I began working the power system for the Burst Alert Telescope of the Swift satellite in 2000. I designed several bits and pieces that wound up in various assemblies in the BAT, but I was primarily responsible for the low-noise voltage regulators that powered the amplifiers in the sensor array. That regulator assembly was called the XA1VR. There were 16 of them in the BAT.
Because house calls are impossible on most of NASA’s hardware once it gets on orbit, reliability is a major design concern. That means that so-called space-class parts are used almost exclusively. This was a significant problem in the design of the XA1VR because there were several key components which were not yet available as space- or even military-grade parts. The pass transistors were a particular problem. Not only were they not available in the sort of hermetic packages normally used, the manufacturer had no plan to make space-class parts available until well after the scheduled launch date for the mission. Because commercial plastic parts had to be used, there were additional handling precautions that had be taken with the assemblies.
All 16 flight units of the XA1VR and all the spare units were delivered on time. They were the only assemblies delivered for integration into the BAT on schedule. They were the only assemblies that required no rework. The service life requirement for the mission was two years. All 16 XA1VRs are still working on orbit after almost a decade.
Image Credits: NASA
The picture shows an XA1VR (mounted in the frame that holds it and 8 sensor blocks) being inspected after handling during system integration. The bundle of white wires carries the high-voltage bias power for the x-ray sensors. The bias potential is supplied by an adjustable regulator in an assembly called the Block Voltage Regulator. The BVR takes power from a low-voltage bus and generates a slightly-greater-than 300 V low-current internal bus. (That part of the BVR and the overall design of the BVR assembly was done by my colleague Lowell Fry.) That bus fed individual 0 to 300 V adjustable regulators, one for each sensor block. I designed the adjustable regulators. Because the currents being controlled were so tiny, bipolar transistors had to be used in those regulators. Here we had the opposite availability problem. Such transistors were common parts back in the days of discrete transistor television sets, but, by 2001, they were no longer readily available as current production and completely unavailable as new space- or military-grade parts. Fortunately, we found some old-stock military parts, and I was able to design a workable solution using them.
From time to time, I’m able to post some new science from the Swift mission. I feels good to have been a small part of getting it off the ground.
UPDATE—Broken link fixed.