US2275023A - Thermionic valve circuit arrangement - Google Patents

Description

E. L. c. )NHITE THERMIONIC VALVE CIRCUIT ARRANGEMENT March 3, 1942.

Filed May 9, 1940 5 Sheets-Sheet l INVEN TOR. E. LC. WHITE MKM ATTORNEY.

March 3, 1942..

E. L. C. WHITE Filed May 9, 1940 THERMIONIC VALVE CIRCUIT ARRANGEMENT 5 Sheets-Sheet 2 BYv INVEN TOR. E. L. C. WHITE ATTORNEY.

ATTORNEY.

E. L. c. WHITE 2,275,023

THERMIONIC VALVE CIRCUIT ARRANGEMENT Filed May 9, 1940 5 Sheets-Sheet 3 INVEN TOR E. L. C. WHITE 75% n I In): 1111 44 44444 1.- 444444 March 3, 1942.

March 3, 1942. E. L. c. WHITE 2,275,023

THERMIONIC VALVE CIRCUIT ARRANGEMENT Filed May, 9, 1940 5 Sheets-Sheet 4 HI .1. ..L 4J L 24 L R8? Q g INVENTOR.

D '68 9b 5.11.0. H/TE ATTORNEY.

March 3, 1942.

E. L. c. WHITE THERMIONIC VALVE CIRCUIT ARRANGEMENT Filed May 9 1940 5 Sheets-Sheet 5 T Fi 30 All INVENTOR.

E. 1.0. WHITE )KZW ATTORNEY.

Patented Mar. 3, 1942 2,275,023 F F l C E.

THERMIONIC VALVE CIRCUIT ARRANGEMENT Eric Lawrence Casling White, Hillingdon,

Eng-

land, assignor to Electric & Musical Industries Limited, Hayes, Middlesex, England acompany of Great Britain Application May 9, 1940g-Serial No. 334,157 In Great Britain May 10, 1939 13 Claims.

This invention relates to thermionic valve circuit arrangements, and has particular reference to coupling networks between valves intended to pass with substantially equal attenuation all frequencies extending over a wide range from zero or direct current.

Coupling networks between valves designed to transmit direct current include a direct current path between the anode ofone valve and the control grid of the next, andthe problem pre sented by such networks is todesign them so as not to discriminate between difierent frequencies, to provide for decoupling of the anode circuit of the first valve, and also to allow for a difierence in potential between that anode and the following control grid suitable for their respective working ranges. Further, the direct current path usually includes a resistance shunted by a condenser which should be of large value compared with the stray capacities represented by the grid of the second valve so -that a resistance connected in the anode lead of the first valve, the value of which is determined by these stray capacities shall be effective on the grid of the second valve, apart from the effect of any circuit elements such as coils, introduced between the anodeand grid for the purpose of boosting high frequencies.

The specification of U. S. Patent No. 2,120,823 describes a direct coupled cascade amplifier for a wide range of frequencies in which the anode load impedance-and anode impedance of a first valve can together be represented by a pure resistance R in series with an impedance Z and in which the first valve is coupled to a second valve through an impedance Z and a leak resistance R is provided between the grid and the cathode of the second valve, the relationship between these elements being-such that the condition The arrangementdescribed in the patent above referred to provides a solution which is applicable to triode valves which have finite anode slope resistance and also to pentodes which have practically infinite slope resistance, but the patent does not describearrangements for meeting the condition which arises when grid bias decoupling arrangements are employed.

The U. S. Patent 2,243,121 also describes a coupling arrangement for amplifiers operating over wide frequency bands with particular reference to and emphasis upon the problem of avoiding phase distortion particularly at low frequencies. In the arrangement described, the grid circuit of a valve coupled to a preceding valve has its constants for coupling capacity resistance and grid circuit resistance, as well as a grid bias filter, made to have a similar characteristic to the anode circuit includingztheanode decoupling resistance and condenser of the: preceding tube. By making the anodeand grid circuit similar, phase distortion is substantially eliminated. Thus, while in the arrangement described in above Patent No. 2,243,121. account is taken of the grid bias decoupling arrangements, the solution applies only to triode valves.

