• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

SOLVED New calling 2080 Owners

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

cajunjosh

New Member
Joined
Oct 14, 2018
So I'm new to water cooling but I've watched a TON of content from big names like JayzTwoCents on water cooling configurations.

I'm running a Ryzen 7 1800x and Asus DUAL RTX 2080 Ti. CPU is not overclocked, GPU power slider has been moved from 100% to 120% but the temperature numbers seem to be the same between both.

My loop:

Pump/Res Combo > GPU > CPU > Radiator > Radiator > Repeat

The Pump is a D5 with 300ml reservoir.
EKWB Water block for GPU
EKWB Water block for CPU
Radiators are Cool Stream XE 480mm x 60mm x2
8 Vardar 120mm SP fans in push configuration

Might be overkill but it was built to be.

So here is the question...

When I'm running an intensive game like Digital Combat Simulator with VR and an auxiliary air conditioner running in my room cooling room temps to around 18.8c I'm seeing some pretty hot temps from the GPU while the CPU is chugging along just fine at normal temperatures. At idle the GPU is running about 6-10ish Celsius above room temperature. When the load is removed from the GPU it cools very quickly which makes me wonder if that's normal behavior or an indicator the water block might actually be seated properly and doing its job.

When running in DCS I'm seeing the GPU temps fluctuate quickly between 67 to 84c with the average while flying being above 84c. When running a non intensive game like RL my CPU temps drop even more and the GPU bounces around between 67c and 84c with the average being in the high 60s. Again the CPU temps are normal and the temps are stable on the GPU. The ASUS Dual has some negative reviews regarding how hot it runs on the air cooler and I'm wondering if the temps I'm seeing on water are just as good as it gets with these toasty 2080 Ti's or if I might have a thermal paste/seating issue with the water block. I used an adequate amount of meh grade Coolermaster thermal paste. Thermal pads were all installed per manufacture instructions and all appear to be seating well when I pulled the water block off to check.

Loop has no kinks and has been purged of all visible air. No longer seeing the occasional bubble working it's way through the system anymore.

Just wanted to get the experts thought's on GPU performance specifically and not looking to get debates started over loop order or capacity unless someone honestly thinks that 2 480x60mm radiators aren't enough to cool this which I doubt would be the case. I'm ordering the Grizzly thermal paste and will try to re-seat the card once that comes in next week. As I type this my room temp is 22c GPU temp is hovering at 44c and CPU temp is down at 26c. Any information is greatly appreciated and thanks in advance!

-Josh

IMG_20181013_205643184.jpg
 
Last edited by a moderator:
I hope you are running 4k 144hz with 2 2080ti's. :)

GPU temps in a loop are typically higher. Do understand you have 2 250w gpus and a 95w cpu, nearly 3x less. ;)

You'll also want to overclock that cpu to get the most out of those cards at any res.
 
Earthdog thank you for the reply and I apologise for the confusion. The 2080 TI branding is called the DUAL but I'm only running one GPU which is the one I'm concerned about. Sorry about that confusion, I should have been a little more clear.
 
Thanks for clarifying. Same story. 250w device and 95w device.


So EarthyDog if I have poor contact between the block and GPU core would I still see decent idle temps and quick cooling (20-30c over the course of a second or two) when loads are removed from the GPU?
 
Seeing as how my 1080 Ti is cooled by itself on a Watercool MO-RA3 radiator and it's load temperature is never above 33c you either have a water block seating issue making good contact on the GPU die, which takes trial and error seating and unseating to see what the thermal footprint looks like.

To see what the true footprint looks like the water block has the be tightened down all the way then removed, and inspect, it is truly a pain in the a$$ to do, but that's the only way to be 100% sure you have the thermal contact you need.

You want even coverage across the entire die using as little thermal compound as possible because you're only filling the microscopic imperfections, and that's assuming that EK has a mirror finish on the die contacting area, as mine wasn't that fine and needed a little extra TIM.

