Separate Water Cooling Loop?

Joined
Jul 6, 2016
Messages
57
Is there a reason why i dont see separate Custom water loops? Cpu/GPU with their own radiator. Especially on Bigger cases. Just Wondering
 
Because there is basically no advantage to doing it that way. You don't always have loads that go to 100% on both components, so having each component on their own separate loop wastes the cooling potential of the other loop that might not be loaded to 100%. In addition, dual pumps in a single loop offer redundancy while a pump for each separate loop has no redundancy. Literally the only reason for having separate loops is for bragging rights with major overkill.
 
Because there is basically no advantage to doing it that way. You don't always have loads that go to 100% on both components, so having each component on their own separate loop wastes the cooling potential of the other loop that might not be loaded to 100%. In addition, dual pumps in a single loop offer redundancy while a pump for each separate loop has no redundancy. Literally the only reason for having separate loops is for bragging rights with major overkill.

Separate loops also require you to double up on everything. You absolutely need two pumps to do that and although not totally needed, a second res would be advisable. All of this adds to the costs in both the aforementioned hardware and the additional fittings. Such a setup also takes up more space, etc.
 
Because there is basically no advantage to doing it that way. You don't always have loads that go to 100% on both components, so having each component on their own separate loop wastes the cooling potential of the other loop that might not be loaded to 100%. In addition, dual pumps in a single loop offer redundancy while a pump for each separate loop has no redundancy. Literally the only reason for having separate loops is for bragging rights with major overkill.

OP you might read this and think: but the water in the loop is hotter because of the other component; however the water is much cooler than either component (if your doing it right) so heat is still transferring which is all we care about.

Think of this hypothetical: you have 2 rads than can dissipate 200w of heat each. A CPU putting out 20w idle/110w under load, and a GPU putting out 10w idle/220w under load. Running a GPU intensive game, this imaginary GPU rad if in it's own loop could be unable to keep up, while your CPU loop would be able to handle another 90w even if it was fully loaded. Of course in reality it's unlikely that both component are fully loaded at one time so it probably be even more out of kilter than that.

With adequate radiator(s) to transfer the heat of your components your better off pooling them in a single loop, there would be no benefit to cooling one excessively well while another is just eking by.
 
The extra expense of a second pump isn't worth the single digit temperature drop TBH. More points of failure, more expense, for a degree or two 'c
 
I can only see two loops being required if you are cooling TECs hot side on one loop and plumbing the chilled water in the other.
 
OP you might read this and think: but the water in the loop is hotter because of the other component; however the water is much cooler than either component (if your doing it right) so heat is still transferring which is all we care about.

Think of this hypothetical: you have 2 rads than can dissipate 200w of heat each. A CPU putting out 20w idle/110w under load, and a GPU putting out 10w idle/220w under load. Running a GPU intensive game, this imaginary GPU rad if in it's own loop could be unable to keep up, while your CPU loop would be able to handle another 90w even if it was fully loaded. Of course in reality it's unlikely that both component are fully loaded at one time so it probably be even more out of kilter than that.

With adequate radiator(s) to transfer the heat of your components your better off pooling them in a single loop, there would be no benefit to cooling one excessively well while another is just eking by.

Yep, I've done the math on this. It's pretty straightforward physics after all.

At 1 GPM (minimum ideal flow rate), a 200 watt component will raise the temperature of the water going through the block by a whopping 0.76 C. A 300 watt component will raise it by 1.14 C. Even if you have a high restriction loop with flow rates down to 0.5 GPM, a 200 watt component will raise it by 1.5 C and 300 watts at 2.3 C. Unless you're at your absolute thermal limit, 2 C isn't going to affect anything.
 
Since day 1 of watercooling (late 90s for me) ive always seperated my cpu and gpu(s) to their own isolated loop. Last time i had cpu/gpu in same loop was fx57 and 6800ultra and temps didnt make me happy.
 
Back
Top