First-Hand:Banging the Large Drum Slowly

From ETHW

Contributed by: William Merton Nellis, IEEE Life Member

In January 1953, I reported for work at Remington-Rand Univac in St. Paul, Minnesota. I had been interviewed for a job at Engineering Research Associates (ERA) several months earlier while still in the Navy but prior to reporting for work the company became Remington-Rand. I worked on a project for the Navy that included recording on large drums as a means of delaying analog signals.

The surface of a drum was coated with a magnetic oxide that could be magnetized with a noncontact boundary displacement recording head. The drum surface was precise and concentric so that the heads could be placed within 1 mil of the surface. The drum was driven at a rather slow constant speed so that the transport time from recording head to pickup head was seconds and adjustable by changing the distance of the pick up on the track. A large drum with approximately ten tracks processed signals from ten hydrophones using this phasing method to turn the hydrophone array into a beamed listening system.

When I recently Googled "boundary displacement recording," it returned an article about ERA and boundary displacement recording that was in Modern Mechanix in January 1953.Another project in 1954 included the ore car weighing system installed at the Great Northern Railroad iron ore docks at Allouez, Wisconsin. A strain gauge track-scale weighed the slowly moving ore cars to obtain their gross weight. The analog gross weight was digitized and used with the ore car tare weight that was held in a digital memory bank and accessed by an operator entering the ore car number visually as the car passed.

The system subtracted the tare weight from the gross to get the net weight of the ore in the car. A flexowriter then typed up the whole transaction to be used in billing for the transported iron ore.

Two things were memorable about this system; First, the tare weight digital memory system for the empty weight of 10,000 ore cars consisted of two six-foot relay rack cabinets full of cross-bar relays. Today a thumb drive or I.C. would do the job. Second, the calculations were done by a mechanical adding machine automated by keyboard solenoids. Today a microprocessor would do that and more!