Note: Most parameters here are given for argon ion tubes since these are most common. However, while physically interchangeable krypton and mixed gas ion tubes are very similar electrically, they will have slightly lower starting and operating voltages. This isn't a problem for starting, but the difference in operating voltage can be significant enough to cause some power supply compatibility problems such as excessive power dissipation in the regulator circuits. This should be kept in mind if substituting tube types. See the section: Comparison of Argon and Krypton Ion Tube Characteristics for a specific example.
The source for the operating voltage is either a direct line-connected rectifier/filter 'front-end' or this followed by a high frequency inverter. (RF excitation is also possible but I don't know of any commercial examples of this approach.)
A series linear or switchmode (buck) regulator may be used with the line connected supply. Some power supplies use a combination of a switchmode preregulator to drop the rectified filtered line voltage to a range where a linear regulator can run with reduced power dissipation and use fewer, lower cost pass-bank transistors. Inverters will use PWM control of their drive.
A simple low voltage power transformer with a centertapped secondary usually supplies the filament current. Fine adjustment of filament current can be done using a tapped primary or Variac. The use of AC is actually beneficial and helps to spread the heat from the discharge over the extent of the cathode (filament) by dithering the arc position.
Note that although the plasma temperature is extremely high, this is only true in the actual bore - which may start several inches away from the filament. Thus, you really do want the filament powered prior to initiating the discharge and at all times while it is running. If the filament loses power for some reason and cools down, the voltage drop between the filament and the discharge will increase leading to much greater power dissipation and hot spots. Depending on the main DC power supply, the discharge may just go out without any harm but for some, this may not be the case and the regulator may keep trying to maintain the specified current as the tube voltage climbs. I don't know how many times you can get away with this and how quickly it will damage or kill a tube but would rather not find out!
A design discharging a capacitor into a high current high voltage pulse transformer is normally used.
Large frame Ar/Kr ion tubes can be over a meter in length and nearly everything about them is, well, much larger. :-) There are even 8 FOOT (2.5 m) long monster medical lasers that output 35 W or more and require over 600 V at 35 A to power the tube. Figure on a direct feed from a local electric utility substation for this kind of power! Ion lasers like these may also have axial permanent or electro-magnets surrounding the tube to concentrate the discharge and other 'stuff' that we will kind of ignore. ;-) They also require several gallons per minute from a tap or chilled water source to prevent a melt-down.
For these reasons, while the offer of a cheap or free large frame ion laser may sound tempting, consider the power and cooling requirements before dragging it home. It will likely end up as a coffee table support or high-tech sculpture if you don't have industrial strength three-phase power at your disposal! Cooling water may also be a problem. Nonetheless, most of the basic information on small air-cooled ion lasers DOES apply to their bigger brothers as well if the numbers are adjusted appropriately. And, the power supplies are quite similar. In fact, the same power supply can often be used for a wide variety of ion lasers by selecting the AC input and changing some jumpers.
Throughout this chapter, references will be made to several commercial Ar/Kr ion lasers systems. Two of the most common are:
A few oddball versions of the 60X (60XB and 60XC) are still manufactured but the 532 is an obsolete model. American and Omnichrome both make/made a few zillion variations on the same theme. All of them look similar and have common features. American still makes replacement tubes for the 60X but they do much higher power with newer generation technology. Call ALC and request a generic brochure if you are curious (or more than curious).
Photos of Various Laser Systems, Power Supplies, and Components has detailed views of various argon/krypton ion lasers including examples of the 60 series from American Laser Corporation. However, note that the ALC power supply shown in the photos will drive the same laser heads, it is NOT the same implementation as the Omni-150R described in detail in the section: Omnichrome 150R Power Supply and 532 Laser Head (Omni-150R/532). They differ physically as well: The ALC (there are a variety of versions) is in an elongated goldish-aluminum ("Alodined") box (as in the photos) while the Omni is shorter and painted black.
Complete schematics of the power supplies and typical laser heads for these units are provided in the chapter: Complete Ar/Kr Ion Laser Power Supply Schematics.
Unfortunately, most of this did not work with Netscape V3.04. Perhaps, it will work with your browser and/or the problems have since been corrected.
See the section: Laser Safety with respect to the optical hazards associated with these higher power lasers. While generally not in the metal cutting class, careless use of Ar/Kr ion lasers can certainly result in instant and permanent damage to vision. There are all going to be at least Class IIIb and some are Class IV.
However, when compared even to large HeNe lasers, there are many additional very real dangers associated with Ar/Kr ion laser power supplies:
In addition:
Some basics:
For small air-cooled tubes, the power supply runs off of single phase 115 VAC and deliver around 150 VDC to the regulator. High power lasers will run off of 230 VAC, 240/208 VAC three-phase, or something even nastier to produce up to 600 VDC at 45 A OR MORE!
Note: Throughout this document, we use 115 VAC as the nominal line voltage in the U.S. However, the actual measured voltage may range from about 105 to 125 VAC and still be considered to be within acceptable limits by the utility company. For this single-phase system, using both Hot legs of the line will then result in a nominal 230 VAC which may actually range from about 210 to 250 VAC.
A buck-boost autotransformer (NOT isolated!) may be included to more closely match the line input to the requirements of the ion tube thus reducing the stress on the regulator (see below).
This is the simplest (!!) sort of approach to take for these tubes. For testing, a 5 to 10 ohm ballast (possibly variable) resistor and a Variac can be used (without a regulator) as long as one eyeball is kept on the current monitor. Of course, the ballast resistor needs to be able to dissipate several HUNDRED watts - the element of a space heater may work well for this. See the section: Constructing Low Ohm High Power Resistors.
This can use a pass-bank of power transistors operating in linear mode (which makes a nice space heater for colder climates) or as a chopper operating as a switchmode buck converter. (Note, however, that the latter type is usually still line-connected - there is NO isolation transformer.) The Lexel-88 and ALC-60X/Omni-532 (150R) power supplies are typical of each of these approaches, respectively.
Such high power inverters can be tricky to design. One approach I have considered is to use the salvaged inverter from the solid state power supply from some model microwave ovens. (However, these modules are sort of rare since the simple half wave doubler operating from a HV line power transformer/HV cap/HV diode is such a low cost reliable design and not very models used them.) The power supply for a full size oven is normally designed for at least 1,000 W (which is ideal) but at too high a voltage (3 to 5 kV pulsating DC rather than the 100 to 110 VDC we need). However, by rewinding the secondary of the ferrite transformer (only a few dozen turns of heavy Litz wire would be required) and adding some primary side filter capacitors, it could be adapted to an Ar/Kr ion laser application. I have not tried this as yet but have one such module sitting around impatiently waiting for its call to duty. :-)
See the section: High Frequency Inverter Type Microwave Oven HV Power Supplies for more info on these modules.
