British Kinematography (1949)

85 the second order differential equation of motion of a continuous position control system. In its simplest form, this coefficient is equal to the coefficient of viscous friction, and a system can generally be stabilised either by adding more viscous friction, which, however, causes wastage of motor power and production of what may be an intolerable amount of heat energy, or by the use of velocity feedback, in which case no appreciable power is dissipated. Another way of stabilising a system is to cause the motor torque to be proportional to the error plus a fraction of its first derivative, the so-called " phase advance " technique. However, this may be troublesome in cases where the error tends to exhibit high rates of change. Mr. F. G. Gunn : Has the lecturer had any experience of operating these servo control mechanisms from recorded data ? In film studios, for example, one might want to carry out a complicated manoeuvre with the camera. One would like to have a simple means of rehearsing that, then of recording the optimum condition and being able to reproduce it automatically, as many times as required. The Author : You want to have a control sequence, which in fact entails having the input quantity di a predetermined function of time. One way of achieving this might be by means of a punched tape. Mr. L. Knopp : It might also be necessary to alter the lens focus, hence there will be two things to control at once. Mr. Gunn : For many years we have used Selsyn controls for focus adjustment. The Author : Are they sufficiently accurate ? Mr. Gunn : They are as accurate as the judgment of the operator. He goes through a rehearsal and carries out a certain number of measurements, then interpolates by using judgment. Some directors carry out very complicated movements with the camera, entailing the movement of tons of equipment, and at present it needs a day or days, to rehearse the sequence. If an automatic control could be devised to overcome these difficulties, it would be a very great advantage. The Author : That appears to be a tremendous problem, especially on account of the effects of the big inertia which is involved. A Member : You remarked that the addition of the devices needed to improve performance adds to the expense of the system. The Author : I was not thinking of the monetary expense of the components used in the special devices needed to improve performance, but of the expense involved in extra design time and probably in time spent in laboratory trials which are generally necessary before the best values of the various parameters in the system are determined. The extra devices such as " phase advance," " velocity feedback," etc., which may have to be incorporated in a high performance system, are really quite simple in themselves, particularly in an electrical system. In a mechanical control system, on the other hand, the production of error derivative, for example, would entail the addition of some relatively bulky and precisely made components. In chemical works, pneumatic control systems are almost universally used, since they do not add to fire hazards in the way that electrical systems do. The many precise controls used in refineries, for example, are invariably pneumatic in their operation. The processes of differentiation and integration are done in a pneumatic system by means of capacities and capillary tubes connected so as to form pneumatic analogues of the electrical capacity resistance circuits which will differentiate and integrate. The latest types of precise process controls use proportional plus first derivative plus integral control in order to improve stability and also to eliminate " droop." One of the biggest problems which face designers of control systems is the reduction of the bad effects of time lags in a system, especially finite time lags such as occur in a human operator. Chemical processes also frequently involve relatively long finite time lags which are due to the low speed of certain fluids in pipes. Mr. Peasgood : We have a problem in developing machines where we want to control the temperature of a large volume of solution within very close limits. The Author : Precise temperature control of large liquid volumes is a most difficult problem, mainly because of the time lags associated with the heat capacities of the various elements. A deviation in temperature from the control point, for example, is not registered immediately, while the-; resultant change in the heat supplied to the liquid does not produce a temperature change at once. The electronic aspects of control systems are discussed in a paper by E. Wilson, " Electronics in Heavy Engineering," in J. Brit. I.R.E., Aug., 1949, p. 278.— Ed.