The object of. the present invention is toprovide a coupling circuit from the anode of one valve to the grid of the next valve which will pass all frequencies down to zero with equal amplitude and substantially no phase shift, and which takes account of the decoupling arrangements for the anode supply and for the grid bias, the coupling circuit being applicable to valves of the triode or pentode type.

According to the present invention an electrical network for coupling a constant-current source of signals to a high, impedance device comprises a pair of branches iii-parallel and fed with current from; said source, thefirst of said branches comprising two parts connected in series, one part being formed by a resistance of magnitude P shunted by an impedance of value a, the other part being iormedby a resistance of magnitude (1?, the second of said branches comprising three parts connected in series, a first part formed. by a resistance of magnitude Q shunted by an impedance of value p, a second part formed by a resistance of magnitude M33, and a third, part formed by a resistance of magnitude cQj shunted by an impedance of value 0,3, the quantities a, b, 0, being real numbers and satistying the relation: a=b'+c and the quantities or, n; being complex or satisfying the relation:

aP bQ I? and it is arranged that the sum of the potential differences developed across the second and third parts of said second branch are fed to said device and the network is so constructed that it transmits said signals over a wide range of frequencies substantially independently of the fre quency. The magnitude Q may be large compared with the magnitude P but in a network in which the magnitude Q is not large compared with the magnitude P a resistance of magnitude W T is connected so as to form a. third branch in parallelwith said pair of branches. In a modification of a network embodying the last mentionedfeature, the resistance of magnitude purely imaginary and' 2% is arranged to be present effectively by modifying the magnitudes of the elements of the second branch of the network.

In a further modification the resistance of magnitude is arranged to be present effectively by modifying the magnitudes of the elements of the first branch of the network.

The source of signals may be constituted by a thermionic valve or high impedance such as a valve of the screen grid or pentode type, but in a modification the source of signals is constituted by a thermionic valve of low impedance such as a valve of the triode type and the magnitudes of the elements of either or both of the first and second branches are modified to allow for the finite impedance of said valve.

In particular forms of network embodying the invention the impedances of values oz, {3 and 0 3 or of values modified therefrom are capacitative reactances, and those having the values on and 05, or values modified therefrom serve the purpose of decoupling.

In a modified electrical network embodying the invention, the impedance of value ea or of value modified therefrom, is in conjunction with the resistance of magnitude P, or magnitude modified therefrom, the electrically equivalent circuit to a network comprising series resistances and shunt condensers serving for the purpose of decoupling.

In a further modification the impedance of value 05 or' of value modified therefrom is in conjunction with the resistanceof magnitude cQ or magnitude modified therefrom the electrically equivalent circuit to a network comprising series resistances and shunt condensers serving for the purpose of decoupling.

Any of the parts of the first and second branches formed by shunt combinations of impedance and resistance may be the electrically equivalent circuit to a network transformable to said equivalent circuit in accordance with the transformation equations to be referred to. Further modifications of networks embodying the invention and circuit arrangements including valves coupled by such networks and intended for operation over wide ranges of frequency in- I eluding zero frequency or direct current and one arrangement for alternating current operation will be described in detail.

In order that the invention may be more clearly understood general theory underlying the alternative methods of designing networks in accordance with the invention will first be described and will be followed by particular examples of circuit arrangements including such networks.

Reference will be made to the accompanying drawings in which:

Figures 1 and 2 are explanatory diagrams,

Figures 3 to are examples of network equivalents,

Figures 16 and l? are further explanatory diagrams,

Figures 18, 19 and 20 are examples of circuit arrangements including particular forms of network,

Figures 21, 22 and diagrams,

Figures 24 to 29 are further examples of arrangements including forms of network derived in accordance with an alternative method, and

23 are further explanatory and readily carried into effect, the

Figures 30, 31 and 32 are examples of circuit arrangements including more complex elements.

