Apple Lisa findings

[max1227]

Hello everyone,

I hope during this period of global uncertainty and crisis that this message finds you well.

I have made some interesting progress with the Lisa that may be of interest to some. I am

also posting to request some help in the hope of getting it up and running fully. 
 

So recently I dug out the Lisa and decided to attempt to diagnose further the issue I was having with it. To cut a long story relatively short it basically turned on whenever power was applied (ie soft power switch not pressed) and error out at I/O error 52. I noticed that if I switched it on let it error, switched it off, wait 10-15 seconds and switched it back on, 100% of the time it would then both correctly with no error.  
 

At the time, a few days down the line I tested the power supply and found that all the rails were as expected. However I noticed that I could remove the back and front panels triggering the interlock switches and the Lisa would not turn off. The switches were working but the Lisa was ‘ignoring’ them. With the machine booted I got it running MacWorks and when I selected the shut down from the drop down menu the system will eject the disk, dim the screen and then nothing. It will permanently hang on a dimmed screen.

So,
 

I then decided to switch a power supply from another Lisa into this one. Now the I/O error permanently went away and the interlock switches were functioning again as expected. Just to test I put the power supply from the 1st Lisa into the 2nd and it caused that Lisa to develop I/O error 52. Therefore there is something wrong with that power supply that is preventing the COP421 chip from working properly. I then tried with MacWorks and asked the system to shut down with the swapped power supply and it did! It fully shutdown so I thought the power button may work again but upon pressing it the computer would not start back up.

So this is the stage I am up to now. Does anyone have any ideas as to what could be causing this? I really hope it isn’t a bad COP421 chip as I know these are incredibly difficult to locate.

Many thanks,

 

[fri0701]

Hi Max,

Are you using the 1.2A or the 1.8A PSU?

I find it very odd that the back interlock switch was not working as expected, since it's internal to the power supply.

 

[max1227]

I believe it’s an equivalent 1.2A power supply. It’s actually a UK 0.7A power supply. Which as previously mentioned I believe outputs a similar power to the American 1.2A power supply.

 

[sunder]

So recently I dug out the Lisa and decided to attempt to diagnose further the issue I was having with it. To cut a long story relatively short it basically turned on whenever power was applied (ie soft power switch not pressed) and error out at I/O error 52. I noticed that if I switched it on let it error, switched it off, wait 10-15 seconds and switched it back on, 100% of the time it would then both correctly with no error.  

This smells like a Lisa 2/5 battery leak issue - check/clean and corrosion and resolder/reflow the I/O board. I've had this myself, in my case, I had a spare I/O board.

It's not necessarily the COP421 itself, though some of the pins might be corroded, or the traces around it might be corroded.

I'd also check the capacitors on the power supply just for good measure. If you see the original large square/rectangular yellow ones, likely you should replace those, as well as any deformed can ones.

Replacing all the big caps is relatively inexpensive and worth doing whether or not they're in spec as you'll likely extend the life of the PS by a dozen years at least.

 

[max1227]

Thank you for your advice and I will indeed check out the I/O board although when I did the inspection a few years back it looked fine. The corrosion whilst bad isn’t as bad as some and many of the traces I tested around the COP421 chip are amazingly still working. 
 

Before I do that I want to swap some parts out between my Lisa’s to attempt to verify where the problem is located. For example swapping out the PS enabled me to narrow that the interlock problem and I/O error 52 was not the I/O board but the PS. I want to put a Lisa 2/10 card cage into the 2/5 but am aware that there may be some incompatibilities between them, such as the Lite Adapter. If I disconnect the Lite Adapter does anyone know if a Lisa 2/10 card cage is compatible?

 

[fri0701]

If you unplug everything in the floppy card cage (lite adapter, drives, etc.), you should be able to swap motherboard assemblies between the 2/5 and the 2/10 with no issue.

 

[max1227]

Thanks for your message. So I swapped everything out and identified something interesting. With a working 2/10 board in the computer it would no longer immediately switch on. That points towards the COP421 chip being to blame on the 2/5. So I tried to start up the 2/10 board but it wouldn’t turn on. This card cage works but the power button wasn’t doing a thing. At the moment I think it could be a trace on the I/O board and the power switch is faulty as well. When the 2/5 board is inserted and the machine powers on, the light doesn’t illuminate and the keyboard isn’t recognised either. 

 

[sunder]

Thank you for your advice and I will indeed check out the I/O board although when I did the inspection a few years back it looked fine. The corrosion whilst bad isn’t as bad as some and many of the traces I tested around the COP421 chip are amazingly still working. 
 

Before I do that I want to swap some parts out between my Lisa’s to attempt to verify where the problem is located. For example swapping out the PS enabled me to narrow that the interlock problem and I/O error 52 was not the I/O board but the PS. I want to put a Lisa 2/10 card cage into the 2/5 but am aware that there may be some incompatibilities between them, such as the Lite Adapter. If I disconnect the Lite Adapter does anyone know if a Lisa 2/10 card cage is compatible?

