Single loop two res/pump combo's

Martini1991 · 7 May 2011 at 15:21

Hey there. I'm fairly new to Watercooling, so bear with me.

Usually I run two independant loops. A CPU loop consisting of my EK Supreme HF, my 120.1 Thermochill RAD and then my XSPC 750lph res/pump combo.
A GPU loop consisting of my two EK 6870 full blocks, 120.2 thermochill RAD and then another XSPC 750 lph res/pump combo.

However, recently I went from an AMD CH4 + 1055T to a 2500k MSI GD65 and again used my two seperate loop, 2 weeks in my board died and I've jumped back to AMD using a CH4 and 1055T again.

Now it's time to put the build back together, however it'd work so much better just use it as a single loop going ;

Res/Pump1 -> CPU -> 120.1 -> Res/Pump2 -> GPU1 -> GPU2 -> 120.2.

Would that work?

Here's the res pump combination ; http://www.overclockers.co.uk/showproduct.php?prodid=WC-011-XS&groupid=962&catid=1845&subcat=
Mine are the rev 2 ones.

Secondly, the way I'm going to do GPU1 -> GPU2 is ;

.In...Out.
.BLOCK1.
.ln...Out.
.BLOCK2.

Block 1 would feed into block 2, and then the water coming out of block two would go out into block 1.
Block 2 would have two of the "stoppers" on, would that work fine? It keeps it simple tubing wise.

Cheers.

Jokester · 7 May 2011 at 15:30

Personally, I would dump the second pump.

Not sure what you mean by your GPUs, do you mean you'll have them in parallel? Doesn't make any real difference, again personally, I would just put them in series.

Martini1991 · 7 May 2011 at 15:34

Jokester said:
Personally, I would dump the second pump.

Not sure what you mean by your GPUs, do you mean you'll have them in parallel? Doesn't make any real difference, again personally, I would just put them in series.

As would I, but I don't think that one of those XSPC res/pump combo 750 lph is good enough to do a loop consisting of 3 blocks and 2 rad's. Flow rate would be fine, but it's the delivery head that's lacking.

And yes I mean parallel for the GPU's. It's because I'm going to have my 120.2 bottom mounted in my HAF 932 (I can't rear mount my 120.1 and top mount my 120.2 as they're too big to fit together)

So with it going parallel it makes it a lot easier.

Jokester · 7 May 2011 at 15:38

What type of pump is it? Is it a Laing DDC type pump or something else?

Running them in parallel should be ok, only concern is whether one block will get preferential flow, but since the blocks are identical it shouldn't be a noticeable problem.

Martini1991 · 7 May 2011 at 15:41

Here's a picture of how I mean with the GPU's they feed in/out of each other (Not my pic, just a google)

And it's the built in one as per link.
I had a quick look around, but I haven't a clue what pump is inside it.
EDIT : For reference, the delivery head is 1.8m.

mike gunnz · 7 May 2011 at 17:17

Pump in pic might be a MCP655, can't say for sure.

Your hardware doesn't require separate loops. A single loop config with a quality fat 360 rad and an 18w DDC or a D5 will suffice surely?

Personally I'd put your existing rads and pump/res combo on ebay and get the above.

Parallel or series will make little difference, but as said put them in series. I can't see how a series config makes tubing more complex.

Jokester · 7 May 2011 at 17:24

He won't be able to tube it up in series if that picture is of a similar set up as there's no space.

I think you're pump would actually be ok, might be worth getting the loop calc spreadsheet that Skinee does I think?

Martini1991 · 7 May 2011 at 18:42

mike gunnz said:
Pump in pic might be a MCP655, can't say for sure.

Your hardware doesn't require separate loops. A single loop config with a quality fat 360 rad and an 18w DDC or a D5 will suffice surely?

Personally I'd put your existing rads and pump/res combo on ebay and get the above.

Parallel or series will make little difference, but as said put them in series. I can't see how a series config makes tubing more complex.

I don't really want to mess around with changing my loops lol.
I have one fat 120.1 and one fat 120.2, it works for me

Jokester said:
He won't be able to tube it up in series if that picture is of a similar set up as there's no space.

I think you're pump would actually be ok, might be worth getting the loop calc spreadsheet that Skinee does I think?

Yeah, that's kind of my reason for the parallel, it's slightly like the picture.
The head delivery is only 1.8m on these pumps.

