Are voltage and rpm directly linked ?

[Gryz]

Last week installed watercooling. With a pump (Laing DDC-1T).
The DDC-1T makes more noise than I would like.
I plan on under-volting the pump. With a fan-controller.

The DDC-1T runs at 12V and uses 10W.
I've read that the pump can be undervolted to 8V or maybe even 7.5V. You need to be careful, because the pump won't start reliably at voltages lower than 7.5V. So I plan to run my pump at 8V.

I don't have a volt-meter.
But I can read the rotations-per-minute via the motherboard sensors. The DDC-1T has a 1-wire connection with a 3-pin plug for reading the rpm via a motherboard sensor.

So my question is:
Is the rotation-per-minute of a pump (or fan) directly and linearly linked to the voltage ?

In other words, if I reduce voltage by 50%, will the rpm also go down 50% ? If I reduce voltage from 12V to 9V, will my rpm go down 25% too ?

Today, at 12V, my pump runs constantly between 3840 rpm and 3900 rpm. If I adjust the fan-controller so that the pump runs at 2600 rpm (66%), can I then assume my pump gets 8V ?

 

[dma0991]

Voltage and RPM should be linked with each other. My Lamptron FC5v2 has readings for RPM and voltage or temperature. If I set it at a lower RPM, the voltage will match the stated RPM. However, this isn't important as you will only need to worry about the startup voltage required. I don't undervolt my pump but my fans do exhibit such behavior where it will not spin till a minimum voltage is applied and I think it is applicable to pump as well.

 

[Gryz]

Anyone else who can confirm this ?

I just don't know enough about old-fashioned electrical laws to feel confident that when I undervolt until my pump runs at 2600 rpm, that the voltage indeed is 8V.

 

[p_monks33]

Be careful using a fan controller on a water pump, the amperage draw of starting the pump will burn up alot of the lower end controllers. I had this happen trying to undervolt a phobya dc12-400, luckily the controller took its own life sparing my other components. You may look into getting a molex connector that is wired to 7v.

 

[Eureka]

Yes and no, electric motors are dependent on voltage when the load is satisfied by the current driving it. If you have enough amperage to run the motor, then the RPM of a motor powered directly (using analog voltage) will spin linearly to the voltage input.

The 7V thing that p_monks33 is actually pretty easy to do, if you have a small knife or pry tool. Just replace the ground wire with the 5V wire (red), and you'll have a 7V pump.

 

[Gryz]

I plan on using a Xigmatek Monocool Fan Controller.
http://www.xigmatek.com/product.php?productid=94&type=feature

It is rated to deal with 10W total output.
It has 3 outputs for 3 fans. But when I use it only for the pump, that 10W rating should be sufficient. (My own judgement. And I checked with the webshop).

I could not easily find fan-controllers that had 10W or more, and did not have the requirement that they needed to be mounted in the cd/dvd-bay. This was the highest wattage controller I could find that can be mounted in a PCI slot. (I would have preferred a fan-controller that I could put inside the case, so I never unintentionally could change the setting).

If I understand correctly, if I turn down the fan-controller, or the pump uses less than the max-wattage of the fan-controller, the rest of the wattage will be turned into heat ? Right ? So if I had used a 20W fan-controller with my 10W pump, I would add 10W extra heat into my case ? And in my case, if I use my 10W fan-controller on my 10W/12V pump, if I turn down the pump to 8V (66%), that means the remaining wattage (33%) will be turned into heat ? So the fan-controller will output 0.33*10W=3.33W heat by itself ?

I know about the 7V trick. Unfortunately I've read on a few places on the net that the Laing DDC-1T does not start reliably when voltage is below 7.5V. That's really unfortunate (missing by half a volt). I just don't dare to put 7V on it. The official minimal starting voltage seems to be 9V even.

My case is reasonably quiet now, and with an undervolted-pump I hope to not hear my pump at all. If that is the case, I will not easily be able to hear when my pump fails to start ! And that is really risky. In theory I could add software to check the pump-speed via the motherboard. (I have it hooked up. Rotations can be seen in CPUID HWMonitor). But it adds more complexity, and thus more ways to fail.

Last question.
I assume the starting voltage is also related to the pressure of my watercooling loop ? In other words: the longer the tubes, and the more blocks and radiators in the loop, the harder the pump must work to start up ? And thus the higher voltage required ?

I have a cpu-block (Zalman) and a gtx680-block (EKWB) in my loop. Plus an old Zalman Resorator-1 reservoir. Today my loop depends mostly on the heat-capacity of the 3 liters of water, and less on the heat-spreading of the fins on the reservoir. Works fine, but after 2 hours of gaming, the reservoir does get warm (35-40C I guess). I ordered a Black Ice Stealth 240 radiator. And maybe I'll put a old, small NB-block on the bottom of my pump. If the undervolting of the pump doesn't work well enough, I plan on wrapping my pump in noise-isolating foam. The 18W Laing pumps supposedly get very hot. The 10W pumps less. I hope an undervolted 10W pump will generate so little heat, it can be removed by the water flowing through the pump. If not, I could add a block to the bottom of the pump. That would maybe add more to the pressure of the loop. And thus require higher starting voltage of my pump ...

I should have bought a 20 meter HDMI cable and USB cables, and then put my PC in another room. Much simpler than watercooling.

 

[EarthwormJim]

It's not a perfect linear relationship. You have non-linear parts within a fan.

It's pretty close though.

Here's a graph Martin produced for one of his fan round-ups.

http://martinsliquidlab.org/


You should decouple your pump before messing with voltage control on it. The majority of the noise a pump makes is it simply rattling against your metal case.

Sitting it on a sponge or foam would suffice.

 

[Gryz]

My pump is already decoupled.
The best test is to just hold it in my hand. That makes it clear that the noise is coming from the pump, not from vibrations.

(And FYI, I also replaced the top of the DDC-1T with a custom top (AlphaCool Plexiglass). That is supposed to improve performance, and lower noise).

Thanks for the chart. I assume that a chart for pumps would look very similar.

 

[Eureka]

If I understand correctly, if I turn down the fan-controller, or the pump uses less than the max-wattage of the fan-controller, the rest of the wattage will be turned into heat ? Right ? So if I had used a 20W fan-controller with my 10W pump, I would add 10W extra heat into my case ? And in my case, if I use my 10W fan-controller on my 10W/12V pump, if I turn down the pump to 8V (66%), that means the remaining wattage (33%) will be turned into heat ? So the fan-controller will output 0.33*10W=3.33W heat by itself ?

Depends on the fan controller, I think any nicer fan controller will have ways of not dumping 10W back into the case

Last question.
I assume the starting voltage is also related to the pressure of my watercooling loop ? In other words: the longer the tubes, and the more blocks and radiators in the loop, the harder the pump must work to start up ? And thus the higher voltage required ?

Amperage. Higher loads will cause a surge current, not a surge voltage, usually.

Replies in bold...

 

[EarthwormJim]

My pump is already decoupled.
The best test is to just hold it in my hand. That makes it clear that the noise is coming from the pump, not from vibrations.

(And FYI, I also replaced the top of the DDC-1T with a custom top (AlphaCool Plexiglass). That is supposed to improve performance, and lower noise).

Thanks for the chart. I assume that a chart for pumps would look very similar.

Can't assume that since they're under far more load than a fan is. Once you drop below a certian threshold, linearity will probably dissapear.

Here's another graph from Martin's. It's not RPM vs. Voltage, but if you compare the % change in voltage to % change in RPM, it's pretty much linear within a reasonable operating range.