Watercooled PSU's

[Denorios]

Does anybody have any thoughts on when (or if) PSU watercooling is going to mature? The few products on the market at the moment are as ugly as sin , and look like nothing more than back-to-front passive designs with hard-drive type waterblocks mounted where the external radiator fins should be. Obviously this avoids the potentially hair-raising problems of carrying water inside the PSU case, but it can't avoid looking like a jury-rigged chop-job (because it is ), rather than a product designed from the ground up.

Obviously you can't really improve the performance of a PSU by watercooling it (PSU overclocking, anyone? ), and if you are watercooling the rest of the system, then the PSU should be perfectly capable of cooling itself, even if it is a passive design. That being the case, is there really a market for them? Overclockers don't need them, because it won't improve performance, and silencers don't need them because passive or low-speed fan designs can cope quite well if they don't have to exhaust the CPU/GPU waste heat as well.

Are we going to see any well-designed products, or will they just continue to be produced in small numbers by modifying existing passive designs. Any thoughts?

[fingers]

I know of one PSU that seems to have been built from the ground up as being water-cooled; aquaPOWER 250W. Can't seem to find a link but you could try googling. It's about time we saw some decent w/c psu designs the benifits are obvious!!

[Denorios]

Not that obvious. Watercooling CPU's and GPU's can allow you to push their performance to the edge - a little more cooling=a little more speed. But a PSU is either stable or it's not. Improving the cooling of an already stable PSU does not improve its performance.

The problem with passive PSU's is that in a system with little case ventilation, they are overheated very quickly by the waste heat from the CPU and GPU. However, once you have watercooled the other heat sources, especially if using a Reserator, the case temperature ought to be sufficiently low for a passive PSU to cope perfectly well on its own.

[Bitter Jitter]

Well in terms of cooling a CPU Moving from worst to best:

Passive
Fan
Water
Phase Change

PSU are mostly Fan cooled because it is low cost, but if more cooling performance was needed than a fan could provide. Then the next step would be water or phase change.
When you think about it, water is the medium which the heat moves in so rather than have the heat localised around the PSU it could be removed by the water to another location outside the case quickly.

Also what sort of cooling do most cars use? water.

I am not saying we are going to see loads of water cooled PSU's but if 700w - 1000w PSU become common place (I really hope not) then it might be a way to cool them. Hopefully Intel and other manufactures have leant a lesson and will produce less power hungry devices, rather than waste power unnecessary.

On the negative it seems like it would be too expensive to start producing them, a fan is a lot cheaper than water cooling setups.

[fingers]

I am not saying we are going to see loads of water cooled PSU's but if 700w - 1000w PSU become common place (I really hope not) then it might be a way to cool them. Hopefully Intel and other manufactures have leant a lesson and will produce less power hungry devices, rather than waste power unnecessary.

It will certainly be interesting to see what both AMD and Intel will do in the future but I have a feeling power consumption with either decrease through design or by pressure from users finding the costs of running becoming an issue! You just have to look at home appliances to see how low power usage sells!

...if using a Reserator, the case temperature ought to be sufficiently low for a passive PSU to cope perfectly well on its own.

That's a very interesting coment!

My system uses a home-made external cooler like the Reserator (full project log here) which means my case temperatures sit at around 9C above ambient on load (Prime95) and much less under normal use. This is with NO case fans and the only air being moved is by the PSU fans.

I have been unsure about using a passive PSU but by your comment it would seem you would see this as a viable route to take? I'd be interested to hear if you have some experience of this??

Sorry for the off topic...

[Denorios]

No, I haven't actually tried it myself, though I have a friend who is fiddling with it. It is more an application of theory. If the heat produced by the two hottest components - the CPU and GPU - is being completely removed from the case by the reserator, then the heat that remains ought to be enough for a well designed passive PSU to cope with - otherwise, what is the point of a passive PSU? Bearing in mind that actively-cooled PSU's inevitably draw in waste head from the rest of the case, and though heat rises, a passive PSU would probably have to deal with less. I'd at least give it a try, though personally, I would still use an undervolted Nexus Real Silent 120, as a rear exhaust, just in case.

[fingers]

...a passive PSU would probably have to deal with less. I'd at least give it a try, though personally, I would still use an undervolted Nexus Real Silent 120, as a rear exhaust, just in case.

Makes complete sense, once cash is available I think this maybe given a go unless someting else prevents itself, at least this way I could have a case fan run if the case temp rises to high!

[pdf27]

Couple of points:
1) CPU/GPU are very dense heat sources - PSU parts tend to be less so, and hence are more suitable for quiet air cooling.
2) PSUs can already practically use 120m fans, and these are the same size that would be used for radiators/heatercores anyway in most cases. These aren't easy to do with CPU/GPUs, but are a doddle with PSUs - so the noise advantages aren't great.
3) The problems with a PSU getting too hot aren't nearly as bad (or expensive!) as a CPU/GPU.