Your 2080 Ti may also be affected by the CPUs temperature added to the loop, many think that doesn't matter but I always run 2 independent loops keeping the CPU and GPU temperatures from combining and limiting each independent overclock potential.

This falls into the department of how much overclock you actually desire to run and even if you don't overclock either, are you satisfied with the temperatures they both are running.

Temperatures can seem to be a non issue until the performance begins degrading from heat.
 
Silver there might be too much thermal paste on there as I was using some cheap and old Cooler Master paste and was assembling the card for the third time trying to figure out a fitment issue. Are there any tricks to the trade when screwing the water block into the card as far as tightening in a certain order or any other tips to getting this right?

IMG_20181014_120141867.jpg

IMG_20181014_120040357.jpg
 
Silver there might be too much thermal paste on there as I was using some cheap and old Cooler Master paste and was assembling the card for the third time trying to figure out a fitment issue. Are there any tricks to the trade when screwing the water block into the card as far as tightening in a certain order or any other tips to getting this right?

Tighten around the die first to be 100% sure your are using the correct screws, EK sent me a package of screws for the 1080 Ti and most were obvious but there was 2 sets of 4 screws that look almost identical but one set was slightly longer and wouldn't allow the die contact to fully tighten. I accidentally discovered that doing my thermal footprint testing and after tightening down around the die the card would freely move. EKs instructions did not cover any information regarding those screw lengths. But after I compared the two types I used the shorter screws and it tightened down properly.

Check it yourself.
 
Silver there might be too much thermal paste on there as I was using some cheap and old Cooler Master paste and was assembling the card for the third time trying to figure out a fitment issue. Are there any tricks to the trade when screwing the water block into the card as far as tightening in a certain order or any other tips to getting this right?

View attachment 201670

View attachment 201671

You want a 'rice size' amount of TIM on chips in general. Tightening in a cross pattern with half twists at the most all around till you can't no more finger tight. It should naturally seal around as you don't want too much TIM on there. It's suppose to be an amount that almost looks like it's 'see through' if you will or close to it as best you can otherwise if it's just too much, it will affect performance.

Tighten around the die first to be 100% sure your are using the correct screws, EK sent me a package of screws for the 1080 Ti and most were obvious but there was 2 sets of 4 screws that look almost identical but one set was slightly longer and wouldn't allow the die contact to fully tighten. I accidentally discovered that doing my thermal footprint testing and after tightening down around the die the card would freely move. EKs instructions did not cover any information regarding those screw lengths. But after I compared the two types I used the shorter screws and it tightened down properly.

Check it yourself.

The longer screws are usually for the rear bracket.
 
After reading all the responses I think I might know what I did wrong. Since the card was brand new and the thermal pads on the stock cooler were brand new and already pre cut I used them with the new water block instead of using the ones supplied by EK which were thinner. I'm betting those thicker pads are restricting the contact between the GPU and the block.

If this ends up being the case I'll add it to the list of all the rookie mistakes made during the build.
 
After reading all the responses I think I might know what I did wrong. Since the card was brand new and the thermal pads on the stock cooler were brand new and already pre cut I used them with the new water block instead of using the ones supplied by EK which were thinner. I'm betting those thicker pads are restricting the contact between the GPU and the block.

If this ends up being the case I'll add it to the list of all the rookie mistakes made during the build.

That's very possible.
 
nice job of troubleshooting gents and welcome to OCF Josh!
something you can do to check how well your block is making contact is after youve mounted the block grab a nice bright flashlight. lay your card on your desk and try to shine the flashlight between the gpu and block. if can you see light coming thru you know something is loose. if not your usually gtg.
 
Thanks Maxfly, I will keep that in mind. Hopefully the replacement pads and washers will be in this week and I can give everyone an update on Project Hurricane when she's back up and running again.
 
Well it looks like it was the stock thermal pads being too thick. Dropped the idle temp by 20 degrees Celsius.

Idle Temp.png
 
Back