HD 4870 X2 Random Crashes - Graphic Corruption / PC Reboot

[slipmat]

Hello,

This topic I could not find directly in the forum, I know there have been some threads related to this issue, which I have tried all of them. The reason I am posting this thread is to make sure I have the BiOS set up properly for the components I have.

I have contacted ATI, they refer me to everyone else saying there is something set up improperly.

The problem can occur in the BiOS with graphical tears on the screen like small colored squares, to the OS running, to a 3D Game. The longest uptime I have had over the two weeks I have had the card is 18 hours.

My computer:
Mobo: ASUS P5N-E SLI BiOS 1101
Processor: Intel Q9550 Yorkfield 2.83 Ghz
Processor Cooling: Tuniq Tower 120
Ram: G.Skill 4 GB (2 x 2GB) in the yellow slots (running dual channel - have also had them in single channel (one black / one yellow) no change of issues.
Ram Cooling: Corsair CMXAF1 Fans
Graphics Card: Sapphire HD 4870 X2
Hard Drive 1: WD VelociRaptor WD3000GLFS (300 GB) - Sata
Hard Drive 2: Seagate Barracuda 7200.11 ST3500320AS (500 GB) - Sata
Power Supply: Rosewill RX950-S-B (950W)
Dvd/CD-Rom: ASUS 20X DVDR with Lightscribe - Sata

Case: Antec Twelve Hundred Black ATX Full Tower - Excellent Air Flow in case someone thinks that is the problem.

Things I have done:

In Operating System:
Changed the ATI Control Panel XML to properly change the fan speeds for cool operations. Random Crashes with small squares still occur.

Outside the Operating System Changes (BiOS):

Changed the CPU Voltage around the full spectrum that is advised by Intel and also have left at Auto. Random Crashes with small squares still occur.

Changed the BiOS to show PCI-e for the video card slot. SLI is disabled (I Have tried with Auto). Random Crashes with small squares still occur.

Adjusted the Memory Voltage to 2.085 Volts and just auto. Random Crashes with small squares still occur.

Played with other voltages yet most of the time it returns a hard crash and requires a CMOS reset.

I have manual set the timings on my RAM which is 6-6-6-18. I have left at Auto. Random Crashes with small squares still occur.

I have tripled checked the power going to the video card, everything is correct, used both Rosewell 12V cables just to make sure. Random Crashes with small squares still occur.

I have yet to try one thing, which is change the PCI-E slot to 105 Mhz or Hz (not sure which) which someone mentioned on another forum in regards to the ASUS mobo with the HD 4870 X2.

Based on my above computer specifications, can someone advise the exact settings I should use in the BiOS?

Have I missed something glaring?

In case you might be curious, the reason I have the ASUS P5N-E SLI motherboard is I ordered a DFI lan Party X48 mother board with the PCIe 2.0v slots, yet the board was DoA, so being impatient the only place I could find a mother board in Philadelphia, the best they had was the ASUS P5N-E SLI. Even though I will more than likely receive a new mother board next week, I would like to figure out what I can do in the mean time to be able to keep the computer stable until then.

Any help would be appriciated.

Thank you,

slipmat

Edit: Changes the thread title
Edit2: Noted the BiOS revision for the mother board. One thing that is odd, my processor was unrecognize for the original BiOS I had. I updated to the 1101 revision and it recognized it. Yet, according the the latest BiOS (now 1201 or 1102), it has added more processors, interestingly enough my processor Q9550 is not on the list. (http://support.asus.com/cpusupport/cpusupport.aspx?SLanguage=en-us&model=P5N-E%20SLI) I don't see any issues directly related to this, yet could this be a/the problem? Every failure has been graphically related, not processor related....

 

[slipmat]

Really? I thought that would be more than enough, ATI for the 4870 X2 says required 650W. I think I have figured it out, the motherboard does not support 45nm processors as I noted above. That sucks, .. hmph.

 

[shortstuff_mt]

strangestranger hit it on the head. This sure looks like a PSU problem. A Rosewill PSU is asking for trouble, especially with a 4870 x2. I would check the ATI site for certified PSU's for the 4870 x2. The best brands to focus on are Corsair and PC Power & Cooling IMO.

 

[pulasky]

ASUS P5N-E SLI BiOS 1101 = Nshitforce chipset = problem with ATI cards

 

[slipmat]

Thanks for all the responses, I suppose I will upgrade that powersupply, that blows as I just bought it. hahaha I thought it was plenty amazing enough, bummer I suck.

 

[slipmat]

I guess some further explaination will help me if you don't mind explaining.

The Rosewill has 950W, but you are saying that this is basically meaningless because it has to do with line amps or something?

What should I look for if not pure voltage, I guess I'm lost a bit on that, I thought I was perfectly safe because it was SLI / Crossfire Rated etc.

Thanks again for your help,

slipmat

 

[roadrunner197069]

I think you got a iffy PSU and I ont think your mobo is 100% compatible with your quad.

Your quad is not on the CPU supported list.

I think your PSU is from a non popular vendor but it looks to have enough amps, and the bios shouldn't get messed up cause the video car isn't under any load at all.

I would replace the PSU and the mobo if it was me.

 

[slipmat]

Not to bump this to the top, I just wanted to understand more about the power supplies and a rated one (Antec) versus the one I purchased (Rosewill).

Here is a link to the ATI Certified Antec backplate: here

Here is a link to the Rosewill backplate: here

From what I see, it has slightly less (20A) per 12V than the certified Antec which runs 25A per 12V.

