• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

GTX570 issue or PSU issue? or Other?

Most of my friends play on consoles (I know, makes me sick too) and wouldn't have a PSU with even 1 6 pin PCI-E connector.

I've know someone that may let me use his PSU (Tagan 500W I think)... I'll try and get a spare PSU to make sure my currect PSU is the problem before throwing money at the problem.

Edit: Thanks Ch3m1c4L, I'll have a geek at the PSU's on offer, Got to make sure my PSU is the problem first though I guess.
 
iB CoMpL3XiTy said:
Most of my friends play on consoles (I know, makes me sick too) and wouldn't have a PSU with even 1 6 pin PCI-E connector.

I've know someone that may let me use his PSU (Tagan 500W I think)... I'll try and get a spare PSU to make sure my currect PSU is the problem before throwing money at the problem.

Edit: Thanks Ch3m1c4L, I'll have a geek at the PSU's on offer, Got to make sure my PSU is the problem first though I guess.
Click to expand...

That'd be the best thing to do, because you could very well end up with a new PSU and it not be the issue, best to be sure if it means spending mo' money.
 
Just called the guy with the tagan 500W and he needs his PSU and isn't willing to lend me it for testing, sad times indeed.

I'll test without the extension cables

I'll test with the cards in different PCI-E slots

I'll try running OCCT and Kombuster at the same time

I'll run Memtest

I'll run Unigene on each card for atleast 3 hours to make absolutely, 101% sure there both fine

Then I'll order a shiny PSU.
 
Just tried to run MSI kombustor Multi GPU and lost signal after about 30 seconds

Was playing music at the time and could still hear the music playing after the monitor lost signal, the PC is still on, just no signal.

Surely if it was a PSU issue, the PC would shut down completely (OCP) ?

Also when I got back into windows after a restart I found that the nVidia driver is no longer installed....

My brain hurtz.
 
That's really bizarre, it could be a case of the GPUs have shut down because of the lack of power, but windows itself is still running?
 
This would cost you probably £20 if it didn't work, but couldn't you buy a psu & if it didn't make any difference return it under the distance selling regulations ?

If it did work, well it wouldn't be money wasted ? :)
 
I wouldn't buy a £20 PSU, even just for testing on one GTX570, If I was going to by a PSU for such a purpose I'd buy an antec earthwatts or something similar.

If I did this and it worked then I'd still send the new PSU back as I don't want two PSU's.

If after my testing I'm still confused then I might just by a Corsair AX1200 or similar, test it out and if it works then I'll keep it, and if it doesn't I'll DSR it.

Thanks for your suggestion :)

I've been Kombusting for 35 minutes with the new card, remember SLI only lasted roughly 30 seconds.
 
I never meant buy a £20 psu :)

The £20 was for covering the return postage due to DSR if the new psu made no difference.
 
98195470.jpg


Set init display first to PCI-E x16 (1) as I wanted to re-test the older GTX570, but the nVidia driver isn't detecting the second card anymore, GPU-Z says the BIOS for the new card is unknown and its revision is now "FF", it was "A1" when I had my PC boot from the newer card.

And I've got a warning sign again in Device manager...

I don't understand it as the new card works fine when I set init display first to PCI-E (2).
 
Why isn't gpu-z showing your pci-e bus interface as v2.0 ?

Does that change if you remove 1 card ?
 
I think the only time ive encountered that was when I had forgotten to plug in a power adapter after reseating the card, either that or I hadn't reseated them properly. But reseating and confirming plugs were in correctly, + reseating the SLI bridge (had somehow got well dirty inside it, during use?) fixed it for me.

What does the warning sign mean? double click the device and see if it says the device cannot start. If its not getting enough power to start though, the lights shouldnt be green on it, and blue is for the card thats outputting to displays. Do the 570's even have lights on?
But if both cards work fine, in different slots, on their own, its def either mobo or pus, if youve updated your mobo bios to the latest and it still doesnt work then its got to be psu right? Id ship you my spare but I think its only 700 or 850w or somthing so wont be much use.
 
