Check out my new water cooling setup (many pics + details)

[Cptn. Foo Foo]

I have been an avid reader of ProCooling and this forum and I thought I would share my design and results with those who could understand/appreciate it



I finally got around to taking some pictures of my most recent cooling design. I made a small webpage which covers everything.

http://members.monarch.net/npoirier/new_box



Some info I forgot to include in my web page. A picture of the pump (with my old Eheim 1046)




And a link to the pump manufacturer's page:
http://www.hagen.com/canada/english/...03004160011001


I bought the pump at a Petland store in Lethbridge, Alberta (it was on sale so I just had to get it )


Comments please

[Kwissus]

With the 120mm fans maxed out the coolant gets as much as 3-5’C BELOW ambient :-)

Eeeeh?
Its impossible to go below ambient with radiators, you cant get the water cooler than the air cooling it.

[Cptn. Foo Foo]

I have two CompuNurse digital temp sensors. One sensor is hanging out the front of my case (measuring the air temp) and the other sensor is taped to the side of the copper pipe piece which splits the 1" flow from the pump. You can see the sensors in the pics if you look hard enough (my cheap camera does not take really sharp pictures unfortunately).

When I turn up the three 120mm fans to max (again each pushes 90 CFM @ 12v) the second temp sensor reports a coolant temp a little below the other sensor measuring the air temp.

I know the two sensors are identical to each other because they both report the same air temp if I have them both measuring the air temp. Plus as an accuracy test I got a glass of cold water with some ice cubes in it and dipped the sensors into the water. They both reported the same temp (about 2'C).

I am pretty sure my temp readings are accurate.



EDIT:

Could the air temp at the floor (where my intake hole into the box is) be 3-5'C cooler than the air 3 feet above it (where my air temp sensor is hanging out of the case) ?

[nikhsub1]

The 90 degree fittings and tee's are killing your flow, they make me dizzy!

[Since87]

Originally posted by Cptn. Foo Foo

I know the two sensors are identical to each other because they both report the same air temp if I have them both measuring the air temp. Plus as an accuracy test I got a glass of cold water with some ice cubes in it and dipped the sensors into the water. They both reported the same temp (about 2'C).

I am pretty sure my temp readings are accurate.

EDIT:

Could the air temp at the floor (where my intake hole into the box is) be 3-5'C cooler than the air 3 feet above it (where my air temp sensor is hanging out of the case) ?

I don't know if the air temperature in your room varies that much, but you can test it right?

Does the CompuNurse use thermistors for sensors?

If so, thermistor self heating maybe the cause of the error. If one of your thermistors is heatsunk to a pipe, and the other is suspended in slowly moving air, the heatsunk one will not rise to as high a temperature as the other, because the power dissipated in the heatsunk thermistor is removed more efficiently.

It's easy enough to test whether this is the case.

Edit: Just noticed that what I'd written earlier made no sense.

[Zhentar]

For the readings, stick the water sensor in the flow (I have my sensor sticking out of my reservoir into the tubing just a bit) and the other one directly in front of the air intake, either near the radiator if the fans are pulling, or right in front of the fans if they're pushing. That should get you reasonably accurate temps.

[Cptn. Foo Foo]

The 90 degree fittings and tee's are killing your flow...

I am very happy with the flow through the entire system. That pump moves about 6 times the volume of my previous pump (Eheim 1048). When I pull one of the 1/2" return hoses out of the reservoir (bucket) the flow coming out is crazy. Like it could fill a gulp cup in a few seconds easy.

When I first got the pump I tested it in the kitchin sink (BIG mistake! - I should have used the bath tub). I was unprepared for the awesome volume of water it was moving. It emptied the sink as much as it could (about 1/2 way down before the inlet was not submerged any longer) in about 3-4 seconds. I just stared at it and grinned

And I know 90 degree bends and such restrict flow but I am not worried in the slightest because I have so much flow to start with

[ralf_c]

hi all

i was looking at your systems picks, i have two questions. did you build your h2o loop for quiet or maximum cooling.

[Cptn. Foo Foo]

I built my latest rig for both performance and for low noise levels. Even with the 120mm fans turned way down my coolant still keeps really cool (no more than 3-4'C above ambient). And my massive pump ensures great flow rates.

When I turn the fans down low the loudest thing in my system is my WD Raptor 10K RPM HD.

Near-silence is bliss

[Hellion_Prime]

I like your implementation. Your box design is better than mine by a fair shot. I like my little giant pump better, tho.

I concur with the others that the 90 degree turns are definately flowrate killers. You might want to investigate a way to get rid of them for your next version. All in all, it looks fairly spiffy.

If you want to take a look at my attempt to get some possible ideas, the pics and such can be found here

[winewood]

How quiet is that pump? How much did you pay for it?

[Cptn. Foo Foo]

Thanks for the comments Hellion_Prime. Your rig looks pretty cool too! I also noticed your pics are a lot sharper than mine. I guess my crappy Logitech Quickcam just doesn't cut it anymore (if it ever did to begin with).

The pump is not silent but its not loud either. The smaller Eheim's I used to use are quieter than this Laguna one I have now. The Laguna puts out a powerfull humming noise but as long as you isolate the pump so its not vibrating the box/case its not a problem. Again, when the 120mm fans are turned down all I can hear is the clicking/ticking of the hard drives really (and the power supply fans).

