Phase Change Cooling System

phase cooling

Chris Morrell has an impressive write up on all the ins and outs of building a phase change cooling system. Vapor refrigeration moves heat from one area to another by changing the phase of the working fluid. Chris used propane as the working fluid in his system. He’s got instructions covering all of the work involved from brazing the copper tubes, to building and lapping the evaporator blocks, to the final tuning. With no load it’s can hit -45DegC.

This story reminded me to check back on extremecorvette’s cascading cooler from last fall. He started receiving parts last month for a brand new design. I can’t wait to see how that turns out.

[via Paul Stamatiou]

29 thoughts on “Phase Change Cooling System

  1. The propane won’t ignite unless the system leaks and there’s an open flame or a spark nearby when the leak happens. So, yeah, not the safest thing around, but if with a little common sense there’s not much danger. (Sure, your PC has electronics in it, but if you have stuff sparking in there… well, I honestly don’t know what to tell you. I can’t even imagine how that would happen.)

    On top of that, it’s a closed system (until it leaks, heh), so the supply of propane inside won’t be very large, and once the amount in the pipes is gone that’s all she wrote. And half of what’s in the pipes will be liquid at any given time, as well.

    Really, the possibility of catastrophic failure is quite remote unless you’ve tricked out your machine with a bitchin’ candlelight casemod or something. But even still the article mentions two alternative, non-explody gases to use if you try this yourself.

  2. > the dilemma was that other refrigerants require an
    > HVAC license to use… anyone can get propane though.

    R-134a (aka the “new” automotive refrigerant) does not require a license. Anyone can walk into most automtive part shops and buy an A/C recharge kit complete with can of regrigerant.

  3. [read read] …and he does indeed mention R134a. I’m very surprised he chose propane when safer alternernatives are available.

    Also mentions CO2. Should be noted that while CO2 does well as a refrigerant, it has a bad habit of soldifying under pressuer (eg. forms dry ice).

  4. if i’m not mistaken, using more flammable/dangerous fluids makes it possible to reach lower temps — provided you don’t manage to do something stupid and blow yourself up, of course.

  5. if i’m not mistaken, using more flammable/dangerous fluids makes it possible to reach lower temps — provided you don’t manage to do something stupid and blow yourself up, of course.

  6. In regards to the danger of the propane exploding, hydrocarbons will not ignite unless exposed to oxygen. Within a closed system such as this, there is no oxygen, so there is no risk of explosion. If the system leaked, then yes I would have a risk of explosion, but to combat this, I pressure tested the system to over 400PSI, which is more than double the operating pressure with ambient temperature of 28* Celsius. If you read the entire article, you will see that I never say you must use propane, you can use r-134a, but as I stated, r-134a has a considerably higher boiling point than propane. -45 is indeed below the spec for the processor, however the chip is already missing 2 voltage regulators and I’m pushing 1.75 volts through it, I’m not exactly concerned for the poor things future. One last note on the use of propane and the quantity, I purposely vented all of the propane from the system and ignited it, and I had a 2 foot flame for about 6 seconds and then it dwindled to a little puttering flame. If done properly and with enough common sense to not kick the thing with a torch heating it up, you will be safe, read the article and don’t do anything stupid.

  7. Propane doesn’t just explode. If he had a leak, and for some reason had something sparking inside his case he would probably just see a ball of flame. There isn’t really enough propane in that system to blow up the room it is in or something like that.

    And besides the point of leak testing is to find leaks before you charge the system. Once you have run a vacuum that holds for 12 hours you don’t have a leak.

  8. Here’s a link to a hydrocarbon based refrigerator that’s even approved by Greenpeace:

    http://archive.greenpeace.org/ozone/greenfreeze/

    (Picture: http://web.mit.edu/dusp/etpp/images/greenfreeze.jpg)

    Hydrocarbon refrigeration is very efficient, uses chemicals that aren’t proprietary, are commonly available from multiple sources, and are dirt cheap. I’m not usually eye-to-eye with Greenpeace, but here’s a bit more:

    [hackaday does not trust me with blockquote, bold or italics]

    Hydrocarbons are flammable. However, their flammability can be easily mitigated through adequate safety measures in production and product design. The content of propane or butane in a domestic ‘Greenfreeze’ refrigerator equals roughly the content of two cigarette lighters. The risk of explosion is minimal: it takes between 17 g/cubic meter and 39 g/cubic meter to create an explosive mixture. Therefore, if the refrigerant were to leak outside the refrigerator, an explosion would be nearly impossible.

    [/hackaday does not trust me with blockquote, bold or italics]

    While many parts of the world either were grandfathered in to being able to make freon, or are using more efficient hydrocarbon refrigerants, there’s been a heaping helping of FUD (http://en.wikipedia.org/wiki/FUD) here in the USA against it for some reason.

  9. Furthermore, propane by itself is the only fuel approved for use indoors (think forklifts and floor buffers). The reason for this is that propane has the property of only being flammable within a very narrow range of mixtures with air. Therefore any leak has to have the perfect mixture with air to ignite properly.

    http://www.google.com/search?hl=en&lr=&safe=off&q=2.15+propane+9.60&btnG=Search

    It looks like the linked article guy knows what he is doing. If you know what you are doing propane is safe as refrigerant.

