P75 System Board
   Indicator Port Pinout (J17) 
Slot Numbers and Widths
P75 SCSI
   Terminators
   Remove SCSI BIOS Chips
   SCSI and System BIOS Wondering
Set SCSI ID
SCSI Cable
Memory 
   Filling SIMM Sockets
   Non-OEM Memory
   64MB in a P75?
   Actions On Memory Errors
Floppy Drive
   Floppy Controller
   2.88MB Capable?
      Attempting 2.88MB Support (Always use the stock refdisk!)
      Undorking a P75 
   External Storage Device Connector
      External Storage Device Cable Source
   Remove Floppy Drive (H1 Model)
Battery Pack
   J18 Pinout
   P75 CMOS Battery Replacement (Dr. Jim)
Processor Card
Video Card
   J1 and J2
   Brightness Program
Riser Swap with P70
Windows 98SE
Built-in Features ADF

J8 Wire Colors
Pin1 on left, pin 10 on right (silkscreened on board) 
Blue (9), Orange (10), Red (11), R (12), R (5), Black (8), Bk (1), Bk (2), Bk (3), White (4). Numbers in ( ) are the corresponding power supply connector 

Indicator Port (J17) Pinout
   This is the Indicator port- three LEDs are driven off this header, Power, HD, and Floppy. Note that the header on the system board and the display card (PN 65X1569) are keyed, so the pins match from 1 to 1, 2 to 2, and so forth. 

Pin   LED
1-2   Power LED
3-4   HD LED
5-6   FDD

Slot Numbers and Widths
   The two upper slots are 32 bit. The slot next to the rear cover has the AVE (dow! case is shut!) connector. The bottom two slots are 16 bit, and they are short. FWIW, Slot 1 is the internal SCSI controller. 

Slot 4 (Outer Top)
32-bit AVE 
Slot 2 (Inner Top)
32-bit MME
Slot 3 (Inner Bottom) 
16-bit 
Slot 5 (Outer Bottom)
16 bit
 
 

Place memory expansion cards in Slot 2!
 
 
 

SCSI BIOS PN 79F3214 (even) / PN 79F3213 (odd). SCSI microcode PN 79F3212 It uses three 8 pin SIPP term packs. 

Up to six SCSI devices can be attched to the SCSI controller.  It supports drives up to 1.05GB.

Terminators
  It uses three 8 pin, seven resistor SIPPs, 145 ohm (150 ohm will work fine). Radio Schlock has the 150 ohm #900-8143. Pin 1 is towards the SIMM sockets.

Removing SCSI BIOS Chips
I want to open my P75 and pull the SCSI BIOS chips. Might have to figure out some way to remove them (Dremel off the carrier, then the leads?).

Dr. Jim sez (in his best bedside manner)
   That's my boy, never use tweezers when a sledgehammer will do.  Removing the SCSI BIOS chips is trivial.  The carrier is a two-piece affair.  Slide the _inner part away from the SIMM socket (you will notice a small gap at that end of the carrier).  It will make a disturbing 'snap' sound, after which you can lift the inner portion out.  The chip will lift right out into your grubby little paws at this point.

SCSI and System BIOS Wondering
Peter wrote:
Today I did a bit experimenting with the 8573-401 P75 portable. I wanted to know if and how drives over 1GB could be used in that machine.

At first I tried using a 2.1GB Quantum Fireball attached to the onboard SCSI controller. The drive was properly recognized with 2.112MB in the setup - but it did not show up with the proper capacity in FDISK. I tried Win95 FDISK - which showed a size of 1.024MB (1.0GB) as estimated even with the advanced large media support turned on.

   Then I installed the short IBM PS/2 Fast SCSI-2 Adapter /A (60E9) in the second lower 16-bit slot. The first slot contained a short 16/4 Token Ring adapter already. While the card has no front tab it just fits between the rear slotcover and the power supply fan without squeezing :-) A very tight fit if you ask me. The internal SCSI cable could be detached from the planar SCSI port and attached to the SCSI-2 card. The length is sufficient.

Now: this seems to work. 