Max load at 840W on the 12V lines on the Antec and max load of 810W on the 12V lines on the Rosewill.

I still have yet to receive my updated motherboard which I should receive early next week (first one was DoA) - so, I don't have much hooked up now, but should I be concerned with the Rosewill still? I guess I am still confused on the PSU concerns everyone has posted. I did read the PSU article in the other section and I don't know how to determine if there are individual "rails" on the 950W Rosewill.

Thanks again for all the help!

-Slipmat

 

[Liderc]

Rosewill is basically a crap brand, that is why everyones mentioning it. I wouldn't pin it right down to the PSU but Rosewill does make crap psu's. Most of these people will look for a weak link in your setup with a problem like this and your weak link is the poor quality psu and the motherboard which isn't exactly compatible with your CPU. Corsair makes great power supplies if you wish to purchase a new one. I personally would try out your new mobo when it gets there and then if the problem continues, swap out the psu.

Should also mention that a poor quality psu can blow at anytime when under load and could kill a lot of your hardware, that's really the main reason everyone mentions a weak psu when trying to pin down a problem.

 

[slipmat]

ahhhh! Now I understand, I thought it was more capabilities and not just the fact that the name = crud.

Fair enough - I'll wait for the new motherboard, if it works super and then i'll look to getting a new PSU in the immediate future.

Thanks again,

slipmat

 

[Liderc]

No problem, hope everything works out.

 

[JAYDEEJOHN]

20 amp vs 25 amp per 12 volt rail isnt slightly, its alot. 25% to be exact. My lil ol enermax rated at 420 watts does 18 amps, "just" 2 amps less than yours, and I wouldnt even consider running a 4870 on it, let alone a x2

 

[slipmat]

Well jaydeejohn,

I was going to attack thinking you might be wrong, but you're not wrong. HAHAHA.

Yet, ATI does say it's okay to run at 19 amps per 12V line, yet they only have one certified PSU that does that and that's an Antec Neo Blue 650W.

Most of them are 25 Amp+ with 2 at 20 Amp that I saw. Now I understand everyone's concern. Thanks for pointing it out again!

I feel slightly enlightened and smarter! I need to find someone to impress with my new found knowledge, I don't think the wife will be interested! haha.

Thanks again!

-slipmat

 

[JAYDEEJOHN]

Yea, she may just ask, does it work or not? heheh

 

[slipmat]

Hi again - starting over - some more issues - yet much less frequent - at least after the system has been "up" for at least 10 minutes?

So here are my final updated specs:

Mobo: DFI Lanparty DK X48-T2RS (BiOS X48DD829 - latest)
Processor: Intel Q9550 Yorkfield 2.83 Ghz
Processor Cooling: Tuniq Tower 120
Ram: G.Skill 4 GB (2 x 2GB) running dual channel
Ram Cooling: Corsair CMXAF1 Fans
Graphics Card: Sapphire HD 4870 X2
Hard Drive 1: WD VelociRaptor WD3000GLFS (300 GB) - Sata
Hard Drive 2: Seagate Barracuda 7200.11 ST3500320AS (500 GB) - Sata
Power Supply: Rosewill RX950-S-B (950W)
VGA Power Supply: ThermalTake W0156RU - (650W Dedicated VGA (HD 4870 X2) Power supply. 30 Amp per rail - running 1 rail to each power input on the card)
Dvd/CD-Rom: ASUS 20X DVDR with Lightscribe - Sata

Case: Antec Twelve Hundred Black ATX Full Tower - Excellent Air Flow in case someone thinks that is the problem.

So ....

I fixed the major issues:

1) Wrong motherboard for processor
2) Insufficient power

Same issues as previously mentioned. Much less frequent but still come up in the BiOS - still come up in the booting process - randomly and still have the issues just sitting in Windows.

I can play Bionic Commando all of 5 minutes before crash, yet GRID, Warhammer Online, WoW, and others I can pretty much play endlessly. So far if system is running and Warhammer Online is playing I have not had a crash?

After shutting off system, sleeping and waking up in the morning computer will take 10 minutes of re-boots before it is stable?

For the record - 950W PSU is pulling from it's on circuit breaker and the 650W PSU is from a different circuit breaker. Just to make sure I don't overload the circuit breaker in the house just in case it pulls more than 20 AMPS. So far it has not.

Could it be a bad card now?

Thanks for your help!

-slipmat

 

[Steve_long]

If the display is tearing in BIOS, then its a card problem. There should be no issues with PSU, since in BIOS, all the hardware should be in an idle state.

I hate it when "experts" automaticallu jump at the PSU!

"Hey my hamster died!.....ohhh, its a PSU problem!"

 

[-713king-]

the same thing happen to me today this morning i just messed with DRAM settings in bios to get my memory running at 1600 and well after i did that it crashed because i put wrong settings and then after that crash my video card just stoped working or something because i would get a gray horizontal line with small squares on it and the rest of the screen was normal then i went into bios and i had dots on spots they were not suppose to be. I did a cmos reset and took out video card and memory and put everything back in place after that and it looked like it worked but then when i started windows on the bottom right corner it said DISPLAY DRIVER crashed and rebooted and then everything went to grayed Pixels not only that but before it crashed i did a Test on video card and it failed .... i had to RMA and get my money back from newegg.. this happen to my HD 4890 Sapphire Vapor-X. I mean putting a wrong setting on BIOS can it mess up my video card? because i was playing fine that night before with no problems and then i just changed those settings next morning and boom it messed up..