Watercooled heatpipes (Give opinons)

[Mond]

hey guys,
i'm currently planning to make a watercooled heatpipe waterblock.
the block consists of nine 14mm round heatpipes (three by three) solder to a 46mm copper block, with 1mm spacing in between the blocks. The copper block serves as a frame that holds the heatpipes together. A coke can will be cutted into half and solder on top of the heatpipe block, cooled by a liquid system. Capable of removing 900W+ per second of heat, and can be used with a TEC.
Basiclly i'm cooling the heatpipes with liquid instead of air. Additonal copper fins will be soldered on top of the block.
Give me some of our opinon.
Please donate or sponsor me a liquid cooling system.

Thanks

[Andrew]

Sounds like an interesting idea. I have been wondering what sort of temps your could get with one of the new HSF modded to use water.

Make sure you post your results.

[Mond]

My block will be done in like a month or two (currently trying to bid a exos liquid sytem on ebay)

[MaxxxRacer]

wont work to well

Having the water as close to the cpu as possible (impingement) is more effecient than heatpipes..

so with that said, the heat pipe system will never be as good of a cooling solution as a regular waterblock.

[Mond]

I decided to solder a plastic bottle on the heatpipe block, so this way i can see what's going on inside and to show off. But would the plastic bottle melt? i'm going to use "solder it" works with a heat of a match.

[Mond]

Dear Maxxxracer, heatpipes transfer heat 1XX time faster than copper.So logiclly it should work better than a normal copper waterblock right?

[Mond]

The heatpipes have direct contact with the cpu, the copper blocks justs holds it in place

[MaxxxRacer]

I recant my previous statement. Did some research and apparently they can have up to several thousand times greater thermal conductivity than copper.

Well if you are serious about this, the best way to handle it is to get the largest surface area on the part of the heat pipe that will be touching the water and to have it flat so that as much of the gas will be as close to the copper tubing walls as possible.

If you are able to manage these two things and have a good flow patter through the block, then theoretically it should work very well.

BUT i still have my reservations about it working better than conventional waterblocks.

[Mond]

Dear Maxxxracer
should i add fins to the heatpipes to increase the surface area? like normal heatisnks, but the fins are watercooled

[Thomas FJ]

Sounds cool!

Regarding the block design I would go with the Apogee design for a large surface.
/Thomas

[Mond]

Dear Thomas FJ, how could i apply the apogee design onto my heatpipe block? do you mean apply it on the base?or the heatpipes? please be specific

[Mond]

Dear Maxxxracer, how long in your opinon should my heatpipes be? because if it's too long the block will be unstable and it also might break the motherboard

[Thomas FJ]

Dear Thomas FJ, how could i apply the apogee design onto my heatpipe block? do you mean apply it on the base?or the heatpipes? please be specific

Please see my suggestion below
/Thomas

[Holst]

Its a nice idea but I think it will be worse than conventional watercooling.

The problem with PC watercooling is that you have a very dense heat source.

We use copper to conduct (spread) this heat to a cooling medium (water, or in your case heatpipe)

If you could attach a heatpipe right onto the core... then you might have something clever. But as you are stuck using copper as an initial heatspreader I dont think you will gain any performance at all.
In fact.. I think your just adding another step in the cooling equation and your temps will be WORSE than conventional watercooling.

But that said, nice idea and its definatley worth a try.

[Bloody_Sorcerer]

the ideal solution would be a "heatpipe" baseplate on waterblocks, instead of a copper one; basically, a IHS or CPU-die sized heatpipe in the base of the waterblock instead of solid copper; only problem is, existing baseplates are so damn thin that it'd be really tough; a much better application would be heatpipe coldplates/hotplates for TECs.

[MaxxxRacer]

I would have the heat pipes in a U shape. The vertical portion of hte U to be cooled by the water. the pipes them selves would need SHORT fins on them to increase surface area. I would say the wetted part of the pipe should be about 1 inch tall to ensure that the block is not putting too much torsional forces on the mobo and cpu. Due to water being incredibly more effecient than air, the part that is cooled by the water needs to be alot less than that of air coolers of the same design (ie. Hyper6 )

[mnewxcv]

subscribed

[Gimmpy224]

maxxs does that mean that this type of block could be better than the storm?

[Mond]

Could it possibly beat the Swiftech Apogee waterblock or the Swiftech Storm waterblock?

[MaxxxRacer]

theoretically I believe it is possible, but as I mentioned before I still have my reservations as to how well this will really work.

I think that with gigantic die sizes like the Itanic and maybe a 90nm quad core, this type of design would have the capability of beating the storm, but the current cores just arnt that large and do best with targeted cooling when talking about watercooling.

[Mond]

Here are my design pics
sorry its all in chinese, ask me any questions if you don't understand anything.

[.sentinel]

I don't think you could use a coke can.....

[Mond]

Both the copper block and the heatpipes will be gold plated

[orkan]

Build and seal a resevoir around a heatpipe heatsink's fin's. The water will pass over the fins... hence cooling the heatpipes.

The PERFECT heatpipe hsf for this would be ones you find with Shuttle XPC systems. They have a flat base, low profile, and swing up to allow a good resevoir seal around the fins.



This way, all the soldering work has been done. All you have to do is figure a way to pass water through the fins. In theory... should cool pretty damn well! The back of the many XPC's that I've built... get DAMN hot under load. So if you had a water loop to remove more heat... perhaps the heatpipes could run more efficiently.

[Serra]

There is really one quick way to solve the question of whether a watercooled heatpipe could be better than watercooling;

Someone with a sub-zero or below-ambient cooling system needs to hook it up the the tip of the heatpipe. Not too close to the block (as the liquid would never vapourize), but just far enough away to instantly cool down all the vapour as close to the die as possible. This would show us the maximum efficiency of a heatpipe. If this was able to cool the CPU better than watercooling, then the use of watercooling on a heatpipe could be eventually shown to be superior (once its well refined). If, on the other hand, it didn't make a terrific difference or was still not as effective as watercooling, then you have your answer.

Someone here should be able to do that within the week, so I look forward to the results (and some credit for the idea of the test).

~ Serra