Therefore, for all of these reasons, the voltage multipliers, flybacks, and HV pulse transformers typically used with HeNe lasers cannot be pressed into service for starting Ar/Kr ion tubes. In addition, the full 10 AMPS or whatever of operating current would need to pass through them once the tube starts (with the HeNe, it is just a few mA).
The usual way to generate the starting voltage is with a high current pulse transformer. One design uses a toroid with 30 turns on the secondary - in series with the anode - and a 1.5 turn primary. Discharging a 1 uF, 400 V capacitor into the primary results in an 8 kV output pulse. There may be resonating and snubber components as well. Care must be taken in design to minimize undershoot on the secondary - which can just as easily blow out the discharge as well as start it!
This type of pulse/resonant igniter is almost always required for older and larger ion tubes. However, for some 'small' short bore designs, other options may exist. See the section: Alternative Starting Circuits for Small Ar/Kr Ion Tubes.
Salvaged microwave oven transformers with the HV windings removed work well in this application. In this case, simply 'adjusting' the number of turns or partial turns on the modified secondary may be adequate for setting filament current. However, it may require an axe, power drill, torch, and several chisels to remove the baked on coil. I *am* exaggerating but only slightly. :-)
Like ion laser tubes, arc lamps usually require a starting voltage (up to 30 kV depending on type) so even the igniter may be built in to the unit! You will need a separate filament supply but that is a small price to pay for not having to build an entire power supply.
I don't know how the regulation on the typical arc lamp power supply compares to a well designed ion laser power supply but it should be much better than nothing! And, these devices will have all sorts of protection and safety features required for commercial applications.
However, not being designed with the particular peculiarities of ion laser tubes in mind, there could be potential instabilities or other regulation problems. Conceivably, additional (fractional Ohm) series resistance may be needed. Be aware of the possible symptoms to know what may happen. See the section: Plasma Oscillations and Other Instabilities in Ion Lasers. h3>Using Common SMPSs to Construct an Ion Laser Power Supply The following is one person's experience with adapting a pair of surplus multiple output switchmode power supplies (SMPSs) to power a Cyonics/Uniphase argon ion laser head.
I wouldn't recommend this approach for someone without a great deal of experience in switchmode power supplies as the most likely result would be lots of smoke and disappointment. However, it does represent a possible alternative to buying an expensive commercial power supply or building one from scratch, particularly for locations where only 230 VAC is available. Surplus SMPSs are readily available, often for next to nothing. With a bit of work (OK, maybe a bit more than a bit of work!), they can be converted into the main portion of an efficient ion laser power supply.
CAUTION: Not all multiple output switchers can be easily modified in this manner. More extensive changes may be necessary or it may simply be impossible to get from point A to point B using the existing design.
(From: Mike Harrison (mike@whitewing.co.uk).)
I've written an article which describes the modification of some cheap surplus switchmode power supplies for powering my argon ion laser tube. Go to: Electricstuff and take the link: "Argon laser power supply from surplus switchmode PSUs".
It's been my experience and that of some medical technicians that tubes ran for long terms on switchers at high currents die from sagged cathodes.
You can/will experience similar problems with a 3 phase variable drive speed drive depending on your line loading.
A medical technician friend of mine travels with a professional Ronk rotary converter, its much smaller then the 10/12 hp motor used in the home made phase converters and you can lift it with two hands. We had no trouble running a 3 phase Lexel-95 supply (yes the huge heavy one) in my garage with the Ronk, its a little noisy, but it seems to be the best solution for single phase.
If you really have no choice, or just like a challenge, by far the most straightforward approach is to use a line connected rectifier/filter with a linear regulator. While not quite as efficient as a switchmode or inverter type, the hassles are fewer and parts are more readily available. For initial testing, a low value high wattage ballast resistor can substitute for the regulator. Just don't be tempted to leave it this way permanently. And, in any case, don't neglect the essential current monitor!
Note that if all you want to do is test a newly acquired Ar/Kr ion tube, there may be a considerably simpler alternative to a continuous high current power supply. See the section: Pulsed Operation of an Ar/Kr Ion Tube.
The igniter (starter) circuit may operate off of this DC+ supply or an additional low current 'boost' source to use a higher voltage (which reduces the turns-ratio of its high current pulse transformer - a hard to wind (or high cost) part.
The current regulator uses bipolar or MOSFET transistors in a linear or switchmode (buck converter) configuration. Some power supplies use a combination of a switchmode preregulator to drop the rectified filtered line voltage to a range where a linear regulator can run with reduced power dissipation and use fewer, lower cost pass-bank transistors.
Feedback based on current, light (beam) sensing, or an external modulation signal, is used to maintain the proper discharge current through the tube. For initial testing, this can be done manually using a high power adjustable ballast resistor and possibly a large Variac on the AC line input to the power supply. However, light feedback is required to minimize optical noise in the laser beam and to maximize ion tube life.
+----------+ DC+ +-----------------+ H o------| |-------------------------| Igniter Circuit |-----+ | Main | +-----------------+ | AC | Bridge | | Line | and | +------------+ Light Feedback (option) | | Filter | DC- | Linear or |<-------------------------+ | N o------| |-----| Switchmode | F1 +-----------+ | | +----------+ | Regulator | +------|-+ | )-+ | | | +------------+ | | ) |-|-------+ | +------+ | | +----|-+ | Tube+ | ) +---|---------+ | F2 +-----------+ | Filament )||( | | Ar/Kr ion tube | Supply )|| +---+ Tube- | | )||( | | ) +---------------+ +----------+This type of supply with a linear regulator would be the easiest to construct. While efficiency is lower (abysmal instead of just terrible - probably about 40% more heat generated than for a well designed switcher), the difficulties of implementing a robust and fool-proof Pulse Width Modulator (PWM) controller are eliminated. Just provide a large enough heat sink and jet turbine driven cooling fans!
Since everything except possibly some logic or low level analog circuits are directly line connected, great care (even more than just considering the 1,500 W or so of raw power we are dealing with!) must be taken in the basic construction, testing, and adherence to ALL safety precautions, implementation of ALL safety, electrical, and thermal interlocks and protective devices. In many cases, even that control circuitry is line-connected as this simplifies the implementation since no isolated interfaces are required.
