Question GPU higher temp than expected, watercooled, help please!

[DeepseaMK21]

Side panels and top panel off, front panel on. During the Heaven benchmark, highest graphic setting, AA x8, 3440x1440. My numbers were as follows:

ambient: 20 c
fans: 1,600 rpm
pump: 3,699 rpm (at about 60%)
Loop temp: 34 c
GPU: 53 c
CPU: 37 c

RTX 2080 TI,-slight overclock at 2055mhz. 8700K, not currently overclocked. 2x360 30mm rads, D5 pump. Pump--->rad--->rad--->CPU--->GPU--->res. 6 EK vardar fans, 1 stock NZXT case fan exhaust, NZXT H700 case.

Currently have all 6 rad fans pushing through radiator and out of case. I should have top pushing out, but front sucking ambient air in. That could be one reason for higher temps when panels are installed but even without panels on temperature is high, as seen in the above numbers.

FPS is fine. Ran Far Cry 5 benchmark as well, just to compare some numbers and FPS was higher than what a few others i saw have posted when on ultra settings at 3400x1440, averaging 108 FPS.

Thermal paste on GPU is Thermal Grizzly Kyronaut. Minus 8 pads on VRAM and such. I am fairly confident the thermal paste application is adequate, perhaps too much but that's better than too little right? I spent a lot of time making sure no bubbles were present in the loop. If any are hiding in the front rad at this point, I couldn't say.

From what I have read, loop temperature shouldn't be higher than 7-10 c over ambient, and GPU should be between 10-15 c above loop temperature, no more. I feel like my setup should cool everything better. Do I have too many bends? I am not sure what I am doing wrong. I know a lot of people go rad directly after component, but would it even matter in this case as equilibrium will eventually happen anyways and ya know....laws of physics and such. What is going on with these temps??

 

[getochkn]

That looks like wayyyy too much thermal paste. Rice sized grain in the middle, tighten, let it smooth it. You have leaking out so much it's all over caps and resistors.

 

[ashdavid]

That looks like wayyyy too much thermal paste. Rice sized grain in the middle, tighten, let it smooth it. You have leaking out so much it's all over caps and resistors.

If the loop is getting hot and the GPU is not, than it would not be a thermal conductivity problem, ie the thermal paste is doing its job by bringing all the heat correctly to the heat sink (and I always thought that you could put more thermal paste on a GPU b/c the surrounding does not matter if it seeps out the side)

That said, where does it say that a loop should only be 7-10 degrees above ambient? If you are running stress tests for a prolonged period, it is only natural that the loop will get hotter with more time.

 

[Gam3r01]

If the loop is getting hot and the GPU is not, than it would not be a thermal conductivity problem, ie the thermal paste is doing its job by bringing all the heat correctly to the heat sink (and I always thought that you could put more thermal paste on a GPU b/c the surrounding does not matter if it seeps out the side)

That said, where does it say that a loop should only be 7-10 degrees above ambient? If you are running stress tests for a prolonged period, it is only natural that the loop will get hotter with more time.

The issue is too much thermal paste begins to act as an insulating material, all its supposed to do is fill in imperfections in the mounting surface, thermal paste itself transmits heat worse than metal, so just enough to ensure smooth contact is all you need.

As for the temps, I personally dont see anything wrong there given the setup.
Are your temps gradually increasing, or do they hit your max temp relatively fast?

 

[getochkn]

View: https://www.reddit.com/r/watercooling/comments/9ufu2d/2080_ti_ekwb_temp_issues/


View: https://www.reddit.com/r/watercooling/comments/9ntbzv/ek_vector_2080_ti_temp_issue/


One it seems to come down to using the right screws. Second, your temps seem fine according to threads like that.

 

[DeepseaMK21]

The issue is too much thermal paste begins to act as an insulating material, all its supposed to do is fill in imperfections in the mounting surface, thermal paste itself transmits heat worse than metal, so just enough to ensure smooth contact is all you need.