   It is however not possible to configure the SCSI device(s) attached to that
controller, if you only copy the ADF to the reference disk. Most likely you
will have to use "copy an Option Disk" and get the updated SCSI.DGS, SCSIHF.DGS and -probably- SC.EXE to the P75 reference disk.  I did not do that ... so far. Maybe later, because I only have a US-P75 reference and only a german "SCSI-2 Option Disk" - and if there is one thing how to voluntarily step in the mud then it is mixing different language versions.

The P75 also lacks the ability to select the startup sequence - but at least it
had no problem determining that the planar SCSI only has an external CD-ROM attached and that the 2.1GB harddisk is on the SCSI-2. FDISK got straight through and installed a 2GB primary partition. After reboot the FORMAT C: says "Formatting 2.012,03MB" ... which takes a while, because this particular drive has some bad sectors, which are not hidden. Strange. But okay.

   Physically the slimline Quantum drive fits into the half-height HD-cage. But the ID-select cable has an "old raster" plug for the large jumper pins, while the Quantum has the "mini jumpers", which are thinner and have only 1.5mm pitch. Doesn't matter much, because once set there is no reason to have the drive ID switched to anything else than the pre-set ID 6 (traditionally I took ID-6 of course). 

   Sadly the P75 has a soldered firmware ROM on the processor card. There is no chance to read out the ROM content *after* the system powered up - and then the ROM is relocated to an area above 16MB and masked from access. Only the shadow-copy remains at the adresses E000 - FFFF and is mixed with the additional codes from the XGA and planar SCSI. 

   This is the same method as used on all PS/2 with 486-class processors (and even 386 cpus) - but these have removeable ROMs that could be read and analyzed before the POST relocates the "pure ROM code" out and only the "compatibility part" remains. I would really like to compare e.g. a Type 2 (old) and 41G9361 Type 2 (new) code for the advanced SCSI support and try to add this to the P75 firmware ROM, which *seem* rather close to the Type 2. Technically the P75 is a sort of Mod. 77 / Mod. 90 mix in a portable case. It has a similar memory subsystem (single modules, unpaired) and the planar SCSI from the "Bermuda" 77, it got the planar XGA-1 from the Mod. 90. So either of these machines -to my opininon- could be used to filter out the additional code that enables the planar SCSI controller to handle drives over 1GB. 

On the Mod. 90 the code can be supplied with the upgrade version of the the
processor board firmware ROM - why not on the P75 ?
   Anyone any idea to this ? I don't have the equipment to unsolder the SMD-ROM undamaged or read it out. My Eprommer takes only DIL-chips ... but that is the least problem, because I can build / buy an adapter. The ROM itself is a 27C1001 / 27C010-version (128K x 8 bit) as far as I remember ... I have PLCC 28C010 Flash-ROMs around here ... :-)
   Hey Jim - wouldn't that be another challenge to push out the P75 performance a bit more ? The advanced SCSI support on the Mod. 77 "B" helps it to even run 8GB drives .... tried that with my old Ultimedia-77 and a Seagate Elite-9 drive, before they were installed in the 9585. No problems getting them to work under NT 4.0.

Note- both  SCSI connectors have the pin holes facing UP. The length of the flat and loose is immaterial, but the OAL should be AT LEAST 6.5".
The end for the drive should use a strain relief.

The Panasonic OEM (IGH0380BA) 4MB is FRU 64F3605 P/N 64F8776
IBM 4MB FRU 92F0105 P/N 71F7010
 FRU 6450128, 927200  are 4MB simms from IBM, the P75 is somewhat sensitive to positioning them, so you may have to move them around until it works.

