LisaList2

For completeness on this subject, the Sapient CPU recreation has among it's many enhancements has a built-in ROM Switcher. Using 256K EPROMS and a special VideoProm that supports four Lisa ROM sets (versions H,D, C, and 3A ).  The ROMs are selected using some switches on the board but these switches can routed externally also.  A photo of this board is attached, (please ignore the XLerator in the photo)

I made a very rough video of the operation here:
https://youtu.be/56KtI2uQaWg

Note: if switching between the older (H, D, C) and screen mod ROMs (3A) the screen pots really need to be adjusted. The original ROM switcher instructions recommend a compromise of settings between the two setups.   Neither the original ROM switcher nor the Sapient board switch in the yoke transformer for the screen mod. 
Rick

Good to know! I had no idea that there was anything on Bitsavers about these things.

Mine exhibited the same symptom when I simply needed to replace the two ROM chips on the CPU board.

Edit: The video board could also be a culprit!

There is more RomSwitcher info on bitsavers:

http://www.bitsavers.org/pdf/apple/lisa/macintosh_xl/

Someone just listed a Lisa on eBay with a ROM switcher installed that lets you change between H and 3A ROMs (as well as the MacWorks screen mod) with the flip of a switch. I've heard about these switchers, but I've never actually seen pictures of one, so I figured I'd post this here in case anyone else wants to take a look!

As a guess, I wonder if that would have to do with the contrast latch being on the I/O board. Maybe tying the CONT signal to some voltage would stabilise it.

Yeah, I was wondering the exact same thing too! It just seemed weird that the contrast fluctuations were really consistent/predictable and not the random behavior that you'd probably expect from a floating contrast signal. But who knows, maybe there's some explanation for that.

As a guess, I wonder if that would have to do with the contrast latch being on the I/O board. Maybe tying the CONT signal to some voltage would stabilise it.

I've tried that for troubleshooting purposes too and my Lisa turns right on and shows the error 41 as well. One other thing to add is that the screen flickers in a weird way too, at least on my Lisa.

Update: a commenter on the thread observes that running a Lisa without an I/O board causes it to power up immediately and show an error 41 (if things get that far along). I take this to mean that we can proceed with our probing plan!

I'm "helping" someone over on 68kmla with their Lisa --- signs point to a problem on the CPU board, perhaps in the MMU, since the problem happens very soon after power-up or reset. We'd like to get in and probe a bit, which of course is pretty hard to do without extenders or other clever tools.

Since the boot process failure happens so soon (or maybe the Lisa isn't even running the ROM code at all), the thought is that the I/O board isn't even really a factor yet --- the CPU might be giving up after trying to write to the MMU registers and failing. This means that we could learn a lot by probing if we could only get the Lisa running without the I/O board in the computer at all.

As discussed in this recent thread, you can force the Lisa on by shorting +5V STBY to ON. The discussion is about how to do that with the I/O board in, but those signals are on the I/O board connector, and you could easily short them while the I/O board is out of the computer.

Is this safe? If so, would the Lisa at least try to start executing the boot ROM in this state? (For example, I was worried that you might need to pull up \NMI, but it seems that both the I/O board and the CPU board have \NMI pull-ups.)