Sign in with
Sign up | Sign in
Your question

GPU not getting enough power? Something else broken?

Tags:
  • Gtx
  • GPUs
  • Power
  • Graphics
Last response: in Graphics & Displays
Share
April 16, 2013 9:42:09 AM

So I'm having some really weird problems. Basically I got a new GPU (GTX 660 ti) because my old one (gtx 470) died in a storm of artefacts.

Now, when I play games sometimes the monitor turns off or just shows a single faint color, like dark green or brown and I basically have to reset. No bluescreen though. I had it even with the old card that the screen turned off and the PC was basically in limbo until you restarted it by switch-off but it was rare.

Now since I have the new card it has gotten more frequent. When I check the event viewer it only says Kernel Power 41, which apparently just means the PC shutdown improperly.

At first, I was trying out GTA 4 on the new card. It ran for around 5 minutes, then the screen turned off and the PC was frozen. I tried Planetside 2. That crashed as soon as I took a few steps.

Then I read online it could be a sounddriver problem and to disable one sound device if it shows several. Okay I thought, lets try that. Tried Planetside 2 again, and it seemed to work. Then I tried Battlefield 3. Worked fine for 30 minutes. Then I tried Far Cry 3, and while that worked for around 15 minutes or so, I got the SAME THING again.

So then I did a clean gpu driver install. Got rid of the drivers with driversweeper in savemode. Restarted, reinstalled, tried Bioshock Infinite in Benchmark mode. Screen turned off after 30 seconds. Then I tried playing it, which worked fine for 5 minutes or so but then the screen turned off. So here is why I was thinking it might be a power problem: In this case, the screen turned off and you could hear the voices still but in a jumbled electric distortion that got deeper and then a bit higher again and then the game crashed and the screen turned back on, which led me to believe it caught itself at the last moment.

I tried Guild Wars 2, and when visiting the city I had a similar experience: Screen flickering and sound getting weirdly distorted, then catching itself again and running normally again.
Also, no crashes when not gaming.

Here are my specs:

- Windows 7 64 bit
- AMD Phenom x4 BE 965 @ 3.4ghz
- 16 GB DDR 3 Ram at 1333 Mhz
- GTX 660 ti
- Intel SSD
- 600 watt coolermaster PSU


I would really appreciate some help here.

More about : gpu power broken

April 16, 2013 9:48:15 AM

haa replace that psu.cm psus are not at all good
m
0
l
April 16, 2013 10:04:00 AM

first i would check if everything is up to date then remove all memory stick and leave only one then do a memtest on each sticks , it could from meroy ,gpu or psu issue .you also wrote that the old die in a storm of artefacts check your motherboard for any sign of bubble capacitor or anything else that could look abnormal ,i use this to see if there is any minidump files in my system http://www.resplendence.com/whocrashed some use this for blue screen viewer http://www.nirsoft.net/utils/blue_screen_view.html
m
0
l
Related resources
April 16, 2013 10:05:56 AM

That Coolermaster PSU isn't good for 600W. It probably wanders out of spec even on the minimal load you're giving it. I'd get something in the 500W-550W range that was made by Seasonic, FSP, or Superflower.
m
0
l
April 16, 2013 10:06:20 AM

You could also have damage to the motherboard. Switch the PSU first, if it still happens it is likely that you got damage to a capacitor somewhere on the board.
m
0
l
April 16, 2013 10:21:04 AM

Is the only way of making sure that it's the PSU's fault by switching it? The only PSU I have here is max 420 watt, which won't be enough for this card. (Says at least 450 watt on the box).

Tried the WhoCrashed program, the last dump file is from when my first graphics card died around 2 weeks ago, it shows a VIDEO_TDR_ERROR with bugcheck code 0x116. The 470 that died with artifacts also only started with minimal resolution and was not recognized by the system anymore.

Now with the new one, when the screen turned off there were no bluescreens or minidumps, but neither were there when I had the same thing with the 470.