The bad news about the corrosion is that it doesn't stop. You should clean the I/O board with deionized water, electronics cleaner and maybe a bit of white distilled vinegar. but it's likely that the corrosion goes inside the traces and keeps eating away very slowly year after year. Reflowing solder should stop it at the solder joints, but not once it gets inside the traces. So it might work today but a month from now it might fail, etc. If that has happened, you'll have to jumper whatever traces are corroded. Use a VOM in continuity/Ohm mode (so it beeps on 0ohms) to test each trace around the battery area. If you get more than a few Ohms, time to jumper that trace or at least reflow the solder joints (they might have cracked too, not just corroded).

There could also be junk alkaline stuff from the batteries under chip legs that might short out some of the pins (likely this is what's re-powering on the Lisa.) or a disconnected pulldown signal.

If you're able to carefully desolder the COP and any chips with corroded joints, clean underneath and solder sockets in their place it would help. This isn't easy and if you don't know what you're doing you can damage the board - and the board could already be damaged from the corrosion. Sorry, yeah, it's bad news all the way down. Tread carefully.

The boards can be repaired and as long as the legs don't break off the COP chip it should be fine. You can tin a soldering iron and very quickly heat each leg of the COP chip - don't over do it as you don't want to melt it off the plastic case of the COP or have the leg fall off. Only do this if you see the legs are corroded and use plenty of flux. And yes, only use LEADED electronics solder.

The 2/10 I/O board is very different but can be swapped. You can swap power supplies without issue. But don't switch the 2/10 motherboard with the 2/5's - the cables inside the chassis are not the same. You can short circuit the power supply if you use a 2/5 motherboard with a 2/10 chassis that has a Widget drive in it. While you're at it look at the edge connectors on the motherboard, it's possible those have corrosion as well. I have at least one 2/5 mobo that has corroded pins on one of the edge connectors - luckily these can be found on eBay for ~$20 and it's the same connector for both the I/O and CPU boards, just offset horizontally, but they're a huge pain to desolder properly. I'd recommend practicing soldering/desoldering on junk hardware first.

The 2/10 I/O board can work in a 2/5 chassis and 2/5 motherboard, but not sure how well. I think I had one this way, but for sure, if there's a Widget internally in the chassis it will short. I don't recall having issues with the Lite adapter in such a setup. All this is from wetware memory, so I might have forgotten something or misremembered.

Best of luck. It's not easy, but go slow and take your time and apply lots of patience. It's worth doing well the first.

 

[max1227]

Thanks again for your help. The motherboard and CPU card have already been treated and cleaned with white vinegar and then IPA to wash off any vinegar left behind. I did that about 4 years ago. So, since the functioning state hasn't changed since then I would like to think the vinegar has neutralised the residue and the IPA washed the contents away. I will test the traces I just wish to narrow down the problem beforehand to make my job a little easier.

I believe all of my COP's chips are socketed so I should be able to use IC removers to re-socket it if necessary.

I did the switch after reading a Lisa FAQ page which detailed that the boards aren't interchangeable, unless you disconnect the Lite Adapter and the Widget drive. In the 2/5, this is because the Lite Adapter is present and has two functions; to modulate the speed of the drive and to adapt the twiggy interface to the Sony 400K microdiskette interface. In the 2/10 the widget is often present which as you say will short the PS as this drive is not present in the 2/5.  So I disconnected the Lite Adapter and used the card cage from the 2/10 in the 2/5, but nothing happened.

 

[sunder]

Nothing happened in what way? didn't turn on? or something else?

 

[fri0701]

Hi Max,

I have an idea about your power supply with the non-functioning interlocks.

If the schematics for the UK 0.7A and the US 1.2A PSUs are the same, I would recommend checking transistor Q5 for damage (shown here in the bottom left): 

http://www.1000bit.it/support/schema/apple/lisa/Lisa_Power_Supply.pdf

If it's broken or has been modified (a leg has been cut, or something), it could prevent the interlock system from properly functioning.

This might also explain why the Lisa starts without the power button being pressed when used with that power supply. 

From the Lisa hardware reference manual (https://archive.org/stream/Lisa_Hardware_Reference_Manual_1983-05-31_Apple/Lisa_Hardware_Reference_Manual_1983-05-31_Apple_djvu.txt):

The power supply is switched on by applying a logic 
1 level to the base of transistor Q5, causing 
transistor Q5 to conduct, which grounds pin 1 of 
U3. The power supply switches on and stays on until 
this level is removed. 

A low level on the base of Q5 keeps the power supply 
in the OFF condition. This can be achieved by 
either driving it low via the software on-off or by 
grounding the safety interlock switch.

 

[max1227]

Nothing happened in what way? didn't turn on? or something else?

It just won’t event attempt to power on. When the button is pressed there is no life. No audible clicks when the CPU and memory etc are being tested. No high pitch whine for the fly back. It just doesn’t do anything with a known working card cage. 

 

[max1227]

Hi Max,

I have an idea about your power supply with the non-functioning interlocks.

