How to diagnose why PC is freezing in 3D Application?

Recently I'm having some problems with my PC. I I'm running applications that depends on 3D graphics card it freezes after some time or throws a blue screen (mostly it just freezes). It's hard to tell when this problems happens but it usually starts 10-30 minutes after starting the game.

It's not the matter of application because I have experienced that in almost every 3D application. I tried on this computer (and that was quite a few). I also checked if the processor it's not overheating but the temperature was normal (around 60 degrees Celsius), I don't have a thermal sensor on graphics card but touching it's radiators didn't make my hard burn as well.

I suspect that it can be the problem with graphics card but I don't want to spend 200 euros on new graphics just to make sure previous one was ok.

Does anyone have any thoughts how I could find what the problems is? How to check if specific parts of my system are working correctly and find the faulty one. Or maybe it's a software problem that I could solve somehow.

I might add that I had similar problems with Windows XP 32 bit that I was running before. I just thought it's a problem with one-year-old Windows with lots of unnecessary garbage on drive C. Now I'm running clean Windows 7 and am expieriencing the same issue. Also it never happend when browsing the internet or reading emails. It seems to be triggered by some 3d-graphics dependent thing.

As for my setup:

• Core 2 Duo E8400 (3GHz)
• ASUS P5QL-E, P43
• 6GB RAM (2x2GB + 1x1GB, all same-series Kingston HyperX 1033MHz DDR2 modules)
• no-name (or I don't remember the name) GeForce 9800 GT 512 MB
• Windows 7 64-bit (system operates on RAID0 matrix)
• latest drivers for all the software installed (I update graphics drivers frequently)

All parts of the computer still are on a warranty but I got them from several places so sending it to a service and having them to deal with it is not an option I'm afraid.

Update

This is what the BlueScreenView gives about crashes.

==================================================
Dump File         : 101209-21356-01.dmp
Crash Time        : 2009-10-12 14:42:17
Bug Check String  : SYSTEM_SERVICE_EXCEPTION
Bug Check Code    : 0x0000003b
Parameter 1       : 00000000`c0000005
Parameter 2       : fffff800`02add105
Parameter 3       : fffff880`0b634cc0
Parameter 4       : 00000000`00000000
Caused By Driver  : ntoskrnl.exe
Caused By Address : ntoskrnl.exe+71f00
File Description  : NT Kernel & System
Product Name      : Microsoft® Windows® Operating System
Company           : Microsoft Corporation
File Version      : 6.1.7600.16385 (win7_rtm.090713-1255)
Processor         : x64
==================================================

One of the things many people think of as last resort is the power supply.

Your power supply has an output of say 550Watt, but thats only to handle peaks. At an average the output is usually a lot lower.

Having said this, my first guess would still be Graphics Card and second the power supply. I thought NVidia graphic cards show their temperature in Extended Graphics Settings (although I'm not sure they all do).

Usually RAM is also one of the main blue-screen trigger hardware parts. I had the behavior you described once when I bought a new module. This would be the easiest check. Try different settings (first 2x1GB, then 2x2GB, only 1x2GB).