alf@xenon.stgt.sub.org (Ingo Feulner) (12/10/90)
Some weeks ago I've installed another 2Meg 32Bit Memory on my A2620
Card. I made some memory tests (using the ROM monitor) and I found
no problems. (The address space for the memory is from $200000 to $600000)
Today I got some very strange system faults and so I checked the Ram
again. There are now memory faults from $400000 to $500000.
I can't believe that suddenly 8 ram chips failed. So I think that the
problem is somewhere in the address logic. Any suggestions? Dave?
My config: B2000, 2090 HDController, Big Agnus, A2620.
Any help is appreciated.
--
Ingo Feulner - alf@xenon.stgt.sub.org
Wolfacher Weg 22 - 7030 Boeblingen - (+49) 7031 272691 - United Germany
Ueberzeuge Dich, dass dies so ist, wie ich es schreibe. (Seneca)daveh@cbmvax.commodore.com (Dave Haynie) (12/11/90)
In article <3589@xenon.stgt.sub.org> alf@xenon.stgt.sub.org (Ingo Feulner) writes: >Some weeks ago I've installed another 2Meg 32Bit Memory on my A2620 >Card. I made some memory tests (using the ROM monitor) and I found >no problems. (The address space for the memory is from $200000 to $600000) >Today I got some very strange system faults and so I checked the Ram >again. There are now memory faults from $400000 to $500000. >I can't believe that suddenly 8 ram chips failed. So I think that the >problem is somewhere in the address logic. Any suggestions? Dave? Chances are that the memory chips themselves are OK. Most likely, you had a cold solder joint somewhere in the third bank, somewhat less likely is a solder short that wasn't quite shorting until recently. One example of what can happen is my little adventure with A1000 piggyback memory expansion. I installed the Chris Erving modifications in my A1000 shortly after they were out in "read-me" file format on the networks (I added a daughterboard hack to locate the memory at $00C0xxxx, though since it was done before the A500 and A2000 were out, I couldn't tell why this worked). Around six months after I installed the thing, it started getting flakey. I finally broke down and examined my handiwork (done at 4AM), and found that one pin on one memory chip wasn't soldered at all, it was simply pressed against the chip below it hard enough to have worked great for six months. The fact that your A2620 worked OK for awhile is a good indication there's no defect in the actual logic or PC board. The best way to look for solder flaws, which on ZIP arrays like this can be REALLY hard to spot, is through a 50x-100x binocular microscope. We happen to have lots of these puppies around here, because the IC folks use them for microprobing their chips. If you don't have access to one, next best bet is a bright light and a good magnifying glass. I use an X-Acto knife tip to push and prod questionable joints, and pay close attention to traces or vias that go near your soldered chips. If you haven't already, try cleaning off the excess flux, too -- occasionally a fragment of metal will get glued to the PCB by the flux well enough to make a short. You don't have to have a great connection to make a short that's enough to zap a bank of memory, and any number of disconnected lines on a single DRAM can not only cause that chip to fail, but it can even drag down the whole DRAM bus. > Ingo Feulner - alf@xenon.stgt.sub.org -- Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy Wheeeeeeeeeeeeeeee...........
peterk@cbmger.UUCP (Peter Kittel GERMANY) (12/11/90)
In article <16421@cbmvax.commodore.com> daveh@cbmvax.commodore.com (Dave Haynie) writes: >Dave Haynie Commodore-Amiga (Amiga 3000) "The Crew That Never Rests" > {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: hazy BIX: hazy > Wheeeeeeeeeeeeeeee........... ^^^^^^^^ Hmm, perhaps you got a new car??? If yes, then good luck with it! -- Best regards, Dr. Peter Kittel // E-Mail to \\ Only my personal opinions... Commodore Frankfurt, Germany \X/ {uunet|pyramid|rutgers}!cbmvax!cbmger!peterk