The RCA Selectron -- The people of IBM
IBM evaluated the Selectron for
use as the memory in its type
701 Electronic Calculator.
Ultimately, the decision was made that the main storage unit would be implemented with a
principle storage cathode ray tube system.
The 706 Electrostatic Store
was the main memory for the 701 and
This type of CRT was the basis for the 701's official logo. (And no, there was no lithium involved -- the graphic of three orbital electrons was just the 1950's way of saying, "We're Modern!" and not a warning of, "Danger! Radioactive Alkali Metals Inside!")
The commonly accepted reasons for the decision not to use the Selectron was that the device itself was too complex, the price RCA set per device was too high for IBMs projections, it was a sole-sourced technology still not proved in field use, and the small storage capacity of 256-bits would require 288 SB256 Selectrons. RCA did (initially) supply the conventional 5-inch CRTs to IBM for the project. IBM began building its own CRTs in house because the performance from vendor supplied tubes was insufficient for both yield and lifetime. IBM has long held sufficient clout in the electronics industry to "force" compliance with their requirements. In circumstances where the external market will not -- or can not -- meet the needs, IBM would do it themselves.* Setting up in-house production of conventional CRTs is not a foolish task; Building Selectrons internal to IBM would have been a daunting and time consuming chore. This may have been the key point that ultimately swayed IBM away from using the proprietary RCA Selectron.
Meanwhile, Arthur L. Samuel was leading the development of tube and tube circuit development at IBM while also leading research into the applications of the newly invented transistor. IBM's Tube Laboratory in Poughkeepsie was housed in a building formerly used for pickle processing by the R. U. Delapenha Company, earning it the nickname The Pickle Factory.
By 1954, and the introduction of the type 704 computer, the electrostatic store had been replaced by magnetic core memories. Although the common attribution for the invention of the magnetic core memory goes to MIT's Jay Forrester , significant and simultaneous R & D contribution to the art was being performed at MIT by William Papian, by L. H. Thomas at Columbia University's Watson Lab, and by Harvard Computation Laboratory's An Wang...and RCA's Jan Rajchman -- the developer of the Selectron.