Change your GPU's thermal paste!

[Fascinate]

I was getting really high temps on my recent RMA from EVGA gtx 760. This is mostly due to how badly they applied the paste:

http://imgur.com/a/eI8wM

But i bet you could drop at least a few degrees just by using a better paste (i had IC diamond on hand).

My valley benchmark temps went from 81c to 66c max lol.

[BetrayedOf52]

I was getting really high temps on my recent RMA from EVGA gtx 760. This is mostly due to how badly they applied the paste:

http://imgur.com/a/eI8wM

But i bet you could drop at least a few degrees just by using a better paste (i had IC diamond on hand).

My valley benchmark temps went from 81c to 66c max lol.

Are those bugs in the top left of the first picture, freaking me out for some reason.

But ouch, how hard was it to take that apart? Curious how mines been applied.

[Denkou]

Are those bugs in the top left of the first picture, freaking me out for some reason.

But ouch, how hard was it to take that apart? Curious how mines been applied.

Pretty sure they're screws...

[chazus]

They're screws with springs on them, pressure tension for the heatsink.

[Rixis]

Yeah. That case is still disgustingly filthy though. I'm betting at least part of the issue was dust.

What case? All I see if an open Graphics Card on a marble (or marble effect) worktop/plate.

[the boar]

What case? All I see if an open Graphics Card on a marble (or marble effect) worktop/plate.

i'd have called it a "speckled shit" effect table-top. but haha yeah either way that is certainly not the inside of a case and those are not bugs :P

[mrgreenthump]

But ouch, how hard was it to take that apart? Curious how mines been applied.

Most heatsinks come off with removing the 4 tension screws around the GPU and another 8 or so around the card. After that it's just seperating the heatsink, though you may need to rip the void warranty seal away first. That also brings us to the topic, that on most cards.. Removing the seal voids warranty.

[Lathais]

Yeah. That case is still disgustingly filthy though. I'm betting at least part of the issue was dust.

????? what case and what dust? There is an open GPU sitting on a speckled counter of some sort. No case, no dust.

[Fascinate]

Are those bugs in the top left of the first picture, freaking me out for some reason.

But ouch, how hard was it to take that apart? Curious how mines been applied.

Haha they are screws with springs, they do kinda look like flies in the picture. Super super easy to take apart (at least evga and asus are) its just 4 screws and the whole heatsink pops off, and a fan power lead.

[Butler to Baby Sloths]

Most heatsinks come off with removing the 4 tension screws around the GPU and another 8 or so around the card. After that it's just seperating the heatsink, though you may need to rip the void warranty seal away first. That also brings us to the topic, that on most cards.. Removing the seal voids warranty.

If you get to the time when the TIM is deteriorating, the warranty period is going to be long gone anyway

[Fascinate]

If you get to the time when the TIM is deteriorating, the warranty period is going to be long gone anyway

Ya this was a RMA card from EVGA i dont expect new cards to have this problem, but like i said i bet everyone could get 2-5c difference by changing to a quality TIM

[Novakhoro]

Not related - what could I used to track the temp of my GPU? I currently use CoreTemp to track my CPU but never worried about my GPU as my rig is water cooled with 2 exhaust case fans on the top and 1 intake fan on the side.

[Fascinate]

I use CPUID's HWMonitor it tracks pretty much everything, and seems accurate.

http://www.cpuid.com/softwares/hwmonitor.html

Just run the setup file.

[Evildeffy]

The funny thing is that if you look at the placement of the cooler of EVGA that this is 1 of their "FAIL" designs.

If you look closely you can see that the 3rd heatpipe is actually not even being used.
It's not the first or the last that way from EVGA

[Fascinate]

The funny thing is that if you look at the placement of the cooler of EVGA that this is 1 of their "FAIL" designs.

If you look closely you can see that the 3rd heatpipe is actually not even being used.
It's not the first or the last that way from EVGA

I usually buy asus cards but after i changed TIM max temps of 66c in any game/benchmark id say its a pretty darn good cooler, i was surprised actually. Its running at 1226mhz so its getting a fair amount of volts to the core.

[chazus]

The funny thing is that if you look at the placement of the cooler of EVGA that this is 1 of their "FAIL" designs.

Most of the heat is transferred from the die to the start of the pipes and plate. The fact it isn't touching isn't a huge deal, as it's the plate that contains much of the heat until the material in the pipes moves outward.

[Evildeffy]

Most of the heat is transferred from the die to the start of the pipes and plate. The fact it isn't touching isn't a huge deal, as it's the plate that contains much of the heat until the material in the pipes moves outward.

That's how vapor chamber cooling works but it doesn't change the simple fact that it's a direct heatpipe contact solution which would most deffo be more effective if it had contact with that heatpipe.
Also the base plate is not made out of copper which means it also doesn't have the thermal conductivity of the heatpipes and the fact the heat isn't just contained in that block, it's contained throughout the entire cooler size to which it is connected and the "back" area is now severely underutilized because the "front" is taking the brunt of the heat forcing a higher fan RPM, noise and you'd have a imbalanced heat spread.

Though in the GTX 760 it is a non-issue for the most part as the GPU is underqualified for most of the larger coolers had it been a bigger chip the story would've been different as the temperature delta would be significant.

They've done it on quite a few and it makes quite a difference because IIRC there were people who had overheating issues and modded the cooler and it dropped temps considerably.

[bbr]

That's a crazy thick layer as well,. you need only the tiniest film of paste, spread thinly with a credit card. - across the whole surface.

[Denkou]

That's a crazy thick layer as well,. you need only the tiniest film of paste, spread thinly with a credit card. - across the whole surface.

The spread method is not the best method of applying TIM, contrary to popular belief. Yeeeeeeaaaaaars ago the spread method was almost universal until people started experimenting and found that it is not quite as good as some other methods. The pea method is generally the most widely used and accepted method of TIM application among enthusiasts.

[Fascinate]

The spread method is not the best method of applying TIM, contrary to popular belief. Yeeeeeeaaaaaars ago the spread method was almost universal until people started experimenting and found that it is not quite as good as some other methods. The pea method is generally the most widely used and accepted method of TIM application among enthusiasts.

Actually method does not matter, literally the only thing that can make a difference is if you use too LITTLE. Two videos showing this:

https://www.youtube.com/watch?v=D-pr4cf70qY
https://www.youtube.com/watch?v=r2MEAnZ3swQ

Mayo ftw