I finally got to install my 3080Ti FTW3 HC and was wondering if the block has a directional input/output? When setting up my loop, I assumed the typical left side (near I/O) as input, right side (near pcie cables) as output. I was thinking of reversing it to have a cleaner tube routing, but would that affect temps/flow or would reversing it be insignificant?
 
Also, are these temps normal under load?
GPU temp: 52C avg, 60C peak
GPU hotspot: 58C avg, 66C peak
GPU mem: 58C avg, 64C peak
 
They seem a bit higher than I expected, but I'm comparing to my last waterblocked GPU which was a 1080Ti FTW3 with an EK block. I have a HWL GTS360 + GTS280, both with Noctua NF-A12x25s as intakes running at about 900RPM, coolant sensor reads around 38-41C. I do have a 3900X in the loop, but it was only outputting about 60-70W on average. 
 
Lastly, is it worth taking apart the block and repasting with some Kyronaut or is the stock paste good enough?

Flint 1760
Your coolant temperatures are a bit high and if that is your idle temperatures they are very high.
 
What is the ambient temperature where your system is located?
 
Are you overclocking the CPU or GPU?
 
What case are you using?
 
Where are your rads in the case, i.g., top, bottom, front, etc.?
 
How many non-rad fans are in the case and are they intake or exhaust?

That's coolant temp at load, during idle coolant is at 32-33C. Difference between idle to load is 5-9C, which is reasonable. 
 
I have the A/C set to 73F, but room ambient is about 75F.
 
CPU has FCLK Enhancement + PBO turned on, but is otherwise stock. GPU is also stock since I was trying to find a baseline for temps. 
 
Case is an O11 Mini.
 
Rads: GTS360 at the bottom, GTS280 at the side, both are intake (5x NF-A12x25s)

Non-rad fans: I have another 4x NF-A12x25s as exhaust (3x top, 1x rear)
 
Airflow is basically setup as a chimney style with a slight positive pressure (5x intake vs 4x exhaust). 
 
I have a Swiftech MCP35X, one of the strongest DDC pumps, which runs at 2K RPM during idle and a little over 3K RPM during load. Even at the max 4.5K RPM, it doesn't make a dent in coolant temps, so flow isn't an issue. 
 

Flint 1760
You have plenty of air flow, btw I have the same fans for both rads and air in my systems.  I still think that your temps are a little high, but the 3080 Ti might be culprit.
 
As a short test, you if can raise the case a couple of inches higher, i.e., put blocks under the feet, run your stress tests to see if there is a difference as the fans have a less constricted air flow from the bottom. 
 
Another question, are you vertical or horizontal mounting the GPU?  If it is horizontal, another quick test is to take the side panel off to see if you have a temperature difference as you have the bottom rad exhausting onto the GPU block in a small case and it may be disrupting the air flow.
 
As far as re-pasting the GPU, it is a viable option.
 
I have not used the O11 Mini, but my wife's new build is a O11 Dynamic and I have a raised vertical mount for her GPU, if and when it finally arrives (waiting in the queue), that allows me to put at least one maybe two fans at the bottom.  It has top and side HWL rads.
 
Strictly a personal preference, I always mount one rad, the thickest, on top as exhaust and the other either front/side as intake depending on the case.  
 
 

Cool thanks for the insight.
 
I do have the vertical mounting bracket for the O11 Mini, so airflow is much better than when it was in the normal/horizontal position.
 
With how the vertical mounting bracket works for the O11 Mini:
-configuration 1, allows rad+fans (60mm) at the bottom and only standard fans (25mm) at the top
-configuration 2, allows rad+fans (60mm) on the top, but there would only be 20mm of allowance at the bottom, preventing me from using standard 25mm thick fans
 
That leaves me with only configuration 1 as the only viable option, so rad+fans have to stay on the bottom. 
 
I guess the 3080Ti just runs super hot despite having a waterblock on. I'm still curious on if there is a proper direction for the Hydrocopper block and if reversing the input/output would affect temps. 
 
 

B0baganoosh

Kokin
I'm still curious on if there is a proper direction for the Hydrocopper block and if reversing the input/output would affect temps. 

 
There is. It is usually with the inlet towards the IO ports. Looking at the LED-ON image from the product page:

(if you open in a new tab it gets much bigger).
 
