Tech help/tools needed

Shougeki

I gott ago fix a computer this weekend.
Here is the problem:
System seems to work fine for everything BUT games. In a game, and i believe its ANY game, the machine reboots after about 5 mins.
SPECs:
Asus P4C-800E Delux Board (800Mhz FSB, 400Mhz Ram ect)
Geforce 6800
400W PSU
1Gig 400mhz DDR Ram
3Ghz 800Mhz Intel Proc

Now my first thoughts here is the PSU cant handle the load.

Any other ideas?

I want to run stress tests on the specific parts of the system also, so i need ideas for:
1. Proc stress test
2. Mem Stress test
3. HDD Stress test (thinking iomonkey is prob best)
4. GPU stress test
5. Is there anything that can test power drain ect?

B

seamus

Something overheating would be my first guess, possibly the RAM.

Can you get your hands on some System Monitor software for the Asus board, and watch the figures as you run the stress tests.

As you say, it may also be possible that the PSU isn't sufficient.

miju

i'd be with you on that one seamus i reckon it's something overheating altough i'd say it would be more than likely the processor overheating than the RAM

i've never heard of a system rebooting over RAM overheating but i've seen a fair few rebooting over processors as most systems are designed to cut out immediatly if the processor overheats AFAIK

is there any vibration coming from the machine that could be causing the fan connection to come loose that could also be another cause and would relate back to the point above.

Shougeki

heh heh

thats my second guess.

Is there any temp guages or anything that i can get for testing RAM and CPU temps? Something that i can put inside the case or run a cable intothe PC ?

B

stevenmu

Yep, really sound like over heating, the Free Software Thread has links for prime95 and 3dmark to stree test CPU and video (altough 3dmark will stress cpu and memory some too), you can also use Memtest86 to stress test the memory.

stevenmu

Shougeki wrote:

Is there any temp guages or anything that i can get for testing RAM and CPU temps? Something that i can put inside the case or run a cable intothe PC ?
B

They're probably built in already, download motherboard monitor (in the free software sticky aswell) and that will probably give you the readings, it supports most motherboards.

the_syco

Ensure that the heatsink is properly fixed onto the CPU.
Ensure that there is space between the graphics card, and the next slot.
Bring directx_9c_redist.exe.

Shougeki

Ok been lookin at this all weekend.

The machines are no longer freezing, BUT when playing games, after a while the graphics go screwy, or the game will freeze for 30 sec or so.
This happens in Battlefield/CS Source (only games ive tried).
The walls in CS Source get messed up. Somethime the sceen does what i can only descibe as "No-Clipping" cheat from the original doom series.

Now here are the machine specs:
Asus P4C-800-E Delux Board
Sata HDD
MSI Geforce 6900GT (256 GPU)
1Gig Crucial 2700 Ram (333Mhz)
400W PSU, with 17a across the 12V Rail.

Now im thinking that
a) Crucial RAM in the first place is dodgy. 2700/333Mhz on a 400Mhz Board is a wrong choice also. I think Kingston is prob a better choice.

b) The PSU should have 22 or 24A across the 12V Rail. Any recommendations on make/model ?

BTW running latest NVIDIA, or older versions dont seem to make much of a difference. Also when installing ne wGPU drivers, is it recommended to re-install DX9 or have DX9 installed before hand?

Any other ideas?

Thanx

RicardoSmith

A P4 will throttle back when it overheats. I've never heard of ram causing these kinda problems. Sound more like the graphics card is faulty.

Webmonkey

Same as above:

Overheating Problems - Processor (Try resitting the heatsink, bring thermal compond)
Dodgy RAM - Bring Memtest, its bootable
Faulty PSU - As said above aswell, none sufficient PSU

After that it gets a little harder to troubleshoot. Bring a spare PSU with you

RicardoSmith

Run memtest, its easy to do. Next I'd try a different gfx card.

WHere are you BTW?

stevenmu

It could be the GPU itself overheating, the nvidia drivers should be able to give you a temp reading in the advanced properties. Also, some makers overclock by default, maybe check what the standard speeds should be and use RivaTuner or something to make sure it's not running over them.