• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Best Routing Method for this Set Up

I

Ingonuts13

[H]ard|Gawd
Joined
Jun 9, 2003
Messages
1,982
Ok so what would be the best routing method then if I were going to use only one pump?

pump ->top rad -> cpu ->bottom rad ->GPU - > GPU ->PUMP

Also will this restric flow too much?

watercool.JPG
 
are you going to be using a T-line in the system?

Basically the only thing you want to worry about is using the most efficient tubing routing. I would say pump > single 120 > double 120 > cpu > gpu > gpu > pump


OR

pump > gpu > gpu > cpu > double 120 > single 120 > pump
 
Erasmus354 said:
are you going to be using a T-line in the system?

Basically the only thing you want to worry about is using the most efficient tubing routing. I would say pump > single 120 > double 120 > cpu > gpu > gpu > pump


OR

pump > gpu > gpu > cpu > double 120 > single 120 > pump
Click to expand...

I am not sure either of those makes sense as the pressure from the pum comes out of the top of the pump not the side. I would think you would definitely want the lead off the top of the pump to either go to the CPU or large Rad first since it will be straight up.
 
Ingonuts13 said:
Ok so what would be the best routing method then if I were going to use only one pump?

pump ->top rad -> cpu ->bottom rad ->GPU - > GPU ->PUMP

Also will this restric flow too much?

watercool.JPG
Click to expand...

id say mount the small 120 on the back fan slot assuming its a 120 and go like this

Pump - top rad - CPU - small rad -gfx cards - pump
 
0mega said:
id say mount the small 120 on the back fan slot assuming its a 120 and go like this

Pump - top rad - CPU - small rad -gfx cards - pump
Click to expand...

That is what I was thinking but I do not have room on the back slot for the rad so it will have to go in the lower front.

What about loop flow rate? Should this be really restrictive?
 
you just want to make the shortest run you can. Going from teh pump to the top, to somewhere, across the case, back across the case, and then back to the pump isn't the best way. Erasumus had it spot on.

The pump just circulates the water, it doesn't do the pressure thing. so it really doesn't matter where it goes from the pump, just that you have the shortest tubing run with the least bends.

flow5il.png
 
The only thing that has me worried about the shortest approach is the almost 90 degree turn i need to make in order to go PUMP-> Bottom Rad.
 
Howbout:
Dualrad -> pump -> singlerad -> gpu -> gpu -> cpu -> dualrad

It will still keep the tubing short and won't create as sharp bends
 
Anyone know how much heat dissipation power I would have with a system like this?

I have a Thermochill 120.2 for the dual rad and still need to purchase the single rad.
 
Suprisingly, those usually coincide with each other. \

As for the 90 degree turn... not a big deal.

For head dissapation, you'll have more than enough for what your doing, at least thats my guess.
 
well if your looking for the best cooling.... to me, it makes since to to get a rad inbetween the GPU's and the CPU. Your trying to cool 2 GPU's, than that hot water is trying to cool the CPU. Not too effiecient.. also having the cool water come out of the dual rad, just to go into the single rad, makes no since.. already trying to cool, cool water? Its redundant.

TO me the best loop for cooling and not too bad on runs is pump->single rad->GPU->GPU->dual rad ->CPU->pump.

as for comparision purposes, my friend and I have almost the same set-up, difference is he has 2 single 120 rads....in basically the same setup i have described about except replace the dual with a single....

I have a dual 120.. so we both have 2X120.. i just have a dual, he has 2 singles... but he put one inbetween his GPU and CPU.. we have the same GPU and CPU... and hes getting 4-5C better temps on his CPU becuase of the placements.

just my 2 cents

Chris
 
-You can cool almost anything with those 2 rads. You'll be just fine.

-The best tubing runs are whatever is shortest. If going pump->singlerad is too much for your tubing, then I would try to move the single rad, because the next-shortest routing (I'm guessing pump->doublerad->singlerad->GPU2) is *much* longer.

