[SOLVED] Rtx 2080 ti doesn't display after deep clean.

[Mpsoldier83]

Read a lot of posts skirting around my issue, but no fix yet.

Fan was making noises like it was hitting something. Took my GPU apart, removed thermal paste and original heat pads, replaced both. Reassembled. One fan wobbles until while card maxes out fan speed, other two are fine.

Now when I turn my PC on, it displays for around 5 minutes, then stops. Faster if I put it under load, any load. It goes from low speed to full and the display stops. Fans still running full speed but no display. Card doesn't feel hot, but I have no monitor on it.

I've taken it back apart to double check the thermal paste and heat pads and everything seems to be right, but this screams overheat issue to be, but I can't see where it would be from.

I appreciate any thoughts.

 

[GhostWPT]

Hello, what thermal compound did you use?

 

[Mpsoldier83]

Noctua.

 

[GhostWPT]

I see, well, I don't think the paste is the problem here, are you running the latest GPU bios?

 

[Mpsoldier83]

It doesn't look like it, I'm on gpu bios 90.2.17.0.7b. Something I may have discovered, I think it's actually crashing my PC, not just failing to display.

 

[GhostWPT]

Try flashing the newest bios, I'm not sure what brand is your RTX card, but here's a list with the GPU's bios collection https://www.techpowerup.com/vgabios...X+3060+Ti&interface=&memType=&memSize=&since=

 

[ko888]

Have you been continuously monitoring the GPU Temperature, GPU Hot Spot Temperature and GPU VRM Temperature? One or more of those temperatures are exceeding the self-preservation trigger points.

If the cooler hasn't been properly remounted you will encounter problems with overheating of GPU, VRM or memory chips.

 

[Mpsoldier83]

In bios my pci-e temp never goes over 35c before the card fails, do you know a better monitor for the temps?

 

[GhostWPT]

In bios my pci-e temp never goes over 35c before the card fails, do you know a better monitor for the temps?

You should have a monitoring tool integrated in geforce experience, just hit Alt+R, but first Alt+Z - Settings - HUD layout - Performance - Advanced

 

[ko888]

The freeware utility HWiNFO64 works quite well.

What is the exact brand and model / model number of the graphics card? Where on the graphics card is the wobbly fan located?

 

[Mpsoldier83]

The freeware utility HWiNFO64 works quite well.

What is the exact brand and model / model number of the graphics card? Where on the graphics card is the wobbly fan located?

Gigabyte aorus 2080ti, I'll grab the rest in a bit, it's the coolest fan to the d-ports.

 

[Mpsoldier83]

You should have a monitoring tool integrated in geforce experience, just hit Alt+R, but first Alt+Z - Settings - HUD layout - Performance - Advanced

GeForce experience keeps crashing on boot, and I can't keep the system up long enough to re-install. Or I haven't been able to yet, will keep trying.

 

[ko888]

Gigabyte aorus 2080ti, I'll grab the rest in a bit, it's the coolest fan to the d-ports.

Half of the GPU chip's VRM circuit is located under that fan. That cooler is also known for its poor/weak contact to the components that it is suppose to be cooling.

When you replaced the thermal pads did you make sure that the new pads were of sufficient thickness and size to aid in transfering heat from the VRM MOSFETs to the cooler?

 

[Mpsoldier83]

Half of the GPU chip's VRM circuit is located under that fan. That cooler is also known for its poor/weak contact to the components that it is suppose to be cooling.

When you replaced the thermal pads did you make sure that the new pads were of sufficient thickness and size to aid in transfering heat from the VRM MOSFETs to the cooler?

They appear to be the same thickness, 1.0mm., and cover the chips completely, but I'll freely admit I'm new to thermal pads.

I'll order a new fan, at worst I guess I'll be out $20 if it doesn't work.

I appreciate all the information.

 

[ko888]

French computer hardware site, Comptoir du Hardware, did a review on the Gigabyte AORUS GeForce RTX 2080 Ti XTREME. The XTREME uses the same WINDFORCE Stack 3X 100mm Fan Cooling System as your card. They reported that the memory chips overheat quickly due to the poor contact of the cooler to the memory chips. They say that the contact surfaces are weaker.

I would interpret that to be insufficient compression force of the cooler against the memory chips or maybe insufficient thickness of the thermal pads used on the memory chips or even insufficient thermal conductivity of the thermal pad material or maybe a combination of any of the mentioned possible causes.

 

[uppercut4u]

Try resetting the cmos on your motherboard

 

[spentshells]

To me this sounds like the cooler has been tghtened to far. Try loosening all the screws to make the heatsink snug, but not tight.

 

[Barty1884]

VRMs shouldn't be heating up that much, in a relative idle state (ie long enough to uninstall GF Experience - although the thermal pads are a very good call. Make sure those are the same thickness as the ones you replaced.

How did your thermal paste application look when you took the cooler off the second time? If not much of a spread at all, the thermal pads may be too thick and stopping the coldplate from making good contact. Flip side of that would be overtightened. Quite hard to do on modern cards with tensioned screws though. You did re-tightend in an 'X' pattern? As in, opposite corners, not 100% down on any one screw before going across and starting the opposites?