Overclock.net banner

1 - 14 of 14 Posts

·
Registered
Joined
·
156 Posts
Discussion Starter #1
What is the best loop direction? I tried res-pump-rad-cpu and currently on res-pump-cpu-rad, +/- 1-3 degrees diff. in temps.

Apogee XT
Jingway DP-1200 aka DD CPX-Pro
Swiftech MCR-220 QP stack
Swiftech Micro Res
2x GT AP-15 push/2x slipstream 2000rpm pull
Primoflex PRO LRT 1/2 ID 3/4 OD tubing
LL
LL
LL
 

·
Super Moderator
Joined
·
19,298 Posts
As long as the reservoir is before the pump at least for filling/refilling, it doesn't matter too much how you do it.
 

·
Registered
Joined
·
829 Posts
I just keep my rad right before the CPU so that the CPU gets nice, cool water. Anything other than that shouldn't impact temps.
 

·
Premium Member
Joined
·
5,945 Posts
Quote:

Originally Posted by Arrec Barrwin View Post
I just keep my rad right before the CPU so that the CPU gets nice, cool water. Anything other than that shouldn't impact temps.
Why do people think loop order matters? It really doesn't. If you put the rad right after the CPU instead of before it, the only extra heat going into the CPU is the heat from the pump. I calculated this in another thread:

Quote:

Originally Posted by charliehorse55
If you have the pump going to the CPU than the heat from the pump is going into the CPU. This is not really a problem through, let's investigate a worst case scenario:

At 1 GPM, (You would usually have more)
50w pump (most pumps use 10-25w)

0.063 L/second of water go through the pump
1 L of water = 1 kg.

0.063kg/second
50w for 1 second = 50j of energy.

The heat capacity of water is 4185j per kg/C. This means it takes 4185j of energy to raise 1kg of water 1 degree C.

So,
4185*0.063 = 263j of energy to heat the amount of water passing through the pump in 1 second by 1 degree.

50/263 = 0.19, so having

Pump>CPU>Rad>Res>Pump instead of Pump > Rad > CPU > Res > Pump would yield a CPU temperature only 0.19C higher.
 

·
Premium Member
Joined
·
8,195 Posts
I bet that the 1-3° difference can be credited to either ambient temp changes, TIM/block application, or a combination of both.
 

·
Registered
Joined
·
214 Posts
Quote:

Originally Posted by charliehorse55 View Post
Why do people think loop order matters? It really doesn't. If you put the rad right after the CPU instead of before it, the only extra heat going into the CPU is the heat from the pump. I calculated this in another thread:
This is very interesting, thanks for showing your calcs... but I am not sure I fully understand (I've yet to take college physics)...

Does this mean that it wouldn't matter if I set up a loop like so:
res>pump>cpu>gpu>rad
or
res>pump>gpu>cpu>rad
or
res>pump>gpu>rad>cpu

All of these would have essentially the same performance? my logic tells me it would be better to cool the water between cpu and gpu, but will it not make much diff?

I apologize if this is a noob question haha.
 

·
Registered
Joined
·
3,085 Posts
after a few minutes of running, the temperature of the water will even out and stay the same, that's the simple way to put it.
 

·
Premium Member
Joined
·
5,945 Posts
Quote:

Originally Posted by Dreamlane View Post
This is very interesting, thanks for showing your calcs... but I am not sure I fully understand (I've yet to take college physics)...

Does this mean that it wouldn't matter if I set up a loop like so:
res>pump>cpu>gpu>rad
or
res>pump>gpu>cpu>rad
or
res>pump>gpu>rad>cpu

All of these would have essentially the same performance? my logic tells me it would be better to cool the water between cpu and gpu, but will it not make much diff?

I apologize if this is a noob question haha.
Nope. Loop order should depend on two things:

1) The reservoir should gravity feed the pump
2) Use a minimal amount of tubing

Also, I've yet to take college physics either. I'm 16 and still in HS.

Basically, water has such a high capacity for heat that loop order does not matter.
 

·
Village Idiot
Joined
·
2,393 Posts
The way you have it now is fine. There have been two threads in the past week that I'm aware of dealing with order. HERE is the big one. Basically it doesn't matter since it's a closed loop.

Side note, I'm happy to see more people using the Jingway pumps.

