Most powerful PWM controllable pump?

[Zarathustra[H]]

So,

I am still in the early feasibility/planning stages of a "sound isolation by pumping water to the next room" build.

I currently have an EK dual D5 pump top with two D5 G2 PWM pumps in it.

I'm concerned this may not be enough power.

Is there anything out there on the market that has more power, and also has pwm control?

I've read of the mythical D5 Strong, but I can't seem to find it available anywhere.

If nothing with pwm control is an option, what are some suggestions for appropriate pumps that are out there?

I always though Eheim pumps were pretty badass, but looking at the specs, they are really not worth writing home about. I had interest in Aquacomputer's Aquastream pumps based on Eheim's designs, but the performance numbers are quite underwhelming.

Appreciate any suggestions.

 

[Zarathustra[H]]

Hmm. The Koolance PMP-600 seems like th emost badass thing I can find.

Requires 24v for it to really shine though.

Koolance has also discontinued the voltage controller for it, but it shouldn't be too difficult to source a motor speed controller on amazon or ebay.

What would be really awesome is if I could find one that would take a PC fan control PWM signal and translate that into 0-24v.

That may have to be a custom soldered job using some timer chips or something though. I have found a few so called "pwm fan hubs" that come with cases, and do something like this for 12v, but nothing 24 v...

Doesn't help that DC motor speed regulators essentially all use PWM internally to achieve the speed regulation, meaning that if the product I want exists, it's going to be lost in the google results...

 

[Ducrider748]

So instead of pumping water into the next room. Pump water back to the PC. That way no need for pwm. Pump can run at 100 percent and you won't hear it.

 

[Zarathustra[H]]

So instead of pumping water into the next room. Pump water back to the PC. That way no need for pwm. Pump can run at 100 percent and you won't hear it.

That is the plan, but I also don't want a 50w pump blasting 24/7, it seems wasteful.

 

[Ducrider748]

As a option is run the dual top in the next room and a single at the PC to pump back. Run them at a lower speed. Should give proper flow depending on water blocks and rads plus distance. I personally have run my pumps at 100 percent unless I was getting to much flow with only 1 block. Not allowing enough time for the water to absorb the heat. Those days are pretty much gone with much better designed water blocks. Noise is not a issue for me though.

 

[Zarathustra[H]]

As a option is run the dual top in the next room and a single at the PC to pump back. Run them at a lower speed. Should give proper flow depending on water blocks and rads plus distance. I personally have run my pumps at 100 percent unless I was getting to much flow with only 1 block. Not allowing enough time for the water to absorb the heat. Those days are pretty much gone with much better designed water blocks. Noise is not a issue for me though.

I appreciate your input, but you do realize, that's not how heat transfer works, right?

You do not need to wait for the coolant to soak up the heat. In fact, if you don't it works better.

Your blocks will be at their most efficient if the coolant flows through them so fast that there is barely any noticible increase in temperature before the block vs after. This is because there is more fresh water filling in right behind it, and continuously cooling the block.

The most effective loop is one where the flow rate is so high that there is no noticeable difference in loop temp anywhere in the loop. The coolant is not hotter after passing through the blocks, and is not cooler after passing through the radiators. It then enters a steady state with a continuous rate of cooling in and out of the coolant.

You are essentially dealing with a differential equation type of math problem, instead of a static heat transfer problem.

All else being equal, there is no case with any block ever made, that lowering the flow rate will decrease core temps, with the following two exceptions:

1.) Your pump is being driven too hard and its dumping its heat into the loop, and dialing it back a bit causes the pump to dump less heat into the loop

2.) Pressures are so high that loop friction is causing noticeable heat, that is reduced when you back off a little bit (this is extremely rare outside of high powered industrial systems)

The block will always be more efficient the higher the flow rate. (though at the temperature deltas usually seen in PC loops you do get to a point of diminishing returns where the improvement isn't easily measurable rather quickly as you get above 1GPM. Thats why 1 GPM is kind of the rule of thumb, as improvements tail off rather quickly above that point. Usually by 1.5GPM its pretty difficult to notice any difference, just because it is very small and difficult to measure.

 

[Zarathustra[H]]

So, some googling suggests that if I want to stick with 12v (which I think I do, as it is the most easily controllable) the Koolance PMP-500 is about the most badass there is out there.

The PMP-600 is crazy at 24v, but when run at 12v it is surprisingly weak.

(For comparison, the PMP-450 is a D5. I'm guessing the PMP-450S is a D5 Strong, but I am not sure. The PMP-400 is a DDC.)

For my application max flow is pretty much irrelevant, so I am looking at the greatest amount of flow at relatively high head/pressure, and for that, at 12v the PMP-500 seems pretty unbeatable. I think I'll start testing with one of them, and if that isn't enough, I'll add a second. I can even speed control them with a PWM signal going into a PWM to 12v voltage adjustment board. Those usually max out at 30w though, and the PMP-500 hits 32W at max. Maybe those last 2W won't matter and there is a little safety margin.

Or I can just configure them to not run above 93.75%

 

[Killerxp100]

I would use the d5 pumps you have now and only look for another solution if they do not provide the flow you need.

 

[cpufrost]

Iwaki RD30 on Meanwell 30V PSU does the trick.