US2410275A

US2410275A - Signaling by submodulation

Info

Publication number: US2410275A
Application number: US422915A
Authority: US
Inventors: Stanley D Eilenberger
Original assignee: HOWARD W TAFT; L G VOORHEES; S E STEEN
Current assignee: HOWARD W TAFT; L G VOORHEES; S E STEEN
Priority date: 1941-12-13
Filing date: 1941-12-13
Publication date: 1946-10-29
Anticipated expiration: 1963-10-29

Description

di f'fw F E c lLJ -T 4 Oct. 29, 1946. s. D. EILENBERGER 2,410,275

SIGNALING BY SUB-MODULATION Filed Dec. 13, 1941 5 Sheets-Sheet 1 AUDIO moou- 03c. MICRO AMI? LATOR' BUFFER CbNVERTL-R 15AM? MIXER I I v V N O F F FILTER D/Sc. H}?

R5 on. AJEc. REPRO- A5 A M F.

L.P. tum; m MIXER FILTER AME y F l Ga 3 E M l L I a 3i 7 0%. (gal M00. HEAHP. w gg AEC. fi/XER REE U N/ o; P Q R 2 T TO NE (3 7 0R A r MODLJI ATOR mam. AJ-r nP. 7 non. o sc.. I R AME Fleas gm INVENTOR SIGNALING BY SUB-MODULATION Filed Dec. 13, 1941 3 Sheets-Sheet 2 is f: ll 51- '5.

F] Ga IO INVENTQR ATTORNEY s; 0. EILENBERGER 2,410,275

fatented Oct. 29, 19 46 UNITED STATES PArENrorFici:

SIGNALING BY SUBMODULATION Stanley D. Eilenberger, Kenosha, Wis., assignor, by mesne assignments, of sixty-five percent to S. E. Steen, Kenosha, Wis., seventeen and onehalf percent to L. G. Voorhees, Macedonia, Ohio, and seventeen and one-half percent tov Howard W. Taft, Balboa Heights, 0. Z.

Application December 13,1941, Serial No. 422,915

nal frequency. The carrier so modulated may bethe radio-frequency carrier by which radio transmission is effected, a low-frequency carrier of a mean frequency less than the frequency of the modulating signal, or an intermediate carrier from which the low-frequency carrier is produced. The low-frequency carrier may be transmitted directly over wire lines or used to modulate a higher carrier frequency for radio transmission as hereinafter described. In such a frequency modulated wave the highest density of energy will always be in the carrier wave.

, Among the numerous objects of this invention are:

First, to provide a method of producing a modulated signal wave where the main energy is confined to the carrier and the amplitude of the sidebands is inversely proportional to the signal frequency, thus reducing the interference potentialities of the sidebands so that the frequency channel required by this modulated signal wave duringtransmission by means of wire line or radio shall be much smaller than the original signal bandwidth, thus providing a means of communication in a very narrow channel.

'Second, to provide a method of frequency modulation of a radio-frequency carrier where the bandwidth occupied b the main energy of themodulated radio-frequency carrier is less than the bandwidth of the original signal wave.

Third, to provide a method of reception ofa frequency-modulated radio-frequency carrier wave as described, that will permit full recovery of the original signal wave both as to amplitude and. frequency range. 7 .Fourth, to provide a method of reception of a frequency modulated radio-frequency carrier wave in which the modulation factor is increased prior to demodulation.

Fifth, to provide a method of reception of a frequency-modulated radio-frequency carrier 22 Claims. (01. 2506) wave in which the modulation factor is increased in successive steps prior to demodulation.

Sixth, to provide a method of reception of a frequency-modulated radio-frequency wave in which the modulation factor is increased at the receiver by reducing the mean carrier frequency to a value less than the frequency of the signal wave recovered by demodulation.

Seventh; to provide a method of communication by frequency modulation by which the modulation factor may be controlled at the receiver.

