US2294863A - Electrical storage and delay circuits - Google Patents

Description

Sept. 1, i942. B. M. HADFIELD 2,294,863

ELECTRICAL STORAGE AND DELAY CIRCUITS Filed March 29, 1941 2 Sheets-Sheet l B TC2 P CY MR5 5 MR2 4 Q4 B Tea 8 l Sept. 1, 1942. a. M. HADFIELD ELECTRICAL STORAGE AND DELAY CIRCUITS Filed March 29, 1941 2 Sheets-Sheet 2 IZbf' Patented Sept. 1, 1942 ELECTRICAL STORAGE AND DELAY CIRCUITS Bertram Morton Hadfield, Harrow Weald, England, assignor to Associated Electric Laboratories, Inc., Chicago, Ill., a corporation of Delaware Application March 29, 1941, Serial No. 385,956 In Great Britain April 6, 1940 4 Claims.

The invention relates to electrical delay circuits and is more particularly concerned with circuit arrangements for time-delaying by storage a sequence of voltage/time functions.

The series of voltage/time functions may be for instance, a number of pulses of uni-directional voltage, the waveform of which may have a constant or variable amplitude with time but the pulse is distinguished by the time interval between points of zer voltage on the waveform. The voltages which'are delayed by storage in the circuit may be a. function of the pulse amplitude or of the pulse time or both, but their sequence is similar to that of the original pulses. It will be appreciated that by delaying the sequence of any number of pulses it is possible to compare with the voltage due to a pulse with the voltage due to any other pulse whether the latter occurred before or after the former in the original sequence.

According to one feature of the invention the circuit arrangements comprise one or more branch circuits each including a uni-directional device, a condenser in series therewith, the con denser being charged by means of the rectifier in dependence upon the voltage/time function and means controlled by the voltage/time function for discharging the condenser to pass on the charge to subsequent equipment, the arrangement being such that the discharging of the condenser is completed at or before the time of arrival of the next voltage/time function at that condenser.

The unidirectional device may be for instance a dry plate rectifier or the grid/cathode circuit of a thermionic valve and a single branch circuit may be employed in which case the voltage developed by the discharge of the condenser is effective on the controlled apparatus for instance an indicating instrument of suitable type or a plurality of branch circuits may be employed in which case the voltage developed by the discharge of the condenser is effective on the condenser of the next branch circuit while the voltage developed by the discharge of the condenser of the last branch circuit is effective on the controlled apparatus. Further where a plurality of branch circuits are employed, the discharging of one condenser to charge the condenser of the next branch circuit prior to or at the time of arrival in United States Patent No. 2,207,513, issued July 9, 1940. This apparatus is arranged to give a visual indication of the percentage make and break p'eriod and/or impulsing speed of an impulsing contact by simultaneously applying voltages substantially proportional to the logarithm of the make and break periods of the contact to the co-ordinate deflection plates of a cathode ray tube. For this purpose it is necessary to delay the application of the voltages until both the make and break periods of one impulse have elapsed.

According to this feature of the invention, circuit arrangements adapted to respond to a series of direct current impulses-and to render available simultaneously voltages proportional to the make and break period of each impulse successively comprise a first group of branch circuits responsive to voltages proportional to the make periods of the impulses and a second group of branch circuits responsive to voltages proportional to the break periods of the impulses,'each branch circuit including a rectifier, a condenser in series therewith and means for discharging the condenser after an appropriate time interval, the' number of branch circuits in the two groups differing by one and the means for discharging the condensers of the first group being effective substantially at the termination of a subsequent make period of the particular impulse in order to pass on the voltage to the next branch circuit while the means for discharging the condensers of the second group is effective substantially at the termination of a subsequent break period in order to pass on the voltage to the next branch circuit.

of the next voltage/time function at the first The invention will be better understood from the following description taken in conjunction with the accompanying drawings inwhlch Fig. 1 shows the invention in its simplest form,