  10. But Chris, it doesnt leak now, I didn’t notice what you used for oil in the system, but even with oil in system, the compressor seals will eventually dry out one day, and you WILL have a leak.

    134a is cheap. You should have went that route. And the r-12 liscense was $15 when i got it in 1997, but yea, r-12 is expensive.

  11. as mentioned above, propane will not ignite if it is outside of its fuel/air ratio. Another time I was building a propane powered pulse jet and my igniter failed and caused a visible puddle of propane to pool around the jet. Propane is one of the more stable and predictable non-regulated gasses that you can use as a refrigerant. Once again, read up and educate yourself, otherwise go buy a vapochill. Read the article and look at how heavily I insulated the board..anything below ambient and your pipes will sweat. read the article

  12. ya,I’d be more worried about it pooling after a leak occures causing board/case damage. I doubt any REAL harm could occure. I personally have been under the hood of a forklift that had pooling vapor when it sparked off.. Even in that large of a volume, under natural air pressure, I only lost some hair. Not even a pink face. Not to mention, if there was a leak, he would notice the smell pretty quick (unless he lives in a dorm with a smelly roommie!)

    The only TRUE danger here is singing his PCB if something catastrophicly sparky were to happen (possible inside the PSU). If he is doing stuff like this to his system, I doubt he cares much if it all ‘blows up’ one day.

  13. RE eyeliner:

    what do you mean by freezing? The processor itself is already a solid – therefore technically frozen to begin with. If you mean it getting too cold to operate, the cooler system brings the chip down to its rated minimum operating temperature, not any colder.

    As for condensation – if you actually read the article, he describes how much foam insulation was needed around almost all of the piping and the processor socket itself. Also that he sealed with nail polish the area of the motherboard around the cpu socket.

    With all this very little if any condensation should occur – and the only place that it would potentially short circuit anything has been sealed against this.

  14. look at what some at the guys over at xtremesystems do.. their forum is insane!

    they have phase changers the size of mini’s. dual and triple cascades.. autocascades and so with crazy gas mixes

  15. Chris, I hate you! Articles like this tempt me to do things I really ought not try… yet.

    @ pretty much everyone who is posting comments, there’s a pretty thorough article written that covers virtually every point mentioned. He took the time to write it, why not take the time to read it?

    As for the long term risks: nobody seems to have mentioned the fact that the components will be out of date well before the cooler is likely to fail. Hence, it will probabbly get rebuilt or replaced at some stage in the near future. There’s no point running bleeding edge cooling inching performance on an outdated chip.

    @Chris: top work and excellent write-up.

  16. I once set my boss on fire with propane, and it wasn’t much worse than a sunburn. This was most of an air compressor cylinder purged with nitrogen, then pressurized to about 50 psi vaporous propane. It was supposed to shoot out of a solenoid valve at the camera. Instead, I hadn’t bagged it enough and it rolled up and over and into his face, about 6 feet away. More damage was done by the ABC fire extinguishers than by the face full of propane.

  17. I am a refrigeration tech, working on commercial kitchen equipment. A refrigeration system is NOT supposed to leak. Some survive 20 – 30 years with no leaks. BUT, There are some units that leak from the factory (Chris did an excellent job of leak checking his system) and some develop leaks years later. One just cannot say “never.” If I may add a bit of humor – – Describe this to your homeowners insurance agent! At any rate, this is a very interesting project! Thanks for sharing it!

  18. Good project, well engineered and tested.
    Here is a little nugget…no system can be ever made 100% leak-proof, it is a fundamental impossibility.
    Leakage and leak rates for systems are determined to be very small compared to the usefull life of a product. Refrigerent will always leak from a system through real leaks such as porosity in material (microcracks etc.) or due to permeation through a material.
    For hydrogen is unstoppable and will go through any material simply becuase the hydrogen atom is much smaller than the structure of the system construction materials so it will diffuse right through the molecular structure but this rate is incredibly small.
    Another example is why helium filled rubber ballons loose the gas…the helium simply diffuses straight through the rubber membrane.
    Hope this little tidbit helps…now where did I put that bottle of propane and match?

    regards

  19. Good project, well engineered and tested.
    Here is a little nugget…no system can be ever made 100% leak-proof, it is a fundamental impossibility.
    Leakage and leak rates for systems are determined to be very small compared to the usefull life of a product. Refrigerent will always leak from a system through real leaks such as porosity in material (microcracks etc.) or due to permeation through a material.
    For hydrogen is unstoppable and will go through any material simply becuase the hydrogen atom is much smaller than the structure of the system construction materials so it will diffuse right through the molecular structure but this rate is incredibly small.
    Another example is why helium filled rubber ballons loose the gas…the helium simply diffuses straight through the rubber membrane.
    Hope this little tidbit helps…now where did I put that bottle of propane and match?

    regards

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.