MarcSam

Mrostom
You can try that your are the best judge since you can see it all.

However try to get a coolant temp sensor to see the coolant temperature.
Also check on the pump speed.
Let us know and I will be able to help you out as much as possible!

I will definetely keep you up to date hopefully with a full detailed report with images and numbers.

Coolant Temp sensor isnt at all necessary.( 8 yrs of WC'ing and never had the need for one, although I do have one in my box of fittings)..a "flow indicator" would be far more beneficial then a temp sensor imo.:
- Coolant temp should be slightly above ambient with your PC idling, so just check your GPU Temp (msi ab)...with an efficient loop your GPU should be idling very close (within a few degrees c of your room temp ((ambient)) ).
 
Here are my temps and information with a 3080 FTW3 Ultra (OC +125 Core / +650 Memory) w EK Vector n Nickel backpplate, room temp (ambient) at 20c (I know coolant is close to the same as GPU Temp idling indicates that):
- pump speed 3200 RPM fixed
- 9900K (EK Velocity all Nickel Block)
- 480 / 360 / 240 Rads all single fans (no push / pull) running at 1250 RPM
- 3080 FTW3 idle temp = 20c - 22c / under full load 1 hr of Cyberpunk 77 = 46c
 
What is your GPU idle temp (idle temp will reflect your coolant temp) vs load temp .....you had said load temp 70c ...which is far too hot for a waterblock GPU...period (unless I'm misunderstanding something).
I know the 3080Ti runs a bit warmer than the 3080, but I would think when I swap out my GPU to the Ti version perhaps the load GPU temp may climb from 46c to perhaps 49c maybe 50c...
 
A number of years ago running SLi 1080Ti's (all under custom loop) was playing Witcher 3 and saw my temps slowly climb and peak out at 68c....from 43c....drained my loop, disassembled both of my GPU Blocks and found some old "crud" blocking the cooling fins (within the block) on the backside of the GPU Heatsink...although had been almost 18 months since a flush..
Cleaned the fins out ....put everything back together and back to normal temps.
 
As I said earlier, with your high GPU temp its either:
- Pump not running properly (or at all)
- Very poor contact between GPU n GPU Block
- Restriction in your loop. (check both CPU and GPU Blocks for any bits of crud plugging up the fins within the Blocks ((you will have to disassemble then to check)) )
 
Please keep us updated as Im quite curious what you find to be the problem....guaranteed it will

^^^ Sorry man..1st ..In 8 yrs of building Loops..I have never seen Coolant temps  10c - 20c over ambient with my system idling, altohugh I know my Loop is very efficient...(if you do you have serious cooling issues in your loop).....as I type this with eveything idling and having shut down a 3 hr session of CB77...my present room temp is 20c...and my GPU is idling at 21c....CPU's (all 8 cores) idling at 25c....so just a glance I know my Coolant is very close to ambient or room temp....most of us that have been WC for several yrs ditched the thermal sensor for liquid.....we know when our loop is running warm or not...the poster w 70c under load has issues for sure!
 
 

KingEngineRevUp

FlavorII
Hey guys, 
Not sure if this is the right place to put this but I'm having trouble with my EVGA FTW3 Ultra 3090 and my EK water blocks. I feel like I'm going crazy. So far I've had 4 of these cards die on me after water blocking them with EK's block. I've been installing water blocks for a long time and I've never had issues like this before. The first time I followed EK's directions for both front and backplates to the T and blocked two cards in preparation for an nvlink bridge that I have yet to actually install. Both cards worked great for about 3 weeks and then one suddenly died. then about 3 days later the other died. I contacted EVGA who RMA'd the cards. I was even more careful blocking the new set, I noticed that the EK backplate didn't actually fully cover the VRAM set towards the back of the PCB so I added a strip of 1.5mm thermal pad so that it would make good contact.  I used 1.5mm because it's 0.5mm offset from where they tell you to put the 1mm pads. After triple checking my work I installed them into the computer and everything booted up and thanks to the additional thermal pad my memory temps dropped almost 10c. However, 2 weeks later....they both died. One first and the next about 24 hours later. 
 
