Macintosh IIsi Success Story

rikerjoe

Tinkerer
Oct 31, 2021
141
216
43
I recently came into possession of a Macintosh IIsi for $30 from Yahoo! Japan Auction via the Buyee proxy service, and I thought it would make for an excellent restoration project. As is often common for items on Yahoo! Japan Auction, the IIsi was marked as “junk”. I figured it was worth a shot.

Thankfully, the IIsi arrived quickly (it took about a week to ship to the US) and intact. I paid more in shipping costs than for the IIsi itself!

55D43FF8-70BA-4116-90CC-FEA9BF89DB5E_1_105_c.jpeg


Overall, the IIsi exterior was in very good shape, with yellowing typical of machines from the era.

EAC103EE-DE59-4460-B757-1740FD819551_1_105_c.jpeg


I did a complete disassembly after watching the video on the Jason’s Macintosh Museum YouTube channel (wish he were still making videos) and pulled out the logic board. Overall, the board seemed fine at a first glance, with a layer of dust and the telltale signs of minor capacitor leakage. It came with four 4MB SIMMs installed, which I later replaced with four 16MB SIMMs.

56B34C51-7811-4D3B-B94B-5E114D993089_1_105_c.jpeg


Hmmm, what is that brown crusty stuff around the onboard memory, near the PSU connector? (Foreshadowing - this is a sign of where I spent most of my time in this restoration.)

B6AEAFE7-D6D9-4ED1-91AF-9F43C4CD48EF_1_105_c.jpeg


Well, those of you familiar with the IIsi know that it suffers from capacitor leakage in the PSU. So, it was off to watch the IIsi PSU restoration project on the @Branchus YouTube channel, then undertake the disassembly and restoration myself.

Yuk. The bottom side of the PSU was coated in nasty brown cap goo, and the PSU connector was green with corrosion. I have never seen the like previously.

2EE982BA-07DB-49A3-A5B3-DF1215F7ECC6_1_105_c.jpeg


Here’s the result after soaking the connector in a vinegar bath and attacking the PSU with 99% IPA and cotton swabs. Much better.

2EF738D2-4C76-413F-9CAE-29F2035A873D_1_105_c.jpeg


After installing the replacement caps on the PSU, it was time for a smoke test. The PSU didn’t blow up, so I took that as a good sign.

1034210E-A889-4AA2-89DA-2E89D3B5DD3D_1_105_c.jpeg


Next, it was time to tackle the logic board. Removing the old caps, cleaning the board with IPA, and installing the replacement caps, I ended up with a nice looking board. You can see I favor the black C body tantalum caps, which fit the pads nicely.

IMG_1151.jpeg


While I was performing the restorations of the PSU and logic board, I soaked the case in warm soapy water to remove the grime and PSU cap goo that had leaked on the bottom case. After retrobriting, the case turned out quite nice.

30A388BF-ADE3-461D-B404-2A815555B4A7_1_105_c.jpeg


Finally, it was time for reassembly and testing. I powered on the IIsi, and it chimed although the speaker was rather faint and garbled, and booted successfully to the blinking question mark floppy. Nice. I was bothered by the garbled speaker, and sure enough when I went to test the floppy drive, it wouldn’t eject. (I did a restoration of the floppy drive and replaced the eject gear, so I knew that wasn’t the problem.). I recalled the last time I had problems with a floppy drive, and it was due to a problem with under voltage on the +12V rail. To test my theory, I tried to measure the voltages via the external floppy port, and theory was confirmed – I measured millivolts on both the +12V and -12V rails. The +5V rail was fine. Hmmm, did I mess up the PSU recap? I didn’t think I did since it didn’t blow up on me for the smoke test, but I decided to pull the PSU and check my work. I went so far as to remove all the components on the secondary side to clean cap goo underneath the transformers, inductors, resistors, diodes, and remaining caps. After reassembly and test, same thing resulted – the +12V and -12V rails were registering millivolts. I wondered if the problem was not the PSU and instead was a problem on the logic board. I pulled the PSU from my Centris 650 and tried it in the IIsi, and registered the same problem. I tested the recapped IIsi PSU in my Centris 650, and it worked fine, with +12V and -12V measured as expected at the Centris 650 floppy port. The problem was definitely not the recapped IIsi PSU and had to be on the IIsi logic board.

