Calling all water cooling gurus

T

the madcow

Limp Gawd
Joined
Sep 11, 2002
Messages
469
I was wondering if i could get some suggestions to improve this setup and how things are routed for some reason i cant see of doing it any other way. Thanks

Specs in order:
D5 Pump
Floppy Dual res
Thermochill 120.3
DD Chipset Maze4
Storm
NV-78 SLI


img08482pj.jpg

Larger Version
 
You want the RES BEFORE the pump, you don't want the pump pushing into the res.. You should do this:

Floppy Dual res
D5 Pump
Thermochill 120.3
DD Chipset Maze4
Storm
NV-78 SLI
 
Or failing that
Floppy Dual res
D5 Pump
DD Chipset Maze4
Storm
NV-78 SLI
Thermochill 120.3
 
I should mention that the order of components really makes NO difference, EXCEPT the res pump deal. The res should ALWAYS be feeding (before) the pump. From there, whatever.

Res > Pump > Whateveryouwantmakesnodifferenceatallreallyihopeyouaregettingdizzynowlmfao :p
 
that looks like a nice setup, what $600 in the best [in my mind] parts available
 
If you talk to Danger Den they would say for the pump to suck from res and the outlet to push on cpu so it gets the most turbulance.

For example
Res pump cpu chipset SLI rad res
 
Get rid of all that evil high flowing dark side stuff and come over to the light side of the force. ;)
 
Thanks guys for all the help Im going to spend this weekend trying to get everything in the right order. Ill post some pics of the new config once its all done. Once again thanks
 
Top Nurse said:
Get rid of all that evil high flowing dark side stuff and come over to the light side of the force. ;)
Click to expand...

It's useless posts like this which are turning the watercooling forum into the "Video Card Forum: Take Two"

While it may be tongue in cheek, it just isn't needed at all.
 
I would keep it the way it's at right now, although if you could make it go from the pump > rad > Storm > maze4 > res, that would be the best solution.
 
Erasmus354 said:
It's useless posts like this which are turning the watercooling forum into the "Video Card Forum: Take Two"

While it may be tongue in cheek, it just isn't needed at all.
Click to expand...

Most people that are concerned with performance don't take Top Nurse's suggestions for small bore systems too seriously. Don't worry. :p
 
Erasmus354 said:
It's useless posts like this which are turning the watercooling forum into the "Video Card Forum: Take Two"

While it may be tongue in cheek, it just isn't needed at all.
Click to expand...
and its posts like this that start the flaming.. Just let it slide man.. it was said in all good humor... chill man.
 
So your saying that 1/2 tubing is better than 3/8 tubing in a restricted sytem .

How much of a pressure increase would be required to produce the same flow using 3/8 tubing compared to 1/2?? If you achieved that pressure increase to get the same flow.. Would the higher pressure with the same flow achieved give better results with a water block such as a Storm?
 
nikhsub1 said:
I should mention that the order of components really makes NO difference, EXCEPT the res pump deal. The res should ALWAYS be feeding (before) the pump. From there, whatever.

Res > Pump > Whateveryouwantmakesnodifferenceatallreallyihopeyouaregettingdizzynowlmfao :p
Click to expand...

This man speaks the truth.
 
You may be able to take some tips from my old system I have just started using again.
It produces some amazing results and I put it down to using 1/2" (internal not external) tubes where possible, no reservoir and a sizeable gap from the fan to the radiator to reduce air turbulence.

The reservoir produces a lot of water turbulence within itself and does reduce water flow rate/pressure so I decided not to use one.
Instead of a reservoir, I have a hosepipe connector that joins the split pipes together on the tube highest from the floor (as this is where most air will collect in the system). Its very easy to top up the system by opening the connector and pouring water in.
A very small amount of air is left after rejoining the hoses but it is no detriment to performance and makes no noise.
I used to top it up about every 3 months and took less than a minute to do.

