TJ07 + PA120.3 + PA120.2 = Photos of my version of this.

[septim]

very nice. thanks for the parts # too.

[sick_g4m3r]

wow! excellent job!

[coolrmaster]

Wow great build and great wiring.

I wonder how the warmer air from one affects the cooling capacity of the other.

what i was gonna say....

[Solarfall]

excellent work AndyM... and the rig seems to work like a charm

[Makubex_GB]

Very nice work! So clean, I love it.

[nealh]

Andy

Is the DVDRW at the very top or one slot down....Corsair really need to make there cable longer...

[AndyM]

Andy

Is the DVDRW at the very top or one slot down....Corsair really need to make there cable longer...

At the very top. Did you get your case already?



On how one rad affects the other...yep, probably does some, but with all that surface area, the overkill aspect kinda makes you not really worry about it too much. Maybe with a quad core at these speeds it might be an issue; so reversing one rad's fans, removing the center fans and having both rads suck in might be an option. Maybe I'll try it that way when that time comes, but for now I like the cross-flow arrangement, keep the upstairs portion relatively cool.

This puppy is pretty toasty now after running Orthos all night in 80F room temperature (makes a nice space heater). Nothing scorching or anything like that, but nice and warm. But, looking at the TAT temp reading, seems to be doing the trick, kinda high for sure, but with 80F ambients, it got its shakedown and survived. All while running the fans at a nice quiet 7 volts, I love it! Time to ramp up the speed some more and see if that works (nice lower voltage 627B chip, but it really belts out the heat, lol)

[nealh]

My case is supposed to be delivered tomorrow...from axiontech
HDD cage fron newegg tomorrow as well
GTX should be here today or tomorrow as well

Going out of town next week so build will wait a couple of weeks at least

[septim]

maybe you need to elevate your case, take out center fans, make air channel in middle of bottom for air to come up and sucked by both rads...

[ColonelCain]

Great job!

All of these amazing builds really make me hope that I can afford to get a TJ07 for my May '08 build.

[elfy]

really nice job on the fan controller really wanna do something similar myself, but demi-worried about putting too manyfans on a single molex. have done electronics for a while so can't really figure in my head how, 3 fans @ 7v running on a 12v molex would work, so am pretty unsure wether to do it or not :S

[Grinch]

great job!

[louie101086]

i want to do samething you did but instead use swiftech rads how good is your flow

[AndyM]

Q&A's

Flow seems pretty good, the water moves from what I can tell when getting the air out (lots of tumbling the case with this loop ).

I dont see any problem with putting lots of fans on one molex...the wire guage is heavier than the thin guage wire on the fans, I think those can handle a couple amps, should be ok. Inside a power supply the connections are all bridged anyway, so wire guage is your only concern.

A little update...seems I got some bum Ballistix, I have really bad luck with memory, second time out of 2 times this happened. Did some testing and the memory doesnt do its rated spec with memtest, so something's funny there. And the random lockup at 423 MHz even after a 8 hour Orthos run is weird. Either that or I dont know what I'm doing, but I tried pretty much all MCH/FSB/Memory voltage combos, straps, timings, and speeds, still doesnt work right. So I put the old Corsair memory with those high-bandwidth Promos chips in them and am just running it at 3.7 GHz now. I will cherish the day I buy some fancy memory and it isnt DOA or refusing to do the rated spec.

Thanks for the complements on my machine guys!

[elfy]

it wasn't really the ampage i was worried about, just really unsure how the voltage would rack up, would they each draw 7v regardless? or with 3 attached would it try to draw 21volts?

[Okda]

ii have a small suggestion

why not make the loop goes NB block then CPU block then VGA block, i think this way u will get much better cpu temps and in the same time ur vga won't suffer much

[Jedda]

Sweet build mate.
Well done.

[AndyM]

Not too worried about loop order, just the neatest possible setup. Amazing how easy it is to get at the BIOS reset jumper now, lol.

elfy...they will each get 7 volts regardless of how many you put across the 12v and 5v leads, up to the current limit of your power supply (think maybe 200 fans or so before you goof something up).

[zgundam]

wow, amazing cable management and layout! great job!

I gave up trying to squeeze my PA120.2 underneath the case and just mounted it on a radbox on the back instead :p

Just one suggestion though - rather than using the 120mm fans on the top to suck air in, how about using the 92mm's as the intakes and the fans on top as the exhaust like it was setup originally?

Seeing as hotter air rises, it makes more sense (for me anyway) to use the 92mm's as intakes and the top fans as exhaust instead...

[Jedda]

Not too worried about loop order, just the neatest possible setup. Amazing how easy it is to get at the BIOS reset jumper now, lol.

lol
On my last box it took a pair of 14 inch surgeons' forcepts to reach and reset the bios jumper.

[AndyM]

Yeah, I can see doing the fans like origional, but then I was thinking the origional is pretty much set up for air cooling. Now, the only cooling I'm doing is the memory, the power transistors, the southbridge, hard drive, etc., with the memory needing the most cooling. And there is a more direct blast of cool outside air on the memory using the top fans going down (easily overcoming convection currents of warmer air). I figure that is better than dragging warmed up case air over the memory.


Amazing how .09 volt and 100 mhz can heat things up. Ambients for this testing run of 3.7 Ghz are one or two degrees less at 78F, but warmed up and running Orthos it is some 8 degrees C less than 3.8 Ghz. Now if I could get this thing to boot at 3.9 to see what temps would be like there, but no luck so far without pushing more than 1.55 vcore or if memory is the reason for the no boot, hmmm Guess I gotta wait for a RMA'ed set of Ballistix that can actually come close to running the advertised 500 mhz to see which it is Yeah, I could change the divider, but not real confident in how accurate that would be to judge the cause, as I'm not really confident of my 975X at this FSB.

[albertyin]

WOW very nice!!
and the temp is good too

[afireinside]

Looks slick! I have two comments:

The tubes going up from the bottom look a bit snug... Why so short?

Your flow must suck with cpu, nb, full cover gpu, and right angles on the rads with a single d5

[AndyM]

Looks slick! I have two comments:

The tubes going up from the bottom look a bit snug... Why so short?

Your flow must suck with cpu, nb, full cover gpu, and right angles on the rads with a single d5

Good eye, when I was looking at those photos I was kinda checking the computer afterwards to see if they were stretched out as they look. Optical illusion. That hose off the chipset really isnt that tight, just that chipset block nipple is so long, then the tubing kind of just falls downwards giving the look of pulling in a photo. I was thinking of ziptieing it to the hose coming up from the pump to the gfx block to get the look right, but I need to find some sort of 1/2" hose spacing block to make that look right also. The water moves pretty good though there looking at the bubbles when purging it, and after Cathar's thread on his flow testing, I'm not too worried that the system might not have some ungodly high flow rate.

[nitteo]

Really nice job!