• Please review our updated Terms and Rules here

IBM RS/6000 7248-132 error 0001000A

davidrg

Experienced Member
Joined
Feb 4, 2015
Messages
91
Location
New Zealand
Last night I tried firing up my RS/6000 43P type 7248 for the first time in over a decade with the hope of exploring AIX a little more and maybe having a go at Windows NT or OS/2 on this thing. I don't remember when I last had this machine running but I do have a photo of it running AIX in 2008 so at the very least it was working fine as recently as then.

When I fired it up I was a little worried about something going bang after it sitting unused for so long, but something potentially worse seems to have happened in its long sleep:
F1oT0MXaYAEYvss


The machine looks fine aside from a little white powder (corrosion?) on the CPU heatsink some of which has fallen on the motherboard - I assume its aluminium oxide and so non-conductive. Only aluminium electrolytics are over by the sound chip and they look fine . None of the tantalums have exploded. Haven't looked inside the PSU but given it reliably gets this far and no further I assume the PSU is ok.

At the debug console the date was showing as 1923 so I replaced the battery and reset the clock with the date and time commands but that hasn't fixed anything except the date/time being wrong. Tried pulling out all expansion cards (S15 video and some Symbios SCSI), removed all the memory modules and installed two different ones in bank 1 (leaving the rest out), put a SCSI terminator on the external SCSI port. I've tried inserting the SMS disk and hitting F1 or F4 but it doesn't attempt to read the disk. All it does is display the keyboard and chip icons, pause a while, then goes to the debugger. Telling it to "go" does nothing - it just sits there forever doing nothing.

Anyone have any idea what this error is and if it can be fixed easily without replacing hard-to-find components like the entire motherboard?
 
Does it trap at the same address every time? If so, does swapping memory SIMMs around change the displayed error detail? If the problem moves around I'd tend to suspect memory or motherboard cache. Otherwise might be a bad CPU. And, not to be Captain Obvious, but have you measured supply voltages and looked with a scope to spot excessive noise? If it's been in storage it wouldn't hurt to clean the power supply multipin and work it in and out a few times. Just throwing things out to see what sticks. These are fun boxes, BTW. I had a 43p 7248 for a long time but ended up having to part with my workstation collection due to space limitations.
 
Yeah, same address every time, swapping SIMMs around has no effect (those weird over-size SIMMs are a real pain to remove!). I've not checked PSU voltages yet as its an odd one I don't have a pinout for, and I don't have access to a scope either unfortunately. I'll try to pull the PSU apart sometime this week though and see if anything looks obviously unhappy in there - maybe I'll be lucky and there is a leaking capacitor or something.

I guess if it turns out to be a faulty CPU perhaps a replacement can be installed in that ZIF socket though I suspect the upgrade processor is probably harder to find than the entire machine at this point.

Really is a shame that this machine didn't survive sitting on a shelf for a little over a decade! I was looking forward to seeing if dual-booting AIX and NT is at all possible, and maybe having a go at that PowerPC port of OS/2.
 
Turns out this was a memory problem after all!

I had tried shuffling the SIMMs around previously with no success. Before packing the machine away as dead/for-parts I had one last go and decided to try swapping all the RAM for some stuff from a DEC Multia and to my surprise it got all the way through self-tests and started asking for the SMS disk!

I then tried putting just two of the original 32MB SIMMs back in and the problem returned. Swapped out SIMM #2 with another one and it was working again! Added all the SIMMs back in and it was still fine! So I guess the problem must have been a dirty and/or faulty memory slot. I did blow the slots clean before putting the Multia RAM in so maybe thats what did the job.
 
Back
Top