High water temperature O11D Evo Reflection 2 Distro

Just resurrecting this old thread.

Out of interest I noticed that at idle the GPU temp is sometimes up to a degree or so colder than reported water temps.

I have been googling, but I presumed that the water should be colder than component temp, otherwise it wouldn't be cooling. Perhaps that applies more to when the PC is under load.

Any scientific thoughts? :cry:
 
A degree or so
Is within margin of error
Especially since got 2 different things
Reporting the temperature
But yeah the principle is the same at idle
As it is under load
Just that you're dealing with much bigger margins
Once it's under load
 
Just turned pc on
Coolant shows 4c lower than gpu
Left sitting at idle for a while that gap
May close a bit
As the cpu isn't totally idle unless I shut down
Every background process in task manager
But still at the mercy of 2 different sensors which
Are reporting from 2 different places
 
Thanks for the responses, that's great info.

I've just been making an offset in Aquasuite as soon as I see the minimum GPU temp go lower than the water temp at idle.

That probably won't give me a completely accurate reading, but I have so far observed about a 2C max difference (that may be in margin of error) but I am going with the theory that water should be cooler than component being cooled (to partly satisfy my insanity!)

That only really tells me that GPU temp equals water temp at idle and if that was the case then my idle temps are only about 5C off ambient temps, which seems way more reasonable.

I suppose this is all without actually putting a thermometer in the water itself and I have promised myself not to tinker with the PC anymore. See how long that lasts!
 
Last edited:
Left it running at idle
And shut down all background processes
To try to remove cpu from equation
as much as possible
Since I posted that 4c difference about 50 minutes
Running at idle now
The difference now is still 3.2 to 4c
All fans at 1,000rpm
If you're seeing 2c difference
That seems reasonable even if I am seeing 4c
Due to may be difference in radiator thickness
Static pressure of fans etc
 
Mcnumpty2323 said:
Left it running at idle
And shut down all background processes
To try to remove cpu from equation
as much as possible
Since I posted that 4c difference about 50 minutes
Running at idle now
The difference now is still 3.2 to 4c
All fans at 1,000rpm
If you're seeing 2c difference
That seems reasonable even if I am seeing 4c
Due to may be difference in radiator thickness
Static pressure of fans etc
Click to expand...
Brilliant thanks.

Sorry, should say 2c difference between reported water temp and what I offset it to (to make the GPU same as water at idle.)

That then equaled 5c above room temp.

Approx. 32c GPU and Water Temp (with offset)
Room temp: 27c (that is a problem in itself as is hot af and probably a major contributor.)

Any idea what your room temp is?
 
Last edited:
Sorry no idea what temperature is in here
Typically batteries have ran out
On my thermometer
And they're some sort of little watch style batteries
Not got any laying around

I would probably set it to make
Coolant 2c cooler than the gpu temperature
Assuming what I am seeing is coolant temperature
Is slightly lower than gpu temperature
Which is sort of the theory
You mentioned earlier about scientific thoughts
Though theres probably more to it
But the coolant being slightly cooler than the component
Does make sense
And gpu is relatively stable it shouldn't suddenly
Be spiking like a cpu does

5c above room temperature doesn't sound
Like it would be too far off though
If I had to hazard a guess
I would guess 24c in here and bbc weather says 22c outside
Not very scientific me guessing temperature I know
But that would make my coolant 7c above room temperature
 
Moved onto putting the gpu under load
Which is also putting cpu under load
Though 100% for gpu
But not 100% for cpu
15 minutes so far looping unigine heaven benchmark
Might not have heat saturated the coolant yet

Very interesting though
The gap between gpu temperature
And coolant has widened an awful lot more
Than I expected
17c of a difference now between them
Though 4 of my 13 fans automatically turned up 300--400rpm
Forgot those were on a curve
All the others are still set manually same speed as before
Bit unfortunate as same speed throughout would be
Preferred
Shows how slow I type/think lol
22 minutes now gpu benchmarking
Coolant seems to be maxed out or very close to it
Can't see it rising more than another 1c or 2c

Other than 3 fans did increase 300-400rpm
That suggests some sort of efficiency curve
Ie at idle the radiators weren't
Being used to their full potential?
More energy/heat being dumped into them
They are more efficient?

