A500 Rev6A apparent memory stability issues

[DaIceMan]

Hi all, I have a Rev 6 board which is very unstable (originally it was completely dead, I don't know what the ex owner did to it) and would like to know if any of you have encountered a similar issue. What happens is after a few seconds or even immediately simply booting the diagrom v1.2 the display gets slowly corrupted without doing anything and the system eventually crashes. Sometimes you can manage to get into some tests (Graphics, memory) and it may even work for a minute or two but it can also crash pretty instantly. Sometimes it just won't even boot the diagrom and show garbled flashing stuff as if powering it on without a memory chip. Only one time I actually managed to boot into workbench but briefly after it reported disk error and hanged. I usually test it using diagrom as it just is not stable enough for the Kickstart.



What I have done so far before reaching this stage:
- socketed and replaced the 74F244s near the DRAMs (one was faulty, I have an IC logic tester) and LS244 and 373 (also a faulty one).
- socketed and checked all 414256 DRAMs, swapped them for 70ns ones (working on another machine) plus the original 100ns ones work fine on other 6A also

- Checked 8372A Agnus working on another machine even swapped the CPU (original one was actually faulty)

- Checked Gary, Denise and both CIAs working on another A500 rev6.
- removed and checked main bus electrolytic caps (no DC leakage) and other 4 100uF caps
- refluxed all main sockets


The stability of the board is not compromised by flexing and/or pushing on any IC or socket (the brief moments it works) so it is *apparently* not a contact issue around sockets. The board has no corrosion and is actually in pretty good cosmetic shape. It is one of the light green boards (vs the dark green series).


It is possible that the board was powered by a faulty PSU, because I have never seen so many failures all in one (1 F244, 1 LS244, a faulty PAULA and CPU). It is possible that some other discrete components have gone out of tolerance but all the resistors appear visually good though I have not tested them all individually. Will continue to diagnose it but if anyone has any heads up I'm all ears!
Cheers

[solarmon]

Have you checked the data and address traces between CPU, Gary, Agnus and the memory chips? Not just for continuity, but for resistance? It could be that one or some of the traces has gone high resistance and is affecting the signal strength. I would inspect the traces with a microscope to see any damage that is not visible to the naked eye.

Also do any components gets hot? Hotter than normal?

I assume the PSU you are using has been checked and voltages are within spec?

[DaIceMan]

Hi and thanks for the heads up. I will check the main address and data traces (I have put the mainboard aside for the moment). The PSU is sound (I have 2, one ATX and one original Amiga switching version which I recapped and use with my "every day" amiga) - that's out of the question. It is certainly a motherboard issue, probably a trace as you say or some defective/degraded/dodgy component. Will let you know my findings! Cheers!

[DaIceMan]

As a follow up I managed to actually boot the Rev 6a board without glitches - here's what I did: I double checked all caps and replaced the ROM socket and found no problems. I could boot and run the diag rom for a while, even a couple of hours and sometimes it would just freeze and would not boot again. This was without PAULA installed. For the heck of it I installed PAULA anyhow and the system became suddenly rock stable (?!) and never froze again. I added another 512K of chip mem (so now it's 1M) and all is ok. I am waiting for the 2nd CIA as I don't want to gut my other A500 again to test this one. I am fine tuning a PS2 to Amiga keyboard converter with my scope (don't have a keyboard for this Rev6a) so I spend a lot of time on the diagrom in the keyboard diag menu so I am sure when it freezes or not. After PAULA was popped in, all problems apparently went away. I know PAULA handles some synch signals but the system should run stable with the diagrom + CPU + EVEN CIA + GARY and CHIP RAM. As soon as I get the other CIA I'll pop in the 1.3Kickstart and try to run some demos/diags and see if I get some glitches.

[solarmon]

Glad you got it working!

But I'm very confused. In your original post there was no mention of Paula not being installed - you mentioned about a faulty Paula, but I read that as you had Paula replaced.

In a recent test I did, when I do not have Paula installed on my A500 rev 6a, it boots to a white screen because Paula handles the interrupts, so it is in a wait state. I have not tried it with a DiagROM installed instead of a Kickstart ROM.

[DaIceMan]

I use the latest diagrom and you *can* boot without PAULA with it and just the EVEN CIA installed (and GARY of course); you can't complete the boot sequence with the regular Kickstart without PAULA as you noticed. I mentioned a faulty PAULA but I did not have a replacemenet YET, which when I received recently I replaced and hence the update!

[solarmon]

I'll try DiagROM without Paula installed and see what the behaviour is on my rev 6a.

[DaIceMan]

Cool - thanks, looking forward to your findings. You should get inconsnstent behaviors, sometimes it boots sometimes red screen or garbled. Sometimes it boots and you can run the tests for some time.

[solarmon]

Originally Posted by DaIceMan

Cool - thanks, looking forward to your findings. You should get inconsnstent behaviors, sometimes it boots sometimes red screen or garbled. Sometimes it boots and you can run the tests for some time.

That would not be surprising because Paula handles the interrupts.

I wasn't aware that you did not have Paula installed, otherwise I would have question that and the stability of your A500.