PC just shut down... ARGH!

More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory and a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

Played Doom3 for a couple of hours last night without any issues. When I
exited the game the CPU temp was around 38'C so I don't think I was taxing
the machine. Opened my OE to read the newsgroups and about 10 minutes into
that the whole PC just shut off. Not fun.

I could not get the PC to start regardless of what I did. Pulled the power
cable and reattached. Disconnected hard drives. Pulled PCI cards. Reset
CMOS. PC would just not post - no beeps - anything. Once I pulled the video
card I got the standard "no monitor" beeps. I dropped in a Geforce 2 card I
have and the PC is working again. Put the ATI card back in and no more boot.
***ARGH*** I killed my card!

I pulled the Artic Cooling VGA Cooler off the card (no shim). No damage I
can see. Compound right where it's supposed to be. No signs of overheating.
I put the stock ATI cooler back on. Still dead. BAH!!! Left the PC in pieces
for the night (it died 5 minutes before going to bed).

This morning I power up with the ATI card in and it works fine again. (Had
to reset the CMOS settings).

What happened? Nothing was overly warm when I was pulling it apart last
night. I didn't do anything this morning except turn the PC on.

On Thu, 19 Aug 2004 12:29:05 GMT, "Noozer" wrote:

More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory and a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

Played Doom3 for a couple of hours last night without any issues. When I
exited the game the CPU temp was around 38'C so I don't think I was taxing
the machine. Opened my OE to read the newsgroups and about 10 minutes into
that the whole PC just shut off. Not fun.


What happened? Nothing was overly warm when I was pulling it apart last
night. I didn't do anything this morning except turn the PC on.

See how clever Carmack is? Thats just one of the scares built into the
game. You get to a certain level and your whole PC shuts down. When
you reach the top level , your keyboard locks up and your power supply
catches on fire.


All I can say is I havent had any problems so far with my ATI 9800
playing the game and it hasnt caused any problems so far in another
system running an ATI 9000.

" wrote in message
...
On Thu, 19 Aug 2004 12:29:05 GMT, "Noozer" wrote:

More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory
and a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all
benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

Played Doom3 for a couple of hours last night without any issues. When I
exited the game the CPU temp was around 38'C so I don't think I was
taxing
the machine. Opened my OE to read the newsgroups and about 10 minutes
into
that the whole PC just shut off. Not fun.

What happened? Nothing was overly warm when I was pulling it apart last
night. I didn't do anything this morning except turn the PC on.

See how clever Carmack is? Thats just one of the scares built into the
game. You get to a certain level and your whole PC shuts down. When
you reach the top level , your keyboard locks up and your power supply
catches on fire.

Bahahaha.... Sounds like I hit the "hell" level early. : )

D3 is actually a bit more enjoying once I found the gamma adjustment in the
cfg file. Still a bit linear for me though and there's no way to sneak up on
the baddies even if you know where they are hiding (ie, toss a grenade down
a staircase or around a corner and it won't touch any baddies there unless
you've seen them and they've seen you).

All I can say is I havent had any problems so far with my ATI 9800
playing the game and it hasnt caused any problems so far in another
system running an ATI 9000.

Something I noticed this morning is that the boot BEEP is pretty delayed
after resetting the CMOS so I may not have waited long enough at boot while
troubleshooting last night.

"Noozer" wrote in message
news:l21Vc.142159$M95.128995@pd7tw1no...
More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory and
a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

Played Doom3 for a couple of hours last night without any issues. When I
exited the game the CPU temp was around 38'C so I don't think I was taxing
the machine. Opened my OE to read the newsgroups and about 10 minutes into
that the whole PC just shut off. Not fun.

I could not get the PC to start regardless of what I did. Pulled the power
cable and reattached. Disconnected hard drives. Pulled PCI cards. Reset
CMOS. PC would just not post - no beeps - anything. Once I pulled the
video
card I got the standard "no monitor" beeps. I dropped in a Geforce 2 card
I
have and the PC is working again. Put the ATI card back in and no more
boot.
***ARGH*** I killed my card!

