VOGONS


First post, by silvervest

User metadata
Rank Newbie
Rank
Newbie

Hey folks, trying to recover a PS/2 Model 70-386 I acquired recently, there have been a lot of repair work so far (PSU was dead, floppy had leaked caps and corroded traces) but I've finally got it booting and reading diskettes!

The problem I'm facing now though, is that when I put the 70/80 reference disk in, it boots into the setup program, does auto-configuration, then restarts but upon restart, it still gives 162/163 and wants me to go into the setup program. When I do so, I get the black/yellow warning screen attached, stating it had a problem reading from battery backed memory and it wants to try to recover or run auto-config again. Either option leads to loading for a bit, then it restarts and goes around in circles.

I can hit F1 without the ref disk in to get into BASIC but nothing else. I've also tried taking out all the cards and HDD out, just to the planer, backplane and FDD and same thing. With the MCA adapters I have installed, it detects them during the auto configuration and needs the option files added, so that seems to be working. The CMOS battery is a brand new one, and I've tried just feeding it 6v from my benchtop PSU as well.

Any ideas?

Attachments

Last edited by silvervest on 2021-11-22, 11:59. Edited 1 time in total.

Reply 1 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

As an update, I was ... somehow able to get into the reference disk configuration menu, went in and set the time/date and manually set and saved the configuration. On reboot, it was still showing a 162/163 error so I'm guessing there's something wrong with the CMOS/SRAM storage chips identified here (https://ardent-tool.com/8570/Planar_T1.html) maybe? I'm pulling my hair out on this one now...

Reply 3 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

As noted, I've replaced the battery with a brand new one, and even used a benchtop supply in place of the battery to be absolutely sure that's not the problem.

Reply 4 of 25, by Rwolf

User metadata
Rank Newbie
Rank
Newbie

Trace corrosion as mentioned was present, might cause voltage drops/breaks from the battery to the chips needing it, so maybe check all affected lines/solder points inbetween?

Reply 5 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

Hmm... the trace corrosion was on the floppy drive from the leaked electros, but I guess that is directly above the circuitry between the battery and the RTC. I couldn't see anything obviously damaged on the planar, but I guess I should measure between the VBAT at the RTC/CMOS directly to be absolutely sure!

Reply 6 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

Okay then, all traces are healthy, no corrosion anywhere on the planar. On the MC146818A itself and during power off, I'm seeing a constant 4V between Vss and Vdd, and 5V when powered on - both are well within the acceptable range for that IC.

Reply 8 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

Yep sorry, should have mentioned voltages there were fine too. I was able to get into the ref disk system checkout and run tests, and it's definitely finding something wrong on the planar. Not quite sure what this error is though, I can't find a definitive guide to what the diagnostic test error codes mean - closest I can find is that 00010300 as a POST error on a 90/95 system means checksum/clock error... Who knows indeed! Best guess, I'll just replace both the SRAM and RTC chips with either new replacements or NOS and see what happens.

Attachments

Reply 9 of 25, by BitWrangler

User metadata
Rank l33t
Rank
l33t

Have you tried a complete clear? Belt and braces style, short jumper, leave battery out overnight style of thing. Once in a handful of blue moons, you get a bit set in the CMOS RAM in a "reserved" area that the setup program can't set, unset or clear, but it screws up the checksum.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 10 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

I haven't shorted the jumper, but the machine was sitting on a shelf for 4-5 years without a battery or power... but yeah I'll give that a try too. Throwing everything at the wall at the moment, hopefully something sticks!

Reply 11 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

No luck on the belt and braces clear, unfortunately. Left it overnight with the jumper in and battery and power out, booted up this morning to a 161/163, went through the usual procedure and on reboot it went back to the 162/163 and configuration data error boot loop.

Reply 12 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

Okay I've had it sitting for a few days now without the battery, and I remade a bunch of copies of the reference disk just in case. Same thing.
I've also switched to a CRT monitor from the LCD I was using above, and the 2401 error I had attributed to the LCD is still there too, so I guess there's definitely something wrong on the planar. I've ordered some replacement RTC and SRAM ICs so will replace those when they arrive, and hope for the best. The 2401 error is making me rethink though and looking more at the video circuitry, but everything there seems physically fine...

Reply 13 of 25, by luckybob

User metadata
Rank l33t
Rank
l33t

can you describe the corrosion damage a bit? That sounds like a probable cause. Does the bios settings stick with a warm boot? if it only doesn't survive a cold boot, it sounds like the battery circuitry is likely to blame.

It is a mistake to think you can solve any major problems just with potatoes.

Reply 14 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie
luckybob wrote on 2021-11-29, 17:12:

can you describe the corrosion damage a bit? That sounds like a probable cause. Does the bios settings stick with a warm boot? if it only doesn't survive a cold boot, it sounds like the battery circuitry is likely to blame.

The corrosion was only on the floppy drive, caused by leaked SMD electrolytics. The main planar has no corrosion that I've been able to find, and I've looked pretty closely now. The main boot loop that I'm referencing here is during warm reboots, after the reference disk restarts during automatic configuration, and I've tested the battery circuits end to end and they're providing correct voltages.

Reply 15 of 25, by luckybob

User metadata
Rank l33t
Rank
l33t

if it were me, at this point, i'd double check and reseat any parts in sockets. looking up the 2401 error...

2401 (PS/2) Planar video error -- it can be the display or the video.
First check for a solid connection -- wiggle the plug at the system end and check for one or more bent pins. Next swap the monitor. If a 2401 error persists after swapping the monitor, then the system board should be replaced next. However if your monitor serial number starts with "72" , is below 72-0374000 and you cannot find the number "61" stamped on the bottom of the display (you may need to remove the display stand), then a replacement monitor will probably fix this error.

If it's definitely on the planar (ie another monitor fails) check the Inmos palette DAC (ceramic DIP module with gold colored middle) for shorts between the pins.

going further, this is beyond my ability. My beard isn't nearly long or grey enough. IBM Museum's majestic facial hair is precisely suited for the task: https://www.youtube.com/channel/UCEfyT_TIXx41xA3U4psqWkw and the secret PS/2 underground where most of us lurk is at https://comp.sys.ibm.ps2.hardware.narkive.com/

It is a mistake to think you can solve any major problems just with potatoes.

Reply 16 of 25, by hyoenmadan

User metadata
Rank Member
Rank
Member

https://www.youtube.com/watch?v=Y6UprDq8LAQ

TL;DR, Seems like soldering on PS/2 planars doesn't age up to IBM quality expectations. As seen in the video, it frequently develops micro-cracks and other subtle, but insidious issues.

Reply 17 of 25, by silvervest

User metadata
Rank Newbie
Rank
Newbie

Just as an update, I ended up replacing both the RTC and SRAM ICs to no avail. After consulting the boffins on comp.sys.ibm.ps2.hardware as recommended, they led me down a good track.

Was able to confirm the RTC functionality perfectly, but the SRAM was not returning anything. Digging and probing with my scope and multimeter led me to find the ~CS line on the SRAM was constantly held high, and I traced that back to the hex inverter MC14069U at U158 which didn't seem to be inverting. I drove it to ground with a jumper wire and all the system issues went away.

Will be looking to replace that hex inverter, but for now the jumper wire seems to be working! Thanks for everyone's suggestions!

Reply 18 of 25, by luckybob

User metadata
Rank l33t
Rank
l33t

I think most of the ibm greybeards have a presence there.

I was watching the CuriousMarc channel fix a very similar system, it was just a borked via.

It is a mistake to think you can solve any major problems just with potatoes.