Macintosh 128/512 board resurrection - checkerboard pattern

[PacificState]

Having figured out the first of two Macintosh 128 boards I acquired (https://68kmla.org/bb/index.php?threads/macintosh-128k-board-with-custom-512kb-upgrade.48789/), I'm now trying to repair the second. I'll add pictures shortly, but it appears to be an 128KB board upgraded to 512KB, which unfortunately boots to a checkerboard pattern.

Few things:

- For testing purposes, I'm running this board with the DosFox ATX + VGA adapter, rather than installing it in the Mac itself. I think the checkerboard pattern can be induced by a low 5V rail, but the first board works fine with this setup.

- The board came with a note saying 'BAD RAM AND ROM SOCKETS'. As I have a desoldering gun and believe bad ROMs can result in the checkerboard, I've swapped out the sockets with new ones and also swapped in the ROMs from the first board - all without effect.

Even if the RAM is bad, I don't know it would cause this problem - from what I can see, this is a symptom of the ROM not getting to run and clearing the screen. I'm therefore wondering if it's either the 68000 or the 6522. I have a scope, so I can start probing lines tomorrow - are there any particular things to check, though? Thanks!

 

[Builder68]

Even if the RAM is bad, I don't know it would cause this problem - from what I can see, this is a symptom of the ROM not getting to run and clearing the screen. I'm therefore wondering if it's either the 68000 or the 6522. I have a scope, so I can start probing lines tomorrow - are there any particular things to check, though? Thanks!

sorry, didn't read that you already did all that.

 

[PacificState]

Hmmm, so it actually looks like the CPU reset line never goes high - it's held at about 0.4V, without ever nearing 5V. If I do hit the reset switch, it drops temporarily to ground but returns to its prior level. Furthermore, the scope shows that there's never activity on the data or address buses, so I assume the 68000 is basically never coming out of reset (the clock, however, is good).

I'm studying the reset circuit around the LM311, and:

- The diode, resistors, voltages, and caps all seem broadly correct
- There's ~4.9V on one side of R9 (with R9 measuring in-circuit as 1K)

On that basis, I cut pins 2 and 7 of the LM311 in the hope that I'd get ~5V (thinking the LM311 was defective) and could manually pull the line down, but it doesn't seem to make any difference.

As far as I can see, with those pins cut, the path is 5V -> R9 (1K) -> the reset and halt pins of the CPU, plus the reset pins of the 6522 and IWM. I don't mind socketing the LM311, but I'm averse to more track and pin cutting otherwise - any suggestions as to what could be going on here? Thanks!

 

[PacificState]

And... after desoldering the 6522 and IWM, the 68000 was dead :s

I managed to socket them all without pulling a pin, and now I get a sad Mac and an 020100 error.

 

[PacificState]

Fixed after looking up the error codes (https://tinkerdifferent.com/resources/sad-mac-error-codes.82/ - it was precisely correct).

Next problem: I have a MacSnap 512KB board which is throwing a similar RAM error when installed.

 

[PacificState]

To close this thread out:

- I couldn't relate any of the error codes to the MacSnap RAM layout (it's a 524E board, for reference)
- Piggybacking a new chip over each existing IC in turn didn't get it booting - but I did notice that two ICs in particular caused vertical black lines to be displayed over the Mac icon
- I then piggybacked *two* new RAM chips over of these ICs, and the Mac got to the floppy screen - although it crashed mid-way through boot
- Desoldering these chips from the board and replacing them with the new RAM got things fully working, and the Mac board now has 1MB.

Posting out loud here helped, and maybe someone else who hits the same problem will find this thread.