exxos's Atari, Amiga & retro forum

stephen_usher wrote: 06 Apr 2023 12:33Bah!

Bah! Indeed

It seams apple products like to keep your trouble shooting skills active. I would start at the 6502 check for clock's halt and stuck line's and work my way back to the ram. Does the unit have Mt brand ram on it ? They always fail

Tom

@rubber_jonnie

Tomswork wrote: 06 Apr 2023 15:18 It seams apple products like to keep your trouble shooting skills active. I would start at the 6502 check for clock's halt and stuck line's and work my way back to the ram. Does the unit have Mt brand ram on it ? They always fail

Tom

@rubber_jonnie

I have tested with another CPU, and the machine passes all Apple diagnostic tests with the 80 column card removed. RAM is OKI BTW, so not really an issue, but I have tested all of it, both in the machine and in a separate RAM tester.

And yes, at the Moment Apples seem to be my nemesis!!

@rubber_jonnie

Hmm the video and system memory are accessed alternatively by the cpu and the video circuit the 80 column card adds bank switching to make things more complex. Adding the 80 column card brings the problem on after 10 minutes or so ?

Tom

So, how does the CPU/video switching work? Is it using a tiny ROM with a truth table for the chip selection?

stephen_usher wrote: 06 Apr 2023 20:50 So, how does the CPU/video switching work? Is it using a tiny ROM with a truth table for the chip selection?

Not 109% sure, I may have to do some heavy reading to try and figure this out.

The 80 column card is little more than additional RAM and logic ICs so everything is already in place in the standard Apple IIe but it requires the extra 64kb.

The Apple IIe diagnostics show the machine as fine, so its something specifically to do with 80 column operations.

I've ordered some freeze spray to see if I can narrow it down to a specific IC, so wait for it to fail, use freeze spray to see if I can get it to recover faster as the current cool down is about the same as the fail time, so 5-10 minutes.

I would also inspect the 60 pin card edge for corrosion for the ram card maybe shorts to. I did pull a 80 colomb card out once the the plastic fram came with it


Tom
@rubber_jonnie

Tomswork wrote: 07 Apr 2023 00:35 I would also inspect the 60 pin card edge for corrosion for the ram card maybe shorts to. I did pull a 80 colomb card out once the the plastic fram came with it


Tom
@rubber_jonnie

Both 80 column cards are new, so there's no edge damage to either, and the fault happens with each.

The socket looks fine under the microscope and it's also been cleaned. I'm going to film it today so I can get a more exact timing of when it fails and how long recovery takes, as I'm convinced it's thermal.

Hopefully the freeze spray I ordered will arrive today.

So, the machine runs fine with the on-board graphics but not with the 80 column card?

If it runs fine in 40 column mode then that rules out the memory, memory bus and video memory bus controlled by UD3 and it's something to do with the slot control circuitry. I would firstly suspect UB2. 74LS245s often fail. I'd also check the MDIN/OUT line as it's quite possible that the 245 is being told to send data the wrong way. Of course, this is being controlled by the MMU, so you could try freeze-spraying that.

stephen_usher wrote: 07 Apr 2023 10:29 So, the machine runs fine with the on-board graphics but not with the 80 column card?

If it runs fine in 40 column mode then that rules out the memory, memory bus and video memory bus controlled by UD3 and it's something to do with the slot control circuitry. I would firstly suspect UB2. 74LS245s often fail. I'd also check the MDIN/OUT line as it's quite possible that the 245 is being told to send data the wrong way. Of course, this is being controlled by the MMU, so you could try freeze-spraying that.

Yep, all test pass with the card removed, but it will run fine for about 5-10 minutes with either of the cards fitted, then there is on screen corruption and the card 'goes missing' and unless I let it sit for 5-10 minutes, it stays missing. After a cooldown period it does come back. This is both cards BTW.

As for UB2 I have replaced that with a new one and no difference, but it could be the socket. I do have a replacement coming for UD3 BTW. When my freeze spray arrives later, I'll be definitely be trying it on the MMU.

While I'm waiting I may see if I can use a compressor and airbrush to cool the MMU down.

I really hope it isn't the MMU...