I bought a Mac IIsi for repair from the web (thrift store site). Under $100 including tax and shipping. It arrived quickly was very well packaged. It was in reasonable cosmetic condition. The logic board was in good shape. The PRAM battery was dead but intact. There were typical signs of SMD cap stress (pcb discoloration). No hard drive and the floppy was pretty beaten up with part of the controller board broken (see later). The machine had 64MB of RAM in addition to the in-built 1MB, though one SIMM was detached and the expansion sockets were in bad shape (more below). There was a PDS ethernet adapter installed but I couldn't immediately determine the vendor (again, see below).
The Sony APS-06 power supply was in bad shape: no +5v or +/-12v, but, encouragingly, the standby +5v was present (indicating a good AC rectification path). Opening it up was disgusting .. cap goo all over but luckily none had oozed out to the logic board. In one spot, the goo had even eaten its way though the legs of some resistors:
These PSUs are apparently renown for such failure. The main high-voltage caps were in good condition and tested fine in-circuit. All low-voltage caps needed replacement. Conveniently, the PSU is included on the Bomarc IIsi schematics and a full cap list is here: https://tinkerdifferent.com/threads/capacitor-list-for-iisi-power-supply.613/ I ordered replacement caps (and resistors) for the PSU and logic board.
While I awaited delivery of the caps, I turned to the Sony MP-F75W-12G floppy drive. It looked like it had been stabbed with a screwdriver. There were gouges, a chunk of the pcb had been broken off, and 3 electrolytics were missing or mangled. Fortunately, some shards of the board turned up under the logic board. Mechanically, the drive looks fine after slening and lubing. I wasn't too optimistic about fixing the controller but I did find the controller schematics for a similar model on the web (e.g. http://absurdengineering.org/librar...Corrected - Sony MP-F75W-01G Floppy Drive.tif).
I managed to superglue two floppy pcb fragments approximately back in place - so I was able to see what traces needed repair. It looked doable and, since I have a microscope, I set about fixing the traces. At this point the patches looked like this:
To replace the missing/damaged caps, I ordered a kit ($1.95 from Console5).
Before any recapping, I was able to test the logic board using a compatible IIci PSU. With no floppy, HD or extra RAM installed, the board booted to a flashing floppy screen. Mouse and cursor were living. There was no boot chime since no speaker was connected but I suspect sound wouldn't have worked because most caps were open-circuit.
When my DigiKey order arrived, I first repaired and recapped the PSU. This involved removing the regulator daughter board to replace a couple of aluminum SMD electrolytics with tantalums. Before reassembly, I verified that it delivered the expected voltages.
Next, I recapped the motherboard, re-installed in the cleaned bottom csse and again tested with the IIci PSU. I had to add expansion RAM to boot macOS and that was problematic .. the SIMMS were reluctant to connect reliably with broken sockets and the extended boot time due to RAM testing was unwelcome. But it chimed and booted perfectly. And soft-power worked.
I then installed the recapped PSU and that also booted. Yeah! But it powered up immediately and soft-power wasn't working. D'oh!
It turns out this is another common failure for the IIsi PSU due to residual cap electrolyte. I disassembled again and first bathed the main board in IPA. Much brown yuck came off but the problem remained. With PWR unasserted, I was measuring close to zero on pin 7 of the regulator IC (M71977) on the daughterboard which enables regulation. Others have reported that electrolyte can accumulate under this chip. I'd cleaned the daughterboard but I hadn't removed this IC to look under it. Ho hum! I once again desoldered the daughterboard .. luckily I have a Hakko sucker! .. and pulled the regulator IC. This is what I saw:
Gag. A ground trace was exposed and I estimate that a few kilo-ohms of leakage from pin 7 to ground would be enough to hold the regulation enabled.
With that cleaned and re-masked, and IC resoldered, and daughterboard refitted, and PSU re-assembled and installed .. soft-power works.
So I returned to the floppy drive... I beeped out the traces around the damage but it just didn't make sense. So I removed the motor cover and this is what I saw:
Yikes. It seems that the destruction stopped just before the motor coil was hit. I patched these gouges and now everything checked out.
I first hooked the floppy drive to my LCIII which had a similar fully functional unit. I popped in a known readable disc and after much seeking, the drive read but very slowly. Odd. When I tried writing (inititiazing) a disk it seemed to start Ok albeit slowly. But eventually it failed with an disk-locked error. So, signs oi life, but weirdness.
Then I checked the detector switches. The disk-lock (SW1) and disk-density (SW3) switches were good but the carriage-down (aka disk-loaded) switch (SW2) was stuck open. On reflection, this was understandable since this switch was on the edge of the destruction. There appeared to no chance of getting a new replacement and I had no potential donor device, so I decided to sacrifice the write-protect feature (which I never use) and move this switch over. The disk-lock switch has a longer plunger that the carriage-down switch but I trimmed it down about 2mm. With this change, and with the SW1 wired closed, the drive is now fine.
The expansion RAM (BankB) SIMM sockets were not in good shape. One socket in particular had broken plastic and bent or missing metal clips. I had one spare socket on an LCII donor board so I extracted that and replaced the worst socket. 64MB of expansion RMM is really too much for the IIsi. Without a custon ROM, the boot-time memory test takes a age. 16MB is a more reasonable amount for this 20MHz machine.
At this point, the machine, equipped temporarily with a spare BlueSCSI2, is haapy to boot macOS up to 7.6.1 and to run NetBSD. But the mystery network card isn't recogized. The BlueSCSI2 has DaynaPort wifi - so I have networking, provided I remove the top case to let the EMI out.
