Overkill 12VDC pumps

[Voodoo Rufus]

I got to thinking recently. I've been considering doing an air to water intercooler setup on my turbo Subaru for a long time, but the gains don't really justify the cost. At the same time, these systems use 12-14VDC pumps.

I think it would be a bit of stupid fun to pick up one of these and see how soundly it beats a D5 or dual D5 setup. Might even best the Iwaki setups of yesteryear. Probably a lot noisier than our consumer pumps I would guess.

Check out the P-Q curve on that EBP40! It can pull a peak of 55W, so I would have to hack up a PCIE PSU connector into a high current connector to power it.

https://daviescraig.com.au/electric-booster-pumps

 

[BugFreak]

I would imagine ambient temps and heat put in the system by the pump itself would come into play at some point. I wouldn't want to stick it in my daily rig in case something leaked or blew out but sure would be fun to see though.

 

[Voodoo Rufus]

Heat dump into the coolant flow from pumps has never been that big of a deal. An 18W DDC only does 2.7W of 'work' on the fluid. The rest is heat from the pump. Pretty minimal once you consider how much dissipation capability we have with the multi-rad setups that are popular now.

A 2000 hour continuous use warranty doesn't sound all that great, either. Only 3 months. I can't find a typical service life for the Davies Craig pumps, or MTBF.

 

[Ben333]

Solar hot water system or vehicle heating \ cooling system would be way more abuse than a computer load. I'd say its worth trying. That said, I know Cathar did testing way back and afik the results said that a D5 or eheim made more sense than a monster iwaki. (Even though the latter had better head & flow)

 

[tRidiot]

The flow of coolant with modern pumps and decently designed water cooling systems isn't really a limiting factor, is it?

 

[Liblinski]

I feel there might be a lot of room for improvement in the pump department. Sure like Ben said, there has been some extensive testing at some point in the past by trailblazers but I think that up to date data is severely lacking. Everyone is pointing back to information that was gathered with very old pump models as well as with old and out dated testing setups/methodology.

Popular belief and most shared advice are contradictory: "1gpm is the golden number, a D5 is plenty and dual d5 is overkill"

… Yeah well most single D5 pumps don't output 1gpm, far from that in most cases actually. The two dudes that recently reported ton their dual D5 Mo-RA build had to run one at full speed at to tickle the 1gpm. And when you point that out people answer is "whatever I am happy with the temps" but keep on repeating the same stale and conflicting rule of thumb.

I think it’s well overdue to think out of the box and try out weird pump!

@Voodoo Rufus What are the price and efficiency ratings of these “overkill” pumps? Noise levels? My main concern would actually be the noise.

Did you take a look at household recirculating pumps? Seriously, solar panel setups and eco-housing have output some really interesting pumps these last years. I got my eyes on one model in particular that is available in Europe: Up to 10m head, PWM control, 7-70W, <43dBa, composite or brass housing, costing less than a dual D5 setup, stellar warranties.... That's some serious arguments.

 

[Voodoo Rufus]

The Eheim pumps are pretty weak with a low performance impeller shape if I remember right, and with running on AC power you had to either run them off a switch or relay. For a simple and low pressure drop loop they'd work fine, but at that point just use an AIO. Aquarium pumps in general don't have great P-Q performance.

I'm curious about these solar heater loop pumps. Any brand names you would suggest, Ben?

The Laing pumps really did the sweet spot for performance, size, cost and noise. There's a reason the D5 has persisted more than a decade now.

If we're all going custom loop cooling and spending the money to get that performance, it may as well be done right. If one is going for style, then your component choice is likely leaving some degree drops on the table. It annoys me when people choose quality blocks and have tons of radiator surface area only to hobble the flow with multiple 90 degree fittings, complicated runs, and even parallel flow, and then just put a single pump in. It would be no surprise at all if the nice D5s people use are working as hard as they can and getting <1GPM because of a poorly executed loop. 1GPM does roughly get most hardware to diminishing performance returns. Even .5 can be adequate.

People like to say dual pumps is stupid except for redundancy, but with automatic throttling and shutdowns on the chips if things get out of hand having redundant pumps is not really necessary. IMO it's more about maximizing performance. Why spend $500 in blocks and rads and fancy fittings if you're not willing to spend another $100-150 on a pump and top to get that max flow?

VSG has probably one of the best review test benches for watercooling. He calibrates his block performance to 1.2GPM to measure the temperature performance. He doesn't show it on his TPU reviews, but he used to show flow rate restriction curves, too.
http://thermalbench.com/2016/12/02/alphacool-eisblock-xpx-cpu-waterblock/4/