It looks like it would also be possible to turn your pump around, so that the inlet faces your single rad, and the outlet points upwards to your GPU blocks. I think that would actually be the best solution, if you're scared of that 90 degree angle.

-There's no need to put radiators between blocks, it will make little/no difference in cooling, and will lengthen your tubing runs.
 
I totally disagree with the "shortest run" is more important than the placement of the rads. I have 3 different WC setups i have access too. 1 of them has 2 runs alone of 5 feet(and about 14 feet of total runs). The temp difference with the one with only about 2 feet in total run? About 1 degree C. for 12 extra feet of hose.

Putting the single rad right after the double rad, does NOTHING for cooling. Nothing. The water going in, will already be as cold as its going to get, the single rad, will do nothing to cool the water further. You might was well take it out, for your "shortest run" theory.

Putting a rad inbetween the GPU and CPU will only increase your "run" by INCHES. Coolng the warm water coming from 2 GPUS, before it goes into your CPU will make a difference.
 
Thanks for the replys so far guys.

The idea was to do somewhat brady has been describing.

The pump up to the dual rad...to the cpu....then to the single rad then to the GPUS and back to the pump.

I have thought about turning the pump so that the inlet faces the rad but then how would I put the single rad between the CPU and GPUs?
 
bradyapba said:
I totally disagree with the "shortest run" is more important than the placement of the rads. I have 3 different WC setups i have access too. 1 of them has 2 runs alone of 5 feet(and about 14 feet of total runs). The temp difference with the one with only about 2 feet in total run? About 1 degree C. for 12 extra feet of hose.

Putting the single rad right after the double rad, does NOTHING for cooling. Nothing. The water going in, will already be as cold as its going to get, the single rad, will do nothing to cool the water further. You might was well take it out, for your "shortest run" theory.

Putting a rad inbetween the GPU and CPU will only increase your "run" by INCHES. Coolng the warm water coming from 2 GPUS, before it goes into your CPU will make a difference.
Click to expand...

All of this is totally wrong...
 
Ingonuts13 said:
Thanks for the replys so far guys.

The idea was to do somewhat brady has been describing.

The pump up to the dual rad...to the cpu....then to the single rad then to the GPUS and back to the pump.

I have thought about turning the pump so that the inlet faces the rad but then how would I put the single rad between the CPU and GPUs?
Click to expand...

You don't need the single rad between the CPU and the GPUs. See this link if you don't believe me.

By putting the single rad between GPU and CPU, your water will be about 1/4 degree cooler than if you left it out, which will make no real difference.
 
haelduksf said:
You don't need the single rad between the CPU and the GPUs. See this link if you don't believe me.

By putting the single rad between GPU and CPU, your water will be about 1/4 degree cooler than if you left it out, which will make no real difference.
Click to expand...


Ah, someone posting something accurate. What a breath of fresh air.
 
haelduksf said:
You don't need the single rad between the CPU and the GPUs. See this link if you don't believe me.

By putting the single rad between GPU and CPU, your water will be about 1/4 degree cooler than if you left it out, which will make no real difference.
Click to expand...

I don't see how this makes sense given that the example you posted uses 50W in it's calculation. The point of adding in the second rad is to account for the amount of Watts needing to be cooled by the system. I do not think a dual 120mm rad is enough to cool the rig I am trying to cool, hence the second rad.

If you can post something that makes sense for the type of system we are talking about then please do so. Everything I have read tells me that I do not have enough cooling power.
 
Ingonuts13 said:
I don't see how this makes sense given that the example you posted uses 50W in it's calculation. The point of adding in the second rad is to account for the amount of Watts needing to be cooled by the system. I do not think a dual 120mm rad is enough to cool the rig I am trying to cool, hence the second rad.

If you can post something that makes sense for the type of system we are talking about then please do so. Everything I have read tells me that I do not have enough cooling power.
Click to expand...

I didn't say that you didn't need the single rad. I only said that putting it between your CPU and GPUs wasn't any more efficient than putting it after, along with your double. Any temp gains you would get by running CPU-singlerad-gpus-doublerad would be more than offset by your longer tubing.