Quote:

Originally Posted by Dreamlane View Post
This is very interesting, thanks for showing your calcs... but I am not sure I fully understand (I've yet to take college physics)...
Maybe you actually learn something at real college, but my community college physics class didn't touch on anything with fluid/heat
 

·
Premium Member
Joined
·
5,945 Posts
Quote:

Originally Posted by pewpewlazer View Post
The way you have it now is fine. There have been two threads in the past week that I'm aware of dealing with order. HERE is the big one. Basically it doesn't matter since it's a closed loop.

Side note, I'm happy to see more people using the Jingway pumps.

Maybe you actually learn something at real college, but my community college physics class didn't touch on anything with fluid/heat

All of my calculations (on this thread and others) are done with a couple of basic formulas:

Heat Capacity = Mass * specific heat capacity for substance

Heat capacity is how much energy a substance can store. That is, how much energy is takes to get a certain mass of the substance by 1C.

Plug in water's heat capacity:
Heat Capacity in Joules = mass of water * 4187

As 1 L of water weighs 1 kg, we can simplify the equation to:

Heat Capacity in Joules = volume of water * 4187

What is a joule? A joule is 1 Watt of power for 1 second. So how do we use these formulas to calculate the temperature change in a loop?

First, you're going to need to flow rate of the loop. Let's assume 2 GPM. However, you're going to need to flow rate in a different unit of measurement, liters per second. To calculate this, multiply by 0.063.

2*0.063 = 0.126 Liters per second

Now, let's plug that into our heat capacity equation:

Heat Capacity in Joules = 0.126 * 4187
Heat Capacity = 527.6 Joules

Since that is the amount of water that flows through a block in a second, and a joule is a watt per second, we can cancel out the unit of time.

Therefore, for every 527.6 W that are dumped into a loop, the temperature of the water will raise by 1C.

Here is a formula if you don't feel like doing all that math.


Flow rate in GPM*263.8 = number of watts required to heat the water by 1C.
 

·
Registered
Joined
·
2,727 Posts
Quote:

Originally Posted by charliehorse55 View Post
Why do people think loop order matters? It really doesn't. If you put the rad right after the CPU instead of before it, the only extra heat going into the CPU is the heat from the pump. I calculated this in another thread:
Yup the water goes through the loop fast enough that it really doesn't matter.^
 

·
Village Idiot
Joined
·
2,393 Posts
Quote:

Originally Posted by charliehorse55 View Post
All of my calculations (on this thread and others) are done with a couple of basic formulas:

Heat Capacity = Mass * specific heat capacity for substance

Heat capacity is how much energy a substance can store. That is, how much energy is takes to get a certain mass of the substance by 1C.

Plug in water's heat capacity:
Heat Capacity in Joules = mass of water * 4187

As 1 L of water weighs 1 kg, we can simplify the equation to:

Heat Capacity in Joules = volume of water * 4187

What is a joule? A joule is 1 Watt of power for 1 second. So how do we use these formulas to calculate the temperature change in a loop?

First, you're going to need to flow rate of the loop. Let's assume 2 GPM. However, you're going to need to flow rate in a different unit of measurement, liters per second. To calculate this, multiply by 0.063.

2*0.063 = 0.126 Liters per second

Now, let's plug that into our heat capacity equation:

Heat Capacity in Joules = 0.126 * 4187
Heat Capacity = 527.6 Joules

Since that is the amount of water that flows through a block in a second, and a joule is a watt per second, we can cancel out the unit of time.

Therefore, for every 527.6 W that are dumped into a loop, the temperature of the water will raise by 1C.

Here is a formula if you don't feel like doing all that math.


Flow rate in GPM*263.8 = number of watts required to heat the water by 1C.
I understand it, it isn't rocket science. Just saying I didn't learn any of that in physics class. Not like I'm super excited for Fluid & Thermal Dynamics though...
 

·
Registered
Joined
·
156 Posts
Discussion Starter #14
Quote:

Originally Posted by jakeface1 View Post
What are your load temps?
20 runs each of linx and IBT, 71 degrees max., used to get 83 degrees max. on air(prolimatech megahalem w/GT AP-15's in push/pull) on the same tests. Is this the usual temp. improvement between air and water cooling?
 
1 - 14 of 14 Posts
Top