This is the third video in my series on building/testing a Pulse Tube cryocooler with the eventual goal of liquifying nitrogen/oxygen.

Part I:
   • Pulse Tube Cryocooler - Part 1  

Part II:
   • Pulse Tube Cryocooler - Part 2 (-75C)  

In this video I investigated the effects of different pulse tube materials, regenerator materials, heat sink designs, inertance tube geometry, and pressurization of the working fluid. While I didn't manage to exceed my record from the previous video of -75C drop (corresponding to a drop below ambient of about 100C), I did gather a lot of information about design factors.

My initial design in this video used a stainless steel pulse tube/regenerator housing, but i found that the temperature difference generated for a given power input was dramatically lower than with PVC parts due to conduction losses.

One of the biggest takeaways from the tests i ran was that performance is almost directly proportional to the pressurization of the working fluid, and having a large average pressure is much more important than having a high compression ratio. This is consistent with how real cryocoolers are built, which are typically pressurized anywhere from 10 to 30 atmospheres, but have pressure oscillations of under 10% of average pressure.

I also experimented with different regenerator materials, such as ceramic beads, glass beads, plastic pellets, and glass fibers, but found that compacted stainless steel wool (which i started with) still performed best.

For configurations both with the 25mm diameter piston and the 40mm diameter, i found that the cooler seemed to hit a wall at around ~100 degrees of temperature drop below ambient, where application of additional power only marginally improved performance. I suspected this was related to limitations imposed by the inertance tube and compared my 1/4" copper tubing against 3/8" flexible silicone tubing of a greater length, but I found that this change reduced performance, most likely due to increased surface roughness and flexibility in the line dissipating energy. In a future video I'll probably try to use rigid copper/aluminum tubing with a 3/8" or 1/2" inner diameter.

I also reconfigured the entire device into an alpha-type stirling cooler, but found that performance was actually dramatically reduced despite the ability to mechanically set the phase angle between compression and expansion. I think this is because the cold expander piston was causing large conduction losses through its thin aluminum walls.

I ran the device with loads disconnected, and with pistons disconnected to determine the amount of power being consumed by mechanical action as opposed to pneumatic power, and found that less than 30% of the input electrical power was actually going into the system.

Finally, I examined the effect of a double-inlet valve, which has the effect of improving phase shift and removing some of the load on the regenerator. While this didn't make a tremendous difference, the difference is very obviously apparent and repeatable.

In my next video, I'm going to build an entirely different test setup using a valve-based (or "Gifford-McMahon") configuration and a standard air compressor as a high pressure source. While this configuration is less efficient, because the input power would be so much larger, it should be an overall net positive. In addition, control over valves allows me to achieve consistent timing via digital control and fine tune it for best performance.