I pulled out the Bomarc schematics and started continuity testing. I wondered if the garbled speaker was a sign of a problem on the +12V and -12V rails, and sure enough the schematics show that the speaker relies on both rails. Working backwards with continuity testing revealed a problem with the PSU connector on the logic board. I should have suspected that, with the brown cap goo on the logic board and green PSU connector! After removal of the PSU connector, it was clear that the +12V line had a broken trace, and the -12V line was hanging on by a thread. I repaired both traces, resintalled the PSU connector, and reflowed the pads on the speaker connector, which had several gouges in it.

04951D14-8A16-487B-B7D2-A2BEEF6E1ECA_1_105_c.jpeg


Now, it was time for a test, and…

CECCEF4E-E6AB-48AB-9A09-6CAD07C142ED_1_105_c.jpeg


…a successful, full-throated chime, boot and floppy drive performance.

One other quick mod I made was to install a BlueSCSI. In case you didn’t know, the modern replacements such as BlueSCSI and ZuluSCSI require power through the berg connector to work on the IIsi. (Ask me how I know.) Easy enough. I also installed a blue LED activity indictor and attached it underneath the power indicator LED with some blue tacky.

3C12F7F4-5B6F-4497-AD29-A5DB3CF92AFE_1_105_c.jpeg


This arrangement works well, as you can see here, comparing when the BlueSCSI activity light is off versus on.

CE1B0D33-9CFA-455B-9185-5F44AC30FAE0_1_105_c.jpeg



6B2B9208-A8BC-4AC8-B7DC-56564F9A2EB1_1_105_c.jpeg


The IIsi came with a Macintosh Display Card and the right-angled adapter, sweet!

327D46B7-7433-4DAA-983C-E23DCD66241C_1_105_c.jpeg


Benchmarks using the on-board graphics versus Macintosh Display Card are revealing.

7C94F567-9046-439D-8DA6-C95804EE1899.png


All in all, this was a fun project. I got to recap my first PSU and do some successful logic board troubleshooting, thanks to the Bomarc schematics. The retrobriting turned out nice, and I figured out I had to supply additional power to the BlueSCSI to make it work. The Macintosh IIsi is not one that I used in the past, but I think it makes for a nice addition to my collection.

And as a teaser, I managed to get my hands on my personal “holy grail” machine - but that is a story that will have to wait for another time.
 
Last edited:

Byte Knight

Tinkerer
Oct 21, 2021
95
74
18
Nice restoration! I recently restored one and had a break in the 12V rail as well. As you probably already know, it's fairly simple to increase the speed to 25 MHz by swapping out a 50 MHz oscillator.
 
  • Like
Reactions: rikerjoe

rikerjoe

Tinkerer
Oct 31, 2021
141
216
43
Yes, I have seen a few references to swapping out the 50 MHz oscillator, so that is on my future tinkering list. I saw your post and was curious about using the fourth connector from the speaker to drive the LED with disk activity. I tried a simple connection from the fourth to third connector but didn't get any disk activity on the LED. I'm not sure if it is because I'm flunking basic electronics circuitry or if the BlueSCSI doesn't send the activity through the SCSI connector. How did you wire from the fourth connector to the LED? Regardless, I'm happy with my solution if a bit janky. Thanks!
 

T-Man

Tinkerer
Oct 30, 2021
31
27
18
Yes....great job. I've done a couple of IIsi and the daughterboard on the power supply slows me to a stop.
 
  • Like
Reactions: rikerjoe

rikerjoe

Tinkerer
Oct 31, 2021
141
216
43
Yeah, I didn’t write about the PSU daughter board in my write up. It brought to mind the board in an Apple 300i caddy loading CD that I recapped. What a cheap, low-quality board! Naturally, I lost a pad when refurbing the daughter board. I thought for a brief second whether the pad repair might have been the source of the problems I suspected with the PSU but instantly dismissed it when I checked the schematics and my pad repair.
 

Byte Knight

Tinkerer
Oct 21, 2021
95
74
18
Yes, I have seen a few references to swapping out the 50 MHz oscillator, so that is on my future tinkering list. I saw your post and was curious about using the fourth connector from the speaker to drive the LED with disk activity. I tried a simple connection from the fourth to third connector but didn't get any disk activity on the LED. I'm not sure if it is because I'm flunking basic electronics circuitry or if the BlueSCSI doesn't send the activity through the SCSI connector. How did you wire from the fourth connector to the LED? Regardless, I'm happy with my solution if a bit janky. Thanks!
The wire actually goes from the fourth connector to the LED. But your solution works too! :)
 
  • Like
Reactions: rikerjoe