PSU? Mobo? Or just a bad series of refubs?

[ICUEDude]

Hello, a bit of advice on defective GPU's

 

Last year I built an i79700k, Gigabyte Arorus Z390 pro, Asus ROG RTX2070 O8G, system powered by a Corsair RM750x PSU. (32g 3200 ram and a MP510 1 gig m.2) Windows 10 pro.

 

My 1st GPU ran fine. This card never "failed" under stress test or heavy load gaming sessions for over 8 months. It had no unusual shutdown issues or other failures. I RMA'ed it because it had pretty bad stutter in games. Turns out this was a known issue with this gen of cards and it failed testing at Asus.

 

After I RMA'ed it, they sent back a refurb. This one started to give me an issue at shutdown where the pc would go to black screen with all fans running forcing a hard shut down. It ran great in games and stress tests under full load but about 2 weeks in, the shutdown issue went to every shutdown. I tried back ups and even a clean win install but it got worse to the point where it would bsod just after windows loaded and or plaster the screen with random pixels. I RMA'ed that one and tested defective at ASUS.

 

I got another refurb. This one worked. No shut down issue. Worked well in games but may have had a bit of stutter again (not as bad as the first card). After 3 days with some long full load game sessions I decided to try a mild overclock to make sure the card was good to keep. I let Afterburner run the Nvidia Curve Oc. Turned it down a bit and applied it with the only change being the power limit set to max . The card began producing green random pixels on the screen and shut down to desk top. I set all settings back to stock. From that point after when ever a load was applied to the GPU it would crash that program to desktop with lots of random green pixels. So I RMA'ed that too and Asus is sending a new brand new replacement card.

 

I'm 99% sure it's the GPU's but I can't help but wonder if there is an issue with my power supply or mother board. What do you think? I'm a bit worried about putting a new card in and it getting damaged. What would you do to test it?

 

I ran the last card in the second PCI slot after it failed and it had identical results (it would run windows but fail under load). I've mem tested and swapped out ram, I've stress tested the cpu with on board graphics, I've checked the idle voltages in the bios (all within spec), updated all bios and drivers of course, and everything checks out fine. It never crashed under full GPU and CPU loads either.

 

Anyway any advice is appreciated. Sorry for the long post, here's a potato {}

[jonnyguru]

If it's the PSU, it doesn't make sense that each card manifested problems in a different way.

[ICUEDude]

Now Asus shipped the repair center another refurb so they are trying to replace my RTX2070 O8G with a RTX2060S O8G model. These guys. I told them to send a RTX2070S. We will see how it goes.

[Vegan]

I have seen lots of dead Asus video cards on various forums. Not sure what is wrong, could be the thermal material is no good or the card has a bad cooler design??

 

Gigabyte is also known for bad Radeon cards.

[ICUEDude]

I'm not sure. Heat was the first thing I checked. The first card rarely got to 60 C, one of the replacements ran a bit hotter but only by 3 - 4 degrees, the other was the same as the first. That's running at 99% GPU usage with no overclocks other than the factory one. My case has good airflow so usually they sat in the mid to high 50's playing COD or BFV.

 

I always run a custom fan curve that is more aggressive than stock but there was no time that it went to full fan speed ever. Not even close. I did read that these cards cooled fairly well.