be in reasonable condition but not perfect. So,
as a test, I will grind off the outer deteriorated (oxidized) layer which
is hard and shiny instead of resilient, dull, and rubbery. Removing the
pinch roller is very easy - just pull up on the plastic cap and the pinch
roller is freed.
I used a #8, 2 inch machine screw and nut to mount it in my drill press.
Then, a file and fine sandpaper were used to remove the outer somewhat
hardened (oxidized) rubber layer. Care must be taken not to upset the
perfectly flat cylindrical shape of the rubber surface.
This appears to result in a substantial improvement. The audio is still a
little weak but much more consistent. It is now possible to play a long
tape to completion without audio problems.
Although resurfacing the pinch roller is generally considered a temporary fix
I have used the VCR this way (though only occasionally) for several years.
I recently obtained the correct replacement from Frank Fendley since he was
able to identify the exact part number.
In addition to replacing the pinch roller, I had to construct a battery
compartment cover for the remote control since this was missing. Can you
believe it? I really expected more for $13! :-)
Comments: When I got this VCR, I wasn't as aware as I am now that old rubber
parts - even if they look good - can result in these sorts of symptoms. Of
course, we all know that erratic audio, random speed changes, and crinkled
tapes edges can result from a hard worn pinch roller. Therefore, the usual
cleaning, inspection, and replacement of dead rubber parts should be performed
first before contemplating exotic mechanical or electronic problems.
Patient: Garage sale HP DeskJet Professional ink-jet printer in pretty good
physical condition. Original problem was described in "Repair
Brief #49 - HP DeskJet Professional Printer - Part 1: Dead" and
Repair Brief $51 - HP DeskJet Professional Printer - Part 2: Missing
Lines".
Symptoms: The printer will be happily going its merry way when within the space
of one line of text, the ink will fade out across the page. After
this, only blowing into the vent hole (with resultant possible mess)
will result in any ink on the paper. This may occur after only a
few lines or several pages.
Testing: A number of cartridges and text files were tried with similar
results.
As you will recall, this printer was dead (and somewhat waterlogged) when
I picked it up for $5 at a garage sale. Drying it out and soldering a few
suspicious connections resulted in everything working except for 2 nozzles
on the print head - #45 and #47. Well, 24/25ths of it works, what do you
want for $5?? :-) This problem turned out to be bad connections to the
cartridge due to the flex cable with the 56 contact points shifting position.
Just when I thought the printer was fully functional, what do you know?
Halfway down a page of text, the type dies out over the course of a couple
of lines. Now, keep in mind that I had been using the printer without
incident since finding the bad connections in the print head.
My first thought was that something electronic was changing - perhaps a power
supply or pulse width - resulting in too much to too little juice to the nozzle
heaters. This would be a @#$% to find unless it were one of the power supply
rails. Even then, a few seconds shift in level would be all that was needed
to mess up the carefully orchestrated operation of the nozzles.
I went to far as to monitor the +20 V while printing without seeing anything
out of the ordinary - just a few mV change in value depending on the load
(number of firing nozzles/amount of black ink on the line). I removed the
print head driver board and examined it for cold solder joint. There were
of course none as is typical of HP's quality manufacturing. I wiggled anything
I could think of but nothing correlated with the drop-outs.
(BTW, never stick anything into the vent hole. I found out the hard way that
this may result in failure of the vent valve and ink all over the place since
it depends on the cartridge being sealed above at normal operating pressure
to keep the ink in place. I salvaged the cartridge (maybe) with a blob of
silicone sealer over the vent hole. I do not really know what the long term
implications of this might be.)
At this point, I even sent an email message to Paul Grohe asking if he had
seen the symptoms described above since from his postings and email it is
obvious that he is knowledgeable on the subject of DeskJet printer repair.
The problem in detecting this was that as noted, whatever was happening would
only need to occur for a second or two to then require manual (by blowing into
the vent hole) priming.
Come to think of it 1, why didn't normal priming work?
Come to think of it 2, did normal priming ever work?
Although it never quite registered until now, normal priming cycles never
seemed to accomplish much of anything. I always had to blow to get anything
to print if a cartridge had lost it 'charge'.
OK, so how does priming work? Very cleverly actually.
There is a kind of 'service station' where the print head is positioned for
priming. When in position, a rubber cup seals against the face of the print
head and connects with a drain tube below. In the base of the printer is a
positive displacement roller and tube pump - a set of rollers (I assume they
are rollers as I did not entirely disassemble the unit) rotates against a
compliant plastic tube (like the blood pump in a heart-lung or dialysis
machine if you have ever seen one of those). The direction of flow is
determined by the direction of rotation of the rollers. This is controlled
by the position of a spring loaded 'feeler' which enables one of three 'shift
levers' to be lifted to engage the sucking pump (Maybe, one of them blows - I
didn't entirely figure out what they all did). The position of the 'feeler'
is determined by print head location which is under control of the print head
servo system. Thus no additional electronics is needed. Slick.
The priming action, which operates off of the paper advance motor takes place
between pages - when there is no paper being fed. With my printer and the
drivers being used, it appeared as though it should prime between every page.
Was it working?
I removed the print mechanism and found a drive gear I could rotate by hand.
With the number one shifter lifted, the pump should have been sucking. Was
it? Using some of my Magic Spit(tm) in the drain tube, I could see it being
pushed out of the tube - away from the print head. Darn, it is working. But
then, when I let up, it appeared to get sucked back in. Huh? Maybe something
was blocked.....
After removing the cartridge, I got a syringe with some water and loaded up
the little cup so the water was even with the top.
Shift, rotate, rotate, rotate. Nothing! It should have sucked that water
away in the first pass.
I got a toothpick and attempted to insert it into the bottom of the cup
and locate the hole for the tube. At first, it would not go anywhere but
eventually, I found the hole. Still nothing. I then rounded off the end
of a resistor lead to use as a probe. With this, I was able to push it
(gently!) down and into the semi-transparent tube so that I could see it.
A little manipulation (sort of like root canal if you have been unfortunate
enough to be familiar with that) cleared out the channel. Well, at least,
any dentists reading this will know what I am talking about.
Finally, rotating the pump by hand resulted in obvious movement of the water
(now mixed with ink). Enough rotation and I was able to clear the cup of
liquid and dispose of it via the drain. Obviously, dried up caked on ink was
blocking the hole and tube connected to the cup. Priming could not possibly
have worked.
Now for the test.
I installed my original old cartridge which had never worked properly under
any circumstances. And, what do you know?! I get quite a nice printout on
the first try without doing any of the manual blow priming I had needed
previously.
I sent another email to Paul Grohe starting something like:
"By now you are probably saying "what a moron". I cleaned out the priming
tube which was totally blocked with dried ink. So far so good."
The next day I received a lengthy reply outlining some other problems of
a similar nature that are common with DeskJets. These will be added to
the FAQ.
So, the priming had not worked for as long as I have been mucking with this
printer!
Now, hopefully, this is the last Repair Brief on this printer!
Comments: This is probably monotonous by now but once again, this was a simple
mechanical problem. Not an expensive chip. Not a power supply. Not even a
bad connection. Just dried up caked on ink. What is amazing is that it worked
at all. Perhaps, the filled cartridge had enough pressure from the weight of
the ink to not be as finicky but when it got used up somewhat began to cause
problems.
Patient: Garage sale Magnavox FD2000-SL01 CD player. This is a classic top
loader. It is as big and heavy as a typical full size VCR. I paid
$2 but also got the service manual as part of the deal - slightly
waterlogged but quite legible.
Symptoms: Power is alive - the front panel LEDs come on but discs are not
recognized though they do spin.
Testing: Tried various CDs, whacking, cursing. Nothing changed.
First, a description of the FD2000-SL01:
While we have become used to CD players of all shapes and sizes with various
levels of feature-mania, let us not forget their roots. The Magnavox Model
FD2000-SL01 is probably one of the earliest consumer CD players. It is a
design more reminiscent of a linear tracking servo locked turntable - which
is probably one of its close ancestors. The engineer was probably told: "We
need a consumer CD player - yesterday. You have designed linear tracking
turntables. Make it operate like one of those."
The service manual (included in the $2 price) has a date of 1983. The manual
was ordered in 1988 (I have the invoice for that as well).
The front panel buttons include the usual PLAY, STOP, PAUSE, >>, <<, and
REPEAT. There are even STORE and CANCEL buttons. However, there is no
time display, only two rows of 15 LEDs.
The top row LEDs are illuminated to show the complement of tracks on the
disc while the bottom row LEDs indicate the current track that is playing
and/or the next track to which it is seeking. The top row LEDs go dark to
show which tracks have been played so this is similar to the traditional
calander or linear displays of modern CD players - but in LEDs instead.
Pressing PLAY while playing serves as a track forward (>>|) button. There
is no track reverse (|<<) seek.
As noted, it is a top loader, about as large and heavy as a full size VCR.
The see-through double action lid permits one to watch the CD spinning - what
a concept! The interior is pretty much packed with electronics - as opposed
to any modern CD player or compact stereo you might encounter!
Internally, the FD2000-SL01 consists of the following modules:
* Optical deck with its own front-end circuit board. It uses a radial voice
coil type actuator.
* Servo board (about 8 x 8 inches).
* Demodulator/audio board (about 8 x 8 inches) with the actual EFM decoder on
a separate mezzanine card.
These two boards are enclosed by perforated metal shields.
* Display/button board mounted behind front panel.
* Power supply board (about 4 x 6 inches) and power transformer.
The pickup uses the Philips rotary actuator technology - little different
from modern Philips CD players or CD ROM drives except that it is much more
solidly constructed with individual parts - the laser and photodiode array -
designed to be field replaceable assemblies. However, while you would think
this design results in rapid access, think again. As you will see below,
this player has by far the slowest seek time I have ever encountered.
We now return you to the present:
With the bottom cover removed, the optical deck comes into view. Pressing
PLAY does result in some action - the pickup bounces to the far end, then back
to the home position - perhaps a couple of times - before giving up. During
this time, one can hear some high pitched whining as the servos attempt to
locate the disc directory.
After a few minutes of wiggling and prodding, I am successful at getting the
disc directory to apparently be read - the number of LEDs corresponding to
the number of tracks on the disc are illuminated. Expectantly, I connect
the audio outputs of the player to my handy-dandy Heathkit compact stereo.
However, the speakers remain disappointingly silent. :-(
At this point, the servo systems appear to be working - if a bit erratically.
I assume there were/are some bad/dirty connections in the cabling or socketed
chips. In fact, in the end, there is still at least one bad connection I have
yet to locate.
Seeking to the next track reminds me of an inch-worm moving along: zeek, zeek,
zeek, zeek, zeek,....., zeek (though inch-worms don't, I suppose, make these
sorts of sounds). I assume what I am actually hearing (the 'zeeks') are
the sounds of track ('track' here referring to the spiral groove or line
of pits on the CD) crossings. If the positioner is moved by hand, you can
hear the same track crossing sound - almost like dragging the stylus across
an old LP - but much more closely spaced of course. With a practiced ear,
it is even possible to count the tracks. Let's see... This CD has exactly
17,243, or was that 17,244? :-)
It takes about 20 seconds (and 40 to 50 zeeks) to seek from track 1 to the
last track on a typical CD using a coarse-fine search strategy. So while
the rotary actuator should result in a very short seek time, the designers
had not taken advantage of their superior technology. Audible search (>>,
<<, at least I think it should be audible), makes a more conventional dit,
dit, dit, dit... sound.
Some careful tweaking of the focus and tracking adjustments shows that these
are probably optimally set already. Since seeks work and the disc directory
can apparently be read, it is likely that the optical components including the
laserdiode and photodiode array are fine. As a double-check, I put my scope
on the RF test point. The 'eye' pattern is stable and free of noise.
All power supply voltages check out.
To go further, I need to get to the top of the circuit boards. The cover
comes off easily enough - the owner had been in there already and lost most
of the screws. Grrrr. The only connection to the cover is the interlock
switch which is easily bypassed. In order to get the CD to spin properly,
I remove the clamper magnet from the lid and set it on top of the CD. The
laser beam is safely blocked by the CD so there is no danger.
Getting to the demodulator/audio board requires removing the metal shield by
prying it out with a screwdriver. The servo board is underneath this (the
foil side is accessible from the bottom after its shield is removed but since
the servo systems seem to be working for the moment, I should not need to
get to it).
Since I have the schematics, I first go for the audio signals out of the
left and right channel (separate) D/A converters - nothing.
Working backwards from the D/As indicates that there is no activity on the
digital lines into the D/As. This confirms a digital problem in the readout
or decoding logic. The audio circuitry is likely in good condition. Woopie!
There is a mezzanine card mounted in the lower right corner of the servo/logic
board which does the EFM decoding. It includes some sort of controller, static
memory, and some logic and I/O buffers. There is no activity on the outputs
of this board - or on its local bus. In fact, although the clock is running,
nothing else seems to show any signs of activity. The main chip is a CX7934.
I have been unable to identify its internal functional block diagram or even
a pin description. So, while all power and as best I can determine, all
inputs are correct including the EFM input (buffered squared up eye pattern),
audio muting, clocks, and power. I do not know for sure if this chip is dead
but it sure appears that way. I believe it is also is running somewhat hot...
Thus, for now, I am shelving the unit. I may return to it in the future. In
the meantime, the only sound produced by *this* CD player is the zeek, zeek,
zeek of the rotary actuator.
Comments: This was an entertaining experience since many interesting signals
are actually accessible with the level of integration used in this design.
Modern CD players use 2 or 3 chips to do all of the processing with serial
data running between them. Here, there are actual bits you can hold in your
hand!
I would still like to get this machine operational. It is a wonderful example
of early CD player design. I would be interested if anyone has one of these
complete CD players or individual components that are likely to be functional
sitting in an attic, basement, or junk pile. I am pretty sure I need the
mezzanine card (possibly designated 30-892-C16) at the very least.
Patient: Garage sale 13" GE color TV. Sitting by itself on the lawn
near end of sale. I think I paid about $5 for it - which was
probably about $10 too much!
Symptoms: Initially, the TV appeared totally lifeless.
Testing: External whacking had no effect. However, some prodding of the
mainboard would occasionally get it going. Once running, it
might continue to work but with erratic messed up color and
occasional vertical jitter.
This is one of those GE chassis affectionately described as 'bad solder
connections held together by copper traces'. I, however, wasn't familiar
with these at the time, so the troubleshooting took some time - make that
lots of time!
Among the problems were:
* Erratic startup.
* Incorrect and washed out color.
* Jittery vertical deflection.
* Dirty tuner and controls (trivial in comparison with the others).
Since prodding had an effect, bad connections were indicated. But where?
Careful examination of the soldering didn't reveal anything. Suspect joints
that were touched up made no difference.
Most connectors seemed to be firmly seated and without detectable problems.
A couple of power transistors in the vertical deflection hanging out in midair
(by design) were connected via 3 pin plugs. These seemed to have some weak
contacts so I extracted each pin and bent the contact to increase the spring
force. This helped some with the vertical jitter but the erratic startup
and color problems remained.
Prodding also had an effect on the color problem but it was not possible
to localize it.
Off to the library to copy the Sams' Photofact for the set. Silly me.
I thought that would actually help!
It took me more time than I care to admit before realizing that this was a
double sided circuit board and those stupid rivet things where not just test
points or wire connections. I had not checked the top side wiring!
Now more about this disaster called a TV:
The wiring on the circuit board uses what are called 'rivlets' - poor man's
through-holes or vias. For each connection through the circuit board or
to a wire, a metal rivet is first set in the board and clinched. Then, the
wave soldering machine is supposed to complete the electrical connection by
soldering the rivlet to the circuit board traces. This has to be done on
the top as well as the bottom but with solder from the bottom. The only
problem is that the temperature probably wasn't high enough or not enough
or improper flux was used. Initially, the connections were fine and the
TV worked reliably for anywhere from a few milliseconds to several years.
However, repeated thermal cycles finally resulted in numerous intermittent
connections.
Removing the mainboard requires unplugged 8 or 10 connectors. Fortunately,
most were keyed and labeled so I got away (by accident) with not drawing
a picture of how they were arranged. I had copied the Sams' Photofact
schematic for the set but not the entire folder with the parts and connector
placement. The mainboard may be removed without disconnecting the CRT anode
as the flyback is mounted to the frame.
What I finally did to more-or-less fix the set was to use a soldering gun
with fresh solder and flux to rework every rivlet I could find on both the
top and bottom (basically going over it twice) of the board. In some cases,
components had to be temporarily removed as some rivlets lurked underneath.
This did cure all three of the major problems.....
Until I buttoned it up. A couple days later, the picture faded to black
over a period of 15 seconds - and then came back over the same amount of
time. Oops, forgot to rework the CRT driver board. Sure enough, I found
an obvious bad connection to the CRT filament - this one I could actually
touch.
A couple of years later, the convergence deteriorated suddenly. You guessed
it. I hadn't reworked the little convergence board.
A while later, the poor color problem returned intermittently. On initial
power-on, the color would be fine but would then drop out and shift toward
blue/green. It was sometimes possible to obtain correct color by turning the
COLOR control all the way up and the TINT control to one end.
I finally bit the bullet and resoldered the entire mainboard yet again.
Now it is working again (crossed fingers). Anyone need a great TV?
Comments: The 'proper' way to reliably repair this chassis would probably
be to remove all the solder from each rivlet, scrape the solder mask from the
traces in the vicinity, and add bare wires through the rivet hole and to
the traces to complete the through-hole connection. Then, fresh solder and
flux with a hot iron. One could spend their waking life on one such set!
There are probably only a half dozen actual rivlets that are the problem
children. However, there is often no visible evidence of the bad connections.
Even with the schematic, locating them based on would be difficult and there
would be no way of knowing where the next one would crop up.
Patient: Garage sale Canon FaxPhone 80 really big and old fax machine. Hey,
it was only $2. (They gave away the supposedly dead microwave
oven free - I still am not quite sure if there is anything wrong
with it other than the missing temperature probe. Maybe the Fax
was dead and they should have charged $2 for the Microwave!)
Symptoms: Buttons and LCD display do seem to work but there is no paper
movement, it keeps insisting on paper problems 'Check Paper'
when attempting to copy. It was also flashing 'Load Ink Sheet'.
Testing: I only have about 2 feet of the semi-special paper this machine
uses. It is a thermal wax transfer type print mechanism with a
full width ribbon. I try copy mode but there is no action. Even
paper advance complains (once I located the correct button with
no manual and those 'universal' icons).
You might call this a plain paper fax - but one using a special ribbon,
a wax transfer type. The ribbon seems to be mostly used. The 'Load Ink
Sheet' message was caused by the marking on the last few feet sensed by
an LED/photodiode sensor. Hopefully, there is enough left to at least test
the machine while troubleshooting. Better yet, I just rewind it a few feet.
So what If I get the previous owner's faxes in negative (like a used carbon
ribbon from a typewriter or impact printer) superimposed on my test! Actually
this could be interesting :-) I don't even know if one can buy this type of
ribbon anymore.
The paper seems ordinary - perhaps a roll of shelf paper would work just
as well as the kind Canon no doubt sold for this unit.
In all fairness, it does seem to want to cooperate - just that the paper
is not moving. So, what operates the paper advance (and the cutter, for that
matter - it makes a horrible grinding noise and seems sick as well - there is
likely a common problem).
In order to access the electronics, I need to remove the cover. This turns
out to be relatively straightforward - 2 screws in the front and 2 on top.
The phone itself unplugs but is not needed anyhow as the keypad remains with
the main chassis.
Now, I can see the gears. My first observation is that one gear is just
twitching when the paper should presumably be doing something. Sound familiar?
Try "Repair Brief #1 - Daisy Wheel Printer - Carriage gets stuck".
It seems that getting to the electronics takes a little more work - two more
screws in front and removing the rear panel. Then, the entire mechanical
assembly swings up enough to get at... the driver board for the paper advance
motor. It is a small separate board, easily detached from the frame. Very
convenient.
Of course, pressing paper advance with the machine in this state results in
the paper actually advancing - what a concept!
Prod, prod, flex, flex. Oops, there it goes into spaz mode again. So, how
about those bad connections? I reseat the connectors - no change. A visual
inspection of the back of the board shows a number of solder pads that may
be suspect but no smoking gun. Running my finger over individual pins has
no effect other than to draw blood. However, flexing the heatsink/board
combination does seem to make the problem come and go. Well, there is only
one way to find out.
About 10 minutes and a hundred or so solder connections later (just to be sure,
did nearly every one), the paper advance function is now stable. Flexing and
prodding the board has no effect. Hurrah!
OK, so do I have a working fax? With my precious 2 feet of paper, I set
out to copy a page - seems like some old assembly language program. Well,
as they say, a page is a page is a page.
Sure enough, copy mode seems to work and the quality isn't half bad.
Will it work with a phone line. Simplest is to transmit a fax to my PC but
that would involve disconnecting from the Net! No way.
I send a one page fax to work - "To sam - Fax Test #1". I am in no hurry.
Next time at the office, sure enough, the fax was sitting in my (physical)
mailbox.
This fax will prove convenient for sending though I will have to find a not
exorbitantly priced source for the roll paper. For now, I rewound the
ribbon about halfway to provide more than enough for any of my needs. Let's,
wasn't that a wedding invitation that just went by? :-)
Comments: The motor driver circuit board does not appear to plated vias (I
cannot even recall now if it is double sided). In any case, there was clearly
inadequate support for the leads of the larger component - just a thin film
of solder. The holes were grossly oversized and with no plating, did not
provide adequate mechanical support for even slight thermal expansion and
contraction cycles. As soon as touched with a hot iron, the solder pulled
away from most of the pin - I had to add a fair amount of fresh solder to
produce a decent bond.
Patient: Player given to me to look at. I was told that it didn't work
and smelled really bad...
Symptoms: Even unpowered, it did smell really bad. :-) I had to borrow a
Sony AC adapter to test it. The unit was totally dead. Applying
power didn't help the aroma either.
Testing: Using the proper Sony 9V adapter results in no action, not even
the display does anything other than all segments being black.
So Sharon (not her real name), our purchasing person (before being laid off),
hands me this Sony portable CD player. "A friend of mine, Larry, would like
to know if there is any hope". Sharon appeared kind of embarrassed to ask me
to deal with this dead fish. :-)
Even with the cover closed, it is obvious that something unfortunate happened.
I don't believe I was actually told that an improper power adapter was used
but this would become obvious pretty quickly. The sorts of catastrophic
failures I found do not generally occur spontaneously.
This is one of the early Sony portables - solid metal construction. The
bottom comes off easily revealing a scorched surface mount part. Hmmm,
maybe this won't be so bad after all. Wishful thinking. A little tracing
reveals that the crater where the part used to live is basically across the
9 V DC input. Protection diode. I scrape its remains off of the board
clean with alcohol, and replace with a 1N4007.
No change - surprise, surprise.
Furthermore, it still smells really bad.... It might even be a little worse.
So, how do manufacturers of portable devices protect against accidents or
stupidity?
There is often some kind of protection in the form of one or more of the
following:
1. Diode across the power input with reverse polarity so that it normally
does not conduct. It will short circuit a reverse polarity supply and
hopefully fail shorted with excess reverse current.
2. Diode in series with power, again for reverse polarity protection.
It will simply not conduct if the polarity is wrong.
3. Maybe a fuse or two in disguise. There are things called "IC protectors"
which look like little black transistors but with only two legs. Hard
to identify on surface mount, though.
4. Overvoltage crowbar - zener triggering an SCR at substantially higher
than normal input voltage. This shorts across the supply and (hopefully)
blows an internal fuse if the resulting current is excessive before the
SCR and circuit traces vaporize.
In this case, only (1) was present - at least I could not locate any fuses,
fusable resistors, or IC protectors.
What next? I still need to locate the source of that really bad smell.
Nothing else on the readily accessible solder side of the mainboard seems
to be in distress. Unfortunately, getting to the top side components
requires unsoldering a bunch of skinny wires (labeling them) and a flat
printed cable (the focus and tracking drive, which I managed to eventually
rip from repeated assembly and disassembly).
This done, I still don't see anything smelly on the top of the board. Then,
I notice *the box*. It is a sealed metal enclosure about 1" x 1/2" by 2"
housing the DC-DC converter which powers most of the player's circuitry.
Although, there is no visible charring (how can metal char?), this is the
only possibility. Fortunately, only 4 or 5 pins anchor this module to the
circuit board. Once removed, it is obviously the source of the aroma. Yum!
I use an Xacto knife and soldering iron to remove the cover - soldered along
the edge. Then the damage becomes clear: the largest transistor is totally
melted, split in two, and unidentifiable. Traces on the little circuit board
are also destroyed. The insulating cardboard *is* nearly charred. However,
other small transistors and discrete components appear to test fine using my
multimeter.
I try a generic NPN power transistor - no output from the DC-DC converter. It
appears not to be enabled as the on-off switch is a logic level going to the
main microcontroller which appeared to be dead. I even tried to substitute
external power supplies for the voltages provided by the DC-DC converter.
Again, no change.
Tracing the input power connections show that they go directly to a large
multilegged chip. Generally, reversing power on integrated circuits is not
a good thing to do as they rarely survive.
Obviously, something more powerful than a typical AC adapter had been used
Usually, such damage is the result of something like attempting to use an
auto cigarette lighter adapter to power the device.
Nothing I have received so far in my quest for dead optical pickups, portable
CD players, CDROM or optical drives, has been in anywhere near the sad state
of this poor Sony. :-(
I sent Sharon email:
"Larry probably tried to power this thing from his cigarette lighter, huh?
