Leak testing new machine
 

http://www.tackyguy.com/images/pc060004.jpg

This is my second unit to build and decided to go copper pipe and use tubing at the connections. Two leaks found: 1 at the pump on the exit connector and 1 at the valve, so looks to be a go. When continue leak test for a few days.

Details

Chiefttech case Yellow Matrix
MB: Epox 8kha+ rev 2
cpu: t-bred xp1800
pump: ehiem 1250
WB: swiftech mcw462-u
pipe: 1/2 id copper
rad: dtek pro combo
Fan 120mm AOC

I guess it's not a "mobile" computing solution. Seriously, I like the looks of the copper piping and by mounting the rad outside, you cut down on the 90 degree elbows. How soon before you have numbers?

In 2 weeks or so
 

Well I just ordered a ps Enermax 430W(P4)ATXEG465P-VE should get that next week. Next time will be mem and I'll fire it up and start taking temps readings.

I really thought about doing an inclosed unit, but I didnt want to add the heat to the case so I just mounted on the outside.

Would it really have been "that" much heat being added if you did an enclosed setup?

I mean...really

You know, I keep think whats the deal with 90 degree bends.

Think about it. Look at your home how many 90's are there to say your shower? I bet more than you think. Do I have a pump in my house to boost pressure? more than like not since its gravity fed, and yeah each city uses pumping stations that distribute water through pipes with countless bends and restrictions. So if this worry about 90 degree is so terrible, I should only have trickle coming my my water falcet.

The effect is so small as to be less of a concern than 3/8 od tubing or a poorly designed water block. Yeah I have increase back pressure but so has having to raise the water 1 ft off the ground.

My goal was to do a clean look, I dislike the dangling mass of tubes I see in other cases. Next up is to tack the cables to the back of the case using copper pipe holds, the same for power cables and such.


Well I have the first results for the test run of the machine I have now dubbed yellow submarine

http://www.tackyguy.com/images/YELLOWSUB.jpg

Not bad for first time after flicking the power on and using cheap generic pc2100 and an old 33mhz hd.

yeah, about those 90 degree turns

say I was to put in, say 3/4 inch ID PVC pipe to neaten the case up a bit... now I'd need several 90's in order to keep it off to the side, etc

would having those 3/4 inch 90degree tursn cause a problem, since they are obviously quite a bit larger than the water which is piped elsewhere, like in the block, etc?

ehensley:

Pretty sure neither wells nor municipalities use hobby pumps of mag drive type.

True, but cities are running how many miles and how many t's and y's dealing with leaks and blockages, and pipe size differences. So I think you'll see the facts are that the impact of this 90 degree crap is not that important and really doesnt impact the final out come.


I'm maintaining 33 to 34 c with 100% load and system has been running for 24hrs none stop.

Here is some relevant test data:

http://www.coolhardware.co.uk/module...owcontent&id=9

And by all means check out Crane tech document #410 if you can find in local library.

I must also comment that the CPU temperature you post is from an in-socket thermistor and may not be the best guage of rather processes such as pressure drop and lowering of flow rate affect cooling capacity.

looks like a real PITA to change motherboard or cpu

PH
I don't think the coper elbow have the same effects as elbow for silicone tube since the claping system for the elbows for silicone tube put some pressure on the elbow, so you need a certain size of plastic to support that pressure.
On the other side, coper elbow have about 1/16" and shouldn't reduce as much the waterflow.
I know it will reduce it, but it's not as dramatic as you seem to point it. Maybe it's more dramatic on silicone tube.

About the water system for cities, did you ever see the (/edit isn't the main pire, it's a secondary pipe) main pipe for the system break ? There was one that broke in montreal this summer, it completly flooded the street (boulevard pie-IX) where it broke and also flooded some homes. The entire boulevard was blocked because there was around a foot of water in a mater of around 2 or 3 hours. There was no pressure in the rest of that section (which was only one of the many section in montreal)

Right around the Stade Olympique?

The restriction should be the same, wether it's copper or PVC (or other plastic) since the inner surface smoothness is about the same.

However...

I do believe that the copper elbows are more rounded than PVC ones, which will reduce the restriction a bit.

Note: the restriction in small diameter pipes is fairly low (versus large ones).

I don't believe that Ehensley would even be able to measure a CPU temp difference, with everything nicely rounded.

...but I will object to the home/shower comment: the flow rates in a house are small compared to what some of us use in our rigs. More flow = more pressure = more restriction.

The 90 degree barbs in that coolhardware thread are indeed a completely different animal to the copper sweeps. It is the difference between a design where two cylinders are connected at right angles and that of a single bending tube. Two 45 degree sweeps are even better. All induce pressure drop and therefore lower the amount available to the wb.

I mainly wanted to make it clear that comments such as "put as many 90 degrees as you want" were properly addressed.

Most hobby pumps are poorly suited to the task of wcing even with detail spent on optimizing flow via low restriction in tubing and rad. Adding changes in flow direction makes things worse. Whether or not this substantially influences CPU temps will depend on the block and pump used.

bb2k yes it was around the stade olympique