Fig. 2 shows the invention applied to the delay of voltages representing the make periods of a series of impulses,

Fig. 3 which should be taken in conjunction with Fig; 2 shows the inventionapplied to the delay of voltages representing the break periods of a series of impulses,

Fig. 4 shows the use of thermionic valves for discharging the condensers, I

Fig. 5 shows an arrangement for delaying the discharge of a condenser through a valve,

Fig. 6 shows an arrangement suitablev for use with an indicating instrument having a time lag comparable with the time spacing of the voltage/ time functions and e Fig. 7 shows the invention adapted for use with the target diagram apparatus and employing thermionic valves for discharging the condensers.

Referring now to Fig. l the branch circuit consists of a rectifier MR in series with a condenser C5 and means for discharging the condenser via a resistance Ri. such means "being, for instance. a contact or other switching device S2.

The nature of the voltage waveform .applied to the rectifier and condenser will be, in general, of impulsive type; that is a rapid change from zero to a maximum with a more gradual decay thereafter to zero. The polarity of the rectifier will be such as to pass the voltage to the condenser. In order to explain the action of the circuit element it will be assumed that the input voltage/time function has been established on a condenser C, and that a contact or other switching device SI under the control of the voltage/time function, switches the condenser on to a resistance R. The voltage waveform on R will be of the impulsive type and is applied to the branch circuit. The contact or other switching device S2 in the latter is assumed to be open at the moment Si closes, and therefore, i

a voltage is established on Cl via therectifier MR. This voltage will remain on CI for a time dependent on the time constant of Cl and the back resistance of the rectifier. This time constant can be made very large, so that for all 1 practical purposes the voltage on Cl can be considered as constant, and the circuit has performed the function of storage and delay by a time equal to the original voltage/time function or pulse. When the next voltage/time function or pulse occurs SI is made to open, and S2 is closed. The voltage on Cl then discharges impulsively through RI, and if applied to a further similar branch circuit can be stored again and delayed by the time interval between the first and second pulses. By placing any number of such elements in series, a series of pulses may be obtained in similar sequence to the original but delayed by any time dependent on the time duration of, and intervals between the original impulse series.

The contact or other switching device SI may not be necessary in general as the voltage on R during the establishment of the voltage on C may be of such polarity as not to pass current through the rectifier. It is essential however, that the contact or other switching device in the branch circuit (such as S2) shall open at or before the charging of the previous storage condenser (such as Cl), and that the dissipation of the charges on all condensers shall be substantially complete.

In order to transfer a reasonable fraction of the initial voltage on C to Cl, it is necessary that Cl should not be greater than C and the time constant of Cl with the forward resistance of the rectifierMRshould notbe greater than that of C and R. An approximate idea of the voltage passed on to C2 may be obtained by considering the forward resistance of the rectifier to be zero, when the charge on C is distributed instantaneously between C and Ci, so that the voltage on Cl will be C/C-i-Cl of the original voltage on C. This value neglects the resistances of the condensers, which is in practice justifiable. It is also found 'in practice that even when the time constant of Cl and the forward resistance of the rectifier is comparable with that of C and that the forward resistance is zero can be obtained.

Figs. 2 and 3 show the application of the invention to the target diagram apparatus. In the form of this apparatus described in said prior patent the delaying of the application of the voltages representing the make and break periods of the impulsing contact to the deflecting plates of the cathode ray tube was effected by stepping switches whose bank contacts were connected to condensers on which were stored voltages proportional to the logarithms of the make and break periods of the impulses. By the present invention it is ossible to eliminate the stepping switches thereby removing a source of trouble as regards dirty contacts, to reduce the number and size of the circuit components and to permit the use of low current supply voltages so that a compact mains unit can be used.

