Heat Pipes

[raptor1956]

Anyone here design anything to be cooled or heated using heat pipes?  I'm working on a project that uses Peltier modules to heat or cool and the temperature difference you can achieve between ambient and the device is a function, among other things, of how well you can remove heat from the hot side of the module.  A measure of that would be the difference in temperature between the hot side of the module and the free air temp and a more efficient cooling of the hot side will permit a lower temp on the cool side as well as higher cooling power.

The thing is ... a really good Peltier module is kind of pricey and it will still need a good heat removal system to function well.  If a modestly more expensive heat removal system using heat pipes permits using lower cost Peltier modules it might make sense to go that route. 

There are really two kinds of heat pipes as best I can tell:  ones with some form of wick to transport the condensed fluid back to the hot side, and the type that does not have a wick and relies on thermosyphonning.  For the temp range I'm looking at (-40C to 70C) it looks like aluminum/ammonia are the materials to use.  A thermosyphon approach would likely be cheaper and in my configuration I think I can depend on the thermosyphon to work though I understand bubbles will interfere with the vapor flow reducing the heat transfer to just the conduction through the Al tubing.  I would appreciate any feedback on the causes and remedies of bubbles.

Just to paint a picture of what I'm looking at I envisige bonding both side to Al blocks with the hot side having heat pipes extend upwards from the block, then travel horizontally through and in contact with an fan cooled Al heatsink, then travel down below the bottom of the Peltier block, then horizontally back to the block, then upwards back to the block completing the loop.  As I understand the system I need to maintain both liquid and vapor at all times and temps and that the minimum temp needs to be greater than about -60C and less than 100C when using Ammonia.  Any idea on what the proper volumetric ration of vapor to liquid needs to be and at what temp this is determined?

Additionally, if I were to power the device backwards so that the hot side becomes the cold side and vis versa with the arrangement as described the heat flow should still work.


Brian

[martinator]

I won't post a direct link in case it's naughty but if you google (will PM you)

heat pipes fifth edition theory design and applications pdf

a couple of hits down you should find a pdf link (not the cern link). I struggled with the maths myself as I'm not a number cruncher but it was a very interesting book (I used to own the hardcover). There's a newer sixth edition that is available from amazon. ISBN-13: 978-0080982663

[raptor1956]

Thanks, the doc is petty big and definitely heavy on the Maths but still helpful.  The  paper is more of a science paper and not so much an engineering work so it isn't going to point the way without delving into this at a level I'm not qualified to do. 

The two approaches, using a wick and not using a wick, is mentioned here as well but not in an engineering sense, again, more of a scientific paper.  I know enough to design and build a prototype for testing so perhaps that's what I need to do.  Logically the thermosyphon approach seem most appropriate but the overwhelming majority of heat pipe system use wicks.  The safe bet would be a gravity assisted wick based system but that would seem to be more expensive.

Among the interesting facts I was unaware of prior to investigating this is that the working fluid for the majority of CPU cooling heat pipe system is none other that good old H2O -- the latent heat a operating temp range work out well for water in this application.  In my use case I need to be able to deal with much colder temps so water is out.  Given the temp range I have ammonia is the working fluid of choice and if you go ammonia then the tubing needs to be aluminum.  That isn't a problem, I think, as the actual heat exchanger will be aluminum as well.  Now ammonia is a bit nasty but the quantities are small and contained in a sealed system.


Brian

[martinator]

I've forgotten most of what I read from that book but I thought that the main difference between a thermo-syphon and a heat pipe was the use of a wick type system. You could have a look at propane as well. It maybe easier to use than the ammonia?

[raptor1956]

It is pretty amazing just how many chemicals can be used as the working fluid in heat pipes and how many of them are flammable.  Ammonia is a good choice for the temperature range I need while water, which is even better within the range it works isn't workable because it will freeze solid long before it gets to the lower temps I need to operate in.

As an example of the Peltier modules I'm working with I built a test rig to test them and in my testing I find the the fan cooled hot side heatsink will tend to go to 16C above room temp to over 24C above room temp depending on the module and the current it draws.  The higher current ones are at the 24C end while the lower current ones tend to have a lower hot side temp.  Also interesting is that while the higher current modules have higher cooling power when close to room temperature that advantage diminishes as the cold side gets down below -20C relative to room temp (20 degrees below room temp).  I tested all my modules at room temp, -10C, -20C and -30C (relative to room temp) and I found that the lower powered models (12706) get colder and have more cooling power at the lower temps than the higher current models (12709) even though they consume about 2/3 the power.

If heat pipes can reduce the hot side temp by 5C that should lower the minimum temp by about 5C and increase the cooling power at all temps.  I can live with a few watts of cooling power at -30C relative but 5W or more would be nice and being able to get to -40C with cooling power would be a treat.  I think a well designed cost is no problem approach could limits the hot side temp to no more than +10C (above room temp) and quite possibly even lower.  So, I suspect a minimum temp of -40C (relative) is doable and doable with cooling power at that temperature difference.


Brian

[calexanian]

I have seen LP (Liquified Petroleum, close cousin to propane) used successfully. Much safer than ammonia. Alcohol is also used in partially evacuated heat pipes as well. I believe it was the most common liquid used in small IC heat pipes because of its inherent safety.

My info may be way way out of date though.

[raptor1956]

Well ammonia can be a safety issue, but for small systems with perhaps only a gram or two of ammonia per tube the risk is actually quite small unless in a very small confined space.  Ammonia is a common fluid in heat pipes and you don't hear of many incidents with it.  It is toxic, but in humans we process and eliminate it pretty effectively so unless the quantities/concentrations are high it poses no major risk to humans.  It is more dangerous to aquatic critters as they are not so capable of processing and eliminating it.  But, again, in my use case I'd be looking at a gram or two per tube.  Cigarette smoke contains ammonia as does many others chemicals in our homes.


Brian

[Cerebus]

Thanks, the doc is petty big and definitely heavy on the Maths but still helpful.  The  paper is more of a science paper and not so much an engineering work so it isn't going to point the way without delving into this at a level I'm not qualified to do. 

For those of us, like me, who consider themselves relatively 'math challenged' don't be put off by the above. This isn't full of partial differential equations, or calculus. I'm 1/4 the way through and I haven't seen anything more scary than basic algebra and trig - if you've done the equivalent of GCSE or O-level maths (exams at 16yrs old in the UK) then you'll be fine.

[raptor1956]

No, the doc isn't Phd level Math but it is more aimed at the scientific end of things than the engineering end.  Still plenty of good info in there and much of it pretty old, not that old is bad. 

The fact that a phase change can produce 100X or more the transport of heat versus a good metal like copper is pretty amazing when you think about it and to do that with very little temperature drop is even more amazing.


Brian