Dead RTX 3090?

[Xapic]

Hey guys, 

I have a ASUS TUF RTX 3090 that crashes after about 1 hour of use, after that i get the q code 97 and a white led on gpu. Sometimes i will boot up again, work for 5 minutes and do the same thing all over again. I cant tell what it is causing that the system boots again, it seems pretty random. It also crashes without any load after about one hour, in the same way it does under load after one hour.

I have already tried:

• CMOS reset
• Using a different PCIE power lane on the PSU
• Updating the MB bios
• Updating the Vbios while the GPU was working
• Using the second Vbios on the card
• Using another PCIE slot on the MB
• Using different screens
• Using different cables to the screens
• Reinstalling windows
• Reinstalling drivers
• Tried playing different games while the gpu was working
• Disconnecting external drive

All of this did not change anything, still behaves the same. I have also send the gpu to RMA but they didnt find anything, and it passed all of the tests.

But what is even weirder for me is that when i change out the gpu to a different one it boots up without any problems, with a different gpu the system runs without any issues. 

Its also important to know that the gpu ran for almost one year without any crashes, isnt overclocked or anything similar, temps are normal.

 

Full specs of the system:

R9 5900x

64gb RAM 3200mhz

ASUS crosshair dark hero X570

ASUS TUF RTX 3090 OC

SAMSUNG 970 EVO (90 TB written, with drive status Good)

Be qiuet Straight power 11 1000W

 

Only thing that is odd to me is that i inspected the card today and saw that two capacitor pairs on the backside of the gpu are touching, i dont know if this is wanted but in all the pictures i have seen until now there wasnt one that looked the same. 

I attached one picture of my gpu. The two marked cap pairs are clearly touching.

 

If anyone has anything i could try i would be happy to try, because im running out of ideas.

Thanks for any replies in advance

[Elijah Kamski]

I'll be honest with you, I don't think touching was intended.
 

Can you confirm to me this is the correct model?

https://www.asus.com/Motherboards-Components/Graphics-Cards/TUF-Gaming/TUF-RTX3090-O24G-GAMING/

Because as you can see, I ain't seeing anything touching XD

[mr fobs]

I had an issue like this and one of the pins on the power connectors was pulled back. I pushed it back forward and it clicked into place. might be worth checking.

[Xapic]

1 minute ago, Elijah Kamski said:

I'll be honest with you, I don't think touching was intended.
 

Can you confirm to me this is the correct model?

https://www.asus.com/Motherboards-Components/Graphics-Cards/TUF-Gaming/TUF-RTX3090-O24G-GAMING/

Because as you can see, I ain't seeing anything touching XD

Yep it is the correct model, i dont know if that was the issue tho, maybe rma didnt test it enough.

[Elijah Kamski]

Contact Asus directly and send them the pic with the stuff touching at the back, since I'm guessing you just RMA'd it via the retailer you bought it from.

[Xapic]

Just now, Xapic said:

Yep it is the correct model, i dont know if that was the issue tho, maybe rma didnt test it enough.

Nope, its not the issue i tried reinstalled the connectors multiple times. And it shows a red light if one isnt connected right

[Xapic]

1 minute ago, Elijah Kamski said:

Contact Asus directly and send them the pic with the stuff touching at the back, since I'm guessing you just RMA'd it via the retailer you bought it from.

Yeah i did, they said it was working fine tho, that is what confused me. I will contact asus directly now tho.

[Deadpool2onBlu-Ray]

11 minutes ago, Xapic said:

Hey guys, 

I have a ASUS TUF RTX 3090 that crashes after about 1 hour of use, after that i get the q code 97 and a white led on gpu. Sometimes i will boot up again, work for 5 minutes and do the same thing all over again. I cant tell what it is causing that the system boots again, it seems pretty random. It also crashes without any load after about one hour, in the same way it does under load after one hour.

I have already tried:

• CMOS reset
• Using a different PCIE power lane on the PSU
• Updating the MB bios
• Updating the Vbios while the GPU was working
• Using the second Vbios on the card
• Using another PCIE slot on the MB
• Using different screens
• Using different cables to the screens
• Reinstalling windows
• Reinstalling drivers
• Tried playing different games while the gpu was working
• Disconnecting external drive

All of this did not change anything, still behaves the same. I have also send the gpu to RMA but they didnt find anything, and it passed all of the tests.

But what is even weirder for me is that when i change out the gpu to a different one it boots up without any problems, with a different gpu the system runs without any issues. 

Its also important to know that the gpu ran for almost one year without any crashes, isnt overclocked or anything similar, temps are normal.

 

Full specs of the system:

R9 5900x

64gb RAM 3200mhz

ASUS crosshair dark hero X570

ASUS TUF RTX 3090 OC

SAMSUNG 970 EVO (90 TB written, with drive status Good)

Be qiuet Straight power 11 1000W

 

Only thing that is odd to me is that i inspected the card today and saw that two capacitor pairs on the backside of the gpu are touching, i dont know if this is wanted but in all the pictures i have seen until now there wasnt one that looked the same. 

