Problems with EVGA RTX 3080 XC3 Ultra Gaming - 2 fans randomly decrease RPM and PC crashes every few days

[Chris Redfield]

Hello everybody,
 
I'm having issues with my GPU for more than a month.
I tried many things to solve those issues withing that time,but with no luck.
I'm using my GPU for mining since 22 February and occasionally gaming.
I'm taking good care of GPU and clean it as well as whole PC often.
Same goes for software and whole OS.
I also always kept my GPU temperature below 55c and memory junction below 92c.
Average temperature for GPU is around 51c and mem. junc. is 87c
 
A bit more than a month ago I started experiencing PC crashes that became often and happened every 2-3 days.
I also started finding 2 of my GPU fans that decreased their RPM speed by around 20-25% randomly while third was working properly.
I always set them all to 88% and only one keeps that speed constantly while other two decrease speed by them self randomly.
It happens every 2-3 days.

I noticed that memory junction temperature always jump to 106c after that happens but never above that despite big changes in ambient temperature or Overclock settings.
I tried changing MSI Afterburner for Precision X1,but it didn't help.
I tried 4 different drivers,setting Power mode to max performance and reinstalling whole system and nothing worked.
Yesterday I took multimeter and tested my PSU and it was fine,around 12.07V under full load during OCCT stress test.
 
I have my GPU overclocked and I found balance between performance for mining and temperature.
I tried to lower OC as well,but it didn't help neither.
It always worked on that OC and it was always stable,but now it crashes even on lower OC.
Reliability Monitor shows Live Kernel Events under codes : 117,116 and 141 while 117 being most often.
If it would help to solve this problem I'll gladly upload minidump that shows more details about erros/crashes
 
Here are my OC settings:
 
Core -150
Memory +1000
Power Limit 68%
Fan Speed 88%
 
Here is my full rig:
Desktop:
OS:Windows 10 x64
Monitor:LG 27UD58-B
CPU:Intel i7 10700KF
CPU Cooler:Arctic Liquid Freezer II 360
MBO:MSI MAG Z490 Tomahawk
GPU:EVGA GeForce RTX 3080 XC3 ULTRA GAMING
HDD:1 TB WD Blue
SSD:256 GB ADATA SX8200 ASX8200PNP-256GT-C, PCIe NVMe, M.2
RAM:2X8 RAM Corsair Vengeance 3200mhz
PSU:Seasonic FOCUS 750 Gold
Case:SilentiumPC Signum SG7V EVO TG ARGB
 
 
I kindly ask you for your help and I hope you'll mange to resolve this issue.
 
 
 
 
 

[alyen]

Are you sure you have set your fan speed properly without the zero RPM mode?

If you have bad fans it could lead to over heating issues with the graphics card while you mining. If that continues it could wear down the graphics card faster.

[twinmat]

try not overclocking it as hard... or at all and see if that helps 

[Chris Redfield]

40 minutes ago, alyen said:

Are you sure you have set your fan speed properly without the zero RPM mode?

If you have bad fans it could lead to over heating issues with the graphics card while you mining. If that continues it could wear down the graphics card faster.

I set fan speed like I always did.

I went into MSI Afterburner set it to custom speed of 88% and hit apply.

I also saved profile profile that does that automatically at startup and it worked for half of a year with zero issues.

Weird thing is that 2 fans acting like that,but third fan is running fine at speed I set.

Is there a way for me to test fan functionality?

I'm worried because I'm having constant PC crashes because of my GPU so I'm not sure if fans themself are the only issue.

I would order 2 new fans from EVGA,but I want to be 100% sure that they are the only problem,otherwise it would be unnecessary cost to buy fans if something else is wrong with GPU and I still have warranty.

Only reason why I didn't sent it to repair is because they don't have them in stock,or should I say in somewhat enough quantity so I'd have to wait for new one for couple of months and I'd probably get newer version that has LHR (blockage for mining) so it would be much less profitable with it and even now it's not some fantastic profit with single 3080,I use it more like small time miner,since I didn't plan to mine with it when I bought it.

I was told from a friend that he used his for mining and he's returned money for GPU with mining,so I did the same.

 

32 minutes ago, twinmat said:

try not overclocking it as hard... or at all and see if that helps 

I lowered OC and nothing changed.

I could try to run in on stock,but it would run hotter and use more power while mining since OC for mining is more of a balance between settings unlike OC that was used for gaming,but I'll give it a shot to see if that's the case.

[JeremyNoller]

I have the same exact card and I am running into the same issues mining ETH.  I can not OC my card with out it crashing at all.  I have also noticed the VRAM Temps Spike quickly to 100 Plus.   I sent the card back to evga and they sent me a new card... The new card is worse than the card I sent them.  I sent them an email today showing them the VRAM temps on it. I am waiting for their reply.  

[JeremyNoller]

I am going to try and replace the thermal pads on it. I dont know if EVGA went cheap on them or what..

[Allan B]

I have the same card and did about 0.001 bitcoin mining ETH in the winter months (up in Canada).  I did similar power and core offsets that you had except I kept the memory around +600Mhz and did not have any issues.  Memory temps were between 92C and 96C (I have not modified the card physically at all).  I am surprised you are running that card a +1000 memory with memory temps around 87C.  You must have modified the stock cooling or have a lot of fans (must be loud), as that does not seem to be inline with the temps i was seeing on mine.

 

I would suggest the same low power and clock options but leave the memory at +0 and try that.  You won't get as much ETH but is a good test to see if your card is stable or if you have cooked your card\memory.