Referring to Figures 1 and 2 of the drawings these show networks which are entirely equivalent if the impedances Z1, Z2, Z3, Z'1, Z'2 and Z3 transform the one to the other according to the transformations In particular it is to be noted that if Z1 and Z2 are real (i. e. resistances) Z'1 and Z'2 are also real (i. e. also resistances), while Z'3 is identical with Z3, except that it has been multiplied by a real number (i. e. it has identical structure and form, but individual elements may have difierent values). Again, if Z1 and Z2 are purely imaginary and of like nature (i. e. like reactances) Z'1 and Z'z are also purely imaginary and of like nature (1. e. like reactances). Thus the networks shown in Figures 3 and 4 can be chosen to be entirely equivalent; likewise the networks shown in Figures 5 and 6 can be chosen to be entirely equivalent. In the same way, the networks shown in Figures 10, 11, 12, 13, 14 and 15 are all equiva lent. Conversions of this nature are explained in detail in the book entitled Transmission Circuits for Telephonic Communications by K. S. Johnson, 1935 edition, pages 267 to 280.

In these equivalences it is important to observe that a decoupling network, such as that shown in Figures 3, 7 and 10, can always be put in the form of two resistances in series and one of them shunted by an impedance. Thus the network shown in Figure 7 can be put in the form of the network shown in Figure 8 or 9, that shown in Figure 10 can be put in the form shown in Figure 13, 14 or 15, to which equivalent networks, other networks may be made to correspond by changing the order of the elements in a given branch.

The general coupling network underlying the applications of the invention can thus be drawn as shown in Figure 16 in which P, Q are resistances, a, b, c are real numbers and a and B are impedances of identical analytical form, but related by a multiplying factor which is a real number. This network is supposed to be driven from a constant current source and if the source is only of finite impedance it is supposed to be replaced by an infinite impedance source and its impedance efiectively shunted across the network input by absorption in the resistances (1?, P. Moreover a, b, c satisfy b+c=a, and on and ii are related according to B 01 These relations form the basis underlying all applications of the invention to coupling networks.

It is assumed that the P-branch is the anode circuit of a driving valve, the bQ, cQ, op part of the Q-branch is the grid circuit of the driven valve and the Q5 part of the Q-branch is the series coupling element between these circuits, the P, and 0Q, 05 parts of which may be interpreted as decoupling networks according to any of the equivalences described above. The decoupling is not limited to first, second and third order as shown.

Separate conditions arise when'Q islargecoinpared with P and when Q isnot large compared With'P. When Q is large compared with JP, it is easily shown'that the potential fed to the grid of thesecond valve due to a currents flowing into the networkfrom the first valve, namely the potential:

bQ+- i fi +n n+ )BQ B+Q reduces simply to zaP, which is independent of frequency.

When Q is not large compared with P, the above calculation of grid potential is not'valid unless it so happens that the presence of the Q branch only modifies the current in the P branch in a constant ratio. In the circuit of Figure 16 this does not happen.

As described, networks including combinations of resistive, capacitative or inductive elements may, subject to the satisfaction of certain conditions, be each transformed into electrically equivalent networks in which the values and/or disposition of the elements may be varied. Thus, the arrangement of Figure 16 may be transformed to that shown in Figure 17.

With reference to Figure 17, it is easily shown that the impedanceof the last branch is:

b m-% (a-l+a-P+1+a -a) so that the ratio of the fourth branch to the second branch is Li 1+0 aP thisisnot the ratio of the third to the firstnamely The last mentioned ratio can be transformed to:

however, by shunting across the third branch a resistance of magnitude may be left across the network; it may be ab sorbed into the first branch so modifying all the elements in the P-branch or it may be absorbed into the third branch. Any combination of these possibilities is allowable, of course.

