Study on time-shared analog computation technique

Abstract

Methods of time-sharing computing elements or cir cuits to achieve economy in analog computing are studied.

The conventional method of time-sharing is to scan inputs and outputs with synchronized switches and to hold the computer solutions after each scan ning. This method is practical only when the scanning interval is short enough compared to the period of solution to reduce error to an acceptable level.

The method presented here allows elimination of hold circuits and keeps good accuracy even with long scanning intervals. In particular, this method is shown to be without phase lag, and this is confirmed by experiments.

For the conventional method, the equivalent dead time (the predominant source of error) is calculated, and several compensation techniques are presented.

Comparisons of all these methods are also made.