Real world pump performance

[morphling1]

Hi, I was pretty bored today so I went and done some water flow test with different pump setups. You may or may not know that pump specs. is very misleading data, and that every water cooling setup is more or less restricting, and all the restrictions(hoses, blocks, bends, radiator,valves ...) are causing preasure drops,and all that preasure drops summs up at the end.Every pump has characteristic of curve in x,y diagram, where x is water flow in (l/h) and y is preasure (bar) or water column in (m) .
This is the only diagram I found on internet: It shows three different pums, and the curve should meet x and y axis



You can see that when we are at the top of the curve the flow drops to 0, and the pump basicaly just maintain that preasure, and that preasure is the one from all the preasure drops in your setup.
So for very restrictant sytem like water cooling setup in computer you need as much preasure as possible to maintain water flow as high as possible. You can do that in two ways, take bigger pump with bigger flow with higher water column, or you can take two smaller pums and run them in serial, which means that you just double the water column while water flows remain the same. If you run them in parallel, water column is the same while water flows doubles, but we don't want that, we need bigger preasure. For instance my Ehem model 1250 pumps 1200l/h at 0m and 0 l/h at 2m while pumping out 28W, a lot better solution would be using two smaller pumps which would probably pump less heat in the water combined.
Ok, so here is the test. I used two small Renna Flow pums 580 l/h at 0m and 0 l/h at 1.1m (5W) and Eheim 1250 (28W).
First I tested with one Renna flow using only my water block(avatar) and 0.7 m silicon hoses 12mm ID , real water flow numbers were 276 l/h, than I add my heater core and another 0.5m of hoses which is the setup I use in my computer, water flow dropped to 257 l/h , here I was surprised how my heater core didn't cause a lot of drop, Next I used Eheim,block and rad. water flow was 370 l/h , pretty dissapointing considering the size and power. Then I hooked two Rena Flows in serial with block and rad. and the flow was 325 l/h that good and at 10W combined still a lot less than Eheim. And for final test I used Eheim + Rena in serial + block+rad. the water flow was 461 l/h wow huge increase just by adding one small 5W pump.
Conlusion, using two or even more pump in serial connection is realy helpful to increase water flow in restrictant system like computer water cooling, I need to test two Maxy-Jet 1100 l/h at 0m and 0 l/h at 1.48m (14W each) I bet it would increase my water flow a lot + two of this pumps is still a lot cheaper than on Eheim 1250 and one more thing if on pump should somehow fail, you still have one for backup.

[Les]

I endorse that this approach may be beneficial in high pressure drop systems where the low flow rate may be,for whatever reason, a concern.It is an extra weapon in the armoury
As part of a broader "look-see" I have even used dissimilar flow-rate(1200l/h and 600l/h) pumps in an attempt to address lamellar flow worries . A definite,but unquantifiable,increase in flow rate over the stand-alone 1200l/h pump was obtained.
http://www.coolhardware.co.uk/module...iewtopic&t=121

[Brad]

morphling, do you think you could try all 3 pumps, just to give us an indication of how well that would go?

[morphling1]

Yeah I could, I'll post results later.

[morphling1]

Ok, here are the result. Eheim+Rena+Rena+block+rad = 490 l/h
And another test just with Eheim and the block, water flow = 453 l/h to compare with Renas 276 l/h.

[bigben2k]

That seems consistent. The Eheim is obviously more efficient.

Each pump will have a different curve. Some pumps, often referred to as "high capacity" will not drop the flow rate as much, as the pressure increases. (Eheim)

All: to convert L/h to GPH, just divide the figure by 3.76 . (There are 3.76 Liters in a U.S. Gallon)

490 L/h = 130 GPH
453 L/h = 120 GPH
276 L/h = 73 GPH

370 L/h = 98 GPH
(for Eheim 1250, block and rad)
That's not bad. It's on the low side though, I would have expected 130 GPH, but if adding the two Renna's do that, then you're all set.

Also, keep in mind that the higher the flow rate, the more restriction there will be, and this is an exponential curve. As a general guideline, keep the pressure under 10 psi, where 5 to 8 is about ideal. Anything higher is not very efficient.

[morphling1]

bigben2k, remember when we discused about that nozzle in my block, seems that you we're right, and it is a lot more flow restrictant, but I already made few changes (wider, deeper chanells ,flow path where two chanells combine was also optimize a lot, plus lose the nozzle), so for my final block it will have much bigger flow on my Eheim.

as for Eheim being more efficient, I don't think so, Renna are 1/5 the size of Eheim plus it runs only on 5W and max flow is 550 l/h. If you look at all that, Renna is much more efficient.
I'll do try and get two Maxy-Jet 1200 MP pums, with combine power of 28W like Eheim and than I'll see some huge flow rates.

[bigben2k]

The Renna's max flow rate@0 is 580 L/h
The Eheim (1250) max flow is 1200 L/h

Under the same load conditions (wb and rad), you measured:
Renna: 276 (52% drop)
Eheim: 370 (69% drop)

So you're right (and I sit corrected!) the Eheim is not all that efficient. I keep lurking the Eheim site for pump charts, but they're not up so far.

I am certainly looking forward to seeing your results for the Maxi-jets. I'd like to see the same with Rios.

One thing to keep in mind: we'll have to watch out for leaks (especially with Rios) because these pumps are not designed as booster pumps: the housing may not have been built to handle the extra pressure.

[Brad]

morph, one more request, do you think you could try changing the positoning of the pumps in the loop? I'd be interested to see if that changes the results at all

[#Rotor]

thit is what I'm doing, I'm cooling 3 pc's all with one Radiator, and am using multiple pumps, as pressure boosters along the flowpath... having the pumps all bunched up will render the weak pump in the group, only a restriction to the more powerful pumps, but moving the weaker one away, to a point where it can actually add to the pressure in the loop, will be more advantageous, theoretically, of coarse

[webmedic]

Hi #rotor I was watching the other thread over at [h] that you posted about flow also.

[morphling1]

Brad, I would test it the other way, but I have already put everything back in my comp. When I'll get two maxy-jet, I'll test again.
I would say that you would probably be better the way I tested (Stronger then weaker pump), but I could be wrong, it could be exactly the same

[redleader]

So you're right (and I sit corrected!) the Eheim is not all that efficient. I keep lurking the Eheim site for pump charts, but they're not up so far.

I don't think you can draw that conclusion because:

1) Resistance is related to flow (higher flow = higher resistance due to diminished returns)

2) You varied flow between the two models.

Therefore you can't comment on pump effciency because we are unable to measure resistance relative two each flowrate (or a common flow rate).

[bigben2k]

Sure I can! The testing conditions were the same, that's the basis.

The pressure drop causes the pump to fall below its 0 rating. If the pump is of a high capacity design, then the flow rate won't drop too far from the 0 rating. If it is not a high capacity pump, then the flow rate will drop dramatically, as it did here.

Neither pump can be qualified as "high capacity", but the Renna does respond better to flow resistance than the Eheim.

More testing would provide a graph of flow rate at different pressure drops, and these two pumps would meet somewhere, in a slightly more restrictive rig, above which, the Renna would achieve a better flow rate than the Eheim. That part is very clear. Do you see it?