As for the price well I cant really remember the exact price. I know the pump was on sale when I bought it. I think it was around $180 (Canadian) regurally but I got it for around $120. The store had a bigger Laguna pump on sale as well (the PowerJet 4000 model I think) but that pump looked absolutely, extremely overkill for PC water cooling.

[Blackeagle]

I'd agree that there are to many 90's in your set up. Just cause you have pretty good cooling as it is, doesn't mean you couldn't get better.

And those 6 restrictive fittings have got to have more than halved your flow rate. Just changing to some long sweep 90's or pairs of 45's would help a good deal compared to the elbow 90's now in use.

Could real easy gain you a C or 2.

[copyman]

Helo friend, could you post a screwshoot of your pc rotating to 2700mhz with the mbm5 cpu z and a prime95 rotating some 20 interactions?

it would also like one to 2800mhz, without prime95, to register new world record of oc for a xp 1700+ with watercooling

[nicozeg]

Originally posted by copyman
Helo friend, could you post a screwshoot of your pc rotating to 2700mhz with the mbm5 cpu z and a prime95 rotating some 20 interactions?

it would also like one to 2800mhz, without prime95, to register new world record of oc for a xp 1700+ with watercooling

That smells like he is speaking about PR numbers and confusing it with MHz.

[pHaestus]

I thought I remembered someone from these forums hovering around 1800MHz with an XP and chilled water?

I really dig that box for the cooling loop. Very nice work. The MCW5000 blocks are apparently pretty low restriction and so I am not sure fiddling with the 90s would gain you any cooling performance. I have a smaller Lagnua pump (the ~350GPH statuary pump I modded to run inline) and it has been a nice unit.

[copyman]

Originally posted by pHaestus
I thought I remembered someone from these forums hovering around 1800MHz with an XP and chilled water?

I really dig that box for the cooling loop. Very nice work. The MCW5000 blocks are apparently pretty low restriction and so I am not sure fiddling with the 90s would gain you any cooling performance. I have a smaller Lagnua pump (the ~350GPH statuary pump I modded to run inline) and it has been a nice unit.

That bomb is exactly what I would not use, shoot off head to 1,7m, imagine what is that value with the pressure drop caused by the reduction of the exit diameter, the fall in the flow rate is brutal

Uniting with the characteristics of 3 blocks, the mcw5000 that has low thermal efficiency at the low flow rates, the efficiency of the system should not be very good.

it is something that I cannot foresee correctly, but his final flow rate should be below 4 lpm after the cpu waterblock

and that is not good for a mcw5000

one provides that I would take, it would be to modify the impeler to reduce the flow rate and to increase the shooot off head, reducing the diameter of the rotor

[copyman]

Originally posted by nicozeg
That smells like he is speaking about PR numbers and confusing it with MHz.

I think was also that...

[Cptn. Foo Foo]

OK here are some older benchmark pics. I have yet to get the time to really put the new cooling design through a good series of benchmarks.


My CPU @ 2500MHz (not PR...real MHz)



And two at 2700MHz






EDIT:

The voltage @ 2700MHz was 2.05v

The details on my CPU: JIUHB 1700+ Week 10, 2003

AXDA1700DLT3C
JIUHB 0310 XPMW
9393056271118

I might have some old 3DMark results saved on their site. I will look.

[Cptn. Foo Foo]

OK. I found an old 3DMark 2001 result I have saved on their site.

I got this benchmark result using one of my older water cooling designs actually (and some older PC parts).

http://service.futuremark.com/compare?2k1=6604347

- I was using a week 8, 2003 JIUHB then
- I had "generic" PC2700 RAM that I was overclocking to 400MHz (it actually went as high as 444MHz !!!)
- I was only using one heater core (and one 120mm fan)
- and I was still using my old Eheim 1048 pump

If memory serves me correctly 2650MHz was the fastest that week 8 CPU ever went. It's in my mother's PC right now (running 2100MHz with some crappy aluminum heatsink )

[Pritorian]

What type of HC is that you are using?

[Cptn. Foo Foo]

What type of HC is that you are using?

HC = Heater Core? The heater cores I am using were bought at Canadian Tire. I do not know which exact models they are - all I said was I wanted a copper/brass heater core so big (I used my hands to convey the size I was looking for). The guy went back to go look and he came back with one that was perfect. Eventually I bought a second one. They were $49 each (Canadian $).

The cores are 8" x 7" x 2" thick.

-----------------------------------------

Here is a picture I just made to show how the cooling system is designed. Note, the picture is not to scale by any means and it does not show exactly how the hoses run. It is just to show what parts are connected to what and in what order.




I might get time today to put a few hours into doing some intense benchmarking.

[Pritorian]

Yes i meant heater Cores

And thanks!

Nice caseing all seems too be a really good performer, keep up the good work.

[copyman]

I feel friend, but results without the mbm 5 are not been worth, because it is the only that cannot be wrong for 2 or 3 malicious sotwares that exist in the moment

Below it shows screens for record approval







Everything that you have to do to order the records of the poor copy here for hell is to post the images

[Cptn. Foo Foo]

OK OK I took some time and ran a stress test.

Here is what I got @ 2676MHz (444MHz FSB)

I will work on 2700MHz+ later in the week.
My CPU water block is still not seating properly.
I know this because my CPU temps are too high.

Hell...maybe I should get a White Water