Abit AN8 suddenly shut down - now POST code 9.0

[exscape]

Hey everybody.
I bought a (complete) computer from a friend a few days ago. It's been running fine since this Sunday, but stopped working today for no apparent reason.

I added two NICs (both of which I know work) and booted up, and then rebooted a couple of times due to OS changes. Anyway, I got the OS to find them, so I set them both to DHCP and rebooted again (to make sure the settings would stick across a reboot).
When booting up, after giving control to the OS, the computer suddently shut down, and now it won't start again.

I get the well-known, horrifying POST code 9.0... Despite doing squat in the BIOS at the time. I have also not flashed the BIOS, ever.
Clear CMOS didn't help (neither did removing the battery).
The only thing I've changed in the BIOS at all was the OTES fan settings (I set i to 8-12v, control at 30-45C or so SYS temp, and shutdown at 55C - IIRC that is). It DID work like that for 30 minuter before dying on me, so the issue is NOT that it froze during save or anything like that. It simply died when the OS was booting up, for no apparent reason.

I've tried removing everything bar the CPU and PSU, still 9.0. I even get 9.0. without the BIOS chip, just to test, but that might be expected, though. All hardware was 100% an hour ago, AFAIK.

I also do NOT have any computer to hotflash with.
In case it matters... The CPU is an Athlon 64 3200+, 4x512MB sticks of RAM (same error without them, though) and a nearly brand-new 420W PSU (18A on the 12v rail). But as I said, it worked flawlessly for 48 hours, under load too, before this happened.

When googling I found something about resetting the CMOS, removing the CPU fan and booting up. I got error 90 when doing that, too, despite expecting a 8E (I think) code.

Still, I can't for the life of me figure out why it'd just... die.
Help me, please!

Edit: What the f**k? Another computer is also dead, I noticed... It was turned off by the time the AN8 computer died, but won't start. I flip the power switch (on the PSU), and it starts up for a second or two then shuts down. The soft-power button doesn't do squat at this point. This isn't my day
Edit 2: OK, got that running again with a simple reset CMOS. No progress on the new, more expensive computer that I *want* to have running... I'm gonna try with another PSU in a bit.

[BUFF]

You didn't have a power surge where you are by any chance?

Anyway 9.0. is most often either BIOS (corrupt , unsupported CPU etc.) or power (failing/inadequate PSU, short etc.) related.

[exscape]

I was thinking about the power thingy too. My dad's computer is in the same outlet and it kept running and still works just fine.

The PSU was not it, btw. I got fine readings from it and another PSU didn't help at all.
I've also tried hotflashing the BIOS in another nForce4 comp, but no success. (WinFlash says "This File is no support Bios Lock Function" (sic) and awdflash from floppy/CD just keeps blinking "Please wait !" forever).

Any advice?
I plan to try UniFlash or similar tomorrow.

[BUFF]

The PSU was not it, btw. I got fine readings from it and another PSU didn't help at all.

consumer PSU testers usually are poor especially as they normally don't test the PSU under load.
What other PSU did you try?

I've also tried hotflashing the BIOS in another nForce4 comp, but no success. (WinFlash says "This File is no support Bios Lock Function" (sic) and awdflash from floppy/CD just keeps blinking "Please wait !" forever).

Winflash is probably comparing the BIOS file against the mobo & realises that there is a mismatch.
Similarly with Awdflash you would need to use the QI switch.

Another thing to check is that on the AN8 series there are 2 holes without a solder ring around them (1 just below the DIMM slots & 1 bottom right of the PCB) & these are not meant to have standoff/screws in them but people often do - this can cause a 9.0. (short)

[exscape]

consumer PSU testers usually are poor especially as they normally don't test the PSU under load.
What other PSU did you try?

More or less a noname, also 420W. That one also works great though (measures nice too), and it currently powers a very similar computer (3000+) with no problems.

Winflash is probably comparing the BIOS file against the mobo & realises that there is a mismatch.
Similarly with Awdflash you would need to use the QI switch.

Another thing to check is that on the AN8 series there are 2 holes without a solder ring around them (1 just below the DIMM slots & 1 bottom right of the PCB) & these are not meant to have standoff/screws in them but people often do - this can cause a 9.0. (short)

Yep, I did try /QI but it didn't help. I'll try a newer version of awdflash tomorrow, I read about /nbl (no bios lock) and /f (force, I think?) switches, that my version doesn't have.

BTW, there was a screw in the hole below the DIMMs - I removed that one, but to no avail.
In any case, thanks for the answers!

[BUFF]

You ideally want a PSU with 24A plus on the 12V but I wouldn't recommend anything lower than 20A.

[exscape]

OK, but still... It shouldn't take that much power simply to boot up the CPU/RAM, with no other components (no HDD/DVD, no fans, no GFX). Besides, it worked fine for two days, doing heavy compiling etc, and now it can't even power on.

Also, more bad news, I'm afraid. I successfully flashed the BIOS (both onto my IOSS Bios Saviour and the original AN8 EEPROM), but it didn't help, I still get 9.0 (I cleared CMOS after inserting the new chip, too). Damnit... Any other suggestions?

