Why are my Graphics cards failing so often ?

[aphexcoil]

Also I am getting minor shock when touching the chassis of the CPU. If I turn off the power of all monitors & keep CPU power on, then I don't get any shock. So definitely monitors are causing this problem.

Wait. You get a constant shock if you touch the chassic of the CPU? Like a constant tingling sensation? Call an electrician if that's the case. Sounds like you could have some sort of fucked up floating ground or god knows what.

Actually, I remember something like this happening to a friend. Usually the case itself is grounded to the ground portion of the plug. That's usually through the outside of the power supply which is in contact with the case itself. If the ground wire went bad from the outlet, you could have a floating differential between true ground and what ground your case is at. This would definitely be bad for electronics.

 

[d08]

Depending on how hard you're hitting the graphics card it may be overheating. I've run into issues where I upgraded the graphics card but the airflow in the main box was insufficient. The only way I could keep it safely cooled was to take off the side panel.

In my case the GPU didn't outright die but would go crazy after a while.

Quite unlikely a card would overheat in 2D usage or even mild 3D like modern Windows.
My guess would be the power supply. PSUs lose power as they age and you might be overloading it constantly, depending on how many and what type on devices you have, it can cause similar issues to yours.

 

[d08]

Wait. You get a constant shock if you touch the chassic of the CPU? Like a constant tingling sensation? Call an electrician if that's the case. Sounds like you could have some sort of fucked up floating ground or god knows what.

Actually, I remember something like this happening to a friend. Usually the case itself is grounded to the ground portion of the plug. That's usually through the outside of the power supply which is in contact with the case itself. If the ground wire went bad from the outlet, you could have a floating differential between true ground and what ground your case is at. This would definitely be bad for electronics.

This is quite common, I get a small shock from my ThinkPad if I brush against some of the side ports. It has caused no stability issues.

 

[stevegee58]

Quite unlikely a card would overheat in 2D usage or even mild 3D like modern Windows.
My guess would be the power supply. PSUs lose power as they age and you might be overloading it constantly, depending on how many and what type on devices you have, it can cause similar issues to yours.

You're right about normal usage. I was using it for parallel processing development which is definitely not normal can can make it run hot.

 

[aphexcoil]

This is quite common, I get a small shock from my ThinkPad if I brush against some of the side ports. It has caused no stability issues.

There's a difference between a static shock and a 60Hz tingling sensation, though. I'm wondering which he meant.

 

[Sig]

It's actually not unusual in your situation to have 2 cards fail at the same time. It sounds like you have identical cards and identical monitor setups and used both cards with the same loading for the same amount of time. If the standard deviation of their mean time between failure is low, you would expect them to both fail at around the same time. Unfortunately the mean time between failure on this particular card apparently happens to be 4 years. This actually means they were manufactured to very close tolerances, ironically, in that both were nearly identical which is harder in manufacturing than you think.

 

[aphexcoil]

If the standard deviation of their mean time between failure is low, you would expect them to both fail at around the same time.

If they rolled off the same assembly line / batch -- I'd say yeah. But if they failed near the EXACT same time, I'd be looking at power surges, static charges, etc.

One thing you don't ever want is the cables from the back of the VGA card getting near carpet. I actually had one blow because there was a high static charge in the carpet and it wrecked the card. Keep them tied up and neat.

 

[Scataphagos]

I am not getting shock with this card & 4 monitors connected to them. So that rules out those 4 monitors & cables.

Don't know how adept you are at troubleshooting, so I offer this...

1. That you have one card and 4 monitors checking out OK is a start.
2. Put 4 other monitors on that card and see how that is.
3. If OK, the last 4 monitors. If all three setups are good, that tells you that the one card, in that PCIE slot, and the monitors themselves are not the problem.

You'll have to do the same with the other PCIE slots.

Then test the whole batch again on the 3rd PCIE slot.

Then, same thing again with 2 of the 3 slots. Then again with the other slots... testing slots 1+2, 1+3, and 2+3.

If you've got a card in 4th PCIE slot, remove it.

That's a lot of fussing around. Hopefully you'll hit on it early in the process.

(I once spent a month troubleshooting a problem. 1st thing I did was run a memory test, which said, "RAM OK". I replaced everything and swapped things around until I was at wit's end. Even had Dell Tech Support send me some different video cards and cables. Eventually got it corralled. It was the RAM that was at fault after all even though the "RAM test" said it was good. Go figure. Troubleshooting can be a real pain sometimes.)

 

[Stocktracker]

Try the following;
Disconnect the card cables> touch the chassis> Do you feel any shock> if the answer is no then> connect the cables without the monitors> touch the chassis again and see if you feel anything? If Not then connect the monitors and touch the chassis again if you feel the shock again then it is your monitors. Exchange your monitors with different ones and touch the chassis again. You see
If you are leaving the Air conditioning running 24/7 then you might be dealing with some moisture built up and what comes with it.

Good luck

 

[qqq]

Depending on how hard you're hitting the graphics card it may be overheating. I've run into issues where I upgraded the graphics card but the airflow in the main box was insufficient. The only way I could keep it safely cooled was to take off the side panel.

In my case the GPU didn't outright die but would go crazy after a while.

My office is air-conditioned 24/7 & air from conditioner is hitting the back of CPU so cooling is sufficient.

The cards are not overloaded at all either.