Why not use heatpipes?

[Thrilla]

My brain is filled with heatpipes these days doing the XS heatpipe thing, I was wondering howcome no one ever made a waterblock with heatpipes to increase surface area thus increase the overall performance?
http://img432.imageshack.us/img432/5311/dsc00092vq2.jpg

[creidiki]

Costs?

I doubt it would be that efficient, heatpipes are pretty good at transferring heat but not so good that the extra cooling area would beta the loss in performance from the extra heat transfer stage.

I could be wrong though.

[mtl_hed]

definatley not effiecent. If you look at all the R&D that goes into WB design, it is focused on getting viscous(inpingment or maze) water flow as close to the die as possible.

Your design would work, but the tradeoff would be higher C/W of the waterblock for more water contact area.

Not work the trouble.


How about a heatpipe cooled TEC? These things are rated for 54W and are 6mm diameter right? So get a 50mm X 50mm TEC and cram 8 U shaped pipes side by side. Would work well on a low TDP CPU.(like the a64 EE's)

[Bloody_Sorcerer]

heatpipes could be positively wonderful for TECs.
Thrilla, how realistic is a cone-shaped (or similar) heatpipe, with a top contact area thats significantly smaller than the bottom? you can guess well enough where this is going ;)

[creidiki]

Ducky! ^_^

How about a heatpipe cooled TEC? These things are rated for 54W and are 6mm diameter right? So get a 50mm X 50mm TEC and cram 8 U shaped pipes side by side. Would work well on a low TDP CPU.(like the a64 EE's)
Have to be extremely careful though, the moment you go over its power you get thermal runaway and the cpu cooks.

[mtl_hed]

Ducky! ^_^


Have to be extremely careful though, the moment you go over its power you get thermal runaway and the cpu cooks.


Yep, would have to be tested on a load tester first. I would not use a real CPU as a guinea pig.

[septim]

with heatpipes and waterblock, it would be bulkier. well for those with smaller
case width this would be problematic, tubing may kink, maybe forced to use
90degree fittings, etc

[Praxis1452]

PCperspective reviewd an air cooled tec... Not great for the price... However an interesting design that does work pretty well.

[mtl_hed]

PCperspective reviewd an air cooled tec... Not great for the price... However an interesting design that does work pretty well.


from what I could see it had an strange(ineffiecient) design. Half of the heatpipes were on the cold site of the TEC the other half were on the hot side, but they were both connected to the same set of sinks.

And they used a crappy LED fan.


With the heatpipes Thrilla is getting, one could make a very usable aircooled TEC.

[Thrilla]

Alright thanks for the input, it makes sense now :P

btw got a link to that thing?

[mtl_hed]

http://www.pcper.com/article.php?aid=273&type=expert

Someone confirm my thoughts on the strange heatpipe design. It seems like this thing would be working against itself. :confused:

[creidiki]

The cooling block looks split in 2.

That dont mean that the pipes hooked up the coldplate make any sense however...

Then again, you have to wonder about aircooling a TEC'd CPU of normal TDP, you need a 226W TEC to be safe so thats an easy 350W of heat (2.5 P4D's worth) you have to dump through a single airsink.

Good luck, is all I'll say... I mean, to have optimal performance for a TEC that size you have to give it a double/triple rad all of its own...

High-performance sink for ultra-low TDP cpus sounds a lot more viable, youd have the same combined heatdump as your avg. OC'ed FX-62, and keep your low-TDP cpu @ ambeint or possibly even below.

Whether its worth it depends on how well thse things OC though.

[Praxis1452]

from what I could see it had an strange(ineffiecient) design. Half of the heatpipes were on the cold site of the TEC the other half were on the hot side, but they were both connected to the same set of sinks.

And they used a crappy LED fan.


With the heatpipes Thrilla is getting, one could make a very usable aircooled TEC.
AS crediki mentioned they need to cool the hot side very well. So they use the cold side.. lol.

However their 25C decison to turn on and off to prevent condensation is bad :-(. $150 for it and they make it so it doesn't go below ambient. Screw that. It's only a 50W tec I think. It cools upto 50W really I think.

However a 35W EE X2 3800+ would work very well with it... only if you managed to just make it run all the time. And then insulated. I would think.

[tw33ter]

What about a waterblock with heatpipes on top instead of the bottom, or would that do any good?

[creidiki]

You mean with a pingrid on bottom/top-plates? Thats brings mainteinance problems - how do you dismantle a block that has heatpipes in a U around it?

I cant think of any cost-viable waterblock that uses heatipes, right now.

But it is 6am here, so I'm hardly in genius-mode ;p

[Thrilla]

That sure is one weird air TEC setup :s
Waterblocks are pretty darn efficient these days, I'm not sure some extra cooling power will make it better or worse.
Maybe have the thin base waterblock in the middle, and on the sides of the base add 2 heatpipes to dump some extra heat, then also have a fan blow on it so it cools the mosfets around it as well.
http://img361.imageshack.us/img361/5265/dsc00094bj0.jpg

[creidiki]

That could work.

Youd want a pretty wide spacing between main block and the airsink part to be able to route tubes without kinking...

[Thrilla]

Or just make 2 seperate air sinks, have the tubes go in the middle like a storm would. But then a fan won't be able to fit without some kind of side mount, lol, seems like a lot of work for that extra 1'C gain if not at all.

[Praxis1452]

How bout heatpipes to do inline nb cooling? Instead of the tubing going all the way down and around for a full 180 you could use the heatpipes to distribute it into a block that kinda goes inline? whatcha think? So nb cooling is easer and less flow restrictve/viable.

[HaxR3]

Pretty good idea

the reasons for that air tec's wierd design are-

1. The Monsoon II active TEC CPU cooler isn’t designed for extreme overclocking, sub-zero temperature operations. The Monsoon II is a high-performance HSF that uses a medium size Peltier device to enhance its overall cooling capabilities. The control module is programmed to maintain a target CPU temperature, which should never drop below 25°C. This eliminates all the problems associated with potential condensation around the cold-plate and CPU socket area.

2. Having heat pipes connected to the cold-plate side of the cooler provides an alternate path for heat to flow out of the CPU and up into the cooling fins. With a light CPU load and relatively cool ambient air temperatures, the Peltier device will not be switched on. The two heat pipes and aluminum fins provide all the cooling necessary to maintain a target CPU temperature. Without the Peltier generating more waste heat, the fan can run at low speed (quiet :)

3. The heat pipes connected to the cold-side also provide a built-in fail safe mechanism for the CPU should something cause the Peltier to fail – no more potential core meltdown.