How's this water cooling config for dual CPU setup?

[Fullmetal Chocobo]

Parts:
D-Tek Fuzion CPU block (2)
DDC-R2 12v pump
Swi9ftech MCR320-QP-K triple 120mm radiator
Swiftech MCRES-Micro reservior
Danger Den fillport - black
Primoflex 1/2" ID PVC clear tubing (10)

Loop:
reservior -> T-line -> pump -> CPU block 1 -> CPU block 2 -> radiator -> resevoir

Notes:
I do realize that this will result in a 1-3 degree rise in CPU 2 temps, but that is fine. I'm not going to ultra performance, just absolute stability.

I'm still searching for parts to do the t-line, but that will be added when I find them.

Water used will be distilled water with additives as described in aigomorla's WCing Guide.

Resevoir is going to be mounted next to a bank of 4 92mm fans while the radiator is likely to be mounted external to the case, probably width wise across the back. Although with the size of the case, it would be easily installed internally.

No GPU water block will be required, as I will be using onboard video for this machine.

Questions:
For proper operation of radiators, what CFM rating is recommended for the fans installed on the radiator? I know it depends, etc, but something of a baseline to work on would be nice.

 

[MotF Bane]

Try a Swiftech radiator.
http://www.jab-tech.com/Swifte...-Radiator-pr-3320.html
It outperforms the Black Ice series, despite costing less.

Koolance? Aluminum? That's a recipe for disaster (and lots of green crap in the loop). Plus, most of the Koolance blocks I've seen the insides of are highly restrictive; that MCP655 only has about 11 feet of pressure head, not sure if that would make it. You should find something else. Try a pair of Apogee GT's, those are universal sockets, so unless 771 is extremely weirdly shaped, it should work.
http://www.jab-tech.com/Swifte...ter-block-pr-3639.html

I don't think you need both a T-line and a reservoir in the same loop, the reservoir should serve the bleeding function.

As for the fans, probably http://www.jab-tech.com/YATE-L...l-Bearing-pr-3772.html these guys would do the job. Yate Loons, 45 cfm for 24 dba. It should be enough to cool it all down happily.

Wait for Aigomorla though, I could be wrong on a few points here.

 

[aigomorla]

Originally posted by: Fullmetal Chocobo
Parts:
Swiftech Apogee GT Extreme (2)
Swiftech MCP655 pump
Swi9ftech MCR320-QP-K triple 120mm radiator
reservoir -- last one I had was aluminum, so I need to find another.
Danger Den fillport - black
Primoflex 1/2" ID PVC clear tubing (10)

Loop:
resevior -> T-line -> pump -> CPU block 1 -> CPU block 2 -> radiator -> resevoir

Notes:
I do realize that this will result in a 1-3 degree rise in CPU 2 temps, but that is fine. I'm not going to ultra performance, just absolute stability.

I'm still searching for parts to do the t-line, but that will be added when I find them.

Water used will be distilled water with additives as described in aigomorla's WCing Guide.

Resevoir is going to be mounted next to a bank of 4 92mm fans while the radiator is likely to be mounted external to the case, probably width wise across the back. Although with the size of the case, it would be easily installed internally.

No GPU water block will be required, as I will be using onboard video for this machine.

Questions:
For proper operation of radiators, what CFM rating is recommended for the fans installed on the radiator? I know it depends, etc, but something of a baseline to work on would be nice.

Okey here is my recomendations.

Ditch the ApogeeGTX's and get D-tek fusions. They will help with the less restrictions. Or use a apogeeGTX on your first and a D-tek fusion on the second. Either way, in a dual cpu single loop, D-tek fusion is much prefered.

Your pump is lacking to push enough flow pressure for 2 gtx's. You'll see more then a 3-5C difference because of this. The GTX block is VERY restrictive. Given its not more then a storm, 2 will most definitely kill your flow. If you can, i would honestly try to split the CPU's in different loops. Or try to Y split the cpu's and run them parallel. You should see less of a performance hit if you do it this way.

Keep the Y's as close to the CPU block as possible, and try to use the least and = amount of tubing to both blocks.

Block 1 Block 2
Y both of inlet and outlets to both blocks. And then run tubes of = length from the Y's to the block.

 

[Fullmetal Chocobo]

Okay, I emailed D-tek to see if the D-tek FuZion supports LGA771 sockets and if it could be used in dual CPU configs.

In reference to the CPU blocks being in parallel, you would split them just before the CPU blocks, then join them again after the CPU blocks?

I couldn't find any metal 'y' connections, so would a 't' fitting like this work? Or should the 'y' be compressed some 1/2" -> 2 x 3/8" to maintain flow?

EDIT: Wow, I received a response in just under two hours--the D-Tek FuZion will work on LGA771 configs.

 

[aigomorla]

Originally posted by: Fullmetal Chocobo
Okay, I emailed D-tek to see if the D-tek FuZion supports LGA771 sockets and if it could be used in dual CPU configs.

In reference to the CPU blocks being in parallel, you would split them just before the CPU blocks, then join them again after the CPU blocks?