I pulled the Artic Cooling VGA Cooler off the card (no shim). No damage I
can see. Compound right where it's supposed to be. No signs of
overheating.
I put the stock ATI cooler back on. Still dead. BAH!!! Left the PC in
pieces
for the night (it died 5 minutes before going to bed).

This morning I power up with the ATI card in and it works fine again. (Had
to reset the CMOS settings).

What happened? Nothing was overly warm when I was pulling it apart last
night. I didn't do anything this morning except turn the PC on.


The card worked fine again after removing and replacing it a few times. So,
most probably the card became unfixed in the socket, or there was a bad
contact at the connector. This happens more (essentially with RAM's).
When replacing the card, I would suggest to use a contact cleaner spray and
to control for a perfect fixation of the card (don't forget the slider!).

" wrote
All I can say is I havent had any problems so far with my ATI 9800
playing the game and it hasnt caused any problems so far in another
system running an ATI 9000.

Quality between different card manufacturers can vary a LOT
even if they have the exact same GPU and other features.
So if you don't have a reference card then saying "Oh, your
ATI 9800 doesn't work? My ATI 9800 works just fine"
doesn't really tell a thing.

"Noozer" wrote in message
news:l21Vc.142159$M95.128995@pd7tw1no...
More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory and
a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

First off, I don't care how fast your memory is...512 is just too little
amount for D3. 1GB of PC400 is better than 512 of PC533. Second, you're oc'd
cpu is always suspect when it comes to PC shutdowns. Again, I don't care
what the third party "stability" apps say, that is where there is risk. I
have the same vid card as you and only a 2.26GHz northwood running stock and
the game runs very smooth at medium(good enough for me). BUT, I have 1GB of
PC3200 ram and that helps a lot. Take this for what's it's worth.


Ron

"Ron" wrote in message
...

"Noozer" wrote in message
news:l21Vc.142159$M95.128995@pd7tw1no...
More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory
and
a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all
benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

First off, I don't care how fast your memory is...512 is just too little
amount for D3. 1GB of PC400 is better than 512 of PC533. Second, you're
oc'd
cpu is always suspect when it comes to PC shutdowns. Again, I don't care
what the third party "stability" apps say, that is where there is risk. I
have the same vid card as you and only a 2.26GHz northwood running stock
and
the game runs very smooth at medium(good enough for me). BUT, I have 1GB
of
PC3200 ram and that helps a lot. Take this for what's it's worth.

More memory coming soon...

As for the OC... It hasn't crashed since December and I couldn't get it to
crash at those settings regardless of what I threw at it. I have clocked it
back a bit today.

BTW, what is the BEST stress test for an OC?

On Thu, 19 Aug 2004 17:06:24 GMT, "Noozer" wrote:


BTW, what is the BEST stress test for an OC?


I've never used it but people say Prime95 is the best stress test.

http://www.mersenne.org/freesoft.htm#downloadPrime95

I have no idea.

"Noozer" wrote in message
news:l21Vc.142159$M95.128995@pd7tw1no...
More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory and
a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

Played Doom3 for a couple of hours last night without any issues. When I
exited the game the CPU temp was around 38'C so I don't think I was taxing
the machine. Opened my OE to read the newsgroups and about 10 minutes into
that the whole PC just shut off. Not fun.

I could not get the PC to start regardless of what I did. Pulled the power
cable and reattached. Disconnected hard drives. Pulled PCI cards. Reset
CMOS. PC would just not post - no beeps - anything. Once I pulled the
video
card I got the standard "no monitor" beeps. I dropped in a Geforce 2 card
I
have and the PC is working again. Put the ATI card back in and no more
boot.
***ARGH*** I killed my card!

I pulled the Artic Cooling VGA Cooler off the card (no shim). No damage I
can see. Compound right where it's supposed to be. No signs of
overheating.
I put the stock ATI cooler back on. Still dead. BAH!!! Left the PC in
pieces
for the night (it died 5 minutes before going to bed).

This morning I power up with the ATI card in and it works fine again. (Had
to reset the CMOS settings).

What happened? Nothing was overly warm when I was pulling it apart last
night. I didn't do anything this morning except turn the PC on.