The make period circuit shown in Fig. 2 will be first described as being the simpler of the two. A relay X, of the high speed type, is used to repeat the impulses which are to be tested, and has a changeover contact XI. The changeover or moving contact is connected to the positive end of a battery, and the break contact (1. e. normally made) to the junction of the charging resistance Ra: and the condenser C1, the other end of Ra: being connected to the negative end of the battery, and the remaining lead of C1: is connected via a discharge resistance R2 to the positive end of the battery. Thus assuming that the break contactof XI is'broken for the time duration of the make period, C: will charge via a Rs: plus R2 to a voltage which can'be made to represent the logarithm of the make period in the mannerdescribed in said prior patent. The discharge resistance is small compared to R1: so that when the break contact restores to normal the time constant of the discharge of C1:

via R2 is some 2 milli-seconds, compared with some 39 mllli-seconds for the charging time constant of Ra: plug R2 and CI. The voltage waveform of the discharge will be of the impulsive type, and is applied to the branch circuit. consisting of a rectifier MRI connected with its negative pole to the junction of C1: and R2 and its positive pole to a condenser C2 whose remaining lead is connected to the positive battery. To complete the branch circuit the junction of C2 and the positive pole of the rectifier is taken via a discharge resistance P to the make contact of XI. Hence theimpulsive voltage on R! when C2: is discharged, is transferred as previously described via MRI to C! so that the latter remains charged to a voltage proportional to the initial voltage on C1: at the moment of discharge, for a time equivalent to the break period of XI. The next make period will dissipate thls voltage via P and the make contact in readiness for the next charge and discharge of Car. Thus there will be available on C2 a I series of voltage pulses of substantially constant B, some 60% of the value obtained by assuming 75 amplitude proportional to the logarithm of the make periods of the impulses, of time duration equal to the break periods of these pulses, and available at the end of the make periods.

The break period circuit is similar, the charge resistance being now Ry, the charge condenser Cu, discharge resistance R3, rectifier MR2 and storage condenser C3, except for theconnections of a relay contact Y-l on a further high speed relay Y operated from the test impulses. The make contact of Yl is connected to the junction of Cy and R11, so that the charge on C corresponds to the break period, whilst the break contact is connected via a discharge resistance R4 to the Junction of CI and MR2 positive.

Thus the voltage on C2 persists for the make period of YI and is proportional to the logarithm of the break period, and is available after the first break period. Since in most automatic telephone systems the impulsing sequence is a break followed by a make, the voltage representing the break period must be delayed until the end of the following make period before simultaneous use can-be made of the break and make voltages. circuits used in the case of the break period will be one greater than the number used for the make period. The voltage on C2 must therefore be delayed by another branch circuit so that it is available after the first make period. A further branch circuit is connected to R4 therefore so that the impulsive discharge voltage of C2 can be passed on to a further condenser C4 via another rectifier MR3 at the moment when YI restores to normal, that is, after the first make period. To complete the second branch circuit, the voltage on C4 is dissipated at the commencement of the next make period by a further make contact BI and resistance Q, the changeover contact of BI being connected to the break contact of YI. The relay B, of similar type to X and Y, is also operated from the test impulses.

It is now apparent that steady voltages are simultaneously available on C2 and C4 for the duration 'of the break period, and of respective values proportional to the logarithm of the make and break periods of the impulses. They can therefore be applied to the deflection plates of the cathode ray tube over leads X and Y provided a common connection can be found. The positive busbar gives such a connection since the break contact of YI will be made at the time these voltages are required.

It will be appreciated that in practice both the circuits of Figs. 2 and 3 are energised from the same busbars A and B, which form two out of three busbars. The object of this three busbar scheme is firstly to obtain a bias to render the deflection zero when the standard impulses are applied, secondly to obtain an adequate voltage to energise the cathode ray tube and finally to enable push-pull amplifiers to be used.