I attached one picture of my gpu. The two marked cap pairs are clearly touching.

 

If anyone has anything i could try i would be happy to try, because im running out of ideas.

Thanks for any replies in advance

They probably tested it for 10 mins and saw it work and sent it back. Demand an exchange. No reason to have a 2k gpu that crashes. 

[GuiltySpark_]

Those components are likely already bridge if they were touching or not, this appears to be a red herring. A multimeter test on the other components would tell you pretty quick.

[Xapic]

Just now, Ryan829 said:

They probably tested it for 10 mins and saw it work and sent it back. Demand an exchange. No reason to have a 2k gpu that crashes. 

Yeah thats true, i will contact asus directly, whats even worse is that i had to pay almost 3k for it due to the scalping.

[Elijah Kamski]

Just now, Ryan829 said:

They probably tested it for 10 mins and saw it work and sent it back. Demand an exchange. No reason to have a 2k gpu that crashes. 

I bet you it must be Newegg.

That's exactly what I thought, they just saw that it works and bam, sent it back.

If it's Newegg again, time to get Steve from Gamer's Nexus on the case XD

[Xapic]

Just now, GuiltySpark_ said:

Those components are likely already bridge if they were touching or not, this appears to be a red herring. A multimeter test on the other components would tell you pretty quick.

Thought of that too, should i check for shorts between the front and end of two different caps?

[Xapic]

Just now, Elijah Kamski said:

I bet you it must be Newegg.

That's exactly what I thought, they just saw that it works and bam, sent it back.

If it's Newegg again, time to get Steve from Gamer's Nexus on the case XD

No its not, its a german reseller. So no need to get steve on this lol.

[Elijah Kamski]

Just now, Xapic said:

No its not, its a german reseller. So no need to get steve on this lol.

Ah okay, hold back on the one man army Steve lads! XD

[Deadpool2onBlu-Ray]

4 minutes ago, Xapic said:

Yeah thats true, i will contact asus directly, whats even worse is that i had to pay almost 3k for it due to the scalping.

Bruh. 3k buys you a lot of McChickens. You could have invested 1.5k in stock market, bought a $500 3070, and then spent $1k on 900 McChickens and fed yourself for like a year AND had like 70% of the gpu performance still. Hopefully you at least have the original receipt otherwise you're screwed. I don't understand giving in to scalpers. 

[Xapic]

Just now, Elijah Kamski said:

Ah okay, hold back on the one man army Steve lads! XD

If its not a intended short it would maybe worth to investigate this more, maybe more people than just me that are having this issue.

[Xapic]

1 minute ago, Ryan829 said:

Bruh. 3k buys you a lot of McChickens. You could have invested 1.5k in stock market, bought a $500 3070, and then spent $1k on 900 McChickens and fed yourself for like a year AND had like 70% of the gpu performance still. Hopefully you at least have the original receipt otherwise you're screwed. I don't understand giving in to scalpers. 

I have the reciept, and the gpu was bought buy stock market returns lol, maybe thats why it didnt hurt that bad, but i get the point you are making, i wouldnt do it again but i did that decision when i bought it.

[Deadpool2onBlu-Ray]

1 minute ago, Xapic said:

I have the reciept, and the gpu was bought buy stock market returns lol, maybe thats why it didnt hurt that bad, but i get the point you are making, i wouldnt do it again but i did that decision when i bought it.

At least it sounds like you were in the financial position to buy it unlike a lot of people 

[jones177]

All my resent GPU issues turned out to be power related.

One was a bad 8 pin PCIe cable and the other was a bad 24 pin motherboard cable.  Both were on PSUs that were less that a year old and both ran perfectly for about 6 months.

 

When I used GPU to test that had less power draw the issues did not show up.

The cards worked fine in other systems(I have 5 in the house) so it was a lot easier to narrow down.

 

What I did was buy a new set of PSU cables and replace all of them. If that was a fix I threw out all the old cables.

 

I am really paranoid about type stuff so the MSI 3090 I have listed below is a spare for one in a dedicated rendering rig.

 

[Xapic]

1 hour ago, jones177 said:

All my resent GPU issues turned out to be power related.

One was a bad 8 pin PCIe cable and the other was a bad 24 pin motherboard cable.  Both were on PSUs that were less that a year old and both ran perfectly for about 6 months.

 

When I used GPU to test that had less power draw the issues did not show up.

The cards worked fine in other systems(I have 5 in the house) so it was a lot easier to narrow down.

 

What I did was buy a new set of PSU cables and replace all of them. If that was a fix I threw out all the old cables.

 

I am really paranoid about type stuff so the MSI 3090 I have listed below is a spare for one in a dedicated rendering rig.

 

The card is drawing normal power, and the one i had when the tuf was out for „repair“ was drawing normal power too.

the tuf also has lights that light up when the power plugs arent connected, which are not lighting up. Plus the way it crashes looks like damage in the gpu core or memory.

Screen 1 goes black, with weird color glitching and screen tearing like artifacts on the second screen.