In lieu of an abstract, here is a brief excerpt of the content:

  • Screen History: The Haeff Memory and Graphics Tube
  • B. Jack Copeland (bio), Andre A. Haeff (bio), Peter Gough (bio), and Cameron Wright (bio)

Haeff-type tubes formed the high-speed memory of Whirlwind I but had their greatest impact in graphics and display technology, remaining in widespread use until the 1980s. Haeff seems to have been the first to store and display graphics and text on an electronic screen for an unlimited period, in 1947.

The ENIAC, which first ran in December 1945,1 highlighted not only the vast potential of electronic computation, but also the need for an economic, high-capacity read/write memory device able to operate at electronic speeds. In the ENIAC itself, as in Flowers’ earlier Colossus,2 high-speed storage consisted either of read-only switch networks of various forms or low-efficiency tube circuits (flipflops, thyratron rings). The per digit cost associated with such tube circuits was too high to permit the storage of more than a modest amount of information (Colossus’s thyratron ring memory held approximately 500 bits). As described later in Early Computer Memory: The Situation in 1945–1947, a number of researchers pursued the idea of using a cathode ray tube (CRT) as a high-speed read/write memory.

Working at the US Naval Research Laboratory (NRL) in Washington, D.C., Andrew V. Haeff—the inventor of the traveling wave tube amplifier in 1933, the inductive-output tube in 1939, and a number of other important types of vacuum tube—designed a cathode ray “Memory Tube” that could, he said, serve as a “computer memory device … to store, and to read out numbers whenever desired” (see Figure 1)4 Haeff applied for a patent on the Memory Tube in August 1947,5 and in September he published his description “A Memory Tube” in the journal Electronics (see Figure 2).7

However, the idea of CRT computer memory was first successfully reduced to practice by British engineer F.C. (“Freddie”) Williams. The Williams-type memory, further refined by Williams’ colleague Tom Kilburn, made its computing debut in 1948, in the Williams-Kilburn “Baby” computer at Manchester University. The highly successful Williams tube was the first commercialized electronic random access memory (RAM), and it served as the backbone of many early generation electronic digital computers.8

Haeff’s Memory Tube was an alternative form of CRT memory, operating on principles very different from those discovered by Williams. Haeff first publicized details of the Memory Tube in June 19479 at the Institute of Radio Engineers Electron Tube Conference in Syracuse, New York, almost a year before Williams publicized the British tube at a meeting of the Royal Society of London in March 1948.10 A photograph of an advanced developmental model of the Memory Tube was published in September 1947.11 In England, Williams and Kilburn also had an advanced developmental model at that time.12 After keen initial interest, Haeff’s Memory Tube was almost universally passed over in favor of the simpler Williams-Kilburn memory. In the end, even the Naval Research Laboratory itself opted for Williams-tube memory in its NAREC computer.13

It is interesting to speculate on the pathway that hardware development might have followed had Williams not made his crucial breakthrough as early as he did of [End Page 9] using repeated read/write cycles to maintain the charge pattern (the data) on the screen of a CRT. Haeff’s 1947 Memory Tube, now a side-branch of high-speed memory’s developmental tree, might have become the main track, with tubes based on the Haeff principle forming the staple CRT memory of early electronic computers—either until the simpler (but much slower) Williams-type tube did eventually arrive or until the arrival of magnetic core memory in 1953 and the consequent general demise of internal CRT memory. In only one case, MIT’s Whirlwind computer, was a Haeff-type memory actually used.


Click for larger view
View full resolution
Figure 1.

A.V. Haeff in about 1949.3


Click for larger view
View full resolution
Figure 2.

The experimental Memory Tube circa June 1947.6 Its evacuated glass envelope is approximately 18 in. long. The...

pdf

Share