US2271078A - Modulating system - Google Patents

Description

Patented Jan. 27, 1942 UNITED STATES PATENT OFFICE MODULATING SYSTEM Walter F. Kannenberg, Rutherford, N. J assignor to Bell Telephone Laboratories, Incorporated,

New York, N. Y., a corporation of New York Application lipril '5, 1939, Serial No. 266,943

I s claims. invention relates to frequency changers which may be either modulators, demodulators or detectors for use in intelligence transmitting systems, and more particularly to such frequency changers embodying non-linear resistance elements.

Heretofore non-linear been used in modulator and demodulator systems resistance elements have as illustrated in the patents of F. A. Cowan No.1

1,959,459 and L. R. Cox No. 1,998,119. As disclosed in the patent to L. R. Cox, supra, arrangements have been devised to improve the efficiency of such systems, that is, to improve the economy of operation, particularly with reference to power consumption so that maximum desired side-band power is supplied to a line for transmission thereover. 7

It is an object of the invention to improve the fidelity of operation of modulator and demodulator systems.

It is another'object of the invention to provide in such systems substantially uniform maximum amplitudes for wave components of all frequen-. cies in desired modulated waves.

It is a further object of the invention to provide in such systems substantially identical overload characteristics for all modulating frequencies.

quencies inthe modulating range of frequencies and thereby establishing substantially uniform maximum amplitudes at whieh'all frequencies in the desired modulated waves maybe transmitted. The nature of the invention will be more fully understood from thefollowing description of a system having it embodied therein and by reference to the accompanying drawing in which:

Fig. 1 is a schematic representation of a modulator and a demodulator system embodying the invention;

Fig. 2 is a schematic representation of a dev modulator circuit embodying the invention; and

Figs. 1 and 2.

Figs. 3A through 3Fv are a -group of curves explaining the action of the systems illustrated in Referring to Fig. 1 a plurality of physically inert non-linear resistance elements III, II, I2and l3 are arranged in theform of a bridge whose one diagonal ll-IS, I6 is connected in shunt relation to. an' input circuit l1 and an output circuit IS. The input circuit includes a modulating input transformer 20 for applying modulating frequencies from a suitable source 2| to the output circuit embodying a band-pass filter BPI which applies a desired band of modulated frequencies to a line 22 for transmission thereover.

It is a still further object of the invention to enable in such systems transmission-of all modulating frequencies substantially with the same amplitude. j

It is'still another object of the invention to provide in such systems optimum impedance terminations to the non-linear resistance elements embodied in a bridge included therein.

In one type of well-known modulator and demodulator system a plurality of non-linear resistance elements are arranged in the mannerof a bridge which-has connected to one diagonal an To the other diagonal 25, 26 of the bridge is con- 'nected a carrier input transformer 21 for applying thereto .a source 28 of carrier waves.

The operation of the bridge in a modulating system is well known and, briefly, consists in "applying modulating waves from the source 2| across the diagonal l4, l6 and simultaneously therewith impressing carrier waves from the source 28 across the diagonal 25, 26. These waves together with the non-linear resistance elements produce an action whereby the moduinput circuit, including a source of modulatingwaves and an output circuit adapted to select a identical overload characteristics for all -frelating and carrier waves are combined to. effect modulated waves having a desired range of v frequencies for selection by the band-pass filter BPF' for transmission over the line 22.

In order to avoid energy losses due to variations and imperfections in the non-linear resistance elements and thereby to promote highest efliciency in modulating systems, inductancecapacity networks connected in the input, output and carrier circuits in the manner disclosed in the patent of L. R. Cox, supra, have been hereftofore employed. While such inductance-ca pacity networks tend to maintain low losses, they have been found to be inadequate from the standpoint that they donot improve the fidelity of the systems. This may be readily understood by referring to Fig. 3D wherein it is illustrated tance and adjustable resistance being .larger' and of one phasefcr larger v that the various frequencies in the desiredmoduiated output have varying modulator losses and different cut-off points; in Fig. 3E that the modulating frequencies have varying modulator losses and different cut-ofl points; and in Fig. 3F that the various frequencies included in the modulated output corresponding to the various modulating frequencies have varying maximum amplitudes and different cut-off points. Obviously, fidelity does not attain its highest degree when the action in modulating systems is in accordance with the for smaller values of resistance.