Jim Shorney's layout:
SIMM1: 71F7010/63F9159
SIMM2: Same as 1
SIMM3: 64F8776/64F3605
SIMM4: Same as 3

From an original system:
64F3605
64F3605
92F0105
92F0105

Load SIMM sockets from top to bottom, largest SIMMs (memory size) first, smaller simms last. The notch in the SIMM goes to the left (towards the outside of the case) 

Filling SIMM Sockets
   When you expand the memory capacity with 2MB or 4MB SIMMs, use one of the following arrangements. Other combinations makes the access times slower. (IBM P75 Quick Ref, Pg. 20)

Non-OEM Memory
From Me
Ran into a good one. I thought, hey, what about dumping 4x4MB sticks in
my P75? Yeah, get some!
   What an education! First, I dumped in 4 70nS PS/2 SIMMs. 201/110 error. Swapped out #3 chip, cuz it barfed at about 1200KB. Didn't do it. 
Swapped out the last two 4MB SIMMs with the original BIG 4MB SIMMs. Barf at 8000KB.
   Well hell, I swapped the original SIMMs into MEM1 and 2, put two PS/2 SIMMs in 3&4. Counts it, it boots. I think it's happy. Odd, does the P75
have special mojo on the original SIMMs or what?

Peter said:
>According to Peter, he's running a later Acculogic SIMMply Ram that can utilizes four 8 Mb 72 pin parity simms with only 8 Mb (two 4 Mb Simms installed on the planar) for a total of 40 Mb.

   That's right. The P75 is -more or less- a "Model 90 with a carrying handle". So the 16MB memory limit does not directly apply on that machine (unlike to Mod. 70 and 80). However: the planar memory mapper seems to be a bit odd and needs to map in the memory on a card ("Channel Memory") into the lower 16MB to have it adressable. That's why I had to remove memory from the systemboard.
   Interesting in this respect was the fact that I had to remove 4MB when the
Acculogic was equipped with 4 x 4MB modules and I had to remove 8MB from the planar, when it had 4 x 8MB. Seems as if the memory granularity only adjusts to full module values (4 or 8MB).

Error in First 512KB
    If POST detects an error in the first 521K of system-board memory, the first 2MB block of system board memory is deactivated and the following occurs:
°  If an additional 2MB block of system board memory is installed, the addresses assigned to the deactivated block are reassigned to the second block of sbm. This is only true for the first two 2MB blocks of system board memory. After the first two blocks, an error code is displayed.
°  If additional system board memory is not installed, no address reassignment occurs and an error code is displayed.
°  The first 512KB of memory address space cannot be assigned to adapter memory.

Error Outside First 512KB
   If POST detects a memory error in any memory after the first 512KB of system board memory, the 2MB block of memory is not deactivated and an error code is displayed. In this event, the customer level diagnostics program can be run to deactivate the 2MB block and reassign the addresses. 
   After the block is reassigned, the defective block is ignored by POST during subsequent power-ons.

   If errors occur one at a time, the system deactivates 2MB blocks of memory. However, if two errors occur at the same time on the same SIMM, the system programs cannot be loaded and an error message is displayed.

The floppy in my P75 is a ALPS DFP723D15C, 12v .47A, 5v .16A

Floppy Controller
  It's a 82077AA. Probably supports 2.88MB floppy drives (yea! got to love the dust shutter), but NOT the " * " marked 2.88MB floppy drives. 

2.88MB Capable BIOS?
This is from the HMM- note the reference to a "4MB Media Sensor".... 
3.5-Inch Diskette Drive  38F5936
3.5-Inch Diskette Drive  38F5936 (with 4MB Sensor) 
 

Installation- This model has the same layout as the original ALPS
DFP723D15C. Looking from above, the drive motor is on the left (same
side as the eject button) and floppy cable is on the right.
   Just swap the plastisized shield to the new drive (nice that the big exposed flywheel is gone on the 2.88), and put on the rails. Because of the black dust shutter, you can't just slap on those rails. Instead, start the rear screw first. Now pull out on the front end of the white rail and start the front screw. Now you can snug it down with little trouble.

Longer 2.88MB Eject Button
   The 2.88 has a little longer eject button that will prevent you from closing the floppy carrier back into the case when you have a floppy in it. BUT since you now have a dust cover on your floppy, you might be able to live with it.
   If it really bothers you, swap the eject button (if you can) with the original floppy. Or file it down (how crude).

Hacking Refdisk for 2.88MB Support
 NO! Don't even try it! Do NOT try to cook up a hacked refdisk using more recent refdisks of other systems that use similar components! I did just that and I almost trashed one of my (very few) P75s!
   The problem is when you hose the POS values with the bogus refdisk, the stock refdisk WON'T WORK! Worse, it comes up as a Non-System disk. Tim Clarke is of the opinion that the later refdisks move the extended CMOS area, and that's what sends the floppy to Nirvana.

