Please help identifying if video card or PSU is faulty

[Mogar]

Hi all,

I've been running my PC after an upgrade for 2 months (changed CPU, mobo and memory, but kept the video card, psu and hdd). Today it started to show an error and lockups:

After some minutes running a 3d application, it appears on screen some artifacts in shape of squares on a repeated pattern, and then the computer locks. Sometimes I have to press the reset button. The squares have always the same size and position on screen, but varies in colors with the background image. This appears only in 3d apps.

The computer is the following:

Core i5-2500k (new), Asus P8VZ68 Pro (new), Gskill ddr3 1600mhz 8gb ram (new), Radeon 4850 (3 years old) and Seventeam 420BKV PSU (3 years old). Running all default clocks.


I've not updated the drivers or anything these days (except for updating Windows 7), and the same applications that were working started failing, and the same squares in different applications.

I would like to know if there is a easy way to figure out if the PSU became faulty or the video card. I've run the Prime95 torture test for 30 minutes, and got no errors. Voltages (except Vcore) have not varied during the test, and temperatures were 87ºC, but running the default heatsink on Brazilian summer, and in games it's below 70. GPU temperatures doesn't look abnormal (below 70ºC in use).

I really want to avoid buying a new psu and a new videocard to see if it solves the problem.

In the past, I was suspicious about this PSU because it used to make some high pitch noises, but since I went from a core 2 duo @3ghz to this core i5-2500k, didn't have those noises, and I've seen a lot of users complaining about high pitch noises on high quality psu brands.

Using Windows 7-64bit with all updates. Catalyst 11.12 drivers (uninstalled and downloaded again and installed again, but no result).


Any tips about what can be this problem or which component is faulty?

 

[chimaxi83]

Take out the video card and run your onboard gpu, since you have a Z68 board. Test and see if your system still locks up. If it does, I'd bet its that PSU. If not, time for a 7970?

 

[Mogar]

Take out the video card and run your onboard gpu, since you have a Z68 board. Test and see if your system still locks up. If it does, I'd bet its that PSU. If not, time for a 7970?

Nice idea, will try that, but I know that the on-board video do not demand so much energy from PSU as the discrete card by far, so I don't know if it will stress the PSU enough to ensure it's ok.

 

[Mogar]

Took off the 4850, installed the intel hd drivers and did some 3dmark 06 runs.... 100% stable.


Put the 4850 back, tried 3dmark 06, couldn't finish the first test. Sometimes, after locking (almost all times with those squares), the driver could recover the computer from freeze, and show on notification area that the video card had stopped working.

Checked the power connections on the videocard, everything looks fine. Still don't know if the PSU is not handling anymore the system's consumption on heavy load (don't know if Prime95 torture can be trusted as a test for the PSU).

Tried underclocking the 4850, no luck, still getting the same kind of freezes.

 

[Magic Carpet]

You could try underclock/undervolt your v-card (say, 30% clocks and 10% volts) and try running those tests again?

Use MSI Afterburner. If you need help, ask.

 

[Schmide]

Dude???? Step on in any trouble shooting. Update your freaking drivers. ALL OF THEM!!!!!!!!!!!!!!!!!!!!!

 

[currahee440]

You could try underclock/undervolt your v-card (say, 30% clocks and 10% volts) and try running those tests again?

Use MSI Afterburner. If you need help, ask.

I would go against doing this... for me I would think if a PSU is unable to provide the power that you need now, undervolting it would just be adding a temporary solution to the problem. What if his capacitors are going bad in the PSU?

It's clear that your PSU is aging. Looking at an online review, it uses OST and capXon capacitors which according to badcaps have known to fail when exposed to heat. Of course we can get into a whole argument about this, including how badcaps is full of people who are overzealous about capacitors...

But I would say your caps are aging, and it's clear that OST and capXon are of so-so quality. Get a new PSU. I would recommend a seasonic or seasonic built unit. Maybe look into Antec, Corsair, Silverstone, XFX, and Enermax

 

[Mogar]

Unfortunately in Brazil we have a very limited access to high quality psu brands, and the few stores pits a big price premium (like over $200 for a 450w antec psu).