For the curious, the kit consists of:
Ford Sierra car heatercore mounted in a sealed Stella carboard box fed by a single 120mm fan running at 7V :D
(its a random fan and I dont have cfm figures. Its almost silent at 7V)
Eheim 1048 pump
1/2" clear tube
Asetek Antarctica 1/2" water block (this uses 1/2" OD tube which was simply inserted into the 1/2" ID tube and gently clamped with a jubilee clip)
Arctic Ceramique paste

I have only just put this on my Athlon64 so havent got any figures for you yet but it worked fantastic on my old Athlon XP 2500+. Using 2.0V Vcore, this ran 24 hours a day at 2.7GHz rock stable! I remember max cpu temp with an external sensor was 37C, ambient 22C.

Another little extra I did that may have no impact but then again...
I put a "tiny" bit of washing up liquid in the water. This provided 2 benefits and possibly a third
1) washing up liquid is a surfactant. This removes the surface tension of water allowing the cpu block and radiator to have first class contact with the water. There are many other sufactants but this is a lot cheaper and appears to work very well!
2) I didnt once get Algae in the system despite running near a window for over a year!
3) its possible the washing up liquid somehow allows the water to flow more smoothly. I cant qualify this, it was thought of while trying to explain why the water cooler I built worked so well.

I hope to get some figures with my venice 3000+ after I've sussed a strange clocking limit.
Good luck getting your system working how you want :)

edit:
for correctness I checked the temps I had from previous posts.
at 22C ambient, idle was 30C, load was 38C
 
Always res before pump. Maze 4's would be better than those ass full coverage blocks as well.
 
It was an Abit AN8 motherboard and the internal sensors hardly changed temp at all despite mad changes in Vcore and clock speed. I didnt believe the figures and the temp change resolution wasnt enough to see what was happening.
This forced me to use an external sensor which gave great results.

The motherboard temps varied between 27C and 32C.
27C was really 25C or lower and 32C was really 37C or higher. It was hard to get it to display more than 32C with the motherboard sensor.
 
So what were those ABIT internal temps. that made little differance . what kind of water blocks do you have and were did you install those reliable external sensors
 
The motherboard temps are the internal sensor cpu temps, my bad.
The external sensor was taped in place as close to the cpu core as possible before mounting the heatsink.

I gave a full description of the water cooling hardware a few posts ago.
 
$BangforThe$ said:
So your saying that 1/2 tubing is better than 3/8 tubing in a restricted sytem .

How much of a pressure increase would be required to produce the same flow using 3/8 tubing compared to 1/2?? If you achieved that pressure increase to get the same flow.. Would the higher pressure with the same flow achieved give better results with a water block such as a Storm?
Click to expand...

Yes you want to reserve all the pump's power for the WB.

Back to topic; here's my suggestion..

Mount the pump on top of the hdd rack with the inlet pulling straight out of the res.

Res out, pump, rad, cpu, nb, bottom vga, top vga, res in.
 
Pherret said:
Yes you want to reserve all the pump's power for the WB.

Back to topic; here's my suggestion..

Mount the pump on top of the hdd rack with the inlet pulling straight out of the res.

Res out, pump, rad, cpu, nb, bottom vga, top vga, res in.
Click to expand...

I like that suggestion, except I would switch the SLI and the nb. There is already a nice section of tubing from the outlet on the storm to the top VGA. Have it go cpu > SLI > nb.

Also, in a very restricted system having smaller tubing doesn't really matter as much. The reason being the added restriction of the tubing is small and doesn't make an appreciable impact of flow rates. Still, larger tubing is always better, if not always by much :p
 
Chernobyl1 said:
The motherboard temps are the internal sensor cpu temps, my bad.
The external sensor was taped in place as close to the cpu core as possible before mounting the heatsink.

I gave a full description of the water cooling hardware a few posts ago.
Click to expand...


What were the M/B temps since a really like Abit M/B I await your ans. Its a simple request. no way were the internal temps readying anywere near as cool as the external temps.
 
You must log in or register to reply here.
Back
Top