Discussion in 'Watercooling' started by WhereHere1, Aug 11, 2017.
my pump is connected to "CPU fan" on my motherboard. Thus when cou gets hotter it pumps faster?
it helps but to a point ... its fine if you have your AIO connected to a cpu fan. If its a custom loop i would rather connect it to a molex/sata
In theory, yes. The more water flow the faster it can transfer heat.
In practice not really to a point it would matter.
It makes little difference once you're above a certain level of flow, which I can't remember off the top of my head. 1.8gpm seems to ring a bell.
Only reason for me to slow a pump down is due to noise, so is it noisy at 100%?
and the less time it spends in the radiator cooling - Theory blown
Didn't seem to make any difference to my temps when I tried.
Only time I run it at 100% now is when bleeding the loop. Other than that it's around 40%
Minimal difference, set your pump to settle at a level where noise is acceptable. If it's pwm on the CPU header, then set the CPU curve in bios to silent.
On a PC setup minimal difference, most people run their pumps slower due to noise levels, I only run mine on 2 / 5 for this reason.
Also the less time it spends over the CPU/GPU getting hot. Theory un-blown.
It makes virtually no difference unless you have a very low flow rate.
It should be a easy thing to test.
Set the pump to full power.
Run a cpu intensive benchmark until everything is up to temperature.
Then slow the pump down and see if there is any temp change.
I bet it's less than 1 degree from almost silent to full power with most modern pumps.
Increasing pump speed reduces delta between the hottest and coolest points in the loop. I have CPU+GPU then rads, with temp probes before and after the CPU+GPU. At the lowest speed the delta can get to about 7C whereas the fastest has a delta of about 2-3C (off the top of my head). High speeds increase CPU temp and decrease the GPU temp by a couple of degrees. I use the lowest speed for silence.
Only time I see a significant impact is if I am running 3/4 GPU's in parallel with the water entering from bottom GPU and leaving the bottom ones. In parallel the flow splits 3/4 ways mostly equally, but with too low of a rate will not make it into the top GPU. Aside from that delta is usually slightly higher if running at very low speeds, but a non issue for most the part.
Personally, Just find the speed at which flow is sufficient without being noisy and done.
The part your missing in your maths is "Waste heat from pump dumped into water"
Thus why it makes very little diffrence as this waste heat undoes any improvement from higher flow.
I've tried multiple speeds with my d5 and haven't seen any notable difference.
Never tried when I was running multiple rads and GPUs but i'd imagine it would of made some difference in that scenario.
Not really no. As long as your getting enough flow then your good to go.
I have tested from 1-5 on my d5 with a degree or two difference between the two.
All I really go on is which speed is the quietest for me.
google Thermal Dynamics.
If i walk a 100 meter track for 10 mins and you ran , you'd reach the end and back more times but we'd both been on the track for 10 mins . So have spent the same time in the radiator .
Both sides there are points to be made, if your planning loads of pends, head pressure will be key for you
If you are going to control the speed of anything for temperature control by using the cpu fan header on the mobo then clearly it should be the fans on the rad and not the pump. If you have PWM fans you can daisy chain them so one fan controls the speed of all, or if they are all normal 3 pin, then splice them together.
but less time for the rad to remove the heat.
Unless you blocks are killing the flow rate (say Cpu + 2-3 cover blocks in parallel) increasing pump speed is unlikely to improve performance but increase noise.
Increasing my D5's pump speed I only do to help get rid of bubbles quicker in some cases.
The only real benefit of faster flow is to make the flow turbulent instead of laminar which helps with heat transfer to the fluid as you remove the static boundary layer at the surface of waterblocks and also the radiators.
Separate names with a comma.