As for the temps, I personally dont see anything wrong there given the setup.
Are your temps gradually increasing, or do they hit your max temp relatively fast?

I'd have to do some longer tests to see where it evens out. Most of the test have been short and end up ending and things cooling down a bit before starting a different test. I understand the concept of too much thermal paste acting more as an insulator, but would it make that much of a difference in this case? I've seen a few tests such as one from Gamer Nexus where he applied a little, an average amount, and too much, and the temps differed by only a hundredth of a degree. I used the pea method on the CPU, but for some reason I was scared to put too little on the GPU, maybe because the chip size being bigger in size, I didn't feel I'd have enough so I used that spreader tip that came with the thermal grizzly.

If the loop is getting hot and the GPU is not, than it would not be a thermal conductivity problem, ie the thermal paste is doing its job by bringing all the heat correctly to the heat sink (and I always thought that you could put more thermal paste on a GPU b/c the surrounding does not matter if it seeps out the side)

That said, where does it say that a loop should only be 7-10 degrees above ambient? If you are running stress tests for a prolonged period, it is only natural that the loop will get hotter with more time.

regarding the 7-10 degrees above ambient. It's something I read while searching through some other posts on the forums. Wasn't a whole lot to back it's validity up, but after seeing other peoples temps for similar setups it seemed to at least hold up to some truth.

View: https://www.reddit.com/r/watercooling/comments/9ufu2d/2080_ti_ekwb_temp_issues/


View: https://www.reddit.com/r/watercooling/comments/9ntbzv/ek_vector_2080_ti_temp_issue/


One it seems to come down to using the right screws. Second, your temps seem fine according to threads like that.

I used the screws that came with the backplate, as they were longer than the ones that came with the waterblock itself, and the backplate and block were bought separately. Same company, but maybe they are too long? Only way to know I guess is take everything apart again.

 

[HWOC]

I don't see any particular issue with the temps, I think those are to be expected. But like previously said, I'd put a little bit less thermal paste.
As to your fan configuration, when your case is closed, you have 7 exhaust fans and no intake fans at all? Like you said before, you should change the front fans to be intakes instead.

 

[DeepseaMK21]

I don't see any particular issue with the temps, I think those are to be expected. But like previously said, I'd put a little bit less thermal paste.
As to your fan configuration, when your case is closed, you have 7 exhaust fans and no intake fans at all? Like you said before, you should change the front fans to be intakes instead.

Yeah, I totally screwed that one up! This weekend I am going to change it up. Plan A, top fans continuing to exhaust up. Front will become intake. Rear fan continue to exhaust.

I have a plan B as well, which would be easier, but I have no idea if it would be as good as plan A or not good at all. What if I kept the top and front fans in the same configuration (pushing air out), but took a much higher flow exhaust fan, and turned it into an intake fan? Bringing ambient cooler air in? Would it be worth a try or would the air going through the rads still not have a big enough delta to have good heat exchange?

 

[HWOC]

For your plan B, where were you planning to bring in the air from? Additional fans on the side panel? In my opinion 6 fans exhausting through radiators would benefit from more intake fans than just one, unless it was really moving a LOT of air. I would definitely go for the front three to bring in air, and the top three to exhaust. And the rear fan for exhaust as well.

 

[DeepseaMK21]

For your plan B, where were you planning to bring in the air from? Additional fans on the side panel? In my opinion 6 fans exhausting through radiators would benefit from more intake fans than just one, unless it was really moving a LOT of air. I would definitely go for the front three to bring in air, and the top three to exhaust. And the rear fan for exhaust as well.

Sounds like plan A wins! I'm going to give it a go. 3 front intake, 3 top exhaust, 1 rear exhaust it is. A little planning on how to reroute wires and I should be able to take the fans out and flip them without have to break a whole lot apart. If that works it shouldn't be too much work. Thanks all!