After digging around on the internet the issue seems pretty common and I found this discussion. The info about the Rev. 1 water blocks was great but I verified that the water blocks I own have the new style stand offs so I'm not sure what could be up. I'm contemplating selling whatever EVGA sends me and my water blocks this time and switching to a different 3090 altogether. I really don't want to block them again and have them die on me again, it's super frustrating. Does anyone know if there's any official statements from either EK or EVGA on this topic? Or if anyone has determined this to be definitely a GPU or definitely an EK water block issue? To be honest I'm not sure what I'm looking to get out of posting here. Maybe some reassurance to give me the confidence I need to try a 3rd time? 

Why don't you play it safe and put kapton tape everywhere around the standoffs? Maybe punch small holes so kapton tape can act as a barrier?

That's a good idea and if I try for a 3rd time I'll probably do that. I took one of the cards apart today to get it ready to send back and took pictures. I was wondering if you or anyone here sees anything I may have done wrong. The grey pads are Fujipoly and the blue are EK pads. 
 

FlavorII

KingEngineRevUp

FlavorII
Hey guys, 
Not sure if this is the right place to put this but I'm having trouble with my EVGA FTW3 Ultra 3090 and my EK water blocks. I feel like I'm going crazy. So far I've had 4 of these cards die on me after water blocking them with EK's block. I've been installing water blocks for a long time and I've never had issues like this before. The first time I followed EK's directions for both front and backplates to the T and blocked two cards in preparation for an nvlink bridge that I have yet to actually install. Both cards worked great for about 3 weeks and then one suddenly died. then about 3 days later the other died. I contacted EVGA who RMA'd the cards. I was even more careful blocking the new set, I noticed that the EK backplate didn't actually fully cover the VRAM set towards the back of the PCB so I added a strip of 1.5mm thermal pad so that it would make good contact.  I used 1.5mm because it's 0.5mm offset from where they tell you to put the 1mm pads. After triple checking my work I installed them into the computer and everything booted up and thanks to the additional thermal pad my memory temps dropped almost 10c. However, 2 weeks later....they both died. One first and the next about 24 hours later. 

After digging around on the internet the issue seems pretty common and I found this discussion. The info about the Rev. 1 water blocks was great but I verified that the water blocks I own have the new style stand offs so I'm not sure what could be up. I'm contemplating selling whatever EVGA sends me and my water blocks this time and switching to a different 3090 altogether. I really don't want to block them again and have them die on me again, it's super frustrating. Does anyone know if there's any official statements from either EK or EVGA on this topic? Or if anyone has determined this to be definitely a GPU or definitely an EK water block issue? To be honest I'm not sure what I'm looking to get out of posting here. Maybe some reassurance to give me the confidence I need to try a 3rd time? 

Why don't you play it safe and put kapton tape everywhere around the standoffs? Maybe punch small holes so kapton tape can act as a barrier?

That's a good idea and if I try for a 3rd time I'll probably do that. I took one of the cards apart today to get it ready to send back and took pictures. I was wondering if you or anyone here sees anything I may have done wrong. The grey pads are Fujipoly and the blue are EK pads.

Are you warming the cards up befo you disassemble them? And when you do, wre you twisting the PCB and heatsink apart? You're not just pulling them apart cold right?

That's the only thing I can think of that can potentially be damaging the cards. Otherwise, kapton tape should be a easy fix for you if you want to play it safe. Kapton tape the front and back.

My 3080ti FTW3 should be here tomorrow and my block and backplate Wednesday.  I was looking over the instructions and noticed on the backplate instructions it says to use different thickness thermal pads in a couple areas depending on if you have a 3080 or 3090.  No mention of the 3080 ti.  So, which thermal pad setup should I use?  
 
Pic for reference...
 

KingEngineRevUp

TahoeDust
My 3080ti FTW3 should be here tomorrow and my block and backplate Wednesday.  I was looking over the instructions and noticed on the backplate instructions it says to use different thickness thermal pads in a couple areas depending on if you have a 3080 or 3090.  No mention of the 3080 ti.  So, which thermal pad setup should I use?  
 
Pic for reference...
 

Copy the 3080

Thanks.  After comparing pics of the PCBs, that is what I was thinking.  Just wanted to confirm.