By all means, use both rads. But if you insist on doubling the length of your tubes, you're taking 1 step forward for 2 steps back.


Oh, and I'm not sure where you got that 50W number, but if you have a look at the calculations, you'll see that your situation isn't all that different.
 
Dude the one dual rad will do you fine. You can add the second rad but you will not see much of a performance increase. Plus you needlessly complicate your cooling loop.

haelduksf is correct...
 
Hahahaha, a beautiful link, and wonder example, but yet, doesn't apply. All your example is doing is cooling 1 CPU, and NO GPU'S, let alone 2! Also, what about OCing? and OC'd CPU throws a lot more heat, than a non OC'd CPU.

I get heat transfer 101, its not rocket science... hehehe, i kill me!

Energy transfer through water (cooling) is pretty easy to understand. How much hotter do you think the water is after going through 2 GPU's overclocked? a Mear .15 of a C as in your example??

What do you think has the ability to move more energy... the water after its been through 2 GPUS, or water out of the Rad?

And through my OWN testing, moving the 2nd rad between the GPU and the CPU got a 3-5 C difference. Same video card, CPU, case, fans, rads, etc.

Why do you think people but their SLI cards on seperate loops? Cant be becuase its more effiecent at cooling, can it?

And moving the Rad to inbetween the CPU and GPU, adds, 1 extra run of about 6 inches on the avg case. Using your theory of heat distrubution, that shouldn't be a problem at all, certainly is 2 GPU's aren't adding enough heat, how can 6 extra inches of tube?


To the OP:

Putting the 2nd rad in after the dual 120 is useless, i think thats agreed upon. So if you want to use it, put it inbetween the GPU and CPU... otherwise don't use it all.

I think still my solution of pump->single rad->GPU->GPU->dual rad ->CPU->pump is your best bet, and doesn't cuase much extra tubing at all.

My 2 cents.

Chris
 
Yeah that makes sense Chris, this is exactly what I was going to say to the last post but I was just too tired.

N E Ways, I just ordered a single 120mm Rad and will measure out some shit when it arrives. The plan is to put the Rad between the GPU and CPU if possible and if it only adds a couple inches to the loop then so be it.
 
bradyapba said:
Hahahaha, a beautiful link, and wonder example, but yet, doesn't apply. All your example is doing is cooling 1 CPU, and NO GPU'S, let alone 2! Also, what about OCing? and OC'd CPU throws a lot more heat, than a non OC'd CPU.

I get heat transfer 101, its not rocket science... hehehe, i kill me!

Energy transfer through water (cooling) is pretty easy to understand. How much hotter do you think the water is after going through 2 GPU's overclocked? a Mear .15 of a C as in your example??

What do you think has the ability to move more energy... the water after its been through 2 GPUS, or water out of the Rad?

And through my OWN testing, moving the 2nd rad between the GPU and the CPU got a 3-5 C difference. Same video card, CPU, case, fans, rads, etc.

Why do you think people but their SLI cards on seperate loops? Cant be becuase its more effiecent at cooling, can it?

And moving the Rad to inbetween the CPU and GPU, adds, 1 extra run of about 6 inches on the avg case. Using your theory of heat distrubution, that shouldn't be a problem at all, certainly is 2 GPU's aren't adding enough heat, how can 6 extra inches of tube?


To the OP:

Putting the 2nd rad in after the dual 120 is useless, i think thats agreed upon. So if you want to use it, put it inbetween the GPU and CPU... otherwise don't use it all.

I think still my solution of pump->single rad->GPU->GPU->dual rad ->CPU->pump is your best bet, and doesn't cuase much extra tubing at all.

My 2 cents.

Chris
Click to expand...

I see you've clearly read the summary, and have clearly not read the rest of the article. The in-depth examples below the summary contain the calculations I was referring to, which do not only talk about 50W of heat.

Fine, let's do this the hard way.