The igniter (starter) circuit may operate off of this DC+ supply, a separate winding on the inverter transformer, or an additional low current 'boost' source to use a higher voltage (which reduces the turns-ratio of its high current pulse transformer - a hard to wind (or high cost) part.
Feedback based on current, light sensing, or an external modulation signal, throttles the DC to DC inverter by Pulse Width Modulation (PWM) - controlling the duty cycle of the drive to its chopper transistor(s) in a manner similar to that for controlling a series pass switchmode regulator. However, there are added design considerations in dealing with the characteristics of the high frequency ferrite inverter transformer.
+----------+ DC+ +------------+ +-----------------+ H o------| |-----| DC to DC |------| Igniter Circuit |-----+ | Main | | Inverter | +-----------------+ | AC | Bridge | | --+ +-- | | Line | and | | )||( | Light Feedback (option) | | Filter | DC- | )||( |<-------------------------+ | N o------| |-----| --+ +-- | F1 +-----------+ | | +----------+ +------------+ +------|-+ | )-+ | | | | | | ) |-|-------+ | +------+ | | +----|-+ | Tube+ | ) +---|---------+ | F2 +-----------+ | Filament )||( | | Ar/Kr ion tube | Supply )|| +---+ Tube- | | )||( | | ) +---------------+ +----------+While still very dangerous to troubleshoot, at least the main tube circuits are isolated from the AC line by the inverter transformer. Aside from the slightly reduced risk of frying yourself, I would probably not recommend the inverter approach unless you already have its foundation such as the HV power module from a solid state microwave oven! For more information, see the chapter: Ar/Kr Ion Laser Power Supply Design. Ion Laser Power Supply on Foreign Power The most common example of this would be to run a USA model power supply (115 VAC/60 Hz) on European power (230 VAC/50 Hz) or vice-versa. By far the most significant difference is the voltage as there is usually very little that is affected by line frequency in an ion laser power supply but see the comments below. The best solution will depend on the model and type of the power supply (e.g., linear or switchmode) so there is no one-size-fits-all answer. Obviously, it is best to check with the laser or power supply manufacturer. However, here are some general comments on the voltage conversion:
As for line frequency, going from 60 Hz to 50 Hz may affect the amount of filtering required after the rectifiers, and heating of the low voltage power transformer(s) and AC relays (if any). However, this really isn't likely to be an issue unless the design was marginal to begin with! The only other issue would be any logic (like the preheat timer of the Omni-150) that uses the line frequency as a clock and would then run slow - likely not a problem in any case. Powering a 50 Hz unit on 60 Hz should be fine except for the possible timer running fast. In either case, if there are already taps and/or jumpers for selecting dual voltage operation, the frequency difference was probably already taken into consideration. If not, then this will need to be evaluated as well but the required changes, if any, should be minor compared to those required to convert to a different line voltage.
Note: Where the use of a panel meter is suggested below, if you salvaged the meter movement from some other equipment or your junk, box, double check the actual sensitivity. There should be a rating printed at the bottom of the meter face or the back of the unit like "fs=1mA", "fs=10A", or something similar. Even though the scale may be labeled with a particular set of units and values, a series or shunt resistor may actually be needed to adapt the basic movement to read at that sensitivity. An AC meter may actually use a DC movement with an external rectifier! It's also a good idea to test the meter to confirm that it hasn't been damaged due to an overload or just age.
A .1 ohm 50 W power resistor will develop .1 V/A across it resulting in 1 V at the maximum likely tube current of 10 A. The shunt must be stable with respect to temperature (which is one reason for using a 50 W resistor when the maximum sustained dissipation is only 10 W). The system should then be calibrated with the actual voltmeter or mA meter to be used.
While a VOM or DMM can be used, it is really best to have a dedicated meter for continuous discharge current monitoring. There are two reasons:
Note: Most cheap VOMs or DMMs won't reliably measure resistances below a couple of ohms. To measure a low value resistor, get a D cell or power supply capable of a couple of amps and wire it to the unknown resistor in series with another power resistor that will drop the current to a value you can measure (or calculate). Knowing the current, you can measure the voltage across the unknown resistor and calculate its value from R=V/I. WARNING: If you do this on a previously powered laser head, just make sure the caps are discharged!
One way to implement a permanent current monitor is to use a basic panel meter. This can be a moving coil (D'Arsonval) or digital type. Add a variable series resistor for calibration using the current shunt. For example, a meter rated at 1 mA full scale will require a total resistance of 1K ohms (for the meter itself and current limiting resistors) connected across the .1 ohm shunt to read 10 A full scale.
Rs .1 50 W Tube- o======+==================/\/\=================+======o DC- | | | +--+ M1 | | R1 | v R2 + +--------+ - | +---/\/\---+-/\/\--------| 0-1 mA |-----+ 940 100 +--------+ Calibrate 10 A Full ScaleSplitting the current limiting resistor into the fixed R1 and Calibrate pot (R2) allows easy fine adjustment without requiring an expensive 10 turn pot.
Make the main current carrying connections to the power resistor. Tap off of this go to the meter. That way, it is less likely that a bad connection can result in the shunt opening - which would fry your meter in an instant!
Calibrate this setup using a DC power supply and known accurate ammeter.
With a dual trace scope, for example, set up the vertical channels to be A-B (A, invert B).
A dedicated panel meter is best for this as well. Obtain a panel meter with a full scale sensitivity of 75 to 100 V or add an appropriate series resistor to a current (e.g., 100 uA full scale) meter. Make sure the resistor(s) can handle the maximum voltage!
+--+ M1 R1 | v R2 + +----------+ - + o------/\/\----+-/\/\--------| 0-100 uA |----------o - 940K 100K +----------+ Calibrate 100 V Full Scale
An isolation transformer is highly desirable for personal safety and to protect your (grounded) test equipment.
Given that these are not usually handy, make sure you understand the safety guidelines provided in the document: Safety Guidelines for High Voltage and/or Line Powered Equipment - AND FOLLOW THEM! Minimize the amount of probing done under these conditions and make connections only with the unit unplugged and large capacitors discharged!
Once you attach the laser head, ideally, one would possess 5 or 6 functioning eyeballs to keep tabs on all aspects of a new Ar/Kr ion tube and power supply combination as it is being tested for the first time.