Could get a headache thinking through this lol
27 minutes gpu benchmark now
Still 17c difference between coolant and gpu temperature

Started at
Coolant 28c
Gpu 31--32c

Now at
Coolant 38c
Gpu 55c
So coolant rose 10c
Gpu Rose 23c
Does that mean in effect the radiators and fans
Are dumping 13c heat difference?
It certainly feels warmer in this room than before
I ran gpu test
Though sadly got no before and after room temperature
And kind of think temperature would vary a bit
From one place in the room to another
Guess you would need to know the volume of air
In the room etc to work out how much energy from the pc
Would raise temperature in every single part of the room

Should say my pc is 2 feet off the floor
And near an open window
So that may also affect my results
But am still surprised the difference between coolant
And gpu temperature widened so much
Once under full load

No idea if all that's of any use
Or will just make you scratch your head even more lol
 
N80VrWK.jpg


What wattage is the GPU putting out?

Just played Resident Evil 4 for an hour.

FE 4090 at 80% power limit. +230 core and +1200 memory.

It can do into the +3000MHz in many games with voltage and power increases. Running at above 100% power limit just seems to send the temps over what I am happy with.

PC right next to window, but not ideal with having it open so close to the PC with unpredictable weather (PC running 24/7). So usually have the door open and try and create some airflow into the room. PC on the desk though.

(Offset of -1.75 applied to water temp to bring it slightly below GPU temp at idle.)

I get a fairly similar 17c difference (as you) between water and GPU temp with these settings.
 
Last edited:
Forgot to look at gpu watts
Normally I have a 10 inch tablet
Sits next to the pc
And have an aida64 sensor panel display on it
But loaded a different windows image
To test something else
And the sensor panel isn't set up on this image

Would probably be 240w--250w
So still a fair bit less than yours even at your 80% limit
So yes you may see a much smaller gap
Than I did under load

Must admit 250w is about as much as I would want
My gpu using
The shove 500-- 600w into gpu to get maximum performance
Doesn't appeal to me
I guess unless they find something new
It's hard to keep increasing performance generation
After generation without just banging up power demand
 
OK, crazy observation...

This is so dumb but I am not sure why I never thought of trying this before!

I use Aquasuite to control my High Flow Next Flow Meter. I just thought I would try another quick reset to default settings.

This time I didn't bother syncing all my silly RGB with the rest of my system. In fact I turned it all off and used the screen brightness on lowest settings.

I observed the water temperature go from around 32C to 30C taking it slightly under the GPU temp (like when I added my offset to make it match GPU temps at idle.)

So, as I believe someone else pointed out previously, something was heating up the displayed water temp (in a shutdown situation as well.)

I have also set it up to turn off RGB when shutdown to see if just the display can heat it up. I mean you need the display at least at shutdown, as you can't see Aquasuite.

Anyway, such a small detail, that I never would have thought would make such a difference!

Anyone with a flow meter with RGB and high screen brightness should know it can seriously heat up the unit itself and give incorrect readings.

I am just testing at full load now to observe how that reacts.

I mean, I should probably partly thank the flow meter for contributing to me reseating the GPU block. I think that needed doing anyway though.

I feel like this should be in the instructions or something though for flow meters!
 
Last edited:
Bit late to the party, so apologies.

I had a similar custom loop setup to you OP.

- Trio X 4090 and 7950x on the same loop
- 1 x EK PE360 and 1 x EK XE360 Radiator with both in a Push configuration
- Had 6 x EK Vardar 120ER RGB Fans
- 2 x DDC Pumps in series as I had a lot of 90 degree fittings in my loop to manage, including a Distro Plate as well
- Flow Rate at 40% on both pumps was around 130L/HR when the PC was on idle. Would go up to 225L/HR during gaming sessions.