Referring now to Figure 18 of the drawings, this shows an example of a coupling arrangement between two valves in which anode and gird bias decoupling networks of the first order are employed corresponding with Figure 3. The valve I is provided with an anode resistance 2, the P element, decoupling resistance 3 and decoupling condenser 4, the or? elements, and is coupled to the grid of valve 5 or other high impedance point, such as a modulatingor deflectingiielectrode of a cathoderay tube, through-the flQelements constituted'by condenser B and resistance"? in parallel. A bias potential from a battery gbdsappliedto the grid of the valve 5 through ia'leakresistance 8, the bQ element, a decoupling. resistance 9 and decoupling condenser IO- the cc elements,ibeing.iprovided.

Figure #19 shows two valves coupled by an arrangement'inclu'ding decoupling networks of the second ordercorresponding with Figures 7, 8 or 9. It will be seen that the impedance 0: of'Figure -16- assumedto be a condenser, has in Fig-. ure -l9 beenreplaced by a condenser II in series with a further'condenser lzshunted by a resistance l3. The BQ elements in Figure 19 are constituted by two condensers l4 and I5 in series, a resistance' lfi being shunted across the condenserliand a further resistance 11 shunted across-the combination. The bQ element is constituted by aresistance l8 and the 0 3 elements bytwo condensers l9 and 20 in series, the condenser ,IB being shunted by a resistance 2|. e 0Q element is constituted by a resistance 22.

Figure 20 is a fu'rther example of a coupling arrangement between two valves employing anode and grid bias" decoupling networks. The anode decoupling networkaP, in Figure 20 includes resistances 23and 24, and condensers 25 and 26. The pQ elements are constituted by the condenser 21 and shuntingresistance 28, together with the additional series connected resistance 29 and condenser 30, also shunted across the condenser 21. The b'Q element is-constituted by resistance 3| and the as elements by resistance 32 and condenser 33; The cQ element is constituted by resistance shunted by condenser 35.

An understanding of the application of the invention to circuit design will be further assisted by the following explanation of the derivation of D. C. coupling'circuits including decoupling for the grid bias and also a derivation of an A. C. coupling which allows a low frequency time constant to be used very much longer than the normal limit set by the anode decoupling condenser and the anode decoupling resistance.

Figure 21 shows a typical anode circuit comprising an anode resistance R1, anode decoupling resistance R2, and decoupling condenser C2, which is shown as going to the H. T. but which can equally well go to earth. This circuit has anlimpedance Z R1+mand his the object of the following derivation to provide circuits such as to give signal anode voltage equal to ZIAso'where IASC is the signal anode current which wouldbe obtained for zero load in the anodeand to provide a grid voltage to the-following valve given by RIIASC. The circuit shown in Figure 21 will only give the correct anode voltage if the anode impedance of the valve isinfinite, but an allowance will later be made 'fora finite anode'impedance. The circuit of Figure-21 can be converted to the circuit of Figure 22, following the conversions previously mentioned book Transmission Circuits for Telephonic Communications by K. S. Johnson at pages 277 and 2'78.

' Figure 23 shows the circuit of Figure 22 split up. The elements R4', R3"' and C3"' are three elements identical in form but many times the impedance'of R4,- R3; and C3. The magnitudes of these-elements are'based on the drain of H. T. currentwhich-will ultimately be passed to the given in the grid bias source. For example, R4!" and 8.3 may each be 100 times R4 and R3, and Cs' may be 100th the capacity of C3. The remnant elements are of such a size that combined with R4, R3' and C3, in parallel, they exactly make up the original impedance shown in Figure 22. The remnant portion of B3, C3 is further broken up into two portions R3, C3 and R2 and C3". In general the portions B3", C2", will be very much higher in impedance than the remnant portion B3 C3 and may as before be uniformly 100 times the impedance, as an example, though the whole order of this factor may be altered at will to give the desired final values of the circuit.

If now the anode impedance of the valve is not infinite, this may be allowed for by replacing R4 by another resistance R4 such that this new resistance in parallel with the anode resistance of the valve equals the previous B4. In Figure 24 such a replacement has been made so that allowance hereafter has been made for the finite anode resistance of the valve.

In Figure 2% only the circuit elements comprising R4 and R3 have been left connected to the high tension source. The remaining elements have been connected between anode and grid bias. Since the grid bias source, like the H. T. source, is ideally unaifected by signal potentials, this will not affect the signal potential on the anode.

