Advice needed - using 2 pumps and 2 radiators

heman

heman

heman

Associate
Joined
21 Feb 2004
Posts
1,325
Location
London
Planning my 1st w/c build. Main purposes are quietness, decent overclock and low temps (same as most I guess!).

Case is a TJ07. Other w/c components are:
360 rad, 240 rad, tube res, cpu block and gpu block (universal or full, undecided yet).

Looking to mount both rads at the bottom of the case, tube res in the middle section.

Pump wise I have x2 D5s and one DDC. I would like to use 2 pumps to maintain good flow, I think 3 may be OTT. Question is where in the loop to position the pumps? I was thinking at opposite points in the loop to maintain flow. Also what combo of pumps to use, the DDC has the higher head pressure so was thinking of one DDC and one D5. Can use a fan controller to reduce the DDC's speed slightly to keep it quiet.

The other question is would it be best to rads in parallel or serial?

Also can anyone recommend a good flow sensor/meter based on experience?

Thanks!
 
Running pumps in series boosts head pressure, it doesn't effect flow you have to run them in parallel for that. Rads it doesn't matter, whatever is easier (serial) as the loop will reach equilibrium.
 
Running rads parallel would reduce restriction, but most of the newer rads have very low restriction any way, so better use what's easiest or neatest.
Since you already have 2 x D5 pumps I would use them, but they do take some space compared to DDCs :)
I use an AQ flow meter (high flow). It is good when used with a AQ fan controller.
Or you can use o.e of the newer flow sensors which uses pressure rather than flow to measure.
 
http://martinsliquidlab.org/2011/04/26/pump-setup-series-vs-parallel/

enjoy

short version - run them in serial, and it doesn't matter where in the loop you put them, one after the other if it's neatest and easiest, you can also run

@Uber... not entirely accurate, boosting head pressure overcomes restriction which in turn does increase flow rate - it won't increase flow rate above the rating on the pump, but it will increase flow compared with running 1 pump

running pumps in parallel does nothing for flow speed (unless you mean running to separate loops, but that isn't what I would call parallel)

mounting rads in parallel might help IF you are getting high temps, the flow will be slower in the rads whilst high in the rest of the loop meaning that the water could shed more heat than it otherwise would... if it is easy to do then go ahead, but if it's neater and easier to run them in serial then go that route :)
 
Thanks for the advice guys. So to summarise run pumps in serial.

Not sure about the rads, Avathar you say running in parallel reduces restriction, but angrybird you say this would reduce flow. This seems contradictory?

Also what pumps should I use, as you said Avathat I would rather use one DDC and one D5 as 2 D5s takes a lot of room.

BTW Avathar do you have a link to the flow meter you have?
 
Last edited:
Why are you looking at running 2 pumps I understand youve got them but that doesnt mean you need to use them Ive got a single d5 running at just speed 3 and its easily keeping up with my loop

If you are going to run 2 then id run 2 of the same ie 2 ddcs or 2 d5s and not mix them
 
Last edited:
I can't speak highly enough of running dual pumps in series, especially on a big loop. I moved to dual d5's at the weekend, filling and bleeding the loop took no time, the increased head pressure seemingly made it impossible for air to get locked, a massive improvement over a single pump in my loop.

You have two d5's, I say use them :D
 
Would like to have a high flow which should help with temps. 3 is overkill IMHO but 2 may improve things.
 
andybird123 said:
@uber... not entirely accurate, boosting head pressure overcomes restriction which in turn does increase flow rate - it won't increase flow rate above the rating on the pump, but it will increase flow compared with running 1 pump
Click to expand...

Yeah sorry thats what I was trying to say.


andybird123 said:
running pumps in parallel does nothing for flow speed (unless you mean running to separate loops, but that isn't what I would call parallel)
Click to expand...

Strange, if you check the MLL article you linked it says that parallel pumps basically doubles the flow rate.
 
Restriction reduces if used in parallel, (just like an electrical circuit using resisters), but flow gets reduced inside a RAD as well(liquid has to be divided in two paths for two RADs. However overall flow of the entire loop will remain the same.

