Help needed - unreliable new built computer

[gregv]

I build a computer from parts recently, and have had no end of trouble with it. Initially one of the 2 1Gb DDR PC3200 has proved to be no good so I ditched it. The situation now is that I have thefollowing set-up;

Mobo - ABIT FATAL1TY AN8SLI
G/C - Nvidia GeForce 6600 - Gigabyte GV-NX66256DP
P/S - Jeantech JNP350 350W (ATX2.0)
RAM - 1Gb DDR PC3200 - not top quality RAM I must admit.
H/D - SAMSUNG SP2504C (SATA 250Gb)
C/D drive - LG DVDRW (can't remember the model exactly)
O/S XP s/p2
CPU - Athlon64 4000+
Floppy drive
Abit micro-GURU front panel device.

Symptoms are that when I start hitting the system heavily (e.g. games or anything that taxes the graphics it seems), then either the system reboots or windows shuts down the executable.

For example I switch on and fire up TOCA3, after 5 minutes to get a windows error message "RD3.exe has had to shut down", and it closes TOCA3. If I fire up TOCA3 immediately again the machine often just falls over reboots after a few seconds.

I'm fully up to date on the drivers and BIOS as far as I know.

I've attached some mini dumps which I can't interpret, if anyone can then I would be most grateful.

The remaining RAM runs thorugh memtest86 (or is it memcheck86) for hours without reporting any errors.

 

[paranoid guy]

That's quite a steup to be running with a 350w psu, can you give us the amps on the 12v rail? Also, seeing as the system shuts up when you run a graphics intensive game, can you check your graphics card's temperature? And you did connect the molex cable to the psu didn't you?(so many graphics card problems at the mo, have to ask)

 

[gregv]

Thanks for the reply,

The 12V rails deliver 15A & 18A, and I'm retty sure everything is plugged in as it should be.

The Graphics card isn't getting hot, I got a measurement of about 45 degrees C on it, so I don't think it is overheating. It's a PCIe card but it does not have a socket for the PCIe power supply, doesn't need it I spose.

Any idea how I could check whether the PSu is up to scratch (apart form binning it and getting a 600W one!), Speedfan (utility) measures voltages but doesn't show anything anomalous, if there was a PSU problem then I might expect to see a bit of voltage fluctuation over time, especially when hitting the system heavily, I don't see any such fluctuation though.

Thanks for your comments anyway, I do appreciate.

 

[iss]

your bios should have a hardware monitor capability. most these dasy monitor voltages. so try booting into the bios and checking the voltages reported in hardwae monitor or hardware health. or you could get a PSU tester for around 12.00. YOur PSU is underpowered for that setup but I would be more lnclined to think it is a ram problem. Nvidia chipsets are fairly sensetive when it comes to ram compaitability. one thing you could try is make sure that in the bios that the ram settings are set to standard and not more aggresive ram settings which decreases compaitability.

 

[gregv]

Thanks for the comment, had a look and voltages look pretty normal, RAM settings are normal too.

Is it possible that another brand of RAM would offer better compatability. If so anyone got any suggestions?

 

[gregv]

Chaps,

Whoever said "RAM problem" hit the nail on the head. I bit the bullet and order a couple of Gig of Corsair XMS and since they've been installed not a single failure.

Power supply probably a bit short of what is ideally required, should be OK now everything is shiny & new, but as components age it might fall short. I've got a 450Watt p/s that came free with a case once but I don't think I'd be better off with it to be honest, it's noisy & lightweight (I like something heavy when it comes to a p/s).

Anyway, thanks for your help.

Regards,
Greg Vaughan

 

[alidabiri]

glad to hear that. also, remember that this same thing can happen if your cpu temp goes up beyond threshhold safe level. the system will shut down. this is by design. so your cpu won't fry.