No differance in temperature with fan speed

Hi I have a H100i cooler in this case

https://www.nzxt.com/product/detail/143-h440-performance-mid-tower.html

rest of spec in sig

I'm using real temp to monitor stress test temps after overclock. The cooler seems to be performing well temps maxing out at 70 degrees. However if while stress testing I change the fan profile and for example reduce the fan speed to 25% the temps stay the same.

What I'm seeing is at 25% fan speed or 100% fan speed there is no difference in cpu temps? Almost as if the fans don't do anything?

Is it the case? is it an airflow issue? Taking the side off the case makes no difference?

The pump as far as I can tell is permanently at full speed

Thanks

Darren

How long did you run your tests with different fan speeds? It can take several minutes at full load for the water in the loop to react to changes in fan speeds. 15-20 minutes should probably be enough to notice a difference.

Ambient temps can also play a huge role in overall cooling performance. If you're in a chilly region you will likely see lower and less erratic temperature swings. While if you have 82F room temps like some of us, a small change in fan speed can make a noticeable impact on CPU/GPU temps.

How hard are you pushing that 5930? If you're running a low OC like 4.1 your chip might not even be producing enough heat to see a big difference in temps with that cooler.

What program are you using to test? LinX or Intel Burn Test will torture that CPU more than pretty much anything else. I still test with Prime95 occasionally to test for basic math stability, but when it comes to slamming the CPU's vector units IBT/LinX will pull much more power and thus produce more heat.

Remember the water removes the heat from the CPU. The fans help remove the heat from the water as it passes through the radiator. The more water you can put into the loop, the less fan speed you need. If in fact the fans are not altering your test results, it would suggest the cooler is operating below it's capacity and before the internal water temperature rises and negatively affects the test, but I would probably need some actual data before making that a definitive statement. Most people run more fan speed than they need on these units. You definitely get diminishing returns at the higher rpms. The noise trade-off generally isn't worth it, unless you are right at your limit.

Also, if this is your first Corsair cooler, note the value of "H100i Temperature" (water temp) is different than your CPU or actual core temperatures. The fans profiles are based on the water temperature. This is a relatively slow to change value, at the bottom being whatever your room temperature is at boot, up to possibly in the mid 40's Celsius after long term load. Besides room temperature, the water temp is also affected by the internal case heat, including GPU waste heat. Hot environments lead to warmer water. This shouldn't affect your load cooling, but does tend to elevate the lower CPU temp values while warm.

Ahh so its all starting to make sense now thanks for your replies i have stress tested the cpu over night but obv not sat and watched it the whole time max temp registered was 72 degrees so imguessing thats pretty good :D

Im running my 5930 @ 1.31v 4500 (actually 4499 or thereabouts) Im happy with the cooler dont get me wrong although it can get pretty loud ha ha

Im in the UK but the house ambient is usually pretty high as the wife cranks up the heating at the slightest hint of a cold snap :laughing:

Thanks again i understand better now ive been out the rig building game for years so its a steep learning curve

Ive just had a thought if the pumpis running faster then the cooling fluid is passing through the radiator quicker does it not give the fans less time to cool it on its way past?

This may sound daft but would it not be better to run the pump slower?

Im a noob bare with me here ;)

No, slower flow isn't better because we aren't as concerned with removing the heat from the water as much as we are about removing the heat from the CPU contact plate. Theoretically, you want each molecule of water to be in contact with the plate for as little time as possible. The longer it is there, the more the molecule will heat up. Your CPU temps will always exceed your water temps, so the temperature of the water is not responsible for the cooling, but rather it's physical properties as a liquid to transfer and carry the heat away. You don't really need to worry about water temperature, but it does represent how much heat is in the loop and can be useful for diagnosing problems. In reality, the flow speed isn't the most significant variable in play. The new GT/GTX series pumps are have a high/low setting, and most people are reporting a nominal difference in CPU temperatures between speeds. I suspect in order to make a noticeable difference, you would need a pump strong enough to substantially increase the flow rate and the trade-off in mechanical noise and cost would be a tough sell.

All that aside, if you got a peak temp of 72 after an over-nighter at 1.31V, that's ^$%@ fantastic. My 5820K only comes in a little lower, but that is with a larger H110 and measly 1.25V (4.3GHz). You are good to go.

No, slower flow isn't better because we aren't as concerned with removing the heat from the water as much as we are about removing the heat from the CPU contact plate. Theoretically, you want each molecule of water to be in contact with the plate for as little time as possible. The longer it is there, the more the molecule will heat up. Your CPU temps will always exceed your water temps, so the temperature of the water is not responsible for the cooling, but rather it's physical properties as a liquid to transfer and carry the heat away. You don't really need to worry about water temperature, but it does represent how much heat is in the loop and can be useful for diagnosing problems. In reality, the flow speed isn't the most significant variable in play. The new GT/GTX series pumps are have a high/low setting, and most people are reporting a nominal difference in CPU temperatures between speeds. I suspect in order to make a noticeable difference, you would need a pump strong enough to substantially increase the flow rate and the trade-off in mechanical noise and cost would be a tough sell.

 

All that aside, if you got a peak temp of 72 after an over-nighter at 1.31V, that's ^$%@ fantastic. My 5820K only comes in a little lower, but that is with a larger H110 and measly 1.25V (4.3GHz). You are good to go.

Thank you Its all making sense now :D