Major challenges

The group’s experience with balloon-, rocket-, and rockoon-launched instruments put us in an excellent position to develop the new satellite instrument. We were well versed in building the types of electronic circuits that would be required, and we had learned how to build them ruggedly enough to withstand the stresses of rocket firing. Nevertheless, designing an instrument for a satellite added new dimensions. In my 26 April 1956 notebook entry, I listed the major problems foreseen in developing the

OPENING SPACE RESEARCH

Vanguard instrument. Foremost among them were (1) learning how to use transistors, (2) developing the in-orbit data storage system, and (3) miniaturization.7

The transition from the vacuum tubes that we had been using in our balloon and rocket instruments to transistors was essential in order to minimize the instrument’s size, weight, and power demand. I worked hard during the initial few months, and, indeed, throughout the entire developmental period, to build miniaturized circuits that would operate at the required low power levels and still be highly reliable over the expected range of operating conditions.

Before starting work on the satellite project, I had never even seen a transistor, let alone developed a circuit that used one, although I had been following their evolution through the engineering literature out of general curiosity. The very few existing books on transistors were not at all helpful. They focused on a network analysis approach to understanding the basic characteristics of the devices and were not useful in designing actual working circuits. I quickly abandoned any theoretical approach for designing circuits and adopted a much more pragmatic experimental approach. Starting with appropriate circuits from our vacuum tube experience, I substituted transistors, and then varied circuit topologies and component values until I obtained the results that I was seeking. That approach eventually gave me considerable insight into the internal workings of the new devices, and I soon developed the same intuitive familiarity with transistor circuit design that I had enjoyed with vacuum tube electronics.

Admittedly, that first U. S. satellite instrument appears trivially simple by the stan­dards of current massively dense fabrication technologies. Today, all of the Explorer III electronics circuits could be fabricated in one or a few very small silicon chips. But those technologies were not available in the 1950s, and we were pushing the then-available state of the art.

The second major challenge was in devising a suitable device for storing the data during each orbit, and for relaying them rapidly to the ground when the satellite passed over the ground stations. As already mentioned, Van Allen’s proposal for the cosmic ray experiment was such that data recovery over a broad range in geomagnetic latitude was essential. Both his late 1955 and early 1956 cosmic ray proposals focused on the use of networks of ground receiving stations for that purpose. Although he did mention the use of “a magnetic storage drum” in his second Ann Arbor paper dealing with the auroral soft radiation, he did not mention onboard storage in the cosmic ray proposal. That was in spite of the fact that, from the beginning of our discussions in the fall of 1955, he and I had talked about its desirability and some early ideas for achieving it. Certainly, from very early, he strongly favored an approach employing onboard storage because of its vastly greater data coverage, even though we knew that development of the in-flight hardware would be challenging.

Van Allen and I increased our efforts immediately following the Ann Arbor meeting to examine various options for the instrument configuration. We continued to keep

CHAPTER 5 • THE VANGUARD COSMIC RAY INSTRUMENT 129

both the no on-orbit storage and the data storage options open while I examined the feasibility of an onboard data recorder. By early May, I was convinced that in-orbit storage was technically achievable, and we committed ourselves to that approach.

As events unfolded, both configurations were ultimately employed, with Explorer I being made as simple as possible with no on-orbit storage. It was followed by the unsuccessful Explorer II and the fully successful Explorer III, both of which carried the onboard storage device. The presence of onboard storage in Explorer III proved to be a critical element in interpreting the unexpectedly high counting rates encountered by the pathfinder Explorer I and was, therefore, a major factor in the discovery of the Earth’s radiation belts.

I began my design effort by inventorying known options, including the counting of shaft rotations, accumulating charge on a capacitor, chains of bistable scalers, magnetic matrix (core) storage, dielectric matrix storage, magnetic and dielectric tape recording, magnetic and dielectric drum recording, magnetic wire recording, cathode ray tube storage, and mercury tank storage. The latter two were discarded outright. By 7 May, I had narrowed the viable possibilities to magnetic drum, magnetic tape, ferroelectric matrix, ferromagnetic matrix, and capacitor bank storage. Two days later, I decided to proceed with magnetic drum storage.

The data storage device eventually passed through four major design phases. The first model, Mark I, was the drum recorder. The Mark II, III, and IV models, all based on magnetic tape storage, incorporated a number of progressive improvements. Mark II used Mylar tape and a mechanical centrifugal governor for controlling the playback speed. For Mark III, metal tape, a magnetic field-eddy current speed control mechanism, and an improved tape advancing mechanism were substituted. The final Mark IV design incorporated a further-improved tape-advancing mechanism.

Today the circuitry on a small fraction of a square millimeter of a solid-state memory chip would provide the same functionality as the Explorer III tape recorder. But that was before the age of integrated circuits, and the only practicable approach was to develop an electromechanical device.

The third major challenge was in miniaturization. Because of the extreme constraints in satellite size and weight, we needed a much more compact method of assembling our electronic components than we had used for the vacuum tube circuits in the balloons and rockoons. Single – and dual-layer printed circuit boards were state of the art in 1956—the Naval Research Laboratory (NRL) and the Jet Propulsion Laboratory (JPL) engineers were using them routinely, and they looked promising for our work. However, there were no facilities for producing them in the Iowa City area. I purchased the necessary supplies, set up a trial printed circuit facility, and produced a few boards to check out the techniques.

Their quality was poor. I turned to an alternate process using terminals that were pressed, or swaged, into holes drilled in fiberglass circuit boards. The component

OPENING SPACE RESEARCH

leads and other interconnecting wires were wound around the heads of the terminals and soldered in place. Although inelegant, that approach proved to be rugged and reliable, and the circuit boards could be assembled wholly within our laboratory by student aides.