Save my marriage! My wife's possessed PC keeps restarting - related to new graphics card?

Chedster

Distinguished
Sep 7, 2015
19
0
18,510
Hi all,

I upgraded my wife's PC to Windows 10 last month, and then found out her graphics card had no drivers for Windows 10. Being an old relic, I decided to replace it with a new PNY Geforce GT 720 card.

The PC now randomly restarts without warning, and having lost her work mid-flow, my marriage is seriously in danger unless I get this sorted soon! :ouch: There is no determined time the PC decides to power down - just randomly.

I have wiped all traces of the old graphics card and reinstalled the new drivers again. This morning, I cleaned up the dust from the inside of the case and the components, removed the old thermal paste from the processor and put some fresh on (the CPU temp was idling at 65 degrees when I was in the bios, which I figured was quite hot).

The motherboard is the MSI P55-GD80, with the latest bios updates.

Any suggestions
 
I agree, Sounds like the power supply is not supplying enough power to keep the system running. Did you make sure to check that the power supply will have enough power to run the new GT 720 before you installed it?
 


The PSU was fine though until I replaced the graphics card :??:
 


The PSU is the Corsair VX550 (550W) so it should easily power the new card, surely?
 


Could still be faulty, also might be the motherboard itself.
 
Have you checked to see what your temps climb up to when it is under load? gergguy is right, 65 at idle is very warm. Could very well be shutting off due to cpu overheating if the temps are getting high enough.
 


Because the old power supply most likely won't be able to handle the power loads drawn by the new graphics card. It's a simple fix when you buy a new power supply.
 


2015_09_07_17_54_36.jpg


Cleaning the system and replacing the thermal paste seemed to bring the CPU temp down nicely, but the PC still restarts at random (not necessarily when being used, either!)
 
Well if you are not overheating, and the problem still persists, it is almost a guarantee that after adding the new GPU your system is trying to pull more power than your PSU can handle.

What power supply do you have in the machine?
 


It would seem so, I agree, but I have the Corsair VX550 (550W) PSU installed. There should be easily enough power to go around :??:
 
Well a 550 PSU would be enough to power the system. Problem is that the corsair VX series is(was) their lowest end of value series PSU (I believed they have discontinued them, now the CX series is the lowest.) Not a single person here would recommend even a CX series psu to anyone. The inner components are low quality to allow for cheap price, but when it comes to PSU you really don't want to cheap out on it. You get what you pay for.

I would get a higher quality PSU, http://www.newegg.com/Product/Product.aspx?Item=N82E16817151093&cm_re=seasonic_m12ii-_-17-151-093-_-Product
 


http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=62

Actually the VX550 was rated pretty highly, all Japanese caps and good ripple. I don't see any reason why that GPU would be pulling too much from the PSU. It has 41A on the 12V rail which means it can push out 492W to that rail, more than enough for a GT720 and that setup.

I can only assume the PSU is possibly failing, no 100% here, or that the GPU is causing a short since, well the issue started with the new GPU, not before. So that would seem to me to say that it has to do with the GPU.

Have you tried a different GPU just for laughs? It is possible the GPU itself is bad and the bad component could be causing short which could in theory cause a system to power down if it over draws fro the PSU.

It could also be the motherboard. Possibly an incompatibility between the board and the GPU. While PCIe is backwards and forwards compatible, for the most part from 1.0 to the current 3.1 and soon 4.0 there was a bit of an issue between 1.0 and 2.1 since they upped the amount of power that can be pulled from the slot itself making it easier to put mid end GPUs in without the need for any additional PCIe connections from the PSU. That board however has PCIe 2.0 so that is probably not the issue.

Only two things to test easily is PSU and GPU. Try new of both just to be sure but make sure the shop/etailer has a good return policy in case you need to return it.
 


Oh I think you are right, I was confusing his VX series with the VS series.
 
I see that you "wiped all traces of the previous drivers", but if you did not also remove the registry entries associated with those drivers, and have not turned off the windows automatic updating of the drivers, it could still be a driver issue. I can agree that on an older power supply like that, the demands of the newer card, while quite low, could easily be enough to push a borderline unit over the edge. Especially since it's rather old and has likely degraded over time.

I'd try these options first, and then replace the power supply if that fails to resolve the issue.

http://www.tomshardware.com/faq/id-2767677/clean-graphics-driver-install-windows.html


http://www.tomshardware.com/faq/id-2763685/stop-windows-automatically-updating-device-drivers.html


I'd also probably reseat both the GPU card and RAM, just to be sure. It's always possible you got a faulty card too.
 
This is a Windows 10 computer with an NVIDIA card. There have been a number of driver issues with Windows 10 and NVIDIA drivers, and I suspect even if the drivers are updated that they are still flaky. That is where I would point the finger as the fact this started with a Windows 10 upgrade.
 
Both AMD and Nvidia products have had nothing but issues with driver structure and problems with the system automatically changing drivers to different versions since the release of Windows 10. Clean installations of the operating system or GPU card drivers, or both, have been fairly successful in mitigating these problems.
 


Hi Jimmy,

Thanks for your reply. The issue is still continuing, so I'm going to try seating the GPU in another socket and cleaning the system of the drivers and starting again.

Fortunately, the GPU is from Amazon, and they have a great returns policy :)

Will keep you updated!
 


Hi,

Thanks for your suggestion. I had actually followed this advice from a suggestion on another thread, yes.

But I did it again last night just in case (including inserting the GPU in another PCI-E slot)... the restarting problem still persists 🙁
 


Hmmmm, it's looking that might be my next step, I guess.

You know what's strange though? This exact same problem has been reported elsewhere online. Look at this (a review for the same GPU as this one):

2015_09_08_21_30_34.jpg
 


Now that is very interesting. I wonder if this could be but a small sampling of a very bad batch of cards?

I wonder if this would not apply to PNY but to all brands who distribute GT720s? Is this just limited to PNY or does it apply to EVGA, Asus, Gigabyte, etc? And if it does, does this mean all GT720s are bad cards or just those limited to certain manufacturers?

I guess what I'm trying to say here is that if you purchase another GT720 from another manufacturer and get the same result then the problem would lie with the GT720, not with the rest of the system. Because after process of elimination the problem most definitely is not the motherboard or the power supply, and your Windows installation seemed to be working fine before the new card.