Martini1991 · 8 May 2011 at 08:37

Here's a rough picture of my GPU's in parallel, this will work right? ;

mathwat · 8 May 2011 at 09:04

I think that will be fine with one pump tbh..

realscot · 8 May 2011 at 12:09

forgive my WC inexperience but in the pic above isn't the water simply going to go straight through the block and carry on to the second gpu as it'd be the path of least resistance and vice versa when it heads back to gpu 1. The water has to go all the way through GPU 2 as it has no choice, but GPU 1 I feel will get very little water going through the block itself.
I could be totally wrong as the block may simply not work like that.

Jokester · 8 May 2011 at 12:15

The blocks are of equal restriction, the only difference is the additional tubing to the 2nd GPU so water will flow fairly equally through both blocks. It's essentially like putting 2 resistors in parallel in an electrical circuit, the current splits in two and goes through both resistors.

An interesting side effect of doing them in parallel is that the overall pressure drop for your GPUs will be about half of a single block and about a 1/4 if you had them in series.

Martini1991 · 8 May 2011 at 12:39

realscot said:
forgive my WC inexperience but in the pic above isn't the water simply going to go straight through the block and carry on to the second gpu as it'd be the path of least resistance and vice versa when it heads back to gpu 1. The water has to go all the way through GPU 2 as it has no choice, but GPU 1 I feel will get very little water going through the block itself.
I could be totally wrong as the block may simply not work like that.

I'm running them in parallel due to my 120.2 RAD being bottom mounted.

Jokester said:
The blocks are of equal restriction, the only difference is the additional tubing to the 2nd GPU so water will flow fairly equally through both blocks. It's essentially like putting 2 resistors in parallel in an electrical circuit, the current splits in two and goes through both resistors.

An interesting side effect of doing them in parallel is that the overall pressure drop for your GPUs will be about half of a single block and about a 1/4 if you had them in series.

Flow rate should be ok even in parallel?
So, I'm assuming the way I've got it set out is perfectly fine then?

Jokester · 8 May 2011 at 12:43

Martini1991 said:
Flow rate should be ok even in parallel?
So, I'm assuming the way I've got it set out is perfectly fine then?

The flow rate will be half what it is elsewhere in the loop, but I don't think it'll be so low that it'll have a significant impact on your cooling performance.

Jokester · 8 May 2011 at 12:46

This was what I was thinking about the other day:-

http://martinsliquidlab.org/pump-and-radiator-optimizer-spreadsheet/

MikeTimbers · 8 May 2011 at 15:15

That parallel pic looks really wrong. I've never tried it but that looks as though the first card block has the inlet and the outlet to the second block directly in line. There is nothing to cause half the water to split and go inside the first block. It appears as if the water will almost all flow through to the second one and then out. I suppose it depends on the internals at the top of the block but all the gpu blocks I've seen would not work as per your pic.

Jokester · 8 May 2011 at 15:19

It'll work fine, the differential pressure across the inlet and outlet of the first block will mean about half the flow will go through that block.

Martini1991 · 8 May 2011 at 16:13

MikeTimbers said:
That parallel pic looks really wrong. I've never tried it but that looks as though the first card block has the inlet and the outlet to the second block directly in line. There is nothing to cause half the water to split and go inside the first block. It appears as if the water will almost all flow through to the second one and then out. I suppose it depends on the internals at the top of the block but all the gpu blocks I've seen would not work as per your pic.

That is what common sense would dictate, however in theory and in practice it works as pointed out on the post above me.

MikeTimbers · 8 May 2011 at 16:20

I seriously doubt that. Water takes the path of least resistance and behaves very like electricity. I see no reason for the water to travel around the first block and suspect that the flow will not be sufficient to cool the first one.

I wait with interest for the actual load temperatures once built.

Jokester · 8 May 2011 at 16:25

MikeTimbers said:
I seriously doubt that. Water takes the path of least resistance and behaves very like electricity. I see no reason for the water to travel around the first block and suspect that the flow will not be sufficient to cool the first one.

I wait with interest for the actual load temperatures once built.

It's because it's like electricity that'll it'll flow through both blocks. Both blocks are effectively resistors, the pressure difference across both blocks in parallel is equivalent to Voltage and the resultant flow is current. The flow will more or less split in 2.