PC Shutdowns. Faulty GPU or insufficient PSU?

Hi, first of all my PC specs

MAINBOARD Gigabyte GA-EP43-S3L
C.P.U. Intel Core 2 Quad q9400 @ 3,2 Ghz With Coolermaster Hyper 212 Plus CPU Cooler
G.P.U. MSI GTX 560 Ti Twin Frozr II OC Edition
SOUND CARD Realtek ALC 888
RAM Corsair 2x1 GB Kit DDR2 800 5.0-5-5-18
H.D.D. Western Digital SATA2 7200 RPM 16 MB Cache 320 GB (X2)
O.D.D. Samsung SH-S203B Serial ATA (X2)
P.S.U. Zalman ZM600-HP 600 Watt
CASE Thermaltake Soprano

With Windows 7 Sp1 x64 Ultimate and with the 270.61 nvidia drivers.

So here is my problem.

I recently upgrade my gpu from the 9600gt to msi 560 ti twin frozr II oc edition. I booted to windows, run furmark and some games. So far so good.

But when i tried crysis warhead i had an instant shutdown of my pc and then it turned on again by itself. The weird thing is that it only does this at the intro logos and the intro movie. Another game that has exactly the same behavior is Crysis 2 and FEAR Combat. Especially the intro movie of Crysis warhead is the movie that shut downs my PC almost all the times when it plays.

The funny thing is that i recently finished bad company 2 and crysis warhead without a single shutdown. I've also tried running furmark stability test at xtreme burning mode, unigine, intel burn test, prime95 as well some other games, again no shutdowns or any other problems. Only at the intro movies of these games mentioned above i had PC shutdowns.

I've tried various things.

- Reinstalling the latest nvidia drivers

- Made sure everything inside my PC case is correctly connected.

- I've made a system image and formatted and reinstall win 7 from scratch

- Reduced my cpu overclocking to default values

- Checked my temps. Didn't notice anything alarming.

None of those things helped at all.

So i was thinking what the problem could be. Maybe something PSU related or possibly a faulty card.

Then i tried to underclock my GPU with MSI afterburner and guess what problem solved! I've tried starting crysis warhead and especially playing the intro movie several times and no shutdowns at all.

Specifically i reduced the memory from 2100 to 1600, the core from 880 to 700 and the shader from 1760 to 1440. Didn't touch the voltage.

So is this a PSU related problem or a faulty GPU card?

Sorry for the long post, i tried to be as detailed as possible.

Sounds like a PSU issue.. Use PSUs that are known name brand, stable, and have enough power to run all of your rig. I have always stuck with corsair for PSUs . All that Zalman stuff aint so great .

That's a pretty decent PSU I would say. I'm running a heavily overclocked 2500K with a GTX 560 Ti (also overclocked) on a 500W PCP&C Silencer. Just benched it at 5.4GHz and 1000/2000/4200 and it ran okay. Then again the Q9400 could be drawing more power than a overclocked Sandy Bridge CPU...

1) Unplug your opticals and your secondary HD and keep your GPU at the stock settings and see if the problem persists. It could very well be that your PSU is maxing out.
2) Lower your CPU overclock to say 3.0GHz and see if that solves your problem.

but thats a zalman 4 rail with weak rails so it may be cause.

try the occt psu test just to check, then i would look elseware like any codec packs u have installed or anything that would break securom (unless it dose it other than crysis 2 and other securom games.)

Thanks for the suggestions, here is what i did.

- Bring my cpu back to stock values (btw nothing else is overclocked in my system - not even the RAM - apart from the gtx 560 ti of course because it's overclocked from factory)

- removed the second hard drive and both of the optical drives

- removed one of the two ddr2 sticks

- removed all the case fans

and unfortunately the problem still exists. After i reseted the gpu back to stock values and tried playing the intro movie of crysis warhead my pc shutdowns and turns on by itself.

Is my PSU really that bad? Btw i have it for maybe 2,5 years now. Or is the gpu the culprit?

I hope that is one of the two and not some really weird and very slight hardware incompatibility between my whole hardware configuration:(

Any other thoughts? Should i return the card or buy a new PSU? Unfortunately all my friends have old PC's with crappy PSU's so i can't really try my card there.

Quote:

---

Originally Posted by zanzabar

but thats a zalman 4 rail with weak rails so it may be cause.

try the occt psu test just to check, then i would look elseware like any codec packs u have installed or anything that would break securom (unless it dose it other than crysis 2 and other securom games.)

---

Thanks for the post. I tried a fresh installation of 7 before installing any codecs (i use only cccp btw) and only things i installed were the drivers and the games and didn't help at all. Exactly the same problem.

