mavity fed/assisted water cooling for wall mounted PC?

[ste91]

If you had a wall mounted PC (or just even an external watercooling system) is it possible to have it running off a large reservoir right at the top, with the water running down through the system into a lower reservoir where the pump just returns the water from the bottom to the top?

Would that work? The idea is that the pump can be run slower (quieter) and the water will naturally want to run down through the system, the pump is not really doing anything other than moving the water from one reservoir to another.

 

[Akagi]

I want to say yes, but in truth im not sure how fast water moves in a traditional water-cooler, or if a mavity fed solution would move the water fast enough for effective dissipation of heat.

 

[d_brennen]

Presumably the flow rate would be similar, what's the point? If it's mavity fed you may well have to run the pump faster.

I'm all for oddball solutions though...

 

[BluD]

Yeah i think you would just put more strain on the pump for no gains. Having a header tank mavity fed is normally so you can get more water pressure than is been supplied but as the speed of the water make very little difference in water cooling it would'nt benefit you really.

 

[Brizzles]

mavity would not provide enough pressure to push the fluid through blocks.

You are using a pump already, albeit in this case trying to pull water through your loop.

Once the loop is closed and free of air bubbles a pump can be run very slowly anyway.

 

[ste91]

mavity would not provide enough pressure to push the fluid through blocks.

You are using a pump already, albeit in this case trying to pull water through your loop.

Once the loop is closed and free of air bubbles a pump can be run very slowly anyway.

That's what I was concerned about, you could use larger diameter pipes to help the flow but you're still bottlenecked by the cooling blocks and radiator.



Also another question while I'm here - is there any benefit in having a "split" loop so that the CPU and GPU have a separate flow of water? It doesn't seem right to me that in a sequential loop the water passes through one heat source and then another, the water is already heated by the time to gets to the 2nd block, so surely the heat transfer at that point is less efficient? I know people say the temperature will equalise anyway and almost all cooling systems use that setup and work fine, but would it not be more efficient to have the flow of water split before the blocks, and then re connected (via Y or T connectors) after the blocks so they share the same pump/rad but get their own equally cooled flow of water, rather than the 2nd block getting water that's already had heat transferred to it?

Or even having an extra radiator in a normal loop i.e. rad-CPU-rad-GPU-res-pump rather than rad-CPU-GPU-res-pump? I'd have thought more components and restrictions i.e. 90 degree bends put more strain on the pump though.

 

[Brizzles]

That's what I was concerned about, you could use larger diameter pipes to help the flow but you're still bottlenecked by the cooling blocks and radiator.



Also another question while I'm here - is there any benefit in having a "split" loop so that the CPU and GPU have a separate flow of water? It doesn't seem right to me that in a sequential loop the water passes through one heat source and then another, the water is already heated by the time to gets to the 2nd block, so surely the heat transfer at that point is less efficient? I know people say the temperature will equalise anyway and almost all cooling systems use that setup and work fine, but would it not be more efficient to have the flow of water split before the blocks, and then re connected (via Y or T connectors) after the blocks so they share the same pump/rad but get their own equally cooled flow of water, rather than the 2nd block getting water that's already had heat transferred to it?

Or even having an extra radiator in a normal loop i.e. rad-CPU-rad-GPU-res-pump rather than rad-CPU-GPU-res-pump? I'd have thought more components and restrictions i.e. 90 degree bends put more strain on the pump though.

No benefit from a split loop as you described - the liquid will always reach equilbrium and remain at a certain temperature. The transfer of heat from block to block would be imeasurable.
2 loops, 1 for cpu and 1 for gpu would be the only way to provide better cooling

 

[Smffy]

My recommendation would be to just do it the way everyone else does to be honest... in a small loop the pump can be run literally 20-35% most of the time so you're trying to solve a problem that isn't there. A decent D5 is not audible at even 50-60% usually.

And as mentioned dual loop is normally for fairly extreme circumstances or because it looks nice with different colours. Reality is the temperature reaches and equilibrium pretty rapidly so it really makes no odds what order you run your components.

 

[Uncle Petey]

Yeah, I think you're overthinking it tbh. Most of us do when it comes to our first loop
Pump noise isn't really a thing these days. Resonance/vibration is usually the thing to watch out for, so just focus on decoupling the pump from it's mount with foam or rubber etc.

 

[cobalt6700]

mavity would not provide enough pressure to push the fluid through blocks.

You are using a pump already, albeit in this case trying to pull water through your loop.

Once the loop is closed and free of air bubbles a pump can be run very slowly anyway.

Eh? A D5 is rated at 3.9M max head, so 0.39 bar (1m of head equals 0.1bar pressure). So if you put a tank 2m above the PC you would defiantly get a flow through your blocks.

Unfortunately the pump would be working just as hard, if not harder, to pump the fluid back from the reservoir (below the PC) to the header tank . By having a closed loop you are likely to have it much quieter as the pump isn't working as hard and you don't have running water plopping into a header tank.