Good morning guys, thank you all for the responses, I spent whole yesterday and this morning doing test, tilting and doing stress testes. Before we go any further here's a quick recap of my liquid cooled components:
 
CPU: 9900K at stock with ekwb
GPU: 3080 Ti FTW3 at stock with ekwb
Motherboard: Aorus Master Z390
 
2 radiator 360 PE series and a ekwb vtx pump.
Here's a photo of the system with a lookup of the cpu and gpu fins in detail:
 
  
 
Lets' start from yesterday with this situation at boot with a thermal probe directly inside pump reservoir and thermohygrometer next to it for room temp readings:
 
 So coolant temp at boot 27°C, with room temp 27 °C throughout the entire test.
 
And after closing the case and wait the temps to normalize they were like this:
 
with GPU temp the same as coolant temp.
 
I started some stress right away and at the end of it here's the situation:

 
 
 
Max gpu temp around 70/71 °C and I would like to point out that pump speed was between 3200 and 4200 rpm depending on the load.
 
Now we are at this morning. Following the suggestion of trying to make escape some air bubble trapped inside the system I tilted the system even 180° in both axis and made the pump run and indeed some air bubbles manage to escape from the top radiator.
This morning we have 30 °C instead and the stress tests were performed with the case opened:
 
 Coolant temp 31°C, room temp 29°C (it will be 30°C in some minutes) and after some times the system idling like this: 
 
Ran 3 stress tests and these are the results:
 
 so max gpu temp around 67/68 °C with coolant temperature like this: 
 
So a quick recap for the maximum values:
 
Close case - gpu temp around 70 °C, coolant temp 45°C and 27 °C room temp.
Open Case - gpu temp around 67 °C, coolant temp 44°C and 30 °C room temp.
 
I hope I explained everything. Do let me know if you need something else. In the meantime I will try to switch fan orientation and see if I make some progress.
 
I will keep you up to date and thank again for your help :D

talon951
MarcSam, that's still a block mount problem. Air to water delta looks fairly typical of 2 360's depending on fan quality and speed. But a 23-25C block delta is pretty high for 400-450w. Higher flow rate would help some. 2nd pump or maybe a D5. But improving the block mount would help the most.

I thought that as well but coolant temperature shouldn't be that high during load otherwise. Anyway I teared down once again the block and here's what it looked like:
 
 at first glance I thought that maybe was too much thermal paste so after cleaned it  I applied a thin layer of thermal paste 
 
I ran two stress tests and in both case coolant temperature at the end of the tests was 43 °C as shows:
 
 
 with max gpu temp reaching about 63 °C with open case. 
 
So at this point either radiators are clogged or the airflow must be changed or I received a defective waterblock.
 
Thank you again for your time.
 
 
 
 

MarcSam
OMITTING YOUR POST TO NOT TAKE UP FORUM SPACE 
ORIGINAL COMMENT HERE: https://forums.evga.com/FindPost/3465436

Can you upload your photos on IMGUR? For some reason they're very pixelated with the service you're using. I think your memory temperatures are around 68C? What is water temperature at? You can probably warm the card up, remove the backplate and I bet you, you'll be able to tighten the 4 screws around the GPU a little more. At most, my memory temperatures are +25-30C above water always, and my GPU temperature is 18-20C drawing 400-400W. 
 
As for your water temperatures, I'll tell you what your problem is right away. It's the QL120 fans. They are horrible radiator fans. How do I know? I also have them. They're beautiful, don't get me wrong, but they are extreme aesthetics and poor radiator fans. They definitely sacrifice performance for aesthetics. I put a 3rd radiator in my system for that very reason, so I can keep my QL120 fans. Look at my post on the first page to see my system build. With this build, I'm at 35C water temperature with 23-25C ambient. Surely, it would be even better if I wasn't using QL120 fans. 
 
The only thing I can recommend for you is to either get a 3rd radiator to lower your water temperatures or get better fans. There's a user with 2x radiators like yours but he's using Lian Li Unifans and they keep has water temperatures around 35C. 
 
For now, warm the card up, turn PC off and PSu, remove the backplate since you're horizontally mounted, tighten the screws a little more and you'll probably go from being +25C over water to +20C over water. 
 

talon951
MarcSam, that's still a block mount problem. Air to water delta looks fairly typical of 2 360's depending on fan quality and speed. But a 23-25C block delta is pretty high for 400-450w. Higher flow rate would help some. 2nd pump or maybe a D5. But improving the block mount would help the most.

