Print

[fri0701]

Hi everyone,
What approach do you all take when debugging a problem with the CPU board in a Lisa? I have two that have issues, but with everything plugged in, it's impossible to reach the board at all to poke around with a multimeter or oscilloscope. I suppose I could connect a lead to a leg of a chip, push the whole card stack in, and read it that way, but that seems pretty cumbersome and time-consuming. The CPU board's chips also face into the computer, which makes the whole thing harder.

Does anyone have a trick to make this easier?

[stepleton]

warmech has designed some extender PCBs that let you move the card cage and/or individual boards out to somewhere more convenient:

https://github.com/warmech/lisa-diagnostic-hardware
https://68kmla.org/bb/index.php?threads/reproducing-the-lisa-2-card-stack-the-cpu-card.41766/#post-465640

Scanning the 68kmla thread, however, I'm not certain the extenders have ever been tried.

These days I'm working on a PERQ workstation from about the same time as the Lisa. The machine has long, TTL-packed boards that slide into deep slots, so you have to settle for this nonsense: https://photos.app.goo.gl/2dxEbZ6UjazQHjnc8

The blue threads are slender wire-wrap wires tagged onto chip legs. It'd work in the Lisa as well but would no doubt be a hassle.

[patrick]

There is a story that the Lisa runs at a boring 5 MHz instead of the full 8 MHz of the 68000 so that they could use bus extenders during development and connect the logic analyzer more easily. Maybe there is something to it. I would just try it.

[warmech]

Small update on those riser card files...

They had a fault that I've finally fixed but haven't had a chance to upload the new gerbers; prompted by this post, I just did. I ordered a new batch and tested them and they appear to work correctly now. With the new ones, though, there's also a second set of gerbers for a spacer insert for the edge connectors. I never had issues with the I/O card making a good connection, but the CPU card wasn't making the best contact with the edge connector on the riser due to the guides the card slides down on the card cage preventing full insertion. The spacer lets you extend the connector on the riser by a PCB width (~1.6mm?), which allows for a perfect fit.

Also, I need to make a set of 3D-printable supports for the cards once they're out of the cage; I've just been using random crap to support my cards, so I should probably do something about that as well, lol.

I'll PM you fri0701, but I have a couple more of these riser cards if anyone is interested. Also, I'll post a video of these things running this evening in order to provide proof of life, lol.

There is a story that the Lisa runs at a boring 5 MHz instead of the full 8 MHz of the 68000 so that they could use bus extenders during development and connect the logic analyzer more easily. Maybe there is something to it. I would just try it.

I had a conversation with Alex along similar lines - the shape and layout of the card cage seems eerily setup for this exact kind of tool. The dimensions for everything almost exactly line up to allow these risers to fit perfectly to the point that we were wondering if Apple didn't design the cage to allow them to use something similar to test/diagnose Lisa hardware.

[Lisa2]

The way I always did this was run the whole card cage outside of the Lisa case with an external monitor.  I  used a spare set of the power and data harnesses ( and spare KB/power switch assembly) with the Lisa PS outside of the case.  This was only possible because I have a Lisa compatible external monitor.   In the past, most folks don't have an external display that works with Lisa, but now anyone should be able to do this using the RBGtoHDMI converter.  https://github.com/hoglet67/RGBtoHDMI

Rick

[sigma7]

Does anyone have a trick to make this easier?

I haven't used this technique, but I speculate it will work for some circumstances:

Without the I/O Board installed, the PSU will (usually) turn on when power is connected, and the CPU board will run through the initial part of the ROM's diagnostics until the point where it finds there is no I/O Board.

So, depending on what kind of problem you are troubleshooting, you may be able to figure it out by probing the back of the CPU Board with no I/O Board installed.

HTH

[AlexTheCat123]

Yep, now that the fault has been fixed, warmech's risers work great! So if I were you, I'd put a couple of those together. They really make working on your CPU board an absolute breeze!