The games i found that have this problem are crysis warhead, crysis 2 and fear combat.

I tried running occt psu check but i only run it for 1 minute. Is it safe to run it and for how much?

you are hitting ocp from 12v2 rail... http://www.hardwaresecrets.com/artic...y-Review/402/9

you can get a molex to pcie converter...to "might" bypass ocp...

Quote:

---

Originally Posted by Hondacity

you are hitting ocp from 12v2 rail... http://www.hardwaresecrets.com/artic...y-Review/402/9

you can get a molex to pcie converter...to "might" bypass ocp...

---

My card came with two of these converters. Should i try putting the converters on both of the 6-pin connectors or in one of them?

Btw my PSU has one 6 pin connector hard wired to the PSU and the extra one is in "modular form" So i use both of these to power the gpu.

To be honest i don't have much knowledge about PSU's but i looked in the box of the card and it says that i need 30A in the 12v rail. The page that you linked says that the 12v1 is 38A and the 12v2 is 18A that's why you think i have those shutdowns right?

Quote:

---

Originally Posted by Haruspex

My card came with two of these converters. Should i try putting the converters on both of the 6-pin connectors or in one of them?

Btw my PSU has one 6 pin connector hard wired to the PSU and the extra one is in "modular form" So i use both of these to power the gpu.

To be honest i don't have much knowledge about PSU's but i looked in the box of the card and it says that i need 30A in the 12v rail. The page that you linked says that the 12v1 is 38A and the 12v2 is 18A that's why you think i have those shutdowns right?

---

if the OCP rails actually are working, then yes, that would be the problem, and adaptor would be a workaround.

Ok so i unplugged the modular pci-e connector (i leaved the first hard wired one to the card) and i used the molex to pci-e for the second one.

And OMG no shutdowns anymore! It was the PSU after all. With the card at stock values of course.

Now my only question is if it is safe to leave it that way or i should buy a new PSU a.s.a.p? I am not planning to installing anything new to the system besides maybe upgrading the RAM from 2gb to 4gb (i was thinking to buying two more sticks of the same RAM instead of buying two new 2x2 gb sticks, so i can keep the old sticks)

Either way thanks everyone for the help.

I also had issues with my PSU - but it did not result in shutdowns, but the 560TI downclocked itself (also in crysis2). Strange, but it was the PSU. 500W, but at 12v it was only rated at ~385 W.

I would still upgrade to a better PSU, as you are still running it pretty much at the limit which might not be to good for an extended period.

I personally recommend seasonic...

Quote:

---

Originally Posted by Haruspex

Ok so i unplugged the modular pci-e connector (i leaved the first hard wired one to the card) and i used the molex to pci-e for the second one.

And OMG no shutdowns anymore! It was the PSU after all. With the card at stock values of course.

Now my only question is if it is safe to leave it that way or i should buy a new PSU a.s.a.p? I am not planning to installing anything new to the system besides maybe upgrading the RAM from 2gb to 4gb (i was thinking to buying two more sticks of the same RAM instead of buying two new 2x2 gb sticks, so i can keep the old sticks)

Either way thanks everyone for the help.

---

its safe:up:

Like i said, i would stick with a corsair PSU .... I had a 1200 watt PSU (forgot the name of the brand tho ) that had the same symptoms as you, few weeks later it actually just fried in another computer setup i have downstairs. My Corsair 750 and 850 watt has been holding its ground, with a i7 950 OCed to 4.2ghz , had my gtx 460s in SLI OCed, and memory OCed all in a air cooled system. Had a different config in a fully watercooled system with same PSUs and ran FLAWLESS :) .. Good luck !

i bet you didn't read the review i linked...

i bet you don't know the oem...

I would say you are pushing that PSU pretty hard. Surprisingly, XFX make very good PSUs I've used a 750w unit from them which has 62A on its 12v rail. Currently I'm using a Corsair 850w unit that seems to be equally as good as the XFX unit. Tagan make very good PSUs as well for the most part as do PC&C and Seasonic units are usually good as well.

Nice to hear someone managed to fix the issue with adapters. I have the same PSU and the same issues, but no adapters to test with. In any case I'm switching the PSU for an XFX Core 750W with one 12V rail so no more rail problems. :up:

I was just wondering if anyone really had any idea why and where the 12 V OCP kicks in? 18A is a lot of current to draw from a single 12V rail, and the GPU draws its power from 3 different rails. I don't see how it could draw that much from a single rail. The other option is that the CPU would draw 18A from its rail, but that seems an awful lot too even for a high voltage OC. And when people say downclocking doesn't help, it doesn't really leave a lot of options left. Strange. :shrug: