GB620140A - Improvements relating to d.c. amplifiers - Google Patents

Abstract

620,140. Valve amplifying circuits. BRITISH THOMSON-HOUSTON CO., Ltd., and OWEN, C.E. March 20, 1946, No. 8658. [Class 40 (vi)] Zero drift in a D.C. amplifier is reduced by associating a series condenser with the input of the amplifier which is charged from the output of the amplifier to a potential which is substantially equal to but of opposite polarity to a voltage representative of the zero drift in the amplifier. In the D.C. amplifier, Fig. 1, when switch S1 is closed, a correction voltage is fed back to charge condenser C1 via a switch S2 which closes at approximately the same time as S1 but opens just before S1 opens, the resistor R3 enables the amplifier to function in the event of a failure of the switches. A modification, Fig. 3 (not shown), enables improved correction to be obtained by applying the correction to a D.C. amplifier having an overall negative feedback. Fig. 5 shows an arrangement to prevent the output of the amplifier falling away during the period when the input is shorted by charging a condenser C2 via switch S3, so that during the period when switch S1 is closed, S3 is opened to isolate C2 from the output of the amplifier. S3 opens just before S1 closes and closes just after S1 opens. An isolating amplifier is as shown provided between the condenser C2 and the output and negative feed-back is preferably taken from the output of the final stage to the input circuit of the amplifier. Fig. 8 shows the first stages VI, V2 of a D.C. amplifier, a voltage being obtained at a point D which bears a substantially constant relationship to the zero drift voltage of the amplifier. The voltage at D is fed via contact S which opens and closes periodically to a corrector unit, the output stage V5 of which feeds condensers C1 and C2 through two reversely-connected diodes V6 and V7, so biassed as to prevent any charge on Cl leaking away. The correction amplifier gain is of such a value that a change in voltage at the anode of V5 is produced equal to or greater than the bias on either diode when S1 is closed and the zero drift error in the D.C. amplifier is the largest that can be tolerated. Condenser C3 charges to a potential proportional to the zero error when S1 opens, when S1 closes C3 is discharged by R10 and either V6 or V7 conducts, depending on the sign of the zero error voltage and a voltage developed across R11 is applied as negative feed-back to the input of the corrector amplifier, the charge acquired by C1 is substantially proportional to that lost by C3, hence the voltage rise of Cl is proportional to the residual zero error of the D.C. amplifier. C2 charges from C when S opens and by choosing suitable values of C1, C2, R10 and R11, the change in potential of C2 due to an operation of S1, is made substantially equal to the residual zero error of the amplifier before S1 operates. As the zero drifts a correction is put in at each operation of switch S1; the combination of R9 with C4 and R8 with C2 smooths out the corrections to obviate sudden changes in the output level. One corrector unit may be used to compensate for the zero drift errors in a number of D.C. amplifiers. The corrector unit being switched from one amplifier to the next by rotary switches, Fig. 9 (not shown).