(From: Steve Roberts (osteven@akrobiz.com).)
Where there is no robust current regulator, for initial testing it is wise to start at 7 or 8 Ohms and work down, even with the pi section CLC or CRC filter as these discharges do possess a negative runaway characteristic despite what the literature says. For some reason it always seems to need a 2 to 4 ohm offset above the calculated value or it will run away.
They also need a minimum current of about 3.5 amps to sustain the discharge with a new tube and about 4.35 A as the tube grows older, this was just measured in my shop with a new tube versus a tube with 2000 hours on it. The new one (an Omnichrome 532 with a buck converter) dropped out on line voltage dips when it came from the factory set for a 3.25 A lower limit. This was on 118.2 V AC.
Therefore when firing up a 'heater' (space heater ballast) based supply, you often have to lower the resistance a little at a time to find the minimum stable current point. It is better to start a high resistance and work down. Of course, you will have a permanent current monitor in the system!
Also see the sections: Measurements of Current and Voltage in Ar/Kr Ion Laser Power Supplies and Testing with a Dummy Load.
It will be necessary to power the logic and control circuits separately for this test so that their supply voltage is constant.
WARNING: Some commercial power supplies simple do not like to power anything but a real tube with its nice non-linear V-I characteristic. This is particularly true of switchmode designs. Therefore, I do not recommend using a dummy load for these as many expensive parts will likely blow.
Provide wired-in monitoring of BOTH load (tube) current AND regulator voltage drop. (See the section: Measurements of Current and Voltage in Ar/Kr Ion Laser Power Supplies.)
Select the load rating, with the Variac at 0, power on the main supply. Bring up the voltage slowly.
The best type for testing probably have internal mirror ion tubes - eliminating another set of hassles. If mirror alignment turns out to be the only problem, you can really win big since this is often correctable. See the sections starting with Problems with Mirror Alignment for internal mirror tubes and External Mirror Laser Cleaning and Alignment Techniques.
Meredith Instruments currently (Summer, 1999) has a limited supply of exactly this sort of item available for $50 each. They are a batch of Uniphase 2214-30SLT argon laser heads that were pulled from graphic arts equipment and were tested with an original Uniphase adjustable current supply at 10A. Output power is (they claim) 10 to 20 mW (should be at least 30 mW) and tube voltage was low indicating low pressure and near end-of-life. However, for power supply testing, what more can you ask for? You won't be putting thousands of hours on it and if/when the tube dies, you still have a nice case and igniter and sense boards, and can install a new or rebuilt tube without fear of it being destroyed by a defective power supply! Send email to info@mi-lasers.com for more details on these as well as slightly higher priced laser heads that are in better condition.
Needless to say, when designing such a power supply and selecting components, err on the side of being conservative. Select parts to run at much less than their rated voltage, current, or power (maximum stress of 1/2 the part's rating isn't too bad a rule-of-thumb). This is especially important for devices that fail in a sudden catastrophic manner like power semiconductors!
Where a high power regulator uses a large series-parallel combination of power transistors, if one shorts, it may overstress others. If there is inadequate protection, multiple devices will fail in rapid succession until the entire regulator behaves like a blob of solder taking other parts with it. Even components that still appear to function may have suffered permanent damage during the overload.
There are also some comments on the repair of specific power supply models in the chapter: Complete Ar/Kr Ion Laser Power Supply Schematics.
For much more information on the servicing of these types of devices, see the following (as appropriate for your power supply):
Before beginning a lengthy troubleshooting session, check the following:
Two types of problems are common:
WARNING: Do not put your eyeball (or any other part of you!) near the tube. Aside from the high voltage, there is considerable UV which isn't any good for organic matter either!
If under light control, the current will likely be pegged at the maximum (e.g., 10 A) attempting to get a proper power beam.
Check the mirror alignment and/or clean and/or realign the resonator as appropriate. See the sections: Checking and Correcting Mirror Alignment of Internal Mirror Laser Tubes or Argon/Krypton Ion Laser Cleaning and Alignment Techniques (external mirror tubes) depending on the type of laser tube in your system.
WARNING: Some or all of these power supplies are line connected - take extreme care with measurements! See the section: SAFETY when Dealing with Ar/Kr Ion Laser Power Supplies.
If there is no evidence of the tick-tick-tick sound, check the (Boost) supply directly at the igniter circuit.
It is also possible to test the output of the igniter by disconnecting it from the tube (with power off and everything discharged!) and then powering up and checking that it will arc 1/3" to 1/2" to a suitable ground point.
For more on startup and igniter faults, see the sections: Omni-150 Power-Up Sequence and Troubleshooting and Igniter Problems and Troubleshooting.
Where the power supply and igniter check out, the tube pressure may be too high (from non-use) or the tube may be misbehaving just because it felt like it - but this may be reversible. It may be possible to the tube to start using an Oudin coil and then run it for several hours to drive down the pressure. See the section: Hard-to-Start Ar/Kr Ion Tubes - Outgassing and Keeping Your Laser Healthy.
I cannot stress enough the need to have a curve tracer around, even a home-built one, for checking pass-bank transistors in-circuit. However, the following procedure using only a multimeter will also work for identifying bad transistors in the Lexel-88 (or most any other) pass-bank in-circuit (where all the collectors are tied together) is:
That's also why the Lexel has the 'stress' meter on the pass-bank. If the meter won't get out of the 10 to 30 Volt range, there is a shorted pass-bank transistor. The correct 'green' range is 10 to 70 volts. For a combination of a power supply configuration and Ar/Kr ion tube that should work, selecting the proper tap on the line input buck/boost transformer will aid in getting the transistors operating in the right range and is the first thing to check when firing up the laser.
In all cases, it's still desirable to obtain matched devices, or at least devices from the same batch number. When purchasing from a large electronics distributor, there is a very good chance any specific order will be filled with parts from the same batch. It may also be possible to request them and/or order (likely at a higher price) devices matched to some specific requirements. Replacing all of them may also be prudent even if some test good since they may have been overstressed and could then be the weak link, blowing themselves and everything else shortly after your repair.
Yes, I can tell what you are thinking: "The designer of the SG-IL1 whimps out and buys a commercial power supply." :)
The Omni-150P is virtually identical to the Omni-150R (I have been informed that it is basically the same design but with a different PCB layout and use of a commercial current sensing transformer/Hall sensor instead of the custom one used in the 150P. There were also a few choice expletives in the description but none affecting performance.) The schematic I have is close enough in any case (as such things go) for troubleshooting.
