View Full Version : Running dissimilar pumps in parallel
karbonkid
03-02-2008, 11:33 AM
Is it OK? ie. will the less powerful pump be forced to work beyond its specifications, or not?
togaman5000
03-02-2008, 11:42 AM
Not a good idea; the lesser pump, if I've got this right, will end up contributing nothing to the flow
STEvil
03-02-2008, 02:07 PM
Parallel on the same loop is useless in PC cooling.
karbonkid
03-02-2008, 03:46 PM
Parallel on the same loop is useless in PC cooling.
I know it won't do anything for the pressure, but wouldn't it increase the flowrate? I think of it like this - if you have two pumps drawing water from a reservoir at the same rate as each other, the water will be withdrawn at double the speed as if you only had one pump. Therefore, flowrate through the system would be doubled as long as the outlets of the 2 pumps are merged back together. Is my thinking flawed?
What I am now suggesting is, what if the two pumps weren't the same? What would happen then?
starlon
03-02-2008, 03:51 PM
What I am now suggesting is, what if the two pumps weren't the same? What would happen then?
Then the flow rate won't be exactly doubled over the more powerful pump.
karbonkid
03-02-2008, 04:15 PM
Yes, but would each pump still work to its maximum capacity (ie. you can just add the flowrates of the two pumps to get an approximate value?) Or would the asymmetry cause problems with pressure, etc?
Fairydust
03-02-2008, 04:27 PM
I know it won't do anything for the pressure, but wouldn't it increase the flowrate? I think of it like this - if you have two pumps drawing water from a reservoir at the same rate as each other, the water will be withdrawn at double the speed as if you only had one pump. Therefore, flowrate through the system would be doubled as long as the outlets of the 2 pumps are merged back together. Is my thinking flawed?
What I am now suggesting is, what if the two pumps weren't the same? What would happen then?
Your thinking is is right in a system without resistance, but in a WC loop the two pumps will need to overcome the flow resistance of the loop at the given flow rate.
Let's compare it to a conveyor belt, in a parallel setup with 2 belts side by side and another belt that has two motors on the same belt to represent a serial setup. Now the parallel belts can transport twice as much volume but the serial setup is able to convey heavier goods. (hope this helps more than it confuses)
With dissimilar pumps the flow resistance of the loop at the given flow rate could be higher than the resistance of the weaker pump, thus enabling the stronger pump to overpower it and effectively pumping through it.
starlon
03-02-2008, 04:28 PM
I don't see why it would impede one pump's performance. It's just like two rivers joining; they both flow at their own rates.
STEvil
03-02-2008, 04:31 PM
I know it won't do anything for the pressure, but wouldn't it increase the flowrate? I think of it like this - if you have two pumps drawing water from a reservoir at the same rate as each other, the water will be withdrawn at double the speed as if you only had one pump. Therefore, flowrate through the system would be doubled as long as the outlets of the 2 pumps are merged back together. Is my thinking flawed?
Yes, your thinking is flawed. You are assuming a free flow system with no restriction which is never the case with PC water cooling.
What I am now suggesting is, what if the two pumps weren't the same? What would happen then?
The same as if they were the same. Nothing.
To apply it to a flowrate/head graph, you'd still be adding the two curves horizontally, I'd think.
Xilikon
03-02-2008, 04:41 PM
Also, don't forget water take the least path of resistance so if one pump is less restrictive than the other, water will pass more thru this one causing the other to just spin in the void.
karbonkid
03-02-2008, 04:59 PM
You are assuming a free flow system with no restriction which is never the case with PC water cooling.
Since the pumps are recieving water from one source, that is then split between them, and then spliced back together immediately after the outlets, the system would see the pumps as just one pump, with a doubled flowrate, would it not? Introducing restriction, yes, flowrate would not be doubled (flowrate being a factor of restriction itself), but nevertheless, increased over just using one pump.
I'm just speculating here, so please excuse my ignorance and correct me if I'm wrong, but my speculation seems to agree with Cathar's findings/knowledge... (http://www.xtremesystems.org/forums/showpost.php?p=2288261&postcount=30)
When we put two pumps in parallel, we double the peak unrestricted flow rate. If our loop is so unrestrictive that with just one pump that the main limiting factor is the single pump's peak flow rate, then yes, it'll be better to put two pumps in parallel.
Now, of course, say I have two dissimilar pumps. I obviously can't connect them in series, as this would result one/both of them being forced to run faster/slower that it is designed for and could cause all sorts of problems, no? So parallel is the only option - but are the same sorts of problems still likely to arise?
karbonkid
03-02-2008, 05:06 PM
Also, don't forget water take the least path of resistance so if one pump is less restrictive than the other, water will pass more thru this one causing the other to just spin in the void.
This pretty much gets to the crux of what I was asking, although I still don't quite understand it. A pump pumps the water, so what do you mean by 'restriction'? If anything, a pump has negative restriction, as it's moving water, not restricting it, so, more water would flow through the more powerful pump, less through the less powerful one - exactly what we want.
Again, please excuse my ignorance. I'm new to all this :).
To apply it to a flowrate/head graph, you'd still be adding the two curves horizontally, I'd think.
In other words, it would make no difference to head, and the flowrates would be added? EDIT: Not quite... http://www.xtremesystems.org/forums/showpost.php?p=2288930&postcount=41
Yes, MpG, you are right the - two curves seem to be added horizontally
What I can read from these graphs is, in practical terms;
Below one pump's max. flow rate, for any given flow rate, the heads of the 2 pumps are added
Above one pump's max. flow rate, the flow rates of the pumps are added???
bundyrum&coke
03-02-2008, 05:36 PM
Running unmatched pumps in parallel will work. The end result will be common zero pressure at the reservoir and a common max pressure at the outlet manifold.
STEvil
03-03-2008, 05:02 PM
You are assuming a free flow system.
I'm just going to stop replying until you realize that no system is ever in a free-flow condition and that if it were unmatched pumps in parallel wouldnt matter at all as neither would have any restrictions to deal with.
Dissimilar pumps in series does not harm either pump as long as the loop max potential flow rate is below or equal to that of the weakest pump (and you do not use the weak pump to feed the more powerful one).
Boogerlad
03-03-2008, 05:20 PM
if you really want to see your theory work, then why don't you just try it?