So I still don't have a definitive identification for this PDS card:
It reports itself as having vendor ID 4010 which seems to be Sonic Systems, Inc. Back in the day, they did make PDS ethernet adapters for the IIsi but I can't find any drivers out there. There's sonic.etherlan.6.1.hqx file that unpacks to a Sonic installation disk but this fails to provide anything for this card. There are some other vaguely Sonic files on the web but they appear to be vacuous or corrupt. Please let me know if you can identify this card or point me to software.
The Sony APS-06 power supply was in bad shape: no +5v or +/-12v, but, encouragingly, the standby +5v was present (indicating a good AC rectification path). Opening it up was disgusting .. cap goo all over but luckily none had oozed out to the logic board. In one spot, the goo had even eaten its way though the legs of some resistors:
These PSUs are apparently renown for such failure. The main high-voltage caps were in good condition and tested fine in-circuit. All low-voltage caps needed replacement. Conveniently, the PSU is included on the Bomarc IIsi schematics and a full cap list is here: https://tinkerdifferent.com/threads/capacitor-list-for-iisi-power-supply.613/ I ordered replacement caps (and resistors) for the PSU and logic board.
While I awaited delivery of the caps, I turned to the Sony MP-F75W-12G floppy drive. It looked like it had been stabbed with a screwdriver. There were gouges, a chunk of the pcb had been broken off, and 3 electrolytics were missing or mangled. Fortunately, some shards of the board turned up under the logic board. Mechanically, the drive looks fine after slening and lubing. I wasn't too optimistic about fixing the controller but I did find the controller schematics for a similar model on the web (e.g. http://absurdengineering.org/librar...Corrected - Sony MP-F75W-01G Floppy Drive.tif).
I managed to superglue two floppy pcb fragments approximately back in place - so I was able to see what traces needed repair. It looked doable and, since I have a microscope, I set about fixing the traces. At this point the patches looked like this:
To replace the missing/damaged caps, I ordered a kit ($1.95 from Console5).
Before any recapping, I was able to test the logic board using a compatible IIci PSU. With no floppy, HD or extra RAM installed, the board booted to a flashing floppy screen. Mouse and cursor were living. There was no boot chime since no speaker was connected but I suspect sound wouldn't have worked because most caps were open-circuit.
When my DigiKey order arrived, I first repaired and recapped the PSU. This involved removing the regulator daughter board to replace a couple of aluminum SMD electrolytics with tantalums. Before reassembly, I verified that it delivered the expected voltages.
Next, I recapped the motherboard, re-installed in the cleaned bottom csse and again tested with the IIci PSU. I had to add expansion RAM to boot macOS and that was problematic .. the SIMMS were reluctant to connect reliably with broken sockets and the extended boot time due to RAM testing was unwelcome. But it chimed and booted perfectly. And soft-power worked.
I then installed the recapped PSU and that also booted. Yeah! But it powered up immediately and soft-power wasn't working. D'oh!
It turns out this is another common failure for the IIsi PSU due to residual cap electrolyte. I disassembled again and first bathed the main board in IPA. Much brown yuck came off but the problem remained. With PWR unasserted, I was measuring close to zero on pin 7 of the regulator IC (M71977) on the daughterboard which enables regulation. Others have reported that electrolyte can accumulate under this chip. I'd cleaned the daughterboard but I hadn't removed this IC to look under it. Ho hum! I once again desoldered the daughterboard .. luckily I have a Hakko sucker! .. and pulled the regulator IC. This is what I saw:
Gag. A ground trace was exposed and I estimate that a few kilo-ohms of leakage from pin 7 to ground would be enough to hold the regulation enabled.
With that cleaned and re-masked, and IC resoldered, and daughterboard refitted, and PSU re-assembled and installed .. soft-power works.
So I returned to the floppy drive... I beeped out the traces around the damage but it just didn't make sense. So I removed the motor cover and this is what I saw:
Yikes. It seems that the destruction stopped just before the motor coil was hit. I patched these gouges and now everything checked out.
I first hooked the floppy drive to my LCIII which had a similar fully functional unit. I popped in a known readable disc and after much seeking, the drive read but very slowly. Odd. When I tried writing (inititiazing) a disk it seemed to start Ok albeit slowly. But eventually it failed with an disk-locked error. So, signs oi life, but weirdness.
Then I checked the detector switches. The disk-lock (SW1) and disk-density (SW3) switches were good but the carriage-down (aka disk-loaded) switch (SW2) was stuck open. On reflection, this was understandable since this switch was on the edge of the destruction. There appeared to no chance of getting a new replacement and I had no potential donor device, so I decided to sacrifice the write-protect feature (which I never use) and move this switch over. The disk-lock switch has a longer plunger that the carriage-down switch but I trimmed it down about 2mm. With this change, and with the SW1 wired closed, the drive is now fine.
The expansion RAM (BankB) SIMM sockets were not in good shape. One socket in particular had broken plastic and bent or missing metal clips. I had one spare socket on an LCII donor board so I extracted that and replaced the worst socket. 64MB of expansion RMM is really too much for the IIsi. Without a custon ROM, the boot-time memory test takes a age. 16MB is a more reasonable amount for this 20MHz machine.
At this point, the machine, equipped temporarily with a spare BlueSCSI2, is haapy to boot macOS up to 7.6.1 and to run NetBSD. But the mystery network card isn't recogized. The BlueSCSI2 has DaynaPort wifi - so I have networking, provided I remove the top case to let the EMI out.
So I still don't have a definitive identification for this PDS card:
It reports itself as having vendor ID 4010 which seems to be Sonic Systems, Inc. Back in the day, they did make PDS ethernet adapters for the IIsi but I can't find any drivers out there. There's sonic.etherlan.6.1.hqx file that unpacks to a Sonic installation disk but this fails to provide anything for this card. There are some other vaguely Sonic files on the web but they appear to be vacuous or corrupt. Please let me know if you can identify this card or point me to software.
