Fixed or variable speed on the pump..?

no_1_dave

no_1_dave

no_1_dave

Soldato
Joined
7 Jul 2004
Posts
8,272
Location
Gloucester
Back when I had a custom setup I just used a fixed speed for the pump because it was pretty much silent and pushed a lot more water through the system than any AIO can even at just 50% (D5)... It was always said that pumps should be at a fixed speed otherwise they wear quicker, but this was 10+ years ago.

Fast forward to today and I'm using a Lian Li Galahad AIO, it's nice and simple and seems to keep my CPU at a respectable temperature. The problem is that being an AIO the pump has a slight buzz to it when running at higher RPMs. It's not noticeable when gaming but is noticeable when just web browsing or working from home. To combat this I've set a custom profile for the pump so it's only running at 50% until the CPU hits 50c, then it increases to 75%, this is to that I get some additional cooling while gaming. It increases to 100% if the CPU reaches 70c so it will run full speed whenever I'm encoding video.

Is this bad for the pump? Are the days of having to run the pump at a fixed speed over and it's now fine to use a curve and have it controlled via PWM?
 
It looks like its a DDC variant - the standard version has a PWM variant so it should be OK unless I'm mistaken.
 
no_1_dave said:
Is this bad for the pump? Are the days of having to run the pump at a fixed speed over and it's now fine to use a curve and have it controlled via PWM?
Click to expand...

I'd say you're fine running it variably. I can't say it with any certainty, but I'd imagine running it at 50% when idling and only ramping it up when needed may actually be better for it long term. Less wear and tear. It may perhaps be not so good for it if it's constantly going up and down every few seconds, but other than that I would guess it's fine!
 
While the water may reach equilibrium
Regardless of pump speed
The temperature of that equilibrium
Won't be the same
So say pump is 25%
Yeah all the coolant will end up equalised/heat soaked
But the temperature of the coolant
Will be higher
Than if the pump was at 50%,75%,100% etc
Usually over approx 80--85% pump speed
There's very little gain by going to 100%
 
Cheers for the replies, it puts my mind at ease so I'll leave it like this. It means I have a pretty much silent PC until I game or encode video which is nice :)
 
Mcnumpty2323 said:
While the water may reach equilibrium
Regardless of pump speed
The temperature of that equilibrium
Won't be the same
So say pump is 25%
Yeah all the coolant will end up equalised/heat soaked
But the temperature of the coolant
Will be higher
Than if the pump was at 50%,75%,100% etc
Usually over approx 80--85% pump speed
There's very little gain by going to 100%
Click to expand...

But then it will spend less time passing across radiators being cooled :)

With more rads you could slow it right down. But not so slow the pressure becomes weak.
 
Last edited:
Shouldn’t be an issue.
Pretty much every Asetek based AIO that I’ve tried presented high pitch noise, some less, some more. Not as loud as a GPU under load, but noticeable because it was really annoying.
I would suggest to set a higher CPU temperature as trigger for the pump to run at higher rpm. 50C can be easily reached in spikes. 60 or 65C is where the CPU is really working and may benefit from higher pump speeds.
Most of the time, increasing the pump speed won’t neg as much benefit as doing with the fans cooling the radiator. Just a tricky balance, I guess, but I managed a decent fan curve without letting the components overheat, far from it. Sometimes to drop 2 or 3C, in unnecessary scenarios, the trade off is not worth. One may be pushed to obsessing about the perfect low temperature. Done that. :D
Most AIO and CPUs won’t require 100% pump speed. If they do, either mounting pressure/thermal paste isn’t adequate or AIO can’t cope with CPU. Also, quite often the motherboard is pushing too much voltage for the CPU.
 
Mcnumpty2323 said:
While the water may reach equilibrium
Regardless of pump speed
The temperature of that equilibrium
Won't be the same
So say pump is 25%
Yeah all the coolant will end up equalised/heat soaked
But the temperature of the coolant
Will be higher
Than if the pump was at 50%,75%,100% etc
Usually over approx 80--85% pump speed
There's very little gain by going to 100%
Click to expand...
I found that running the pump at 100% was only 1-2 degrees cooler than running at 25% so I have a curve set so that it runs around 25% to 45degrees and then ramps up from there if needed.
 
Gavster said:
I found that running the pump at 100% was only 1-2 degrees cooler than running at 25% so I have a curve set so that it runs around 25% to 45degrees and then ramps up from there if needed.
Click to expand...
Yeah will depend on what pump
What rpm 25% is
What your actually cooling etc
That was just a ballpark figure as an example
 
The other thing to consider with variable speed pumps, or anything variable speed for that matter, will be if it goes through a resonant frequency as the speed ramps up. You obviously don't want your custom pump curve to ever sit at a resonant frequency if there is one. However, you can usually change the actual resonant frequency by changing the mounting pressure a tad, since the frequency is really fundamentally related to the geometry/stiffness of the whole arrangement, so tightening up or slacking off changes the stiffness. Adding mass will also change an assemblies resonant frequency, but probably not really possible to do in practice on a pump/cooler.
 
You must log in or register to reply here.
Back
Top Bottom