It would probably have survived for a while on the 12 to 15 V of the auto
battery instead of 9 V from the adapter. But, he probably got the polarity
reversed. With the virtually unlimited current capacity of an auto battery,
the microcontroller was toast before it knew what hit it. Any fuse would have
been too slow to prevent terminal damage even if the circuit traces didn't
vaporize first.
If Larry wants it back, no problem. It still looks like a CD player. If he
takes it in for repair, the technician (while holding their nose) will even
probably agree that it was a CD player at one time."
Actually, with the blown up parts removed and the circuit boards cleaned with
alcohol, the odor has mostly disappeared. I still have it as a reminder that
AC adapter connections and ratings are not the same as data cables which can
often be reversed without damage. But not always - like those 44 pin min-IDE
connectors used on laptop harddrives - they smoke really expensive parts if
plugged in backwards because power is also on the connector at one end. :-(
Sharon forwarded my email to Larry who confirmed my suspicion. He was
attempting to use the player in his car.
I did remove the optical pickup and tested it. The laserdiode, focus and
tracking coils, and motors were fine (I have not tested the photodiode array
but expect it to be undamaged as well).
The DC-DC converter isolated most of the circuitry from damage. Unfortunately,
at least that one large IC, presumably the main system controller, ran on the
wall adapter DC voltage directly and appeared to be toast.
I also have a bag of smoked digital clocks, a smoked clock radio, and a smoked
cordless phone (which I have since repaired) from Larry (now Sharon's former
friend)..... I will discuss those in a future Repair Brief. Stay tuned for:
"Clocks, Clock Radio, and Cordless Phone - Smoked." Till then, you can attempt
to guess what happened. :-)
Comments: those voltage, current, and polarity ratings marked on portable
equipment are there for a reason. The voltage rating should not be exceeded.
Using a slightly lower voltage adapter will probably cause no harm though
performance may suffer. The current rating of the adapter should be at least
equal to the printed rating. The polarity, of course, must be correct. If
connected backwards with a current limited adapter, there may be no immediate
damage depending on the design of the protective circuits. But don't take
chances - double check that the polarities match - with a voltmeter if
necessary - before you plug it in! Note that even some identically marked
adapters put out widely different open circuit voltages. If the unloaded
voltage reading is more than 25-30% higher than the marked value, I would
be cautious about using the adapter without confirmation that it is acceptable
for your equipment. Needless to say, if you experience any strange or
unexpected behavior with a new adapter, if any part gets unusually warm, or if
there is any unusual odor, unplug it immediately and attempt to identify the
cause of the problem.
Some devices are designed in such a way that they will survive almost anything.
A series diode would protect against reverse polarity. Alternatively, a large
parallel diode with upstream current limiting resistor or PTC thermistor, and
fuses, fusable resistors, or IC protectors would cut off current before the
parallel diode or circuit board traces have time to vaporize. A crowbar
circuit (zener to trigger an SCR) could be used to protect against reasonable
overvoltage.
Not this one, unfortunately. All the smoke has been released....
Patient: Trash-picked 25" Zenith System 3 Model SC2569W in pretty good
physical condition.
Symptoms: Occasionally, mostly during the first few minutes but possible other
times as well, picture is replaced will solid blue screen with
retrace lines.
Testing: Running the set resulted in occasional blue flashes. Tapping the
cabinet did not have any effect.
My cousin (the one with the 3 kids, 2 dogs, and 13 goldfish), was walking the
LARGE poodle when she came across this TV sitting on the curb. The owner was
about (or perhaps she knocked - that wouldn't be out of character - so the
symptoms were available. "Screen occasionally turns all blue".
Upon my return from garage sale-ing, I get this call. "Are you interested
in a Zenith TV"? How big? "Big". OK, I will check it out. My usual rules
are that (1) I don't tend to bother with consoles, (2) TVs with knob tuners
unless they fall into my lap, (3) those that are excessively abused, or (4)
those that are really old but not old enough for antique status.
This one turned out to fit the requirements - a table model only slightly
larger than the CRT. No remote - darn. Have to send her back for that!
Fortunately, I have a half broken remote from one of Bill's dead TVs (you
know, the one that had a totally white picture that I may talk about in a
future Repair Brief, then again, maybe not, it isn't pretty) It is (was)
a Zenith A-line chassis TV but fortunately the remote seems to work. Of
course, Bill lost the battery compartment cover - typical. :-(
Getting it home wasn't too bad - only half a block away, fit the back seat
of my car.
At first, the supposed symptoms didn't show up. Great! Maybe the ride home
has cured it!
Then, after screwing around with a broken CD player for a while, came back
and turned it on - and - what do you know? Picture came up all blue and
flashed to normal.
Removing the cover is easier than some - 8 screws, back slides off and set
is fairly stable though I wouldn't want to sit on it.
Now, of course, no amount of tapping or prodding can make the picture turn
blue. Turn it off and wait.
A few minutes later, powering on results in a few flashes of blue - and -
more importantly, tapping on the CRT neck seems to affect it. Prodding
the CRT driver (neck) board or anywhere else has no effect.
The detailed symptoms appear to be:
1. Blue gun turns full on but with Red and Green still visible under the fog.
2. A second or so later, red and green disappear and only blue remains.
3. Tapping or waiting results in screen flashing to black (presumably as
some capacitors charge or something) and then normal picture appears.
This is almost certainly a heater-cathode (H-K) short in the picture tube.
Since it is intermittent, I cannot simply measure the resistance but must
check it while running.
I fabricate a temporary isolated filament (heater) winding by wrapping 3 turns
of insulated bell wire around the flyback core. Carefully, I unsolder the
picture tube filament connections to the CRT driver board and jumper the
temporary winding to this. My multimeter is connected between the filament
and signal ground.
Now, applying power should result in the blue gun video drive voltage appearing
on the filament when the short occurs.
Sure enough, after a few on-off cycles with no problems, the meter jumps to
a reading greater than 100 V. With the heater isolated, no apparent change
takes place to the picture.
To do a permanent repair, I first determine that 3 turns is satisfactory:
Reducing it to 2 turns results in a dull orange glow from the filaments
and slightly fuzzy slightly dim picture. 3 turns was correct - the glow is
bright orange. (I could also have compared it with the original.)
I locate a couple of feet of well insulated wire and make a more permanent
winding, routing the wires well away from the high voltage connector and any
hot components. I carefully cut the traces on the CRT driver board right at
the filament socket - maintaining the spark gaps to a couple of other socket
pins. Then, the new winding is soldered directly to the filament pins.
I could also have probably reused the winding on the flyback and just isolated
it from the ground on the CRT driver board but this would result in more stress
on the flyback - probably irrelevant but the homemade winding was easy enough.
While the set is open, I adjust the position of the CRT as well - for whatever
reason, perhaps since new, it is tilted and off center on the bezel. If I
hadn't removed so much dust, I would have suspected it to be a replacement
CRT but probably not.
I call my cousin: Do you want the TV? It is fixed. "No kidding, how?". So
I told her: There was a short in the picture tube but I rewired the set.
"It won't blow up or anything?" Nope, just not something every TV repair
shop would do. "Sure, I can give the one in the living room (Repair Brief
#12 - Sylvania TV with no Horizontal Sync) to the nanny and this one will
be fine for use with the VCR (Repair Brief #21 - Sharp VC7864U VCR Erratic).
The kids will like a bigger screen." This is also convenient as there is
already a universal remote at that location being used with the VCR which
will be fine with this set.
Comments: The owner probably called a TV repair shop (or perhaps lugged it
in) and was told that with these symptoms the CRT was shot and fixing it would
not be worth the time and money. However, there is really nothing wrong
with the isolated filament. In fact, the stress on the red and green guns is
actually less as the difference in voltages between the cathodes and heaters
is on the average smaller and the maximum voltage difference is less as well.
As long as an H-K short does not occur with the red or green, there should
be no problem.
Patient: Ralph's Sony boombox AM/FM/CD/Cassette.
Symptoms: CD player erratic - sometimes won't recognize discs, sometimes
shuts off and resets in the middle of a disc.
Testing: Tried multiple discs (classic to rock) just to make sure it isn't
just a matter of the player not liking Ralph's tastes in music (which
most other humans don't like either.
Ralph was our 'Nedrie' type (from the movie 'Jurassic Park'). Visualize
a Sun workstation with Coke stains all over the keyboard, empty Coke cans
and half eaten week old pizza, manuals and other stuff cluttering his desk
and the immediate area. He is, however, an excellent software engineer which
is why all this is tolerated.
Anyhow, Ralph treats his boombox about like the two vicious dogs he keeps
locked in his basement at home - firmly :-). Perhaps, I will discuss the
ruggedness of old Realistic portable CD players at some point in the future.
One of his (previous) dogs thought it was a bone. The CD player looked pretty
beaten but continued to run. I don't know about the dog.
He used to have a pet iguana as well.
Anyhow, Ralph keeps a typical small Sony boombox in his office (2 doors from
mine). Thankfully, the volume is turned down and/or he uses headphones.
Otherwise, I *would* be very creative about seeing to it that the CD never
worked quite right. Oh, your CD is acting up - again? That is really a
shame.... ;-)
I had taken a look at this boombox once before, even partially disassembling
it, but then the problem went away. This time, the top is squashed in about
1/2 inch due to Ralph's constant beating in an effort to get it to cooperate.
The main problem is gaining access to anything. In order to remove the CD
player assembly, one must remove the front cover (six recessed screws - at
least the locations are more-or-less marked), LCD display/LED/button card,
cassette deck, power supply, and part of the audio/radio section. The CD
player including the optical deck and electronics board can then be extracted
but actually running it in this condition is difficult or impossible. However,
I did manage to set everything carefully propped up and run the player sitting
on the table. It, of course, ran fine all day in this configuration. Well,
maybe Ralph would not notice the additional disarray......
At first, I thought the erratic problem was due to marginal power as the
display LEDs fluctuated slightly in brightness while the CD was seeking.
However, the line transformer is built into the boombox rather than being
an external adapter so at least there is no way to use an improper adapter.
Testing voltages internally seemed to indicate that power was fine.
Next, I removed the CD player completely and went over the solder side of
the electronics board carefully looking for bad solder connections. On one
connector, there was a suspect joint or two which I repaired. These were
probably the source of the erratic shutoff problem. I also tested the
interlock switch for intermittents - it was fine. I then ran it for another
few hours without incident which, of course, proved little.
Reassembly required another 15 minutes or so but was straightforward.
The interior is relatively open so cable routing, at least, was not a problem.
I installed only two cabinet screws and left the cassette deck in my desk
drawer for a few weeks so access would be easier should the problems return.
Ralph never noticed.
It has been a couple of years now and I have not heard of any complaints.
One time I thought there was a problem when I went into his office and
attempted to play a CD but it must have been a bad CD because he has seemed
happy (at least with respect to the boombox). Ralph has recently left for
greener pastures though I am sure I will hear from him if the CD fails again!
Comments: One wonders why product engineers cannot take a few more minutes
to design an enclosure that is easily serviceable. I suppose the quick
answer is that these are deliberately designed to be throwaway products
in any case. There is no way that anyone can justify the cost in time
to gain access to fix a miserable cold solder joint. The whole thing probably
cost $100 retail. Of course, one wonders further why cold solder joints are
still so common even on ordinary (not large pins) components. Then again,
Sony's manufacturing engineers haven't met Ralph!
Now I see that an AM/FM/cassette/CD boombox is on sale for $60. I wonder
what kind of construction it uses....
Patient: This is one of those portable tape players often used by the severely
visually impaired to listen to 'books-on-tape'. It has oversize
buttons with tactile function impressions, a wide range speed
control, and the ability to play any of the 4 channels on a standard
cassette.
Symptoms: I was told that the player would not finish the tape. In fact, what
was happening was that it would not play channels 3 and 4 resulting
in a high noise level but no proper audio.
Testing: I tried the tape provided as well as a regular stereo audio cassette.
Only 2 of the 4 channels worked properly.
This 'customer' has provided me with quite a lot of 'business' in the past
(keep in mind, I don't charge for anything beyond parts in these cases).
Mostly, previous problems were due to unintentional abuse. These have
included bad solder connections on a classic as well as more modern Radio
Shack multiband radio and a sprung reel clutch on a cheap boombox.
This tape player is a nice well designed unit with large rugged controls.
There is no sign of external damage but that may not mean much.
Since I at first thought that this was a case of the tape not finishing
rather than not playing properly, I first checked for takeup reel and capstan
motion and torque. Both were fine and thus I rather doubted that it simply
stopped playing a tape. Then I tried the tape channel selector.
Recall that a standard stereo cassette has 4 channels of audio - L+R forward
and L+R in reverse. Book tapes use all 4 channels individually. A switch
selects between the L (1 or 2) and R (3 or 4) channel is being played. (OK,
it might be the other way around.) Flipping the tape over selects 1,3, or 2,4.
Anyhow, after a little testing, it was obvious that channels 1 and 2 were
loud and clear but 3 and 4 resulted in some background noise - greater than
the tape hiss of channels 1 and 2 but not hum or oscillation.
What could it be?
Keeping in mind that there is only one channel of amplification in this
unit and the switching is done mechanically, not much.
One might suspect the electronics - but this is probably impossible as the
same electronics is used for both channels.
One might suspect the tape head - easy to check.
One might suspect the wiring - quite likely given a well used possibly abuse
tape deck.
To test for output from each tape head channel, all that is needed is a
small pointed tool like an awl or jeweler's screwdriver. Touching each
of the 4 terminals on the tape head should result in hum or buzz on at
least one of them - depending on the channel selector setting. (This also
works for normal stereo cassettes - you should get a hum or buzz out of the
appropriate channel when its tape head terminal is touched.)
I didn't even need to go that far as the cause came immediately into view
once I was able to access the tape head terminals.
The designers of this tape player did a good job. Unlike many similar devices,
there is a pop out plastic piece which covers the tape head - presumably to
allow access to the azimuth adjustment screw. As soon as this was removed,
the broken wire came into view. Since the entire tape head assembly moves,
some slight flexing must take place at the point of soldering. It should not
be much as the cable is clamped to the moving part but there may still be
some flexing so over time, the wire simply broke off.
Someone searching for an electronic fault would have had the guts of this
thing strewn all over the workbench and shop floor before finding the true
cause! It was tight, but just barely possible to strip and solder the wire
back in place. I probably should have put a drop of semi-flexible sealer
over the connection to stabilize it but did not. Well, in another 10 years
I will know what to do!
Comments: This is not one of those 'cast of thousands' repairs. Nonetheless,
for those who depend on talking books for their education and pleasure, a
working tape player may be more important than a high tech entertainment
system.
Patient: Aiwa 'Compact Disc Stereo Radio Cassette Recorder CSD-707'. Wow,
what a mouthful! Don't the designers of these things give people
credit for at least minimal powers of observation? It's a boobbox,
for goodness sake. Geez!
Symptoms: Everything works fine except for the CD player. Insert a CD and
close the door. The disc starts spinning clockwise (correct),
then changes its mind, reverses direction, and goes into warp
drive in the incorrect (counterclockwise) direction. Eventually,
it shuts down with a 0:00:00 display.
Testing: Multiple discs behave in a basically similar way. None succeed
in reading the TOC.
Jeff brought this beauty to work. "Sam, I have something to keep you busy".
Not the printer...? (The previous week, he showed me this Apple LaserWriter
NT that was making a horrendous noise. Since I concluded it was from the fan
in the power supply underneath, I really did not want to deal with it. It
would require half the day just to get at it Apparently, the printer was
behaving - Jeff now only turned it on when needed. "No, my dad's boombox."
This boombox appears to be in excellent physical condition (unlike the one
described in "Repair Brief #63 - Sony Boombox CD Player - Erratic Operation"
which was slightly crunched. Indications are that it just stopped playing
CDs without an provocation. No scratches or dents.
Well, since it is not even possible to view the lens without some disassembly,
I warm up my screwdrivers...
It requires removal of 8 long screws to separate the front portion with the
all the CD player components, speakers, and cassette door (yes, just the
door) and the rear half with the rest of the electronics. Four cables link
these two sections: speakers, CD power, CD control, CD display/button panel.
Four more screws and it is possible to separate the optical deck from the CD
electronics board. It is just possible to arrange everything such that the
unit can be run with access to solder side of the CD electronics board. This
is definitely much easier than that Sony boombox which required removal of
almost everything to gain access to the CD player.
With the system disassembled, it is possible to observe the following
as it attempts to read the TOC of a disc:
* Sled moves about 1/4 of the distance along the track and then resets
to the inner stop. Sled motor stops. This indicates that the limit
switch is working.
* Disc starts spinning clockwise, hesitates, then reverses direction. Sled
motor spins attempting to move the pickup *past* the inner stop. This
continues until shutdown takes place. The limit switch is being ignored.
* It is possible to hear the 'gritty' focus/tracking noise which indicates
that the optical pickup is at least partially working - focus must be nearly
correct. Tracking may be messed up. Only a scope will determine this.
With such behavior, my first reaction is that something must be really really
screwed up since the disc spins at warp speed in the wrong direction and
the sled isn't bothering to recognize the inner limit switch when it gets into
this state. However, I have come to realize that small problems can result
in bizarre behavior from CD players!
So, first I clean the lens -- no change.
Next, I disconnect the optical deck from its intermediate circuit board and
go over all the solder connection on it -- no change.
OK, bring in the heaving equipment - I drag a scope into my office.
There are marked test points including RF, TE, FE.
The RF test point is - well - strange. When power is applied, it jumps,
then momentarily is of high amplitude but then decays to a much lower
amplitude signal with somewhat random characteristics. In fact, I cannot
make any sense of it.
At this point, I decide that other tasks are of more immediate concern
(like lunch) and will button it up and take it home to work on at my leisure.
I tell Jeff. "So what are you going to do." Oh, probably stare at it some
more, a little cursing, then give up. Jeff knows better but just laughs.
Next day, I start by disassembling it once again (of course, I had only
installed 2 of the 8 cabinet screws) but going further this time:
* Remove the front panel as well (in hindsight, this wasn't needed).
* The clamper disk with the magnet so that I can have access to the disc
itself.
* Stuff a wad of paper into the interlock to keep the switch closed.
* Solder wires to the RF test point and a convenient ground for my scope.
Now I can work without having to juggle the front panel, electronics board,
optical deck, and scope probe all at once.
The connections to the optical deck look to be the same as for some of the
optical pickups I have been accumulating but no cigar - nothing that I would
risk swapping until I get desperate. I could easily blow the laserdiode or
electronics if they are not quite the same and I don't even know for sure
that those other pickups work. After all, they were guaranteed to be dead :-).
I will postpone that decision.
Confirming that nothing has changed - nothing has. The RF test point which
should have the eye pattern is still strange. With easy access to the disc
to stop it from spinning by hand, the 'gritty' sound is unmistakable - the
focus servo, at least, appears healthy.
My test disc is one of those zillions of AOL free 15 hours CDs. While I
don't expect to be able to listen to the data, most CD players will recognize
the TOC of a CDROM - though the resulting display will be meaningless. Thus,
I don't need to worry about damage to the disc - there are a semi-infinite
supply of these! Of course, should I succeed reading the TOC, I will need
to use a music CD for audio tests and adjustments.
OK, well, Mr. AOL CD behaves the same as the others.....
Now, do I dare tweak anything? There are three readily accessible
adjustments: Laser Power, Focus Offset, Tracking Offset. (There is
also a Tracking Gain pot which I noticed later. However, the gain controls
are usually of secondary importance so I never do bother touching it.)
I generally do not touch laser power adjustments until I am sure there
are no other options.
What about focus or tracking?
This smells like a tracking problem. However, the fact that the eye pattern's
amplitude doesn't remain fairly constant suggests that there may be some sort
of focus problems. Toss a coin.
Normally, this is the point at which I would insist anyone doing any
adjustments mark the positions of all controls. Yep, you guessed it,
I violate rule #1. What the heck...
Starting with T.OFS, just a hair in each direction - and what a change!!
The RF signal remains at high amplitude for much longer AND the disc now
tries to spin clockwise for a much longer period of time. My goals now are
twofold: (1) to get the disc to spin in the correct direction at roughly
the correct speed of 500 rpm and (2) to maximize the amplitude/length of time
that the RF test point signal remains at high amplitude.
Then I realize something else that is in the FAQ: I really should use a
non-metallic screwdriver as the slot seems to be electrically connected
to high impedance circuitry and the behavior is changing when I touch
it with my cheapo jeweler's screwdriver. A quick remedy which seems to
be adequate is to wrap electrical tape around the handle.
A few minutes of going back and forth between F.OFS and T.OFS finally - I see
the display change to what is obviously it pathetic attempt to interpret the
time/tracks of the AOL CD. Of course, not thinking, I press PLAY and am
greeted with a loud hum/buzz - the 80X86 object code of the AOL SETUP program
or something. Hit STOP.
OK, so how about a proper music disc.
I have an 'I don't really care much about this disc' disc for just this
purpose. One of those 'Intro to the ballet' or something. You know, it
sounds like the Salvation Army orchestra on a bad day. But serves my
purposes....
Although the TOC is read with a bit of hesitation, the player seems unable
to locate track one. OK, so the current settings may not be entirely
satisfactory.
A little more tweaking. Still just some whining and clicking sounds once
I hit PLAY. Well, maybe it will like an outer track better. (I have no idea
why it had no problem starting to 'play' the AOL disc.)
This seems to do it. It is finally successful at locating track 5 or 6 and
starts playing - a bit scratchy. Now that it is stable, I can carefully
adjust both F.OFS and T.OFS for maximum amplitude of the eye pattern. After
a couple of false starts where it got confused and shut down, I have both
set optimally and the player now seems to work normally.
I also note that touching the metal frame seems to cause some static but
I assume this is due to the poor ground through the cable - normally the
electronics board is screwed directly to the frame of the optical deck.
I finally did risk tweaking the laser power to obtain a 1 V p-p eye pattern
at the RF test point. Without service information to know what the proper
setting is, this can be risky (even a dirty pot could conceivably blow the
diode). I just adjusted it up from about .9 V to 1 V p-p. It was clear
from the response to my screwdriver - I could have increased it much further,
probably to the point of burning it out - that the laserdiode was not weak.
If the laser was dying, the power would have topped out about where it had
been set as the feedback would be doing all that it could to maintain
constant power.
After fine tuning the focus and tracking adjustments yet again, the behavior
seems relatively normal. I rather suspect that there may still be some kind of
problem as the tracking adjustment did not have as distinct a peak as I would
expect, but who knows?
I will have to embellish the repair description for Jeff. I cannot just
tell him I adjusted something. Maybe: "Well, I transplanted the laserdiode
from an extra 25X CDROM drive - I figured you would consider the boombox more
important than such antiquated technology" or something juicy like that. :-)
Comments: Would Aiwa Service have replaced the optical pickup? Perhaps.
I have no idea why adjustments were so far off as to render the CD player
useless or why this happened suddenly (as far as I know) without any warning.
Perhaps it was dropped and there was indeed a change in the optical alignment.
Perhaps a bit of dirt caused one of the controls to change value and my
twiddling really just cleaned the track on that pot. Perhaps the original
adjustments were marginal. Whatever the cause, It now works and appears
stable. The eye pattern looks fine. The music sounds fine.
Patient: A Sony 13" color TV model KV-13TR-20 in excellent physical
condition.
Symptoms: Initially, all channels resulted in a white screen. After some
warmup, overdriven video (AGC bad) appeared on one channel. After
some more warmup, a couple of channels were fine.
Testing: Let it run for a while. Whacking on the cabinet didn't seem to
have much effect.
A little tag came with this TV: "Tuner/IF Block, $99". I assume the owner
took it in for an estimate (or maybe just called someplace on the phone).
Obviously, $100 to repair a $150 TV is a bit steep. Now my usual rates,
on the other hand....
These are classic symptoms of the Sony bad solder joint problem afflicting
many models of middle aged TVs - not to be confused with the GE/RCA tuner
bad solder/EEPROM problems which are similar :-). See the document: "Sony TV Tuner and IF Solder Connection Problems".
Fortunately, on this set, getting to the tuner and IF boxes is pretty easy
once opening the case is overcome - snaps are used in four spots rather
than screws. Give me screws any day! On the other had, you cannot lose
snaps. (On some larger screen Sonys, the engineers in their infinite wisdom
placed reinforcing plastic at exactly the wrong place to block access to the
pins on the solder side of the mainboard :-(. )
The tuner module has about 8 thick pins and 5 shield connections to the
mainboard. The IF module has about 11 thin pins and 4 shield connections.
Both come off easily (no bad connections visible on the mainboard pads.)
I resolder all the edge pins (though most did look fine), all shield
connections, and anything else that appeared marginal. I wasn't convinced
that I had located the cause of the problem. There was no smoking gun...
Indeed, the symptoms are essentially unchanged.
Now, I start prodding each of the metal boxes. The tuner box is rock solid
but I can get the problem to come and go by pushing gently on the IF box.
Out it comes again. Now I get out the heavy artillery - bright light and
magnifier. Then I spot *it* - one lone pin not particularly noteworthy
in any respect except that there is clearly a broken solder connection and
the pin even moves visibly. This is apparently a common location for these
problems - one pin of a coil. I check over everything else once again but
I am confident that this is *the* cause of the erratic behavior.
Indeed, once reinstalled, reception is flawless - even playing bongo on the
TV has no effect.