You can see underneath the top of the "E" from the EVGA logo that there's a little triangle/arrow thing, as well as underneath the bottom of the "G" in "GEFORCE". The points of those go towards the inlet. They aren't a flow indicator per-se, but they are made as flow-control so that the liquid is pushed from the point-outwards, not pushing against the flat side of the triangle. They are much clearer in the KP-block:

They go the same way in this block^. inlet is on the left, outlet on the right. The little triangles spread the fluid out as it flows from left to right in the image above.
 
Edit: I remember reading from another post that it will still function if you hook it up backwards, but the flow is more restricted and performance may not be as good. The kingpin block installation guide indicates this by saying left-to-right is "recommended" and right-to-left is "acceptable".

Whoa I never noticed those triangles! The front of the block is smoked, so it's hard to make them out even on my own block. This is a nice bit of info, thank you for confirming the flow direction.
 
Do you happen to know if others have found swapping out the thermal paste worth it? Before I installed the GPU, I was able to verify that the memory thermal pads were installed properly from the factory through the cracks between the block and PCB. 

Your results are when you have the PL set to 112%? What happens when you run a timespy or port royal stress test, what are the temperatures then? 

KingEngineRevUp
Your results are when you have the PL set to 112%? What happens when you run a timespy or port royal stress test, what are the temperatures then? 

These results are just from running the stock 100% PL, HWiNFO and GPU-Z are showing the GPU board power hitting 400W.
 
Runnng TimeSpy on loop as a stress test, I still reached the same load temps as 2-3 hours of Apex Legends.
 
Ambient ~24C (75F)
Coolant Sensor
Idle 32
Load 40-41C
 
GPU - 3080Ti FTW3 Hydro Copper (stock paste/pads)
Idle 31C
Load 54-55C
 
CPU - 3900X with TechN AM4 block + Kryonaut applied thinly with applicator
Idle 37-40C (probably lower but Ryzen is sensitive to monitoring S/W)
Load 50-55C
 
Noctua NF-A12x25 fans were set to run at 1100-1200RPM and pump was running at 100%. HWL GTS360 at the bottom and GTS280 at the side, 5x intake fans on the rads, 4x exhaust fans at rear/top. Removing case panels helped with the overall heat soak, but load temps didn't budge down. 
 
I got a graphics score of 20,154, which seems to meet stock performance. GPU clock was running between 1905/1920/1935 throughout the loops. Mem was fixed at 1187.7
 
 

rjbarker
FYI
I just installed a Velocity block on my 3080Ti:
Ambient 19c
Idle GPU 21c/22c
Full load GPU looping Port Royal approx. 1/2 Hr 43c / 44c - pretty much the same temps I was getting with the non Ti version (perhaps a few degrees c higher, but negligible)
 
All stock pads and Kryonaut (spread with applicator very thin)
 
480*360*240 Rads (RX XSPC low FPI)
Noctua Fans as well but I run them around 1250 RPM (All intakes as well).
 
Your load temps seem high, but I have never used a Hydro Copper, so no idea what the expectations are.

Yeah these load temps are about 5C higher than I expected, which was more towards 45-50C. I understand it's a 400W+ GPU and it's also the first time I've used a Hydro Copper block so maybe my expectations coming from previous EK blocks were a bit too high? My last waterblocked GPU was a 1080Ti FTW3 with an EK block and that was running a 330W vBIOS, which ran between 40-45C. 
 
That's why I was also curious if replacing the stock paste would benefit me, since Kryonaut is slightly better than most standard OEM paste. 
 
GPU-Z (open in new tab to see a bigger res)

talon951
That looks about right for water and GPU temps to me. A 14-15C block delta at 400w is about as good as people do without going to significantly more effort. Re-paste might get 2C.

The higher wattage GPUs increase both the water and block deltas, so it gets increasingly difficult to cool them. I found that out last winter when I built a loop for my 3090. Probably took me 6 months to finally get where I wanted to be. Re-mount, more rads, 2nd pump, better fans.

I wouldnt be happy at all if my 3080Ti / 3080 was 55c - 60c load temp under water. Then again I have no idea what to expect from a HC Block....or from less Rads, BUT my eyebrows would immediately go up and I would be thinking "****" if my GPU temp reached even 50c under full load ;)
I have yet to see anything more than 46c (full load 2 hrs into CB77) ......430w !
Oh....and I forgot again to install that little Water Temp Sensor when I drained and swapped out GPU's :)
 

Kokin

rjbarker

talon951
That looks about right for water and GPU temps to me. A 14-15C block delta at 400w is about as good as people do without going to significantly more effort. Re-paste might get 2C.