Let's assume that the OP has a massively overclocked processor (or is kinda slow and bought a Prescott ;)) that outputs 125W of heat. That 655 should be able to give at least 5 LPM of flow, and I have no idea how restrictive the OP's blocks are, so I'll stick with the flow rate calculations in the article. That means (with a perfect waterblock) that the water will be 125 / ( 5 * 69.77 ) degrees Celcius (or 0.358 degrees) warmer coming off the CPU block. Since I know for a fact that the OP's waterblock is not perfect, the actual rise in temperature after the CPU is likely to be closer to 0.3 degrees. This will have little or no effect on the temps of his GPUs.

As to your theory that "moving the Rad to inbetween the CPU and GPU, adds, 1 extra run of about 6 inches on the avg case", the OP has stated explicitly that he can't move the rad there, and that it has to stay at the front of his case- which would add at least 2 feet of tubing.

I have presented facts and calculations. Please reduce your personal attacks and hearsay.
 
1st let me start off my saying if you took it as a personal attack, that was not my intent. I apologize for anything considered an attack, i am just trying to help the OP. 2ndly, nothing I present as "fact" is hearsay, is a direct result of my own testing. The OP can take it for whatever it is worth. Why would i try to decieve him? Just trying to help.

Now, i did read your article. Its flawed. There is no GPU in the test. Only 1 CPU. It has nothing to so with what the OP is doing.

Now when you say "the actual rise in temperature after the CPU is likely to be closer to 0.3 degrees.This will have little or no effect on the temps of his GPUs" Your right. Why? Becuase after the cpu the water goes to the DUAL Rad before it gets to the GPUS. How much heat comes out of the CPU was NEVER in question. It was always going to the RAD before the GPU's.

The question is, how much heat gets transfered into the water from the 2 GPU's as the water then tries to cool the CPU after all the energy from the 2 GPU's are in the water. What will effiecently cool the CPU...that water.. or water that has had the the chance to have the GPU energy taken out of the water from the single rad??

And your right again, it would be 2 feet of hose, but i was under the assumption he was going to use the single rad... SO, what i meant was, with my run it would be an extra 6 inches if he used my run, vs the run of running the single rad after the dual rad. It is 2 extra feet if you dont use the single rad at all.. but i was under the assumption it would be used.

And my real world tests showed me the hose had little effect on temps, where as moving the Rad did. Those are my tests...and i hope my tests have helped the OP make a decision.

Enjoy the water cooling!
 
bradyapba said:
1st let me start off my saying if you took it as a personal attack, that was not my intent. I apologize for anything considered an attack, i am just trying to help the OP. 2ndly, nothing I present as "fact" is hearsay, is a direct result of my own testing. The OP can take it for whatever it is worth. Why would i try to decieve him? Just trying to help.

Now, i did read your article. Its flawed. There is no GPU in the test. Only 1 CPU. It has nothing to so with what the OP is doing.

Now when you say "the actual rise in temperature after the CPU is likely to be closer to 0.3 degrees.This will have little or no effect on the temps of his GPUs" Your right. Why? Becuase after the cpu the water goes to the DUAL Rad before it gets to the GPUS. How much heat comes out of the CPU was NEVER in question. It was always going to the RAD before the GPU's.

The question is, how much heat gets transfered into the water from the 2 GPU's as the water then tries to cool the CPU after all the energy from the 2 GPU's are in the water. What will effiecently cool the CPU...that water.. or water that has had the the chance to have the GPU energy taken out of the water from the single rad??

And your right again, it would be 2 feet of hose, but i was under the assumption he was going to use the single rad... SO, what i meant was, with my run it would be an extra 6 inches if he used my run, vs the run of running the single rad after the dual rad. It is 2 extra feet if you dont use the single rad at all.. but i was under the assumption it would be used.

And my real world tests showed me the hose had little effect on temps, where as moving the Rad did. Those are my tests...and i hope my tests have helped the OP make a decision.

Enjoy the water cooling!
Click to expand...


