Just completed my WC setup...

[ChillPhatCat]

I have just one question, I'm now running the comp and it's doing 37C Idle/ 43C Load @ 1.95V... I'm just wondering if this is about right or what? Considering there are still a couple air bubbles in the system and I haven't made a fan shroud yet.


http://chillcat.fortunecity.net/almost.jpg

Maze4, Mag 3 and Caprice HC on the back side... The case is a Lian Li PC61

I've got (2) 120 mm fans @ 60 - 95CFM ea. on the HC as you can see in the pic, I didn't have the HC installed in the pic yet but it gives you a pretty good idea.


I took out the 3 1/2" drive cage and put in a 92mm Fan which fits pretty well.
http://chillcat.fortunecity.net/casefront.jpg

[CGR]

Looks pretty good. I like that you used all brass fittings. Makes it look nicer IMO..

One thing I noticed was you have the water going to the CPU first then to the radiator. I beleive water goes out the top of the Mag pump. Might get better performance having it goto the rad first.

[ChillPhatCat]

yeah I am going to do that eventually... just have to figure how to mount the pump... I'm at 2V right now and 45C under load which isn't terrible... I'm still purging the system.

[obsolete]

I have that pump. The water does go out the top of the pump & in on the side (from the way it is positioned in that picture). I would switch your first "out" from the pump to go into the radiator. The radiator will cool your water & then have it go out to the CPU as the water will be cooler & provide better temps (more then likely).

I should have got a brass fitting for the intake of my pump. I used a plastic fitting that was the correct size. But unfortunately the intake on the pump cracked on top which caused me to purchase a Hydor L30, lol.........As much time as I spent with leaks & fittings & buying adapters. I pretty much told myself I just want a pump that has 1/2" fittings.

Anyways, change your loop & have the radiator as the 1st component to receive water from your pump. Should help some!!

[sandman001]

It really makes virtually no difference what order your setup goes in.

That being said, nice setup.

[CGR]

I wouldnt say it makes no difference... That pump heats the water up a fair amount and if you have a fair size rad with good airflow it should drop temps by a couple degrees if you place it between the pump and block..

[sandman001]

Cathar figured it all out using I think it was mag3 or eheim 1250, and the pump added .067C or around there.

[Soja]

Originally posted by Cathar
How about we inject a little bit of science into the equation eh?

Let's assume we have a Danner Mag 3, being a 35W pump. In actuality the pump will draw a low less power, but let's run with 35W.

Let's assume that the pump dumps all 35W of its power as heat into the water as the water flows through the pump.

Water has a thermal capacity of 4186 J/kg°C. Water has a density of 1.00

Let's assume that the flow rates through the system is 2GPM, or 7.5LPM, which is fairly reasonable given that pump and a moderate-high restriction block, or two low-moderate restriction blocks, and a low-restriction radiator.

Per second, 7.5/60 = 0.125l, or 0.125kg of water flows through the pump, for a total thermal capacity of 4186 * 0.125 = ~523W/°C

Keeping that the pump dumps 35W of heat into the water, then the water will rise by 35/523 = 0.067°C as it flows through the pump.

So the difference between having the radiator before/after the pump is just 0.067°C

So there's the mathematical/physics way to disprove the misconception.

[sandman001]

yep, .067C.

[CGR]

Well i dont know about his findings... but what i do know is when i put my system together and had the pump going to test for leaks and no air through the rad, the entire rad was warm. In order for that to happen it has to be more than .067c.

[sandman001]

if you left your system running, then yes it would warm up the water.

But, that isn't relative when you have fans going, and the whole system running.

[CGR]

point is the water raises by more than .067c because im sure cathar didnt take fans into consideration either.

[sandman001]

No, he did. The water is heated up by .067C each pass through the pump.

So, it only changes the temp by that much before it hits the rad and cools back down.

So, it only changes by more than that if the cooling in the loop can't dissipate the heat.

[csidd]

Aside from the pump things look nice. As for 0.67c difference, the only way to find out for yourself is to switch things up. You only need to change the in and out on the pump. I know it's not an easy change, but if your happy with your temps who cares.

I did notice you saying that there is still air in your tubes. When I set my system up I used a T that was about 12" filled up about 6". After a few days I lost another 2". Your T looks pretty short. I would keep an eye on your H2O levels to make sure you don't drop into the T and introduce more air into the line.

[ChillPhatCat]

It's almost all purged now... it's running a nice clear green color now. I have been watching the water level carefully.

Right now I'm at 2.025V and 2662 MHz... I'm going to have to stop here until next week since my PSU just can't handle the load... 480W true power coming in monday.

FYI AIB doesn't know how to make a stable PSU.

My 12V rail is at 11.4V when the CPU goes to 100% load which causes a hard crash... I don't know the load temp, but we're idling at 38C right now... I still have to make a fan shroud which should help a bunch...


I'm pretty happy so far at an 829 MHz overclock.

[sandman001]

Nice overclock, that has to be a screaming system :)

[ChillPhatCat]

Thanks :D Now I just need a good video card...

[TheNewbie]

Nice looking rig.

[mata2974]

good job

[ChillPhatCat]

I left the pump off all day and when I came home I fired them both up... I haven't passed 38C for the last 2 hours... It probably bled a bit more air out while I was at work... IDK I'll have to run more tests... I've got a better PSU coming in Tues so hopefully it will help me to push the limits a little more.

[Posidon42]

I have pretty much the same system as you. I also found the mag's hard to bleed when using a T-line. So for my next project, I am going to submerse my Mag5 in a FRIGGIN HUGE reservoir. More for looks than anything, but it will take about 3 seconds to bleed the system. I tested it out with a water bottle with some fittings (ghetto, I know), but it was great.

I also have the Maze4 - two of them actually and they do pretty ok. Not great, but ok. You are definitely getting better o/c than I was. I could get it stable at 11x200@1.8V but I didn't want to push the system anymore than that. I wanted to keep the fsb at 200 just to keep things simple and it seemed that 100mhz increments of o/c was a bit too much to ask. I think I tried 1.9V once or twice trying to get it higher, but to no avail.

So are these mobile versions better than the normal cores? I might have to do some reading...

[PolarExpress]

Kinda of a hijak but ....
Has anyone tried a submerged mag?
Would have zero flow restriction on the head no?

[ChillPhatCat]

I'll bump it with my newest finding... When I open the door to this room the idle temp drops to 35C since the rest of the house is at about 23C

The final internal setup:
http://chillcat.fortunecity.net/complete.jpg

I'm making a fan shroud today:
http://chillcat.fortunecity.net/rad.jpg

The 1/2" tubing w/ 1/8" sidewall fits perfectly inside the 3/4" inlet... a little silicon and tape seems to work perfectly:
http://chillcat.fortunecity.net/radfrt.jpg

I just love those 5 1/4" -> 3 1/2" adapters...
http://chillcat.fortunecity.net/frtbare.jpg