I will test the RAM now, but having run Prime95 for 4 hours now has not produced a crash yet.
m
0
l
April 16, 2013 10:45:49 AM

Yeah I'm fairly sure the old card was just toasted so right now I'm just trying to figure out why my new one doesn't want to work either. (Or what is keeping it from working rather) Since everyone says the powersupply sucks it seems likely that that might be it, doesn't it? Are the symptoms I described something that would happen when the PSU can't keep up with demand?

I would also appreciate a recommendation for a good PSU that is capable of handling the card (GTX 660 ti) and some extras like harddrives and fans and isn't too expensive, I am really bled dry on the budget side.
m
0
l
April 16, 2013 11:26:42 AM

If your PSU is wandering out of spec under stress, it could indeed cause crashes.
About the cheapest you'll find that's good enough and big enough would be a Rosewill Capstone, Hive, or even an older 80+ "RG" unit. All have received favorable competent technical reviews. The Antec BP-550 is another inexpensive, well-reviewed unit.
m
0
l
April 16, 2013 12:30:37 PM

With what Watt number? Also can't find a Rosewill Capstone for sale in my country, neither one from Hive. Not sure what you mean with 80+ RG unit. Sorry. :/ 

m
0
l
April 16, 2013 12:41:33 PM

www.80plus.org tests PSUs for efficiency. They do it at the artificially low temperature of 23C, so your actual efficiency will be lower at 35C-40C, but they run PSUs at 100% of their label, which would kill cheap junk. The RG-series is another line from Rosewill, but may not be in your country either. While it's no guarantee of quality, an 80+ designation (preferably Bronze or better) will have at least weeded out most of the junk. Check their site to see which of the PSUs available to you are actually listed.
A good 500W PSU is sufficient for your PC; i.e. one that is actually capable of producing 500W, and doing it cleanly.
m
0
l
April 16, 2013 12:44:37 PM

link us to a store where you could buy a psu where you live .
m
0
l
April 16, 2013 1:33:08 PM

i seen some good on the hardwareversand what is your budget .
m
0
l
April 16, 2013 3:15:53 PM

I would like to stay well under 80 Euro if possible, better somewhere around 60. My wallet is absolutely ravaged for the month.
m
0
l
April 16, 2013 7:35:40 PM

I have a couple of the tx750's and 1 of the hx750's from corsair. Never had a single problem with either. Good units that run stable and quiet.
m
0
l
April 17, 2013 7:42:04 AM

Does it need to be 650-750 watt though? And do they use more energy than one with a lower watt number, or is that just what they can provide should it be needed by the system?

Also scout, what do you mean with the last part where you say "and your card specs who need to pci-e connector"?
m
0
l
April 17, 2013 7:44:09 AM

no you just need good 500w psu

i suggest seasonic 520w psu
m
0
l
April 17, 2013 9:08:30 AM

A PSU is most efficient when providing 40%-60% of its actual rated capacity, but it will only pull the amount of power it needs, based on the demands of the attached equipment.
m
0
l
April 17, 2013 1:53:40 PM

+1 to what onus says. Only uses what you demand from it but try to keep it around half. It'll save you money in the long run paying a couple dollars more for something a tad larger. Also it will run quieter being not as large of load on it and last longer because not as large of a load.
m
0
l
April 17, 2013 7:16:59 PM

(to ) was a wistake error it means your card need 2 power cable from psu with 6 pins each .
m
0
l
April 23, 2013 7:29:45 AM

After I now ordered a replacement PSU I had to find out: It was NOT the PSU. The problem is still occurring and I am getting a bit desperate. The RAM isn't the problem either, it passes all tests and switching out either one of the modules in either one of the 4 slots does not help. The card works just fine on my other PC.

I now managed to narrow down the problem, when the PC didn't freeze but instead crashed the running game. The error is: "Display driver nvlddmkm stopped responding and has successfully recovered."

Another event before this one in the event viewer is: The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.

If the event originated on another computer, the display information had to be saved with the event.

The following information was included with the event:

\Device\Video7
An uncorrectable double bit error (DBE) has been detected on GPU (02 02 00).