[BUFF]

It won't boot with no fans & no gfx - it'll error out before that.

It still sounds like a power problem to me - PSUs do fail so the fact that it ran for 2 days doesn't rule that out.
About the only other thing that I can suggest is an out of the case install to rule out shorting on the case.

[exscape]

Yeah, they do fail, but it should've booted with the second PSU (that I know works, I used it in the hotflashing computer for one) if it was "simply" broken.
What I meant was that I find it unlikely (although possible) that it simply doesn't provide enough power to POST.

I thought about an out-of-case test as well, gonna do that tomorrow, it's getting late and I'm tired like hell.
In any case, thanks again, and I'll report back later. Hopefully with some good news.

[Agent]

use /tiny when hotflashing. It will bypass all the checks.

You will need to leave your old chip in right until the point of it saying "are you sure", then swap, then press Y.

[BUFF]

Yeah, they do fail, but it should've booted with the second PSU (that I know works, I used it in the hotflashing computer for one) if it was "simply" broken.
What I meant was that I find it unlikely (although possible) that it simply doesn't provide enough power to POST..

what are the specs of the other system? e.g. an nF4 system needs more power than a VIA K8T800 system.

Plus with modern PSUs you can actually have problems if you don't have enough load on the rails - we had 1 the other week on the abit USA forum with a barebones installation - once he added an HDD it powered up OK though.

Anyway the below is some stuff that Sean has previously said.
re. 9.0 on an AN8 32X
"Before you start clear the CMOS.
1. Remove all parts except the CPU and Heatsink.
2. Power on the system. Should see a code C1 or 04.
3. Then Add one stick of RAM, Should run through a series of codes and stop at 7F.
4. Now add your display card. At this point providing steps 1-3 are OK you should get a display."

re. 9.0 on an AN9 32X
"remove all other components excluding the CPU.

Power on the board and see what code is generated. If you see any other code than 9.0. you know its the RAM. If it still stops at 9.0. it only leaves three pieces of harware.
-PSU
-Motherboard
-CPU"

[exscape]

Sleep, who am I kidding?

use /tiny when hotflashing. It will bypass all the checks.

You will need to leave your old chip in right until the point of it saying "are you sure", then swap, then press Y.

I don't mean to sound condescending at all, but are you really sure? The other M/B isn't an AN8, so shouldn't I switch before starting awdflash to make sure that it detects the correct chip? I switched after booting, but before actually doing anything (eg. starting awdflash).

what are the specs of the other system? e.g. an nF4 system needs more power than a VIA K8T800 system.

The other system is also nF4 (MSI K8N Neo4), with a Venice 3000+, 2x512MB RAM, 1 HDD, 1 CDRW.

Plus with modern PSUs you can actually have problems if you don't have enough load on the rails - we had 1 the other week on the abit USA forum with a barebones installation - once he added an HDD it powered up OK though.

Anyway the below is some stuff that Sean has previously said.
re. 9.0 on an AN8 32X
"Before you start clear the CMOS.
1. Remove all parts except the CPU and Heatsink.
2. Power on the system. Should see a code C1 or 04.
3. Then Add one stick of RAM, Should run through a series of codes and stop at 7F.
4. Now add your display card. At this point providing steps 1-3 are OK you should get a display."

re. 9.0 on an AN9 32X
"remove all other components excluding the CPU.

Power on the board and see what code is generated. If you see any other code than 9.0. you know its the RAM. If it still stops at 9.0. it only leaves three pieces of harware.
-PSU
-Motherboard
-CPU"

Good info, thanks. CPU, eh... I suppose I could try with the 3000+ (that'd be easier than buying a new PSU, possibly in vain).

[BUFF]

1 thing to bear in mind is that the AN8 32X & AN9 32X mentioned above are both uGuru 3 mobos whereas your AN8 uses uGuru 2 - this may possibly make a difference to codes/behaviour.

[exscape]

CPU was OK, it worked just fine in the other comp.
I guess that leaves two options, power and MB.

[exscape]

OK, now I've tried it all.

In dad's computer:
* My RAM
* My PSU
* My CPU
* My GFX
Everything worked.

In my computer:
* Dad's RAM
* Dad's PSU
* Dad's CPU
* Dad's GFX
9.0 in all cases.

Time to get a new mobo or is there a tiny possibility of something else?

[Agent]

I don't mean to sound condescending at all, but are you really sure? The other M/B isn't an AN8, so shouldn't I switch before starting awdflash to make sure that it detects the correct chip? I switched after booting, but before actually doing anything (eg. starting awdflash).

Yes. I have lost count of how many hot flashes I have done via this method
Most of the time, it will work. I've had the odd failure, but this has usually been down to other factors.
That includes doing Intel BIOS's on Nforce boards, and the opposite way around.

Its probably best to try and flash your current working chip with its normal BIOS, so you can see at which point you need to swap.
You *must* use the /tiny switch.