Dude, I didn't link the article because of the "1 CPU + 1 Pump" example they have posted. I made no reference to the example at all, except for using their LPH estimate. The point of the article is to back up my calculations.

As for the water flowing from the GPUS to the CPU, I was working under the assumption that the whole system was setup in the opposite manner. namely pump->120.1->120.2->cpu->GPUs->pump, as in the picture in Punx's post. In any case, if the order were reversed, you get more or less the same results- the water dT isn't great enough to warrant running tubing all over your case. IF the OP were able to move the 120.1 to the back of his case, and the extra tubing were minimal, it would be technically superior to what I've been talking about, if only by making his CPU a fraction of a degree cooler. If, however, he HAS to mount the 120.1 at the front of the case as he says he does, daisy-chaining it and the 120.2 directly will give you the same or better results as stringing hose all over the case, and will look better to boot.
 
And i disagree with you. I think 2 OC'ed 7800 gtx's have increased the energy of the water enough (dT) to warrant putting the Rad there. My tests tell me it does.

running Folding at 100% for 8 hours.. moving the rad too inbetween the GPU and CPU dropped the temp from around 49 C, to 44 C on the CPU. Moved the rad back to next to the other rad, temps back to 49 C.

Those are my results. The OP can take it for what its worth.
 
bradyapba said:
And i disagree with you. I think 2 OC'ed 7800 gtx's have increased the energy of the water enough (dT) to warrant putting the Rad there. My tests tell me it does.

running Folding at 100% for 8 hours.. moving the rad too inbetween the GPU and CPU dropped the temp from around 49 C, to 44 C on the CPU. Moved the rad back to next to the other rad, temps back to 49 C.

Those are my results. The OP can take it for what its worth.
Click to expand...

You are 100% full of crap. Results are worthless. If you had the most poor of flow rates, you might experience the difference you claim.
How many watts do you think 2 gtx's put out while running folding at home? Remember now, volts and clock rate are dropped during 2d.

Also, you keep talking about only adding 6 inches of tubing. Don't forget the rad will add restriction also.

I have a opty clocked at 3ghz, 2 7800gt's overclocked and volt modded. All on one dual rad with max temps reaching around 40c for each component. Ambient temp around 22c.

Not trying to be a jerk, just wanting to see decent info hear...
 
a simple typo... i was typing fast.. when i post during the day i am at work, so trying to work and post during breaks and lunches....

it should read folding at home, and 3d mark 03. I normally run those 2 to test.. simply missed it when i posted.. just in a rush.

And my results are what they are. Believe or not believe. Maybe there is another factor not relating to the position of where i was putting the rad that came into question that made the temp difference. All i know is readings i got.
 
bradyapba said:
...as for comparision purposes, my friend and I have almost the same set-up, difference is he has 2 single 120 rads....in basically the same setup i have described about except replace the dual with a single....

I have a dual 120.. so we both have 2X120.. i just have a dual, he has 2 singles... but he put one inbetween his GPU and CPU.. we have the same GPU and CPU... and hes getting 4-5C better temps on his CPU becuase of the placements.
Click to expand...
OR because of better block mounting, better TIM interface, better IHS/core contact....etc.,etc., etc.
Maybe the rad placement is a contributing factor, maybe it's not, but your anecdotal evidence does not prove it either way.

Now, let's talk about your method of gathering temp readings....
 
I do not think adding the second rad will restrict things too much. The MCP655 is plenty beefy to push enough water through that.

I do agree that increasing the overal length of tubing will have some sort of negative effect on the system. Where I do not agree is that the extra rad is not going to help that much. I am thinking that it definitely will.

Would love to see some hard facts one way or the other.
 
Also, I just measured and figured out that I actually might have room for a single 120.3 Radiator instead of a 120.2 and a PA160.

If you were considering this scenario, what would you do?

Keep in mind that I already have an HE120.2 and would be adding the PA160 or selling the HE120.2 and getting a 120.3 instead.
 
You must log in or register to reply here.
Back
Top