starters of resistance and larger but of opposite phase Such critical value of resistance exists regardless of the application of carrier waves to the bridge. In adcurves shown in Figs.'3D, 3E and 3F, as it is apparent that the amplitudes of certain frequencies in the modulating range of frequencies are cut off ahead of those of other frequencies in the same range. This means that the maximum amphtudes of different modulating frequencies are varyingly reduced so as to be transmitted over the line 22 as if they were originally produced in the source 2! with such variations therebetween. In other words, the ultimate maximum amplitudes at which different modulating frequencies are transmitted are different, regardless of the maximum amplitudes to which it is possible to produce these frequencies in the source 2| This does not provide for high fidelity transmission as the maximum amplitudes of the modulating iredition, it has been found that the amplitude of the fifth harmonic of the modulating frequencies likewise has a minimum value substantially at harmonics of the modulating frequencies such that the bridge diagonal embodying the carrier source possesses the critical value of resistance quencies are not transmitted over the line with regard to a maintenance of the relation therebetween as produced in the source 2| and further as the maximum amplitudes at'which the different frequencies may be transmitted are different. Hence, the different modulating frequencies have different overload characteristics. v i

invention contemplatesimproved fidelity in modulator systems by utilizing an arrangement for substantially equalizing the maximum amplitudesat which all modulating frequencies, and hence all frequencies corresponding thereto and embraced within a desired modulated band of frequencies, may be transmitted, and therefore provides for the establishment of substantially identical cut-ofl oroverload characteristics for such frequencies.

In accordance with the invention, a tuned circuit ll is inserted in one side of the modulatinginput circuit embodying the secondary winding of the input transformer 20, a tuned circuit 3! is' inserted in one side, of the bridge output circuit, and a resistance 32 is inserted in one side necessary to eliminate or 'balance out the harmonics as described above. Insertion of resistance 32, particularly when a large value is required,*interferes with the sharpness of overloading and, in addition, tends to raise the impedance of the bridge by changing the operating points of the individual non linear elements. This is so for the reason that the insertion of theresistance 32 cuts dow thefiow of rectified current in the bridge there y changing the direct current bias on the individual non-linear elements from some average value. a corresponding change in the impedance of the bridge as seen from the modulating input and the modulated output circuits. Introduction of the shunting network 33 restores the low resistance path for rectified current, hence the direct ing the resistance of the network 33, the effective bridge impedance may be raised or lowered at will as such adjustment varies the operating 1 of the carrier input circuit embodying the secondary winding of the carrierinput transformer and a network 33 comprisingin series an induc-v is connected in bridge ofthis winding.

' A modulating frequency entering the bridge "through the input circuit seesthe bridge as a full-wave rectifier. In general, for any sinucurrent bias on the ndividual elements of the bridge is restored to the average value. This low resistance direct'current path controls the sharpness of the knee of the overload or cutof! characteristic of the bridge without appreciably disturbing the impedance of the carrier path for the harmonics of the modulating frequencies as hereinbefore mentioned. By adjustpoint of the individual non-linear elements as previously mentioned. In this manner control of the average impedance of the bridge is obtained.

The modulating input circuit and the wanted side-band or modulated output circuit, particularly the channel band filter serve to terminate the bridge. As abovedescribed, the network 33 may be used to improve the impedance match soidal voltage wave applied to the bridge from themodulating source, thewave shape of the currentwill not be sinusoidal. As the return path through the bridge includes, in addition to the two conducting non-linear resistance elements, whatever. resistance is included in the bridge diagonal therebetween,l' it isfound that for relatively large voltage amplitudes the current wave tends to be peaked when this diagonal resistance approaches zero and flat topped when it approaches a relatively large value. At some critical value'of this resistance the current wave "will departa from a sine'form.. For

of the bridge and the several circuits connected thereto. As the resistance of the network 33 should be low to provide sharp overload characteristic as mentioned'above, itis important that .the nominal impedance of the modulatedoutput and modulating input circuits should preferably approximate" the. bridge impedance under such condition. Also, as the .level of the carrier waves-affects the impedance of the bridge the such'critical valueof resistance, it has been found modulating frequencies has a minimum valu ues that the amplitude of the third harmonic olfithe above impedance match should obtain at the level of carrier waves necessary to-provide the bend of the overload characteristic at the desired level for either the modulating input-or modulated output.

The circuits connected to the bridge need be matched thereto in'the abovemanner only at those frequencies which are considered to be This causes,

useful. Thus, the bridgemust approximately match the modulated output circuit at the frequencies of the wanted modulated output and the modulating input circuit at modulating frequencies. 'Such impedance match is a necessary but not a sufficient condition. If these circuits have finite impedance atcertain other frequencies initially present or produced at relatively large levels, then impedance networks must be interposed therein to so modify the impedance 1 of the bridge at the certain other frequencies that an appreciable flow thereof is precluded in the respective circuits.

The impedance of the modulating input circuit tends to be low at frequencies of the order of second harmonic of carrier. As the bridge appears to the carrier input circuit like a halfwave rectifier, it should be expected that the bridge diagonal opposite the carrier input circuit should therefore tend to contain second harmonics of carrier of relatively large amplitudes. The network 30 inserted in the modulating input circuit is tuned to be anti-resonant at twice carrier frequency, thus providingthe modulating input circuit with an impedance characteristic which is sufficiently high at twice carrier frequency to reduce the flow thereof in the modulating input circuit to .a negligible amount. 7

At modulating frequencies, however, the modulated output circuit tends to have a relatively low impedance. The network 3|, which is em- Figs. 3A, 3B and indicate, respectively, that all frequencies in the modulated output have substantially the same modulator loss and cut-off point, all frequencies in the modulating range of frequencies have substantially the same modulator loss and cut-off point, and all frequencies in the modulated output corresponding to the frequencies in the modulating range may 4 attain substantially uniform maximum amplitudes and therefore have substantially identical cut-off points. As a result the fidelity of the system is substantially improved as it is readily apparent in Figs. 3A, 3B and 3C that'all fre- -quencies in the modulating range may attain the same cut-off points. This means that all frequencies of the modulating waves have equal opportunity for transmission over the line 22 at the same maximum amplitude and therefore such frequencies may be transmitted with particular regard to a maintenance of the relation between the amplitudes thereof as produced in bodied therein and tuned to be anti-resonant at the middle of the range ofmodulating frequencies, tends to raise its impedance over the range of modulating frequencies substantially to a sufficient amount to reduce the flowof modulating frequencies in the modulated output circuit to a negligible amount.

In the manner indicated above an impedance adjustment is achieved whereby the bridge is fitted to the-connecting circuits from an impedance standpoint such that an approximate match is simultaneously obtained at certain desired fre- I quencies, a maximum mismatch is provided at certain undesired frequencies which tend to be strongly present and cannot readily be balanced out, and a balance is obtained at the third and fifth harmonics of the modulating frequencies.

In other words, the impedance networks interposed in the modulating input, carrier supply and modulated output circuits are proportioned to provide terminating impedances for the nonlinear bridge which have the most favorable magnitudes and frequency characteristics for certain wanted frequencies and certain unwanted frequencies tending to flow in the respective circuits. As such impedance adjustments are set for a given carrier level which is required to fix the knee of the overload curve at the desired input and output values, these adjustments shall provide (a) fiattest transmission of all desired frequencies, (b) sharpest overload or cut-off characteristic, (0) overload or cut-off characteristic most uniform at all frequencies transmitted and (d) a loss approaching the minimum consistent with high quality performance. .These characteristics are illustrated in Figs. 3A, 3B

and 3C.

Non-linear bridges of different characteristics may be compensated for bythe network 33 which serves to raise or lower the average impedance of each thereof until it matches substantially the connecting circuits in the manner afo'redescribed.

the source 2 i A condenser 34 interposed between points I4 and I5 tends to balance the bridge so as to preclude a leak of unmodulated carrier waves to.-

the output circuit and also to prevent a circulation of rectified current in the input and output circuits.