Trust me, I've spoken to over 20+ EKWB FTW3 owners now and +20C over water seems to be the norm.

KingEngineRevUp

MarcSam
OMITTING YOUR POST TO NOT TAKE UP FORUM SPACE 
ORIGINAL COMMENT HERE: https://forums.evga.com/FindPost/3465436

Can you upload your photos on IMGUR? For some reason they're very pixelated with the service you're using. I think your memory temperatures are around 68C? What is water temperature at? You can probably warm the card up, remove the backplate and I bet you, you'll be able to tighten the 4 screws around the GPU a little more. At most, my memory temperatures are +25-30C above water always, and my GPU temperature is 18-20C drawing 400-400W. 
 
As for your water temperatures, I'll tell you what your problem is right away. It's the QL120 fans. They are horrible radiator fans. How do I know? I also have them. They're beautiful, don't get me wrong, but they are extreme aesthetics and poor radiator fans. They definitely sacrifice performance for aesthetics. I put a 3rd radiator in my system for that very reason, so I can keep my QL120 fans. Look at my post on the first page to see my system build. With this build, I'm at 35C water temperature with 23-25C ambient. Surely, it would be even better if I wasn't using QL120 fans. 
 
The only thing I can recommend for you is to either get a 3rd radiator to lower your water temperatures or get better fans. There's a user with 2x radiators like yours but he's using Lian Li Unifans and they keep has water temperatures around 35C. 
 
For now, warm the card up, turn PC off and PSu, remove the backplate since you're horizontally mounted, tighten the screws a little more and you'll probably go from being +25C over water to +20C over water. 
 

talon951
MarcSam, that's still a block mount problem. Air to water delta looks fairly typical of 2 360's depending on fan quality and speed. But a 23-25C block delta is pretty high for 400-450w. Higher flow rate would help some. 2nd pump or maybe a D5. But improving the block mount would help the most.

Trust me, I've spoken to over 20+ EKWB FTW3 owners now and +20C over water seems to be the norm.

I changed all the images with imgur now you should be able to see them full size.
I'm adding the 2 results for the previous stress tests of last post I made:
 
Port Royale stress test Time Spy extreme stress test 
 
The two spikes (gpu and cpu) are sensors misreading (gigabyte sometimes sucks...)

Keep in mind MarcSam showed an ambient temp of 30C. That's only about 15C air to water delta. Pretty typical with moderate fan speeds.

KingEngineRevUp

talon951
Keep in mind MarcSam showed an ambient temp of 30C. That's only about 15C air to water delta. Pretty typical with moderate fan speeds.

The QL120 fans hold him back the most. They're not great fans to begin with and it's probably worse when you put them in a pull situation for aesthetics. I'm speaking from doing the same myself lol.

I was accounting for that in my assessment.    I haven't seen anything that tells me fans are worse in pull though.  EK has some tests on their website showing an improvement in pull.  They were using Vardar fans though.
 
https://www.ekwb.com/blog/push-pull-or-push-pull-on-radiators/
 
It's difficult to get really good temps with a high wattage GPU.  At one point I had slim 360/280 combo rads with good fans (Noctua A12's and Arctic P14's).  Even running those fans near full speed, the best I could do was a 10-12C water delta with a 390w GPU. 
 
Then it went up from there trying to run the KP 1kw bios at 500w or so.  About 15C water delta or a little more.  Combine that with a so-so block delta of 20C @500w, the core temp would creep up close to 60C.  It took adding a 480mm x 64mm thick external rad (6C water delta now) and improving my block delta by about 5C with a remount and second D5 pump to get it to stay below 50C.  This is with about 25C ambient.

 

talon951

KingEngineRevUp

talon951
Keep in mind MarcSam showed an ambient temp of 30C. That's only about 15C air to water delta. Pretty typical with moderate fan speeds.

The QL120 fans hold him back the most. They're not great fans to begin with and it's probably worse when you put them in a pull situation for aesthetics. I'm speaking from doing the same myself lol.