In Figure 25 the elements R2, R3' and Cs' have been reconverted (as by the conversion methods referred to above) to elements R1", R2"

and C2". Similarly elements R4, R3 and C3 have been converted to elements R1, R2 and C2. Element C3 has been converted to two condensers C5" and C6 whose values are such that, connected in series they give a resultant capacity equal to C3" and the ratios of capacity are such that the impedance of Ce" compared with that of C5" is equal to the impedance of R1 compared with R2. This gives Call C3 The potential of the anode A has been unchanged by all these modifications since the impedances are still the same as in Figure 24:. Now the elements R1", R2 and C2 are identical in form to, but uniformly a multiple of the impedance of the original R1, R2 and C2. Hence the potential across R1" i. e., at the junction point G, is the potential that would have been obtained across R1 in Figure 1, that is RlIASC. Now the potential across the elements C5", C6" and R3" is again ZIASC and a calculation based on the derivation of RsCz will show that the potential of the junction C2 of C6" and C5 is again equal to RIIASC. We now have two points G1 and G2 which both have the correct wanted signal voltage. Both of these points may therefore be connected to the grid of the following valve and such an arrangement is shown in Figure 26. The volts on the grid will now be exactly those required.

Figure 27 shows the capacities C2 and C5 combined in a single capacity C25". Also the elements R1, R2, Co" have been converted (as by the conversion referred to in the reference above) to elements R7, R8 and C8. These conversions will not alter the signal potential on the grid.

and

Figure 28 shows the elements C2 and Ca connected to earth instead of to the high tension and grid bias. If, as was hypothetically assumed, these potential sources are fixed as regards signal voltages, this change of connection will not alter the anode signal potential nor the grid signal potential. The circuit of Figure 28 therefore provides the required coupling system.

A variant on the above system is shown in Figure 29 wherein the impedance values selected for R4, R3' and C3 in Figure 23 are all infinite. The remnant grid coupling circuit therefore only comprises C5, C6" and R3, the other impedances in parallel with these being now infinite. This coupling provides true D. C. coupling to the grid of the type described above, but does not provide any D. C. connection to the grid. In Figure 29 a grid leak R9 is shown connected to the grid which provides the necessary D. C. connection and therefore converts the coupling system into a purely A. C. coupling. However, the time constant of this coupling is in practical cases approximately given by R9 which can be given a value independent of the time constant R2, C2. The more conventional coupling where there is only a condenser between anode and grid such as C5", has the time constant controlled by B2, C2, if the coupling condenser and grid leak are so fixed as to neutralise the low frequency boost, that would otherwise be given by R1 in parallel with R2 and C2.

More complex decoupling than that shown in Figure 21 may be used in the anode circuit and Figures 30, 31 and 32 correspond to Figures 1, '7 and 8 except that a more complex circuit is used. The method of deriving such a circuit is to replace the condenser C2 in Figure 21 by a complex net of condensers K2 in Figure 30. The whole method proceeds similarly, K2 (considered as an admittance so as to change one value numerically in the same manner as C2) becoming K2, K25 and K8 in Figure 31. However, in deriving the final circuit, the conversion of R2 in parallel with K2 proceeds to give two separate decoupling resistances with two decoupling condensers as described in the book referred to above von page 278, for the conversion of Figure 200 to Figure 203. The two condensers thus obtained connected to H. T. become connected to earth as previously described. A similar conversion is effected for Ba and K8 providing the double grid bias decoupling shown in Figure 32. Also K25" has been converted to a form requiring smaller total capacity.