Referring to Fig. 1, the carrier waves may traverse two parallel paths through the nonlinear bridge, each path constituting a half-wave rectifier. Thus, one such path would comprise the non-linear elements I0 and -II included in series and poled in the same direction between the terminals 25, i5, 26 while the other such path would comprise the non-linear elements l2 and 13 included in series and poled in the. same direction as the previously mentioned non-linear elements between the terminals 2 5, i 6, 26. These potentials of the terminals ll-IS and I8 and as this is zero, no unbalance current would tend to flow between the terminals 14-45 and I6 during the conducting part of the carrier cycle. Accordingly, no unbalance current would tend to flow in the input and output circuits for the reason which will now be explained.

However, it may happen in the situation as-, sumed .above that commercial non-linear elements, due to impurities and methods of manufacture. may possess dissimilar resistance characteristics so that a difference of potential may exist across the junction terminals and 18 when a voltage is applied across the terminals- 25 and 26. that is, rectified and carrier. hereinafter explained, may tend to flow Consequently, unbalance current; currentas will be in the circuits connected. to the terminals H-ll and IS. The parallel path embodying non-linear elements having the greater dissimilarity of resistance characteristics will determine the direction of flow of such unbalance current in the circuits applied to the terminals ll-l5 and II.

- characteristics.

As the alternating and rectified current resistance characteristics of the non-linear ele- Y ments during the conducting part of the carrier cycle tend to exhibit similar trends, it is found that a change of the rectified current characteristics due to a control of the magnitude of the rectified current tending to flow in the two aforesaid parallel paths will cause a. corresponding change in the alternating current resistance From the above it is understood that substantlally no rectified current would tend to fiow in the input and output circuits when the non-linear elements possess substantially identical rectified current resistance or voltage current characteristics and rectified current would tend to flow in the input and output circuits when the non-linear elements possess dissimilar rectified current resistance or voltage current characteristics. In-the latter case the flow of rectified current would be of such direction as to tend to increase the amount thereof in the low resistanceelements and to decrease the amount thereof in the high resistance elements over the amounts that would tend to be present if no rectified current could fiow between the junction terminals and ii. In other words the resistances of the non-linear elements carrying the aforementioned increased and decreased amounts of rectified current are tended to be decreased and increased, respectively. Acc'ordingly, this tends to decrease the rectified current difference of potential between the teracross these terminals. Hence in the rectified current case, the otential across any non-linear 5 element is not proportional to the resistance thereof whereas in the alternating current case,

the potential across any non-linear element is proportional to the resistance thereof. The insertion of the condenser 34 as above ondary winding is connected a suitable translating end of the system, not shown, is one.

apparent that the alternating current difference of potential across the terminals ll-IS and I8 substantially approaches zero. This tends to diminish substantially 'or to obviate entirely a leak of unmodulated carrier current to the circuits connected to the junction terminals and l8.

Fig. 2 represents a; demodulator system in which modulated waves received over a transmission line are applied through receiving band-pass filter BF!" and T-shaped resistance pad '49 to diagonal il, 52 of anon-linear bridge 53 comprising a plurality of physically inert non-linear resistance elements that are similar to those embodied in the non-linear bridge in Fig. 1. Impressed across the other diagonal 54, 55 is the secondary winding of a carrier input transformer 56 whose primary winding is connected'to a source 51 of carrier waves. Included in one side of the carrier circuit is a series resistance 58 and bridging the secondary winding of the carrier input transformer 56 is a series network 59 consisting of an inductance and an adjustable resistance. The series resistance 58 and network I! serve the purpose of eliminating the harmonics of the modulating waves as described -above in connection with Fig. 1. Also connected to the diagonal il, 52 in shunt of the demodulator band-pass filter is a receiving lowpass filter LPF whose output is applied through a receiving transformer across whose secapparatus 86. a

