WC - Twin pumps: best positions?

[loic2003]

Hi all,
I've got a Reserator 2 which I've found to be a pretty good system. However, I have CPU, GPU and chipset watercooling blocks, so there's quite a lot of resistance. Every so often (curiously, when the ambient temperature in the room is cold) the pump won't really start when I fire up the PC. I'll have to shake the rad to get movement, and as the coolant heats up, the flow rate improves significantly. I'm pretty sure the system has been bled properly.

I'm also thinking of adding a second internal radiator, so resistance will be even higher.

Because of this, I'm thinking of adding a pump into my case, but I'm not 100% sure where about in the system I need to place it. Currently, the flow is as such:

Reservoir -> Pump (inside Reserator) -> CPU -> Northbridge -> GPU -> reservoir

Do I:

a) Place it such that the output of the pump on the Reserator goes straight into the input of the new pump:
Reservoir -> Pump (inside Reserator) -> New pump -> CPU -> Northbridge -> GPU -> reservoir

b) Place the new pump in the middle of my current system, IE after the CPU or GPU sections, so each pump has to push water through some restrictive pieces:
Reservoir -> Pump (inside Reserator) -> CPU -> Northbridge -> New Pump -> GPU -> reservoir

c) Place the new pump at the end of the system so the new pump 'pulls' water through:
Reservoir -> Pump (inside Reserator) -> CPU -> Northbridge -> GPU -> New Pump -> reservoir

I'm thinking a) is pretty much certain, but I want to be sure.

Thanks in advance.

 

[WoodButcher]

I'll bet he resistance is corroded aluminum, toss the reserator.

 

[Billb2]

Flow wise, it doesn't make any difference where you put the second pump. Heat wise, it should be before the rad.
...and, get a real WCing system.

 

[loic2003]

what's wrong with the reserator? Seems to work alright, despite the fairly high load on it.

the system is only a few months old and has treated coolant, so I doubt there'd be too much corrorsion.

 

[WoodButcher]

Originally posted by: loic2003
what's wrong with the reserator? Seems to work alright, despite the fairly high load on it.

the system is only a few months old and has treated coolant, so I doubt there'd be too much corrorsion.

Originally posted by: loic2003
However, I have CPU, GPU and chipset watercooling blocks, so there's quite a lot of resistance. Every so often (curiously, when the ambient temperature in the room is cold) the pump won't really start when I fire up the PC. I'll have to shake the rad to get movement, and as the coolant heats up, the flow rate improves significantly.

Hmmm

 

[aigomorla]

first off dont.

Dont tweek a zalman kit expecially. If your happy with it, keep it as it is, if your not, replace the entire thing. Dont ever try to tweek it.

That being said, in pairing pumps you must follow golden rule #1. You must match the pump.

The zalman uses a very weak pump. Ehiem if i remember correctly, and i also believe that pump is a subermissible pump, so pairing the pump would be very difficult.

Second the entire system is composed of ALU. This is why u dont tweek it.

The safest mod ive seen was one guy buying 3 reserators, and making a tripod out of them. He then had 3 pumps which passed the coolant though the system.

I thought that was kinda neat.

 

[loic2003]

- I believed the occasional poor flow of coolant was due to quite a lot of restrictive parts, and just one pump. Once the fluid had got some momentum, it works a treat. I'd still like to be sure and place a serious pump in there, plus stick an extra Radiator in there for good measure.

-You're right about the pump - I took the system apart one day when it misbehaved and the R2 has a submersible centrifugal pump. I'm curious about this golden rule of pumps - do you mean putting them in parallel rather than series with a couple of T- junctions? I can see how you'd need to have matching pumps in this scenario. I'm thinking of placing the pumps in series, however, with the weaker pump at the back. Even if the Reserator pump wasn't active, it's just a centrifugal pump, so coolant would still be drawn through it without that much resistance.

-why is Alu bad for a cooling system?

 

[WoodButcher]

In order to prevent corrosion an inhibitor must be used, the fluids are "thicker" than water and more difficult for the pump to move, less flow in this case means less cooling. The best, most readily available material for waterblocks is copper so this means mixing metals in the loop, the aluminum will corrode first, clog the blocks and pumps or worse leak. Why risk this in a pc? I don't know how old you are but back in the '70s aluminum became the metal to use in cars for the weight / cost savings. The antifreeze wasn't what it is today and there was water! everywhere..... The additives are better, granted, but aluminum has not changed.

 

[aigomorla]

pair pumps...

its so one pump doesnt do more work then the other.

If you have off pumps, and one is higher raited, theres a chance you will burn out your weaker pump from overworking it.

This is why u need to match pumps. Its hard to match a subersible pump because you have no option to serial it.

My friend pulled his off by serialing the reserators. So he doubled his cooling capacity by having 2, trippled it by having 3.

Watercooling on a closed loop works on a principle of equalibrium, so dont get this concept confused.

But anyhow you cant really pair that pump without another reserator unit, which is where i was going.

 

[Billb2]

Pumps in Parallel or Serial Connection

Pumps in series = same flow, but add their pressures.
Pumps in parallel = Same pressure, but add their flow rates

This is very a simplistic explanation, as the resistance of the loop also affects the resultant flow and pressure of the combined pumps. Just knowing the maximum pressure and maximum flow rate of the pumps isn't enough to predict the outcome of combining them. You have to look at the P-Q curves for both pumps also.

 

[loic2003]

Ok, thanks for the replies.
I'm going to give it a try putting them in serial, turning down the speed of the new pump so the flow rates match better. I'm not sure how this could burn out the current pump in the reserator considering it's a centirugal pump which would allow water flow through it even with the pump off. There'll still be resistance on the motor too with plenty water going through.

Doing a little research, it looks like it's actually the Zalman coolant additive that gathers in chunks eventually and messes with the flow indictor.

I'll let you all know my results.

Thanks for your responses again - most helpful.

 

[WoodButcher]

Thanks for the follow up, maybe you could find better additives to prevent corrosion, I don't know. Did you , or will you filter the coolant to see exactly what it is? A gel or slime would be the coolant and a solid or flakes is corrosion, either way, not good for pc cooling.