The seller even agreed to include the most likely components to have blown - the power MOSFETs in the main chopper. So, what the heck, it will make a nice toy and provide much information for the FAQ!
The replaies are from Steve Roberts.
I received the Omni PS from Dave. It, to put it mildly, needs some work :-(.
Dave sent along replacement MOSFETs - the originals had been pulled as had the snubber diode (D21), and all socketed ICs except the CD4040. So, I need to find replacement for those. The power supply cooling fan is also missing.
It is also kind of beat up.
While not totally disappointed, this isn't quite what I expected.
When I asked Dave about its condition originally (back when I made the decision to buy it), he had indicated that it was given to him along with another similar power supply to test a head, and he hadn't tried this supply at all since the other one worked. Now, perhaps, someone had eviscerated it before he got a hold of it but at least part of its state would be quite obvious from the outside.
(From: Steve.)
This is the normal condition of used laser equipment, and if there are no scorches, your lucky. Knowing Dave quite well, if he had the chipsets, you'd get them, he's not that kind of guy. Odds are somebody did a amateur hour attempted repair on them with out a schematic or scope, which is suicide, then Dave got the job to resurrect one for the club that uses it in a projector. So now you get to play detective, finding what was originally wrong and finding the new added problems.
Does it smell like pot smoke or cigarettes? Its usually one or the other, if its the later, clean the covers and fins good as they usually have high levels of nicotine oil on them, as a club laser is near the ceiling and sucks all the smoke through its cooling ducts and the PSU. Its highly toxic in such concentrated form if your sweating. I speak from a personally blinding experience. Temporary but scary. :-(
(From: Sam.)
Actually, inside it is quite clean - dusty, but no great gobs of nasty coatings. I don't know if I would recognize pot smoke residue but it just smells like old electronics - not burnt, just dusty. I spray painted the outside now so it doesn't look too bad there either :-).
There is even dust on the empty sockets so I assume it has been like this for quite a while.
I haven't tested anything on it yet but at least there aren't any obvious scorch marks or holes blown through the circuit board!
(From: Steve.)
The PWM chip is $3.40 at Digikey or about $6.50 at any place that sells the ECG or NTE line of TV repair parts, and the op-amps are no more then $1.95 at Radio Shack.
Speaking from experience, you usually need to replace the op-amps and PWM chip anyways for sanity reasons. You have no idea what the last guy did his measurements with, possibly something destructive like a MEGGER :-). This is a mission critical unit, and its better to replace than miss something stressed and have it pop again.
First stop: The +15 and -15 V regulators, 95% of the time its one of them has let out its precious internal smoke and one of the microfuses (look for olive drab, chartreuse, or cyan unmarked resistors) is popped, these are generally 3/4 amp picofuses. This surge takes out a op-amp or two, but the discrete parts are usually fine, save for the MOSFETs and fast diode. The MOSFETs are usually $17 to $24 each.
If a picofuse is popped, look at the 110 V and anode leads where they go onto the board, odds are these are misconnected, either that or the external fan wires were shorted. There are a couple of possible combinations, but only one works and doesn't pop a fuse.
REMEMBER TO ISOLATE YOUR SCOPE AND DMM FROM GROUND USING A PROPER ISOLATION TRANSFORMER OR A ISOLATED SCOPE OR BATTERIES.
(From: Sam.)
Yep, God knows I have enough warnings in the FAQ about this! I was actually thinking of hooking up a couple of big transformers I have back-to-back as an isolation transformer for the duration.
Actually, for initial testing until I actually fire the tube, my small isolation transformer should be fine.
I think I will also rewire the fan for operation off the -15 (4 diodes to drop it to 12 V) for a DC fan - I assume the one in there was for 115 VAC - not easy to find in my junk box.
Didn't you also mention a cap that blows along with that stuff?
Assuming it matches the Omni-150R - Power Subsystem, which cap would it be? C35? Why would that blow? Or is there another one I missed? Or is it C27 which I couldn't really determine from the schematics you sent?
(From: Steve.)
Yep, its the electrolytic snubber in the PWM circuit, 200 to 300 uF if I remember right, C35 on your schematic. I can't remember ever seeing two the these supplies that are exactly the same, usually the changes are minor. Check the components around the PWM chip, see if it matches, other then that you should be fine.
(From: Sam.)
The Cyonics tube will be fine for testing and I won't worry too much about damaging it. I never did get it to lase with my pulsed power supply, maybe with a real one it will do better but it does have a bad case of cathode sag....
Note: This was before I had succeeded in getting the Cyonics tube working with SG-IX1.
(From: Steve.)
See if you can get a hene beam down it cleanly.
(From: Sam.)
I did that. I checked mirror alignment best I could with a long baseline HeNe setup. Yes, I know, not really valid (at least at one end because of wedge) but it was TOO perfect by any test I could perform to totally write if off. I did try torquing the mirror mounts at each end in all directions while it was pulsing - nadda. Current was well above threshold (probably peaked at 15 A) for a millisecond or more.....
Yes, I know, it should have worked.
(From: Steve.)
Ah, perhaps a millisecond is too short to establish a stable plasma, pulsed lasers are usually filled about 1/10 to 1/100th of the fill pressure of a cw unit. in a week or two you'll be able to ram 11 amps down that tube if you want to.
(From: Sam.)
HeHeHe... More likely with a space heater supply and your ignite board :-). Things don't happen THAT quickly around here! I wish. :-)
Let's see.... I was running a 200 uF cap with a 15 ohm ballast if I recall correctly - probably more like 3 ms.
P.S. Will the ignite board alone produce laser pulses on a 60X? What is the pulse length when you get a tube like the NEC that doesn't start but produces pulsed laser output (as with Ben's)?
(From: Steve.)
A 60X will generate a faint pulse with just the ignite board and boost, that's how I tell if the anode voltage isn't there, it's just a faint click, only a faint whiff of lasing on higher power tubes, however if you have anode volts it's a banging or popping sound and a long bright flash if it doesn't start. If its faint you'll see a flash of purple from the plasma in a dark room.
Like I have said in the past, it would not surprise me if that pulse is 250 to 500 mW at its peak.
(From: Sam.)
Well, the ignite board and boost is a pulse of unknown width and RF content from the SCR discharge and 5 uF at 500 V or so discharging into a short circuit :-) so that should be pretty brief.