13023468.jpg


This is strange, I know for a fact that the first cards stock voltage is 1.050V, and the newer cards stock voltage is 1.037V, In the SS above its displaying the right temps and when I change the fan speed nothing happens as you'd expect (its watercooled), but its saying the voltage is 1037mV, when it should be 1050mV

Why is it diplaying the stock voltage for the second card, but everything else for the first card... MSI AB won't let me select the second card.

This is driving me to insanity lol
 
I'll double check all cables, get rid of extensions and I'll reseat the cards.

I'll also double check that I have the latest BIOS for my mobo.

Thanks for all the help.

EDIT: when I double click the warning sign in Device manager it says "Windows has stopped this device because it has reported problems. (Code 43)"
 
maybe its because they are different volts that is causing the problem, one is trying to increase/decrease volts and that is causing an issue? Each card should be able to just clock up/down etc to the master one and be fine, but it could just be your unlucky with these cards.

They are both the same 570 right? not one thats OC'd its nuts off and one thats stock?
 
Its the card with a stock voltage of 1.050V thats the master card, so in theory it shouldn't be undervolting itself right?

Both cards are POV GTX570 Beast Editions with exactly the same stock clocks.

I removed the extensions and reseated the top card and moved the newer card to the third PCI-E lane, the only difference this made is that the warning sign in Device manager has moved from the top card to the bottom.

I checked and I have the latest BIOS (F13j BETA)
 
iB CoMpL3XiTy said:
I'll double check all cables, get rid of extensions and I'll reseat the cards.

I'll also double check that I have the latest BIOS for my mobo.

Thanks for all the help.

EDIT: when I double click the warning sign in Device manager it says "Windows has stopped this device because it has reported problems. (Code 43)"
Click to expand...

Someone on here yesterday had problems with the same code 43 error & a GTX 460, and it's related to usb devices iirc.

Here's the thread :

http://forums.overclockers.co.uk/showthread.php?t=18237154
 
its certainly weird, since you cant select it in MSI afterburner you can force different volts on them either so thats a no go, but I have no idea about volts on gpu's and what is/isnt normal/safe etc....really should learn that so I can get more out of my 295's.

Do what the guy in th eother thread did and unplug your usb's/change em over etc. may as well unplug any extra hdd's as well just to rule out other driver conflicts. I did get a random blue screen the other day (USB Driver conflicting with something) after my computer telling me my usb was unlugged/plugged in (new g19 is only change really), then ive been getting driver crashes since updating my nvidia ones. Maybe there is some incompatability in the new drivers?

I can send you any old drivers if you want to try them as well, got around 2gb of nvidia drivers lol.
 
ok, alittle update

I set the init display first back to PCI-E x16 (2) which allows both cards to be recognised perfectly and also allows me to enable SLI, Set the voltage for both cards to 1.050V and ran MSI Kombustor Multi GPU and the monitor lost signal after a minute and a half, at stock volts it lasted roughly 4 minutes, then after a reboot I set the voltage for both cards to 1.1V and re-ran Kombustor, this time the monitor lost signal after about 5-10 seconds... So to me that indicates a PSU problem.

Although I am still confused as to why when I boot from the first PCI-E lane the second card isn't recognised at all, but when I boot from the second PCI-E lane everything is recognised as it should be... I don't understand the logic in that.
 
what lanes does it use for PCI-E when in SLI mode? maybe if the master socket is used it does 16/8/8 and when the secondary is it uses 16/16 or something weird like that? Thats a total stab in the dark though, as I expect a x58 to use 16/16/16

Are the the correct slots? My board has 2 pci-e 16x 2.0 and 1 pci-e 16x 1.1, So it says to use slots 1 and 3 for the first 2 cards?

It could also just be on how its splitting the load on the gpus or something?

Id just buy a nice new shiny PSU of around 1200 as you said (meaning tri SLI wont be out the picture) and it should last for SLI for another gpu upgrade or two. and then if its not that, just return it under DSR, or just return it normally (lots of PSU's aren't sealed and ive done this to other retailers before when I bought a Silverstone one and it only had an 8pin, not a 4pin+4pin, and the board only had 4pin...doh)
 
You must log in or register to reply here.
Back
Top Bottom