• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Balancing a water flow in a parallel loop

Joined
May 24, 2023
Messages
1,204 (1.65/day)
How match a water flow in a parallel loop to a heat output of a CPU and GPU? The answer is simple - with a valve!

 
The loop is done:

5080 WF WB parall.jpg
5080 WF WB parall 2.jpg

I closed the valve so much that the CPU runs at 70°C with 100% Cinebench load, during normal use it is much cooler.

There is one downside of this arrangement though: If I used pump regulation between say 20 - 70% and regulated it with water temperature, and then hit the PC with some intensive CPU load, it would start overheating because of the low water flow in the CPU branch.

So at the moment I have the pump at fixed 50%. and everything runs fine.

If I had conventional serial connection of the components and all water would be flowing through the CPU block, I could use the pump regulation with low flow without any problem.
 
Last edited:
This is actually a rabbithole I don't want to dive in, but...

As far as I know, most waterblocks (CPU & GPU) have an optimum efficiency with approximately 100 ~ 120 l/h coolant flow.
In series, you just have to pump 100 ~120 l/h around for both blocks, parallel you have to give each block 100 l/h so you have to pump 200 ~240 l/h around to have a comparable efficiency.

But if you like the aestetics of a parallel loop, nobody is stopping you. Instead of a valve you can also use a orifice or an inline filter to create the needed flow restriction on the CPU block.
 
As far as I know, most waterblocks (CPU & GPU) have an optimum efficiency with approximately 100 ~ 120 l/h coolant flow.
In series, you just have to pump 100 ~120 l/h around for both blocks, parallel you have to give each block 100 l/h so you have to pump 200 ~240 l/h around to have a comparable efficiency.

I understand that a quick flow can enhance turbulence and heat flow from the block in the water, but on the other hand, it seems to be cooling just fine even with the slower flow. The only downside I see is what I mentioned, it limits variability of the pump regulation.

EDIT: Note, that a parallel combination of both blocks has lower water flow restriction than just the GPU block alone.

But if you like the aestetics of a parallel loop, nobody is stopping you. Instead of a valve you can also use a orifice or an inline filter to create the needed flow restriction on the CPU block.
I have a solid metal door and I do not need to see the (ugly) pink hose spaghetti.

The flow restriction needs to be variable.

In some design high end loop this enables very nice and simple connection and minimal tubing in the motherboard area.
 
Last edited:
Back
Top