Question about splitting flow percentage.

Danger-Darryn · 01-14-2003, 05:57 PM

I'm setting up my WC system but looking for more flow though the CPU WB then the NB & GPU

Using inline restricters in the tubing would I acheive 75% flow through the CPU WB and 25% through the GPU & NB? All tubing it 1/2" ID.

Bignuts · 01-14-2003, 06:37 PM

What I would do is just put a valve to restrict flow on the Northbidge line just before the rejoining of the two lines.

It would seem to be a easier way to restrict flow than trying to compute some thoretical tubign size/line length/water block restriction type of solution.

Personally, I just plumbed the whole stinkin mess in series, and I still get good results, bu I could see why you would want it that way (http://forums.procooling.com/vbb/sho...&threadid=5188).

Danger-Darryn · 01-14-2003, 06:43 PM

Yea, but I want to try this out before purchacing a ball valve. I'm just wondering now how many ppl on here have the Atlantis up and running in their systems? There doesn't seem to be much feedback from users at all reguarding performance.

Bignuts · 01-14-2003, 06:46 PM

Then just go and get a Hemostat and lock it down to a acceptable flow level of squish in the tubes to test flow, or use a pair of vice grips, etc.

Danger-Darryn · 01-14-2003, 06:52 PM

I think, pending results, I'll just go to a dual pump in a res setup, keeping the Eheim 1250 for the gpu & nb and something BIG for the cpu. If the cpu block wants flow, then flow it shall get.

Blackeagle · 01-14-2003, 07:44 PM

Why not run the whole flow through the CPU first, then split the flow to go to the NB & Gpu blocks after the CPU has had the advantage of all the coolest water and flow first? And I'd suggest having the rad between the pump and the CPU.

Did that rad turn out to be all SS? It sure looked like the fins and tubes in it were copper in that pic you posted on a differant thread.

Danger-Darryn · 01-15-2003, 12:57 AM

Yea, I did put the arrows going the wrong way, my bad.

The SS/Cu rad is in this thread and my Al/Cu here

8-Ball · 01-15-2003, 03:08 AM

Parallel pumps is not really a good idea, unless you have some kind of one way valve after each pump. If one of them dies, then it is likely that some of the water will go back through the other block loop, without passing through the rad. This would be bad.

However, if you had two pumps in series, and one of them fails, you will still be getting flow through all of the components.

I agree with Blackeagle, run all the flow through the CPU, then split off to the northbridge and the GPU.

8-ball

01-14-2003, 06:37 PM	#2
Bignuts Cooling Savant Join Date: Mar 2001 Posts: 155	If I were King... What I would do is just put a valve to restrict flow on the Northbidge line just before the rejoining of the two lines. It would seem to be a easier way to restrict flow than trying to compute some thoretical tubign size/line length/water block restriction type of solution. Personally, I just plumbed the whole stinkin mess in series, and I still get good results, bu I could see why you would want it that way (http://forums.procooling.com/vbb/sho...&threadid=5188).

01-14-2003, 06:43 PM	#3
Danger-Darryn Cooling Neophyte Join Date: Nov 2002 Location: Liverpool, UK Posts: 40	Yea, but I want to try this out before purchacing a ball valve. I'm just wondering now how many ppl on here have the Atlantis up and running in their systems? There doesn't seem to be much feedback from users at all reguarding performance. __________________ P4 1.8a @ 2888MHz 1.85v CPU Load/Idle 55/36oC 256meg PC2100 (generic crap) 2.5-5-3-2 2.65v Abit BE7 RAID (R.I.P BD7-II) CPU:MEM 1:1 160MHz Cheapo Coolmaster HSF Full Atlantis system under construction

01-14-2003, 06:46 PM	#4
Bignuts Cooling Savant Join Date: Mar 2001 Posts: 155	Oh Then just go and get a Hemostat and lock it down to a acceptable flow level of squish in the tubes to test flow, or use a pair of vice grips, etc.

01-14-2003, 06:52 PM	#5
Danger-Darryn Cooling Neophyte Join Date: Nov 2002 Location: Liverpool, UK Posts: 40	I think, pending results, I'll just go to a dual pump in a res setup, keeping the Eheim 1250 for the gpu & nb and something BIG for the cpu. If the cpu block wants flow, then flow it shall get. __________________ P4 1.8a @ 2888MHz 1.85v CPU Load/Idle 55/36oC 256meg PC2100 (generic crap) 2.5-5-3-2 2.65v Abit BE7 RAID (R.I.P BD7-II) CPU:MEM 1:1 160MHz Cheapo Coolmaster HSF Full Atlantis system under construction

01-15-2003, 12:57 AM	#7
Danger-Darryn Cooling Neophyte Join Date: Nov 2002 Location: Liverpool, UK Posts: 40	Yea, I did put the arrows going the wrong way, my bad. The SS/Cu rad is in this thread and my Al/Cu here __________________ P4 1.8a @ 2888MHz 1.85v CPU Load/Idle 55/36oC 256meg PC2100 (generic crap) 2.5-5-3-2 2.65v Abit BE7 RAID (R.I.P BD7-II) CPU:MEM 1:1 160MHz Cheapo Coolmaster HSF Full Atlantis system under construction

01-14-2003, 07:44 PM	#6
Blackeagle Thermophile Join Date: Oct 2002 Location: U.S.A = Michigan Posts: 1,243	Why not run the whole flow through the CPU first, then split the flow to go to the NB & Gpu blocks after the CPU has had the advantage of all the coolest water and flow first? And I'd suggest having the rad between the pump and the CPU. Did that rad turn out to be all SS? It sure looked like the fins and tubes in it were copper in that pic you posted on a differant thread.

01-15-2003, 03:08 AM	#8
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	Parallel pumps is not really a good idea, unless you have some kind of one way valve after each pump. If one of them dies, then it is likely that some of the water will go back through the other block loop, without passing through the rad. This would be bad. However, if you had two pumps in series, and one of them fails, you will still be getting flow through all of the components. I agree with Blackeagle, run all the flow through the CPU, then split off to the northbridge and the GPU. 8-ball

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)