As mentioned in the general description of the branch circuit the necessity for a contact or switching device preceding the first circuit element, does not occur in this case, as the voltages on R2 and RI during the establishment of the voltages on Ca: and C11 are of such polarity as not to pass current through the rectifiers MRI and MR2. that the charge dissipating contacts in the circuit elements (1. e. XI make, YI break, and BI make) shall open at or before the charging of the storage condensers (i. e. C2, C2 and C4) will automatically be satisfied in the case of XI and YI owing to the inevitable transit times of the contacts, whilst in the case of BI this result can be secured by reduction of its release leg with respect toYI. The latter condition will. in general, be facilitated by theunnecessary ad- Justments to the Y relay in order to make the YI make contact distortionless, since its. release leg will have to be increased to compensate for the transit time.

Hence the number of branch In addition the first requirement dissipation of the charges on the condensers shall be substantially complete, is catered for by making the discharge time constants (i. e. Ca: and R2, C2 and P, Cu and R2, C2 and R4 and C4 and Q) not greater than 2 milli-seconds. As the minimum discharge period will be the make time which may be as low as 10 milli-seconds, this time constant will reduce the original charge to 0.005 of its value in 10 mill-seconds, which is quite satisfactory.

The frequency of operation of the circuit elements may be increased by replacing the contacts by switching devices such as thermionic or gas-filled valves, controlled by the, incidence and polarity of impulsive voltage inputs. Fig. 4 shows a type of circuit arrangement adapted for use with-a voltage/time function of impulsive nature in which a positive half-wave is followed by a complementary negative half-wave. The circuit employs four storage and delay circuits I, II, III and IV in which four thermionic valves VI, V2, V3 and V4 are used, the valves being alternatively switched by the incidence of nega tive and positive wave-fronts applied to the input terminals I and 2. The voltages on the impedance Z, which are to be transferred along the chain of branch circuits, are made to represent the desired voltage/time function in any wellknown manner. The circuit operates in the following manner. The voltage/time function is applied to the terminals I and ,2 and when ter-- minal I is positive with respect to. terminal 2. the condenser Cl is charged. The negative potential on terminal 2, however, makes the grids of the valves VI and V2 negative with respect to the cathode so that these valves are closed i. e. in a non-conducting condition to prevent rectifiers MR4 and MRI in conjunction with condensers CI and 01 from discharging. When terminal 2 is made positive with respect to terminal I, valves VI and V! are opened i. e, made conducting so that condenser Cl discharges through the valve VI and resistance RI2 to cause the charging of the condenser C4. The charge is maintained on the condenser C8 owing to the rectifier MRI and owing to the fact that the valve V2 and also the valve V4 are closed due to the negative potential on the grids of the valve. when the next voltage/time function arrives at terminals I and 2, terminal I is again made positive with respect to terminal 2 so that condenser C4 discharges and condensers Cl and C1 are charged. In this manner successive voltage/tim functions are transferred successively from one circuit to the next. It will be appreciated that the number of circuits may be added to, whilst the output may be taken between terminals 8 and 4, or 4 and I,

- according as to whether an impulsive or substan- The second requirement that the -Il m tially steady output is required.

As previously explained it is an essential requirement of the branch circuits that its switching device shall open at or before the charging of the previous storage condenser, that is, preferably before the closure of the previous switching device. This can readily be achieved in the case of thermionic switches by delaying the application of the positive grid potentials which close the valve, with respect to the negative potentials which open the valve. A typical circuit is shown in Fig. 5, in which voltage wavefronts applied to M and N, and making I! positive to N are only applied to the grid/cathode path of the valve after a time dependent on the time constant r2 and cIl' (the rectifier J being when the voltage wavefront makes M negative to N, the greater portion of it is immediately applied to the grid via Ml, across the rectifier and resistance H, the latter being needed so that cill may discharge during the negative pulse in readiness for the next positive pulse.

