Pentium MMX suddenly won't post -- suggestions?

Posted by Plus-Dust@reddit | vintagecomputing | View on Reddit | 8 comments

I bought a cute desktop-style case system last weekend -- it's a Pentium P55C-200mhz, 80mb RAM, two 350Mb hard drives and DOS 6.22 preloaded. An S3 Virge PCI VGA card and an ISA SCSI card were included (one of the drives is actually a SCSI disk from a 68K Mac and the card does some kind of option ROM thing to make it work). Full system working great. Since I couldn't find any of my other floppy drives I did some QBASIC trickery and bootstrapped Procomm Plus and PKUNZIP onto it over the COM1 port and then started XMODEM-ing some more software onto it. I played with it for a couple days not really having dived deep yet, ran a couple of simple Apogee games such as Commander Keen, Cosmo's Cosmic Adventure etc just to check it out, then last night I figured that since it should be able to handle more than 386-era stuff I would try copying a bigger program, so I started Procomm back up and started sending Blackthorne and Descent. The transfers were only running at 9600baud so I walked away for a few minutes and then about 5 minutes later came back and noticed the transfer had stalled. I waited a while and then tried to escape and realized it was locked up. I progressed through ctrl+alt+del, reset button, etc and finally just turned it off and on again. All of a sudden it didn't come back up. Hard drives and fans spin up but black screen and no floppy activity, and the Turbo light was off which it had been on all the time it was working. It never made any kind of POST beep in normal operation so I can't gather anything from the absence of any sounds. It was late so I went to bed and my BF messed with it and reseated the video card and got it to boot up again. Then he said it made a loud beep while booting and froze again and it hasn't come back since. What I've tried so far: * Reseating and de-ox all the cards * Reseating and de-ox all the RAM (four 72-pin SIMMs) * Try with just one pair of RAM or the other installed in Bank 1/2 only * Removing all the extraneous drives, SCSI card etc not required for POST * Checked the power-supply rails, it was like 5.1 and 11.8V * Reseating and de-ox the CPU socket * Found [the manual](https://theretroweb.com/motherboards/s/fic-pa-2007) for the motherboard and checked all the jumper settings. One of them was suspicious actually, CLK1/CLK2/CLK3 were set to up/up/up when according to what I'm understanding in the manual it should've been up/up/down for a P55 @ 200mhz. I actually don't see "up/up/up" as even a selection in the manual. The "VR" voltage-select block was also set slightly wrong to 2.9V for an AMD-K5 instead of 2.8V for the Pentium. Fixing those didn't change anything though. * I don't usually do DOS machines so I don't have one, but I've ordered one of those POST-code test cards. The machine was working fully and then with it coming back briefly I'm a little dubious if something might have actually blown up and that's why I went so after oxidation even though it seems to be in pretty good shape. But until I get that POST card I just don't know what else might be going on. The machine came from another enthusiast who seemed fairly knowledgeable so I don't think it was abused too much. FWIW, I'll note that there were a couple of weird things going on with the configuration that I noted down to look into later while it was running: * Despite a post-it on the case that says 20Mb RAM, the BIOS test was reporting 80Mb RAM. I don't understand how 4 SIMMs adds up to 80Mb, a very non-power-of-two value. * Although I chalked it up to the SCSI card, nothing seemed to agree on the hard drive configuration. The Award BIOS POST screen would report no drives found, then the SCSI option ROM start and would find one disk. Then the little summary page on the next screen before DOS started would say something else. Then in DOS you had two drives C: and D:. In the BIOS setup utility it had listed one drive with a custom CHS config which I wrote down. * This was the first time I had tried to push the RESET button, and I had thought at first that maybe I had accidentally hit Turbo and messed something up somehow as the buttons look very similar, but actually under the case both buttons are "soft" non-latching momentary buttons so I don't think that anything I might've pushed on the front panel should've been persistent.

8 Comments

[-]

Plus-Dust@reddit (OP)

UPDATE: I dropped a Pentium 90 in it, reset the jumpers, and it booted right up. Therefore, I'm going to chalk this up as the CPU having melted itself as soon as I started doing more than just leaving it at the DOS prompt, due to the PO neglecting to apply any thermal paste, for reasons I simply cannot fathom - who builds computers and hasn't heard you're not supposed to just slap the heatsink on without paste?

is_reddit_useful@reddit

> Turbo light was off which it had been on all the time it was working This seems interesting because often these are controlled by simple hardware, and not dependent on the CPU running code. If the switch and its connections are okay, I wonder if all the power supply voltages are present. Maybe the power supply voltage that would provide power to the LED isn't present.

I tested the 5V and 12V rails (off of a free molex connector), but looking [at the pinout](https://allpinouts.org/pinouts/connectors/power_supply/at-motherboard-power/), I see that there is also a -12V bias supply and "Power Good" signal going to the MB. I will check those and the other pins on the actual motherboard connector too just to be sure.

misterhinkydink@reddit

My ASUS PCI/I-486SP3G board did this same thing last week. Everything was working fine after 30 years dormant. At first I suspected the BIOS was corrupted after I updated and didn't move the VPP jumper back to 5V but verified the BIOS was OK. After probing the 486 have settled on a defective 486 after finding nothing that would hold it up.

boluserectus@reddit

20Mb's of RAM is simply 2x8 + 2x2. At a later stage the 2x2 was replaced by 2x32, which makes 80. That kind of CPU cooling is not strange, but make sure the surfaces touch each other properly, maybe apply a little paste, and (when it boots again) check the temp with your fingers. When it hurts, it's too hot. The BIOS setup has one drive listed, but the post screen reports no drives, this probably means your IDE hard drives died. SCSI drives are never listed in BIOS setup or POST screens. SCSI cards have their own BIOS, which you see at the summary page. Most probably the SCSI drive is partitioned and has C: and D: Turbo buttons do nothing in this era, the mother board has no connection for it. I am not completely sure in the way you described it, but my advice would be to start again with minimum config. Only some ram and a VGA card and see if it boots. Maybe try different slot for mem and VGA. Also try to reset your BIOS. Still nothing? Remove the CPU and try to listen for beep codes. If you never heard beep codes before and they start after removing the CPU, that probably means it's faulty, but I never ever had a broken CPU (with proper handling).

Both drives actually worked though -- although it was a little confusing because they both had almost the same contents and directory structure, there really was an extra D: drive once you got to DOS. So I dunno. Tried only RAM and VGA, changing slots, etc and nothing. I guess I need the post code card since the Award BIOS supposedly supports it I think, but manual says nothing about beep codes and it doesn't make any. It just seems weird that it came back for a little bit.

exjwpornaddict@reddit

If you do get it running again, it should make a good 98se computer. 80mb of ram is probably too much for dos 6.22, but should be suitable for win95/98se. I'd probably replace the 5.25 floppy drive with a cd-rom drive.

NiknakSi@reddit

Try another CPU if you have or can get one. I recently had a K6-2 system do something similar, suddenly no beep no post. Swapped out everything incl. PSU before finally trying a different CPU and that did the trick. Also check the caps on the board for any swelling or leakage.

Reply to Post

8 Comments