If the anode supply is live, I would guess you get some follow through on that to lengthen the pulse even if it doesn't catch?
HeHeHe, I found the op-amps and PWM chip at Mouser:
I can't find that diode (D21 - MUR1560) at MCM, Digikey, or Mouser.
I discovered another pair of smoked resistors - R4 and R5 - pullups to the PWM drive. :-(. It may be the only one actually fried but the other was in a somewhat sorry state - both measured correctly however.
Also, D11 and D19 have been replaced with a single diode - there may be other changes in that area as well.
My 12 VDC fan is now running happily off of the -15 V supply through 4 diodes. :-) It only draws .125 A so the 7915 regulator should handle it easily - while the circuitry for the positive and negative supplies are similar, there is less of a load on the -15 V (since the relays are powered from the positive supply) so there should be ample margin. I added the following circuit off of the 7915:
Analog Gnd o----|>|---|>|---|>|---|>|-----+--------+--------+ D1 D2 D3 D4 | | | C1 +_|_ C2 _|_ +----+----+ D1-D4: 1N4002 22uF --- .1uF --- | 12V FAN | 25V - | | +----+----+ | | | -15 o------------------------------+--------+--------+The purpose of C1 and C2 are to prevent any fan motor switching noise from propagating BACK to the analog power supply.
I have disconnected the controller AC from everything else so will be able to troubleshoot it without worrying too much about Hot AC and meltdowns!
(From: Steve.)
The last diodes I bought were from Digikey, IXYS makes a equivalent, so does IR, it's a 20 to 100 nS diode. The driver resistors are always smoked but functional, in every unit I've ever seen. Usually somebody monkeys with the clamp diodes on the hot side of the driver. Most people assume they are shorting out the driver, when in fact they are a pulse forming network, so take a look there as well.
(From: Sam.)
OK, I found some IXYS diodes in Digikey: The most likely options are: 14A, 35 ns, 600 V or 17A, 40 ns, 600 V. MUR1560 which is what was on the schematic is 600 V, 15 A 35/60 ns (whatever that "/" means).
My Mouser order arrived - replaced the 220 uF, 250 V cap and the 10 uF 350 V one as well just for good measure. I have 5 each of the SG3525AN and TL074ACN. count 'em! I am now ready to fire up (hopefully not literally!) the control subsystem with the power circuitry disconnected (and the MOSFETs and D21 not even installed).
I see that R30 in this unit is 150 ohm, 5 W instead of the 330 ohms listed on my Omni-150R schematic. Hmmmm.
And, someone has jumper/bypassed the filament tap switch - it tests bad (no continuity in any position) so I will remove it and replace with a machined pin IC socket and insulated 2 pin male jumper.
I have now set the mainboard up so that power is applied only to the low voltage power supply transformer via a Variac on an isolation transformer!
Anyhow, after confirming that the pinout made sense, I plugged in an SG3525 (without any op-amps) to see what would happen.
Fortunately, I still have some ancient DIP clips that fit 14 and 16 pin ICs!
Poking around its pins shows nothing of interest except on pins 5 and 7 which have a ~100 KHz sawtooth and pin 16 which has 5.1 V - this is promising. However, there is only a pegged at +15 rail voltage on pin 13, the output. OK, so maybe the op-amps are needed to at least provide a valid input signal. What the heck - in go the TL074s (after confirming the +/-15 V power to their sockets).
Now, I have a reasonable looking signal from pin 13 and a similar drive signal on TP2. Although I don't have the control panel hooked up, I assume this corresponds to the Standby preset pulse width setting.
Next, as a change of pace, I want to see if the Preheat Timer - the CD4040 - is functional. An initial test seems to show no activity here but then I realize it will only do something with the interlock chain grounded. Since this is normally provided via the control cable, I add a jumper from pin 11 to pin 8 of the CD4040. At first I thought it was still dead but then I heard the main relay (RLY1) kick in - just about 30 seconds after applying power. Sure enough, completing the interlock after applying power results in a 30 second delay and then relay activation. There is something a bit strange about the behavior - the order of things matters - but that may be normal. It basically works.
So I sent some email to Steve:
Greetings:
I installed my PWM and op-amp chips. Running with the mainboard by itself (with only the low voltage power supply transformer powered via a Variac through an isolation transformer :-) I have drive pulses on TP2 at about 100 KHz I think and the preheat timer seems to work enabling RLY1 after 30 second - though it doesn't always appear to reset properly by just interrupting the interlock chain.
Do you have any advice on minimizing the chance of blowing expensive MOSFETs? I haven't installed Q8/9 or D21 yet. I know - don't try to power a resistive load, use a real tube. What about adding a light bulb or space heater in series with the power feed to limit current? What blows these MOSFETs besides trying to deliver 10+ amps into a short circuit?
(From: Steve.)
What kills 150s are: Back spikes from heads with no blocking diode, and bad MUR1560 diode. If you inject a signal to the PWM chip does the duty cycle change? Have you set the initial voltages and 40% PWM duty cycle up yet?
The blocking diode is a 11RA90 or similar in series with the tube. They are traditionally big heavy can diodes for two reasons: (1) They seem to survive much better then plastic packages, and (2) they can handle the CW 10 amps passing through.
(From: Sam.)
I'm pretty much ready to fire this thing up (hopefully not literally) but still need to find some TOP3 size mica insulators for the MOSFETs. Stay tuned.
You'd need a very high voltage source (like 5 to 10 kV or more) to charge a fairly large capacitor to get them to lase cold cathode. Plus, a spark gap or thyratron switch to keep the tube from just glowing. Aside from running a high chance of poping the glass to metal seals on every pulse, the sputtering and gas cleanup will kill it rather quickly. You're also going to see discharge down the gas returns on Spectra-Physics tubes (at least). And once the discharge goes down the returns once, it rarely returns to the bore. That hot cathode makes a world of difference in getting a good clean arc.
And, even though the metal end-bell might look tempting to use as a large area cathode, this really isn't a viable option. It isn't heated, so the sputtering at high current will be very severe, rapidly coating everything with a metallic coating. Consider that a HeNe laser tube runs at only a few mA, not 10s of AMPs. And, when the cathode coating in HeNe laser tubes is depleted, even this low current results in a sputter overcoat in a few hours of operation.
Tubes designed for pulsed operation run at 1/10th to 1/100th of the pressure in their CW counterparts. The gas cleanup rate is huge.
