Upgrading water cooling - some questions

tophee · 5 Aug 2008 at 18:03

Hi guys,

I'm about to rebuild my system and want to upgrade my water cooling solution. At the moment I'm using a Zalman Reseterator v2 (the black one) which I originally bought to cool my media centre pc. For reasons I can't be bothered to go into the mdeia centre is currently cooled by air and is way too noisy, so the Zalman system is going back into that.

So, my first question: Which is better parallel or series?

If I were to put it into series it would be:
>Res>Pump>CPU block>GPU block>Rad>Res
In parallel it would be:
>Res>Pump>Tbar>CPU block>Tbar>Rad>Res
...................>Tbar>GPU block>

Is there any benefit to this second setup? Pitfalls?

My current rig (in bits atm):
Q6600, 4 Gb Corsair XMS2 ram, Radeon 3850 Pro, Abit XI38 QuadGT motherboard, Akasa Eclipse

My proposed watercooling:

XSPC Edge Acrylic CPU Waterblock
or
EK Supreme Clear Top

XSPC Dual 750 Bay Reservoir & Pump

XSPC RS240 120mm Radiator

and associated bits: tubing clips, barbs etc.

Which cp block is better? I saw on another thread that the edge block is a better pairing with the res/pump. But it's out of stock atm and I need to get on with getting my rig back in use. I intend to do some reasonable overclocking.

Oh, I've already got a suitable block for my graphics card.

Many thanks

Chris

robbiecfc · 5 Aug 2008 at 18:13

The best block of them all is the EK Supreme but generally in a CPU only loop. It is a bit too restrictive otherwise. You could also look at the D-Tek Fuzion V2 which is less restrictive and is very well regarded.
As to pumps, have a look at the Laing DDC3.2 series and mate with a XSPC Resevoir Top for best performance.

Ben M · 5 Aug 2008 at 18:42

series =\ don't use parallel for a water loop.

WJA96 · 5 Aug 2008 at 18:52

robbiecfc said:
The best block of them all is the EK Supreme but generally in a CPU only loop. It is a bit too restrictive otherwise.

I'm not sure that's 100% correct.

robbiecfc said:
You could also look at the D-Tek Fuzion V2 which is less restrictive and is very well regarded.

What back-pressure measurements are you using to make these statements?

I'm starting to get excessively annoyed by all the folks saying that such and such is too restrictive when they usually don't actually have any experience of what they are talking about.

I have a CPU/Chipset/Dual graphics loop which features an EK Supreme Acetal and the whole lot is driven by the Eheim 100 240V pump in a Zalman reserator. That system is specced to drive a graphics block, a chipset block as well as Zalman's own CPU block.

robbiecfc said:
As to pumps, have a look at the Laing DDC3.2 series and mate with a XSPC Reservoir Top for best performance.

Everyone seems to forget that a Zalman Reserator generates 0.4m of head just in the water in the reservoir. You won't need anything like a Laing DDC to drive that loop, especially if you're leaving the pump running in the reserator. Ideally you have the pump at the lowest point in the loop because mavity helps push the water through the pump, but if you're having a dual pump system then the dual bay 750 should be fine. I'd also have a good look at that radiator as it's a dual 120mm (240mm), described as a 120mm.

The order you choose is going to be rendered largely irrelevant by the fact that the water in the reserator will be the temperature your system will run at. The reserator actually works best when it's hot to the touch in a cold room (the bigger the differential, the more efficient it is) and in many ways the active radiator is much the same. Whatever order you put the blocks in, it'll all end up the same temperature anyway. I'd route my piping for the least possible length of tubing, and that does generate back pressure, and maybe pump from the reserator through the blocks into the Bayres then pump from that into the active radiator and back to the reserator.

The IX38 has really good PWM fan control and I would be very tempted to use good PWM fans daisy-chained off the CPU fan header on the radiator and set it to switch off the fans to the radiator while the CPU is running relatively hot (55C?) then come in 100% to cool the fluid down ASAP. My bet (from a similar system I'm running here) is that the fans won't kick in very much at all, if ever.

If you want the system I'd really recommend then get a another decent passive external radiator eg. Innovatek Konvekt-o-Matik or Cape-Cora and use that instead of the 240mm active radiator.

tophee · 5 Aug 2008 at 20:38

Hey WJA96, You make some very interesting points. I actually quite like my Reserator, it's just I need to put that back in my media centre... though it could wait. I think you may be confused slightly... only in respect that the reserator et al will be moved off system to a different system and I am looking for a replacement. I have to admit the passive cooling of the Zalman reserator it quiet and appears to do the job well. Perhaps a new pump is in order and a second Zalman for the media cenre.

WJA96 · 5 Aug 2008 at 21:08

tophee said:
Hey WJA96, You make some very interesting points. I actually quite like my Reserator, it's just I need to put that back in my media centre... though it could wait. I think you may be confused slightly... only in respect that the reserator et al will be moved off system to a different system and I am looking for a replacement. I have to admit the passive cooling of the Zalman reserator it quiet and appears to do the job well. Perhaps a new pump is in order and a second Zalman for the media cenre.

Well, at £126 incl. a very good CPU block the current V2 that OcUK have on special is quite a decent deal at the moment.

robbiecfc · 5 Aug 2008 at 23:37

WJA96 said:
I'm not sure that's 100% correct.

What back-pressure measurements are you using to make these statements?

I'm starting to get excessively annoyed by all the folks saying that such and such is too restrictive when they usually don't actually have any experience of what they are talking about.