Comments: As I have noted before, you would think that after several decades
of manufacturing consumer electronics, soldering would not be an issue. But,
this certainly seems not to be the case. The demands of the bottom line as
well as manufacturing issues of mixed through-hole and surface-mount technology
combine to make reliable assembly quite a balancing act. At least the Sony and
RCA/GE bad solder problems are well documented and straightforward to repair.
Patient: This is a Magnavox 'Stereophonic Phonograph Model 3P2515', age
unknown but likely at least 25 years old. I bet the feature
length name sounded really impressive at the time.
Symptoms: Record changer load and eject functions totally inoperative.
Testing: Cycling mechanism results in no change. It does play records
if the tone arm is positioned manually. The stylus also wants
to eat the vinyl.....
Since the changer did not respond, I directed my attack to this mechanism.
Unfortunately, the owner was not available to explain if this was the main
problem - it could possibly be something else.
On an older mechanical device, be it a dinosaur such as this or reel-to-reel
tape deck, lubrication, deteriorated rubber parts, or broken parts, are the
prime suspects.
But how is this one supposed to operate?
A large metal cam gear actually moves and raises and lowers the tone arm. This
is engaged by a smaller nylon idler gear which is moved into position by the
OFF/ON/EJECT knob or the movement of the tone arm to the end of the record.
In this case, the idler was failing to engage most likely due to gummed up
grease. Once it was freed, the large cam gear was so tight that either the
motor slowed and stopped or the gears teeth were skipping - thankfully without
lasting damage.
To free up the large cam gear, required removing a bracket and using some
WD40 to loosen the dried grease. Once this was accomplished it was worked
free. The shaft and bearing surfaces were cleaned. That seems much better.
Darn, the eject is now ignored.
Oops, not the correct timing - there are two possible positions for the
relative relationships of the cam and idler. Try the other one.
Well, although it now cycling correctly, the record does not drop. The
spindle mechanism is - what? Totally frozen!
Upon closer examination, it would appear that the load/eject mechanism had
been glued into a fixed position - the parts of the spindle that normally
drop the record were immovable. At first I thought this was simply dried
up grease but I am quite sure it is really some adhesive - possibly Epoxy.
Disassembly requires removing the entire platter bearing assembly from the
changer, clamping it in a vise, and twisting and pulling to free the spindle.
It would appear that this, too, was glued in place!
At first, I try freeing it up with WD40 but this proved inadequate so I use
a punch to remove the roll pin. The assembly now comes apart easily - and
I promptly forget how it goes back together! One would think that something
so simple would not be a problem but it took another 10 minutes or so to
come up with a logical arrangement of all 6 parts and several more minutes
on the floor searching for a 1/16" shaft that popped free.
However, once cleaned, lubed, and back together, it does appear to be much
happier.
Replacing the platter bearing assembly goes smoothly except for accidentally
stripping the hole of one of the 3 mounting screws - replaced with machine
screw and nut.
At first, everything appeared to work correctly but then the motor shut off
before cycling. In fact, it will only stay on erratically, sometimes only
when the control knob is held in the EJECT position. While I was tempted
to suspect that I didn't get something back together correctly, sanity
prevailed and I examined the motor switch. It seemed that the lever that
operated it did not quite push far enough and the contact was marginal.
After disassembling the switch and cleaning the contacts - and bending one
slightly - operation is reliable.
Perhaps, this was the actual complaint since there is no way the changer
could have ever worked in modern times!
Another quirk seems to be that when cold, the motor takes a few seconds
to come up to speed. At first I thought this was an idler rubber problem
but then discovered that the motor seemed to be running at half speed even
with no load.
There is no starting or running capacitor to go bad, it is simply a shaded
pole motor. Interestingly, it acts as the power transformer for the amplifier
with an isolated center tapped low voltage output. I cleaned and lubricated
the motor but I do not believe this is the problem. The speed does seem to
recover so I am not going to lose much sleep over it.
One final adjustment: The tone arm tracking force was set so high that I
was afraid the stylus would dig a trench entirely through the record on
the second or third playing. It took a minute or two but I located the
tracking force screw and set it to a more reasonable value - high by
audiophile standards but a 3000% improvement.
Comments: There are still many people who depend on this sort of ancient
technology for their listening pleasure and are probably better off than
many of us who worry about signal-to-noise ratio, dynamic range, and whether
Shannon and Nyquist really knew what they were talking about. They are much
happier to have their old record player restored to health than to have to
learn or afford new-fangled technology like compact discs and (gasp!) DVDs.
Repairing this equipment is generally going to be a seat-of-the-pants affair
due to lack of easy or convenient availability of documentation - which is
generally not needed in any case if a few brain cells are assigned to the task.
Causes are nearly always simple in principle - dried up grease, deteriorated
rubber parts, broken parts (hope you can fabricate replacements!), or previous
attempts at repair.
Patients: Two LED clocks, 1 clock radio, and 1 cordless phone, all really
and truly dead.
Symptoms: Dead. What more can one say?
Testing: Plugging into the wall outlet results in no action of any kind.
No display and no sound.
So Sharon walks in with this bag of stuff. "I have some presents for you.
Larry was working on the wiring in his house."
Is Larry the same guy who toasted the CD player? (See Repair Brief #61 - Sony
D14 Portable Compact Disc Player - Smoked) "Yep, he accidentally connected the
220 to the 110...."
Geez, you really have to work at screwing up to achieve such a spectacular
disaster. Then again, Larry does have a track record :-).
What happens when 110 V devices are connected to a 220 V line?
Actually, it is quite predictable as I found out. The transformer primary
melts and opens. That is it. Period.
I tested each device by providing my own low voltage AC from a Variac feeding
a 24 V center tapped transformer substituting for the original smoked power
transformer.
In all cases, there was absolutely no damage to the electronics despite the
application of an input voltage twice what was specified. In the case of
the clocks and the clock radio, there are is no internal voltage regulation.
I can only assume that either the components were able to withstand the excess
voltage - which must have lasted long enough to burn out the transformer
primary - or that the transformer core saturated and limited the output voltage
to something much less than double.
The cordless phone was too good to pass up. Instead of attempting to
locate a UL approved transformer to install inside like the original,
I located a 9 VAC wall adapter of suitable current capacity and wired this
into the phone circuit board. It has been operating this way for several
years now.
I didn't bother actually repairing the clocks and clock radio. All I did
was confirm that they worked with my jerry-rigged power. The clocks I
keep as reminders. I cannot easily use the wall-wart trick on them as a
center tapped transformer is required. The clock radio required at least
two separate center tapped power inputs.
Comments: I, at least, would have expected more damage from such an
overvoltage. However, it would seem that just-good-enough design has
its merits if indeed core saturation saved the electronics.
Obviously, this should be a lesson to anyone doing major rewiring: don't do
it live - both for your own safety and the risks to your appliances! Have
the main power disconnected (by a licensed electrician or the utility company
if necessary). Turn all the breakers to the 'off' position or remove the
fuses. Before applying power, double check the wiring. With main power
restored, check the voltages on the circuit feeds and subpanels. Then
energize the branch circuits one at a time and check for proper operation.
I can only guess that one of the power line Hot cables touched the Neutral
bus feeding the branch circuits. It wouldn't need to be there very long.
Interestingly, at least one VCR survived with only a blown fuse - similar
to an incident where a lightning strike totally obliterated a TV but only
blew the fuse in a Panasonic VCR.
Patient: GE Model 13GP235 color TV in generally good physical condition but
Susan explains that it fell off of a bookcase.... I don't ask.
Symptoms: Dead. Normally, I would not plug a piece of dropped equipment
in until I had inspected it but apparently, this had already been
done. What the heck. In any case, nothing.
Testing: NA. Inspection comes next.
Indeed, it was not necessary to go very far to locate the problems, many
problems....
First, one of the case fastening tabs was broken off. Otherwise, there is
no visible external damage. At least it doesn't rattle!
Next, the mainboard was visibly bent around the area of the flyback transformer
which of course is the heaviest component. Depressing a lock lever allows
the mainboard to be slid about half way back and out of the chassis before
stopped by some cabling. But that is quite enough! The major damage comes
into view - a crack clear through the board running about 3 inches from the
rear near the center. It is obvious that numerous traces have been severed.
At this point I drag Susan in to see the extent of the damage. I explain
that even if I can repair all the traces - which I expect would take several
hours - there is no assurance that the CRT has not suffered some trauma which
would result in the color purity being messed up (popped or deformed slot
mask). She says "keep it". Lest you think I am taking advantage of her,
Dave and I have just repaired a nice 19" Sony of hers with the classic bad
tuner solder connections problem. Now, if she would only get it off our
lab bench....
Of course, you know, I am not going to let it go to the dumpster (actually
my parts bins/boxes/cabinets/piles without a fight!
About 6 months later, I finally manage to drag it home. I remove the back and
them turn it upside-down with a cardboard box under the rear to prevent a,
shall we say, unfortunate accident. To get at the full extent of the damage,
I only need to pull off the degauss connector and the mainboard slides out far
enough for my soldering iron.
Where through-hole pads connected to the broken trace are conveniently located
on either side of the crack, I run a short length of #24 tinned wire to the
pads. Where there is any chance of shorts, I use insulated wire.
If this is not possible because the pads are nowhere to be found - too far
away - I scrape off about 1/2" of soldermask (the green coating) on either
side of the break and use a 1" length of tinned wire to bridge the gap.
This isn't as hard as it might sound since the surface tension of the molten
solder tends to align the wire on the bare copper. These should be quite
strong. Since the board is actually supported from the rear edge by the
plastic cover (at lest under normal conditions), I believe no additional
reinforcement is needed.
Where there is a convenient pad close to one side of the break, I wrap the
wire around that but still remove the soldermask and solder to the copper on
both sides of the break.
I think I have found everything. One last inspection. No shorts. Time for
the smoke test. In order for the front panel power button work, the mainboard
must be pushed all the way back into the cabinet. Grrrr.
And - yes! We have snow. With the addition of a pair of rabbit ears, we
even have a - totally messed up picture :-(. The colors are all wrong and not
uniform across the screen. Could the CRT be ruined? Well, not to give up
yet. It is still sitting upside-down (though I did remember to plug the
degauss coil back in). There are no field adjustable ring magnets for purity
and static convergence. Better hope these are not messed up! Waving a speaker
magnet around the screen seems to help some. OK, turn it over. That seems to
be much better. Some more waving (I am too lazy to drag out a degaussing coil)
and the picture is not half bad.
Is the repair solid? I start pressing on the circuit board and - oops - the
set goes off. Then comes back on when I release it. I must have missed at
least one.
Out comes the mainboard.
Oh, the crack passes under a surface mount resistor apparently lifting one
side loose resulting in a bad connection. To fix this, I add a bit of wire
looping over the resistor end cap and connect it to a nearby through-hole pad.
Lights! Camera! No change. Pressing on the mainboard results in exactly the
same behavior, maybe a little worse. @#$%@
Out comes the mainboard once again.
Now, how could I have missed that?! A second crack, nearly as long as the
first, runs from the edge of the board near the flyback under the flyback
severing a half dozen more traces. For that matter, how did it work at all?
A half hour later, I am sure it is finished.
Sure enough, pressing and prodding now have no effect. The edge of the board
even seems straight enough to not require reinforcement. The plastic back
of the TV has prongs which will support the board. I am confident that the
repaired traces will resist anything short of another 5 foot fall - and perhaps
even that (but I don't intend to find out).
When testing the next day, I thought a new problem had developed - I was
getting 2 bands of snow drifting up the screen mostly on VHF channels. What
is this? A power supply problem? After poking, prodding, and cursing, I
went to plug it into my isolation transformer to do some probing and guess
what? The snow disappeared! It didn't return even when plugged directly into
the wall. Flakey EEPROM? No, it turned out to be interference from a power
massager being using elsewhere in the house! That device must use a non-RFI
protected interrupter.
Comments: You have probably seen the TV advertisements - I don't recall what
they were for, an insurance company, perhaps - where a late model TV is dropped
out a many story window on a bunjie cord. The set rebounds once undamaged and
without hitting a baby in a stroller passing under the window but then smashes
to smithereens on the sidewalk once the stroller had moved on. Needless to
say, this is generally not a recommended way to treat a TV set!
Ok, so your set survived in slightly better condition.....
If you take it in for service, the estimate you get may make the national debt
look like pocket change in comparison. Attempting to repair anything that has
been dropped is a very uncertain challenge - and since time is money for a
professional, spending an unknown amount of time on a single repair is very
risky. There is no harm is getting an estimate (though many shops charge for
just agreeing that what you are holding was once - say - a TV, or was it a
fishtank?)
When making repairs on printed circuit boards that have been cracked or broken,
do not be tempted to just bridge the breaks with solder even if the board has
been glued and reinforced. Solder is not compliant, bad connections will
develop over time and you *will* be returning to your handywork many times
in the future. In the case of this set, the total time to repair was about
3 hours - time well spent to assure reliability.
Stay tuned for "Magnavox 31 Inch TV - Fell on its Face" :-(.
Patient: Nutone RF Wireless Chime model LA-99N - otherwise known as a
wireless doorbell.
Symptoms: No response to button. Owner claims new batteries installed in both
button unit and base station.
Testing: No amount of fiddling with the buttons has any effect. The battery
contacts seem springy, clean, and free of corrosion.
This unit is similar in many ways to a garage door opener Tx/Rx pair. The
button unit has a five position DIP switch, presumably to set codes (and as
I discovered later, to select between chime sounds for front and back door
use). A 9 V battery powers the code generator (a single chip Motorola
SC41343) and the RF transmitter (all contained in a little shielded box)
except for a trimmer capacitor (which I don't dare touch).
After checking to make sure the batteries were indeed fresh and checking that
the button was indeed operating its contacts, the next step is to determine
whether the problem is in the button unit or base station.
How to do this?
The operating frequency is not marked and I don't really have anything to
conveniently detect a low level RF signal at no doubt hundreds of MHz anyhow.
Therefore, I punt and go on the the base station!
This is powered by 4 D cells and a single AA - what this is for, I have no
idea. There is no backup memory as far as I can tell! Removing 4 screws
from the bottom allows the cover to be removed exposing the solder side of a
2-1/2" x 7" circuit board. The antenna is connected to a point at one end.
Running my finger along some pins near the other end results in some faint
clicks from the speaker. OK, it is not totally dead.
Time for Magic Spit(tm).... Moistening my Mark 1 thumb and using this as a
probe evokes some more clicks. Placing it on suspect transistors finally
results in triggering the chimes. Therefore, I expect that this transistor
is the gate from the RF/decoder circuitry. I can fairly repeatably trigger
the chimes in this manner. A 56 K resistor from the base to collector will
also do this. Thus, I know the chimes are working.
Now for the front end. There are several ICs on the non-chimes side of the
circuit board. The two that seem to be interesting are an MC1776 which is an
op-amp and a Motorola SC41342 which appears to be the decoder mate to the IC
in the transmitter. It is connected to a four position DIP switch.
That op-amp looks promising. Even with just a DMM on pin 6, the output
clearly jumps when the remote button is pressed. Time for the scope.
The waveform at pin 6 is clean and stable when the button is pressed. It is
9 low pulses followed by 7 high pulses repeating. How does this change when
the DIP switches settings are altered?
Answer: Not much. At least, not every switch setting produces a unique bit
pattern. In fact, most of them are the same as this one which probably isn't
really even valid.
The op-amp signal goes into pin 8 of the SC41342. Pin 10 goes via a resistor
to the base of that 'trigger' transistor. As expected, there is nothing on
this pin.
Next, I decide to try all combinations of the four common DIP switches on both
transmitter and receiver as well that lone fifth switch on the transmitter.
The majority produce the same waveform. However, 3 other combinations of the
4 switches prove interesting:
* Two settings result in codes being transmitted only when the button is
released. The bit rate of this code is lower as well.
* Several settings produce 9 low and 3 high pulses.
* One and only one setting produces an alternating pattern of long and short
pulses - AND a response from the chimes!
It is erratic, however. As is apparent from the scope display, the bits are
not stable for this one 'useful' pattern. However, a good percentage of the
time, the chimes will trigger. It is at this point that I discovered the
effect of the 5th DIP switch - to change the pitch of the chimes presumably
to distinguish between the front and back doors.
There are only 6 components in the code generator - the SC41343, 3 resistors,
1 capacitor, the pushbutton, and the DIP switch. Three of these parts
obviously determine the bit rate - jumpering them with similar valued
components affects only this. I check them all anyhow. Nothing seems amiss.
The soldering is perfect.
I decide to check the battery once again - it is a little low, perhaps 8.5 V.
The owner did mention something about putting in a new battery but the button
being stuck in the depressed position. So, maybe the battery isn't so fresh
after all. This really should not matter but what the heck. However, a new
battery seems to clean up the erratic pulses.
So, in all likelihood, the SC41343 is bad resulting in a whole wad of missing
codes - only one pair seems to work - and marginal operation with a slightly
weak battery. Can a chip wear out from being on for too long? :-).
Unfortunately, an initial check shows that my usual haunts don't seem to carry
this chip. Then I email Paul Grohe: I was just wondering if you had any
datasheets for either of the following Motorola chips: SC41342 and SC41343?
"AH! Never blindly trust Motorola's search engine! (Ya gotta know where to
dig!). Here ya' go:
http://design-net.com/books/dl136/pdf/mc145026rev1.pdf
This datasheet contains the RC Tx/Rx parts. The 'SC' #s are low-power versions
of the 'MC' parts."
Sure enough, the pinouts match perfectly and Dalbani and Allied list the
MC145026 transmitter chip. With an actual datasheet in-hand, I go back
and scope all the pins but this only confirms that the chip is almost
certainly bad.
It is likely that these chips are used in a number of garage door openers as
well. With a total of 9 possible input bits, each of which may be set to any
one of 3 states (low, high, open), this results in 19,683 possible codes. That
number of combinations seems to ring a bell (no pun...) with ads I have seen.
Comments: This is not the first time I have come across a situation of this
type. My cousin's Sears garage door opener decided to ignore at least one
of the 8 DIP switches. It would work fine as long as that switch in the
transmitters and receiver was set to a 1. In the case of this doorbell, it
would appear that a much more drastic failure has occurred.
For now, I will return it to its owner but will order the replacement part.
The only problem with selecting the only working code would be if a neighbor
has the same model set to the same code! At least, it is not the default.
Patient: Tandy color TV/monitor model #16-232. This unit was described
in "Repair Brief #14: Tandy Color TV/Monitor with hum bars".
Since that initial repair after being rescued from curb-side, it
has been in daily use for about 5 years.
Symptoms: The set will randomly shut off and then come back on after a few
seconds. There appears to be some correlation with high temperature
or high humidity but not much.
Testing: The appearance of symptoms is so infrequent that I do not do anything
until it seems to quit for good....
Finally, it goes off and stays off. Oh well, something finally died....
So I unplug it and take it down to the workbench.
Plug it in - still dead.
Whack it once, whack it twice - and - it springs to life! Ok, where are
the bad connections on this one?
The previous problem had to do with startup drive to an SCR based power supply.
Once started, however, the circuitry involved was not active so the new problem
must be, well, new.
I remove the back and start the hunt.....
Soldering is pretty good except for a few spots - one is a just visible crack
around the base pin of the horizontal output transistor. Hey, that looks
promising....
Unfortunately, I am unable to confirm that soldering this pin is *the*
solution since it is so erratic. Although I was able to restore operation
with my special whack(tm), I have never been able to make it go off in
this manner. While I am in there, I also resolder a few other power
components and the video source select switch because it had such an ugly
soldering job :-(.
I also replace my startup resistor since the one I was using seemed to have
a power rating just a tad too small.
However, the cause of the erratic behavior was almost certainly the crack
at the HOT.
Operation *is* now solid but only time will tell if the problem returns. FWIW,
the TV has been running for several months without incident.
Comments: This isn't a standard bad tuner solder shield problem but similar
nonetheless. Large components heat and cool stressing the solder connections
ever so slightly each thermal cycle. After awhile it is just enough to break
the solder bond to the single sided circuit board. Then, this expansion and
contraction cause the two parts to shift ever so slightly sometimes breaking
the connection. With some kind of stress relief or reinforcement, this would
not happen. But that might cost a few cents....
I was initially prepared for a long troubleshooting adventure with this set.
There is no Sams' Photofact for this model - none even close. I just checked
the Tandy info at http://support.tandy.com/video.html and a model that might
be this one (the number is slightly different - 160-232A - but everything else
checks out) is there. However, only a parts list seems to be available.
I was thinking at first: HV shutdown or heat or humidity sensitive component
and was not at all disappointed when it turned out to be something simple.
Patient: Braun table-top quartz clock (don't they manufacture coffee makers)?
Symptoms: Clock does not run even with fresh battery. No audible signs of
life either. These don't go 'tick-tock' but there is usually
some faint click each second.
Testing: Battery is fresh and contacts are springy and clean. Gently
whacking and twisting will not start it going.
I know, most people would toss a clock of this type when it started acting
up. This one is important to the person who owns it. Some people have rabbit
feet; she has a favorite clock!
The problem with these things is that everything is small, made of plastic,
and fragile.
Anyhow, To get at the guts requires removing the back (2 screws) and popping
out the plastic 'crystal'. This done, the hands can be removed and set aside,
Now, the solder side of the circuit board is accessible.
First test: check the solenoid for continuity. I had one where the fine wire
broke - thankfully at the terminal and it was repairable. This solenoid is
fine.
Without it mounted in the case, I need to locate a substitute battery holder
to apply power.
Next, I look for pulses on the solenoid. I have already noted that the alarm
works by moving the hands to set it off. This confirms that the quartz crystal
is oscillating and the huge 8 pin chip is working, at least to some extent.
Putting a scope across the solenoid terminals reveals alternating positive
and negative nearly 1.5 V (the battery voltage) pulses at what I presume to
be a 2 second rate for an entire cycle. This would seem to indicate that it
should be working.
Therefore, the problem must be mechanical. In fact, gently rotating the
second hand shaft seems to point to a problem with the gears.....
I have to go deeper! Popping the gear cover finally reveals the problem: the
rotor of the motor is a teeny tiny cylindrical magnet glued to a plastic
pinion gear. The glue has failed. Can you believe it wasn't constructed
to military standards! A 1/4 drop of household cement and it should be as
good - no - better than new.
It is interesting that while this is a 2 pole single phase motor, the starting
direction is not ambiguous. How? It seems that there is a little plastic
tab on one side of the rotor. When the clock is mounted in the normal upright
direction, this is just enough unbalance to force the rotor into a known
orientation when unpowered and between pulses (it will be settle with this
tab at the bottom or perhaps just balanced at the top but not on one side).
From this, the N-S pole relationship is guaranteed to force the proper
clockwise rotation of the hands (at least in this part of the universe).
Clever and simple.
Reassembly is not as difficult as I feared - the gear shafts are guided to
the holes by conical bevels on the cover.
And - it works! I even got that little plastic tab oriented correctly.
Comments: OK, so you are saying "Why bother?". Maybe it is the challenge.
Maybe it is not being able to say no. Maybe I have too much junk already
and I would never throw it away if I couldn't fix it. The owner will be
happy. And, I (and perhaps, you as well) have learned something about quartz
clock movements. Not a bad deal for 1/2 hour's work.
Patient: Sony D88 'Dual Size' portable CD player. This really cute
miniature CD player was designed mostly for CD 'singles' and
is only slightly larger than a 3-1/2" disc. Normal CDs hang
out on 3 sides.
This is the first of the "Request for (dead) optical pickup series"
of Repair Briefs. This classic Sony was donated by Zev Berkovich
(ah392@freenet.toronto.on.ca).
Note: I am still looking for additional samples of this model and
other portable CD players, CDROM drives, optical pickup assemblies,
and other optical disc technology technojunk. I will pay slug mail
shipping (e.g., UPS Ground Commercial).
Symptoms: Inserting a disc and pressing PLAY results in the pickup moving
toward the center but the disc directory is never read and the
player gives up after 10 seconds or so.
Testing: The behavior is the same with several discs. It doesn't appear
as though it is even focusing correctly....
First some history - here is the dialog between Zev and myself:
Sam: Original posting (summary) -
Newsgroup: sci.electronics.repair
Subject: Request for (dead) optical pickup assemblies.
I am interested in complete or partial optical pickups or optical decks,
portable CD players, CD ROM or optical drives for parts. I will be slightly
disappointed if they are fully operational :-( Obviously, I do not want to
pay much so no offers of new Sony pickups at retail prices please! Since
I do not not make money on most of my repair activities (family, friends,
charity, the Internet, personal amusement), I am only willing to pay for
slug mail shipping and perhaps a wee bit more if there is enough fascination
value in it. In return, you *will* get priority replies to your repair
problem postings! :-)
Zev: I have a non-working Sony portable CD player (it is designed for 3"
CD's, but will also play 5" CDs with part of the disc sticking out of it).
I have the complete service manual with schematics.
Since this unit would be of little use to me even if it worked, I have
not really put too much time into it. However, I have used it to play
around with and try different things on my scope. The laser, BTW, is
fully operational and is drawing exactly the right amount of current. I
am pretty sure with some time put into it the thing will work perfectly,
I just have no use for it, and I would be happy to let you have it.
The laser unit is not compatible with later sony discman models.
Sam: Thanks, that sounds perfect - it satisfies my criteria for optical
pickups and would appear to be a really interesting item. What is its
current status?
Zev: Originally I had very carefully adjusted the focus and gain settings, and
I did manage to get it to work. After it played for a few hours though,
it just quit on me. I may have gone a little wild after that tweaking the
dials, so it's possible that they are all out of adjustment.
After I received the package...