It will be appreciated from the foregoing explanation and description of particular forms of circuit that the invention may be applied to the derivation of other forms of coupling network than those specifically mentioned,

I claim:

1. An electrical network for coupling a constant current source of signals to a high impedance device comprising a pair of branches in parallel and fed with current from said source, the first of said branches comprising two parts connected in series, one part being formed by a resistance of magnitude P shunted by an impedance of value a, the other part being formed by a resistance of magnitude aP, the second of said branches comprising three parts connected in series, a first part formed by a resistance of magnitude Q shunted by an impedance of value ,8, a

second part formed by a resistance of magnitude bQ, and a third part formed by a resistance of magnitude cQ shunted by an impedance of value 0,6, the quantities a, b, being real munbers and satisfying the relation: a=b+c and the quantities a, 3 being complex or purely imaginary and satisfying the relation:

aP bQ T is connected so as to form a third branch in parallel with said pair of branches.

4. An electrical network according to claim 1, wherein the magnitude Q is not very large compared with the magnitude 1?, and a resistance c is connected so as to form a third branch in parallel with said pair of branches, the latter resistance arranged to be present effectively by modifying the magnitudes of the elements of the second branch of said network.

5. An electrical network according to claim 1, wherein the magnitude Q is not very large compared with the magnitude P, and a resistance is connected so as to form a third branch in parallel with said pair of branches, the latter resistance arranged to be present effectively by modifying the magnitudes of the elements of the first branch of said network.

6. An electrical network according to claim 1,

wherein the impedance of said source of signals is finite and the magnitudes of the elements of said first and second branches are modified to allow for the finite impedance of said source.

'7. An electrical network according to claim 1,

wherein the impedances of value a, p and 0e are capacitative reactances, and those having the values a and 0 3 serve for the purpose of decoupling.

8. An electrical network according to claim 1, wherein the impedance of value a is in conjunction with the resistance of magnitude P the electrically equivalent circuit to a network comprising series resistances and shunt condensers serving for the purpose of decoupling.

9. An electrical network according to claim' 1, wherein the impedance of value cc is in conjunction with the resistance of magnitude cQ the electrically equivalent circuit to a network comprising series resistances and shunt condensers serving for the purpose of decoupling.

10. An electrical network according to claim 1, wherein any of the parts of said first and second branches formed by shunt combinations of impedance and resistance is the electrically equivalent circuit to a network transformable to said equivalent circuit.

11. An electrical network according to claim 1, wherein the second of said branches is composed of a pair of parallel branches, one of which includes a resistance connected in parallel with a condenser and a second resistance in series with that parallel combination, and the second of said parallel branches includes two condensers and a resistance in series, the arrangement being such that the junction point between said parallel combination and said second resistance is at the same potential as the junction point between said two condensers, one or the other of said junction points being connected to said high impedance device.

12. An electrical network according to claim 1, wherein the second of said branches is composed of a pair of parallel branches, one of which includes a resistance connected in parallel with a condenser and a second resistance in series with that parallel combination, and the second of said parallel branches includes two condensers and a resistance in series, the arrangement being such that the junction point between said parallel combination and said second resistance is at the same potential as the junction point between said two condensers, one or the other of said junction points being connected to said high impedance device and wherein the first of said pair of parallel branches is effectively absorbed in the first of said branches and said high impedance device is connected to said junction point of the second of said pair of parallel branches.

13. A circuit arrangement for amplifying substantially uniformly all signal frequencies within a wide range down to and including zero frequency, comprising two thermionic valves and a network interconnecting the anode of one with the control grid of the other, said network comprising a pair of branches in parallel and fed with current from the anode of said one valve, the first of said branches comprising two parts connected in series, one part being formed by a resistance of magnitude P shunted by an impedance of value oz, the other part being formed by a resistance of magnitude aP, the second of said branches comprising three parts connected in series, a first part formed by a resistance of magnitude Q shunted by an impedance of value c, a second part formed by a resistance of magnitude bQ, and a third part formed by a resistance of magnitude cQ shunted by an impedance of value 0 8, the quantities a, b, 0 being real numbers and satisfying the relation: a=b +c and the quantities a, 5 being complex or purely imaginary and satisfying the relation:

I wherein it is arranged that the sum of the potential differences developed across the second and third parts of said second branch are fed to the control grid of said other valve.

ERIC LAWRENCE CASLING WHITE.

Images