30 W at 50 K Single-Stage Coaxial Pulse Tube Cooler with Tapered Buffer Tube

The performance of large-capacity single-stage pulse-tube coolers at temperatures below 70 K is often hampered by Rayleigh streaming, or the natural boundary-layer convection that occurs in the buffer tube (‘pulse tube’). In Olson and Swift’s landmark 1997 paper1, they explained how a slight taper in the buffer tube could suppress this streaming. They also showed how the streaming could be suppressed by the proper phasing of pressure vs. flow in the buffer tube, which is enforced by the proper choice of phase-shift mechanism (i.e. length and diameter of inertance tube, etc.). This is the approach usually taken because a straight buffer tube is simpler, especially when considering a coaxial construction (where a tapered buffer tube would imply a tapered regenerator). In addition, the phasing which suppresses Rayleigh streaming coincides with efficient cycle phasing for many applications. At temperatures below 70 K, however, this is less true, and for a 50 K machine a significant benefit may be realized by decoupling the streaming suppression from the cycle phasing. At the same time, we have found that coaxial coldheads can successfully use nonconductive buffer tubes, if the material with the right thermal expansion coefficient is selected. This enables the use of a tapered buffer tube in a coax design, as the material can be thick enough to have a constant outside diameter (OD) and a tapered inside diameter (ID). This paper will discuss the results obtained on a high-capacity cooler using a tapered buffer tube and includes some measurements showing the importance of having the right taper angle. INTRODUCTION Large (input power 2 kWe) high-frequency (>40 Hz) pulse tube coolers2 have had increasing success in recent years, reaching capacities as high as 1000 W at 77 K. In such large coolers, the Rayleigh streaming can be a very significant parasitic loss, which becomes more significant as the temperature approaches the cooler’s ultimate temperature; in this situation the gross cooling of the cooler is relatively small and the streaming penalty is greater. Thus, suppression of the Rayleigh streaming is absolutely necessary for operation near a large pulse-tube cooler’s ultimate temperature. At the same time, the phase shift necessary for efficient operation increases at lower temperatures. It is therefore at least theoretically advantageous to suppress the streaming by using a tapered buffer tube, and choose the cycle phasing independently. The awkwardness of a tapered buffer tube in a coaxial coldfinger can be overcome by using a thick-walled, nonconductive material, since the buffer tube is not part of the pressure vessel. We have found that Garolite XX is a suitable material for buffer tubes, as its thermal expansion coeffiat