Question Graphics card crashing during load after Kraken G12 install

[kymearra5]

Hey guys,

Hope everyone is well... I'm not

I have recently just purchased a new PC and have been running it for a month without any issues.
The temperature on my GPU was been sitting around the 40 degrees celcius on idle and ramped up to about 60-70 on load (3dmark / Kombuster on MSI Afterburner). And my CPU around the 30-40 mark during idle.
I ran 3dmark without any issues and also Kombuster on MSI Afterburner without any issues. Managed to OC the Graphics card to 100Mhz Core clock + 400Mhz Memory clock.

So then anyway. A few days ago I decided to purchase a Kraken G12 + Kraken X62 to mount onto my GPU.

So I teared down my MSI 1080 Ti Duke OC and mounted the Kraken G12 + AIO without any issues.

Booted up the PC and it turned on without any issues. GPU was idle'ing at 27 degrees celcius. WOW!

So then I launched 3dmark and as SOON as the benchmark begins, probably 2-5 seconds in, the benchmark freezes immediately and the PC goes black/sometimes red and dies. I have to switch it off and on again to get into Windows. I did this several times and literally everytime it fails at the same point. As soon as the benchmark begins, I hear a little 'vvvVVV' noise coming from the graphics card and then it freezes and I stop hearing the sound. I have no idea what this sound is.

So then I started up Kombuster and same again. Literally 5 seconds in, the program crashes. I even turned off all of my OC's and all the clocks are set to default.

Absolutely dumbfounded and frustrated at this point, I reinstalled a fresh copy of windows. Reinstalled Nvidia drivers and installed 3dmark and launched it and AGAIN same thing. (no programs/OC programs or anything have been installed).

I then set the X62 pump to 100% and fan to 100% and was averaging 25degrees celcius on the GPU and same again. Freezes within 5 seconds.

So then I figured.. maybe the Benchmark is forcing my graphics card to start ramping up the fan speeds because it is under load. (but obviously there are no fans connected to the graphics card), so maybe my PC thinks there's nothing connected, that it should freeze and pause everything.
So then I decided to take off the G12 and put back the heatsink/fans that came with the graphics card and mounted it back on. I didn't plug in the fans and see if that would cause it to freeze. Low and behold.. with no fans connected.. the benchmark ran start to finish without any problems. Wtf???

So now I know its not a fan issue, I thought maybe my PSU didn't have enough W after installing a second RAD. So yesterday I upgraded to Corsair RM850x (from a RM650x). Plugged everything back in. Mounted the G12 back on (with new thermal paste).. ran the benchmark.. bang same issue. Freezes after 5 seconds. #@$&.

It seems no matter what I do.. with the G12 on, the benchmark freezes.

My only last thought of what it could be is the VRAM temperature or something... Because the tiny fan on the G12 doens't offer the best cooling for the memory on the graphics card?Could this be the issue? Why does it fail immediately though without even a chance for the temps to go up?

Can you guys think what might be the issue?

PC Specs:
ASUS Z390-F
Intel Core i7 9700K (with Kraken X62)
MSI 1080 Ti Duke OC
Kraken G12 (with Kraken X62)
G.Skill 16GB F4-3200C16D-16GTZR
Corsair RM850x
Samsung 860 EVO 1TB M.2

 

[rubix_1011]

Sounds like you might actually have something shorting a component causing the instability. Do you notice any marking near where the cooler mounted to the card? You're lucky it didn't just fry the card, if this was occurring.

Other than that, it seems the cooler was working correctly, there was something else going on. I've installed a lot of waterblocks on GPUs and simply unplugging the GPU fan doesn't cause a problem. I still think there is a mount issue or something making contact that shouldn't.

 

[digitalgriffin]

What's the pump RPM? Could it be the pump is clogged/failing causing the GPU to overheat? Even at 3.5GHz Idle, if my pump runs off I can overheat my 3770k.

EDIT: Never mind. If your pump was foobar it wouldn't have low idle temps. How are you powering the fan and pump?

 

[kymearra5]

hey guys thanks for the replies..
I'm not sure if something is shorting, it is a relatively new graphics card and I installed the G12+AIO as instructed, I even tried tightening the G12 on loosely and also tightly to test both, didn't make a difference. Still get the same issue.

That was actually an interesting thought though about it shortening. So I decided to plug in my G12+AIO whilst the graphics card is in (with his generic heatsink/fan) on and ran the benchmark and it finished to the end without freezing. They're both running and spinning and the G12 is just hanging outside my case watercooling nothing without any issues.

I set the pump to 100% and it still froze, I don't believe its a pump issue because as soon as I start the benchmark, the program freezes out, like within 5 seconds. It doens't even give the graphics card a chance to heat up and bug out.

I am at a complete lost. The only thing I can think of is that nothing is making physical contact with the VRAM to keep it cool when the G12 is on it. As compared to when the generic heatsink is on, the heatsink is making contact with the VRAM and pushing very direct airflow on it.

Do you maybe also think it is spazzing out because the GPU does not have the AIO mounted on it perfectly? Would this cause it to automatically shut down and freeze after 5 seconds? Hmmm I'm not sure about this also because I've seen the spread of the thermal paste (just using a line of it) and its spread pretty evenly and flat after I took it off.

Oh man.. this is annoying!!!

 

[kymearra5]

All sorted! Turns out I didn't tighten the G12 mount/bracket and AIO on tight enough!!

 

[despawn]

Hello, I am suffering the same problem. I am sure I tightened the AIO and Bracket. Do you know any other solutions?

 

[kymearra5]

Hello, I am suffering the same problem. I am sure I tightened the AIO and Bracket. Do you know any other solutions?

Hey there, sorry i can’t take any photos at the moment because i am overseas but I cleaned off the existing thermal paste andy put on a line and when i put the bracket back on. i didn’t use any of those rubber pads they give you because it was too chunky.
placed the bracket on and really made sure it was making contact with the gpu and tightened it pretty firm.
you’ll notice that the corners of the bracket kind of sit off the gpu. you want to make sure that it’s not sitting off it and making good contact and tighten those screws on tight! that’s really my only tip i can give you. just make sure it’s on right and tight.

then run msi afternoon and load up kombuster and if it doesn’t immediately crash then you’re all sorted! trust me when i says it’s no other external issue. it’s not the vram or anything. it’s totally the tightness of the bracket being fitted to the gpu.

i’ve had it on for months now and with the graphics card overclocked and haven’t had any issues.

 

[despawn]

Hello I just replaced thermal paste and added some small heatsinks on the vram and remove the rubber pads then tightened all the scews and bingo! Works flawless. Thanks