Chasing down Lisa Error 52 (COP, I/O Board, etc)

[pl212]

I have a Lisa 2 that worked for many years, but has recently started giving an Error 52 -- and also immediately starting up when powered, instead of waiting for the power button. (Interestingly that light remains unlit despite the machine turning on.)

Here's what I've tried so far:

1) Replace COP chip with (allegedly) known-good chip.
2) Replace I/O Board with a different (allegedly) known-good board.
3) Swap another COP chip to both the original and replacement board, for a total of three COP chips tried.
4) Swap in an (allegedly) known-good CPU board
5) Test continuity between the internal keyboard connector and the 1/4" keyboard plug.
6) Made sure all contacts are clean and undamaged

All of these result in the exact same error, 52, and the same instant-on behavior.

Spelunking around for Error 52 shows these interesting discussions:

• https://lisalist2.com/lisalist1/0569.html (Swapping CPU boards fixed it.)
• Other ideas from that thread: bad 6522 for the keyboard?
• https://yesterbits.com/2012/04/25/error-52-and-mystery-of-the-missing-cops-chip/ (Swapping COP chip fixed it.)
• https://68kmla.org/bb/index.php?thr...-help-troubleshooting.7248/page-2#post-412231 (Fixed by repairing a bad trace)
• (Another user in the same thread also fixed by repairing a bad trace)
• https://68kmla.org/bb/index.php?threads/apple-lisa-findings.35503/ <- User found an out-of-spec power supply was causing the COP chip to malfunction.

At this point my theories are:

• Problem in the actual motherboard backplane that is blocking communication.
• Problem in the female jack where the keyboard plugs in, or in that larger module.
• Failing 1.8 power supply is no longer supplying correct voltage (see last link below)

I think at this point my next move is to recap the power supply, just to eliminate that from the equation, but would appreciate any other ideas...

 

[fri0701]

Has your Lisa suffered any corrosion damage from batteries over the years, even minor?
I've had similar problems (not sure of the exact error code) in the past, and it all boiled down to bad traces on the motherboard backplane. I jumped many bad traces (on the underside) and finally got things working again.

Another thing you might want to check are the female connectors inside the case that the card cage plugs into. I've worked on a Lisa that actually had hidden corrosion damage *inside* that connector - severe enough that some pins were loose and later fell out entirely! Even slight corrosion might cause the pins to not spring properly into place. Just be careful messing with it too much - the pins are fragile and poking at them can cause them to break.

The 1.8A PSU is generally pretty reliable in my experience. If you do decide to recap it, I'd urge you to read my thread on PSU recapping, especially this post about choosing capacitors with the proper ripple current rating: https://68kmla.org/bb/index.php?threads/lisa-ac-interference-coupling.35554/post-384411

Good luck!

 

[pl212]

An update -- fully recapped the 1.8A PSU, mostly to get rid of the risk from the RIFA's. I've measured voltages on the Widget power connector and can see a nearly-perfect 12 and 5 volts. (Except for pin 3, but 5v doesn't seem to be necessary there).

I've created a page on TechTangents caps.wiki site just to document what I found... someone else on that site has started updating it with Mouser part numbers for replacements.

I've measured continuity through the "motherboard" at the bottom, and can confirm the signal traces from where I can reach the fingers to the back of the board. (Of course, there could be corrosion deeper inside the connector, but this board seems pretty clean.) Just to sure, I cleaned the "motherboard" ultrasonically, changed the caps on the I/O board, and inspected/cleaned all the socketed chips on the I/O board.

Still faced with Error 52 however. I'm beginning to think I should look for a spare "motherboard" on eBay, as that's literally the only component that isn't either known-good or that I don't have two of...

I suppose the other move would be to measure voltages at the COPS chip, etc. Kind of a pain to do that without a harness however...

Thanks for the tips on the motherboard traces and the internal case connectors, will take a look at that!

 

[pl212]

Realized the I/O board is facing the right way to do easy testing with the back cover removed and interlock defeated

I've tested each of the I/O boards I have, and both seems to be showing the correct voltages on the COP chip:

• Pin 1: Ground
• Pin 9: 4.86 volts
• Pin 14: 4.86 volts

I think this means voltages are OK, and it's time to start testing signals somewhere (IO Board, CPU Board, both?)

 

[pl212]

I've worked with another Lisa owner to narrow the problem down to... the bare metal chassis itself! More specifically, there must be a bad connection somewhere around the female port connectors that receive the motherboard assembly, and/or the power switch, as all of my components (motherboard assembly, power supply, keyboard) work when in this other Lisa chassis. Interestingly my power switch bulb does not light either, even after replacing the bulb, which is another clue that the problem must be somewhere there.

 

[pl212]

The ultimate answer here was that the momentary push-button power switch had failed - metal fatigue, judging from the loose parts I found inside the switch. This switch is supposed to momentarily interrupt a connection to the I/O board, which then triggers soft power-on. I believe my switch had failed to be open all the time, causing the Lisa to think I was holding the button down all the time.



This explains why the machine immediately powered on, with the lamp lit immediately - and why it then complained about an I/O Error 52.

The same exact switch, now branded Honeywell, is available for the princely sum of $40. The model number is AML21CBA3AA, and it’s important to get exactly that specific one. There are dozens of variants - this one has the printed circuit / push-in style terminals. Interestingly Apple themselves used the larger terminal variant, designed for soldering wires to the end - and apparently trimmed the terminals down to fit though-hole. Not something I would advise you to try.

The other thing to watch out for is to not damage the through-hole traces, both when desoldering the old switch, and when installing the new one. They’re pretty delicate.