Pumps in series
 

I've seen alot of talk about running pumps in series lately. HAs anyone done any testing on this to see what the actual increases are? Wouldn't parallel pumps be better?
Won't the 2 pumps being not of 100% equal specs cause some excessive parts wear? How about pressure on both sides of the impeller when in series or parallel?
Seems to me a "holding tank" or "equalization tank" with both pumps pumping into it with one exit line to eqaulize the pressure coming out without causing undue stress on either pump would be the ideal setup?
Any thoughts on this?

I have been thinking about this myself as i dont want everything messed up if a pump fails.

so far the only ideas i can come up with to completly remove undue stress on the pumps is to have them running in seperate loops through the same waterblock (a concept i am working on...) or to have them set up so as they are electronically managed to produce exactly the same output as each other.

the problem with storage tanks is that they pumps will still be fighting with each other to puch water into the tank.

i'm gonna phone my uncle, he makes multi pump systems for central heating.

I think the tank idea will work fine actually. Any "fighting" would be be one directional in that it may be fighting more "head" essentialy with the resulting pressure coming out of the tank at a sum of both but slightly less due to any extra resistnce from both pumps pressurizing the tank at the same time.
In series or parallel pumps "fight" each other and cause stress on both sides of the impeller. That's what will cause undue wear on the impellers.
I'm actually most curious about the benefits of 2 pumps really. Unless you're using weak pumps to begin with, holding what Cathar has recomended as pump overkill before to be true, then the advantage of 2 pumps would only be beneficial to those with weaker pumps. In which case 2 pumps could add up to the cost of a much better pump pretty easily. Plus you have the heat issue which is already been discussed and would only serve to be doubled by adding a second pump. Seems that may offset any benefit from flow rate increase.
I think redundancy would be the biggest attribute to running 2 pumps. In that case a series condition would be less benficial due to the fact that if one pump fails then the reamiining pump would be stuck driving the impeller of the dead pump. Parallel mode would seem better for redundancy sicne if one pump fails then it's essentially left with running the loop itself with no real ill effects from the dead pump.
For the cost I haven't quite understood why anyone would consider two pumps a benefit. Unless that person happens to have 2 pumps laying arond that are the same and of lower "power".
I'm curisous since I'm not sure I may not be missing something with the 2 pumps. Hopefully someone will answer that question.

I fed my Eheim 1250 outlet into the inlet of my Iwaki MD30-RZ. Near as I could tell it added close to exactly the pumping pressure that the Eheim 1250 produces at the resultant flow rate that it was pumping with.

It doesn't matter if the pumps are mismatched so long as the resultant flow rate through the system is not higher than the maximum flow rate that the weaker pumps can push in open flow.

The Eheim 1250 can push 20LPM maximum, but realistically around 15LPM after accounting for the tubing and fittings on the 1250. The flow rate through my system with it in series with the MD30-RZ was around 12LPM in the test I did (WW + raidator), which is less than what the 1250 can push in wide-open mode, so it was safe to add it in series. Without the 1250 in series the MD30-RZ was pushing just over 11LPM by itself through the same loop. However, since even 11LPM is more than the Eheim 1048 can push in wide-open mode (10LPM max) then it would've been a bad idea to stick the Eheim 1048 in series with the Iwaki MD30-RZ.

Sticking pumps in series can work well so long as you don't grossly mismatch.

I have 2 L20 in series, performing better then one L30. Price was just about the same. Combined wattage also. Noise level is quite a lot lower.
And the big bennefit is that if one of the pumps fail, you are still up and running :p

I just orderd 2 MCP600 to replace the L20. I hope it will give me a bit better results :D
Currently My tems are Idele 30c and load 37c with the WW (temps from the IC7-Max3, and that board is known to give strange temp readings)

Serial is good if you have pumps, such as many of the aquarium pumps, which have a relatively low head capability compared to their free flow rating. In this case, the head will be roughly doubled for the same flow rate.

With parallel, you cannot really use non identical pumps, unless you use a complicated set of valves, preventing the stronger pump from pumping water the wrong way through the weaker pump.

Redundancy is a benefit, obviously, but for the most part, it is when used in serial to provide the kind of pressure often associated with more expensive industrial pumps, rather than aquarium pumps.

8-ball

I understand what you're saying 8-ball. I still think the "tank" idea to be the best of both. Even with uneven pumps, which even identical pumps are not going to be equal, wouldn't both pumps pumping into a mutial tank with one combined outlet give you both added pressure and redundancy? That is with both pumps having seperate supply lines to the reservoir. Maybe I'm wrong on the pressure being combined on the tank outlet. Don't see problem with that setup other than if one pump fails but then all you'd have to thave is a one way shutoff valve on each pump.
Obviously series is the easiest setup, I'm just considering doing a similar thing with multiple loops and a distribution manifold. If the tank idea is not totally off base then I could run multiple lines out to the 2 loops.

Regarding "pushing through a stopped pump". This is really only a bit of an issue with the closed-impeller (high-head) pumps. With regular open-impeller pumps, it's actually fairly easy to push water through them when they are stopped.

Basically it's a case of if you have two cheap-ass open-impeller style pumps, then sticking them in series is a good idea as even if one breaks down, it won't offer much resistance at all to the one remaining working pump.

Thanks Cathar :)
Answered my questions I needed answered, as always.

So Cathar, you mean, i would be able to put a swiftech mcp600rev2(700L/h stand specs and 3,2m head with FOUR DOUBLE heatercores and 4 m of tubing) in series with 2 IWAKI MD-15's(50hz-version,less powerfull, so 2,4 m head and about 960l/hour each)? without wearing the pumps out?Have recently bought 2 iwaki md-15-R's for about 70 euros(=$92) and want to put them in the loop.

bump

cathar?

Yes, you could do that.

Not sure why you would want to do that though. The MCP600 will probably be running somewhat towards the end of its PQ curve and not be contributing much at all. It'd make sense if the Iwaki's in series were pushing less than, say, 7LPM through your system, but if the Iwaki's by themselves are pushing above that, I'd be tempted to give the MCP600 a miss as it'd add so little to the flow rates to make any appreciable difference to the waterblock's performance, but will be dumping extra heat into the loop (not that pump heat is much of an issue with 4 double heater-cores).

Just a single Iwaki MD30-RZ would serve you better in all scenarios, and dump less heat into the water as well.

Hi,

I don't think the iwakis will be pushing more than 7 l/m through my system.

The setup will be

2 external boxes under my desk.The boces will be on top of each other.

will be using a cascade wb
each external box will contain:

-iwaki md15-R (mind you, on 50 Hz, so a lot less powerfull, about 2,4 m head and 960 l/h
-2x xlr120 jpiperformance rads running in series

the computer case itself will be about 70 cm above the lowest box and about 35 cm above the upper box

the setup will be in this order
lowest box-higher box-computer case(cascade)-lower box, etc.
all will be done with 1/2 inch ID tygon tubing.

Just couldn't resist to buy the two iwakis for 70 euro's(for both of them!)! If i buy them new in the Netherland, i would have to pay 210 euros EACH!

My watercooling sytems seems to be little bit over the top, and it is.Will be using 1 external box per system in the future, just waiting for the new system to arrive and i wanted to try this until the second system comes in.

My external boxes look like this:

http://www.teamster469.com/images/quadra2.jpg

http://www.theforumisdown.com/upload...5-0540_IMG.JPG

Cathar, waht about the other way around? 1xiwaki md-15 and 2x swiftech mcp 600 in series?