Sam: When one compares the newer models to these, it is amazing how far the
manufacturing has come - a couple of ribbon cables/connectors and the
entire optical deck pops out - none of this flying lead stuff! BTW,
that pickup looks to be similar to those used in a number of quite new
non-Sony model portables and boomboxes including Emerson (whoever makes
them -- Mitsumi?), Aiwa, etc.
The other portables I have received based on my request for dead optics
have all had good optics! :-( They have seemed to have power problems - one
was smoked by too much voltage, another has an overheating problem.
Zev: Hmmm. I never thought about comparing the optics to other brand names.
I just compared it to the newer Discman models (D121 and up), which use
a totally different connector (one ribbon cable and a couple of two prong
snap on connectors) which I can pop in and out in a matter of seconds.
What is also quite amazing is that when you look at the evolution of
Discmen even over the past 5 years, they have gone from multiple circuit
boards that take up all available internal space, down to a tiny board
which only inhabits a fraction of the interior.
Initial testing confirmed that it failed to access the disc's directory (TOC).
An internal inspection, revealed that the unit was still set to test mode but
was behaving more like normal mode. Interesting....
Since Zev sent along the original Sony service manual along with supplements,
I at least have a starting point. Given that it seems to not want to go into
test mode now, I figure I will check some basics from the service manual.
Resetting the test mode jumpers was a royal pain! This was just the start of
pain. The jumpers are itty bitty solder pads, already slightly abused from
Zev's previous efforts through no fault of his. I proceeded in further
damaging one trace and had to add a jumper wire to bypass it.
Changing the jumpers made absolutely no difference! I double checked the
soldering... Everything appeared correct.
It didn't take long to identify an anomaly: One of the adjustments is for
+5.4 V. Once I located the test point, it measured only 4.1 to 4.2 V and
the adjustment had no effect. Hmmmm.
Sam: I started to measure some test points and found that the 5.4 V was not
being generated and the 5.4 V Adj. has no effect. This is produced by a
charge pump running off of the servo chip - it seems that the servo chip
thinks the voltage is enough but it is only 4.1 V. Reducing the input
voltage results in the DC-DC converter running but never getting up to any
reasonable voltage. Unfortunately, since it will not now go into test mode,
operation is only for a few seconds at a time.
I checked the jumpers. You left it in service mode but I guess with the
inadequate +5.4 V it behaved as though in normal mode and putting it into
normal mode made no difference.
BTW, don't you love how much pain it is to go back and forth resulting
in lifted pads etc. after awhile? Why couldn't Sony put in a switch or
at least a single jumper?!
Zev: I did have it in test mode to repair it. When I got it running again, I
put it back in normal mode. Then when it stopped playing, I remember trying
to achieve test mode again, but it was stuck in normal.
How about providing an external +5.4 V supply to see if that helps?
I guess you can see now why I got rid of it. I've repaired quite a few
of the newer Sony Discmen, but those older ones are a real pain.
I considered the idea of testing with an external power supply but decided
to use this as a last resort - I have no idea what would happen if main power
were removed with the external supply still on. There could be, shall we say,
unfortunate consequences with sneak paths through the chip protection diodes.
How are the various voltages generated in this thing? Quite cleverly in
fact: After some careful examination of the schematics (that is, once I
located the set that corresponded to the PCB revisions for this model),
it appeared as though a little circuit boosted the 4.4 V source to
5.4 V. This simplified circuit is shown below:
+4.4 V in
o +---o +5.4 V out
| |
C +-------+ +----||----+
C L703 | | | |
C | \ R708 +---/\/\---+
| D701 L704 | / 62 K | |
2SD1048 +--|>|--+---CCCC---+ \ | |\ |
Q704 | | | | +5.4 V +--|- \ |
|\ R706 |/ C _|_ C705 _|_ C706 \ Adj. | | >--+---o P
D o---| >o--/\/\--+--| X-6 --- 10 uF --- 4.7 /<----------|+ /
|/ 1K | |\ E | | uF \ | |/
IC701 / | | | | R701 /
TC7S04 R707 \ +-------+----------+-------+ 22 K \ (The amp is
10 K / | | | / part of IC504)
| | _|_ | |
+----+ - +--------+
One of the large chips is the controller for this switching DC to DC converter.
Of course, I at first suspected this many legged surface mount chip! However,
I should have known better. The voltages associated with the converter were
way off and the drive output to the boost circuit appeared mostly dead. This
came directly from the large chip, IC504.
The regulator includes a difference amp (reference-fraction of output voltage),
pulse width modulator (sawtooth clock and comparator), and a driver. I finally
was able to convince myself that the amp was working since it output and
feedback components are external and I could see the output change when the
relative values of the two inputs were adjusted.
The leaky integrator provides the control voltage to the pulse width modulator
that drives the DC-DC converter. The op amp circuit is internal to the motor
driver chip, IC504). The output, P, is compared with a sawtooth derived from
a clock signal. The result is the charge pump drive.
Then, I thought the driver was dead. However, upon close examination, there
appeared to be an occasional burst of activity - but no corresponding change
in output voltage. The inverter was the complement of the driver output. So,
maybe that boost transistor circuit is not working.
Unfortunately, Q704, R706, R707, and C704 are all mounted on the underside
of the circuit board - all surface mount and inaccessible from the top.
Getting to the bottom is a real treat mainly because two printed cables and
numerous thin wires interconnect the main logic board with the optical
deck and microprocessor board (the control panel in the hinged top of the
case). However, after undoing a bunch of black tape and pulling one of the
printed cables, I can *just* get to the required components without needing
to unsolder anything.
Tests with DMM: Resistors - A-OK, Capacitor - Not shorted.
Transistor - What have we here??? Finally something that is a bit strange:
The B-C junction has a normal .682 V drop on my DMM but the B-E junction is
1.29 V? I double check. Yes, indeed, the junction seems to be bad. Joy!
It is an SO23 surface mount part, barely visible let along replaceable.
Now, where can I find a replacement? It is probably just a vanilla flavored
transistor - a 2N3904 would probably be fine. (In fact, I checked while
writing up this Repair Brief and the 2SD1048 is a general purpose switching
transistor). However, I would rather use an exact match to be sure.
Now, where can I find one? Hey, remember that smoked Sony D14 (Repair Brief
#61 - Sony D14 Portable Compact Disc Player - Smoked)? I bet it has some
usable body parts that were not damaged. After all, most of its circuitry
was on the output side of its DC-DC converter module (but not the controller,
unfortunately.
Sure enough, there are numerous similar transistors - type X-6.
I manage to mangle one of them because they are glued in addition to being
soldered to the board. Finally, using a combination of my soldering iron,
SoldaPullit, jeweler's screwdriver, Xacto knife, and needlenose pliers, one
comes free in good condition. Confirming this with a multimeter is fun as
well since it is barely possible to even hold the thing without it popping
out of existence.
The broken transistor is removed more easily since it is not glued - heating
the solder while prying on one side frees it.
I use a set of tweezers to position the replacement and a wooden clothes pin
to hold it in place while soldering. You would have to see this setup to
believe it :-).
Now for power!
At first I thought there was no change. Did the new part blow already? But
then I realized that I had been playing with the +5.4 V Adj. pot and left it
where the two inputs to the op amp were just about equal - which would yield
around 4.1 V even with the DC-DC converter operating properly.
And, indeed, now the +5.4 V adjustment has an effect. I would have been
surprised if the player had actually started working as you will recall that
the internal adjustments are probably way off. In fact, the behavior doesn't
appear much changed at all :-(.
However, now, I can try some experimenting with the adjustments. I solder a
wire to the RF test point for my scope.
Nothing. OK, let's see if Focus Bias (offset) does anything.
Nothing. What about Focus Gain?
Nothing. How about back and forth?
Oops, what was that? Is that an eye pattern? Well, of sorts. Can I
persuade it to stick around.
Adjust. Adjust. Adjust. Adjust. Adjust.
Tweak. Tweak. Tweak. Tweak. Tweak.
Finally, it seems to be fairly stable though not the prettiest eye pattern
I have ever seen. Will it play? I have been using my trusty AOL CD which
really leaves something to be desired in the audio department... Of course,
it isn't much use for getting on-line these day either :-).
Sure enough, my 'Strauss Greatest Hits' disc does play - a bit noisy but
recognizable. It seems to have some trouble locating and playing outer
tracks so a little more tweaking is in order. A slight shift to the PLL
free run frequency and it now is somewhat happier.
Someday, I will perform the proper servo alignment though I do not think any
of the adjustments are far off. Nonetheless, I might as well take advantage
of the luxury of having an actual service manual!
Unfortunately, another disc played with a lot of noise and it seemed to be
touchy about my even going near the player! I checked to make sure the +5.4 V
had not died again - nope, exactly correct. Oh, well, enough for now. Stay
tuned for: "Sony D88 Portable CD Player Part 2 - Erratic Audio Noise".
Since I did not attempt any adjustments prior to replacing the transistor,
I really don't even know if the correct voltage on that +5.4 V bus really
made any difference but I am not complaining. I also do not know what
caused the original transistor to blow - it might fail if the drive were
on for more than a few microseconds as the inductor is the only thing
limiting collector current. I left if playing on repeat mode to see how
robust it is.
Comments: This is a really cute CD player - not really very practical but
probably one of the neater examples of solidly built Sony technology which
still contains some metal in its structure!. I would be very interested in
obtaining other samples of this or similar vintage CD Sony players.
Would a repair shop ever have taken the time I did to analyze the circuit and
locate and replace a surface mount part? No way. Only a Sony service depot
would likely get that deep into anything of this sort and then again, probably
not even there.
Patient: Sony D88 'Dual Size' portable CD players. This really cute
miniature CD player was designed mostly for CD 'singles' and
is only slightly larger than a 3-1/2" disc. Normal CDs hang
out on 3 sides.
This is the second of the "Request for (dead) optical pickup series"
of Repair Briefs and is the continuing saga of this classic Sony CD
player donated by Zev Berkovich (ah392@freenet.toronto.on.ca).
Symptoms: After locating and replacing a bad SMT transistor and retweaking
the servo adjustments, the player works but playback with some discs
in particular is noisy at times - sometimes very noisy. Seek times
are sometimes excessive. Moving near the player seems to have a
significant effect as well - often to the point of killing tracking
and playback entirely.
Testing: I substituted another AC adapter just to make sure it wasn't at
fault. Tweaking the servo adjustments has minimal effect.
As you will recall in "Repair Brief #73: Sony D88 Portable CD Player Part 1 -
Does Not Recognize Discs", this player was totally dead when given to me. Zev,
the original owner, said that he had tweaked the servo adjustments and it
worked for a few hours and then died. He then did some more tweaking without
success.
After a feature length troubleshooting session, I was able to locate a bad
surface mount transistor - part of the +5.4 V boost DC-DC converter - and
replaced it. I was then able to get the player to work somewhat even if it
does appear to have an attitude problem......
A few days after the initial success I decided to try another disc and was
greeted with terrible noise which seemed to correlate with the disc rotation
rate. What's up? The eye pattern didn't look great. In fact, come to think
of it, the eye pattern looked terrible. Therefore I decided to do a little
more investigation. (As it turned out, a lot more investigation - an entire
afternoon's worth!).
With service manual in-hand, I checked the laser current and eye-pattern
amplitude;
* Laser current was 50 mA - the sticker said 47 mA - well within the
acceptable range of +5,-11 mA).
* Eye pattern amplitude was .7 V - just a tad short of the lower limit of
.75 V but probably acceptable.
Since there doesn't appear to be anything wrong with the optical deck or
pickup, there is no point in pursuing it. The laser power is at its correct
setting (there is no adjustment in any case) based on the current measurements.
The lens and its suspension appears well within normal limits. I did test the
spindle motor which seems fine with no evidence of shorts or dead spots (the
symptoms do not really apply to a bad spindle motor anyhow).
However, on closer examination of the eye-pattern with a better scope (350 Mhz
bandwidth), its ugliness became even more apparent. In fact, there seemed to
be a 100 MHz oscillation superimposed on the signal. Is this an artifact of
my scope probe? I do not think so. Could the laserdiode actually be
oscillating at 100 Mhz? Oh come on, that is grasping at straws.
Another interesting behavioral quirk was that the disc didn't like me! Recall
that this is a strange player - 5-1/4 inch normal CDs hang over the side. If
I moved my finger *near* the spinning disc, the audio becomes noisier and will
cutoff entirely when close enough. This did not appear to be mechanical - I
could be 1/2" away from the disc and still detect a significant change in the
audio. At first I thought this was a problem with a component in the player
but it was the disc - placing a strip of aluminum I was holding near the disc
produced the same result! Well, after some thought, the only conclusion could
be that my finger was capacitively coupling to the aluminum of the disc's
information layer and *that* was capacitively coupling to the metal parts of
the optical deck. Shouldn't these be connected to an analog reference point
or analog ground? Of course - and there was even a hint staring me in the
face - a taped unconnected wire. However, there is no mention of this in the
service manual or schematic!
There is a shield which is part of the printed flex cable which carries the
laserdiode power/feedback and photodiode signals and this is soldered to the
metal cover plate of the optical deck. However, it wasn't connected anywhere
else! At first I ignored that dangling wire and soldered it to what I thought
was a suitable ground.
Wrong! That connection didn't work at all - it must have been a noisy digital
or power ground. In fact, I wasted another hour tweaking the servo adjustments
and then ripping the player apart once again suspecting (incorrectly) that my
efforts to attach the ground wire and clean the optics while I had the unit
disassembled had damaged something else! All was not well (at least as well
as it was before) until I removed my original ground wire completely. The
player came back to life but as expected was just as noisy and tempermental.
If I had followed Rule #45123 - never change or fix more than one thing at a
time without testing, I could have saved a lot to time and aggravation.
Finally, I came to my senses and connected a wire to the proper solder pad
and - presto! - the audio noise completely disappeared and the RF test point
now looked like a much more respectable eye pattern - quite nice in fact.
Obviously, the sheet metal of the optical deck cover was picking up all kinds
of crosstalk from the electronics - perhaps even forming a feedback loop of
sorts - and totally corrupting the RF signal (and who knows what else).
A little more tweaking to optimize the amplitude and appearance of the eye
pattern (With the player in the upright position and the controls accessed
from underneath - joy!) as well as adjustment of the PLL free-run frequency
(according to the service manual, no less!) and I do believe the player is
in pretty good operating condition. Since even the service manual doesn't
provide procedures for adjusting focus and tracking gain, I left these at
their midpoints. This may not be optimal but seems to be quite acceptable
for all the discs I tried. (Gain of these servos affect mostly the
performance with dirty, scratched, and/or warped discs.)
Comments: I still do not know if the original problem with the +5.4 V would
result in a totally dead player. It seems to mostly feed the audio circuits
and somehow also is involved in the selection of TEST mode. Maybe someday I
will tweak it back down to 4.1 V and see what happens (yeh, sure). Perhaps
it would totally kill audio or simply result in screwball behavior.
This is a case where the previous attempted repair probably resulted in this
entire problem - no fault of Zev's). Without a mention in the service manual
or schematic, it required a visual inspection to finally identify the missing
shield connection. However, this does underscore the nature of possible
difficulties in servicing equipment that has been worked on by others.
I have seen erratic behavior caused be a missing shield in CDROM drives
(see Repair Brief #52: Aztech CDA-268-01A CDROM Drive - Drawer Continuously
Closing".) In that unit, even audio play was noisy until a circuit board
metal shield cover was replaced. Thus, whenever you repair a CD player,
CDROM or similar device, pay attention to the ground strap, jumper, or
spring clip when you remove the optical deck!
Patient: Old but solid state at least - Panasonic 19 inch color TV model
CT-714.
Symptoms: Totally dead.
Testing: Tried different outlets, nothing.
I didn't locate the circuit breaker until much later - probably just as well.
Other components may have let out their smoke if I had pushed the button.
First, I searched for any fuses. I only found one on the mainboard and it
tested good.
Then, I tested resistance to ground at HOT - dead short.
Removed HOT - tests good. Still dead short at socket.
Removed red wire (whatever that it) - still shorted at socket.
In order to get at the components on the HOT socket requires removing the
HV multiplier/focus block (4 screws, 3 push-on connectors - better remember
where they go!) and popping the socket from the heat sink where it snaps in
place.
Once this is done, I unsoldered damper and snubber - these are shorted.
Removing the capacitor shows that the damper is shorted.
I tested other power semiconductors - a transistor near the main filter
capacitor tests good. A T03 package device which turned out to be an SCR
(probably part of power supply regulator) tests strange G-K low in both
directions; A-K, high impedance in both directions. Maybe it is good.
Perhaps, there is an internal resistor that is confusing my meter.
Before proceeding, I go get the Sams' Photofact from the library.
Unfortunately, this set is at least 20 years old and the Sams' folder is on
microfilm - an actual spool! (Or so I was led to believe. A later visit
resulted in the hard copy magically appearing from a back room. Grrr.)
This was actually not too bad as the microfilm reader had variable zoom and
pan so with the assistance of the person in charge, I could get exactly what
I wanted (after waiting for all those other people to copy old newspaper and
other articles). The bad part was that the copies were only marginally
legible but probably good enough though some identifying numbers could only
be guessed.
I decide to replace the HOT and damper with a transistor that has a built in
damper diode. I keep some in stock for general replacement in TVs (though I
would not do this for an SVGA monitor). One nice one is a 2SD871D. The
advantage of this is that if there is still a problem, I don't have to take
out the tripler and socket - just pop the transistor.
I will power it using a Variac AND series light bulb and check the HOT for
excess temperature rise periodically.
However, first I will leave the HOT out entirely just to make sure the
front-end of the power supply works.
I connect the the Variac to the output of my isolation transformer in series
with the light bulb.
At first - nothing. Series light bulb remains dark.
Then, I found the circuit breaker and pushed the reset button. There was
a satisfying click :-).
Now, my series light bulb (100 W) glows brightly (degaussing) and then nearly
goes out. This is the correct behavior with almost no load (horizontal
deflection not running).
Checking voltage at the regulator with my multimeter - it doesn't come on
at all until the Variac is 2/3 of the way up but then stabilizes close
enough to what it should be - 110 V - that I feel confident in proceeding.
I pull the plug and set about installing my replacement HOT/damper combo.
At least the main filter capacitor bleeds off in only a few seconds. Just
double check - no need to discharge it.
It is somewhat difficult to get at the pins to solder in the cramped quarters
between the CRT yoke/convergence board. This is one of those TVs where there
are zillions of convergence adjustments which I have no intention of touching!
Finally, it is done (so I think). Snap the socket back in place, install the
HOT and tighten the screws securely. No heat sink compound was apparently
used - perhaps before I button it up, I will add some if it seems to be needed.
Replace the HV multiplier/focus block. Oops, where did THAT lead go. Oh,
but there are two pins, now who added that second pin? :-) I was just able
to make out scrape marks on one and stuck the wire there. (It seems there
are two possible locations for the focus wire, for whatever reason.)
All set?
Power - slowly ramp up the Variac. Light bulb glows brightly and then settles
back to still glowing somewhat brightly though not full on (which would have
indicated a short). The voltage on the power supply now reads 60 V.
I am not ready to try a larger light bulb - 100 W should work.
I try disconnecting that red wire - still don't know what it is - but that
makes no difference. Removing the HV multiplier/focus block also makes no
difference. Could the flyback be shorted?
There is one thing - my combination of the Variac powering the isolation
transformer is not what I usually use so I put the isolation transformer first.
Now, for whatever reason, at nearly full voltage, I suddenly get a burst of
static and a few seconds later, just a hit of a raster and picture - about
1/2 size and rolling.
I power down and check the HOT - cool as a cucumber.
OK, what about a larger light bulb?
I find a 150 W PAR floodlight bulb.
Now, I get something much closer to normal. A little squashed top and bottom
but nearly or totally full width.
Go for it!
Leaving out the light bulb but still using the Variac, a not too terrible
picture appears at about 90 V or so and is fairly independent of input
voltage. An antenna helps quite a bit but the picture is still pretty ugly.
The vertical is squashed top and bottom and pulsating apparently along with
some hum bars rolling up the screen.
I let it run for a few minutes and pull the plug.
HOT is still cool. OK, so it, at least, is happy. What about that ugly
picture.
Adjusting V-height and V-linearity are not able to get it full screen top and
bottom simultaneously and there is still that pulsing behavior.
Well, probably a bad capacitor.. There are only about a half dozen
electrolytics in the vertical output stage. Locating part numbers is a bit
of a pain with the poor quality copies but the forth one I test reads way low
on my capacitor checker (part of one of my cheap Radio Shack DMMs). It is a
bypass cap on the collector of the vertical output pair (yes, discrete
transistors).
Jumpering a good capacitor across this returns the vertical deflection to
something respectible and the adjustments now are able to completely restore
normal operation of the vertical.
Then, I notice a slight double wiggle moving up the screen. No obvious hum
bars but just a 1/8" or so wiggle. I check the obvious - main filter capacitor
(actually after the SCR regulator) and a couple of other bypass caps in the
power supply but jumpering across them with good capacitors has no noticeable
effect.
There is a linear regulator (called Automatic Picture Stabilizer - APS - or
something like that) after the SCR regulator. The output of the SCR regulator
according to the Sams' should have about 6 V of ripple. I check it and indeed,
there is about 5 V of ripple. However, there is exactly the same ripple on the
*output* of the linear regulator. As a matter of fact, the output of the
linear regulator measures exactly the same as the input! I thought I checked
that transistor.....
Testing on the circuit board, B and C are shorted. So, either it is now bad
or some other component is shorted. It didn't take long to locate the problem.
A zener diode apparently there to protect against too high a voltage drop on
the transistor was a dead short. At first, I was not able to identify the
part from its markings but a 5.1 V 1 W zener seems to work fine and ran
cool. The wiggle is gone. Ripple on the 110 V output is almost undetectable
on the scope. (Later, I went back to the library and confirmed that it was
supposed to be a 15 V 1 W zener - which fortunately, Radio Shack stocks - and
even a genuine Motorola part!
I was all set to return it to its owner when the next morning I noticed what
could only be called an instability - for the first couple of minutes after
power-on, the raster would tend to jiggle or perhaps change size and brightness
erratically. After it warmed up, everything would be fine as though the TV
was taunting me - "You think I am repaired. HeHeHe". I hate those sorts of
problems! Not enough time to probe anything.
One reason I did not notice it before was perhaps that I had been switching
power to the TV rather than using its on/off switch. This is one of those
'instant-on' TVs. The picture really does appear within 2 seconds of pulling
its power knob as it keeps the CRT filaments hot at all times. Using an
external power switch, of course, killed power to the filaments as well and
thus the worst of the jiggling was probably gone before the picture appeared.
Only when I put the set on a Variac was I able to actually see the supply
going into a low frequency oscillation (about 10 Hz) at high line voltage
even after it warmed up (though it was less susceptible after warmup). And,
I had actually seen some hint of it before on the scope but it went away
quickly and there were no operational symptoms. Only when this added ripple
really became large was there any evidence of it in the picture.
Assuming it was a bad capacitor, I systematically removed and tested each and
every electrolytic in the power supply regulator. Well, it wasn't so bad -
there are only half a dozen or so - except that since my photocopies of the
Sams' sheets are almost illegible, it takes a little creative parts
identification to find each one. Of course, it was the last one! A 1 uF,
150 V capacitor was reading between .01 and .3 uF depending on what capacitor
range my meter was on. An organ donation of another dead set took care of
that. Regulation stabilized throughout the relevant range of my Variac.
There is still a little bit of video noise - a slight pattern of vertical
undulations - at the left side of the screen on some channels but I could
spend the rest of my life checking capacitors in this thing. I suspect more
are on their way out......
Comments: It could be that these four problems were totally unrelated though
the zener probably failed at the time the damper shorted. Certainly, the
bad capacitor in the vertical deflection just died of old age and the owner
probably never noticed. If there was a cause for the failed damper diode
in the first place, it could have been the bad capacitor in the power supply
totally messing up the main B+ output. Perhaps, it really got out of control
if the set were off for an extended period of time. In fact, the TV had not
been used for about a month prior to its failure though I do not know if it
died the first time it was turned on after its vacation.
Patient: Emerson 14" VGA Color Monitor. This was Ed's PC monitor until
he was forced to 'upgrade' to a NEC 3D. It then was used on the
PAL burning PC which meant that it was often forgotten and left
on all day and all night.....
Symptoms: While on and unattended, this monitor started making a high pitched
whine with loss of picture.
Testing: Connections to the PC were verified and a substitute monitor
worked fine on the PC. This is standard DOS boot and VGA on
a nothing-to-write-home-about video card so there is no issue
of incompatible scan rates.
Most often, a dead monitor emitting a high pitched whine indicates a
short circuit in the secondary of the power supply, horizontal output
transistor, flyback transformer, or one of its secondary loads. In
many cases, this would also result in failure of the SMPS chopper and/or
horizontal output transistor but not always.
This monitor has a separate switching power supply connected to the
mainboard via a removable cable. Although no schematics were available,
measurements were made of the voltages on this connector both with
the mainboard connected and removed. On average, the values were about
half the unloaded values when attached to the mainboard. The whine
disappeared when the mainboard was unplugged.
While this most likely indicates a fault on the main board, it could also
indicate a power supply that has (very) poor regulation.
A quick check was made of all the visible semiconductors on the mainboard
including the horizontal output transistor (HOT), I did not locate any shorts.
One diode that was initially suspect turned out to be in parallel with a 4.7
ohm resistor as part of the HOT base drive circuit.
Without schematics, there was no way of knowing the correct voltage values.
