R9 290x RMA no fault found advice

[glokon]

Hey,

I'm looking for some advice, I sent in a Asus R9 290x back in mid January to SCAN, the first test they did they found no fault, the second test also found no fault, but I am 100% sure there is a fault, it was hard to replicate on any of my systems but always managed. I'm not sure what to do, I bought 2 of the same R9 290X's in April, one goes faulty the other is perfectly fine, I tested each individually, the problem I got was that my system would shutdown or lock the system with a red screen randomly through playing games or running furmark.

I not sure what to do in this situation, pay the £20+VAT and just give up or ask for another re-test and hope that it breaks.

Any advice would be greatly appreciated.

EDIT: I should mention I tested different PSU's one 1000w and a 1200w the card still caused the same problem.

[pastymuncher]

What are your system specs including the make and model of the psu's?

[glokon]

MB: Asus MAXIMUS VI EXTREME
CPU: Intel i7 4770k (No OC)
RAM: Corsair Vengeance 8gb (x4)
PSU (1000w): EVGA Supernova
PSU (1200w): Corsair AX1200i

[pastymuncher]

That rules out a power problem then as they are both quality units. Have you tried updating to the latest motherboard bios? Sometimes a bios update can cure gpu/display problems. Have you checked temps? 290X's can get quite warm especially the top one when running in crossfire.

[watercooled]

Have you clearly explained the steps necessary to reproduce the problem? If you've just given a generic 'faulty' description, they might not pick up on something more obscure.

Also, it might be helpful to post in the Scan sub-forum itself, so you can discuss it with one of their support team.

[glokon]

Yes I clearly explained the steps to reproduce, I run triple monitor at 5760x1080, I tried to run furmark on the faulty card for an hour, but it failed within 10 minutes, however SCAN say this didn't happen to them, however they only tested BF4 on the triple monitor setup not furmark, they only got the triple monitor setup in the 2nd test, but they said nothing failed when BF4 ran for 60 hours over the weekend.

I asked them to test the memory with something like MemtestCL, which showed errors on the VRAM for me, but SCAN said its not certified for testing so they couldn't use it, its the only program that I have found that tests 4gb VRAM, OCCT is limited to 2gb and VMT is limited to 1gb.

Not sure what to do in this case.

EDIT: @pastymuncher: Yeah it was updated to the latest BIOS when I had the card, I double checked everything before sending it away. When testing for Temps they never went above 70-80c in Crossfire and 55-65 in single card setup. (Except FurMark burn-in test got both setups up to 85-90c area)

[watercooled]

How did Furmark fail? AFAIK it doesn't have error detection capability, you'd just see visual corruption on-screen or in severe cases, maybe a BSOD.

However other programs including OCCT do actually check for errors, MemtestCL might too but you'd have to check.

[glokon]

Apologies, I didnt mean fail, I mean it shutdown my system, not because of heat either, it just stayed a steady 85-90c for about 5-10 mins, then just shuts down.

Like I said I ran OCCT, it said it had a 2gb VRAM limit, which isn't enough for the 4gb card, I ran MemtestCL and found several errors, however when I told SCAN to use MemtestCL when I requested the second test, I was told MemtestCL is not certified for use as its "too new".

EDIT: To point out I did the exact same burn-in test on FurMark on my other card (While the other was removed) and it ran fine, no crashes for the entire time I ran FurMark (about 2-3 hours)

[watercooled]

Ah sorry I missed that part about OCCT.

I'd consider a shutdown a fail, I just meant, was it a shutdown or image corruption.

It's a shame about the memory limits in the other programs. What about this one? http://www.evga.com/ocscanner/

[glokon]

I tried ocscanner sadly it limits you to Nvidia and EVGA cards.

Also its a shutdown not image corruption, sometimes but a lot less frequent is a system lock with a red screen.

[GoNz0]

Simply ask scan to send it back to the manufacturers as it is well documented throughout the history of scan that the RMA dept is rather poor (understatement) at times.

You do not owe them a NFF fee until it comes from the manufacturers.

[glokon]

They said if I did that I would most likely get a previous customers card that has been repaired instead of a new one.

[watercooled]

How long have you had it exactly? If it's under 28 days IIRC you can just cancel the sales contract under DSR (or whatever its replacement is), faulty or not, and get a refund possibly minus shipping costs.

[GoNz0]

They said if I did that I would most likely get a previous customers card that has been repaired instead of a new one.

Nothing stopping you asking that this does not happen once asus find fault.

[glokon]

Alright, I'll ask them to send it to manufacturer, thanks guys, been a big help.

Also I purchased it in April last year sadly within 12 months warranty but not 28 days.