If pulsed output is desired, keep the cathode hot and simmer the tube with a low current arc, then flare up the current when you need it? That's the way some medical systems work, and it's a lot easier on the tube.
For determining if a new or used ion tube is good, there may be no need to run it on a full blown power supply. That way, you can hold off committing the time, money, and other resources to obtain or build one until you know that you have a working tube. (Once you read the chapter: Ar/Kr Ion Laser Power Supply Design you will know why this is worth considering - these are not like little HeNe types!) A simple pulsed circuit created using the tube as the active device in a relaxation oscillator will suffice. This has a number of other attractive benefits as well:
However, there are several caveats:
WARNING: Despite the low pulse rate and short pulse duration, the optical output from the laser running on these circuits could be enough to result in eye damage. For example, a 60X at startup may produce almost a half watt of peak laser output power for a brief period as the igniter and multiplier caps discharge into the tube - which is a similar situation to what is presented here (especially when you decide to increase the value of the storage capacitors to boost output!) - and it is banging away at the tube repeatedly! Take care and NEVER look down the tube bore while the circuit is active - even if it appears to be dead as a brick!
Argon ion lasers have been built to operate in pulsed mode. See, for example US Patent 3,555,451, granted January 12, 1971 (this may be too old for some of the on-line databases). Output was 6 W peak/2 mW average pulsed at 60 Hz A design that builds on this one can be found in in U.S. Patent #3,967,214: Stabilizing Electrode for Cold Cathode Gas Ion Laser. (The US Patent & Trademark Office currently has a search facility with free access to the full text and graphics.)
Having gotten that out of the way, here are several circuits that should be adequate for a typical small air-cooled Ar/Kr ion tube.
+-------------+ + R1 R2 Vin+ o----| |------/\/\-----+-----+---------/\/\-------+ | HV DC Power | 400K | | 140 | | Supply | 10W | | 10W |Tube+ | 2 kV, 10 mA | - | | .-|-. Vin- o----| |---+ | / | | | +-------------+ | C1 _|_ \ R3 | | | .25uF --- / 10M | | LT1 | 2,500V | \ | | | | | | | | | | ||Z.| | | | o - Test + o '+-+' | | | | Rs | F1| |F2 NC o-+ T2 +-----------+-----+---+---/\/\---+ | | )|| _|_ 1 | | | AC o----------+ || - | | | )|| | | | Variac )<--------------+ T1 | | | 0-140V )|| )|| +-----------------|----+ | 1A )|| Filament )||( Tube- | | )|| Transformer )|| +-----------------+ | +--+ 3VCT,15A )||( | | )|| +------------------------+ AC o-------+-------------------+
This will form a relaxation oscillator using the tube with current limiting to about 10 A. Adjust the pulse rate by either varying the input voltage or changing R1 (with power off!).
Running this at a few pulses per second for a reasonable length of time (i.e., not for days on end) should result in no significant damage to the tube or shorten its life by any detectable amount. You shouldn't need to run it this way for very long in any case - just don't think that this setup can be used in place of a REAL power supply!
As long as the peak current exceeds the tube's lasing threshold, there should be visible flashes of laser light from its OC (output coupler) end if it is working and aligned correctly.
WARNING: This circuit is still dangerous - just less so than a full blown ion laser power supply. The anode of the tube (including the mirror mount at that end!) will have a voltage of up to 1.5 kV with respect to ground (for this example). While the amount of energy stored in C1 is fairly small - less than 0.5 J (W-s), it can still be lethal under the wrong conditions. The HV power supply itself can deliver up to 5 mA through R1. Either of these are at least enough to evoke a reflex response which may ruin your whole day even they do not kill you. Take care.
Note: I show the entire setup earth grounded including the tube cooling fins and support structure. This makes it safe to touch everything BUT the tube anode (and of course, the HV power supply). Floating the entire affair is also possible but most of the same problems exist since portions of the tube will still be at the negative potential of the power supply and, if you use the scope monitor points across Rs, will be grounded through the scope (unless you isolate that as well - which is not recommended).
It may also be possible to use this approach for starting small to medium size tubes since it provides a 'boost' voltage like that used by the igniter of the ALC-60X/Omni-532, SG-IT1, and SG-IL1. See the section: Pulsed Operation of an Ar/Kr Ion Tube.
What this design provides is two power supplies driven from a single 650 VRMS center-tapped transformer (T3). Many other approaches for the power sources are possible. See the chapter: HeNe Laser Power Supplies for ideas.
Note that D7 can be built from 4, 1000 V, 2.5 A diodes in series since the single cycle (pulse) rating for these is much higher than the approximately 10 A (15 A without the regulator) peak that is required.
The (optional) constant current regulator allows the tube current to be set at a high but safe value. If a simple resistor is used, either the current would be lower than desired for most of the discharge, or higher than the maximum tube specs for some portion of it. If you don't use a regulator, change R2 to 33 ohms at 10 W (for 12 A peak).
R4 D3 +---/\/\---|>|---+----+----------------+ | 10K 1kV | | | | 10W C3 _|_ / R5 | | 75uF --- \ 220K | | 350V | / +-------------+ | | | | Constant | | +----+ | Current | | | | | Regulator | | C4 _|_ / | (Optional) | | 75uF --- \ R6 +-------------+ | 350V | / 220K | | | | D5 | | +----+ +---|<|---+ | _|_ | 3kV | - | 2.5A | | T3 R1 | C1 D1 | R2 +--/\/\---|---||---+---|>|---+----+----+-------/\/\-----+ ||( 2M | .01uF | 3kV | | 5 | ||( | 3kV | | | 10W |Tube+ AC o--+ ||( 325V | | | | .-|-. )||( | | C2 _|_ / | | )||( | | .01uF --- \ R3 | | )|| +---------+ | 3kV | / 20M | | LT1 )||( | | \ | | )||( | D2 | | | | AC o--+ ||( 325V +---|<|---+ | ||Z.| ||( 3kV | | o - Test + o '+-+' ||( | | | Rs | F1| |F2 +----------------------------+----+----+---/\/\---+ | | 650VCT 1 | | | 50mA NC o-+ T2 | | | )|| | | | AC o----------+ || | | | )|| | | | Variac )<---------------+ T1 | | | 0-140V )|| )|| +--------|----+ | 1A )|| Filament )||( Tube- | | )|| Transformer )|| +--------+ | +--+ 3VCT,15A )||( | | )|| +---------------+ AC o-------+--------------------+
Setup and operation is similar to that described in the section: Ar/Kr Ion Tube Pulse Test Circuit 1. Adjust T2 to obtain the proper filament voltage for your tube and modify the value of R1 to vary the pulse rate.