I decided to do a little reverse engineering of the primary output of the
switching supply - the one connected to the feedback optoisolator. This took
roughly 10 minutes and revealed that the open circuit voltages were correct.
The reference was supplied by a 6.2 V zener diode and a simple transistor
controlling the optoisolator - turning it on when the output voltage
exceeded 120 V. While the supply could still be bad, the probability
of this is small based on this new information.
Next, the flyback transformer was removed from the circuit (after verifying
that the CRT anode was discharged). Using my chopper based flyback tester,
no faults were found. However, since this device does not test at full
voltage so a breakdown failure was still possible.
Next, the flyback was connected via a pair of wires to only the HOT collector
and B+ source to guarantee that no secondary loads were the source of the
failure. Behavior in this case was unchanged.
Probing the collector of the HOT with a scope showed a grossly abnormal
waveform. (Note that probing of the collector of the HOT is usually to be
avoided in a working monitor or TV since the flyback spikes can exceed 1 KV.
However, with the reduced power supply voltages due to the fault condition,
this was considered safe). The 'on' portion of the waveform starts out near
ground (correct) but starts ramping up toward the end of the sweep. The 'off'
time shows a great deal of ringing. It was thought that this now pointed back
to a bad flyback.
The base drive signal looked fine in terms of shape and levels. With normal
drive signals, the ramped up 'on' portion of the waveform indicates a flyback
of greatly reduced inductance and Q permitting excessive current to flow toward
the end of the sweep - no doubt due to a short or internal arc.
Although the possibility of a bad HOT is small, it was an easy thing to
substitute a similar transistor. There was no change.
At this point by the process of elimination, it was thought that the flyback
transformer was the only other possibility. A closer inspection - actually
a closer listening - revealed that in the instant that power is applied,
there is a brief snap as though something is arcing internally following
by a very faint sizzle indicating a continuous internal arc. Perhaps it
was wishful thinking....
On the basis of this, a new flyback transformer was ordered. With great
expectations, the new flyback is installed but what is this - no change.
Same low voltage, same screwy waveform - darn :-(.
What next? Well, we have not tried substituting parts around the HOT.
So, one by one, the caps and damper diode are removed. Using the Variac
should be safe as an indication of a major improvement should be obvious
before the HOT can be destroyed by lack of the snubber caps or damper diode.
There, however, does not change in behavior in the least.
Give up for a while....
OK, there is only one more major item that has not been checked - the
deflection yoke. A shorted yoke is always a possibility. How to test?
Well, it is on a plug - remove it, no interlocks.
Running the voltage up on the Variac - poof. Oops, too late realizing that
the parts around the HOT have not been replaced. However, this is progress.
If there was no problem with the yoke circuitry, there should not have been
a major change in pulling the yoke. Now, there was enough flyback kick on
the HOT to have blown it.
OK, rummage around for a replacement. A BU508A should be an acceptable
replacement for a BU508V at least for testing.
Replace all components and solder in new HOT. Monitor B+ while running
up line voltage. B+ climbs nicely through the 60 V barrier and stabilizes
at 120 V. While watching both front and back of CRT - after all there is
no deflection and the screen can get a permanent navel quite quickly - there
is now filament power with normal brightness of the filaments. This indicates
that the flyback is probably running at a normal power level. Nothing on the
screen yet. Maybe the monitor has a shutdown circuit that prevents anything
from showing up without a video signal. Connect PC. Still nothing. Try user
brightness and contrast. Still nothing. Finally, try the flyback screen knob.
Now I can get a spot indicating that there is HV. Quickly turn it back down to
protect CRT.
Next: deflection yoke testing. (Actually, about a year passed.)
So, how does one test a deflection yoke?
Here is a brief tutorial:
The deflection yoke consists of the horizontal coils and vertical coils (wound
on a ferrite core), and mounting structure. There may also be some additional
windings on the same assembly in some designs for various functions like
pincushion correction and raster rotation. The following deals only with the
actual deflection coils - the others can be tested in a similar manner.
If possible, compare all measurements with a known good identical deflection
yoke. Of course, if you have one, swapping is the fastest surest test of all!
Note: it doesn't have to be mounted on the CRT which would disturb purity and
convergence adjustments but see the caution above about drilling holes in the
CRT face plate!
* Horizontal - the horizontal section consists of an even number of windings
hooked up in parallel with half of the windings on each of the two ferrite
core pieces. In rare instances there may be other components (like the
S-correction capacitor) on the same assembly.
The horizontal windings will be oriented with the coil's axis vertical and
mounted on the inside of the yoke (against the CRT neck/funnel). It may be
wound with thicker wire than that used for the vertical windings.
- Resistance check - This may be possible without removing the yoke from
the CRT if the terminal block is accessible. Disconnect the individual
windings from each another and determine if the resistances are nearly
equal. Check for shorts between windings and between the horizontal and
vertical windings as well.
Typical resistance of the intact windings (at the yoke connector assuming
no other components): TV or NTSC/PAL monitor - a few ohms (3 ohms typical),
SVGA monitor - less than an ohm (.5 ohms typical).
- Inspection - Look for charring or other evidence of insulation breakdown
due to arcing or overheating. For the horizontal windings, this will
require removing the yoke from the CRT since little if any of the windings
are visible from the outside. However, even then, most of the windings
are hidden under layers of wire or behind the ferrite core.
- Ring test. See the document "Testing of Flyback (LOPT) Transformers".
This deals with flyback transformers but the principles are the same.
Disconnecting the windings may help isolate the location of a fault.
However, for windings wound on the same core, the inductive coupling
will result in a short anywhere on that core reducing the Q.
* Vertical - The vertical section is usually manufactured as a pair of windings
wired in parallel (or maybe in series).
The vertical windings will be oriented with the coil's axis horizontal and
wound on the outside of the yoke. The wire used for the vertical winding
may be thinner than that used for the horizontal windings.
- Resistance check - This may be possible without removing the yoke from
the CRT if the terminal block is accessible. Disconnect the individual
windings from each other and determine if the resistances are nearly
equal. Check for shorts between windings and between the horizontal
and vertical windings as well.
Typical resistance of the intact windings (at the yoke connector assuming
no other components): TV or NTSC/PAL monitor - more than 10 ohms (15 ohms
typical), SVGA monitor - at least a few ohms (5 ohms typical).
- Inspection - Look for charring or other evidence of insulation breakdown
due to arcing or overheating. The accessible portions of the vertical
windings are mostly visible without removing the yoke from the CRT.
However, most of the windings are hidden under layers of wire or behind
the ferrite core.
- Ring test - Since the Q of the typical vertical winding is very low, a
ring test may not yield much useful information.
End of Deflection Yoke Testing 101.
In the case of this Emerson, the horizontal winding is actually 4 pairs of
windings in parallel. It measures about .6 ohms. I use the first approach
and unsolder the individual windings at one end to measure resistance. This
is easy since the terminal block is at the top of the yoke under a plastic
cover which snaps off.
Almost immediately, there is indication of a major anomaly: 4 of the windings
measure about 4.7 ohms but the other 4 vary from 3.3 to 3.7 ohms. OK, so
there ARE some shorted turns somewhere. Now, this is progress!
Since nothing is visibly charred on the outside, I draw a diagram of the
precise position and orientation of the yoke and purity/static convergence
magnet assembly and remove the yoke.
Wow! That is surely the problem. In one corner, where the windings are
possibly near the tube DAG coating, there is a black charred spot. It
seems to involve a number of individual wires.
Can I repair it? Possibly. With a pointy jeweler's screwdriver, I was able
to separate approximately 8 or 10 wires from the windings that appear to have
their insulation burned off. Oops! One broke - possibly due to excess force,
possibly due to prior damage. Oh well, I will have to jumper that one.
Once they were all separate, I used a Q-tip and alcohol to thoroughly clean
everything in the vicinity. There did not appear to be any other damaged
wires. Then, I wrapped each one with a layer of plastic electrical tape. I
repaired the broken wire with a piece of bare wire and then wrapped it with
electrical tape. The repair is now well insulated though I do not know how
well the plastic tape will hold up to heat, if any. I may go back and replace
the tape in the future if this repair turns out to be successful.
Getting the yoke and magnet assembly back in place is no problem - there was
some glue originally holding the yoke to the convergence wedges and the break
in this showed exactly how it should be positioned. The magnet assembly is
probably close enough. In any case, these can be adjusted later.
Ready? Power!
Guess what? No whine and big bright screen - way out of focus - appears
almost immediately. Well, I don't know what it should do without a signal,
so I move the whole affair over to the PC on which it was on originally.
Boot the PC, then power the monitor! Same behavior. Then, it occurs to me
that the focus and screen (G2) pots on the flyback were never really adjusted
after the initial test of the flyback because there had never been a raster
with the replacement flyback.
Sure enough, a little twiddling and there is a picture. Then some fiddling
with the yoke orientation to get it straight and the convergence magnets to
eliminate fringing. Ouch! What was that? A stray wire bit me... It looks
like one of those windings was never reconnected. (Maybe the picture was a
bit narrow..) I will fix that and try to be more careful.
For the first time in a couple of years, this monitor is working - better
than new.
Comments: Had a substitute flyback transformer been available, this process
could have been shortened considerably. In this case the tester should
have been believed.
One nice feature of the design of this monitor (whose specs are really
nothing to write home about) is that the switching supply is fairly well
short circuit protected and current limited. Therefore, it is probably
virtually impossible for any fault to destroy the HOT through overcurrent
(though apparently not through overvoltage) or other power components on
the mainboard. Even with the shorted HOT, the power supply just whined
a bit but was perfectly content despite that fact that the B+ line was
dragged down to 3 V.
What caused the yoke to arc? There is no way of knowing but the location
at a corner suggests damaged insulation. Magnet wire insulation is very
fragile to begin with so any kind of scuffing during manufacturing is a
failure waiting to happen. Perhaps it was bent too far at some point and
then straightened out or scraped against something. Perhaps, expansion and
contraction from thermal cycling caused it to rub against the CRT. Given
the characteristics of magnet wire, it is quite amazing that these types of
failures are not more common.
Patient: Really BIG fairly old 25 inch Sony console.
Symptoms: Pressing the power button results in a relay click and some evidence
of high voltage (a bit of static on the screen) but no picture and
only a slight hum from the speaker. Releasing the power button
results in it shutting off. The standby LED is flickering.
Testing: No additional tests possible without getting inside.
When Frank asked me if I would take a look at a TV for him, I had no idea it
was such a BIG TV. How will we put it up on the bench? And, for that matter,
how will be hide it in the lab when those VIP types come around? After all,
TV repair is not supposed to be our main occupation. (The latter problem was
easily solved at least - it ended up in my office!)
Than I find out that it is Frank's landlord's TV....
This is most likely a problem in the standby power supply. Best to go get
the Sams' Photofact before poking around. Checking the Howard Sams' web
site (of course everything is done on the Internet nowadays even if checking
the hard copy Sams' index is easier) reveals that there is no Folder for
the KV-2675R. The closest is the KV-2670R. Well, that will have to do.
Hopefully, the Sony engineers didn't get carried away with 'improvements'
between those two models. I only photocopied the AC input/standby power
supply and microcontroller pages assuming that since there were indications
of HV, the deflection and signal circuits are probably fine. Hopefully.
First problem: how to get at the circuit board? Well, it isn't finding it or
even the available free space around it - the cabinet is *mostly* empty space.
Rather, working on the floor is just not convenient! So, Dave and I clear
some bench space and man-handle this console up onto a regulation lab bench.
I put some blocks underneath to prevent it from rolling off (it is on wheels).
Now, it looks kind of strange but access is perfect. At least no VIPs are
expected that day!
A couple of snaps, release a few cables, and the solder side of the mainboard
is accessible. The components are even labeled.
Hey Dave, check the voltage on C114. "Uh, I don't see a C114". How about
C115? "No, wait, yes. Around 1.5 V". It should be +5.1 V. OK, how about
across C609? "Wait...... 47 V." Are you sure you are on C609? It should
be 129 V. "Yep, 47 V." What is connected to it? "Just a sec.... From the
plus side, D603 bar, C608, C114 plus,..." What about the minus side? "Let's
see, L604, C114 minus,.." OK, that checks out. Here, I brought some caps
just in case... Put this in place of C609. 10 uF, 150 V.
A couple minutes later. Ready, plug it in.
At once, there is lots of static on the screen and few seconds later, snow
appears. As usual, reception even with our rabbit ears is terrible in the
lab. We finally find a UHF channel that has a signal-to-noise ratio greater
than about .1. A broken VCR (at least the blue screen wasn't broken) worked
fine. Later, we would leave it on Channel 57 or something playing the
usual afternoon cartoons.
I later tested the original C609 and found it to be totally open. Its rating
was 10 uF, 250 V. A dead TV contributed one that was 10 uF, 200 V. That will
have to do since Frank is leaving the country in a week (I wonder why!) to
return home and there is no time for an MCM order. The maximum voltage on it
should be less than about 160 V under any conditions so this should be fine.
At this point, the only remaining problem is that the TV seems to be finicky
about powering on and off. Not due to any power supply problem but rather
it appears to interpret the button press erratically. Suspecting contact
bounce due to a dirty pushbutton switch, I had Dave swap two of the switches
but this made no difference. Once it comes on, it is solid and once it
goes off it is solid. Just, sometimes it will decide not to cooperate.
This seems to be worse when it has been on for a while. Another capacitor?
I asked Dave to bring in a Sony remote control to see if it behaved the same
way with that. I wasn't around when he did so but claimed that "it 'worked
well enough for me' so I left it at that and told Frank to get the BIG TV out
of the lab"! Well, at least Dave knows when to quit.
Comments: The flickering standby LED kind of guaranteed that it would be
a problem with the standby power supply. Since there was some power, it
had to be a simple problem like a dried up capacitor. Fortunately, this
Sony was close enough to the model for which I has service data - most of
the part numbers seemed to match - that finding the culprit was very easy.
However, even if it had not been the case, some quick checks of the parts
in the vicinity of the AC line input would have found the bad cap in short
order.
Patient: Magnavox 31" TV. This is what one would call one unlucky set :-(.
First, it was used for a jungle gym by a colleague's 2 year old and
fell off a low platform smack on its face (fortunately, no one was
injured). About 6 months later, a truck hit a substation transformer
resulting in excessive voltage on the 110 VAC lines. The resulting
county wide power surge blew electronic equipment in dozens if not
hundreds of residences. I am sure the insurance companies just
loved it! (No claim was filed on this set, however, because it was
already in bad shape - there are still some honest people in the
world.)
Symptoms: Originally, colors were messed up across the screen resulting in
areas of colors shifted between red, green, and blue. After the
power surge, it was totally dead... I am not sure which was really
the worse situation.
Testing: The front bezel was slightly cracked and there was an ominous
rattle from inside. After the power surge, well....
When I first saw this disaster - the first time - the complaint was that all
the colors were messed up. Once I was told the story, it was nearly certain
that the cause was something mechanical - unfortunately likely to be in an
inaccessible location inside the CRT.
First step: Remove back. Actually, let Tom remove the back :-).
The only visible damage appears to be that the mainboard broke loose and is
just sitting on the bottom of the cabinet. No real damage to the board itself
or the circuitry.
Powering the TV reveals serious purity problems. A roughly vertical band in
the middle has the correct colors which then transition in a rainbow pattern
toward the upper left corner (red-green-blue). Similarly, toward the right
edge, the colors transition once or twice. Not promising.
Manual degaussing does absolutely no good.
External degauss was attempted with power on to confirm that it was not a
magnetization problem. Similarly, this does absolutely no good.
The yoke and purity magnet assembly seem to still be solidly locked in place
and adjustment of the purity magnets has no useful effect.
The only conclusion can be that the shadow (slot) mask inside the CRT either
deformed or popped loose due to the impact.
At this point, the TV was pronounced incurable and sent home to be used for
video games hoping the kids wouldn't notice the weird colors or think that
they were abnormal.
Six months later, a large truck hit a substation transformer. This apparently
resulted in a 12,000 V feed falling across a low voltage line sending mucho
excessive voltage to an unsuspecting neighborhood. It made the news in a big
way.
Many many houses were affected with lots of blown stuff. Geez, if I could
have driven around there for the next few weeks on trash days, I could
probably have collected all sorts of late model electronics - much of it
(as you will see) having easily corrected problems (as opposed to this TV).
Tom comes in one morning mentioning the power surge and I volunteer as usual.
Sure, bring the stuff in (not really knowing what to expect).
So, this unlucky TV shows up again. This was a couple weeks after his toasted
Panasonic VCR was brought in only requiring a new fuse.
In the case of the TV, it turned out that a fusable resistor sacrificed its
life to protect the fuse. For testing, a jumper was used in its place but
a proper replacement was ordered for the permanent repair. Sure enough, the
set comes back to life. Unfortunately, the colors are still messed up...
There is one other possibility to at least reduce the severity of the color
problem: refrigerator magnets. If a compensating magnetic field can be
created, the beams might be convinced to hit the proper phosphor dots! The
next day, Tom shows up with a bag of all sorts of small magnets and by the time
I take a look, there are numerous warts pasted around the perimeter (as well
as further back) of the CRT.
"Hey Sam, why do these magnets vibrate when they are brought near this thing
on the tube"? This of course is due to the 60 Hz field current being pumped
through the vertical deflection coils. The effect is very noticeable up to
several inches from the yoke.
"Hey Tom, did I mention that the fat red wire here has about 25,000 V, this
thing with the coils (the yoke) has 1,200 V pulses on it while the set is on,
and that you probably don't want to let the TV fall over again with the
picture tube all exposed like that?"
Well, Tom survived somehow.
However, poor Tom was only using a VCR's blue screen to position the magnets
for a pure blue screen. What he didn't realize (and I forgot to mention) was
that the geometry was being quite thoroughly messed up while the purity was
being repaired. Thus, when we played an actual tape, the colors were now more
or less correct but shapes were distorted. Oh well, you can't have everything.
Tom will have to decide which will keep him (or his kids) happier.
Comments: On a 31 inch CRT, the shadow mask is a thin sheet of metal 2 feet
across. Drop the tube on its face and it is likely to distort or pop free.
This wasn't a slight bump as the bezel was cracked and the mainboard broke
free. It has been suggested that in this case one should then drop the TV on
its back to reverse the effects of the original fall. I don't think Tom is
quite willing to try this....
Patient: Basic Tone/Pulse dialing 'Princess' phone clone made by Conair,
Model SW-204. Don't they make blow dryers and back massagers?
Symptoms: One can just barely make out a dial tone - very faint. Buttons
do seem to produce tones or pulses depending on the switch setting
but these do not make it to the phone line. Side-tone does work.
Testing: Tried different cords (both base and handset), whacked and pounded
it - no change.
Rich brought in both a Fisher VCR that eats tapes (boring - bad idler tire)
and this cheapo phone. Well, it takes all types.... He said he was trying
to determine why the phones in his house went dead and tried in the 'test'
jack, wherever and whatever that might be. Then this phone stopped working
everywhere else.
Indeed, plugging it into a known good phone line results in just barely being
able to make out the dial-tone. Nothing else works.
I figure it is probably not worth spending a lot of time on this thing. After
all, it is not what you would call valuable in any sense of the word. Check
for bad connections, test semiconductors for shorts, clean the keypad. That is
about it. Maybe I will get lucky. It has happened in the past.
First, the base unit. One screw to remove the cover. Not much in here. Just
the bell (a real genuine gong, none of this piezo buzzer stuff). There is
nothing here that is the least bit suspicious - the on-hook switch seems to be
in good condition and does function even though you cannot really hear it too
well.
The handset is more interesting. A bunch of transistors, diodes and
rectifiers, other stuff, and a single chip.
Prodding the circuit board has no effect. Even my Magic Spit(tm) is utterly
worthless - can you believe it?! No change except a couple of pops from the
earpiece.
I check all transistors and diodes for shorts - none.
I remove the keypad rubber pad and check for stuck keys - none.
I was about ready to pronounce it dead-dead when a lone part on the bottom
of the circuit board caught my eye. What is this? A blue diode? What sort
of diode would dress up in *blue*. Checking across it reads about 50 ohms
on my VOM. That doesn't sound right so I unsolder one end. Still 50 ohms.
I can just make out the part number - ITT ZPD120. My ECG Guide shows this
to be a 120 V 1 W zener diode.
For testing, I just leave it out since I don't expect any phone calls that
might produce a ringing voltage (which is the only situation where there
could be anything approaching 120 V outside of a lightning strike).
Sure enough, the phone now works perfectly.
I don't have anything like a 120 V zener in my junk box. However, a neon
lamp would probably work in a pinch to protect against the ringing voltage or
voltage spikes. An NE2 would have a breakdown voltage of about 90 V (and
sustaining voltage of about 60 V) - close enough for now.
Later on I had second thoughts. Was that really a 120 V zener and not a 12 V
zener? Unfortunately, I misplaced the bad diode on the basement floor and by
the time I located it, the printing had been obliterated. Therefore, I did a
little circuit tracing.
The zener was across the output of the bridge rectifier connected to the phone
line. From the types of transistor (HV - 300 Vcbo sort of things) and high
value resistors (100s of K ohms) I conclude that indeed the 120 V is probably
correct. In fact, it would appear that a voltage divider is set up (100K/2K)
such that one of the transistors would just turn on in response to the ringing
voltage.
Just to be sure (well, for peace of mind - other parts would have blown if my
conclusions had been incorrect), I dialed the phone from my computer line a
couple of times to check that it rang correctly and didn't explode.
Comments: There is no way of knowing what caused the zener to go bad. Maybe
ESD (Electro Static Discharge), maybe Rich's electrifying personality. Nah,
forget that option! More likely he somehow managed to plug it into a power
outlet.
The diode actually looks like an afterthought - mounted on the bottom of the
circuit board across two pads with no component marking of any kind. The
designers probably found out the hard way that such protection was needed!
Patient: Nice little older 13 inch color TV. Nothing fancy but probably worth
some time, effort, and (someone elses) money.
Symptoms: Dead-dead-dead.
Testing: Not a cord or outlet problem.
Actually, Dave had been talking about buying a bunch of horizontal output
transistors (as in 3 just in case) from MCM Electronics for several weeks.
Seems this TV blew the HOT and a fusable resistor before I saw it. Finally,
we placed the order....
The day after the parts arrive, he comes in. "Well, the flyback is arcing".
So I asked him if he had tried taping over the location of the arc.
One week later. "Well, I put a bunch of layers of electrical tape over the
cracked area of the flyback and then it started smoking".
OK, bring it in. Without knowing whether anything else has been fried by
the bad flyback, it is probably not worth investing in a flyback. If the
arc got into the signal circuits, anything could be fried.
The following week, the TV shows up. Without powering it up, I open it and
then connect my series light bulb AND Variac gizmo. I really was just doing
this for the entertainment value (not the shows) rather than seriously fixing
the set. The friend of a friend of Dave who was our 'customer' had already
decided at that point not to pursue a repair since there was no assurance that
a new flyback would not be just the first of a long line of replacement parts.
My plan was to confirm that the flyback was beyond redemption and strip the
carcass for parts (like the still surviving replacement HOT).
Now for the action! As the voltage is brought up, the light bulb glows
brightly (from the degauss circuit) and then goes dim as it should but sure
enough, the set starts smoking. Some closer examination shows that the smoke
is not coming from the flyback but from somewhere *under* the mainboard. What
do you know!? There are a couple 1/4 watt resistors tack-soldered in place.
It looks like the location where a fusable resistor would live. What the heck.
I remove the resistors and put in a jumper wire. As long as I use my series
light bulb, it should not be a problem. And this is just for fun anyhow!
Now, no smoke, but there is indeed arcing under the bandage (electrical tape)
Dave has installed on the flyback.
Now I cannot resist the challenge. I would really like to be able to get it
working well enough to confirm that the rest of the set is undamaged. Well,
first step is to rip off the tape and see what is actually there. Now, when
power is reapplied, it appears as though the arcing is from within the area
of the focus and screen pots.
I then unsolder the flyback from the set and pull off the jerry-rigged focus
connection (no solder of course!).
So, I get out my hacksaw and Vise-Grips (OK, no Vise-Grips but a rather large
screwdriver used as a pry-bar. You guessed it. I cut around the periphery
of the focus/screen pot cover and pry it off. Interesting low cost design.
The 'pots' are just screened on resistor material. The knobs just seem to
have a sort of conductive plastic or rubber for wipers. Probably good enough.
After a little cleanup, I give it another try. Now, no arcing but of course,
no picture either as the CRT screen (G2) is not connected. I can solve that!
Some careful positioning of the focus and screen wire ends (hanging off of the
CRT neck board) in contact with the resistor material on what is left of the
flyback! This should work. However, adjustments will be really tricky :-).
Sure enough, it is now possible to power up the TV and even tune in our
favorite channel 57 broadcasting the afternoon cartoons.
I must have guessed pretty close - the focus isn't too bad and the brightness
is about right.
Hey Dave, the TV is working... "Uh, yeh, sure." Really. Adjustments are
touchy.... BTW, what was the value of that fusable resistor. "Probably
2 or 3 ohms." Why were there a couple of 22 ohm resistors installed?
No wonder they smoked! "I noticed that.....a little too high, huh?"
Of course, Ed came over to watch all of the clowns at play (but from his usual
safe distance, being a 5 volt TTL type).
As it turns out, the 'customer' reconsidered and has now decided that it would
be worth springing for a new flyback now that we knew nothing else is broken.
Comments: No, I would not recommend this as a permanent repair. Even taping
or insulating a flyback is somewhat risky if a new breakdown path develops.
