4

Something in my computer is damaging any graphics card that is put into the computer. I would like to know what components have the capability of damaging a graphics card, and how to determine which component is the culprit.

Some Specifics:

  1. I know the graphics card isn't the problem, as multiple have been fried by my system.
  2. I don't think that the card is underpowered, but you can look at the list of parts below and be the judge of that yourself.
  3. The computer doesn't boot or post, and the motherboard instantly displays a VGA error light.
  4. Cards that don't need to be plugged into the PSU (such as my ancient Radeon 4650) don't get damaged, although it could also be that the system is underpowered and that the 4650 doesn't draw enough power to be a problem.
  5. After the card is damaged, it won't work in any other system (tried with 2 others).
  6. I had just installed a new Corsair H100i CPU Liquid Cooling system. Before that installation I experienced no problems.
  7. When installing the H100i (Corsair Link), my mobo didn't have enough USB connectors, so I had to unplug the cable leading to the USB 2.0 ports in the front of the case.
  8. After running Eyefinity for an hour and then disabling it, the center monitor displayed "input not supported". When I turned the computer off and attempted to restart, that's when the above problems happened.
  9. Resetting the BIOS settings didn't fix anything (had to access them through the 4650 mentioned above), although I didn't update the BIOS when an update was available.
  10. Putting the GPU in another PCI-E slot yields the same results.
  11. The 2 fans on the GPU spin properly.
  12. Turning off the computer once it displays the error requires holding down the power button for ~3 seconds.

Also, one more detail that I think is very important: My computer will constantly show me random voltage warnings that make no sense. I assumed the program that displays the errors is bugged (it would show things like the motherboard being at -1 degree centigrade, the CPU voltage being 0.0v, etc). Now I"m not so sure that it's a software problem causing those misreadings.

Parts

  1. Intel 3770k OC'd @ 4.2ghz
  2. 2 x 8gb Corsair Vengeance @ Stock 1866mhz
  3. Radeon HD 7970 3gb Sapphire Vapor-X
  4. 120gb Kingston Hyper-X SSD
  5. 1tb HGST 7200rpm HDD
  6. 850w Platinum Rated Modular OCZ PSU
  7. Corsair H100i Closed Loop Liquid Cooling
  8. 6 x Blue Cold Cathode Case Lights
  9. 4 x 240mm Blue LED Case Fans
  10. 3 x 140mm Blue LED Case Fans
  11. ASUS DVD/CD Burner/Reader
  12. Windows 7 64-bit

If you need more information, please let me know.

  • 1
    The PSU or motherboard can cause a short-circuit fault to damage the GPU – magicandre1981 Jan 22 '15 at 16:39
  • @magicandre1981 Any way to check which without purchasing a voltmeter? – user3144349 Jan 22 '15 at 16:46
  • @user3144349 - You can go out and replace the PSU and Motherboard I suppose. Even if there is a short it will be extremely difficult to find since the short doesn't always exist (intermediate short). Yes; I know that's not the exact word I am looking for.. – Ramhound Jan 22 '15 at 16:52
  • @Ramhound That's... Unfortunate. However, the 4650 was fine (used it for about a month while waiting for the RMA) and was not plugged directly into the PSU. Would that potentially mean that there is a larger chance of the PSU being the problem rather than the motherboard? – user3144349 Jan 22 '15 at 17:07
  • 1
    Honestly. If its either the motherboard or PSU that is causing this. Its likely around a 50/50 chance of it either being one or the other. If everything is under warranty pick with enough time to do both, pick the one, that will result in the fastest turn around time. The fact it only happens when you need to use a PCI-E power cable might sway that to a 60/40 chance in favor of the power supply though. If there is a short on the motherboard, you are better off just trying to RMA it, since obviously that can't be fixed easily and without considerable amount of soldering skills and docs. – Ramhound Jan 22 '15 at 17:22
0

An unstable/badly wired power supply and/or power supply connections can damage anything.

You mentioned that only video cards plugged straight into the PSU get damaged.

From what you posted, it definitely looks like a power system error. It's likely that your video card is mistakenly getting too much power. At the very least, I would recommend tossing out your computer's power supply before it damages anything else. This might sound expensive but then again, so is throwing good video cards away on a bad computer.

My guess is that the liquid cooling installation might have destabilized the power supply or you might have damaged it while installing the cooling.

Even if your video card is fried, the fans can continue working anyway. "Fine circuits" are easy to damage, fans are relatively tough. I wouldn't be surprised if giving a fan too much power just caused it to spin even more than normal.

In my experience, video cards are the only thing on any of my computers that ever break even when I never did anything that could damage them.

  • My first Macbook Pro (2008 type) had a video card defect that caused it to burn out, resulting in a black screen (no display) until I got a new logic board (under warranty).
  • My next Macbook Pro (2010 type) had a similar video card defect after around 1.5 years of use. I was able to get a new logic board under warranty, however the new logic board's video card burned out within 40 days (for me, that's the record shortest amount of time it has taken for any of my electronics to break). I got yet another logic board and it works fine now.
  • A video card in one of my old desktops burned out and I ended up putting an antiquated "better than nothing" video card in there instead.

I specifically asked for a heavy-duty Intel CPU and integrated graphics in my newest laptop. No more video card burnouts/defects/other crap to put up with. Good riddance.

-1

A few things are likely to damage a card: overclocking, under-cooling or a bad power supply. Since you state "I had just installed a new Corsair H100i Liquid Cooling system. Before that installation I experienced no problems," it would seem that the cooler is defective or improperly installed. Check the temperature of the GPU: http://www.computerhope.com/issues/ch001504.htm. Also use an accurate meter to check the voltages of the power supply and check for AC leakage due to a bad capacitor; see http://www.pctools.com/security-news/faulty-capacitor-test/.

  • The new cooling system is for my CPU, and is working exceedingly well (25 degrees idle, 50-60 degrees load). The GPU temps are also standard (35 degrees idle, didn't check temps under load unfortunately). I will also check the voltage of the PSU shortly, I need to buy a voltmeter first though. – user3144349 Jan 22 '15 at 16:45
  • 1
    This answer seems to be focused on the wrong information. An after market CPU heatsink isn't going to cause this problem with a graphics card. The user seems to have taken the proper diagnostic steps so far. Furthermore it really doesn't attempt to explain how the user can solve his question besides using a voltmeter. – Ramhound Jan 22 '15 at 16:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.