In article l21Vc.142159$M95.128995@pd7tw1no, "Noozer"
wrote:

More of a comment then a question... but comments are welcome.

PC is an ASUS P4C800EDlx mainboard running 512meg of OCZ DDR533 memory and a
128meg ATI Radeon 9600XT. CPU is a 2.6 P4 @ 3.19 (stable in all benchmarks
since December - CPU never hits 50'C) Enermax 350w PSU.

Played Doom3 for a couple of hours last night without any issues. When I
exited the game the CPU temp was around 38'C so I don't think I was taxing
the machine. Opened my OE to read the newsgroups and about 10 minutes into
that the whole PC just shut off. Not fun.

I could not get the PC to start regardless of what I did. Pulled the power
cable and reattached. Disconnected hard drives. Pulled PCI cards. Reset
CMOS. PC would just not post - no beeps - anything. Once I pulled the video
card I got the standard "no monitor" beeps. I dropped in a Geforce 2 card I
have and the PC is working again. Put the ATI card back in and no more boot.
***ARGH*** I killed my card!

I pulled the Artic Cooling VGA Cooler off the card (no shim). No damage I
can see. Compound right where it's supposed to be. No signs of overheating.
I put the stock ATI cooler back on. Still dead. BAH!!! Left the PC in pieces
for the night (it died 5 minutes before going to bed).

This morning I power up with the ATI card in and it works fine again. (Had
to reset the CMOS settings).

What happened? Nothing was overly warm when I was pulling it apart last
night. I didn't do anything this morning except turn the PC on.

Maybe the shutdown was the PSU overheating ? Did it actually
shutdown or just attempt to reboot ? Shutdown could be the
PSU itself doing it, or the thermal protection on the motherboard.
If, on the other hand, the machine just rebooted (i.e. power was
not lost), then that is consistent with losing communications
with the video card, and the video driver taking a dump.

As for the lack of recovery of the video card, I have run into
a phenomenon like that once, but there is no matching root cause
in your case.

The GPU on the video card, could have gone into micro-latchup.
But, I cannot see any mechanism to trigger it, in your case. For
example, I've worked on a particular chip in the lab, where the
chip appeared to be dead. At the time, I only had 15 prototypes,
and couldn't afford for them to die. I turned off the power on
the hardware I was using, and the chip would not recover. When I
left the hardware unpowered overnight, the next morning the chip
ran as if nothing was wrong. The circuit had no protection
mechanisms, no overload protection etc. The root cause was bus
contention (two chips driving a bus at the same time), which is
what killed the chip. The consistent recovery of the chip, by
waiting overnight, is still a mystery. A phenomenon called
micro-latchup is the only thing that fits the symptoms, and
before the latchup will release, the power supply has to be
completely drained - something that may have occurred by waiting
overnight.

If this happens to the video card again, try shutting down the
PC and then unplug it. Wait 10 minutes and plug it in again.
Maybe that will be enough time for it to recover.

If you read some of the reviews and info on Doom3, it seems to
be quite stressful for the video card. It almost suggests that
video cards are not being tested well enough when they are
designed, to prevent surprises when a new video game is
introduced. (I.e. A program like Prime95 or CPUBurn needs
to be written for 3D graphics, by ATI or Nvidia, to make sure
their video card is designed well enough for any application
that gets used. There have been cases, where even processor
designs have internal noise problems, that can be triggered
by certain assembler language sequences, so ATI and Nvidia
are not alone in this. Any chip with millions of gates can
suffer the same fate.)

On my ATI9800, I've noticed that one inductor (coil wound on a
vertical cylinder) gets very hot. Using an after market cooler
for a video card, may prevent air circulation around such parts.
I recommend doing a touch test of the video card, whether you
plan on modding the heatsink or not. While running 3DMark or some
demo loop in a 3D application, run your hand over the video card
components and see if anything is hot enough to burn you.
Depending on how a third party heatsink is designed, it may
prevent air from circulating near the hot component, actually
making life for that component worse than with the original
heatsink. In my case, I placed a spare 80mm fan I had in the
adjacent slot to the video card, and the air from that fan keeps
the inductor at a comfortable temp no matter what app is running.

HTH,
Paul