Had I known exactly where the arc was located, I probably could have done
less of a hatchet job on the flyback retaining use of the focus and screen
pots - knobs and all. The actual path of the arc seemed to be from a corner
of the cover via the surface of the rest of the flyback to the core.
It was important to confirm that nothing else had been blown by the arcing
(the original HOT and missing fusable resistor had been the first casualties).
Having done this, one can confidently order parts knowing what to expect.
Patient: Ravi's Sony color TV. Symptoms: No reception. A/V input works fine. Testing: Tried with antenna and output from VCR. There apparently was a history to this set.... Ravi came in and said "I have a Sony TV that doesn't work." Right, so what else is new? This isn't the same one Dave and I repaired before, is it? "No, a friend had borrowed the TV and it broke about six months ago. He attempted to repair it by replacing the IF box, whatever that is." How did he know to do that? Did he see it on the Net? "Uh, perhaps. Now it doesn't work at all." OK, bring it to the lab. Sure enough, there is nothing when using the RF input. Connecting it to a VCR's RCA jacks results in a perfect picture. So, maybe the problem wasn't the IF box or maybe the repair wasn't done correctly. I, of course, expect that the original cause were the infamous bad solder connections Sonys are known for. It didn't take long to determine the problem - nearly every pad to the IF box had been ripped off of the circuit board! Hey Ravi, what did he use to repair this thing? A blow torch? I have never seen a worse soldering job. It took a minute or two, but Ravi finally realized what I was talking about. Hey Dave, you have to see this.... Hopefully, the new IF box wasn't actually damaged by the attempt or by being only partially connected (mostly not connected). It required about 20 minutes to install jumper wires to the nearest component pins on each trace. Even where there was some sort of pad remaining, I added jumpers to be sure there would be no problems in the future. I just hope this IF box doesn't develop internal bad solder connections. Whoever ends up working on it will be in for a real treat. Guess who that might be? :-). And, presto! Reception is now perfect - at least to the extent that can be tested in our lab which is to say, at least there is now snow and the VCR's channel 3 and 4 output works fine. Ravi later told me everything else was fine as well. Comments: One assumes that those reading the repair guides have some minimal desoldering equipment. However, this episode prompted me to add a section on the need for proper tools to both the GE/RCA and Sony FAQs. In this case, I am sure Ravi's friend didn't get his advice from the Net as there was a problem report form from "Tom and Joe's TV Repair" or some such place (hope not anyone reading this). So, the set was probably taken in for an estimate but Ravi's friend figured he could save some money..... Too bad I don't charge for this stuff - I could have cleaned up!
Patient: Andy's friend's cast-off Emerson Model 1400EMA RGB as in CGA monitor.
Symptoms: Dead as a brick.
Testing: Tried with and without a source of video. Of course, coming up with
a CGA output was a bit of a challenge but a classic Toshiba laptop
came to the rescue.
So I found this thing sitting in the middle of my office floor one morning.
As usual, I just stepped over it and went on with my business until the donor
came around to say something.
So Andy, what is this thing??? "A friend of mine was cleaning up and was going
to throw it out. Of course, I said I knew someone who would probably take it.
He thought it could be fixed." How? "I don't know and I haven't tried it."
Of course, Andy is a MAC person so the last comment, at least, is not at all
surprising :-).
The monitor is indeed very dead. Getting at the mainboard proves to be quite
a challenge. There are many shields. Even just being able to probe the bottom
of the mainboard requires 20 minutes to remove countless screws.
No fuses blown.
B+ appears to be present on the output of something that looks like a linear
regulator. It is also present at the collector of the horizontal output
transistor (HOT). Nothing appears overheated. Since there is absolutely no
response of any kind - no initial static or even a tweet - it is not likely
in HV shutdown.
Checking at the base of the HOT, there is no drive of any kind at any time.
To determine how base drive is obtained requires tracing back from the base
of the HOT. This goes to the usual driver transformer. The primary side of
the driver transformer appears to go to a multilegged chip - likely a combined
H/V deflection processor.
There is nothing on the pin of this chip that goes to the driver transformer
but there is apparently power to the chip. The output appears stuck low.
Wait... There is just a hint of a pulse at the horizontal frequency (around
15.7 KHz) free running. What about on the supply side of the driver
transformer. Is the chip shorted? I unsolder the pin. Still almost
nothing on either the isolated pin or the transformer. That cannot be
right - that transformer pin should now be pulled high. Double check - yes,
there is power to the other side of the winding.
We have an open transformer primary!
Unsoldering the transformer (4 pins) and testing out of circuit confirms
an open winding. So, how could that happen? It really could only be a
case of a manufacturing defect. There are no signs of overheating or other
stress.
Now for the repair.....
This is a simple straight core ferrite design so it is easy to get at the
windings, at least.
The primary of the transformer is wound *under* the secondary. First, I
peel off the mylar insulating tape and unsolder the start of the secondary.
After noting the direction of the winding, I remove the wire saving it as I
will just replace it when the primary is fixed. It is only about 20 turns.
I then peel off the mylar insulating tape between the two windings.
Fortunately, the break is just about at the terminal so that it is not
necessary to unwind the entire primary (which has many more turns - the
ration is probably at least 4:1, perhaps much higher.)
I inspect the other connection as well but it seems fine.
After unwinding one turn (no one will ever know) I carefully resolder the
wire making sure not to nick or damage it.
Then, I replace the mylar tape, secondary winding, and outer insulation.
That should be as good as new.
Sure enough, the monitor comes back to life.
Just when I thought I was done and about to button up the case, I try is once
more. What is this? No vertical sweep. OK, must be a service switch inside
somewhere that got knocked to the service position. Yep, after a couple of
minutes of panic, I find it toward the front of the mainboard. I probably
bumped it when replacing the transformer from the top of the board. No, I
really did not want to disassemble this thing again.
Comments: The only possible explanation for this sort of failure is that
the wire was nicked during assembly and eventually, the small but persistent
vibration of the transformer at the horizontal frequency and/or thermal cycling
finally caused it to break free. Considering the wire size (probably #34 at
least), such damage could happen quite easily. At first I was concerned about
some cause like a shorted chip resulting in overcurrent but I do not believe
this was such a problem.
Although this monitor is much much simpler than modern SVGA monitors, this
sort of problem could occur with either.
Patient: My cousin's VGA monitor. OK, actually my cousin's 3 kids' VGA
monitor. OK, actually my Tatung CM1495 on loan to my cousin's
3 kids.
Symptoms: Dead-dead. Not much info. "The computer doesn't work" is about
as much as was forthcoming.
Testing: No response of any kind when plugged into computer or wall outlet.
I had put this state-of-the-art 386-20MHz system together from spare parts for
my cousin Kathy's three kids (ages 8, 8, 9) to use. Kathy knows next to
nothing about computers though she does somehow manage to use a Mac Plus for
letters and rumor has it that she has even logged onto AOL at least once from
the PC.
I originally found the Tatung monitor along the curb with some cast off Atari
PCs and an EGA monitor. I grabbed the monitors, too bad about the computers!
Both monitors worked fine - no Repair Briefs material at that time.
About 3 months after giving them the system, I get a frantic call from the
oldest: "The computer doesn't work. The screen is dark but it makes some
noises like it wants to work." Did you see what happened? "No, it just
wouldn't come on." OK, I will be over. Just in case (actually a sure thing),
I take a spare NEC Multisync II to be used in place of the likely dead monitor.
Sure enough, not even the power light is lit. Swapping the monitors makes
the kids happy and Kathy as well since she doesn't need to keep hearing their
whining.... I have offered repeatedly to give them another pair of computers
but there really is no room......
Getting inside this monitor isn't too bad and the power supply is a separate
unit which is easily removed - 4 screws or so and 4 or 5 connectors. It will
also be possible to test the supply outside of the chassis which will be
convenient.
One problem appears immediately - a blown fuse on the AC input to the power
supply. Testing shows a dead short. The posistor is on a separate fuse
and it is fine - darn, that at least would have been easy :-(.
It takes a few minutes of unsoldering various components to discover the
problem - the main switching transistor of an STK7406 switching regulator
is shorted. Unsoldering that single pin restores resistance to normal.
There are actually two virtually identical sections to the power supply board.
Both use similar STK parts and have their own apparently independent feedback
for regulation. However, only one of these is controllable via a signal from
the mainboard - this is likely the deflection B+ which must vary based on scan
rate. I considered swapping the STK7406 for its mate but decided that not only
was that risky but I really would be sure of what I was seeing with only one
side active.
So, I at least need to order a new STK7406 but what caused it to go bad?
This is where not knowing the actual events that led to failure will always
leave a lingering doubt. There are no other obvious problems. Was it left
on overnight? Perhaps, without video input? Would this matter? Did it
overheat? Were the kids screwing around with Windows or switching screen
resolutions? (The last one I rather doubt as they have strict instructions
which appear every time the PC boots not to mess with it!)
I test nearly every electrolytic capacitor on the power supply board with
an ESR meter - all appear fine. No bad solder connections are in evidence.
There are no other shorted semiconductors.
OK, I will order the parts and just be very careful in the initial testing.
Three weeks later......
The new STK7406 and fuse are installed. I am going to start with a Variac
and 100 W light bulb to be safe. The power supply board is propped up next
to the monitor with suitable insulators to prevent shorts.
Power! The light bulb comes on brightly and then dims out as it should.
There are some erratic flashes on the screen at around 100 V - jumpy vertical
as the scan seems to start up and die. Well, that could just be too small a
light bulb.
Next, I try a 150 W PAR floodlight. Not much different. At low line voltage,
bulb glows quite brightly as though there is a bad region of input voltage
where excessive stress may exist. It doesn't seem to recover at normal line
voltage.
I will give it one more chance before declaring there to be other problems.
Now, with a 300 W light bulb, there should be no excuses! Hear that?
Sure enough, bringing up the voltage slowing results in the bulb glowing
brightly at lower voltage but it recovers when full line voltage is applied
and a normal raster appears. Watching the glow of the bulb is instructive
as my test PC boots - it increases in brightness just a bit. Then, when
switching into Windows at 800x600, the brightness increases another notch.
Still just barely glowing.
I put a clip-on AC ammeter in the line circuit but this really doesn't show
anything more than the light bulb. Current is well within specifications.
Only a comparison with an identical monitor could truly show up any difference.
Removing the Variac and using the power switch results in no sudden increase
in bulb brightness except due to the degauss coil (the first time after having
been off for a while). Thus, it really doesn't like low line voltage. Could
a brownout have been the original cause of the failure? How about flipping
the power switch quickly off and on again?
Comments: This is a case where the failure is obvious but the cause may never
be known. I am tempted to leave the light bulb in place all the time when it
is used. Whether that would save the STK7406 from subsequent failure should
the same conditions arise again I do not know. In the meantime, the kids are
using the NEC which interestingly appears to use a very similar design as I
found out when swapping a mainboard on one of these for my friend Bill. The
external controls and internal construction are too alike to just be a
coincidence. The Tatung is probably a reduced cost knockoff of the NEC MSII.
Patient: Arnold's portable CD player (I don't recall the exact model but it
is one of those that still would set off a metal detector). The
optional battery pack is missing.
Symptoms: No response of any kind - not even the display - using the AC
adapter.
Testing: The AC adapter tests fine. I even tried another one without any
better success.
The first real test is to determine if the CD player will work from the battery
input rather than the AC adapter. The normal battery pack is a 4 V rectangular
lead-acid type.
Four volts is close enough to 5 V minus a diode drop so I connect the battery
contacts to a logic supply in series with a handy diode I found sunning itself
on the floor...
Sure enough, all functions are normal. Hey Arnold, it works fine with this
nice 100 A power supply but it may be just a bit too heavy to carry around :-).
So, the problem is in the power circuitry. I suspect that the AC adapter
output is supposed to be converted to 4 V from which all the CD players
circuits operate.
The bottom comes off after removing just a few micro-screws but this is one
of the all metal super thin Sony Discmen - and impossible to access for repair.
The power input/regulator/converter is all crammed on a 1/2" x 3/4" circuit
board with lots of very very tiny surface mount parts. Only the solder side
is visible. Thus, the surface mount parts cannot be probed without additional
disassembly. There are no obvious cold solder joints or other evidence of bad
connections. The DC jack seems in good condition and checking voltages with a
multimeter reveals that power is making its way onto the board. Something is
faulty in the power conversion circuitry.
So, I have two choices:
1. Spend a semi-infinite amount of time riping the player apart and attempting
to check microscopic surface mount components, reverse engineer the circuit,
or attempt to locate a service manual or :-),
2. Hey Arnold: do you care if the battery pack can be used at all - ever???
"No, It died after a few weeks and I never bothered with it again. Why?"
I can rig a regulator to fake out the thing into thinking it is running
off battery power when it is actually running from the wall adapter. "Sure
that would be fine. Whatever works."
So, I added the following regulator circuit and mounted it inside the battery
compartment.
D1 I +-------+ O D2
Vin (+) o----|>|---+-------| 7805 |-------+----|>|----o Vout (+)
1N4007 | +-------+ | 1N4007
| | C |
Sony 9 V +_|_ C1 | _|_ C2 Battery
adapter --- 10 | --- 1 uF terminals
- | uF | |
| | |
Vin(-) o-----------+-----------+-----------+-----------o Vout (-)
An aluminum plate (a piece of roof flashing) was attached to the regulator to
serve as a heat sink. D1 provides reverse polarity protection. D2 reduces the
output of the 7805 by one diode drop. This produces about 4.3 V under load
which is close enough to the output of a fully charged batter pack.
Wires were soldered directly to the DC jack and routed into the battery
compartment. The original connections to the internal voltage converter
were cut.
OK Arnold, it seems to work fine. "What do I owe you?" Only charge is that
if you ever decide to dump it, send it to me. It is a nice example of Sony
equipment when the Sony name meant something. "Sure."
Comments: I could have made this into a major time consuming repair or opted
for a short cut. Since the battery was never used, why spend a lot of effort
when a simple regulator could replace the functionality of the defective
circuitry? In principle, the original wiring could be restored - and then
the original problem could be remedied. For now, it works fine as far as I
am aware.
Somehow I don't think I will ever see this player again as Arnold has since
moved to Japan :-).
Patient: Surplus Lambda LUS-8-12 12 V 1 A switcher - a cute little supply that
I was using to test an automotive radio/cassette when something went
horribly wrong :-(. (There was nothing wrong with this unit that
a new belt didn't cure - boring.)
Symptoms: Fuse blew. Of course, what did I do? Put in a larger fuse! Then
other stuff blew.... I just figured it blew due to the overload,
not to its parts shorting out. Right :-(.
Testing: It doesn't take a series light bulb, Variac, or rocket science to
now see that this is a very blown power supply with a dead short on
the power line!
OK, so this was from my younger days and was the first switchmode power supply
I had ever attempted to repair other than bad connections and the like. It was
also the first (and hopefully, last) such supply that I blew so completely and
in such a spectacular manner! (I won't mention that other one that billowed
smoke from its main filter/doubler caps after blowing the switchmode power
transistor due the slip of a scope probe....).
The original problem was a result of miswiring of the power inputs to the
radio/cassette: a clip lead was shorting +12 to ground. No problem - these
supplies are short circuit protected, right? WRONG! Or, at least, not
entirely. I may have been operating it on a Variac and increasing the voltage
gradually without realizing that there was no voltage across the power inputs
to the radio/cassette. At some point - POW, the fuse blew with a bright flash.
It was all down hill from there!
This is a really cute little cased switcher which is the only reason I spent
several hours and $10 or so for parts to repair it. It was more a challenge
than anything else. I may never use this supply for anything again before the
Sun becomes a red giant :-).
Known casualties in the parts department:
* Line fuse.
* Switchmode power transistor.
* Fusable resistor.
* HA17339 (LM339) quad comparator.
* 5.6 V zener diode.
* 1N4148 signal diode.
* 2SD467 transistor.
There may have been others - I lost count.
I suspect that only a couple of parts blew at first but the rest of the
components let out their smoke when powering up for the second time. There
is no way to know for sure at this point and I am not about to recreate the
disaster!
All the fried discrete semiconductors were dead shorts so it just was a matter
of checking each with a multimeter. The HA17339 wasn't totally shorted but had
a low resistance on its power-ground pins so I guessed that it, too, was bad.
I traced out the entire circuit which wasn't fun or quick. It took a while
to determine that the HA17339 was just a common LM339 clone - Internet search
engines didn't exist back then.
I replaced the HA17339 with an LM339 in a socket. Removing the original part
wasn't easy even with my SoldaPullit - I damaged a couple of pads which needed
to be jumpered.
For testing, I substituted a BU406 which I had handy from a long stripped
video display terminal.
Using a Variac (no light bulb at that time), the supply came up properly and
ran without smoking. I then installed the proper transistor. Next power-up
it worked fine but when I removed the Variac and plugged it directly into
the wall, it was lifeless again. I feared the worst but as it turned out,
only an underrated resistor had opened (the fusable resistor I had temporarily
replaced with a 1/4 watt film type). After installed the proper resistor, it
seemed happy as before. I just will now know not to attempt to drive short
circuits!
Comments: This was my first exposure to SMPS repair and also to 2S part
numbers. When I first ordered the 2SD467 it was a guess - that the part in
the MCM catalog was really what I wanted. Well, everything has to have a first
time. If this weren't such a cute little supply - and I wasn't to angry at
myself for blowing it up, the time and effort would not have been justified.
In fact, it was relegated to the scrap bin until I saw the light with respect
to parts identification.
Many SMPS designs are robust only near full line voltage. At lower voltage,
the pulse width modulator struggles to achieve regulation and succeeds or
kills everything unless the designers specifically took these conditions into
account and provided appropriate current limiting. Just claiming that a
supply is short circuit or overload protected is not enough. This is one
reason brown-outs can be hard on equipment.
Patient: Derek's friend's Fisher component 80 W/channel power amp. It looked
kind of nice with those LED power meters. No comments about Fisher
quality, please :-).
Symptoms: The right channel is stone dead. No hum, no noise - not a thing.
Testing: Double checked inputs and tried several, whacked it (to wake up any
speaker relay) or bad connections, switched input selector back and
forth. No effect of any kind.
With absolutely no hum or noise on the dead channel, the problem is probably
not in the input circuitry or preamps. There would be some evidence of life
in those cases - some low level hiss at least.
The nice thing about stereo equipment is that there are two of nearly
everything. Thus, comparing signals, voltages, and resistances can be
effective even without a schematic.
Unfortunately, these assumptions led me astray......
Big audio amps are often designed along the lines of power operational
amplifiers. They have several voltage gain stages and a final power/current
gain output driver - with negative feedback around the entire thing. One
implication of such a design is that intermediate signals may look strange
even when the unit is operating properly.
In this case, the intermediate signals on the good channel looked like
hum and noise while the similar signals on the dead channel appeared to be
normal audio.
Among the problems found were bad solder connections on the pre-driver
transistors due to their running hot (no heat sinks) but repairing these
had no effect on the original problem.
Fuses were all good. Checks of major power rails showed them to be healthy.
Finally, I was convinced that the most expedient approach would be to swap
the finals - STK0080 hybrids. There is always a slight chance that swapping
power components can blow the previously good part but since everything else
appeared normal, what the heck. They are only 10 pin devices and less than
$20 in any case.
Five minutes later, sure enough, the right channel seemed to be working
normally and the left channel - was, well, missing a hybrid!
However, now there appeared to be some noise and erratic behavior when powering
up. One or both channels would not come on as soon as the speaker protection
relay clicked but at some random time later. The input signal source selectors
did not appear to have any effect on the noise but tapping the relay revealed
that it was at fault. Fortunately, the top poped off easily and some contact
cleaner and burnishing restored it to health at least temporarily.
Comments: As noted, most stereo equipment has a distinct advantage when
troubleshooting in having duplicates of nearly everything but the power
supplies. Therefore, it is pretty easy to locate obvious shorted discrete
parts by simple comparison measurements. Even testing of the hybrid modules
may have been possible in this way though I did not think to try. However,
with a design depending on a closed high gain feedback loop around nearly the
entire system, actually looking at signals can be misleading unless you are
aware of the possible implications of what is being observed. In this case,
the power supply rails had substantial hum (normal) and signal points did not
have anything resembling a signal on the working channel - rather it was just
low level noise and hum. Buried in there somewhere was the actual signal. The
similar test points on the bad channel had a reasonable audio waveform because
without proper feedback they were operating at a much higher level than was
normal. It was probably highly distorted but was more recognizable as an
audio signal.
Patient: Garage sale Mac Plus with original Apple HD20 (non-SCSI). It was
marked $50 but before I could make an offer, the price somehow
dropped to $25. OK. No keyboard or mouse, however. Well, can't
have everything. At least with a Mac, basic functionality can be
determined without any user input devices!
Symptoms: Powering without keyboard or mouse seems to result in attempt to boot
from diskette which is stuck in the internal drive. However, without
a keyboard or mouse, it is not possible to attempt to eject from
software. The paper clip in the eject hole doesn't work either - it
is jammed. The hard drive does not spin up.
Testing: I borrow a mouse and determine that the computer itself seems ok
except for the uncooperative diskette drive. The hard drive remains
dead.
This poor little Mac was sitting by itself along the side of the driveway
at this garage sale that also had other interesting items but the Mac seemed
almost useful. I probably would have gone higher than $25 but wasn't about
to argue. (Recently I picked up a Mac Classic at a flea market for $5 but
that is another Repair Brief!) They did say the hard drive probably didn't
work either....
In fact, it came in handy when my cousin was given a Mac Plus system with a
dead flyback. A flyback transplant confirmed that nothing else was broken and
got her up and running nearly instantly :-).
First step is to see what is wrong with the diskette drive.
Of course, opening any of the single piece Macs is always a joy unless you
have the proper foot-long Torx (or hex) wrench to get at the two screws under
the handle. I did not and so was forced to improvise. I ground down the
end of a triangular metal file to a shape approximately a Torx. As they say
"If it works, use it." And, no, you don't need an Apple approved 'case
splitting tool' to get the back off. With care, a pair of wide flat blade
screwdrivers work fine without damaging anything.
With the cover removed, it is a simple matter to remove the diskette drive.
A close examination reveals that a little bracket that is part of the eject
mechanism was somehow bent out of shape - short work for my needlenose.
The grease - a common problem with older Mac diskette drives - appears to be
in as good as new.
It doesn't appear to be defective or weakened in any way. My guess is that
someone attempted to pull a diskette out of the slot by using a pair of pliers
rather than the paper clip or software. I cycle it a few times using the
slightly abused but still serviceable diskette and a paper clip - as good as
new!
After replacing the diskette drive I confirm that it will boot with a known
good system disk. It even ejects on command. What more can you ask?
Next, the hard drive. There seems to be no sign of life from this. The power
switch does nothing as far as I can tell. Possibly the ready light flashes on
momentarily, but that is about it. No spinup sounds at all.
After removing the cover it is necessary to blow out several years worth of
accumulated dust before checking anything electronic. The controller and
drive itself do appear to be receiving power - the chips get warm at least :-).
Could it be sticktion? This system probably has not been used in a quite some
time. Fortunately, this is one of those drives ('some user serviceable parts")
where the spindle motor flywheel/magnet is readily accessible once the drive is
removed from its mounting.
The flywheel is quite stiff. Actually, this is probably not actual sticktion
where the heads glue themselves to the platters but just dried hardened grease.
I spin the flywheel back and forth a few times by hand. This seems to loosen
it up considerably. I then reinstall the drive on its mounts.
Sure enough, it now spins right up and the Mac even boots - slowly but what
do you expect? - revealing some useful word processing programs, tons of games,
and a database of horticultural information. Not bad for an hour's work.
I can hear you all saying: "Joy, a computer with the power of my digital watch
and the features of a boat anchor!".
Sure, but (1) this was a few years ago and (2) this became the one of the Macs
for my cousin's three kids (who now have a state-of-the-art 386-20 as well)!
I still have the hard drive but really don't know the location of the system
unit as we have played musical chairs with the Macs several times at this point
(but I would have heard if it stopped working).
Comments: Whether you like Macs or not, one cannot deny that the Mac Plus
was a cute little computer that could do useful work and had a user interface
that surpassed anything from Bill Gates at the time - perhaps even today.
Most of the time, problems with these computers were easily diagnosed and
easily repaired - burnt Molex connector pins and blown flybacks are two
popular failures.
I have not had a reoccurrence of either the diskette or hard drive problem
despite it not being used all that much but I guess 'not all that much' beats
'not at all' by a wide margin.
Patient: Color version of my HP DeskJet Professional. This is an older but
nice little printer. The separate black and color cartridges make
it much more convenient than some of the newer models.
Symptoms: Goes through all the motions but nothing appears on the paper.
Cleaning/priming doesn't help. Both cartridges are not likely
empty!
Testing: I use the built-in test initiated by holding down the FONT button
while powering up. Nothing on the paper.
So Joe comes by and mentions that this printer has been acting up for several
months now. This is on his son's computer so I imagine Joe is not that eager
to have it working properly - considering the high cost of color cartridges.
He did tell me that he called some place and got a flat quote of $140 to
repair it excluding any needed cartridges.
"OK, drop it off but I don't know when I will get to it."
The printer shows up a few days later - no scrap paper to test it with. At
least he provided the power pack!
Fresh from my experience with my DeskJet, it takes about 30 milliseconds to
locate one problem: the rubber boot that seals the priming station is nowhere
to be found. I am not at all surprised that printing is erratic without this.
Priming would be useless without the suction seal. Now, I wonder where it
went.....
