Relay computerswhy so few 13

Peter Flbutt

I don't know about power consumption. As mentioned, certain relays only use power during actual switching. Other relays can operate on very low levels of power. I don't know how that compares to the tube filaments. Tube computers needed lots of ventilation or air conditioning; telephone central offices equipped with relays did not and I don't know of any heat problems with relays. (Researchers on early Univacs worked in their underwear in the lab because of the heat and had to be reminded to put Clothes Line on if they went to the front office. IBM's labs were air conditioned and dress was much more formal--they had to put on stiff collars when they went to the front office.)

Relays are far slower than tubes. That's one reason they weren't used after tube computer electronics was perfected. My question refers to the time before tubes were perfected in that application. However, keep in mind that relays remained cheaper than electronics for many years, at least until the 1960s

However, relays are far more reliable than tubes. Tubes burn out. In large tube computers, they had to be tested daily to replace all burned out tubes. Also, IBM found that in digital applications tubes had to be free of impurities. Impurities wouldn't be noticeable in audio applications but would cause a false bit in digital.

By 1940 the telephone industry (like the railroad industry) had a lot of experience with relays and knew how to build them to be very rugged and reliable over a long period of time. Bell Labs already knew all sorts of things such as the optimum metals for contacts, spacing, metals for the spring part, and coil design. A relay could be designed to act fast or slow.