The operation of the. demodulator system is well known and is, briefly, that modulated waves incoming from the transmission line 50 are applied to the bridge diagonal ll, 52, while at the same time carrier waves are impressed across the'opposite bridge diagonal ll. 56. These waves together with the non-linear resistance elements produce an action whereby the incoming modulated waves and the carrier waves are combined to eifect modulation products of which the modulating waves that originated at the transmitting These modulating waves are applied through the receiving low-pass filter and the transformer i! to the translating apparatus a. As pointed out so may discriminate in such fashion that the varistated interrupts the flow of rectified current to the input and output circuits and. in addition is charged with such polarity as to oppose the increased amount of rectified current in the non-linearelements of decreased resistance and to aid the decreased'amount of rectified current in the non-linear elements of increased resistance. Such charge therefore tends to increase the resistance of the former non-linear elements and to decrease the resistance of the above regarding Fig. 1, the non-linear bridge ous frequencies in the modulating-range have different overload characteristics, being similar to those shown "in Figs. 31), 3E and 31". Hence the fidelity of the system is relatively low.

g5 Therefore, it is contemplated by the invention explained above withrespect to Fig. 1 to promote fidelity in the system and" thereby to provide in Fig. 2 the action represented in Figs, 3A,

latter non-linearelements. In other words such charge tends to regulate the fiow of rectified current in the parallel paths 2!, ll. 28 and 2!. ll. 2! so as to bring the bridge moreclosely to a resistance balance therefor. As the alternating current resistance characteristics tends to follow closely the rectified current resistance characteristics of the non-linear elements during fected. As .the alternating current voltage is 33 and 3C. This is accomplished by serially connecting a capacity ll incne side of the demodulated wave or side-band input between the bridge terminal II and the T-shaped path 40 and also serially connecting a capacity II in a corresponding side of the modulating wave output between the same bridge terminaland thereceivlng low-pass filter. 'I'he T-shaped path serves to improve the impedance presented to both the non-linearbridge and the band-pass filter while Y Q proportional to the resistances of the non-linear elements as herelnbefore pointed out, it is now the receiving low-pass filter prevents modulated waves from passing into the translating apparatus I.

Capacities n and II interrupt a fiow of 'rectified current in the respective circuits thereby preventing a leak of carrier waves thereinto.

waves to their respectivecircuits is desirable so as to avoid unnecessary dissipation of energy. This is accomplished by making capacity 10 relatively small so as to have low impedance for modulated waves of high frequencies and high impedance for modulating waves of low frequencies, and by making capacity II relatively large so as to have low impedance for modulating or signaling waves of low frequencies. While such conservation of energy is a primeconsideration in modulating-demodulating systems, another consideration of vital importance is the attainment of high fidelity transmission therein.

In accordance with the operation of the invention as explained above in connection with Fig. 1, high fidelity is achieved in Fig. 2 by further adjusting capacities I andll so as. to terminate the input, output and carrier circuitsin optimum impedance relative to the non-linear bridge to obtain the action represented in Figs. 3A, 3B

rier waves connected to the other diagonal of said network so that the modulating and carrier waves are combined in a desired band of modulated waves and rectified current isproduced in circuit with said network, and an output circuit selective to the desired band of modulated waves and connected to said one diagonal of'said'network, means to provide for said network terminating impedances which have substantially the most favorable magnitudes and frequency characteristics for transmitting wave components of all frequencies in the desired band of modulated waves with substantially uniform maximum and 3C. This means that all modulating waves have equal opportunity for reception at the same maximum amplitude. Thus, the unfavorable discrimination of the non-linear bridge for certain frequencies in the modulating range is obviated and all such frequencies have the same overload characteristics. This therefore promotes a substantial improvement in'the fidelity'of the modulating system as explained above in connection with Fig. 1. It has been found that whatever energy is sacrified in the further adjustment of capacities l0 and H it is more than adequately compensated for by the improved fidelity.

' What is claimed is:

1. In a modulating system comprising a source of modulating frequency waves, an input-circuit selective to the modulating waves, a load circuit selective to desired modulated waves, a plurality of non-linear resistance elements arranged in the form of a bridge which has one diagonal connected to the modulating input and load circuits, and a source of carrier frequency waves connected to the other diagonal of the bridge so that the modulating and carrier waves are combined in the desired modulated waves; means comprising a capacity connected in the bridge diagonal to which the modulatinginput and load circuits are connected to interrupt a direct current path to both latter circuits to balance thebridge against a leak of unmodulated carrier waves into the load circuit.

