I posted this in another topic, but figured it'd be very interesting to discuss on its own
 
Always been freaked out about liquid cooling. If it's set up right, it does an AMAZING job. But MAN, it's pricey!! And if something goes wrong, the potential for everything to be fried is huge.
 
But I've been thinking about a system - not sure it'd be cheap, nor risk-free - where the PC is elevated a bit, and a Stirling engine would react to the heat from the processor - somehow - and that engine would then be used to draw the water through the tubes into the cooling system - helped a bit by gravity - and then be pushed back upwards into the system. Perhaps the cool-water side could also be used as the "cool air"-side of the Stirling engine, producing an even greater difference between the hot and cold points. Of course then the system would probably need a small kickstart when booting every time.
 
It seems like a system like this'd take a LOT of tweaking and be potentially VERY big given the large difference of energy/speed that would probably need to be generated to move a constant supply of water like that, pls the faact that it seems it might be a bit delayed because the engine has to react TO the heat of the processor which would be pretty dangerous. But if it could be made to work it seems like it'd be a great, cool-looking semi-mechanical system. Question is whether it'd be cost-effective.


Thoughts?

Considering you would have to engineer​ and build it yourself to even find out if it worked.... I'd say not very cost effective

When I was building my current PC I purchased one of these all-in-one water coolers:
https://www.pccasegear.com/products/15098
But when I tried to fit it, it was proving too awkward. The tubes were very rigid and would not bend sufficiently for it to sit comfortably in the case. In order to fit properly, the tubes would have needed to be bent and twisted to their maximum extent and I considered that to be risky. I ended up returning it and used the stock Intel CPU cooler instead.

In my self-built floortop, I've been running this Zalman water cooler for the last 5 years on a Xeon 1620 - no problems at all.  Very easy to install, the only part I wasn't comfortable with is when tightening down the screws for the part stuck to the CPU.  You sort of can't know when to stop, when enough is enough.
 

davdud101
I posted this in another topic, but figured it'd be very interesting to discuss on its own
 
Always been freaked out about liquid cooling. If it's set up right, it does an AMAZING job. But MAN, it's pricey!! And if something goes wrong, the potential for everything to be fried is huge.

But I've been thinking about a system - not sure it'd be cheap, nor risk-free - where the PC is elevated a bit, and a Stirling engine would react to the heat from the processor - somehow - and that engine would then be used to draw the water through the tubes into the cooling system - helped a bit by gravity - and then be pushed back upwards into the system. Perhaps the cool-water side could also be used as the "cool air"-side of the Stirling engine, producing an even greater difference between the hot and cold points. Of course then the system would probably need a small kickstart when booting every time.
 
It seems like a system like this'd take a LOT of tweaking and be potentially VERY big given the large difference of energy/speed that would probably need to be generated to move a constant supply of water like that, pls the faact that it seems it might be a bit delayed because the engine has to react TO the heat of the processor which would be pretty dangerous. But if it could be made to work it seems like it'd be a great, cool-looking semi-mechanical system. Question is whether it'd be cost-effective.


Thoughts?

 Only reason for liquid cooling is for extreme overclocking.  I don't overclock.

kitekrazy1
 Only reason for liquid cooling is for extreme overclocking.  I don't overclock.

It depends on what level of liquid cooling you're talking about.  The main reason for my small radiator is to avoid having a massive passif heatsink hanging off my motherboard (140w CPU).  Neat and the computer can be knocked around a bit without worry.
 

kitekrazy1
Only reason for liquid cooling is for extreme overclocking.  I don't overclock.

Trudat - nor do I. In fact, my current CPU runs at about half the tempurature of my previous while executing the same tasks WAY faster. I've been able to get my recording latency as low as 4ms recording a full 3 mintues track straight through with no stops, no dropouts or errors, no NOTHING. Haven't tried any lower than that, but at that point there won't be a noticeable difference.
 
But as far as this system goes, is there any merit in attempting to engineer something like this? It'd be costly, but MAN it'd be cool, for science's and engineering's sake.