Higher Flow dual pumps
- Thread starter ai_01
- Start date
I'm about to add my second pump to my loop, but there are a few things you should take into consideration:
* 2 x identical pumps running in parallel will still only give you the same amount of flow V = F. I'm actually including the second pump for redundancy, which in turn creates a potential problem should one give up the ghost. The 50Z pumps that you and I have will allow water to travel either way including backwards (backflow) should one stop, which necessitates the inclusion of a backflow valve of some sort in between the two 'y' connectors). Water will always take the shortest route.
* If I'm reading the included graphs correctly, these are somewhat flawed for the purpose with which we need them. The difference in volume flow rate (delta V squared) is measured according to free flow. We are not using a free flow system (or we would have any fluid left in our loop rofl), but a closed one.
* Further you need to take into account the 'K' factor, which is the increased restriction or friction value placed upon your loop by the inclusion of extra connectors, for example the backflow valve and 'y' connectors.
Hope this helps
PV
* 2 x identical pumps running in parallel will still only give you the same amount of flow V = F. I'm actually including the second pump for redundancy, which in turn creates a potential problem should one give up the ghost. The 50Z pumps that you and I have will allow water to travel either way including backwards (backflow) should one stop, which necessitates the inclusion of a backflow valve of some sort in between the two 'y' connectors). Water will always take the shortest route.
* If I'm reading the included graphs correctly, these are somewhat flawed for the purpose with which we need them. The difference in volume flow rate (delta V squared) is measured according to free flow. We are not using a free flow system (or we would have any fluid left in our loop rofl), but a closed one.
* Further you need to take into account the 'K' factor, which is the increased restriction or friction value placed upon your loop by the inclusion of extra connectors, for example the backflow valve and 'y' connectors.
Hope this helps
PV
..and today's oatmeal cookie goes to PV 
If i'm reading that right, and the pumps can't cope with backflow, then i'm guessing series would be the best option? Further to that, if you were to run pumps in series, wouldn't it be best to place the pumps strategically in your loop like this:
pump>rad>cpu block>pump>gpu block
rather than
pump>pump>rad>cpu block>gpu block
So that the second pump would act almost like a signal booster?
If i'm reading that right, and the pumps can't cope with backflow, then i'm guessing series would be the best option? Further to that, if you were to run pumps in series, wouldn't it be best to place the pumps strategically in your loop like this:
pump>rad>cpu block>pump>gpu block
rather than
pump>pump>rad>cpu block>gpu block
So that the second pump would act almost like a signal booster?
ROFLname='XMS' said:
Running pumps in parallel is still more efficient than in series, but the benefits are almost negated by the restrictive nature of the included connectors required for two pumps.
Purely running the pump>pump (series) scenario is pointless. I'm actually going to set mine out very similar to what you have described below, and see how it performs (I'll post my results). Essentially, it brings the cooling efficiency of the radiator back into play. If you take the included heat dump into your loop from 2 x pumps, and the extra connectors, the cooling surface area is now critical because of the airflow required to effectively dissipate that extra heat.
Once I get my loop set up on the weekend, I'm going to run some major tests and I'll post my findings from various configurations.
PV
underclocker
New member
I would like to see those results myself as I'm giving consideration to running two pumps.
Rastalovich
New member
XMS said:
Effectively a pump for the cpu and a pump for the gpu, but within the same circuit and using the same water ?
Or if u like the 2nd pump drawing from the 1st ones work, instead of the ciruit relying on the flow from the 1st to complete the cycle.
??
I`m thinking that if u had, for example, 1 meter of tubing, 1 @ 0m and 1 @ 0.5m would be optimum ?
Rastalovich said:
Yeah basically. The second pump can help to bump up the flow rate again after the water has passed theru the first block. Only really useful if the block is restrictive like the MP5/Storm.
Rastalovich said:
Nah not really, you want the pumps to be placed where ever they are needed. All depends on the blocks used.
Rastalovich
New member
IC, so u wouldn`t necessarily be measuring the tubing in-between to make sure they`re a good distance, or certain distance apart.
I understand.
I understand.
Woops, forgot to add...
Another interesting point to make is that if you're looking for increased velocity from a dual pump setup (especially for impingement type blocks), then you don't always have to use the largest I.D tubing that you can find. Many high pressure mains or fire hydrants utilise big slow pumps; but where they achieve the high p.s.i rating is through smaller pipes or hoses (forcing a great amount of water through a narrower passage). Hence the reason that the water exits at a greater pressure.
It's actually the right hand setup for parallel config. The left is actually in series.name='NickS' said:
Another interesting point to make is that if you're looking for increased velocity from a dual pump setup (especially for impingement type blocks), then you don't always have to use the largest I.D tubing that you can find. Many high pressure mains or fire hydrants utilise big slow pumps; but where they achieve the high p.s.i rating is through smaller pipes or hoses (forcing a great amount of water through a narrower passage). Hence the reason that the water exits at a greater pressure.
No not exactly mate. My Swifty MCP600's/ AquaXtreme 50Z's are both high flow and high head rated for 700 litres/h and 3.2m (10.5 ft). Running both of these pumps in parallel or series is still only going to move 700 litres of coolant per hour, it's not doubled. But, with a parallel setup the load is now shared between the two, and the 'y' connectors act as a balancer. In all honesty, dual pumps is way overkill for the majority of wc'd rigs considering that there is no way that any fluid in my rig will be travelling vertically (according to gravity) for 10.5 ft until it reaches 'shut-off' head pressure. But again, I'm including the second pump for redundancy should one fail.Llwyd said:
Philly_boy
New member
Not knowing what CPU block we're working with doesn't help making the best recommendation. Alex @ Petra's Tech Shop did a comparison review on different tops for the DDC2/MCP355. In that review, he noted that adding a second similarly configured pump in a restrictive loop (with impingement blocks like the Storm G4) will only net you about .5 gallon per minute increase in flow, but your effective head pressure will multiply to 1.5 times the head of just one pump. Parallel pumps will net you only about .1 gpm increase in flow and no increase in head pressure. I agree that the "Y" connectors used in a parallel config would also be restrictive themselves.