Shorting the MC146818AF RTC
   Getting to it is half the fun. Undo the 4-40 posts from the ports, then undo the two screws holding the left end of the MCA slot bracket, unplug the header from the planar, unscrew the screw at the top where the braided ground comes in from the plasma display, unscrew the screw at the bottom next to the SCSI ID header artifact, then remove the single screw that is in the center of the "T". Now rock the riser out of the planar. Wasn't that simple?
   With power off and the battery unplugged, short pins 12 and 24 together for 15 seconds. That's the certain way to clear it, yet I am not totally convinced it clears everything.
   Pin 1 on the RTC faces left, so 12 is the rightmost pin on the bottom, and pin 24 is the leftmost pin on the top. (Thanks to William Walsh for sending a snippet from the newsgroups).

   Stick the MCA slot bracket back on, replace all screws and posts.

Boot with the PC DOS 7/refdisk hybrid. Run SC.EXE and SETCLOCK.EXE, now the system should (maybe, possibly) respond.

Floppy Drive Removal
 From Dr. Jim (I'm an engineer, not a doctor) 
   OK, this is not rocket science, especially if you have an H1 floppy. 

Reach in behind the display panel with your left hand. Slide your index finger in behind the floppy housing on the left side  about 3/4 of the way up.  Then slide your finger downward along the edge of the housing until it stops on the catch.  At this point, hook your finger behind the left side of the housing and pull the side of the housing to the left, maybe 1/4" or so.  Once the little post on the floppy cover clears the catch, the floppy will flop forward. 

Continue to rotate it until it is level, then pull it straight out. 
 
 
 
 
 
 
 
 
 
 
 
 

Removing Floppy from Drive Carrier, H1 Model

  Unscrew both screws from the upper side of the carrier. Pull the floppy rearward until the white plastic piece on the right side is just to the front of the catch.
   Now pull the side of the carrier away from the drive and pull the drive up on the right side and pivoting it on the left side. NOTE: The floppy will NOT slide out of the carrier. 

   Warning! There is a aluminum ground shield across the bottom of the floppy drive. It is attached with four screws through the bottom mounting holes (which are not used by the plastic "rails"). It is coated on one side with a plastic non-conductive coating. To properly attatch this shield, the grounding strap "tab" MUST be on the same side as the motor! A simple test- use an ohmeter on the shield, one side conducts, the other, not. 

Reassembling Floppy into Drive Carrier
   This had me stopped for a little bit. You have both white plastic pieces that screw onto the front sides of the floppy. Which goes on which side?

   If you look into the recess that the drive carrier fits into, you will notice a black catch in the upper right corner. Now look at the two white plastic pieces. Notice that one has a little "arm" sticking up? This part gets screwed on the floppy on the same side as the eject button.

  Slip the left side of the floppy into the drive carrier. Pull the right side of the carrier out so the floppy "rail" will clear. Slide drive forward until it stops. Install both self tapping screws.

J18 Pinout

Battery Card plug  3 is Black, 1 is Red, 2 is NC. The battery card socket has a plastic plug in the position of the missing pin.
 
 

IC5 486DX-33 
OSC1 66.6666 MHz
ROM PN 64F9924 
P75 CPU Upgrade  (Dr. Jim Shorney's site) 
 

A lot of these are used on the XGA adapter. 

J1 and J2
Both J1 and J2 are jumpered on 2-3.
Any further info on the jumpers or chips? Tell ME! 

Brightness Program 
 To install Brightness program, run Brtinst.com from the refdisk. This installs brt.com. 
   To have emphasized text brighter than standard text, type "BRT" and press Enter. To have standard text brighter than emphasized text, type "BRT /H" and press Enter. 

The P75 has XGA-1 with 1MB VRAM.

The internal display however supports only the 640 x 480 mode and turns off if any of the higher modes is used. This will explain why you can get a picture on the external screen but not on the plasma. 