As with the original DeskJet, the top pops off after releasing 4 catches near
the corners. Ah ha! There it is... Shading itself on the mainboard! Did
some overzealous manual cleaning cause it to pop free? Nah, I doubt anyone
in that house would realize that such things need cleaning. It may have just
popped off from one of the automatic cleaning cycles - it does seem a bit
lose. However, in all my subsequent activities, it stays put just fine.
I clean the contacts on the cartridges and the ribbon cable as well.
Since this hasn't worked for a few months, I clean the cartridges by hand
just to be sure - some warm water, patted dry with lint free paper towels.
A little blowing in the vent hole on top of the black cartridge, no vent
holes visible on the color cartridge.
Ready?
Sure enough, the basic test patterns come right up. Only a couple of nozzles
not firing. More cleaning, blowing, priming. OK, now B/W seems to be fine
but there is only a hint of color - maybe yellow.
Even when executing the ALIGN command, almost no color (magenta) shows up.
I try more washing of the color cartridge. Now, there is a hint of the
magenta alignment marks but they are spread from here to kingdom-come.
In other words, a vertical line prints as a scatter diagram of dots!
I try some cleaning cycles. These prime the cartridge and then should print
stripes of each color. At first, only yellow shows up. After 4 or 5 attempts,
there is some evidence of cyan (blue) and magenta (red) appearing at the
start of each line but they then tend to die away.
OK, that could just be an empty cartridge. Yellow seems to be fine and there
really isn't anything different about it. (This is not surprising as yellow
is the least used color.)
So I call Joe: "You know how they were going to charge you $140 to service the
printer (hint hint - never works with him! Maybe if I didn't make it sound so
easy....), well it appears to just be that a rubber part popped loose and the
color cartridge appears dead." You mean it prints in black? "Yep, as good as
new."
Two hours later, a new color cartridge shows up.
And, sure enough, this works just fine out of the box.
I recommend that he obtain replacements for the two rubber boots, at least,
as they do seem to be on their way out.
Comments: Once again, a simple cause with dramatic consequences. I am still
amazed that ink-jet technology works at all given the simplicity of these
printers. Of course, that simplicity masks some very sophisticated and
precision technology that goes into the cartridge/nozzle construction and
controlling firmware.
Nothing serious ever seems to go wrong with these older DeskJets. Now all
I need to do is locate a parts source for the plastic guide bar that broke
in the DJ855C at the office!
Patient: Garage sale Craftsman Eager 1 - 4 HP, 22 inch, self propelled,
electric start. $10. Seems to have had some hard use but not bad,
I was thinking.
Symptoms: Although there is plenty of gas in the fuel tank, it will not start
and shows no signs of even wanting to start.
Testing: Doesn't start at all - no way, no how. There is no sign of even a
weak attempt at turning over.
The seller mentioned that all it needed was to have the carburetor cleaned.
This made sense though I wouldn't have trusted him as far as I could throw
him :-).
After yanking on the starter cord for a while, I removed the spark plug - dry
as a bone. Squirting a couple teaspoons of gas in through the spark plug
hole, replace the plug, then crank - and it will sputter to life and run for a
couple of seconds. So, the carburetor is a good bet.
Removing the float carburetor on a Craftsman engine is a simple matter - two
screws, the fuel hose, and the throttle/governor linkage.
Before taking it inside, I unscrew the bolt at the bottom and drain the
remaining gas in the float bowl.
On the bench (actually, on top of a wad of newspapers on top of the washing
machine), I disassemble the carb. What a mess! This must be 10 seasons of
neglect - icky disgusting green caked on muck. After removing the rubber
gasket (as it may be damaged by harsh solvents), liberal use of carburetor
cleaner and a roll of paper towels are able to remove most of the build-up. I
recall that there are critical holes - orifices - in that bolt. Using some
wooden tooth picks, I am able to clear it down to the shiny brass.
There, that should do it.
Reassembly is straightforward - or should be. Now, which hole did that
throttle governor wire go into???? Well, that one looks correct but I
wonder...
Tighten the screws, replace the fuel hose. Ready?
Wow, that started right up. Might it be running a little fast?
A quick check of the oil. Maybe a bit low. Let me test the mower a bit.
Then I can change the oil and check out the air filter, spark plug, etc.
So I mow the back yard. This is a self propelled but the drive engage-lever
doesn't seem to stay in the selected slot. Well, something else to take care
of later.
Is it running fast??? Even on slow speed, it sounds more like my other mowers
on fast. Try fast. That isn't right - sounds more like a 2-stroke. Back to
slow.
Bang! Ding! Clunk! Arggggg! Ever get that sinking feeling? Well, I got it.
Something bad happened. Or, as Luke Skywalker liked to say in Star Wars "I
have a really bad feeling about this." Or was it C3PO. Well, no matter.
After checking for something under the deck (I didn't really expect to find
anything), I had to come to the conclusion that there was a major mechanical
failure inside :-(.
That is when I took a closer look at the oil level. Dang - it was much worse
than I thought at first. In fact, one could say that all that was left really
was the sludge! Geez, what a disaster.
Without even moving it from the spot on the lawn where it died, I start to do
the disassembly (or perhaps, autopsy would a more accurate term).
Off comes the cylinder head. Rotating the blade or flywheel no results in no
movement of the piston and it is not possible turn it through a complete
revolution. I never did find one of the head bolts amongst the weeds....
Unbolting the engine and taking it into the basement:
Off comes the blade. As expected, the blade shear key is broken. Later I
discovered that the flywheel shear key was also - sheared.
To remove the front wheel drive pulley requires some careful filing of the
burrs on the keyway slot. This will also prevent damage the main bearings
when the crankcase (sump) cover is removed. At least I as thinking rationally
enough to take these precautions.
Unbolting the cover and a little tapping - darn, the gasket tore. Oh well, it
probably would need to be replaced anyhow.
The extent of the damage is now clearly evident (amid the icky disgusting
dripping black sludge). Maybe more newspapers would be nice.
The connecting rod is partially broken off from the cap. One bolt is still
hanging loosely but the other bolt is laying in the bottom of the crankcase.
There is a chunk of the rod laying there as well and a big ding has been taken
out of the soft aluminum of the crankcase when the broken end impacted (but
that is just cosmetic damage).
Arggggg. This won't be cheap.
Was it lack of oil? Was it the apparent overspeeding? Was it both?
Close examination of the broken parts reveals:
* There was some overheating - there are brown deposits on the rod and cap
in the vicinity of the crank bearing.
* The bolt that was sitting by itself had backed out on its own and just
dropped off - the threads and hole where it mounted were undamaged. It
was not ripped out.
* Without the second bolt, the constant pounding of crankshaft/piston actually
distorted the cap until it mashed into the crankcase and broke into three
pieces.
* There is scoring on the crank pin but it does not appear to have been caused
by the current problems but would seem to be a result of normal wear and
neglect. There is no evidence of overheating (blue tinge, etc.).
So, I tried Sears for a replacement. Unfortunately (or perhaps, fortunately,
as we will see), I didn't have the 27 digit engine number so I couldn't get
the time of day let alone an estimate of what a new rod and cap would cost.
Probably about $25 was the closest they would come. Though, I wasn't exactly
sure they even knew what a connecting rod was! Fan belts was more their
speed.
Well, I am not going to spend $25 on a $10 lawn mower.
I also tried a lawn mower repair shop. They were even worse (still would't
give me the time of day) but quoted $45 anyhow.
That evening I emailed to a fellow Netter among other things that I was
thinking of writing a lawn mower FAQ (not actually describing the disaster
of the afternoon).
"Speak of the devil.... I swapped cranks in an 'Eager-1' (Craftsman) mower
this weekend for a friend. His wife ran into the curb and, well, the shaft
was about 9 degrees bent :-/ He picked up another mower that was 'stripped'
but the crank was there... So, it was pretty simple actually. Started the
first pull too :-)"
Hey, where did you get the replacement? What did you do with the rest of
the carcass? You wouldn't happen to have a connecting rod by any chance???
"I'll check. Does it have a slant gate or a perpendicular to rod? I ask
because there are two types, the straight and the slant type."
Thanks. It is the slanty type. Here are the dimensions....
"I got the piston+rod for you. The part # is 25-0-25 (rod), 30-0-?4 (cap),
5-0-66 (piston, includes shaft but not clips, sue me :-) It's the slant-gate
one. Let me know if the numbers jive..."
Perfect. The '?' in the 30-0-?4 even matches. The piston isn't the same but
my piston is fine (send the piston along anyhow, though, you never know when
these things will come in handy.
Well, that wasn't so bad. The price was right - just shipping.
The parts that arrived were in perfect condition!
Hey, how many hours on that mower?
"Only about 25 hours before the incident with the curb :-(."
I did some rough measurements of clearances using the new rod+cap on my beat
up old crankshaft. The crank pin clearance is marginal (large) but no way I
am going to hunt down a new crankshaft. It will have to do.
In order to assure that no metal chips are hanging around anywhere, I
completely strip my engine, wipe down the inside of the crankcase, and blow
out the oil passages, etc.
Everything goes smoothly until I realize that I didn't make note of which way
the piston goes (there is a right way and a wrong way) but the Chilten small
engine manual comes to the rescue. (Pistons are not symmetric - the piston
pin is slightly off-center to account for the direction of crankshaft rotation
and the direction and center of force. If it is installed the wrong way, at
the very least, there will be excessive piston slap; at worst, parts will just
bind outright.
Tightening the rod bolts - the most critical as far as proper torque is
concerned - is always fun. It is just possible to get my torque wrench with a
3/8" to 1/4" adapter and a 1/4" socket in place into the confines of the
crankcase with the crankshaft in one particular position and alternately
tighten the bolts to specification.
All the other parts go together easily. I obtained new gaskets (actually
bought them at that lawn mower repair shop) but reuse the old head gasket
until I am sure everything inside is fine. Later, I will install a new one.
Timing is obvious since there are marks on both the crankshaft and cam gear.
To assure that the oil pump passages are clear and the pump is working before
total reassembly, I add a little oil to the sump, remove the test port screw,
and frantically crank the engine with no spark plug installed. After what
seemed like the 199th or 200th pull, I realized the darn thing had electric
start. Plug it in, push the button, oil starts trickling out of the hole
almost immediately. I then replace the screw and then tighten the sump bolts.
I also repair the front wheel drive shifter while I am at it - mostly some
careful rebending of the sheet metal but getting to it meant drilling out some
rivets and that was not fun.
I replace the carburetor now actually comparing it to one of my other mowers
(what a concept) to determine which hole the governor linkage goes in. What
do you know, it was the OTHER hole :-(.
Total damage: Broken rod and cap, broken blade shear key, broken flywheel
shear key, gasket set, bruised ego.
Finally, I top off the oil.
Now for the test: Not too shabby. It started on the second pull. It seems to
run fine (now at a more normal speed) without any excessive vibration or
noise. Only time will tell.
Comments: Whether it is a lawn mower or an automobile engine, oil is its life
blood. That is why even if you don't do any sort of maintenance on your car,
at least you should check the oil periodically and head the oil idiot light
should it even come on - pull over, stop the engine, get a tow. Running
without oil for a couple of minutes can ruin the engine. Ever seen that TV ad
where they drain the oil and then run a bunch of engines till they seize?
And, no, synthetic super-duper oil won't save you.
This whole episode was likely preventable if I had actually taken the time to
check the oil and had realized that the overspeeding was a danger sign.
How long will it run? I don't know. If it weren't for that clearance on the
crank pin, it would be as good as new.
In the end, there was some benefit to this whole affair. It was the stimulus
to write "Notes on the Troubleshooting and Repair of Small Gasoline Engines and Rotary Lawn Mowers".
Patient: Our first stereo receiver - Harman Kardon model 520.
Symptoms: Burnt out Stereo Beacon(tm), noisy left channel, weak right channel,
tuner stuck on one station, intermittent stereo reception, erratic
controls (not all at the same time).
Testing: Where applicable, checked multiple audio sources or stations.
This receiver is from the days when one could buy stereo equipment at record
stores (yes, records, vinyl, remember those?). Yes, it is solid state :-).
Fortunately, Harman Kardon 25 years ago was willing to sell service manuals.
I do not know what their policy is today. And, even in 1970 dollars, it was
cheap - $3 including postage!
This series of problems occurred over the course of 25 years or so - it isn't
as though the whole thing turned to you-know-what one day :-).
Problem 1: Stereo Beacon(tm) doesn't work.
This one is easy - the usual underrated incandescent lamp has bit the dust.
Rather than replacing it with (1) an overpriced part from Harmen Karden that
will (2) burn out in a couple of years, I locate a bright LED and 220 ohm
resistor. Since this actually runs on DC of about 5 volts from the stereo
detection logic, no additional components are needed. The LED will last well
into the next century.
Problem 2: Noise on FM only, left channel.
This took a bit little more work. Since only one channel is affected but no
other audio sources, this limits the possibilities greatly. It cannot be in
the tuner itself nor in anything common to Aux/Phono/Tape inputs. However, it
appears as though there is a clump of circuitry unique to FM mode just beyond
the tuner. It is apparently an amplifier between the tuner and the source
select switch. Why would they need an additional amp? I have no idea but in
any case, it is only 4 or 5 components for each channel. The active element,
a transistor marked with an HK house number, seems to be sensitive to
vibration. At that time, I was more conservative than I am today (and didn't
have an ECG cross reference book either) so I actually ordered the exact
replacement. Can you believe it? Problem solved.
Several years later, a similar problem developed with the other channel (or
maybe it was the same channel - faulty replacement) and guess what? I just
threw in a 2N4401 or something like that. Indeed, it sounds every bit as
good!
Problem 3: Weak right channel.
This one developed over the course of several months and affected all audio
sources but not the Tape Out suggesting a fault in the power amp. The fact
that it didn't happen suddenly meant it was probably something like a dried up
electrolytic capacitor. Since there were only a few of these in each channel,
I simply jumpered across each with a known good cap until the problem went
away. (Actually, I picked the big ugly one first - which happened to be the
culprit!)
Problem 4: Tuner stuck on unknown station.
While just about to settle into the Evening Concert, what happens? Reception
of my classical station is lost and some rock station replaced it. In fact,
the tuning knob had no effect whatsoever.
My initial thought was that the tuner was hosed which would be bad news since
I really don't like troubleshooting tuners! The service manual does include
actual tuner schematics. Since other audio sources work just fine, this would
point to the FM tuner.
However, first things first: check power supply voltages. It turns out there
is one that is unique to the FM tuner. And sure enough, what should be 10 V
reads 0 V!
Well, this isn't anything fancy: resistor->10 V zener->capacitor filter. The
zener is a dead short. The resistor doesn't look so healthy either. In fact,
while it tests within tolerance, it would appear to have seen better days
being somewhat discolored. There are signs of long term overheating on the
circuit board as well.
Since nothing else really seems faulty, I replace the zener and resistor with
ones rated at twice the power just to be safe. Like the Stereo Beacon(tm)
light bulb, these appear to be somewhat underrated as well. I wonder how many
cents they saved!
Problem 5: Intermittent stereo.
At first I thought the station I was listening to was not broadcasting in
stereo for some reason but after checking a few others dial locations, it
would seem that the stereo selection circuitry in the receiver wasn't working.
However, prodding and pressing of components on the MPX circuit board resulted
in the stereo coming back and staying there. As best as I can determine, this
turned out to be a polystyrene capacitor with an intermittent internal
connection. Pulling on this component only resulted in restoring correct
operation the next time the stereo dropped out.
Problem 6: Erratic volume, mode, speaker select, etc.
Another easy one: contact/control cleaner in the various pots and switches.
Comments: If you get the idea, this receiver has been a problem, you would
actually be wrong. It dates to 1965 and has been in continuous daily use
since then with no plans for early retirement. While not trouble free, name
something you could buy today that would have a 30+ year service life with so
few problems and no major failures? My total cost of maintenance over the
years has totaled less than $10.
Yes, new audio equipment tends to have control panels that would look more
at home in the cockpit of a 777 but that really doesn't make the music sound
any better. This receiver is a case study in simplicity and I would be
happy to have something modern which is as easy to use and maintain. Speaking
of which, my Yamaha R8 now has a case of amnesia - forgetting its settings
when powered off. Probably a bad SuperCap but that is for another Repair
Brief!
Patient: ExecuFone(tm) multifunction desk phone - 20 button memory, clock
with alarm and timer, speaker phone, etc.
Symptoms: Pickup of handset results in static rather than a dial tone.
Repeated attempts may result in dial tone eventually.
Testing: Tried replacing cords and using different phone jack - no change.
This rather nice full featured phone was given to me by the original owner in
working condition - this wasn't a problem originally. (The owner, a typical
CEO type, isn't happy if his toys are more than 6 months old. OK, he probably
didn't actually pay for it either). However, the erratic behavior developed
over a period of a few months (a couple of years later) to the point of the
phone not being usable.
No doubt a bad connections problem somewhere but where?
Internally, there are several small relays as well as leaf/microswitches for
on/off hook. It didn't take long to determine that one of these microswitches
was at fault.
Under the lever operated by the handset is a bank of 3 separate microswitches.
Two of them seem to be fine but the other would appear to have burnt contacts.
In an ideal world, I would hunt up a replacement - and of course there would
be a service bulletin on a modification to prevent the same problem from
reoccurring - but this was not to be.
Instead, I carefully unsoldered a pair of the microswitches from the circuit
board and popped the top off of the bad one. After cleaning the contacts
as best as I could, I replaced the original switches swapped in position.
(Unfortunately, I also ripped a couple of pads off with the switches so this
complicated the reassembly slightly.)
This did restore operation for a couple of years. I knew I shouldn't have
passed up the same model phone at that garage sale! It would have yielded
two good microswitches - no doubt the reason it was being sold was this same
problem.
I still use the phone for its autodialer and clock but have another (ATT)
phone on the same line to grab the dial tone! :-).
On a separate note, interestingly, the phone's instruction manual recommends
replacing the NiCd battery pack used for memory and clock backup every two
years. That same battery pack has been just fine for the last 13 years!
Comments: This isn't rocket science but how many phones are cluttering up
land fills due to just these sort of simple problems? In this case, it is
almost certainly a design fault - there is nothing funny about my phone lines.
Since the problem didn't occur suddenly, it is also not due to a lightning
strike or other one time event. Just, gradual deterioration of an underrated
set of switch contacts.
Patient: Slightly battle worn Aiwa Sports AM/FM/Cassette boombox. Well, it
doesn't appear very 'sporty' but I suppose styles change...
Symptoms: The cat did it! Fell off window sill. The cassette works fine
but there is absolutely no sign of life from either band of the
radio.
Testing: Sure enough, not even a peep (or meow) out of the radio on AM or FM.
Jeff comes by my office: "Wanna fix a boombox." This isn't that Aiwa disaster
is it? (Repair Brief #65 - Aiwa CSD-707 Boombox CD Player - Doesn't Recognize
CDs). "No, this is mine. Fell off a window sill. The radio doesn't work."
OK, sure bring it by.....
So, a couple weeks later, Jeff walks in with this vintage boombox. You really
want to fix that thing? It is also an Aiwa, but probably at least 15 years
old. "Yep, I opened it up but didn't see anything wrong." Sure Jeff, open it
up again. "A couple of the screws are missing and one of the plastic post
thingies broke off in the fall." Right.....
At first, there didn't appear to be anything obviously broken so I decide to
power it up (after all, Jeff had already done this so the additional risk is
minimal). Fortunately, it can be powered easily with the two halves of the
case separated.
Sure enough, the radio is dead as a brick. Pressing and prodding produces
absolutely no change.
Then I notice something... Uh oh, it looks like the corner of this circuit
board is cracked off. OK Jeff, you will have to remove it and it doesn't
look like fun as the dial cord seems to run from the main case to the variable
capacitor pulley on the circuit board. Better make a detailed diagram of
exactly the way this is run so that we can replace the string if it pops
loose. "Yeh, sure."
I come back 5 minutes later to find the board removed. "No problem, the dial
cord didn't need to be disturbed." Chalk one up for Aiwa - intelligent design.
It turns out that the pulley engages a keyed shaft so the string and all the
other dial cord stuff remains in place.
With the board removed, the extent of the damage is evident - not really that
bad, just two traces that run around the corner of the board. These are easily
jumpered from wider expanses of copper near their destinations.
Then, Jeff notices that the other corner is also broken - another pair of
traces, possibly extensions of the same ones.
Inspecting the remaining two corners indicates that these survived intact.
Power! Sure enough, the radio now appears to work fine.
Jeff, close it up.
"Sure boss :-)."
Five minutes later: "I even managed to anchor one of those corners with a the
broken piece so that at least 3 of the corners are attached AND I won't let
the cat near it again!"
Comments: It could have been much worse - many fine traces run really close
to the corners but they were spared. A better mounting arrangement to reduce
stress on the circuit board could have been designed (either by Aiwa or for
the repair) but these things are only supposed to be subject to reasonable
bumps - not total abuse.
I don't know about Jeff. A while ago, there was another boombox - or maybe
a clock radio - that had been dropped with similar consequences.....
Patient: Mid size Kenmore microwave. 'Customer' just said it doesn't
work and is willing to pay 'whatever it takes'. Really must like
it - nothing special.
Symptoms: About half the buttons on the touchpad are dead and even when it is
started, there are normal sounds but no heat.
Testing: Not much more can be done without removing the cover.
As usual, the cover comes off easily. It is somewhat gunked up from lack of
cleaning but not excessively so. There is a schematic of the microwave
generator but as expected, not of the controller.
Since the unit is not totally dead, the main fuse must be ok.
A quick examination of the power circuitry reveals the likely cause of the
no-heat problem: a deteriorated lug and terminal on the filament of the
magnetron. This likely occurred over time as a slight resistance due to
corrosion or the lug just being a little loose resulted in heating and
eventual failure.
Rather than attempting to install another FastOn(tm) type lug, I elect to
drill a hole in the tab on the magnetron itself and then use a ring lug,
nut, and bolt, to attach the wire. There is ample metal for this and once
filed clean, it results in a secure connection to a magnetron that would
otherwise require replacement.
I inspect the other connections to the magnetron and high voltage transformer
(secondary and primary sides) but they appear to be fine.
A quick test with a cup of water shows that the microwave generator is now
functional.
To prevent accidental operation of the HV circuits while testing the touchpad
and controller, I unplug the connector for the AC input to the HV transformer
from the controller board.
Now, for the touchpad. This is not going to be nearly as easy. In order to
determine if the problem is in the touchpad (as expected) or the controller
(which would be bad news), I need to reverse engineer the touchpad matrix.
For some unfathomable reason, manufacturers never seem to like to use a nice
orthogonal logical layout. Each 'row' or 'column' line snakes its way all
over from here to yonder. Of course, there is also printing or other opaque
paint in random locations to further stymie this activity.
However, there do appear to be two distinct sets of buttons with different
behavior when pressed:
* The working buttons result in a dramatic drop in resistance.
* The dead buttons result in erratic or no change in resistance.
Eventually, it would appear that one of the row wires is damaged somewhere
in an inaccessible location inside the touchpad. In fact, careful examination
of the front of the touchpad reveals a very tiny dent - possibly a break - in
the plastic cover. Perhaps, a knife was thrown at the panel? While it
doesn't seem to account for the problem, this is the only evidence of damage.
To confirm the diagnosis (though I am pretty sure at this point), I go back
to the ribbon connector and use a 10K resistor to manually jumper a row and
column pin that correspond to one of the bad buttons. Sure enough, that
function now responds.
The touchpad is only available from Sears parts at a cost of about $30 but the
'customer' is quite happy to pay this even when informed that the magnetron
may be on its way out of this world in the not too distant future. (As it
turns out, I have not heard of any problems with this oven and it has been a
couple of years now.)
Comments: I have not really figured out the sequence of events accounting for
these failures. Perhaps, the no-heat symptom occurred first and the user
attempted to convince the oven to cooperate by whacking at the touchpad.
If the touchpad problem happened first, the entire oven would have been
virtually useless as some of the numbers and other functions were affected.
Patient: Paula's vacuum cleaner. One of those newer 'mostly plastic' things.
This is not like the good old days where a vacuum was so heavy that
a fork lift was needed to carry it up the stairs. OK, so maybe they
weren't such 'good old days' after all :-).
Symptoms: No air flow at all. This vacuum really doesn't suck.
Testing: The official Corn Flake test is not required to evaluate this one!
"Sam, do you want to look at a vacuum cleaner.". Sure, what is wrong? "It
doesn't seem to pick up the way it used to." OK, maybe it is just a belt or
clog. "OK, it is in my car. I will get it later."
Trying the vacuum on Paula's office floor, it indeed really doesn't do much of
anything. The brush does spin - so the belt is good. The bag is not full.
In fact it has nothing in it. The air passages appear clear. So much for
those theories.
This will require some disassembly.....
Removing the bottom (or was it top?) cover doesn't reveal anything new. For
all intents and purposes, it would appear that while the motor makes the
normal very impressive 'mega horsepower' sound, there is no airflow - zero,
zappo, nadda.
Naturally, Hoover uses Torx screws for all internal parts. At least it is
modular and easily disassembled - two screws to remove the entire motor unit
and four more to remove the blower cover.
Ah Ha! Paula, Paula, where are you? Look at at this: The blower wheel has
come loose from the motor shaft and indeed is not spinning at all despite the
motor noise.
"Can you fix it?"
I can attempt to reattach the blower to the motor but eventually, a new blower
will probably be needed - I have no idea how much that would cost.