I couldn't find any metal 'y' connections, so would a 't' fitting like this work? Or should the 'y' be compressed some 1/2" -> 2 x 3/8" to maintain flow?

EDIT: Wow, I received a response in just under two hours--the D-Tek FuZion will work on LGA771 configs.

danny is a GREAT guy.

I love chatting with him on designs. And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.


And you can use nylon Y's. Why use metal? and yes you want to keep the Y close as possible to your blocks.

http://www.jab-tech.com/Wye-Fitting-1-2-pr-3276.html x2 one for both inlets and 1 for both outlets. :]

 

[MotF Bane]

Originally posted by: aigomorla
And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

What do you mean by accelerators? I can't find a picture of them anywhere.

 

[aigomorla]

Originally posted by: MotF Bane

Originally posted by: aigomorla
And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

What do you mean by accelerators? I can't find a picture of them anywhere.

there kinda like nozzles. You put them in the inlet portion of the d-tek, and its suposedly adds restriction, at the cost of water speed.

This then helps disapate more heat. Think of the storm.

 

[MotF Bane]

Originally posted by: aigomorla

Originally posted by: MotF Bane

Originally posted by: aigomorla
And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

What do you mean by accelerators? I can't find a picture of them anywhere.

there kinda like nozzles. You put them in the inlet portion of the d-tek, and its suposedly adds restriction, at the cost of water speed.

This then helps disapate more heat. Think of the storm.

That could be useful for the long-term. Storm means nothing to me, but I understand the rest of it. Thanks.

 

[SRoode]

Originally posted by: MotF Bane

Originally posted by: aigomorla

Originally posted by: MotF Bane

Originally posted by: aigomorla
And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

What do you mean by accelerators? I can't find a picture of them anywhere.

there kinda like nozzles. You put them in the inlet portion of the d-tek, and its suposedly adds restriction, at the cost of water speed.

This then helps disapate more heat. Think of the storm.

That could be useful for the long-term. Storm means nothing to me, but I understand the rest of it. Thanks.

So the accelerator reduces water speed?
IMHO, an accelerator should increase water velocity while decreasing total flow (nothing is free).

 

[aigomorla]

Originally posted by: SRoode

Originally posted by: MotF Bane

Originally posted by: aigomorla

Originally posted by: MotF Bane

Originally posted by: aigomorla
And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

What do you mean by accelerators? I can't find a picture of them anywhere.

there kinda like nozzles. You put them in the inlet portion of the d-tek, and its suposedly adds restriction, at the cost of water speed.

This then helps disapate more heat. Think of the storm.

That could be useful for the long-term. Storm means nothing to me, but I understand the rest of it. Thanks.

So the accelerator reduces water speed?
IMHO, an accelerator should increase water velocity while decreasing total flow (nothing is free).

I ment to say increase in water speed. Not reduction. Sorry i was busy at work.

But the accelerators causes more turbulance in the chamber. This allows a greater heat transfer and more efficiency of the block. Only problem is accelerators are concentrated in a single area. With Dual DIE quads, This is a VERY VERY bad solution. However on a dual core, Cathars G7 waterblocks pretty much dominate the market. If you can get your hands on one.

 

[nealh]

I love chatting with him on designs. And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

So the Fuzion with an accel will equal a GTX on dual core but the GTX will still exceed Fuzion with accel on quad core???

 

[aigomorla]

Originally posted by: nealh

I love chatting with him on designs. And what is even more tempting about the fusion is that the accelerators will be out next week for them. This should keep it above or very near = to a ApogeeGTX.

So the Fuzion with an accel will equal a GTX on dual core but the GTX will still exceed Fuzion with accel on quad core???

This is what im assuming neal.

Because the only real person to be able to test this is nikhsub1, and most likely a few other people who have the prototype nozzles.

But if you take physics into account, the accelerators will only hit he middle pin section. Quads are divided up like this

0 1
2 3


So your cooling concentration on the block will be the middle. The gtx however flows across from core 0 to core 3 diagonally. Hence the whole cpu is covered.

Does this make sense? But i will be getting the accelerators as soon as they become available. Then the fun will start on my end. :]

 

[MotF Bane]

Glad I have a dual-core; I can just get those with the rest of my loop. The DDC-2 is going to laugh at these. Wonder if they could make accelerators to aim specifically for each of the four cores?

 

[aigomorla]

Originally posted by: MotF Bane
Glad I have a dual-core; I can just get those with the rest of my loop. The DDC-2 is going to laugh at these. Wonder if they could make accelerators to aim specifically for each of the four cores?

i didnt recomend the DDC-2 because, you cant find it anymore.

The current mocks which run around pretending there DDC-2 are in fact DDC-3.1's and 3.2

They DO NOT perform as well as a DDC-2. This is why i recomended the D5.

The D5 should put out near the same head pressure, and be more reliable.

 

[MotF Bane]

I just got a DDC-2 pump; it has the orange impeller, torq head screws, and a label saying DDC-2TPHP and Mfr: 1/07. I bought that one in a store, so I checked the screws and label right there; I have a second coming from their warehouse, which I obviously am just trusting their markings. Once I confirm that one is also a DDC-2, I'll let everyone over at XS know where they can be found.