@Svangool, now that is an interesting thread! Thanks for linking that for me. And yes the symptom is kind of the same (its even the same GPIB address) except with mine, the communication between the analog and digital boards is still working. The failure mode on mine is that initially the unit works and then after a few minutes, it freezes and/or starts displaying a scrambled display and then freezes (meaning that it does not change any more and does not respond to keypress input). On restart it often gets to the HPIB adress and stops there. You have to let it sit "off" for 10-15 minutes and then it will start up again.
On that subject though, I did consider that (the optocoupler) could be the point of failure and was probing the output with a scope, but not really seeing the serial bus communication I expected (all I saw was low level noise). Yet the meter was working, at least for a little while. So that is a problem I am interested in also, why I couldn't read that traffic with a 'scope. (I was probing on the Digital side of the connector using the Digital ground, which is chassis ground as well. Is it basically a low impedance node at that point?
Meanwhile, I ordered a replacement uP and while I am waiting for it to arrive (actually I ordered two slightly different parts from two sellers) I decided to install a 40 pin socket, so that if things didnt work out I would not have to keep stressing the PWB with rework. And obviously, I put the processor back in to test the socket, and see if being elevated a bit off the PWB made a difference. And it did. The meter worked for about 10 minutes (at least 2x longer than before), tested voltage, current, resistance, range changes, etc. But then froze up like before. So I think it is the uP but these things are tricky and as Kleinstein pointed out (Thank you Kleinstein), the point of failure may be nearby the uP instead. Fingers crossed, it looks like Friday I should have a replacement. BTW these things (Motorola 68B09 type) are getting rather hard to find, which is one reason why I bought a couple. Also that one vendor was offering the "industrial" version (MC68B09CP) which has a wider operating range, and these were actually cheaper than the original MC68B09P part from other vendors. So I will try them all out. (The readily available E suffix is not compatible with this circuit)