|
| |
Manchester Differential Analyser
From: Science Museum
| By:
Peter Turvey |
EDITOR'S INTRODUCTION |
The differential analyser was an important achievement in the development of modern computing. Its creation enabled researchers to work out complex differential equations through a series of graphs. This feature looks at the Manchester University differential analyser, which was based on the design of Massachusetts Institute of Technology's (MIT) first ever differential analyser and helped contribute to contemporary computer designs. |
 any problems in science and technology can be described by differential equations; in the 1920s and the 1930s, early analogue mechanical computers, 'differential analysers', were built to solve them. Modern digital computers are essentially counting devices which can handle numbers, but these differential analysers solved equations by modelling mathematical relationships with shafts and gears. |
This reconstructed section of the Manchester University differential analyser of 1935 is one of the surviving relics of this era in computing. It is about half the size of the complete machine, built by the local engineering firm of Metropolitan-Vickers under the supervision of Professor D. R. Hartree (1897-1958). He first used it to solve problems in atomic physics, and during the Second World War it was used for military research, including work on the cavity magnetron. |
Its design was based on the first differential analyser, completed in 1930 at the Massachusetts Institute of Technology by a team led by Vannevar Bush (1890-1974), an electrical engineer who later marshalled America's scientific effort in the Second World War, and also played a key role in the decision to drop the first atomic bomb. |
The drive to develop the differential analyser came from the growing telecommunications and electricity-supply networks of the 1920s. To design power-transmission networks which would not be liable to blackouts, engineers had to solve complex differential equations which could take months of painstaking work. Bush realised that 'better ways of analysing were certainly needed'. His solution, like that of Charles Babbage one-hundred years before, was to mechanise the process of calculation. In the 1920s the only feasible way of making such a computer was to use mechanical rather than electrical technology. Relatively complex analogue mechanical computing devices had been developed to aim the guns aboard battleships, but Bush's machine was the first civilian analogue mechanical computer. The idea of the differential analyser had been put forward by the British physicist and entrepreneur Lord Kelvin in 1876 (he did not attempt to build one), but Bush's team did not know of Kelvin's ideas until after their differential analyser had been completed. |
The Massachusetts Institute of Technology differential analyser was widely copied in Europe as well as in America. It was the best means of solving differential equations until the development of the electronic computer in the 1940s. During the Second World War, differential analysers were chiefly used to compute firing tables for guns, as well as to design radar control systems and electronic systems and electronic circuits. However, setting up a differential analyser could take several days. Although hybrid, part-mechanical, part-electrical differential analysers were built during the war to speed up the process, in the end they were a dead-end technology, made obsolete by electronic digital computers. |
Hartee wrote that 'my first impression on seeing the photographs of Dr Bush's machine was that they looked as if someone had been enjoying himself with a super-Meccano set'. Hartee began trying to build a Meccano model 'more for amusement than with any serious purpose', which was so successful that, with the help of a student, Arthur Porter, he built a small differential analyser using many standard Meccano parts. It was capable of useful work, and gave good practice in 'programming' whilst the full-size analyser was under construction. |
The differential analyser was a purely mechanical device -- an 'analogue computer' which worked by representing each element of the equation by a rotating shaft. Gearing allowed the shafts to be set to rotate relative to each other, reproducing the relationship of elements of the equation. Experienced operators could follow the mathematics of solving an equation by watching the turning shafts. |
The first step in using a differential analyser was to set up the shafts to reproduce the equation to be solved, a time-consuming operation. Graphs were prepared showing how changing variables in the equation altered with respect to one another, for example in the case of a train's speed changing with time. These graphs were placed on input tables, and as the analyser ran, the curves were followed by human operators who moved pointers geared to the mechanism, feeding in the varying values. |
Output tables on the analyser produced graphs showing the solution to the equations set up on the differential analyser; in our simple example, the output would be a graph of distance travelled by the train versus time. |
|
| |