I have a CPU/Chipset/Dual graphics loop which features an EK Supreme Acetal and the whole lot is driven by the Eheim 100 240V pump in a Zalman reserator. That system is specced to drive a graphics block, a chipset block as well as Zalman's own CPU block.

Everyone seems to forget that a Zalman Reserator generates 0.4m of head just in the water in the reservoir. You won't need anything like a Laing DDC to drive that loop, especially if you're leaving the pump running in the reserator. Ideally you have the pump at the lowest point in the loop because mavity helps push the water through the pump, but if you're having a dual pump system then the dual bay 750 should be fine. I'd also have a good look at that radiator as it's a dual 120mm (240mm), described as a 120mm.

The order you choose is going to be rendered largely irrelevant by the fact that the water in the reserator will be the temperature your system will run at. The reserator actually works best when it's hot to the touch in a cold room (the bigger the differential, the more efficient it is) and in many ways the active radiator is much the same. Whatever order you put the blocks in, it'll all end up the same temperature anyway. I'd route my piping for the least possible length of tubing, and that does generate back pressure, and maybe pump from the reserator through the blocks into the Bayres then pump from that into the active radiator and back to the reserator.

The IX38 has really good PWM fan control and I would be very tempted to use good PWM fans daisy-chained off the CPU fan header on the radiator and set it to switch off the fans to the radiator while the CPU is running relatively hot (55C?) then come in 100% to cool the fluid down ASAP. My bet (from a similar system I'm running here) is that the fans won't kick in very much at all, if ever.

If you want the system I'd really recommend then get a another decent passive external radiator eg. Innovatek Konvekt-o-Matik or Cape-Cora and use that instead of the 240mm active radiator.

Just quoting general sentiments expressed in the Watercooling uberforum that is extremesystems. Or perhaps you should have a look at http://www.martinsliquidlab.com for unbiased testing and a couple of excellent tools for calculating flowrate.
I have the following: TFC240, EK Dual Turbo Top with 2 DDC3.2 18vs, TFC Shrouds and 4 Xilence fans so do have experience of the EK Supreme. For BEST PERFORMANCE use on its own with a flowrate > 1.5GPM. Thats why i'm sitting here in front of my pc (Athlon 64 X2 4200 clocked to 2805 mhz) and i'm averaging 24 Degrees C with an ambient temp of 19 Degrees C.
I'm sure that you can utilise an EK Supreme with a NB, Dual GPUs and use a Zalman to drive it - If that makes you happy cool. Its just not the way I would do it.

platinum87 · 6 Aug 2008 at 00:23

Ben M said:
series =\ don't use parallel for a water loop.

why not? it would be cooling the water before moving on the to GPU, rather than the CPU warming it up, then having warmer water cool the GPU.

tophee · 6 Aug 2008 at 17:04

As a short term measure, which pump would be recommended to replace the pump in the reserator? Preferably one that would sit where the current one does (ie in the bottom of the tube).

Thanks

Ben M · 6 Aug 2008 at 19:11

platinum87 said:
why not? it would be cooling the water before moving on the to GPU, rather than the CPU warming it up, then having warmer water cool the GPU.

A parrallell setup will have rubbish flow. Water won't split its flow the way electricity does...

MikeTimbers · 6 Aug 2008 at 20:59

Ben M said:
A parrallell setup will have rubbish flow. Water won't split its flow the way electricity does...

Actually, I think you'll find that it is exactly like electricity. Say there are two pipes, one is ¼" and one is ½", two thirds of the water will go down the bigger pipe and one third down the small one. If there was an electrical circuit with two wires of different resistance, the flow would split between them in exactly the same way.

The problem with parallel loops is being able to determine exactly what the resistances of each loop are otherwise the cooling will be a complete guess.

***F1ZZY*** · 7 Aug 2008 at 14:24

ignore

platinum87 · 8 Aug 2008 at 13:53

Ben M said:
A parrallell setup will have rubbish flow. Water won't split its flow the way electricity does...

then how about better pump, or actually 2 pumps?... would it be better then?

1kiwi · 8 Aug 2008 at 17:11

Ben M said:
A parrallell setup will have rubbish flow. Water won't split its flow the way electricity does...

Yes - but the problem with putting everything in series means that the component closest to the rad gets the worst cooling. If you split you get equal temps.

Mekrel · 8 Aug 2008 at 17:33

beh* said:
Yes - but the problem with putting everything in series means that the component closest to the rad gets the worst cooling. If you split you get equal temps.

Order of blocks in a loop doesn't affect temperatures, so that's not true.

fornowagain · 8 Aug 2008 at 17:49

beh* said:
Yes - but the problem with putting everything in series means that the component closest to the rad gets the worst cooling. If you split you get equal temps.

That's rubbish I doubt a typical system with a reasonable flow rate would see much difference anywhere in the loop.

Just checking my numbers 1.5GPM and 250W heat load is around 0.5C maximum water delta, or a slower 4 Ltrs/min and 500W load is 1.8C.

1kiwi · 8 Aug 2008 at 21:03

fornowagain said:
That's rubbish I doubt a typical system with a reasonable flow rate would see much difference anywhere in the loop.

Just checking my numbers 1.5GPM and 250W heat load is around 0.5C maximum water delta, or a slower 4 Ltrs/min and 500W load is 1.8C.

Well i stand corrected

ZedPowered · 8 Aug 2008 at 22:08

i have tested my loop with a themometer (infared) and it is +/- 0.5 c throughout