Magnetic drum computers of the 1950s

We were cataloging text materials recently at the Computer History Museum and came across documents describing some early computers from the 1950s. These machines were being sold during the time when “small” machines used magnetic drums as memory. A few years later this type of memory was replaced in the market with core memories.

The Elecom 120 computer was sold by Underwood. Its magnetic drum memory contained storage for 1000 words of memory with an access time of 20 milliseconds. Programs for the machine were written in machine language. The Elecom could also attach paper tape and magnetic tape devices.

Another machine from the same era was the National Cash Register model 102-D. It was similar to the Elecom and also used drum memory.

National Cash Register 102-D computer

DEC archiving project at the Computer History Museum

The Computer History Museum (Mountain View CA) recently completed a large archiving project that consisted mostly of documents and photos donated to the museum by the Digital Equipment Corporation (DEC), which was headquartered in Boston and existed as a company from about 1950 through about 1998. DEC was one of the computer giants of its day, second only to IBM.

Dick and Anna participated in the project for its entire duration, which was about two years. The project contributed 20,000 new items to the museum’s database.

Table set up to do archiving in the Computer History Museum collection
Table set up to do archiving in the Computer History Museum collection
One of the photographs from the DEC collection: participants in a sales training class in the 1980s
One of the photographs from the DEC collection: participants in a sales training class in the 1980s

Dumb terminals at Stanford

In 1976 (before home computers became available) I bought an ADM-3 ASCII terminal kit. The terminals were meant to be educational and workplace tools: a way of connecting students and system programmers, such as myself, with the mainframes at Stanford University and the Stanford Linear Accelerator Center (SLAC), where I worked. SLAC arranged for us to purchase the terminals at a discount.

The kit had over 2000 solder joints and took many hours to complete. When it was finally working, I could log on to the SLAC computer from home using a 300-“baud” (30KB) acoustic coupler modem. Now people think 50KB is slow! Primitive as it was, this terminal saved me many trips to SLAC to fix problems.

However, in our household (and in many others in the Bay Area computer community), the terminal was quickly co-opted by the family teens and preteens for game-playing, which was also pretty primitive at the time!

Above is a picture taken in the late 70s of my daughter Ellen playing a Star Trek game while logged on to the SLAC computer.

The ADM-3 was what was known as a “dumb terminal.”  The display held 24 lines of 80 column text. It had no graphics capability. Imagine playing a game called “Star Trek” (another popular one was “Adventure”) with no graphics! The player had to imagine the background scene (for example, a cave), and who or what was there. Instructions to the computer and the resulting actions (if any) were typed in text.

For example, if the player wandering through the cave had just found a magic lamp, her instructions to the computer might be “Rub the lamp.” The computer might respond with “A door opens.” A common response, which was very frustrating to players, was “Nothing happens.”

In early 1977 Lear Siegler sold the ADM-3 as both a completed terminal and also as a “do-it-yourself” kit. By the end of 1977 the first personal computers began to appear on the market and people started to lose interest in having only a terminal at home.

“Vintage” ADM-3s still exist today. Below is a picture of one stored in the Computer History Museum archives (Mountain View, California), where I saw it in 2014, when I was cataloging artifacts.

ADM-3 terminal in the Computer History Museum collection
ADM-3 terminal in the Computer History Museum collection

Anker-Werke banking machine

One of the more interesting objects in the Computer History Museum’s SAP collection is a banking machine made by Anker-Werke, a German company with an interesting history that used to be headquartered in Bielefeld, North-Rhine Westphalia.

Anker-Werke logo
Anker-Werke logo

The company was founded in 1876 by Carl Schmidt as a knitting machine factory. Besides knitting machines [Nähmaschinen] the company has also made cash registers [Registrierkassen], ticket and reservation machines [Buchungsmaschinen], small motorcycles [Kleinmotorräder], and bicycles [Fahrräder].

Banking machine
Banking machine

After World War II Anker-Werke focused exclusively on office machines. In 1976 the company was liquidated. Part of the business was sold to the British firm Thomas Tilling, and the rest now belongs to technology giant Oracle.

Chuck Fillerup and the IBM 650

One of the items in a recent donation to the Computer History Museum was an IBM 650 computer. IBM announced this machine in 1953, and it used a magnetic drum for memory instead of core.

My first job when I graduated from UCLA in 1965 was at IBM Service Bureau Corporation. My mentor there was Dr. Chuck Fillerup. Chuck taught me a lot about programming while we worked together on several projects.

Chuck told me many stories about how challenging it had been for him a few years earlier to program the IBM 650, since memory was actually a rotating drum. The way a program transferred control from one piece of code to another could have a big impact on performance, especially if it caused the program to have to wait too much for drum rotations. I thought of Chuck when I set my eyes on the 650 sitting in the Computer History Museum warehouse.

IBM 650 computer in the Computer History Museum collection
IBM 650 computer in the Computer History Museum collection

IBM 7094, an ancient mainframe

In 2014, the Computer History Museum got a large donation from Paul Pierce. The donated collection included an IBM 650, IBM 709, and an IBM 7094.

When I first graduated from UCLA, the first machine I was paid to write software for was the IBM 7094. At the time the language being used to program it was Fortran II, since the 7094 was a “scientific” machine. IBM began selling it in 1962 and it used transistors instead of vacuum tubes. When it was new, such a machine cost about $3,000,000.

The image above is part of a donated IBM 7094 sitting in the Computer History Museum warehouse where Anna and I volunteer to help catalog and curate artifacts.