Does anyone know what this means?
m
0
l
April 23, 2013 8:05:59 AM

Remove ALL of your video drivers. Use CCleaner to make sure they're gone. Reboot, then re-install the latest drivers for your card.
m
0
l
April 23, 2013 8:57:19 AM

Did that several times now, with different drivers. Helped nothing. I put the card in an old PC of mine which has not crashed so far (far longer than it took in the other one to crash so I think it runs just fine in the other PC.) However, I'm using the same HDDs and same windows installation.
m
0
l
April 23, 2013 9:03:46 AM

Same HDDs and same Windows? On a different mobo? Your drivers could be COMPLETELY hosed. Although it is sometimes possible to move a Windows installation from one mobo to another [similar] one, it usually fails outright, and when it doesn't, in general it will never be "right."
m
0
l
April 23, 2013 9:57:36 AM

Yeah I just hooked up the same hdds to the other mobo, it didn't install anything besides the GPU drivers. (The other mobo is an older version of my main-mobo so that might be why it works so well).

I have now several times uninstalled both audio and graphics drivers on the computer where it fails, deleted the stuff left from the drivers via driver sweeper in safe mode, and reinstalled them again. I did all windows updates, tried it with having windows updates disabled and installing manually after having uninstalled cleanly, checked and monitored temps, tried it with the onboard sound disabled in the BIOS, loaded opimized defaults in BIOS, loaded fail save defaults in BIOS, tested the RAM, switched the card to the second PCI slot on the mobo, changed the PSU, tried it with only the main hdd, did an sfc/scannow just for good measure, ran virus and malware scans in safe mode which both found nothing, disabled/enabled nvidia HD audio drivers, set the system to "Maximum performance" in the power saver settings, nothing works.

When the first card broke and I used an old ATI 4200 card for 2 weeks I had no problem of this sort with the computer on which it now doesn't work. These problems occurred the exact same time when I put this card in. (No ATI drivers are left on the system either, uninstalled those in safe mode and driver sweeper'd those as well)

So I would think the card was at fault, if it wouldn't work in the other computer/mainboard just fine.


Can this still be the fault of the card? Or is that not possible, since it seems to work on the other (old) mainboard?
m
0
l
April 23, 2013 5:42:51 PM

i would rma that graphic and if you still have issue with the new card i would carefully check the the motherboard and the pci-e slot on this board,what exactly happens to the ati card and did you use memtest for your memory test one stick at the time ?
m
0
l
April 23, 2013 11:37:55 PM

Yes i tested the memory one stick at a time. Nothing happens to that Ati card, it works fine without any crashes. The new one however crashes in both pci-e slots, so I don't think the slot is faulty. I also have not found anything on the mainboard that looks like something is wrong with it, no damaged capacitors or blisters or anything that looks suspicious.
m
0
l
April 24, 2013 8:00:42 PM

Nothing happens to that Ati card, this means you use a ati card remove the driver from ati then put hte new gtx card in install the driver there is could be conflict with the ati driver and the gtx card.
m
0
l
May 2, 2013 6:23:25 AM

As I said above, I used the ATI card before I got this one and had no problems with it. I already did a complete driver uninstall and went over it with driver sweeper in save mode.

I just now got a new card back from the shop and it is still the same problem. Power supply I already changed as well. I also just installed the nvidia beta drivers 320.00. Didn't help .
m
0
l
May 2, 2013 7:47:11 AM

One thing you might try is re-installing DirectX. Many games include an installer for it, and I think you can get it from Microsoft's web site.
Try a complete re-install of Windows and a sample game or two (use a spare HDD if you want to just test this; if no improvement you can just put your old disk back in). You're using a different motherboard, so there will be differences.
m
0
l
May 20, 2013 9:19:54 AM

It turns out it's neither the GPU nor the power supply. The issue persists. I switched out the PSU for a new one, switched out the card for a new one, sent my processor in for warranty and reinstalled windows fresh to make sure it is not messed up drivers acting up. The only thing I haven't switched out yet is the motherboard.

Now I was thinking I should probably buy a new one. Can someone help me with that? I'd like to get a not too expensive one for this Phenom x4 965 Black Edition. If possible I'd also like to overclock it so I get some more performance out of it until I upgrade to Haswell. But I guess that clashes with the "not too expensive" part.

I'd appreciate if someone could help me once again. Thanks for all the advice so far.
m
0
l
!