Neither in its visits nor in conversations did the WTEC panel hear of any cryopackage activity for wireless subsystems that approaches the level of sophistication that has been demonstrated by the four small U.S. companies Conductus, ISC, SCT, and STI. In the advertised products of these companies, integration of the cooler with a number of filters (six or more) and low noise amplifiers has been achieved, complete with a vacuum enclosure for long life, lightning protection, and default to a conventional system in the event of refrigerator failure. Figure 7.2 shows a picture of a complete package from Conductus.
Fig. 7.2. Part of a commercial HTS wireless subsystem (Conductus).
In Japan, the WTEC panel was shown single filters in a research-style package for use on the bench top (Fig. 4.3, p. 26). These are similar to the state of the art of the four U.S. companies about two or more years ago, which are still used by these companies for device testing purposes. The rapid progress of these U.S. companies on a wide variety of packaging issues over the two years prior to this WTEC study demonstrates the progress that can be made when system-level objectives are defined and are driven by the prospect of market applications close at hand. Similarly rapid progress towards system goals was made in Japan at the Superconducting Sensor Laboratory, as discussed in Chapter 6.
Almost exclusively, superconducting quantum interference device (SQUID) systems have been cooled either by liquid helium or nitrogen. The liquid cryogens are ideal for this purpose, as they contribute no magnetic noise and little vibration. As most SQUID products to date have been sold to research laboratories, availability of the liquids has not generally been a problem. However, Prof. John Clarke of the University of California at Berkeley has suggested that the difficulty of obtaining liquid helium in remote locations around the world was one reason why low temperature (LT) SQUIDS were not accepted for geophysics surveys.
If SQUID systems are to develop into large markets, the use of a cryocooler will be necessary for some applications. By far the most comprehensive and detailed study of the issues involved in such use has been that at the Daikin Laboratory (see the site visit report in Appendix B). The most demanding application was chosen, a 61-channel system for magnetoencephalography (MEG), and a 32-channel magnetocardiography (MCG) system also was built, both using low temperature superconductor (LTS) Nb sensors. A systematic characterization of the vibrational and electromagnetic environment created by the cooler was performed. Ways were found to create a "template" of the cooler noise, which could then be subtracted from the signal. Figure 7.3 shows a schematic of the 61-channel system, and Fig. 7.4 shows a cardiogram taken with the 32-channel system. Clearly, many of the problems associated with using SQUIDS with cryocoolers have been solved at Daikin.
Fig. 7.3. A schematic of a 61-channel SQUID system; cross-section of the cryostat (Sata et al. 1997).
Fig. 7.4. Noise-cancelled signals of cardiogram taken with 32-channel SQUID system (Fujimoto et al. 1997).
An alternative approach, which is being investigated in Europe, is to use a remote refrigerator and a cold gas flow to the SQUIDS. This removes the magnetic noise and vibration from the vicinity of the sensors. A cooling system of this type, required to cool the tuned HTS coil to below 30 K, is sold in the nuclear magnetic resonance (NMR) probe made by Varian and by Conductus (whose technology has since been sold to Bruker). The group at Daikin considered this type of SQUID system and decided that it was not sufficiently challenging!
No SQUID research groups or companies in the United States have yet tackled the issues of SQUID operation near to cryocoolers.
Towards the end of the Josephson Computer Project, the group at Fujitsu demonstrated a hybrid digital system in which a Josephson digital signal processor chip was connected to semiconductor circuits at room temperature. The custom cable made for the interconnect, of thin copper on polyimide, went directly through the wall of the Dewar that separated the liquid helium from the vacuum, and also through the wall between the vacuum and the air. The Josephson chip was cooled directly by liquid helium, which was replenished by condensation on the cold head of a refrigerator. This system still appears to be the most complex Josephson/semiconductor hybrid digital system assembled to date in Japan. In the ongoing hybrid project funded at NEC and Hitachi by the Future Electron Devices Research and Development Association (FED), there are plans to use both Josephson switches and semiconductor components. But the panel did not see any construction of the cryogenic package for such systems during its visit.
Meanwhile, in the United States there are similarly two Josephson switch projects nearing completion. Both of them use refrigerators to cool the Josephson chips and semiconductor components, without the need for liquid cryogens. The package becomes quite complex in such cases. Fig. 7.5 shows the final demonstration system from the DOC-funded Advanced Technology Program hybrid switch (4 x 4, 8 Gbit/s) project completed in the summer of 1997 by Conductus, NIST, Stanford University, TRW, and the University of California at Berkeley.