|
After doing this transistor
circuit development at RCA labs in Princeton, in 1951 I teamed with a senior engineer in the design and
construction of a 5,120 bit (not byte) memory system for the Air Force
Cambridge Research Center using Jan Rajchman’s SB256 Selectron tube. (The senior engineer – remember that I
was only 23 at the time - was Igor
Grossdoff, who was a Russian,
perhaps 55 or 60 years old.
He was a friend of David Sarnoff, another Russian émigré). In the late 1940s, the principal
random access memory device was the Williams tube in which 256 bits were
stored as charges on the phosphor of 5” cathode ray tubes. RCA’s Rajchman invented the Selectron
tube as the first “fully digitally accessed” memory device, to replace the
analog-accessed Williams tube. We
developed a 5,120 bit memory system for the Air Force Cambridge Research
Center for $750,000! These weren’t
really “computer-reliable, since the MTBF wasn’t very good. The $750,000 that the Air Force paid
works out to about $146.48 per bit.
The pain of the high price was offset, however, by the joy of its 16
microsecond high speed access.
Returning to Camden in 1952 I
was manager of an engineering team which developed the "highs
speed" random access magnetic core memory for BIZMAC, RCA's first
commercial computer. The system
stored 14,336 bits in seven 6' relay racks.
______________
|
