TJ07 + 2x MCR320 + MCR220 + 2x MCP655 = Complete

[NicePants42]

I just finished this build on Monday, and am now working on overclocking and testing the hardware.

The complete project log, from start to finish and including everything, can be viewed at Techgage.

I want to thank:
AndyM for posting his TJ07 project, featuring the best cable management I’ve seen.
MartinM210 for his MCP655 porting guide (you can see how I did mine in the log), his flow rate estimator, and his research on elbow performance.

Finished pics:













It weighs in at 63lbs

which is actually 4lb lighter than my previous setup.

Specs as follows:
Hardware:
Q6600 (lapped)
ABIT IP35-Pro
EVGA 8800GTX
4x1gb OCZ Reaper DDR2 800 (only $50/2gb AR!)
Cooler Master Real Power Pro 1000W PSU
2x WD Raptor 150s
and some other crap.

Cooling:
2x Swiftech MCR320s
Swiftech MCR220
2x Swiftech MCP655s
D-tek Fusion CPU Block
Swiftech MCW60 GPU block
EK Multioption 250 reservoir
0.5” x 0.75” tubing in the base, 7/16” x 5/8” tubing on the motherboard

In the future I may attach the fill port to the top of the reservoir instead of on the side, and I might try to hide the PCI Express cables a little better (for AndyM), and if I get really ambitious I might try to make a cleaner looking aluminum plate to hide cables/mount the res, but for now, I think it looks pretty freaking nice.

[disruptfam]

wow that set-up is sweet

is that all one loop??how are your temps??

Very well done my friend!!!

[NicePants42]

Yeah, it's all one loop. ATM, temps are questionable. At 3.4Ghz @ 1.55V fully loaded Coretemp shows 66C, but at a more modest 3Ghz @ 1.35V, loads are ~54C. Abit Uguru reports temps about 9C lower.

In either case, one thing I like is that as soon as load is removed, the temps instantly drop back to idle - there is no gradual drop. The radiators have never gotten hot either, so I know the loop is working.

At this point I still need to do a lot more reading up on how to OC the IP35-Pro - I'm starting to think that it's not stable at FSB over ~360.

[Spawne32]

That doesnt make sense, with all those radiators you shouldnt be at 66c, even at 1.55v, is the mount between the block and the CPU ok? somthing doesnt seem right to me.

[disruptfam]

is there such thing as having too many rads?? in 1 loop? could that be affecting his temps,I was reading somewhere on here that at some point adding more rads actually makes temps worse...

I guess having 2 d5's would help that though

[RockfordFosgate]

Good work bro

I love TJ07 with watercooled systems

[NicePants42]

is there such thing as having too many rads?? in 1 loop? could that be affecting his temps,I was reading somewhere on here that at some point adding more rads actually makes temps worse...

I guess having 2 d5's would help that though

I agree that 66C seems high, even with high voltage. I did test the block mount before securing it and everything looked good, but I may try to tighten the screws down some more even though they're nearly bottomed out.

The only way radiators will hurt temps is by adding restriction, so yes, the double MCP655s should eliminate that problem.

And like I said, the reported temperatures change instantly once the load is off, so it's not like the heat is hanging around.

[Nickel020]

The CPU temps seems high - what's your water temp and flow rate?

It could be that all the rads and tubing restrict the flwo too much, which is bad for the rads (can't properly cool water beacuse of low flow) and/or the CPU waterblock (can't cool CPU because flow is too low, even if water temp is low).

I'd check first if the the water block is properly seated thtough, that's most likely the problem.

Other than that, *very* nice system, I want!

[ceemic]

Very nice and clean setup

[Baleful]

Those temps are way to high, keep remounting untill you get the desired temps.

[AndyM]

Congrats! Talk about massive cooling capacity . I like the very neat setup, good job! I never played with cooling a quad core, but assuming double thermal output compared to dual core, temps look good at the speeds and voltages you are running at. You have the radiator surface area...now to work on the thermal interface between the block and IHS. I'm a big fan of Coolaboratory LM, and since you have a lapped IHS...you're good to go trying that TIM out.

[Spawne32]

The only thing i could think of is contact to the CPU is bad, or flow rate is bad.

[Sparky]

concave IHS perhaps

[dunn2953]

i get temps around that with MCR320>apogee gt>MCP655

[Martinm210]

Awesome looking build, very clean and professional

I really like the clean tube routing

[disruptfam]

i was just thinking do you have both d5's @ the same speed?

[shazza]

Very nice ... the res looks great in that position. I do like the TJO7!

[genec57]

I get about the same temps with 1.58 volt. I have a mcp320 and mcp 220. You can only extract so much heat no matter how many radiators you have.

[johnanderson]

Umm...is it me or is that you're cooling the gpu before the cpu. If its the other way around then I believe you have the inlet and outlet of your fusion mixed up. I would assume that both situations would account for your "high" temps.

[dr4gon]

Very nice and neat! That thing is a monster!

[Jedda]

Umm...is it me or is that you're cooling the gpu before the cpu. If its the other way around then I believe you have the inlet and outlet of your fusion mixed up. I would assume that both situations would account for your "high" temps.

GPU is before the cpu for sure, assuming the res is correctly placed on pump inlet.

[Waterlogged]

Bravo! Very clean build, I especially love the neat little trick of flipping the EK upside down.

[NicePants42]

Umm...is it me or is that you're cooling the gpu before the cpu. If its the other way around then I believe you have the inlet and outlet of your fusion mixed up. I would assume that both situations would account for your "high" temps.

Order of components does not matter in a closed loop. The GPU is before the CPU.

Bravo! Very clean build, I especially love the neat little trick of flipping the EK upside down.

Yes, however, one problem with this is that even though I now have both the return line and the fill port at the 'top' of the res, since they are both coming in from the side, there is still a small air pocket up there, and turning the pumps up to 2 or 3 will suck that air into the system. I've had some success using the 'top' plug as a bleed screw, but I think that the way I should've done this was to put the fill port coming out the very 'top' on the same side as the return line. I plan to make this adjustment in the very near future.

I tightened down the mounting screws on the CPU block to well past the spring's limit, so that now the spring coils are starting to offset and the whole stack looks wavy. This brought temps down ~6-7 degrees C, so I will have to remove the motherboard tray and remount the block with a little more care for level mounting and a little less care for the springs