The higher wattage GPUs increase both the water and block deltas, so it gets increasingly difficult to cool them. I found that out last winter when I built a loop for my 3090. Probably took me 6 months to finally get where I wanted to be. Re-mount, more rads, 2nd pump, better fans.

I wouldnt be happy at all if my 3080Ti / 3080 was 55c - 60c load temp under water. Then again I have no idea what to expect from a HC Block....or from less Rads, BUT my eyebrows would immediately go up and I would be thinking "****" if my GPU temp reached even 50c under full load ;)
I have yet to see anything more than 46c (full load 2 hrs into CB77) ......430w !
Oh....and I forgot again to install that little Water Temp Sensor when I drained and swapped out GPU's :)
 

That was my knee-jerk reaction too. I even moved up from a 240 rad to a 360 rad at the bottom in anticipation of the higher heat load. It's a 400W GPU with a CPU that pulls anywhere from 60W-150W depending on usage, so I guess I've approached the limits of my thin rads/overall loop setup.
 
Unfortunately, I'd have to give up the vertical mount if I wanted to reuse the 240 up top or do 2x 360s + 1x 280 and that might net me another 3-5C improvement. It doesn't seem worth it for me right now and might consider trying undervolting as an alternative in the future. 

Yes the thin Rads are great for room....but not not the most efficient heat exchangers.....I have had these XSPC RX Rads (480 / 360 and 240) in my loop forever (since maybe 2014), just keep recycling them.....
https://www.xs-pc.com/radiators-rx-series
 
I have mentioned it in some other threads, for optimal heat exchange low FPI (Fins Per Inch) rads along with low RPM high static pressure fans are best and set up rads as intakes.....fans at 1250 RPM is a good speed, my pump sits at around 3000 - 3200 RPM (although its a variable speed pump, there is no need to change the speed).

Kokin

rjbarker
Yes the thin Rads are great for room....but not not the most efficient heat exchangers.....I have had these XSPC RX Rads (480 / 360 and 240) in my loop forever (since maybe 2014), just keep recycling them.....
https://www.xs-pc.com/radiators-rx-series
 
I have mentioned it in some other threads, for optimal heat exchange low FPI (Fins Per Inch) rads along with low RPM high static pressure fans are best and set up rads as intakes.....fans at 1250 RPM is a good speed, my pump sits at around 3000 - 3200 RPM (although its a variable speed pump, there is no need to change the speed).

I have an older RX240 V2 + EX240 from XSPC, but I found the HWL GTS series to be comparable to the XSPC RX series, which is insane for the thickness difference (30mm vs 56mm). 

Thats right 56mm thick....but look closely at the FPI vs other Rads....
Its another reason why I have held onto this "old" Corsair 900D case...with those 56mm rads...I need all the space I can get !!
480 up top...the 360 n 240 down below sides.

Is 45c bad? I'm asking for real because I just got a 3080 ti xc3 hc, and it'll hover around the same temps as yours.

Yea, I upgraded from a 2080ti, and since the 3080 ti is more power hungry, it was only natural temps slightly raised.

Man I guess I should be happy with my 38-41C under load temps with the DDC pump running around 2800rpm & fans(Arctic P14's in push) running around 900rpm. Granted I have two EK 420mm CE rads for just the 3080Ti HC.

You really can’t use coolant temp alone as a yardstick of performance. There is significant variability in indoor ambient temps due to geography . So that right there is going to change your GPU baseline temperatures significantly just because of the coolant ambient temperature. If you’re using sub ambient coolant that is a whole different animal.

A better quantification of block performance and contact is coolant to gpu temp delta. Depending on what you’re doing you might look at gpu to gpu hotspot or gpu to gpu memory junction as well. So for example, if your coolant to gpu temperature delta is 30C, that is not good. If it’s 15-20c that’s not bad. The HC block is good but it’s not full coverage nor is there an active backplate. Less than 15c delta? Nice job tell us your secret.

I think the only time it makes much of a difference is if the block has a jet plate like most CPU blocks have.  Or like this Heatkiller GPU block,
 
https://www.techpowerup.com/review/watercool-heatkiller-v-for-nvidia-rtx-3080-3090-ebc-backplate/4.html
 

Idle seems high but load seems ok-ish. Maybe repaste?