Hmmm, that does not seem to be a short.

A short would typically manifest itself as:

a) PSU shutting down immediately once power applied.
b) PSU/Motherboard/component getting very hot, followed by fire and sparks, then the magic smoke being released, then nothing.

The fact you are getting post codes is a good thing.

Use contact cleaner on RAM/CPU/AGP sockets and try it again.

I tried pulling, cleaning, and reseating the CPU, and the RAM slots are empty, but I'll try anything more than once if there is a chance it can help. As for why I was thinking it was a short, the auto-starting of the power supply reminded me of something like this, https://www.instructables.com/How-to-power-up … y-without-a-PC/ but I haven't checked if there is continuity between these pins on the motherboard. I'll try that next and report if I get any weird results.

Computer won't post without memory installed.

Ideally, for mimimal post on this motherboard, you want:

a) CPU
b) 1 x RAM stick
c) 1 x PCI video card
d) Your post card.
e) Keyboard plugged in

Without all that, diagnosing POST problems can be difficult. Your previous post codes indicate the motherboard was having trouble initializing the memory:

https://uxd.zendesk.com/hc/en-us/articles/360 … OST-Error-Codes

As for the PC instantly turning on when power is applied, that can happen if it is programmed in the CMOS to do this, or if the motherboard is in a undefined state when you apply power. I wouldn't generally take that by itself as an indication of any fault.

Also, the ISA slot you plugged the POST card into doesn't supply 3.3V, so that is why the 3.3V indicator is dark.

It would probably light up if you plugged it into a PCI slot and turned the system on.

Ok I tried plugging back in the keyboard, video card and a stick of RAM and the error (and the code on the POST test card) persists (this is consistent with how I first noticed the problem, when the computer out of the blue started turning on whenever I flipped the switch to the power strip that it was plugged into). I didn't have the bios setting for auto-power on enabled on this machine, but just to be sure I moved the bios reset jumper just in case the bios settings got corrupted. I also pulled the CR2032 battery out to test its voltage (a solid 3 volts). Any other places anyone would suggest? This computer was built with all new-old-stock parts a few years ago, I feel like any part of I can preserve would be worth the effort. Also thanks for the tip about ISA 3.3v, I moved the test card and while the error code is the same at least the lights are all lighting up 😀

EC -> Have you tried different sticks of RAM ... also try them in different sockets. Still stuck on memory init.

I can try to reproduce this error when I get home with bad or missing memory.

I found this page: https://www.intel.com/content/www/us/en/suppo … s-and-kits.html and it seems to suggest that meaning of the post codes are related to "recovery" though I can't find the exact codes of 3C and 3B the 30 to 3F range is dedicated to recovery messages

Your latest post code is EC .... and based on the list I posted earlier for this type of BIOS, it indicates a memory problem.

Could also be memory problem in the CPU.

sndtst wrote on 2021-10-20, 03:04:

As for why I was thinking it was a short, the auto-starting of the power supply reminded me of something like this, https://www.instructables.com/How-to-power-up … y-without-a-PC/ but I haven't checked if there is continuity between these pins on the motherboard. I'll try that next and report if I get any weird results.

a) The PowerOn signal to the PSU is provided by the chipset, not the CPU.
My personal guess is that it will do this even without CPU inserted. Am I correctly guessing?

b) A defective chipset that does "autostarting" like this might also fail to provide a clean reset, which also would reset the POST card. The POST card is apparently not being reset/initialized, as in normal operation it would not display "<blank> b", but "0b". Thus the "EC" code probably can be disregarded as meaningless, probably just what the POST chip "defaults to" when powering up.

c) Do you have means to rule out a PSU problem, like multimeters or a known-good PSU that can be used for testing?

retardware wrote on 2021-10-20, 08:18:

b) A defective chipset that does "autostarting" like this might also fail to provide a clean reset, which also would reset the POST card. The POST card is apparently not being reset/initialized, as in normal operation it would not display "<blank> b", but "0b". Thus the "EC" code probably can be disregarded as meaningless, probably just what the POST chip "defaults to" when powering up.

Actually, as the OP said, that is just the camera not picking up the refresh of the LEDs right. As far as I know, all 4 POST card ids are being displayed correctly, the camera just doesn't capture them.

canthearu wrote on 2021-10-20, 08:28:

Actually, as the OP said, that is just the camera not picking up the refresh of the LEDs right. As far as I know, all 4 POST card ids are being displayed correctly, the camera just doesn't capture them.

oops, my bad 😀 Need more ☕

My gut feeling says it's the chipset or the PSU which doesn't behave... after CMOS reset the mobo shouldn't start by itself.

I misread the POST codes they are EC and EB (not 3C and 3B), looking a the intel table of post codes only has a definition for EB ("Calling Legacy Option ROMs"). As for isolating the power supply as the culprit, I tried a different power supply from a known working system, same behavior, and the original power supply on that known working system and it worked. So this behavior seems to be isolated to the motherboard and CPU. Also to answer another suggestion, I tried no RAM, single RAM, and different RAM, all the same. I also tried pulling the CMOS jumper (the motherboard manual describes this as putting the BIOS in recovery mode in case a BIOS update fails).

Just got a replacement Motherboard (which also came with a CPU and RAM). The original CPU and RAM all work in this new board, so I guess this narrows it down to the motherboard. I may try and reflow the chipset chips to see if it helps but with a working system I will likely just put the board into a growing pile of "doesn't work, needs more testing" hardware that I have collected. Thank you everyone for the ideas of things I can test. I learned a lot 😀