2. In combination in a wave combining system, a plurality of non-linear rectifier elements arranged in a bridge network, aninput circuit including a source of modulating waves, an output circuit selective to a range of desired output frequencies, a source of carrier waves, said input and output circuits applied to 'one'diagonal of said network said source of carrier waves onnected to the other diagonal of said network to cause side-band wave products of said carrier waves and modulating waves to flow in said output circuit and to produce rectified current fiow in circuit 'with said network, and means to selectively control the magnitude of said rectified current flowing in said network comprising a shunt path around said other network diagonal, of low impedance to said rectified current-but of high impedance to harmonics of said modulating waves;

3. In a modulating system comprising ai amplitudes, said terminating means comprising an impedance in series with said modulating source to suppress waves having the frequency of ing the desired band of modulated waves, a resistance in series relation to a portion of said network and to said source of modulating waves and in circuit with said source of carrier waves, said resistanceproportioned to minimize distortion of the wave form of the modulating current and tending to impede the flow of said rectified current and an' inductive impedance in shunt relation to said and other network diagonal to aflford a path for the flow of said rectified current to limit to a desired amount the flow of said rectified current in said network.

4. In a modulating system, a source of input waves of a band of frequencies, a four-terminal network of non-linear rectifier elements arranged in bridge form and having one pair of diagonally opposite terminals connected to said source of input waves, an output circuit selective to a different band of frequencies connected to said one pair of terminais, and a source of carrier waves connected to the two remaining diagonally opposite terminals of said network, said network serving to produce modulated carrier waves in said output circuit, representing modulation products of said carrier and input waves, said network being also traversed by direct current I resulting from rectification of carrier waves by said non-linear rectifier elements, and means to control the magnitude of said direct current comprising an inductive impedance connected across said two remaining terminals of said network.

5. In combination, an input circuit selective to waves in a certain band of frequencies, an output circuit selective to waves of a different band of frequencies, a plurality of non-linear rectifier elements arranged in a bridge network and having one diagonal connected across said input and output circuits, a source of carrier frequency waves appliedto the other diagonal of said network so that the input waves and the carrier waves are combined in output waves in said different band of frequencies, and rectified current flow is produced in said network, and means -to terminate said .network' with impedances of such magnitude and frequency characteristic that all waves in the different band of output frequencies are transmitted with substantially uniform attenuation, said terminating means comprising in each of said input and output circuits means'to suppress waves having a certain -order'of frequency and to pass a wanted band of frequencies with substantially uniform attenuation, and in circuit with said carrier source 'a further network comprising a resistance embodied in series: with-said sourceto suppress waves having harmonic frequencies of the input waves of said certain band of frequencies, and an inductance and a variable resistance in series applied in shunt relation to said source and said other diagonal of said network to control the amount of rectified current flowing in said network.

6. In a wave translating system, comprising a plurality of non-linear rectifier elements arranged in the form of a bridge network, a plurality of circuits for supplying alternating current waves to said network to be translated into 'waves of certain frequency and rectified-current,

frequency, and circuit means to'connect said a load circuit to utilize said waves of certain frea plurality 'of non-linear rectifier elements arranged in the form of a bridge network, a plurality of circuits for supplying alternating current waves to said network to be translated into waves of certain frequency and rectified current, a load circuit to utilize said waves of certain supplying circuits and said load circuit to the diagonals of said network, said rectified current fiowing in said network and one of said circuits;

'means comprising anr'impedance network between one network diagonal and said one circuit to control the magnitude of the said rectified current flowing in said network, said im pedance network comprising a resistance in series in one side of said one circuit and a resistance and an inductance in series applied in shunt of both said onev network diagonal and said one circuit.

8. In a wave translating system, comprising a plurality of non-linear rectifier elements arranged in a bridge network, a plurality of circuits for supplying alternating current waves to said network to be translated into waves of certain frequency and rectified current, a load circuit to utilize said waves of certain frequency, and circuit means to connect said supplying circuits and said load circuit to the diagonals of said networkzsaid rectified current tending to fiow in said network and said. circuits, means comprising an impedance network to control the flow of said rectified current in said network and said circuits, comprising a capacitor in series with one network diagonal, a resistor in series with the other network diagonal, and an inductor and a resistor in series applied in shunt of said other network diagonal.

WALTER 1".

Images