The brands you find here as "premium" PSUs are Thermaltake and Coolermaster (the latter I already had on past and became very disapointed since it last 2 years on a pentium d 2.6 ghz with no special video card).

Could the psu make this fault with this described pattern (the same squares on image happens everytime)?

 

[Mogar]

Dude???? Step on in any trouble shooting. Update your freaking drivers. ALL OF THEM!!!!!!!!!!!!!!!!!!!!!

Well, I did that as first measure. I was already running latest drivers, and as I described, downloaded them all again and reinstalled to ensure there was no corrupt file.

I used to frequent Futuremark forums to help users, and I'm quite experienced on solving computer problems, but I've been for 3 years without doing that and became quite outdated regarding pc hardware. When you get stuck on a problem that is not easy to solve and don't have spare parts to do tests, it's always good to have a 2nd opinion.


Update:

Got a borrowed PSU, low quality one from Huntkey 450W, but sufficient to see if it is the PSU. Just did two runs on 3dmark 06 with the radeon 4850 without faults. With my psu, couldn't get the first test finished withouth artifacts and lockups.

Will do some more runs, but looks like I've found who was faulty. Nice since I'm unemployed right now and need to avoid spending money.

This time I will not save money on PSU quality, whatever it costs, as soon I find a quality one to buy.

 

[Mogar]

Bought an Corsair 650TX psu (650W, 53A on 12v rails). Enough from having problems with psu (not the first time), and have plenty of reserve power for upgrades.

Got impressed with the package, build quality and connector quality (they fit firmly without need to make too much force). No issues/artifacts yet, and think the problem is solved.

 

[Magic Carpet]

Glad you resolved it, happy gaming

 

[toyota]

Take out the video card and run your onboard gpu, since you have a Z68 board. Test and see if your system still locks up. If it does, I'd bet its that PSU. If not, time for a 7970?

what? his problems happen during 3D loads. the psu would have hardly any stress at all running onboard so that means nothing really.

 

[toyota]

Bought an Corsair 650TX psu (650W, 53A on 12v rails). Enough from having problems with psu (not the first time), and have plenty of reserve power for upgrades.

Got impressed with the package, build quality and connector quality (they fit firmly without need to make too much force). No issues/artifacts yet, and think the problem is solved.

is that the newer V2? if so I just bought the same psu a couple of weeks ago and like it so far. the Antec Neo Eco 520 I previously had seem to struggle and would ramp the fan up to load at more that 65-70% load.

 

[chimaxi83]

what? his problems happen during 3D loads. the psu would have hardly any stress at all running onboard so that means nothing really.

Whoa buddy, no need to buff your post count making a separate post just to knock my advice.

I had the exact same issue as OP which degenerated into my system not even booting up, and figured it out to be my 9800GTX thru onboard testing first, and then a different video card when onboard showed no issue. Sue me for trying to help.

Glad you fixed you problem OP :thumbup:

 

[superjim]

The 650TX is a nice PS and big seller on Newegg. I'd pit Corsair power supplies against anything else out there in terms of quality and reliability. That should hold you over for awhile....... at least until you splurge for CrossfireX 7970s!

 

[Mogar]

Well, started having problems again.


After some days with the computer working flawlessy with the new PSU (Corsair TX650), I've got the same problem. This time I could take a picture with my cell phone of that is happening, while driving on iRacing (a great racing simulator).


Uploaded with ImageShack.us

You can see the "mighty squares" as well some strange polygons (like this piece of grass surface near from the right tyre) on the image.

One fun thing that I figured out is that the screen was locked with those artifacts, but the sound kept going, also the force feedback, and I could note the physics of the simulator responding to the controller inputs.

Last time I could check core temperature on my 4850 the core was at 75ºC. Ambient temperature around 28ºC today with high humidity (brazilian summer and a rainy day). Doesn't seem to be a big issue.


Anything I can do else than buying a new video card?