You will have to stay at the low-res mode here, since the display is from the principle only a "generic VGA screen" (same as on the P70 BTW).

Serial Port
   The built-in serial port connector can be assigned as Serial 1 through Serial 8, or disabled. 

   <SERIAL_1  03f8-03ff int 4>, SERIAL_2 02f8-02ff int 3, SERIAL_3 3220-3227 int 3, SERIAL_4 3228-322f int 3, SERIAL_5 4220-4227 int 3, SERIAL_6 4228-422f int 3, SERIAL_7 5220-5227 int 3, SERIAL_85228 -522f  int 3, Disabled 

Parallel Port
      The built-in parallel port connector can be assigned as Parallel 1 through Parallel 4 or disabled. 
     <PARALLEL_1 03bc-03bf 1278-127f int 7>, PARALLEL_2 0378-037f int 7, PARALLEL_3 0278-027f int 7, PARALLEL_4 1378-137f int 7, Disabled 

Parallel Port Arbitration Level
   The built-in parallel port connector can be assigned any one of the available arbitration levels 0 through 7. Select <Disabled> to use the parallel port in compatibility mode. 
   <Level 6>, 5, 4, 3, 1, 0, Disabled, Level 7 

Preempt Enable/Disable
   Allows the system board processor to preempt continuous data transfers by other devices for its use of the Micro Channel. 
     <Enable>, Disable 

Video I/O Address
   This selects a particular I/O address range for the Display Controller Registers. This field also affects the exact location of the video coprocessor registers. 
      <Instance 6: 2160h - 216Fh>, Instance 1: 2110h - 211Fh, Instance 2: 2120h - 212Fh, Instance 3: 2130h - 213Fh, Instance 4: 2140h - 214Fh, Instance 5: 2150h - 215Fh 

Video ROM Address Space
    This defines the memory address range used for the system video ROM. 
      <C0000 - C1FFF>, C2000 - C3FFF, C4000 - C5FFF, C6000 - C7FFF, C8000 - C9FFF, CA000 - CBFFF, CC000 - CDFFF, CE000 - CFFFF, D0000 - D1FFF, D2000 - D3FFF, D4000 - D5FFF, D6000 - D7FFF, D8000 - D9FFF, DA000 - DBFFF, DC000 - DDFFF, DE000 - DFFFF 

Video Arbitration Level
       The video sub-system can be assigned any one of the available arbitration levels 8 through 14. 
     <Arbitration level 13>, 12, 11, 10, 9, 8, 14 

Video Fairness
        Video Fairness indicates whether or not the video sub-system coprocessor will follow the fairness algorithm for bus usage. 
     <Fairness On>, Fairness Off 

Auto-Dim Time
        Auto-Dim function automatically turns the plasma display off after a specified period if no input from the keyboard or auxiliary device is received. Auto-Dim Time can be assigned from 1 to 120 minutes. 
     <10 Minutes>, 20 Minutes, 30 Minutes, 60 Minutes, 120 Minutes, 1 Minute,5 Minutes 

Auto-Dim Reset
   There are two ways to turn on the display again: 
           1. Pressing SHIFT 
           2. Clicking the mouse. 
        If the mouse picks up vibration on a desk which turns on the display again, change Auto-Dim Reset from <Keyboard and Aux. Device> to <Keyboard Only>. 
        <Keyboard and Aux. Device>,Keyboard Only 

Color to Gray Mapping
     There are some applications that are developed for color displays but can be used on monochrome displays. When using such applications, reset Color-to-Gray Mapping to <Green Signal Only>. Setting <RGB Mixed Signals> may result in an unpredictable display image. 
  <RGB Mixed Signals>, Green Signal Only 

Turned-on Display
      Primary turns on only one display that is considered by the   system to be the primary display.  When no external PS/2 display is connected to the system unit, the plasma display is considered as the primary display. Only VGA mode is supported.  When the external PS/2 display is connected, it is considered as the primary display. VGA or XGA mode is supported for the external PS/2 display. 
         Plasma and External turns on both the plasma display and the external PS/2 display. Only VGA mode is supported. 
   <Primary>,  Plasma and External