Dual Pumps.

You don't have to, many folks just placed them in two different spots in the loop, that's best for redundancy rather than extra water flow since the flow increase is linear, you would get better pressure though.
Anything past about 1GPM is pretty negligible in cooling difference unless your loop is really restrictive and a single D5 does that easily.

Edit: Whats your goal with your loop?
 
Last edited:
I'd like to be able to have two video cards and one CPU on a single loop. Overclock both of them as far as they'll go without doing any hardware modification and be able to cool all three to 30-50 C.\

Right now I've got 3 560 radiators, will be going push. I've got plans to go to 4 or even 5 560 radiators. It's an external watercooling loop for the most part.
 
I just piped 'em together. I do have 4 radiators in this rig, but mainly did it to keep the flow up without running the pumps wide open.

194458_RVIEO-1.jpg
 
The best part about dual pumps is the redundancy... Though most D5's run for years, and years, even decades, before failing.
 
I just piped 'em together. I do have 4 radiators in this rig, but mainly did it to keep the flow up without running the pumps wide open.

Looks like one pump leads into the other? Would it not be better to have Pump > Component > Radiator > Pump > Component > Radiator > Res?

I feel like having Pump > Pump, that the second pump would resist the water flow going into it somewhat, therefore, nullifying the effect you were going for.
 
Nope, Martin's Liquid Lab did the tests - the 2nd pump can be turned off and even then it offers very little flow resistance. Pumps can be run at different speeds (even different pumps), all sorts of things and it's nothing but positives.

https://martinsliquidlab.wordpress.com/2011/04/26/pump-setup-series-vs-parallel/

The format is a little strange on that page - the 2nd and 3rd page links are just above the "Comments" tag at the bottom.
 
Nope, Martin's Liquid Lab did the tests - the 2nd pump can be turned off and even then it offers very little flow resistance. Pumps can be run at different speeds (even different pumps), all sorts of things and it's nothing but positives.

https://martinsliquidlab.wordpress.com/2011/04/26/pump-setup-series-vs-parallel/

The format is a little strange on that page - the 2nd and 3rd page links are just above the "Comments" tag at the bottom.

Good read, thanks for that, did think there would be higher than 2% resistance. So to OP you don't need a special pump and looks like you can run both without any issues.
 
Nope, Martin's Liquid Lab did the tests - the 2nd pump can be turned off and even then it offers very little flow resistance. Pumps can be run at different speeds (even different pumps), all sorts of things and it's nothing but positives.

https://martinsliquidlab.wordpress.com/2011/04/26/pump-setup-series-vs-parallel/

The format is a little strange on that page - the 2nd and 3rd page links are just above the "Comments" tag at the bottom.

Something to take from this as well, he actually RECOMMENDS using pumps of different speeds and types due to noise harmonics.
 
I tried running them at different speeds off 2 different fan headers on the mobo, but I never did get them operating just like I wanted - they never tracked with each other very well and the differing speeds were more distracting than just using a Y splitter. If I had a nice controller I still think it'd be a good idea.
 
Nope, Martin's Liquid Lab did the tests - the 2nd pump can be turned off and even then it offers very little flow resistance. Pumps can be run at different speeds (even different pumps), all sorts of things and it's nothing but positives.

https://martinsliquidlab.wordpress.com/2011/04/26/pump-setup-series-vs-parallel/

The format is a little strange on that page - the 2nd and 3rd page links are just above the "Comments" tag at the bottom.

The only issue I take with his testing is "redundancy" for two pumps in parallel. In the field, you always have a discharge check valve that prevents flow from one pump "short circuiting" backwards through a non-operating pump. He didn't have that, and found out that, yup, the flow will short circuit through the non-operating pump rather than go through the circuit.

That isn't exactly pertinent to the OP though - other than unless it's a mission critical high availability system, you don't really want pumps in parallel anyway.

Only reason I would run pumps in series - one pump alone isn't getting you the flow rate you need. You need an awful lot of flow restriction for one D5 (or similar) to not provide what you need in a PC. They are overkill pumps as it is for most PC setups. And even then, rather than running two in series, I would just get a single larger pump...
 
The best part about dual pumps is the redundancy... Though most D5's run for years, and years, even decades, before failing.
I got two D5 varios second hand in 2005. One of them died around 2014, the other is still going.

So I will admit that the value we place on redundancy might be slightly overblown.

However, when that first pump eventually died, it was really, really nice to be able to continue using the PC and replace the pump when it was convenient. It's also nice to be able to use speed setting 2-3 on the varios rather than 4-5; a bit less noise, and a positive effect on pump life also.

When I first set them up they were simply piped into each other. Then, once Martin came along, I decided to dremel out the ports for a bit smoother flow. Then, I got super lucky on [H] when a hail-Mary PM led to me buying a nice dual D5 top. In all cases, these upgrades, while nice, were mostly psychological - although the dual top I have now has the pumps facing the same direction for fast and easy speed knob access.
 
Looks like one pump leads into the other? Would it not be better to have Pump > Component > Radiator > Pump > Component > Radiator > Res?

I feel like having Pump > Pump, that the second pump would resist the water flow going into it somewhat, therefore, nullifying the effect you were going for.

It doesn't happen because the restriction of the loop is far greater than the pump.

In any case, two pumps in series adds head pressure but does not increase maximum flow, i.e. the maximum theoretical flow is that of a single pump with no restriction. Pumps in parallel double the maximum theoretical flow but do not increase maximum head pressure.
 
I just piped 'em together. I do have 4 radiators in this rig, but mainly did it to keep the flow up without running the pumps wide open.

View attachment 146288

Dude... I'm just digging around to see if I need to start thinking about running dual D5's.

Can I rent you to come hardline my 900d? haha. Also, where you got the drives hiding at? I'm still using the cage with the hot-swap board. Would love to delete it and get one of those snazzy filler plates.
 
Dude... I'm just digging around to see if I need to start thinking about running dual D5's.

Can I rent you to come hardline my 900d? haha. Also, where you got the drives hiding at? I'm still using the cage with the hot-swap board. Would love to delete it and get one of those snazzy filler plates.

All the drives in that rig are M.2 There are two under heatsinks on the board and now two more in a little card that plugs into that slot that looks like another DIMM slot just to the right of the DIMM slots so 4 970 Pro M.2 drives altogether. That pic was taken before it reached it's final configuration, now it has the M.2 add on board and a 2080Ti with a Heatkiller block on it (was a Titan X Pascal w/EK block in the pic). The midplate is from ColdZero in Portugal https://www.coldzero.eu/corsair-900d-midplate-long That thing is even hard piped under the midplate and has a 480 and 240 in the basement, all P/P except I had to leave one fan off the inner side of the 240 so I could get the plumbing to fit.
 
Man, I wish we could get a hold of this guy (Martin). Or is there an expert here with a site that can help with air cooling?
 
Back
Top