776,974. Electric analogue calculating systems; power measurement. ENGLISH ELECTRIC CO., Ltd. March 11, 1955 [March 12, 1954], No. 7261/54. Class 37. An electric multiplying circuit whose output is proportional to the product of the amplitudes of two continuous input signals comprises a generator of a sequence of pulses whose individual amplitude time-integrals are proportional to the amplitude of one input signal and whose recurrence frequency is proportional to the amplitude of the other signal, and means for developing an output signal proportional to the mean amplitude of the pulse train. In Fig. 2, a D.C. motor 20 is energized from variable input voltage e1(αy) over contact 21 of relay 22 energized in series with resistance 23 and normally open contact 24 of relay 25 from positive voltage source E, also connected through the switch and relay 29 to the anode circuit of thyratron 28, whose grid is connected over a limiting resistance to the junction of resistance 34 and capacitance 35 in series across the thyratron anode circuit; the grid being earthed over normally closed contact 37 of relay 25. The anode circuit of thyratron 26 containing relay 25 is energized over normally closed contact 27 from voltage source E, and the grid thereof is connected over a limiting resistance to the junction of resistance 30 and capacitance 31 across a source of input voltage e2(ax), and is connected to earth over normally open contact 33 of relay 29. Initially contacts 33, 34 are open and 27, 37 closed, so that relay 22 is de-energized and the motor is inoperative. Capacitance 31 charges exponentially from voltage source e2 until thyratron 26 strikes 1 after a time interval proportional to -, relay e2 25 is energized, contact 24 is closed, and contact 37 opened. Relay 22 is then energized to close contact 21 and excite motor 20, which runs at a speed proportional to e1 while capacitance 35 charges exponentially from source E until thyratron 28 strikes to energize relay 29 after a time delay. Contact 27 opens to cut-off thyratron 26, and contact 33 closes to discharge capacitance 31. Relay 25 is de-energized, contact 37 closes to discharge capacitance 35 and switch 24 opens to de-energize relay 22 and arrest the motor after a time interval a cycle then repeating. The thyratron time constants are adjusted for linear charging rates and the motor is then energized intermittently by a square wave of amplitude y and constant duration a separated by space intervals inversely proportional to x, so that its angular rotation is proportional to the mean amplitude of the pulse train and approximately to x.y. (Fig. 1, c). When the striking voltage of thyratron 26 is appreciable but less than e2, the charging voltage of capacitance 31 over resistance 30 increases non-linearly and the time elapsing before the thyratron strikes is A/e2 - B (A and B constant), B is adjustable by the time constant of 30, 31 to equality with a, so that the motor is then energized by successive square waves of amplitude y and duration a recurring at a period inversely proportional to x, and its angular rotation is proportional exactly to a, xy (Fig. 1, b), the mean amplitude of the pulse train. In a further modification (Fig. 3), a D.C. motor 40 is energized from variable input voltage el (αy) over switch 41 while variable input voltage e2 (ax) energizes over switch 43 a coil 42 heating bimetallic spiral 45 which in combination with bimetallic spiral 46 responsive to ambient temperature rotates spindle 44 to open switch 43 and close switch 41 when the temperature difference between the spirals reaches T1 and to reverse the switching when the temperature difference falls to T2 after a constant cooling time a. Initially, coil 42 carries a current proportional to e2 and the spiral is heated at a rate proportional to e2<SP>2</SP> until after a time proportional to 1/e2<SP>2</SP> during which the motor is stationary, switch 41 closes and switch 43 opens for the duration of the cooling time, after which the cycle repeats. Motor 40 is thus energized by successive square waves of durations a and amplitude y, separated by time intervals proportional to 1/α<SP>2</SP> (Fig. 1, d), so that its angular rotation is proportional to the mean amplitude of the pulse train and approximately to xy. The device is applicable to a maximum demand K.V.A. meter in which e1 and e2 are respectively proportional to maximum voltage and current. Specification 649,779, [Group XXXVIII], is referred to.