The remaining details are left as an exercise for the student! A switchmode buck converter will be needed for the optional regulator unless you have a bank of really high power transistors gathering dust in your junk box. :-) The problem with using a linear regulator is the peak power dissipation and keeping inside the SOA (Safe Operating Region) for the transistor(s). A common BUT12A would handle the current and voltage individually for this example but not the peak 4,000 WATTs - 400 V AND 10 A at the same time!
WARNING: Take care as C3 and C4 can pack quite a wallop - especially once you increase their size - as I know you will. ;-) And, both supplies can deliver dangerous levels of current continuously even without the capacitors!
An alternative which may work for some small tubes like the Cyonics (those which will start without help from a boost source) is to use a line powered (non-isolated or 1:1 isolation transformer) supply for the pulse current source followed by an (optional) linear or switchmode regulator.
Without the regulator, it would look like the following:
R1 D1 R2 +2 kVDC o----/\/\-----------+----|>|-----/\/\---+--------+ 100K | 3kV 100 | |Tube+ C1 _|_+ 1A 10W | .-|-. 1uF --- | | | | 3kV | - | | | R3 | D2 R4 | | | +150 VDC o----/\/\-----+-----|----|>|-----/\/\---+ | | LT1 100,25W | | 3kV 4 | | C2 _|_ + | 6A 10W | | 500uF --- | | | 200V | | ||Z.| | - | '+-+' DC RET o-------------+-----+----------------------+ F1| |F2 | | | NC o-+ T2 | | | )|| | | | AC o----------+ || | | | )|| | | | Variac )<---------------+ T1 | | | 0-140V )|| )|| +--------|----+ | 1A )|| Filament )||( Tube- | | )|| Transformer )|| +--------+ | +--+ 3VCT,15A )||( | | )|| +---------------+ AC o-------+--------------------+
Details of this, too, are left as a exercise for the student!
A 600 VCT power transformer (T2) charges the energy storage capacitor (C1) to approximately 425 VDC and also drives the parasitic voltage multiplier to generate an additional starting voltage of up to more than 2,500 VDC. When the Ar/Kr ion tube starts, C1 discharges through D9 with a current limited to about 10 A by R3. The uF value of C1 may be changed to provide the desired discharge energy. Adjust the values of R2 and/or C3 to assure that C1 charges in a shorter time than it takes for the HV to build up to the point at which the tube starts.
My first version of this circuit was built as an all-on to a 30 year old home-brew tube-type bench power supply (remember the 5U4GB rectifier tube?). I never thought I would ever find a use for that again but it did have all the connections required to attach the output and voltage multiplier conveniently located on front panel binding posts!
C2 C3 C4 +------||-------+------||-------+------||-------+ | D3 | D4 D5 | D6 D7 | D8 R3 / +--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+--|>|--+ 1M \ | C5 | C6 | C7 | R4 / +------||-------+------||-------+------||-----+-+--/\/\--+ | D1 | R1 D9 R5 | 100K | T2 +--+--|>|--+-+---/\/\-------+----+--------|>|------/\/\--|----------+ ||( | 1K | | 3kV 30 | | ||( | 10W C1 +_|_ / R2 2.5A 10W _|_ C3 |Tube+ ||( 300V | 10uF --- \ 470K --- .01uF .-|-. ||( | 450V - | / 1W o - Test + o | 5kV | | | ||( | | | | Rs | | | | || +------------|--------------+----+-------+---/\/\---+----+ | | ||( | 1 | | | LT1 ||( | T2: 600VCT, 50mA | | | ||( 300V | D1-D8: 1N4007 | | | ||( | C2-C7: .01uF, 1.2kV | ||Z.| ||( D2 | | '+-+' +-----|>|----+ AC o--------+ T1 | F1| |F2 )|| +-----------|---------+ | (Ac input and Filament )||( Tube- | | T2 primary Supply )|| +-----------+ | not shown) )||( | )|| +-----------------------+ AC o--------+
Like the other pulse supplies, this can also be used as a starter for some small ion tubes. All that is needed is a high voltage high current blocking diode between the ion tube anode and the DC+ output of a the normal ion laser power supply.
All laser PSUs work over a much wider range then the tube needs, to accommodate for tube wear. No two tubes ship with exactly the same operating characteristics due to minor differences in the cathode, gas pressure, bore size, and Brewster window quality. So the PSU is built wide range to accommodate them all. Often the same PSU is designed to work a whole series of head models over a wide power range. It's desirable that the PSU cover a wide range in case your tube is high pressure and hard starting. It also helps in diagnosing problems if you can sweep your tube over a wide range of currents. Short periods of low current wont hurt it, but attempting to start at low currents is very hard on the tube and PSU. Running forever at below rated idle current will damage a tube as fast as overcurrent.
So you need to hit the magic number, amps-wise. For example a SP-168 likes to run at 30 to 35 amps, but not much below 25 amps, then you're doing damage, and 40 amps is way into overcurrent, but the PSU can ram 45 amps into the tube if the tube is new and you don't mind buying pass-bank transistors often.
There are also some cases in the lab when you need to dial down a laser, either to get it going in "single mode" oscillation or to reduce background noise in the beam, at a known sacrifice of tube life. Also some lines are sensitive to both magnetic field and current, at a certain product of I x B they drop out of lasing. So you need a wide range. For example, on the krypton yellow, green, and deep violet lines.
At the other extreme, overcurrent: For example, my Lexel-88 argon ion laser with its brand new tube can sustain 3.5 watts (21 amps) for a very short period of time, then drop back to idle at 10 A. But, at about 9.5 A it drops below lasing threshold, and at 9 A or so the discharge is highly unstable and probably is damaging the tube. Nominal running range is a watt at 13 to 15 amps. But it was designed as a pulsed eye surgery laser, and thus needs a wide range of available power. On the back of the PSU is a duty cycle warning saying at a 5% duty cycle I can way overcurrent the tube without much damage. 1000 hours from now based on past experience with other tubes, I'd need 18 to 19 A or more to sustain a CW power of one watt and could not obtain anywhere near the 3.5 watts, as the cathode would be depleted and the pressure would be low and internal dust on the Brewsters would be obscuring much of the beam. hence the wide PSU range.