Project 173K: the journey to -100C
Page 2: Design of the -100C refrigeration system

The basic setup for Project 173K is very similar to that of Project 210K. It is also a 2-stage classic cascade, thus with 2 separate refrigeration systems running a different refrigerant (if you have no idea what that means, check this page). Project 173K is so similar to Project 210K that, instead of describing the details, I describe the differences. If you haven't read the design part of Project 210K, I would recommend to do so first. Click here to visit this page.

Project 210K vs. Project 173K:

Project 210K Project 173K
High stage compressor Old retrofitted R22 A/C compressor Tecumseh CAJ2464Z 1.5hp low-temp R404a compressor
High stage refrigerant R507 R507
Interstage heat exchanger Standard small heat exchanger Custom, better heat exchanger needed
Low stage compressor 1/3hp low-temp R134a (Danfoss NL11F) 3/4hp low-temp R502 (Danfoss SC21B)
Oil separator Custom made oil separator 'real' oil separator with float
Low stage refrigerant R410A (boiling point: -52 °C) R1150 (boiling point: -104 °C)
Low stage temperature controller None Partlow MIC1160 intelligent PID process controller

The hugest difference is the low stage refrigerant, for which I intend to use ethylene, also known as R1150. Ethylene has a very low boiling point, and features very high pressures to work with, even at low temperatures. The result is that the gas features a high volumetric efficiency, and has the capability to drop evaporator temperatures to -120 °C.

The rest of the changes comes from the use of ethylene. Ethylene features very high condensing pressures, even at low temperatures, and thus the high stage must be capable of maintaining low interstage HX temperatures. Hence the quite powerful high stage compressor (powerful enough to keep about 7 household deep freezers cool). Also, load on the low stage compressor is higher, hence the more robust SC21 instead of my old thrusty NL11F.

When incorporating all these changes into the Project 210K piping diagram, we get this:


Project 173K piping diagram


How does it all work? Well, we have two separate stages.
The 'high stage' cools the interstage heat exchanger down to about -40C (-40F). It does so by evaporating R507 refrigerant in a standard TEV based system. The compressor compresses the R507 vapour, which is then condensed in the aircooled condenser. The condensed liquid goes into a storage vessel which serves the purpose of coping with varying refrigerant needs. Liquified refrigerant from this vessel is dried and filtered in the drier, from where it is fed to the thermostatic expansion valve (TEV). The expansion valve meters the refrigerant into one loop of the interstage HX where it evaporates, extracting heat. The TEV also adjusts the amount of refrigerant metered in by sensing the temperature of the vapour exiting the interstage HX. From there, the vapour is fed to the compressor, and the cycle starts again.

The second stage, called the 'low' stage, works the same, but this time with ethylene. The compressor compresses ethylene vapour to a pressure of about 16 barg (~230 psig). This hot vapour passes an oil separator, which separates the oil from the refrigerant. This is necessary since most refrigeration oils freeze around -60 °C, so leaving the oil circulating with the etylene would result in a clgged evaporator. The oil is directly fed back to the compressor. The hot, but now oil-free ethylene vapour continues it's journey to the interstage HX. There, the vapour is cooled down, and it will start condensing into liquid. This liquid is fed to another small filter/drier, where the ethylene leaves all the water and possible debris. Then, it is expanded into the evaporator by means of a capillary tube. In the evaporator, the etylene evaporates, extracting heat from our processor. Although tuning a capillary tube is a major pain in the ass, it is used due to the reliability and simplicity.

To prevent disasters, a few safety interlocks are added. The first one is the Eliwell EWPC-901. This is an electronic thermostat which is set up to switch on the low-stage compressor when it senses a temperature of -40C. It switches off the low stage compressor again when temperature rises above -30C. This prevents an unlimited rise in low stage condensing temperature (and thus pressure).
Since electronics are utterly unreliable, there is another safety device added: the high pressure cutout switch. This switch senses the discharge pressure of the compressor. If, for whatever reason, this pressure rises above 24 barg (~350 psig), the low stage compressor is switched off also.
If both safety interlocks fail, the last resort is the break plate in the compressor itself. This breaks at 30-40 bar (450-600 psi), and renders the compressor unuseable. To use this last resort, high side piping must be able to withstand at least 50 bar (~750 psi) pressure before it bursts.

The CPU evaporator will be connected to the system with a flexible braided full stainless steel hose. This hose type can withstand low temperatures, high pressures and chemicals, and it has a low permeability (leakage through the hose). Just what I need. Almost all other hose types are unsuitable for this purpose. They either won't resist the pressure of the low temperature. And since ethylene imposes a potential fire hazard, I need a hose that won't get me into trouble.