Electronics > Projects, Designs, and Technical Stuff
guidebook on medium freon compressors interfacing?
<< < (3/4) > >>
Rerouter:
Start from the simplest approach, you need a pump, a heat load, a heat dump (radiator) and an expansion valve to drop the pressure,

The cooling effect you get is from a pressure drop initiating a phase change, if the pressure drop is enough that it wants to change phase then you also get a boost from the latent heat of the phase change,

When a material reaches a point where it wants to change phase, generally it involves energy from the outside, e.g. water becoming steam needs additional energy from its surroundings, and vice versa when it condenses it gives off a similar amount of energy, this is how your refrigerator works.

So you design your loop around this phase change, just pressure vs temperature, pick a chemical and you can likely find its temperature / pressure phase chart online. So you need a pump that can reach a pressure where it becomes a liquid at your desired temperature range. but is a gas when not under operating pressure.

Most refrigeration loops do not use pressure feedback, only temperature feedback, run the pump for X time, then start to crack the expansion valve, you should see your heat load start cooling quite rapidly, then regulate the valve position to balance how much you want the inside to be cool, vs not heating up the radiator so much your fluid boils back to a gas, or so cold on the valve that it freezes over (that valve will be the coldest part of your loop in general)
coppercone2:
is it possible to have continuous control rather then the on/off control? like fully analog
Rerouter:


Cheap expansion valves are self regulating. Focus instead on varying the speed / power to the compressor if you want to regulate it, be aware most are rated at a certain duty cycle, you can drive them at the equivilent power duty cycle constantly, but dont try and run the name plate rating full tilt or you will cook it,

As the pump compresses the gas it heats up much hotter than simply the electrical power fed in, same reason why the valve gets cold from the pressure dropping over it.

You will have probably noticed with just about every fridge and aircon out there that they pump up and then they cool, there is pressure for the valve to drop, and the compressor can usually provide way more flowrate than the system needs, so the simplest control method is have the pump cycle every so often.
dmills:
Note that not all compressors do well with variable speed, some need to be running at pretty much full song to get the lubrication to work correctly.

There are plenty of non obvious refrigerants, some of which have issues of their own:

Propane for example works well as anyone who has had a high flow application from a too small tank can tell you, but leaks can be bad news.
CO2 is workable, just got to keep the pressure high enough to get a liquid phase.
Anhydrous ammonia, dangerous as hell, but it works....
I am guessing that nitrous oxide would work (oxidation hazard however), being as it is physically similar to CO2 in most respects.

However, water has a HUGE specific heat capacity, and unless you are running a **BIG** load for long periods, a big tank and passive cooling usually does well and has none of the annoying issues of lines freezing and similar nonsense that refrigerated plants can suffer from.
coppercone2:
LOL nitrous oxide. The movie Bad Sheep comes to mind.

Why is this fucking spark moving so slowly man..............?

Is there a compressor that does not have problems with lubrication at low duty cycles?

Can you maybe cut oil grooves into it or modify the oil grooves with a chisel to change the lubrication properties?

The main reason I wanted to do the custom heat exchanger is because I thought I can minimize the tank size. If you can some how match the thermal extraction of energy by water from the freon (using flow) then you can get away with the minimum tank size, which would be enough to fill the hoses, pump and DUT with a bit extra to spare. Then if you get a leak in the lab its maybe like 2 gallons tops? I imagine with a on/off compressor you would need a big thermal average.

I know how to do this with peltiers but the problem is I would be too cheap to turn it on. If I was rich I would.

Another idea might be to use a small water volume with LOTS of copper to act as a non fluid thermal averager. Given how ridiculous this already is, buying a few copper bricks might not be the worst idea. It might be the cheapest and most reliable option actually. Like build a tank around a jenga stack of copper sheets.

Then you can add a high resolution mode where you turn on the peltier heat exchanger and interface that with some inline RTDs connected to the DUT (like voltage sense).

I am just thinking now if you can use the compressor to cool the peltiers too in addition to cooling the process.



The problem is though, you would still want a custom radiator, even if you gut a water chiller unit.

Am I correct in thinking the model of this is basically a SCR-switched preregulator (the compressor is gonna have a really really crappy time constant compared to the peltiers) connected to a shunt regulator? What a weird electrical supply that would be (analogy wise).

Then you can use those shady-mixed hydraulic/electrical connectors to just retrieve a RTD signal and provide flow on a nice manifold. (who would use this to actually carry power o_O). There is a mil-spec one that looks like a DSUB with coaxials but the coaxials are supposedly 0 leak plug connectors (yea ok maybe in a fucking clean room). But it would be harmless with a RTD and a 1/4 inch coolant line for small prototypes.

If you do this put a little mesh under your manifold that goes to a drain tank so you don't get drips or leaks on your work table.

I am saying this because 90% of thermal products look like they belong in the loading bay of the nostromo. And they are equally likely to house evil creatures.

youtube.com/watch?v=NFiLVrux4V8 . Yea thats why you don't use evaporative cooling. 
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod