Relays from the era of the IBM ASCC (Harvard Mark I)

Home Up	Contact Legal Stuff

Just a couple of relays

My collecting tends to focus on representing technological trends in time; and the farther back in time you go, the harder it becomes to have an unbroken timeline. I didn’t even think I’d find anything from the brief period of electromechanical computers...

But I did: I got these two relays.
They’re nothing much to look at, just a couple of electromagnetic switches the likes of which were used extensively in all sorts of equipment throughout the 20th century, notably in telephony apparatus. But these two are special. They were built specifically for computing -- they have “IBM” stamped right into their metal frames. What’s more, the instruction manual for IBM’s first relay-based computer -- the Automatic Sequence Controlled Calculator (ASCC), more often

Click photo to enlarge

remembered as the Harvard Mark I -- shows photos of practically identical components. It would be silly to expect these items to be from the Mark I itself, but they could certainly have served it, or other IBM gear from the same decade, as spare parts. They fill the hole in my timeline as fair representatives of the brief era in the 1930s and 40s when relays were used to construct a few lumbering computers just before vacuum tubes ushered in the era of electronic computing.

A relay consists of an electromagnet and a set of switch contacts; when current flows in its coil the electromagnet pulls the contacts to close the switch. This allows one to control current flow in one circuit by passing a current in another -- precisely what we do today with the transistors in our silicon chips. Thus, interconnecting these relays allowed the early pioneers of modern computing to construct the logic circuitry required for building a computer. This idea occurred to a number of innovators in the 30’s, notably Konrad Zuse in Germany and Howard Aiken in the US. Aiken, working with IBM and Harvard University, completed the ASCC in 1944. It was a monstrous machine: 15.5 meters long, 2.5 meters high, and weighing over 4 tons, it contained close to 800,000 components. Although utterly underpowered in any modern sense, it was the largest computer built up to that point, and possibly the first to be put to “serious” use.
The two relays I have are configured as 12 Pole, Double throw and 4 pole, Double throw switches.

Their construction is elegantly straightforward, even frugal; the fixed contacts are made of plain bent metal strips that feed right through to become the external connection terminals. Both come with sockets -- clearly, a machine as huge as the ASCC would require this to allow the fast replacement of failed parts...

Click a photo to enlarge

		Exhibit provenance: An online antique dealer in the US. More info: - ASCC information from IBM. - ASCC Operation manual. - A present-day Relay Computer built by Harry Porter of Portland State University.

Just a couple of relays