I was accounting for that in my assessment.    I haven't seen anything that tells me fans are worse in pull though.  EK has some tests on their website showing an improvement in pull.  They were using Vardar fans though.
 
https://www.ekwb.com/blog/push-pull-or-push-pull-on-radiators/
 
It's difficult to get really good temps with a high wattage GPU.  At one point I had slim 360/280 combo rads with good fans (Noctua A12's and Arctic P14's).  Even running those fans near full speed, the best I could do was a 10-12C water delta with a 390w GPU. 
 
Then it went up from there trying to run the KP 1kw bios at 500w or so.  About 15C water delta or a little more.  Combine that with a so-so block delta of 20C @500w, the core temp would creep up close to 60C.  It took adding a 480mm x 64mm thick external rad (6C water delta now) and improving my block delta by about 5C with a remount and second D5 pump to get it to stay below 50C.  This is with about 25C ambient.

I believe the QL120 fans are not that good in pull for these reasons
 
 
 

 

MarcSam
For now thank you so much for your time and knowledge guys, the block is mounted properly and at this point what I will do is flushing radiator and change fan completely, as my corsair ql are already showing bad leds and it's not the first time that I have trouble with corsair environment not to speak all the cable mess that they force you to do if you want their stuff to work as intended. So as soon as I will have time I will switch to the lian li al fans and perform a deep clean of the radiators.
 
Just for comparison
QL : 1.55 mm and 41.8 CFM
AL: 2.62 mm and 64.5 CFM
 
Not to mention that QL fans are round shaped with lots of air gap if mounted to a radiator while the square frame of the AL ensure a tight fit so a better use of the airflow.

 
Yep, just saw your post after I posted from above. I too am debating about changing fans one day when Lian Li releases the infinite mirror uni fans. 

Coolant temp will only be within 5C of ambient with a lot of rad area when running 400-500w gpus. Even with the 3 rads I have it only holds 5-6C with quite a bit of fan speed (around 1500 rpm). All 15 fans.

rjbarker

talon951
Coolant temp will only be within 5C of ambient with a lot of rad area when running 400-500w gpus. Even with the 3 rads I have it only holds 5-6C with quite a bit of fan speed (around 1500 rpm). All 15 fans.

And thats perfect....mine is around the same....likely closer to 3c...
9 Noctua Fans on Rads + 3 Case Front Intakes + 1 Rear Exhaust, so 13 Fans in all (all Intake (( except top rear)) set to 1250 - 1300 RPM which is around 75% of Max, quite quiet) ....I used to run push / pull on my Quad RX...but ditched them several flushes ago.(so was at 16 Fans back then)..the drop of 1-2c hardly worth a hassle installing them...
I have had a Bits Power Coolant sensor / plug in my box of Fittings since 2014/15...perhaps (if I dont forget) I'll plug it into a spare port and confirm what I already know ;)
Again, a flow meter (or some way to check flow) is pretty important, for me its quite technical....I leave my filling tube coming off the top of the Rez, just have to stick your ear to it to confirm the "hum" of the D5 Vario...
 
....as said several times....you will know if your Loop isnt very efficient, just by monitoring your GPU/CPU Temps...plugged cold plate fins (most common) on either of your blocks (GPU or CPU) will result in higher than "normal" temps...this is usually gunk from the inside walls of flex tubing and +12 months since last flush.
 
I am notoriously lazy and have gone anywhere from 12 months to 20 months (my record) between flushes...Even at 20 months my temps were good, slightly higher than initial fill (maybe 5c).....
Water Cooling n Loops in general are far more robust and "idiot-proof" than what newbies give them credit for...
 
The gains over air are so dramatic...last Air Cooled Card I ran was SLi 680 Ultras....they ran hot, as in low 70's....my 780Ti's on air were the same SLi 780Ti's running around 75c....threw a couple of Blocks on em and I was blown away, as in 39c under full load gaming...crazy...have never considered Air since!
 
Enjoy your Loop ;)

No one is telling you to plug your temperature sensor in... But there are benefits to having one. Fan profiles should be made off of water temperatures for the best control over sound and taking care of those water temperatures. It's a cheap $10 investment. I just disagreed with your statement that they're not needed. Water cooling in general isn't "needed" but there are better ways to do things in the hobby. Having a fan profile based off of temperature sensors is one of them.