Hardware malfunction... Graphics card or power supply?

[blarney]

In the last few days my desktop has become increasingly hard to boot up. The first signs that something was amiss was when it hung whilst in my Debian partition. I'd just installed drivers for my Nvidia 8600GT so I suspected I'd messed up somewhere. I booted back in to Pclinuxos kde 2010 (Fully upgraded) and everything seemed fine.... Until it too hung on me.

From then on, it started to play up in the boot process. Sometimes getting to GRUB and halting there, other times I could hear it booting but nothing came up on the monitor (Light didnt turn green) and I had to ALT-SysRQ- REISUB. Sometimes that worked and it booted up properly but then even that stopped working.

Normally, when I hit the power on button it turns blue and immediately starts up but now it seemed to take ages to even get to the blue light, like it was struggling to find enough power.

So, I pulled the graphics card and Hey Presto! it booted fine with the onboard Intel graphics... But I still don't know whether the Nvidia card is malfunctioning or the power supply is about to fail? If I continue with a failing psu I'm guessing it could effect other components?

Cheers,
Blarney

[OMSkates]

Maybe your graphics card needs a new cable?

[menotu]

Maybe your graphics card needs a new cable?

Have you tried reseating the nvidia card again and rebooting?

[blarney]

Thanks for all the suggestions. When I get home tonight I'll
give them a go... 

I believe you can also check voltages in some BIOSes? A poor
power supply could play hell with the M/B I'm guessing...


cheers,
Blarney

[exploder]

Checking the voltages in the BIOS is a good idea and it should help you figure out where the problem is.

[John Bee]

Hi;

 If your video card uses a hd power connector for additional power, make sure it is the only thing using that power cable/connector. They don't share well at all with other peripherals.

 

[ThirdOfSix]

I had similar odd behavior for quite a while on one or two of my machines that had large fan-less heat sinks on the display card.

This occurred at a time when I was experimenting with multiple monitors and display cards etc.

Sometimes one of the cards outputs would send weird display frequencies that caused the monitors to appear defective.

I did a little bit of testing and comparing using a non-contact thermometer.

It turned out that over time, the heat-sink compound had hardened and then as I inserted and removed the cards multiple times, I caused the hardened compound to break loose from either the chip or the heat-sink. This resulted in an air gap of varying thickness depending on when and where I last touched the heat sink.

This caused different areas of the chip to run at different temperatures as I changed cards around. Because the thickness and location of this gap would change as I swapped cards around, the symptoms would also change as different portions of the chip overheated.

The real problem is that some of these cards just use some simple plastic clips to hold the heat-sink in place and it takes very little to cause it to move.

In the end, the fix was to simply remove the heat sink and clean both surfaces and then re-install with fresh heat sink compound.

That card has been working fine for many months after the fix with no further problems.

If by applying a gentle twisting motion to the heat sink, said heat sink moves, you have probably found your problem. But if you apply too much force and if that was not your original problem...it is now.

Good luck.