On idle and with fan speeds of around 500-600RPM, my Ambient Temps were around 23c and Water Temps were around 29-30c. At 800-900RPM, the delta would be around 4-5c.

With loads of around 550 - 600w (Combined CPU and GPU) during heavy gaming sessions, water temps would go up to around 42c and GPU would be around 60c with Fans around 1300RPM (60%).

I recently installed another 120mm Hardware Labs GTX 120 (54mm) Radiator in a push/pull at the rear of the case and it brings all the above temps down by around 2c during the same loads.

With thicker rads, ideally you want to setup a push/pull configuration if you can to help drive the heat out of the radiators. I would do it myself but I don't have the room in my case.

I now run my 4090 with an Undervolt of 0.945mv @ 2730Mhz with a +1500Mhz OC on the Memory, so my GPU and CPU rarely draws more than 420w total during gaming. That keeps the GPU at 46-47c and the Water at 33-34c.

From experience, I've found that EK GPU blocks can give you a water-GPU delta of around 15-20c. I've now bought blocks for a 6800XT, 3080Ti and 4090 and used the same setup as above and all fell within that delta range at really high loads (450-600w). However, if your GPU is jumping straight from idle temps to 60c immediately once you start a benchmark, that usually indicates a contact issue as the heat isn't efficiently being transferred out of the block.

Otherwise, I don't think you're doing much wrong in terms of optimising what you have for efficiency, you may just need more radiators and more fans to help keep your water temps down.
 
Last edited:
xenosys1983 said:
Bit late to the party, so apologies.

I had a similar custom loop setup to you OP.

- Trio X 4090 and 7950x on the same loop
- 1 x EK PE360 and 1 x EK XE360 Radiator with both in a Push configuration
- Had 6 x EK Vardar 120ER RGB Fans
- 2 x DDC Pumps in series as I had a lot of 90 degree fittings in my loop to manage, including a Distro Plate as well
- Flow Rate at 40% on both pumps was around 130L/HR when the PC was on idle. Would go up to 225L/HR during gaming sessions.

On idle and with fan speeds of around 500-600RPM, my Ambient Temps were around 23c and Water Temps were around 29-30c. At 800-900RPM, the delta would be around 4-5c.

With loads of around 550 - 600w (Combined CPU and GPU) during heavy gaming sessions, water temps would go up to around 42c and GPU would be around 60c with Fans around 1300RPM (60%).

I recently installed another 120mm Hardware Labs GTX 120 (54mm) Radiator in a push/pull at the rear of the case and it brings all the above temps down by around 2c during the same loads.

With thicker rads, ideally you want to setup a push/pull configuration if you can to help drive the heat out of the radiators. I would do it myself but I don't have the room in my case.

I now run my 4090 with an Undervolt of 0.945mv @ 2730Mhz with a +1500Mhz OC on the Memory, so my GPU and CPU rarely draws more than 420w total during gaming. That keeps the GPU at 46-47c and the Water at 33-34c.

From experience, I've found that EK GPU blocks can give you a water-GPU delta of around 15-20c. I've now bought blocks for a 6800XT, 3080Ti and 4090 and used the same setup as above and all fell within that delta range at really high loads (450-600w). However, if your GPU is jumping straight from idle temps to 60c immediately once you start a benchmark, that usually indicates a contact issue as the heat isn't efficiently being transferred out of the block.

Otherwise, I don't think you're doing much wrong in terms of optimising what you have for efficiency, you may just need more radiators and more fans to help keep your water temps down.
Click to expand...
That's an interesting comparison thanks.

The 4090 really generates a lot of heat towards maximum wattage.

That's a good temp you have now and will never tell the difference.

I have just overclocked the GPU as far as it can go and limited the power to 80%. I had an undervolt on my 3090 but haven't bothered to do that this time.
 
You must log in or register to reply here.
Back
Top Bottom