FIC AU31 No longer posting

[ProdigyS10]

I have an FIC AU31 i'm hoping to resurect to be a win98SE retro pc. has a athlon xp 3200+ cpu.
it had bad capacitors on it causing reboots (or so i thought), replaced the capacitors 4 of them had opened up. we didn't do a very clean job. but it worked and the system seemed stable. during this testing though i had a pci card that was up slightly and without thinking i pushed it down and
the system hung. i tried rebooting and wouldn't get any post beeps after that. ive since upgraded my solidering/desolidering station and cleanly mounted the capacitors nicely hoping that it was just one of the caps got loose.

still no luck so i've started trying to diagnose it deeper
removed ram and cpu and power on the board it turns on and fans spin but zero post beeps. i can push and hold the power button and it will still respond and shut off that way so i don't think its totally dead yet.

using https://cdn.badcaps-static.com/pdfs/...ef6a0f5689.pdf
on the socket A cpu i looked up pins wanting to see if the cpu was getting power since it didn't seem warm ever after trying to post.
took a reading from B2 (vss) got like .01v and then D2 (vcc) but got 0v.

any thoughts on where to look next?

[ProdigyS10]

Re: FIC AU31 No longer posting

https://theretroweb.com/motherboards/s/fic-au31-l is the board in question if its any help

[ChaosLegionnaire]

Re: FIC AU31 No longer posting

i had a pci card that was up slightly and without thinking i pushed it down and the system hung.

i think doing ^this^ might have resulted in esd that fried the southbridge.

do u have another spare known working cpu and ram to test? to rule out faulty cpu and ram, or another spare socket 462 motherboard to test the cpu and ram? to prove the fault is indeed with the motherboard. i know athlon xp cpus of that era are "fragile and brittle" and are prone to failure: case #1. if u run it without the heatsink, it would often overheat and fry the chip. not sure if it lacks proper esd protection as well, if u tried hotplugging things in the expansion and/or ram slots, it may fry the cpu...

on the socket A cpu i looked up pins wanting to see if the cpu was getting power since it didn't seem warm ever after trying to post.
took a reading from B2 (vss) got like .01v and then D2 (vcc) but got 0v.

umm... u take cpu vcore readings from the body pad of the cpu vrm mosfets, the cpu vrm coils or the positive leg of the caps on the cpu vrm output. NOT on the cpu socket pins! they are fragile and so many of them packed closely together. what if u accidentally short the adjacent pins? that would most likely fry/short-circuit the cpu!

[ProdigyS10]

Re: FIC AU31 No longer posting

i think doing ^this^ might have resulted in esd that fried the southbridge.

do u have another spare known working cpu and ram to test? to rule out faulty cpu and ram, or another spare socket 462 motherboard to test the cpu and ram? to prove the fault is indeed with the motherboard. i know athlon xp cpus of that era are "fragile and brittle" and are prone to failure: case #1. if u run it without the heatsink, it would often overheat and fry the chip. not sure if it lacks proper esd protection as well, if u tried hotplugging things in the expansion and/or ram slots, it may fry the cpu...

umm... u take cpu vcore readings from the body pad of the cpu vrm mosfets, the cpu vrm coils or the positive leg of the caps on the cpu vrm output. NOT on the cpu socket pins! they are fragile and so many of them packed closely together. what if u accidentally short the adjacent pins? that would most likely fry/short-circuit the cpu!

hoping not to have fried the southbridge but its definitely possible.

lots of spare ram but only the one cpu. but in theory it should give the no ram and no cpu beeps if its getting to that point.

pulling the voltage from the socket seems to be fairly common on a lot of diagnostic youtube channels. necroware frequently will place a small wire in to the right pin point to confirm a voltage or frequency for the cpu just like i did. infact thats where i got the idea from.

i do have one of those pci/isa post code cards coming that may shed some light on things once it arrives as well.

[momaka]

Re: FIC AU31 No longer posting

still no luck so i've started trying to diagnose it deeper
removed ram and cpu and power on the board it turns on and fans spin but zero post beeps. i can push and hold the power button and it will still respond and shut off that way so i don't think its totally dead yet.

Can't test board for POST codes without CPU. Without RAM - yes.

Install CPU and CPU heatsink (with proper thermal compound) and heatsink fan attached to motherboard (some socket A motherboards have a protection feature where they won't POST if they don't see a tachometer readout for the CPU fan.) No need to install memory. All other peripherals and hardware removed, including any PCI cards. See if board gives memory beep code. If not, unplug PSU from the wall and remove CMOS battery for a minute. Re-install CMOS battery and re-test again. If you do get a memory error / beep code, install memory back (1 stick at a time.)

pulling the voltage from the socket seems to be fairly common on a lot of diagnostic youtube channels. necroware frequently will place a small wire in to the right pin point to confirm a voltage or frequency for the cpu just like i did. infact thats where i got the idea from.

Too risky, IMO - very easy to slip with the probe/wire and short something out.
Doing it from a low-side MOSFET's Drain pin is easier and gives acceptable accuracy usually. Doing it at the output of the filter toroids is better (may get slightly more accurate voltage.)

i do have one of those pci/isa post code cards coming that may shed some light on things once it arrives as well.

Good.
Those do come in handy every once in a while with older boards, at least to see if a board is doing something or not.

i know athlon xp cpus of that era are "fragile and brittle" and are prone to failure: case #1. if u run it without the heatsink, it would always overheat and fry the chip.

Fixed that for you.

And yes, socket A CPUs are also fragile in the literal meaning too - very easy to crack the core or chip a corner if not careful during heatsink installation.

not sure if it lacks proper esd protection as well, if u tried hotplugging things in the expansion and/or ram slots, it may fry the cpu...

Older CPUs are actually more resilient to ESD, in or out of the board.
Also, on these older PCs like socket A, socket 370, and socket 478, pretty much all of the pins on the CPU go through the Northbridge. So hot-plugging a PCI card can't really fry the CPU. PCI signals on old boards go through the Southbridge, which then connects with the Northbridge, which communicates with the CPU. So more likely the Southbridge could have taken a hit... or the PCI card in question (hence the suggestion to test with all PCI cards and other non-essential hardware removed.)

[ChaosLegionnaire]

Re: FIC AU31 No longer posting

pulling the voltage from the socket seems to be fairly common on a lot of diagnostic youtube channels. necroware frequently will place a small wire in to the right pin point to confirm a voltage or frequency for the cpu just like i did. infact thats where i got the idea from.

huh? this means he measured by putting the wire in with no cpu in the socket. this is meaningless and doesnt do anything because the cpu needs to be in the socket to send its vid signal to the vrm controller ic to tell it what voltage the cpu requires to start. different cpus are programmed with different vid values at the factory. without a cpu in the socket, the vrm controller doesnt know what voltage to send to the cpu so it just turns itself off as a safety feature by default so no wonder u measured 0V at the vcc pins... therefore, this doesnt tell us anything useful about why the board wont post at all...

[Per Hansson]

Re: FIC AU31 No longer posting

pulling the voltage from the socket seems to be fairly common on a lot of diagnostic youtube channels. necroware frequently will place a small wire in to the right pin point to confirm a voltage or frequency for the cpu just like i did. infact thats where i got the idea from.

Adding to what ChaosLegionnaire said: the quoted procedure only works on older CPU's where you set the voltage via jumpers on the mainboard.
Then it it a good test to make sure you don't fry a 3.3v CPU with 5v for example.
For more modern boards that set the voltage automatically by detecting the CPU this is not useful...