Radio age research, manufacturing, communications, broadcasting, television (1941)

This electronic memory device NEVER FORGETS In a Few Millwnths of a Second it Accepts, Holds and Releases any Part of 10,000 Units of Information By Joseph L. Blotner RCA Laboratories Division, Princeton, N. /• L the short time since World War II, the electronic computer has grown from infancy to a healthy adoles- cence. Its next step to full-fledged adulthood will be helped along by one of the research projects under way at the David Sarnoff Research Center of RCA at Prince- ton, N. J. The need during the war for fast, accurate fire con- trol equipment and radar networks pointed clearly to electronic tools as the solution. Targets sped too fast for the old mechanical fire directing computers, just as the acoustical air-raid warning system was unable to cope with the new tactics of air warfare. The computers which resulted from this war-spurred development took two basic forms called, respectively, the analog and the digital. In the analog system, voltage or current is varied to represent different quantities in the complicated equations to be solved. In the digital system, all the information is converted into numbers rather than varying voltage or current. The numbers are coded in terms of "on-off" signals. In the binary code, numbers take the form of a sum of powers of two. In the decimal code, numbers take the form of a sum of powers of ten. The code is chosen by balancing efficiency and versatility against practicality. With the end of the war, research recognized the great value of these instruments. They could relieve scientists of long, wearying computations. In some cases, they could be used as a sort of crystal ball, to find out whether or not equipment would work even before it was built. Progress in their development was rapid and a fairly standard computer design soon took shape. The Input, perhaps in the form of a teletypewriter and tape, puts information into the Memory. The Con- trol executes orders stored in the Memory. Next, the Control actuates the Arithmetic Unit to perform the cal- Minute size of the Myriabit elements is emphasized here by the cigaret which spans more than three groups of wires and cores comprising 300 units of information. culation, and then transfers the result to the Memory. Finally, the Control transfers the result to the Output. It was soon obvious that the range and value of the computer depended largely upon its memory. The memory could be made very large, but if it was, it took too long to get information into and out of it. If this access to the memory's information was speeded up, it meant a sacrifice in capacity. To evolve a large memory unit with speed of operation was the problem which Dr. Jan A. Rajchman and his colleagues at the Research Center set out to solve. It is possible to construct memories of various types. They can be tele-typewriter perforated paper tape, mod- ern magnetic tape, or a rotating magnetic drum. In all of these types, unwanted information, often a lot of it. has to be passed over to get to the desired information. To eliminate this delay, Rajchman and his group devel- oped a selective electrostatic storage tube which gave immediate access to any specific storage element without traversing many others. This was the first practical random access high speed memory. 72 RADIO AGE