I used Aqua Computer High-Flow Flow Sensor

But you will need it to be connected to a Aquaero to get the readings.

Flow is over rated in a loop by many. Most of the sites say you need an optimal flow of 1GMP for good temps, but I found it is not that much of an issue even with 0.2GMP!
I currently have a single D5 running around 0.5GPM and I see no difference when I ran it at 0.2GPM.

I once toyed with the idea of placing 2 x D5s in the front drive bay area of my TJ07.
 
If I can be honest here, 2 pumps is overkill, I have 1x D5 running a system with 5 blocks, triple and double 120 rads and fans running at 600rpm and the temps are fine. I fail to see what actual real world benefit this would bring to the table at the expense of making your loop that much more complex, Just my opinion.
 
Only advantage I can see is redundancy(in case of a pump fail), but that can be achieved by other means like monitoring pump RPM and auto shutdown.

Oh and bragging rights ofc :)
 
Ok, radiator bit makes sense. Because their individual flow will be halved running in parallel and that they are not inherently that restrictive I will probably have them in series.

Will try 2 pumps as well. The flow meter will be useful in recording results.
Avathar flow meter looks interesting but I don't have an Aquaero. Will look at others.
 
SpeckleD said:
If I can be honest here, 2 pumps is overkill, I have 1x D5 running a system with 5 blocks, triple and double 120 rads and fans running at 600rpm and the temps are fine. I fail to see what actual real world benefit this would bring to the table at the expense of making your loop that much more complex, Just my opinion.
Click to expand...

One pump could run it, but would your temps be better if you had two pumps (unless you have tried this?).

As mentioned one of my aims is decent temps, I have 2 pumps so will probably try out with one and pump and add a second. Wanted to know what other peoples experiences of 2 pumps were.
 
My experience here is the ability of your rads to disipate the heat generated is the factor here, the temps even out throughout the loop no matter how fast the flow rate (within reason) you may gain a degree or two but not huge ammounts, hope this helps.
 
Hi speckle, which pumps did you use and what were you components please?
Want to get an idea for comparison.
 
ubersonic said:
Strange, if you check the MLL article you linked it says that parallel pumps basically doubles the flow rate.
Click to expand...

kind of yes, but only on a very low restriction loop, as you increase restriction the pumps will still stop working at the same point as a single pump effectively, where as series pumps overcome restriction to deliver a higher effective flow

dual pumps are for high restriction loops, if you have a low restriction loop then you don't need dual pumps anyway,
if you have lots of blocks, 90's and rads you are looking at say 6psi of drop, the parallel pumps will work no better than a single pump, looking at where the curves cross over

also, in terms of redundancy, a single pump on a parallel loop loses 83% of it's pumping power, where as a serial loop loses a tiny amount, the single pump still works as effectively as if the other pump was never there, giving you some semblance of redundancy that you don't get access to with a parallel setup

parallel basically offers no advantage over a single pump, where as serial gives you good benefits in pumping power to overcome block restrictions, and redundancy if something fails
 
I am currently running an i7 930 at 4.2 through a swiftech Apogee XT waterblock, a rampage III extreme M\B with a EK waterblock, Dominator RAM with an EK waterblock, 2x HD5870 GPU's with Koolance VID-AR587 waterblocks. These are cooled by a Thermochill TA120.3 and a TA120.2. I did have 2x D5 vario pumps a while back but as stated the difference was minimal for the extra complexity. I now run with just one of the pumps running on setting 4.
 
heman said:
One pump could run it, but would your temps be better if you had two pumps (unless you have tried this?).

As mentioned one of my aims is decent temps, I have 2 pumps so will probably try out with one and pump and add a second. Wanted to know what other peoples experiences of 2 pumps were.
Click to expand...

I run 2 pumps, EK DCP 4.0's (4m head, 800lph)
I have 3 GPU blocks, 1 CPU and a phobya 1260, all in series

I've improved temps by about 5C by going to dual pump (for an extra cost of £40)
 
You must log in or register to reply here.
Back
Top Bottom