If the schematics for the UK 0.7A and the US 1.2A PSUs are the same, I would recommend checking transistor Q5 for damage (shown here in the bottom left): 

http://www.1000bit.it/support/schema/apple/lisa/Lisa_Power_Supply.pdf

If it's broken or has been modified (a leg has been cut, or something), it could prevent the interlock system from properly functioning.

This might also explain why the Lisa starts without the power button being pressed when used with that power supply. 

From the Lisa hardware reference manual (https://archive.org/stream/Lisa_Hardware_Reference_Manual_1983-05-31_Apple/Lisa_Hardware_Reference_Manual_1983-05-31_Apple_djvu.txt):

The power supply is switched on by applying a logic
1 level to the base of transistor Q5, causing
transistor Q5 to conduct, which grounds pin 1 of
U3. The power supply switches on and stays on until
this level is removed.

A low level on the base of Q5 keeps the power supply
in the OFF condition. This can be achieved by
either driving it low via the software on-off or by
grounding the safety interlock switch.

That sounds like a fantastic suggestion! Thank you so much for looking into this. I shall check this in the morning! 

 

[max1227]

So I checked the transistor and no legs clipped or anything visibly wrong. Will check to see if it is shorting and if that is the source of the problem. 
 

Meanwhile I’ve been investigating the widget drive I have with the Lisa and have attached an audio recording of the drive when in use. It doesn’t make the characteristic squeak noise when it seeks but rather this juddering noise. Could this be a sign of a suck actuator arm? Has anyone ever heard a similar sounding drive when it’s attempting to seek. The brake does release beforehand.

View attachment Apple Lisa Widget Drive.m4a

 

[sunder]

It sounds like the spindle motor is moving, so that's good, so the two most common failure modes (spindle motor freezing due to lack of use/hardening lubricant) and head actuator brake are not there.

It could be an issue with the half moon board on the bottom of the drive, or the top controller board, you might be able to find replacements on ebay or from Vintage Micros, but some may not work due to different versions, etc.

You'd have to notice the LED lights on the drive as well and the timing of what happens when as these can give clues, but I'm not familiar with the exact details of this.

I'd also use either BLU or NeoWidEx to try to get status from the drive. I'd also split this into a new topic, and also ask about the Widget separately on LisaList2.com as more knowledgeable folks might be able to chime in.

A working drive should self identify with the string WIDGET-10, if you don't see that, it could be a controller error on the top board. A non working drive might provide status to NeoWidex, so NeoWidEx should provide more details if the drive works well enough to report status, for example, if the drive has a lot of bad sectors the spare table will be full and it will help you low level format it.

see: https://github.com/stepleton/NeoWidEx

If something else is wrong at the hardware level, you'll need more work to help identify it.

Please post the results of whatever you find either here as over time I'd like to update the LisaFAQ with as much details about these kinds of repairs (not just the widget ones).

 

[max1227]

It sounds like the spindle motor is moving, so that's good, so the two most common failure modes (spindle motor freezing due to lack of use/hardening lubricant) and head actuator brake are not there.

It could be an issue with the half moon board on the bottom of the drive, or the top controller board, you might be able to find replacements on ebay or from Vintage Micros, but some may not work due to different versions, etc.

You'd have to notice the LED lights on the drive as well and the timing of what happens when as these can give clues, but I'm not familiar with the exact details of this.

I'd also use either BLU or NeoWidEx to try to get status from the drive. I'd also split this into a new topic, and also ask about the Widget separately on LisaList2.com as more knowledgeable folks might be able to chime in.

A working drive should self identify with the string WIDGET-10, if you don't see that, it could be a controller error on the top board. A non working drive might provide status to NeoWidex, so NeoWidEx should provide more details if the drive works well enough to report status, for example, if the drive has a lot of bad sectors the spare table will be full and it will help you low level format it.

see: https://github.com/stepleton/NeoWidEx

If something else is wrong at the hardware level, you'll need more work to help identify it.

Please post the results of whatever you find either here as over time I'd like to update the LisaFAQ with as much details about these kinds of repairs (not just the widget ones).

Thanks for your response. Unfortunately I am not doing well on any front. The Lisa is still having issues with powering on. Although I have narrowed down the issue somewhere on the I/O board. So I guess that’s progress.

As for the widget. I genuinely have no clue where to even start. It’s such a shame because it sounds fantastic. The drive runs very quietly as if it has seen so little use. It seems every widget drive that is broken has its own unique problem which makes it very hard to track down. All I can really say is some of the chips on the top board do get quite warm to the touch after not very long. 
 

I decided since I wasn’t having too much luck with the Lisa or the Widget to take a look at the ProFile. It does have stuck bearings but a few turns loosened it up. It however has a more noticeable fault. It’s exactly the same as this YouTube video. This is not my drive but the movement of the stepper is identical. It cannot pass the self test but just goes forward and backward and stops. Then after about 1 minute it will attempt again.
 

I tried loading it into BLU and much like the widget BLU doesn’t see it.

And similarly to the Widget, the Z8 chip gets warm quite quickly.