It will be noticed that in all the foregoing description of the action of the branch circuit, the voltage pulses available at the end of a chain of circuits'only last for either the duration of the original voltage/time functions or the time interval between such voltage/time functions, according to the number of circuits employed. In certain cases where the indicating instrument has a time lag, such as for instance an ammeter, the desired readings may be confused by the motion of the instrument tending to, or returning from zero- In such cases the pulse time, or time interval between pulses, is comparabl with the time lag of the meter, and the terminating circuit element may be switched as shown in Fig. 6. The voltage/time function is applied to Al and BI and the desired representation of it in the form of an impulsive voltage is generated across the impedance Zl in th normal manner, and transferred via any number of branch circuits between E and F to a low impedance source L which may be momentarily shorted without affecting the input voltage. A typical form of such source is a cathode follower valve with a high resistance in the grid lead; in this case L represents the cathode resistance. The pulse voltages on L are passed on to C9 via rectifier MR8, and are available at terminals '1 and U for operation of the meter. The thermionic valve V5 is arranged to short C9 only for the time necessary to dissipate substantially its voltage via the internal resistance of VI. This time depends on the time constant of C9 and V5 and may be made a fraction of a miili-second, if desired, In order that this time may not, be exceeded, the positive grid pulse is differentiated by the grid resistance Gi and condenser Kl, so that a very short positive grid pulse is obtained; it is obvious that the time constant of Ki and GI must be of the same order as C9 and V5, or higher, depending on the form of the applied voltage/time waveform. Both the condenser C9 and impedance L will be shorted momentarily, therefore, and hence the need for a source which can be shorted without affecting the input voltage.

When the short is removed, C9 will charge up again to the new input pulse voltage, and hence the cut-put across '1 and U will consist of a series 01" voltages of nearly the same amplitude as would have formerly been obtained, but separated only by a very short time gap. The meter will not, of course, respond to these gaps and the indication will therefore be much steadier. A source of grid bias, such as a battery 132 will in :e erel L:- needed. so as to prevent the valve from n'ghest voltage desired on on to the grid condenser 11;," point in the preceding Wavefront exists with at the moment a to L. It be end of C3, in oer n3 effect over some 200 milli-seconds being effected by an electrical circuit, instead of by the meter movement. A quicker and less tiring estimate of the input level is obtained.

It is obviously practicable to use thermionic switching, as described above, for the target diagram apparatus, with the advantage that the relays can be eliminated together with the necessity for special adjustments to secure distortioniess operation. Fig. 7 gives a typical circuit arrangement, in which the main components used for measuring the break and make periods and the storage and delay groups are lettered to correspond with Figs. 2 and 3.

The impulsing contact TI to be tested, is con nected via a positive battery B3 to the input resistance R33 of a valve V6, so that closure of TI increases the anode current, grid current being limited by R32. The make period charge condenser Cr is thereby caused to discharge'from the previous anode voltage towards the much lower voltage on V8, due to the increase of anode current, via resistances R2, Rx, R23 and V (the rectifier MRI! being non-conducting). Of these resistances it is arranged that R1: shall have predominating effect. At the termination of the make period, V6 becomes a very high resistance, so that C: charges from the busbar voltage and via R22, R2 and the forward resistance of MRI2, within say 2 miili-seconds. The resultant voltage pulse on R2 is positive to the grid of V1 and of magnitude corresponding to the logarithm of the make period in the usual manner. V1 is designed so that it acts substantially as a cathode follower, so that the voltage on R25 is in phase and of approximately the same magnitude as the input grid pulse (grid current produced by the fortuitous long make periods being limited by RSI). The voltage pulse on R25 is applied to the usual circuit MRI, C2, and discharge valve Vl, so that the make period voltages are available between terminals 6 and l. The switching of V! at the beginning of the next period is obtained by applying the impulsive voltage change on the anode resistance R24 of V1, due to a negative pulse applied to the grid across R1: and R2 during the discharge of CI, to the grid of V8 via a condenser CI2 and resistance RIB. Provided the latter time constant slightly exceeds that of C2 and V8, and the latter is not greater than some 2 milli-seconds, C2 may be thus substantially discharged in the minimum pulse time of 10 milliseconds. A grid bias BI is needed for V8 sumcient to prevent conduction when C2 is charged to the maximum test voltage.

