Modified Heat Pipes

[janas19]

Hey folks.

I have a question about heat pipes. I know that retail CPU cooler heat pipes work by evaporative cooling. The heat of vaporization transfers the heat of the processor to the water molecules, which break away from the liquid and carry the heat as a vapor. When the cool off in the top part of the heatsink, they condense back to liquid and are pulled back to the heat source via the capillary action of a wick. Heat pipes are passive of course. And they are able to achieve cooling from 60-70 degrees C using water.

I understand that part. My question is theoretical. I want to gather information on the possibility of building a modified heat pipe. So this is in the early stage.

If you used a different working liquid, one with a lower heat of vaporization, this would offer more cooling, but since it would evaporate quicker, you would have to use more volume of that liquid than the water. My question is, how would that greater volume affect the pressure? It seems like it would make the pressure greater. Therefore, would using a non-water working liquid in a passive heat pipe be an improvement? Would this change in liquid necessitate a greater vacuum inside?

Appreciate any help you guys can offer. Thanks

 

[Automata]

Heatpipes do not use water, because it wouldn't start working until 100c. They already use a liquid that has a lower boiling temperature.

Adding more will also not increase cooling capacity. Given enough pressure, it can still be liquid past the boiling point. That is because "boiling point" normally assumes atmospheric pressure. If it stays liquid, the capillary action can't happen and cooling will cease. The same works in reverse. You can boil water at room temperature.

These companies have a ton of money to throw into research. I'm sure the heatpipes that are produced perform optimally. They don't just guess.

 

[ssjwizard]

Heatpipes used in current heatsinks use a compressed low molecular weight gas that boils below zero. When compressed they maintain a liquid state up to whatever temp the mfg has designed it for. Adding more coolant would raise the pressure inside of the heatpipe and therefor raise the minimum working temperature of the heatsink.

 

[Bobnova]

Heatpipes actually do use water, very pure water.
The key is that inside a heatpipe is only water. There is liquid water and water vapor and nothing else.
This means that the water is always at it's boiling point (lower pressure, lower boiling point. Higher pressure, higher boiling point).

When you apply heat to a heatpipe that area is hotter, being hotter than the whole the water there boils, this absorbs heat and raises the pressure and puts the cooler parts of the heatpipe at well below boiling temperature, causing condensation. Condensation releases heat and results in liquid water. That liquid water then travels back through whatever wicking mechanism is inside the heatpipe to evenly distribute itself (surface tension does this), and the cycle continues.

Use of different fluids will make the heatpipe more efficient in different temperature zones, water works really well from around 30c to around 100c. Above 100c efficiency drops and below 30c efficiency drops. Under 0c the heatpipe shuts down almost entirely as the water freezes, this can also crack the heatpipe letting air in and destroying it.

The only fluid that has a chance of beating water at the temps we need them to work at is ammonia, and it isn't that much more efficient and requires an awful lot more taxes to be paid for using hazardous materials, plus every time some doofus cut open a heatpipe the company would be at risk of a lawsuit due to that person dying.

Wikipedia has an excellent heatpipe page for further information.

EDIT:

Thou shalt not mess with thine heatpipes. The manufacturer already uses an optimal amount of water in 'em, and they are at a vacuum far purer than one you're likely to be able to make or maintain.

 

[Automata]

I stand corrected, they do use water, but at a vacuum.

http://en.wikipedia.org/wiki/Heatpipe#Structure.2C_design_and_construction

Examples of such fluids include water, ethanol, acetone, sodium, or mercury.

 

[ssjwizard]

Humm my knowledge of heatpipes is obviously outdated. Thanks for the linkage.

 

[Bobnova]

They're absolutely fascinating, the mercury ones are staggeringly efficient in the right temp range. Just don't break 'em

 

[Diggrr]

Wasn't Joe (from the front page here) experimenting with a heatpipe coldplate that used freon as the working fluid?
It was years ago, but him and a friend were working towards a commercial version.

Yeah, they balance the vacuum to make the water boil at the target cooling point.
It would seem easier to use pressure with a freon fluid to achieve the same.

 

[Bobnova]

I expect it is easier, I know some people have made home made ones with R134a. I expect the manufacturers use water because it's cheap, and in manufacturing you can't simply vent r134a to get rid of the air in the tube, you have to pull a hard vacuum in it. Once you have the hard vacuum water isn't much, if any, harder to get to work than r134a is. I think water has a higher specific heat anyway when you get down to it.

I've been meaning to build one just for giggles, haven't gotten around to it.

Now the plus side to r134a is that you could use it in sub-freezing temps without it freezing up on you, very useful for benchers who live in the frozen north!

 

[Diggrr]

Propane, Butane and Methane are also refrigerants, and can be handled without a license.
Theoretically, of course...

 

[janas19]

Ok Bobnova, I fully understand and appreciate the advice not to mess with the heat pipes. However I'm not a noob exactly. I intend to do my homeeork before making any pipes.

A guy named Jim made his own heat pipes on Benchtest.com. This is the only evidence I have found on the internet of someone building their own passive (no vapor compressor) copper water heat pipes. In fact, his pipes went through 2 revisions. All things considered, I think he did a pretty good job. He didn't fabricate a wick structure on the inner surface of the pipe (I imagine this would present some technical difficulties for us amateurs), rather he relied on gravity. I think in the final revision, he was able to achieve slightly better temps than a stock cooler.

You mention ammonia though, and heat of vaporization. Any particular reason why ammonia works better here, for the temps we are shooting for?

(I've also eyed chloroform as a potential working liquid)

 

[Bobnova]

Ammonia has a very high working heat, and a freezing point much colder than water.
For a normal heatsink in room temp conditions it probably wouldn't be much, if any, better. It would however retain usefulness at a much lower temp. For extreme benchers, working in colder temps is a plus. For everybody else it isn't relevant. I wouldn't mess with ammonia, personally. That stuff is nasty, really nasty.

I've been wanting to make some copper pipe+water heatpipes for a long time, if you attempt such a thing please keep us informed!

 

[janas19]

Thanks so much for the info, I appreciate it. Like I said, I'm only in the information-gathering stage atm. If I move to the next stage I will definitely keep you guys updated

 

[ivanlabrie]

I remember reading a guys site where he posted his homebrew heatpipes for a fanless computer and fanless PSU.
I'll try to recall the site and post it here. It was really impressive!

 

[ivanlabrie]

http://www.silentmods.com/section2/item247/
Gotcha! It even feature modified flexible heatpipes!!!
Check it out...

 

[janas19]

http://www.silentmods.com/section2/item247/
Gotcha! It even feature modified flexible heatpipes!!!
Check it out...

Nice find!

Definitely a rough prototype to be sure. Using a coin on one end is a little ghetto lol. But the cool thing about what this guy did is he demonstrated (sort of) how to create vacuum in the pipe WITHOUT incorporating/buying a valve. Sure, pinching a tubule sounds easy, but not everyone knows that method.

Just my thoughts...

 

[ivanlabrie]

Yeah...I really liked some of the stuff he posted there. I dig that site and some of the stuff at SPCR also.
I wouldn't try to build heatpipes myself, BUT i liked his idea of buying heatpipes from that Zalman hdd cooling bundle -He mentions that in the fanless heatpipe cooled psu project log-