Delay line storage in EDSAC - pick your poison

As the name EDSAC (Electronic DELAY STORAGE Automatic Calculator) suggests, the 1949 computer depended upon delay lines for both the main store and for the internal registers of the machine.

In the original EDSAC, the delay lines were constructed from long steel tubes ('tanks') filled with mercury. It was not practical to replicate this in the reconstruction because of cost, safety and reliability, so a nickel delay line using steel wire as the storage medium was built.

Delay line technology was used in many early computers of the 1950s, but the engineering knowledge and skills required to build nickel delay lines have been lost for many years.

In this video, one of the EDSAC volunteers, Professor Peter Linington, explains how he researched the construction of nickel delay lines and set about manufacturing them for the EDSAC reconstruction. His delay lines are now fully working and ready to connect to the rest of the machine, replacing the temporary modern 'semiconductor delay lines' that they have been using during the commissioning phase of the project.

Turing and gin delay lines – a rum suggestion?

When delay lines were being considered in the late 1940s, Alan Turing is said to have advocated the use of gin, which might contain just the right amount of alcohol and water, as the delay line medium.

Technically, it might produce the necessary zero-temperature coefficient of propagation at room temperature.

The gin story is a little hazy and perhaps Turing's proposal was more in jest than reality. Others may have highlighted the claim for their own interests – or for a little mischief.

Nonetheless, Peter Linington recently looked into the issue and concluded that perhaps sherry would have been more appropriate than gin - although the extra sugar might cause problems!

On balance, Peter thinks that gin would not be suitable from an engineering standpoint. Gin might have transmitted sound a bit too quickly; would have had too high an attenuation rate, leading to high power loss; has a lower density than mercury so less effective acoustic matching; has an unknown rate of sound dispersion (gin is quite a complex liquid); and its interaction with other elements of the container are untested. However, gin might have had one advantage: it would wet the transducer to give more reliable operation.

Some of today's EDSAC team preferred other uses for gin in the reconstruction process.