The circuit for the break period is of similar nature, except that two branch circuits are used, and the discharge valve Vii must include a resistance RlB, in order to pass the voltage on to the second element. The discharge of C4 is eiiected by applying a positive pulse to the grid of VI2 from R29, since at this moment Vii is nonconducting. The input charge and discharge valve V9, corresponding to V6, is operated from the latter by a. high resistance potentiometer R24, R25, since Cy is required to discharge via Ry at the end of the make period, 1. e. at; the beginning of the break period.

It will be noticed that the difficulty previously mentioned when using thermionic discharge valves, that valve shall open at or befoie the charging of the plBViOLlS storage condenser, has been complete,- overcome without using the cir cuit shown in Fig. 5, making the discharge eiifect impulsive, so that the previous gondenser is discharged and open clrcuited before the next pulse. Thismethod is of course superior to the former method, since it can invariably be used whatever the nature of the available voltages for causing the discharge valves to function.

In certain circumstances, the condensers C2 and C! may be placed directly in the cathode leads of V1 and VII, thus eliminating RII, R", MRI and MRI.- I'he grid/cathode path takes the place of the rectiflers. This method has the advantage of giving a greater voltage output on 1 C2 and C3, by causing the condensers to charge from the supply voltage during time grid current iiows. MRI can likewise be replaced by a valve with C4 in its cathode lead. 1

It will be appreciated that since the circuit in Fig. 7 measures the halt-cyclic periods of the,

input waveform, it may be used as a frequency meter, by designing it for operation at the highest required irequency, and causing VI to act as a "peak chopper." By using a cathode ray tube as the indicator, the "instantaneous Irequencyoi' theinputmayberead.sothat, for instance, the rrequency spectrum oi speech or music may be displayed; The remaining co-ordinateaxisoi'thetubemaybeused forthe peak value or time duration of the input.

I claim:

1. A circuit arrangement comprising a main circuit in which direct current impulses are generated, a series or branch circuits ted therewith, certain of said branch circuits having means responsive to the make period 0! mid impulses to produce voltages proportional thereto, and others of said branch circuits having means rsponsive to the break period'oi' saidimpulses to produce voltages proportional thereto, each branch circuit including a rectifier and a cond nser in series and means for the condenser after a certain tune interval; said meansforthecondensersintheflrst group 01' branch circuits being eflective substantiaily at the termination of a subsequent make period to pass on the voltage to the next branch circuit, the means for discharging the condensers in the second group of branch circuits is eii'ective at the termination 01' a subsequent break period to pass the voltage to the next branch circuit.-

2. A circuit arrangement as claimed in claim 1, in which the means for discharging said condensers comprise thermionic tubes each connected across one of the condensers, the voltage of the grids of the valves being controlled by-the impulses to render the tubes conducting or non conducting to control-the discharge or the condensers.

-3'. A circuit arrangement such as claimed in claim 1 in 'whichthere are thermionicvalves for controlling the discharge oi. the condensers. means for placing positive voltage on the grid of the tubes and for then placing negative voltage thereon, and mean for delaying the application or positive voltage to a tube to ensure that that tube is made conductive at or beiore the previous tube is made non conductive.

4. A circuit (arrangement for simultaneously obtaining voltages proportional to the make and break periods or an impulsing contact including a main circuit having the contact therein and two groups of branch circuits associated therewith, a pair of thermionic valves associated with the main circuit, a condenser and a uni-directional device in each branch circuit, means for causing the first one of said valvesto pass current to cause charging of the condenser in the first branch circuit oi! one group when thecontacts make, said second valve having an input comprising a potentiometer connected across the first valve, and means tor causing the second valve topasscurrent-tocausecharsingofthe condenserintheflrstbranchcimiitortheother the break of said contact. BERTRAM MORTON HADFIELD.

Images