So, I pile the remains of the vacuum in a corner of Paula's office and will
take the motor/blower unit home to work on it. Unfortunately, since the
Hoover isn't in any condition to do its thing, the pile of dirt and dust
dislodged during disassembly will have to be cleaned up the old fashioned way.
Is there a broom in this place? :-).
Well, I thought it was just the plastic blower. Wrong! After some additional
disassembly of the motor itself, it would appear that the bearing at the blower
end of the motor has entirely disintegrated - balls, race, covers, everything.
This was likely the actual cause of the blower failure causing it to rub and
then bind resulting in its loosening and becoming detached from the shaft.
Well, I don't keep a collection of ball bearings for nothing!
Rumage..rumage..rumage.... Ah, here is one that has the correct ID and OD but
it is sealed on only one side. Oh, well, cannot have everything. I locate a
metal washer and file out its hole so it will fit the shaft and protect the
bearing - somewhat. I orient it so that this improvised shield in toward the
inside of the motor - hopefully, a less hostile environment. I check end
play (because of the additional washer) to be sure that there is adequate
clearance - there is.
Now for the blower. Due to the effects of the shaft spinning inside the stuck
blower, the mounting hole is totally distorted. Therefore, I use a reamer to
enlarge it symmetrically and then make a bushing out of some roof flashing to
center the blower on the shaft. It was mounted by a nut and this will still
work but for good measure, I drip some windshield sealer (similar to Duco
Cement) into the assembly to prevent the assembly from working free and then
tighten the nut securely. Then some more adhesive to seal the nut.
After allowing this affair to dry overnight, I am ready for the big test!
Using a Variac, I can run the speed up from 0 through normal line voltage
to the 140 VAC limit of the Variac with no obvious problem and no excessive
noise (well, in a relative sort of way) or vibration. The air flow is really
quite impressive!
A few days later when I return to reinstall the motor: Paula, I think it will
work for a while - have no idea for how long. If it dies again, we will need
to order the proper part. "Great! Now I can clean my apartment."
After putting everything back together, I then proceeded to vacuum up the new
pile of dirt and dust dislodged during reassembly as well as my office which
was in much worse condition! :-).
This vacuum really sucks now!
Comments: At first, I was going to blame the excessive use of plastic for this
problem. However, it would actually appear that the cause was the failure of
the blower-end ball bearing. This, then caused the blower to bind and be
literally ripped from the motor shaft.
How could a ball bearing fail? This isn't subjected to excessive loads of
any kind and I doubt that the vacuum had seen that much use - it just isn't
that old (no reflection on Paula's housekeeping). I have a 45+ year old
Filter Queen (well, they haven't existed for decades - built 'em to last!)
that still runs fine. Ball bearings do not fail that often! Thus, one can
only assume that either the bearing was defective or its seals permitted dust
to enter.
The much greater use of plastic in modern appliances has indeed reduced weight
and cost (or boosted profit margins depending on your point of view). In this
case, flimsy plastic probably wasn't a factor but one can point to many
instances of parts being under designed for the application resulting in
premature failure.
Patient: Paul's wife's car stereo. Subaru stock in-dash AM/FM/cassette.
Symptoms: Reception would come in and out. Paul noted that BIG bumps seemed
to affect the radio. Also problems with rear speakers.
Testing: No BIG bumps handy. Tapping it didn't seem to do much.
Paul had been bugging me for several weeks to look at this problem. In
general, I don't like dealing with car stereos as removal is a pain, powering
on the bench is a pain, and reinstalling is a pain. The key work here is
'pain' :-).
However, I reluctantly agree to at least take a look at this one.
The first step is to figure out how to remove the radio (I tend to call
anything in a car dash a radio even if it will mix drinks and is DVD-ready).
Fortunately, the 'radio' is held in by four Philips head screws. On the first
go-around, that is far as I got. For whatever reason (I don't recall
exactly), I was unable or unwilling to take it to the lab (perhaps, the
connectors weren't obvious or there were no connectors).
At the same time I took a look at the rear speakers which Paul had also been
complaining about. The right speaker was kind of mushed with a damaged cone.
I asked Paul if he wanted to repair or replace it. "No, as long as it makes
some sound." (He wanted to sell the car eventually.) I kind of unmushed it
but don't expect any crystalline highs. The left speaker seemed to have become
detached. Paul, exactly what kinds of bumps have you been going over lately?
Pushing the connectors back in place restored its sound.
This doesn't cure the main erratic problems so inspection of the interior
of the radio is next on the agenda.
The next week, we are prepared for action. I have Paul park the car near
our rear exit for convenience if tools or parts are needed. Then, I remove
the radio again from the dash and pull off the top cover. Nothing is
immediately visible but touching some vertically mounted daughter boards
results in the sound coming in and out.
Sure enough, a careful inspection of the solder side of the main board from
the bottom reveals some hairline cracks in the connections between one
daughter board and the main board. How the manufacturer expected something
like this to survive road bumps and potholes escapes me. However, it would
seem that this sort of construction is used in many types of equipment subject
to mechanical shock. (I will tell you about the Tandy multi-band radio with
the erratic volume control someday.)
OK, so I go back into the office, fetch our Weller soldering station and a
pair of 50 foot extension cords...... It doesn't take long to touch up the
two rows of pads.
That did it. The reception is now solid and totally immune to as much
whacking and wiggling as I could provide. This was actually a lot less
traumatic than I had expected.
The clock in this car developed a bad connection problem several months later
but that, alone, would not qualify for a Repair Brief though it did take a
couple of attempts before the actual problem was found - a corroded cold
solder joint that required removal of the old solder, scraping of the pin and
pad, and new solder to repair.
And, yes, Paul did unload the car, mushed left rear speaker and all!
Comments: As noted, car radios are not my favorite repair activity. Perhaps
they just don't have that entertainment value in addition to being difficult
to service - did I mention how everything about them is a pain? :-).
Aside from road bumps and potholes, the electrical system in an automobile
is prone to dips and spikes - some quite spectacular. The environmental
conditions may be horrible - from arctic to tropical temperatures. There can
be condensing humidity and tobacco smoke as well.
Thus it comes as no surprise that many common failures with car radios relate
to mechanical problems and to a lesser extent, electrical abuse. In some
cases, locating the cause is easy. In others, it can be very frustrating as
Dave is currently finding out with a friend's car radio. Stay tuned (perhaps)
for a report on that one in the future.
Patient: GE top freezer 21 cu. ft. refrigerator.
Symptoms: Rising temperature in fresh food compartment. Frozen food may be
softening up also but this is not as obvious.
Testing: Checked that compressor was running (continuously). Condenser coils
(underneath) have been cleaned. They are moderately warm.
Some possibilities for inadequate cooling:
* The door is not properly closing for some reason.
* Someone messed with the controls accidentally.
* Something is blocking the passageway between the evaporator and the
fresh food compartment.
* The defrost cycle is not working and ice has built up in the evaporator
coils. This could be due to a bad defrost timer (most likely), bad
defrost heater, or bad defrost thermostat.
* The interior light is not going out when the door is closed - that small
amount of heat can really mess up the temperature (remove the bulb(s) as a
test if you are not sure.
* Low Freon can result in problems of this type but that is a lot less likely.
(These refrigeration systems are hermetically sealed (welded). Slow leaks
are unlikely.)
Although various convoluted (and expensive) faults are possible, the most
likely problems have to do with the defrost system - defrost timer, heater,
and defrost thermostat. Door seals are intact and the interior light is out
when the door is closed. Controls are set for lowest fresh food temperature.
With the compressor running and clean condenser coils, there isn't much else
that is external to the sealed system.
The most common type of defrost system on a no-frost refrigerator or freezer
usually consists of:
* Defrost timer - motor driven (typically) switch which selects between the
compressor and its associated devices (like the evaporator fan) and the
defrost heater (located adjacent to the evaporator coils). The timer motor
likely only runs when the main thermostat calls for cooling.
* Defrost heater - resistance element located in the evaporator compartment to
melt ice built up on the coils
* Defrost thermostat - closed when the temperature is below about 32 degrees F
to allow current to flow to the defrost heater. Shuts off once the ice melts
as indicated by the temperature rising above 32 degrees F.
The terminal numbering and wire color code for the defrost timer in a typical
GE refrigerator is shown below:
Black (4)
Gray (3) /o---------o Normal position - Compressor, evaporator fan.
H* o-----+------/
| o---o Blue (2)
Timer | Defrost heater Defrost Thermostat
Motor (3180 o------------/\/\/\------------o/o----------+
| ohms) 31 ohms 32 F |
| |
| Orange (1) |
o---------------------------------------------------------+--o Common
* H is the Hot wire after passing through the main thermostat (cold control)
in the fresh food compartment.
The entire timer unit is readily accessible once the kick plate is pulled off.
By turning the finger access shaft in the defrost timer, I was able to get
the reassuring click indicating the start of the defrost cycle. Sure enough,
this was followed a couple minutes later by a variety of melting and sizzling
sounds. These are normal as the ice melts and parts expand. Then, a steady
trickle of water could be found dripping into the pan below.
I also noted that the timer motor did not seem even moderately warm as it
would normally be - running as it should about 90 percent of the time. Could
the motor be bad?
Two screws and the timer assembly comes free. After writing down the
positions of each of the colored wires, I disconnect them and take the timer
to a more convenient location for inspection. The motor is stone cold....
The timer assembly consists of two parts: A synchronous gear motor and a
plastic housing with the cam and contacts. Two more screws and the cover to
this part comes free. Initially, nothing looks amiss but then I notice a
wire just hanging in mid-air. And, it is one of the two wires powering the
motor!
Apparently, the wires were tack-welded to the metal strips with the contacts
on their end. And, guess what happens when something flexes even a small
amount a few thousand times? It breaks!
A quick soldering job and we are back in business.
Once reassembled, I force several consecutive defrost cycles (until no more
water comes down into the drip pan) to assure that all the built up ice is
gone.
Five years later:
Guess what? Same symptoms. Now, the diagnosis is even quicker. This time
the wire broke near my soldered connection. To make sure this doesn't happen
in another five years, I solder the wire to a location just behind the outside
terminal - a place where there is no movement of any kind!
(The built up ice did a number on the styrofoam insulation between the bottom
of the freezer and evaporator compartment but there is no way I am going to
pay GE $75 for a 50 cent sheet of styrofoam! A little universal mending
material, a. k. a. duct tape, takes care of that!
About 5 years later, 1 week after an extended power failure (1-1/2 days! At
least it was the dead of winter).
Same symptoms. However, this time the motor is warm so my previous repair is
still intact :-). The motor is receiving power but nothing visible is turning
as confirmed by rotating the cam by hand to the defrost position - where it
would happily remain indefinitely.
Upon removing the cover of the gear reducer, the cause is obvious - a split
plastic gear hub most likely due to gummed up grease (it hardened after
cooling off while power was off) caused excessive stress when restarting.
Well, bite the bullet. Calling in a serviceman would result in a $135 bill.
Our local appliance store had the timer assembly for about $30. Just after
this episode, MCM Electronics started carrying universal defrost timers for
about $12.
Comments: A properly designed and manufactured refrigerator or freezer is a
very reliable appliance. The hermetically sealed compressor and coils can
last virtually indefinitely as long as they are not abused (like using a sharp
instrument to pick ice off of the soft evaporator coils!). There have been
instances of early failures (and possible recalls) of some models but in
general, several decades is not an unusual life span for the refrigeration
components.
The defrost system is another matter. Its parts are exposed to the elements
and get a workout several times a day. However, diagnosis and replacement is
usually straightforward (except, that is, for finding a place for a freezer
full of thawing food!). With luck (of sorts), the timer is the problem and
this is a 5 minute repair if you are willing to just replace it.
Patient: Panasonic VCR model PV4820.
Symptoms: Just blinking --:--. No response to any front panel buttons or
remote control.
Testing: Nothing much without going inside.
This should have been a 15 minute repair (if you don't count gaining access to
the power supply):
* If I had been able to read my ESR meter correctly :-(. (No fault of the
ESR meter.)
* If I had simply installed a power supply rebuild kit.
* If I had elected to use my scope to check a signal or two sooner instead of
later.
As it turns out, it was more like a two hour repair at the cost of a pair
of smoked zener diodes.
The blinking --:-- with no response to any buttons is a classic symptom of a
power supply problem. In the case of Panasonic PV48xx and other models using
similar power supplies, the culprit is usually C21 or C16 (though not all may
use the same numbering), or some other electrolytic capacitor in the secondary
side of the switchmode power supply. In fact, this turned out to be the case:
C16 showed high ESR and I should have found that in about 2 minutes but did
not and was led on a wild goose chase.....
Anyhow, back to the story.
With the cover off, the power supply connector is readily accessible. Although
there are a number of variations on the power supplies used in the various
models, all should show a couple of pins with a voltage of around +5 VDC.
Probing carefully reveals a bunch of much higher voltages (which are expected)
but two pins have something around +3.6 to +4 VDC. Checking my reverse
engineered Panasonic power supply schematic, these pins should have the +5 V.
No problem I say, remove the power supply, check C21. It should test with high
ESR or low uF, replace, and be done. (Actually, the common failure with C21
is to have excessive leakage but no matter, I would just try replacing it
anyhow.)
Removing the power supply module in these VCRs isn't too bad (as opposed to
some older Panasonics which have apparently been designed specifically to make
this difficult). A bunch of well marked screws holding the main board, lift
free, unbend and unsolder a couple of tabs. Then unbend a couple of tabs on
the sheet metal cover and - presto - access to both sides of the power supply
circuit board.
To fully appreciate what is going on in the discussion below, it is
recommended that you refer to the Panasonic VCR power supply schematic
available at:
* http://www.repairfaq.org/filipg/REPAIR/F_Sam_Schem.html
Can you locate the errors in the schematic at this site (unless it has been
corrected by the time you see this)? Hint: Compare IC1 and Q4 to the diagram
below which shows the relevant portions relating to the +5 output and the
feedback circuit:
T1 D8 (+3.6) L3 ==== ICP1 _
+---|>|---+--------+-----+---------+----------+---CCCC---+---_ ---o +5 Out
|:|( | | | | | |
|:|( D11 _|_. R14 / \ R17 __|__ IC1 +_|_ C16 +_|_ C17
|:|( 5 V '/_\ 220 \ / 10K _\_/_ Opto --- 330 uF --- 1000 uF
+--+ | / \ | LED - | 6.3 V - | 6.3 V
_|_ | C21 | | Q4 |/ E _|_ _|_
- | 1 uF +_|_ +-------| 2SB641 - -
| 50 V --- Q3 |(+2.8) |\ C (PNP)
| - | |/ C |
+--/\/\--+---| 2SD636 / R22
| R15 (-1.5)|\ E (NPN) \ 220
| 220 | /
R16 \ (-.7)+----+ _|_
5.6K / | _|_ -
\ R18 / /_\
| 18K \ _|_
| / -
| |
-V source o---+--------------+
Note: Measured voltages shown in ().
C21 does show slightly high ESR, so I find one in my stock and replace it.
No change. In fact, if anything, the output voltage has *decreased* slightly.
(It turns out that the replacement cap had lower ESR and actually would be
expected to make the problem worse given the actual cause.)
Next, I check all caps in the vicinity with the ESR meter. I even specifically
checked C16 which would result in excess ripple and this would feed through C21
if C16 had high ESR. Apparently, I missed it - probably ignoring the absence
of the huge decimal point thinking it read .20 ohms when in fact it actually
was 20 ohms.
Next, I figured, "Well, maybe my replacement C21 was bad as it wasn't a new
cap". So, I removed it entirely for testing.....
Now, C21 serves at least one function I know of and that is to limit the rise
in voltages as the power supply is powered up. It couples any AC component
of the +5 line back to the regulator to reduce output if there is excess ripple
or if the voltage is increasing rapidly. With C21 removed, this protective
function is not present......
But I realize this and put the supply on a Variac so I can increase the input
slowly and avoid any problems. Right :-(.
Guess what? As I am increasing the input, I hear that high pitched whine
indicating an excessive current fault somewhere in the power supply.
Well, maybe it is just the absence of the cap (not really thinking it through).
Replace it.
No change.
OK, what happened? Is the supply now badly fried? Could the VCR have been
affected as well? Well, at least the latter is unlikely.
It would appear as though something has shorted. Not the main switchmode
power transistor as that would blow the main fuse and really mess up my day.
Well, I got lucky - at least in the time to locate the problem. Measuring
across the first part I tested - a big fat diode - showed 0 ohms. Unsoldering
it made no change but unsoldering a nearby zener diode (D15) showed that it was
shorted. D15 is on the +15 V output and protects (by sacrificing itself
apparently) against overvoltage since it does not really conduct during
normal operation. (It is an 18 V, 1 W zener as I found out later).
Fortunately, I had a replacement known to be good which was removed from a
similar supply when I installed a rebuild kit.
After the zener is installed, we are back to square one - 3.6 V on the 5 V
line.
With the caps supposedly eliminated, that leaves the feedback network.
A bit of analysis shows that regulation is accomplished by feedback from the
+5 line through a 5 V (approximately) zener (D11) to the base of a transistor
(Q3). When the output exceeds 5 V, this transistor turns on which turns on
another transistor (Q4) which supplies current to the LED of the optoisolator.
Its photodiode then conducts and reduces the pulse width of the switchmode
power transistor. Got that?!
Checking at the base of Q3: -1.5 V which is nicely .7 V more negative than
the emitter. Q3 should be solidly off. However, its collector reads about
.8 V less than its supply... Hmmmmm. I replace Q3.
No change. In fact, I replaced Q4, the optoisolator, AND D15 for a second
time after I experimentally shorted the input to the optoisolator to confirm
that it was capable of increasing the output - it was :-( and D15 blew again.
Not knowing the rating of D15, I replaced it with a 1N4742A, 12 V zener.
However, this caused the +5 to read +4.1. At first I thought this was a
clue but then realized that it simply was allowing more drive by sucking some
of the power by conducting. Installing a 15 V zener remedied that. Back to
square one again. (18 V is the correct value but at least the 15 V zener
isn't loading the circuit.)
Finally, I do what I should have done early on - turn on the scope! Looking
at the base of Q3 it is now obvious that I must have missed something in
checking the caps. As the input voltage is increased, a very significant
pulse waveform can be seen riding on top of the +5 V (now +3.6 V) output.
This is coupling enough signal to the feedback circuit to reduce the output
voltage. Since C21 has been eliminated as a possible suspect, the only other
possibility is C16. Sure enough, a quick test with the ESR meter and - what
do you know - 20 ohms. That is over 100 times what it should be!
Installing a replacement and all is well - only a couple hours more than it
should have required!
Frank Fendley identified the correct value for D15 - 1N4746A, 18 V, 1 W but
to top it all off, our local electronics distributor is out of zener diodes!
Some prototyping house bought up their entire stock or so the story goes.
And, Radio Shack only goes up to 15 V! I will obtain an exact replacement
rather than cobbling something together from two lower voltage zener diodes
since D15 does serve as protection and I don't want to affect that.
Comments: This repair should have taken about 15 minutes. I have no idea now
I missed the ESR reading on C16 as I specifically went to that cap as it is
one of the two most likely causes of this problem. I guess we all have bad
days. On the plus side, the confusing situation forced me to analyze the
operation of the power supply in more detail. I even found a couple of errors
in polarity on the schematic. (It has only been available for the entire world
to see at the Sci.Electronics.Repair FAQ site for about 3 years now. You would
think someone else might have noticed!)
In the end, the only damage was about 40 cents worth of blown zener diodes and
my ego in return for a little knowledge gained.
Patient: Kenmore 26" window fan - two speed bidirectional with thermostat.
Symptoms: Does not spin in either direction on either speed.
Testing: Spinning the blades gets it going a bit but it never reaches
anything approaching normal speed and makes a clicking sound.
This is my cousin's fan and has been in use during the summer months for the
past 15 years. I had lubricated it once before but never really properly
disassembled it entirely. (That was after the floors had been refinished - the
sanding dust didn't help matters any.)
Everything structural in this fan is made of plastic except for the motor, and
speed and thermostat trim plate.
Eight screws must be removed to take off the rear grill.
Ah - now that is a problem.... It seems that the blades in this are molded
as a single unit with a section on the shaft for the clamp. Well, that section
is totally broken off so even if the motor wanted to get up to speed, there is
nothing holding the blades to the shaft! The clicking sound was probably the
result of the shaft slipping inside the hole.
However, first the motor must cleaned and lubricated (at least) as it seems
tight. Disassembly requires the removal of a front cover plate with the
controls (speed and temp.) and then 4 nuts holding the motor itself. Then,
4 long bolts which clamp the motor halves together.
The closed-end seems fine - adequate lubrication and no damage.
However, the open-end is dry. It appears to be in basically good condition so
a thorough cleaning and then some electric motor oil on the bronze bushing and
the felt oil reservoir would appear to do the trick. I add a few drops to
the other end as well just to be safe. Testing shows that the motor works
as well as the day it was built.
Now for the broken blade problem.....
Repair with adhesive is not possible as the broken pieces could never be
reliably reattached. Sears doesn't stock a replacement blade for this model
(can you believe it?) and it would probably cost more than the entire fan if
they did. So, I will have to improvise.
At first I figure on finding something in my junk boxes that would fit the
shaft and screw or clamp to the hub portion of the blades. However, nothing
seems to fit or lend itself to modification.
Therefore, I decide to fabricate a mounting assembly from some sheet aluminum
and a hose clamp.
This consists of 3 strips of aluminum stock about 3/8" x 2" x .050, bent over
3/8" at almost a right angle at one end. Holes are drilled near the opposite
end which will be used to screw these to the hub of the blades.
The source of this material was a long ago cannibalized HDS video terminal
keyboard cover plate. You probably already know that I don't throw away
much - somewhat less than gets wasted from a typical cow - never mind :-)
The right angle portions were then attached to the shaft with the hose clamp
and while in position, holes were drilled through the hub of the blades (a
roughly flat portion about 5 inches in diameter. Screws, nuts, and washers
made from the same material were then added and the entire affair was
tightened in position.
Since only the clamp portion of the blade hub was missing, there was still
a reasonably tight hole for centering and a layer of tape removed any sign
of movement. However, a bit of adjustment was needed to correct a slight
wobble. Translation: I rotated the blades and noted the high side - and then
pressed down on it!
Now, the fan runs smoothly at both speeds in both directions though I will
recommend using low as much as possible.
Comments: This represents yet another case of an appliance snatched from the
jaws of the dumpster and land fill. There is nothing high tech about either
the problem or the repair but these projects are much more satisfying than
simply going and buying a new fan. Cheaper as well - $0 in parts at least.
How long will it last? The motor is as good as new. The blade mount seems at
least as sturdy as the original - only time will tell.
Where the nights are cool, an exhaust fan is a much more energy efficient
solution than either window air conditioners or central air conditioning. The
fan sucks out stale house air which is replaced by outside air through any
open windows.
We still use an all metal window exhaust fan - it is just about 45 years old at
this point and runs fine. A few drops of oil every so often (though I must
admit to being negligent in this department of late) and it keeps chugging
along.
Patient: Dave's neighbor's JVC Model AV20TP3 20 inch color TV.
Symptoms: Single horizontal line - video appears to be present.
Testing: External whacking has no effect, no service switch. Flexing
the mainboard, however.....
Dave had been talking about this TV for a couple of weeks. So, finally,
I found it with the back off on the bench in the lab. I wander over. Of
course, there is just a dim horizontal line (Dave DID turn down the
brightness). Naturally, I cannot resist a little exploring...
I flip a switch on the edge of the board that could have been a service switch
but it had no effect (I later found out it was the CATV/TV/ANT/etc. switch).
Tapping and whacking don't do anything.
However, pulling up on the mainboard in the vicinity of the flyback transformer
results in momentarily restoring a full raster.
DAVE??? Do you know that flexing the board has an effect?
"Um, yes, I think."
OK, let's put the board up on its side - can you do that so it doesn't short
out?
"Sure."
So, I rummage around for something insulating - a flux removal brush would seem
to be appropriate.
Dave, you watch the screen while I run this over the pins. Oh, I see you have
touched up a few joints, huh?
"Just a few."
OK, here goes....
First, I attack the area of the flyback on the chance that the bad connection
is for the power to the vertical circuits which probably are scan-derived.
No change.
Next, the area of other power semiconductors - hopefully one of these is the
vertical driver.
Still no change.
Finally, just systematically over the entire board.
"STOP - there was something".
Backup, slower this time.
"Ooooo - you had it. Gone now".
This? "No." This? "No." This? "No." This? "No."
This? "Wait, yes. Good. Bad... Bad... Good.. Bad... Good."
It is an IC. Let me try each pin....
"Bad. Good. Bad. Bad. No effect."
Pin 2 of this chip. Let me try to locate the crack. Is there a magnifying
glass in this place? It took very close inspection but there was a definite
crack that would widen as the pin was pressed to one side.
Dave, on Pin 2, jumper a wire to another pin on the same trace. Also, touch
up all the other pins on this IC.
"I guess I wasn't forceful enough."
Five minutes later:
"They were all somewhat short of solder - what was there kind of disappeared
when I touched them with a soldering iron but I think they will be fine now."
What is the chip ID? 'Uh, IC421, LA7838.' LA7838 is a vertical deflection
processor. Pin 2 is 'Vertical trigger in'. Yes, that would explain a lot!
Sure enough, the TV is now solid.
Comments: There was no evidence of heating or other damage. With the meager
amount of solder present, just the very slight effects of thermal cycling were
likely enough to cause this failure in set which is something like 2 years old.
Dave's neighbor was about to toss the set because someone else had told him
it was likely a bad picture tube! Simply amazing :-).