What to do after Unsuccessful RMA

[Bahazbz]

I was having some trouble with a GPU that was under warranty so I sent it in for RMA. I just got back my RMA from Gigabyte. Upon receiving the RMA part back, I checked the status online and found the boilerplate response "No trouble found after tests with multiple configurations". This was surprising because the problem is fairly evident if any testing was done at all. Even more disappointing, the card now has some pretty intense coil whine and thermals seem to ramp significantly faster than they did before. I repeat, the RMA GPU came back worse than I sent it!

What are my options as a consumer?

Should I send it back in and ask them to reexamine it?

Is there a way to specifically request a replacement instead of a repair?

Do I just have to wait until it finally dies all together in order to have them actually fix it?

A more detailed post on the original problem can be found here. I would really like some advice on what to do next.

Any and all answers are greatly appreciated!

 

[Lutfij]

1| Return it to the seller, if possible.
2| Possible!
3| If it's found to be faulty, it'll be replaced.
4| I wouldn't ask you to go that far.

Not to doubt your knowledge but can you specify what the other system's specs was(on which you tested in multiple configs)? You may want to take another PSU that is fairly brand new of of higher wattage courtesy of a friend or neighbor to see if the issue lies with your PSU.

 

[Bahazbz]

1| Return it to the seller, if possible.
2| Possible!
3| If it's found to be faulty, it'll be replaced.
4| I wouldn't ask you to go that far.

Not to doubt your knowledge but can you specify what the other system's specs were (on which you tested in multiple configs)? You may want to take another PSU that is fairly brand new of higher wattage courtesy of a friend or neighbor to see if the issue lies with your PSU.

My electrical engineer friend and I check my PSU with a multimeter. We checked each of the rails for tolerances of +/- 0.03v. We are also fairly certain we are not exceeding the PSU's limits, as power draws from the wall reads < 250W when under synthetic loads according to a kill-a-watt wall monitor. Thanks for the answer anyway, the PSU may be worth testing again regardless.

The other system was my friend's machine. Can't remember the exact spec but it was a 700W PSU I believe. So I doubt I was pushing it in the power draw department.

Anyone else have any advice?

 

[renz496]

Some manufacturer probably did not consider coil whine as a "defect" on a product. While the noise is annoying it should not prevent the card to operate normally. So when asking for RMA you should say the issue is exactly about the coil whine and demand the replacement if the noise can't be fixed. If they said they will not going to replace your card because of coil whine issue then time to look for manual fix. In some cases the coil whine will be gone after few hours of stress test. About the heat issue try re-apply your thermal paste.

 

[Bahazbz]

Some manufacturer probably did not consider coil whine as a "defect" on a product. While the noise is annoying it should not prevent the card to operate normally. So when asking for RMA you should say the issue is exactly about the coil whine and demand the replacement if the noise can't be fixed. If they said they will not going to replace your card because of coil whine issue then time to look for manual fix. In some cases the coil whine will be gone after few hours of stress test. About the heat issue try re-apply your thermal paste.

Right, the thermal paste! I forgot about that little detail. Gigabyte may have taken off the heatsink during their tests, and that would explain the sudden thermal differences. I can try reapplying thermal paste. Is there a special compound for GPU heatsinks, or will something like Arctic Silver 5 suffice? I'll also see if stress testing helps for the whine. I really wouldn't mind the new noise if the GB team had actually fixed the original problem, lol.

 

[renz496]

AS 5 should be good enough.

 

[Bahazbz]

Some manufacturer probably did not consider coil whine as a "defect" on a product. While the noise is annoying it should not prevent the card to operate normally. So when asking for RMA you should say the issue is exactly about the coil whine and demand the replacement if the noise can't be fixed. If they said they will not going to replace your card because of coil whine issue then time to look for manual fix. In some cases the coil whine will be gone after few hours of stress test. About the heat issue try re-apply your thermal paste.

Just got done re-applying thermal paste (arctic silver 5) and replacing the thermal pads with fujipoly top quality pads. Temps dropped by about 1 degree, a far cry from the 10°C delta I saw upon return. Oh well, that wasn't the issue. Any other advice guys?

 

[renz496]

make sure you have "just enough" when applying the new thermal paste. too many it will hamper the heat transfer. this is a mistake that can happen even at factory assembly. other than that double check your case air flow. you might want to use custom fan curve as well instead of manufacturer default setting.

 

[Bahazbz]

make sure you have "just enough" when applying the new thermal paste. too many it will hamper the heat transfer. this is a mistake that can happen even at factory assembly. other than that double check your case air-flow. you might want to use custom fan curve as well instead of manufacturer default setting.

Yeah, I've had a custom fan curve set up for a while, just can't be thing get too loud, I share a room with others rn. I always do a test run and check my thermal application after supplanting the heatsink once, just to make sure I am using an appropriate amount. I might send the card back and have them look at it again, or else I'll just deal with the high temps and the defective card. Doesn't seem like there is much else I can do at this point.