Eighth, to provide a method of communication which sha'llprovide a high degree of freedom from all undesired signals, such as noise, static, etc.

Ninth, to provide a method of secret communication.

Tenth, to provide a method of controlling the depth of modulation in a frequency-modulation system so that the efiiciency of recovering the original signal at the receiver may be uniformly high.

Eleventh, to provide a method of demodulation in a frequency modulated system so that a small deviation ratio at the transmitter is made to appear as a large deviation ratio at the receiver.

Twelfth, to provide a method of frequency modulation in which a high order of frequency stability is obtained.

Thirteenth, to provide a method of combining two or more of the above advantages into a single system.

It is understood that a basic purpose of this invention is to provide a means of controlling the energy concentration ratio in an electric wave asset forth in object one above, and that such control is always obtained in combination with one or more other features of this invention,

For the purposes of this application, the term sub-modulation is understood to mean a condition where the lowest frequency of the modulating signal wave is higher than the carrier wave; this low frequency carrier is referred to as a subcarrier, and the process of modulating such a carrier is referred to as sub-modulation, Such a sub-modulated subcarrieris always obtained in combination with one or more other features of this invention.

The following mathematical analysis of the frequency modulation process and the accompanying explanation of the manner in which carrier various factors are treated in this invention will serve to show how this invention differs from prior frequency modulation methods, where l T:21.-n denotes the signal frequency.

F=21rf, the frequency modulated carrier.

Ar the amplitude of the unmodulated carrier F. m the modulation index defined by the equation.

It is well known that Equation 1 can be expanded in a series to be represented by the Bessel function A=Af cos [FM- sin Ni] 71:00 A=l:l (m) sin Ft+22[;,,(m) cos 2nFt+ 2211 (QM-1) (m) sin (2n,l)Ft:|A; (1a) In this series the term with the coeflicient Io represents the carrier corresponding to the Bessel function of the zero order. The side bands are represented by the terms having the coefficients of the Bessel function of the first and nth order. It can be shown that if the argument m becomes small then the series converges very rapidly and all the terms having the coefficients of the higher order function will disappear. In such a case the coefficient In of the first term representing the carrier will contain the greater part of the energy contained in the entire series.

For purposes of illustration Equation 1 can be transformed into A=A [cos (Ft) cos (gen Nt)sin (Ft) sin (N01 2 itself because the sine function is under all circumstances smaller than unity. Hence, the value of the cosine function S 2 sin N t) can be approximated very well with unity without making any great error. For similar reasons, if m is small, the function sin sin N t) can be written as in substituting the sine function by the angle itself. With this approximation Equation 2 will be transformed into A=A [cos Ft% sin (Ft) sin (N01 A cos Fz A, [cos p mps- 51, [cos (FN)t] Equation 3 represents a familiar expression of a modulated wave containing both the carrier and the two side band frequencies at a distance corresponding to the signal frequency. In contrast to amplitude modulation, the side band components have opposite signs and are out of phase with the carrier. The amplitude of the side bands is in the first approximation proportional to the modulation index as defined above. Hence, for any given width of the frequency swing between the two edge frequencies fa and ft, the amplitude of the side bands will become smaller as the signal frequency becomes higher. This is obvious because the modulation index 112 for a given frequency swing is inversely proportional to the signal frequency.

It can therefore be stated: in a frequency modulating system where the modulation index is small the main energy will be confined to the carrier and furthermore, the amplitude of the side bands will become lower with increasing signal frequency. In contrast .to ordinary amplitude modulation where the amplitude of the side bands is only determined by the depth of modulation and is independent of the value of the signal frequency and where, therefore, the interference of adjacent channels is solely dependent upon the highest frequency present in the sideband, a frequency modulated system in which the modulation index m is small presents a case wherein the amplitude of the interfering signal from the adjacent channels will continuously decrease as the frequency limits of the side band are approached. In the commonly used systems of frequency modulation, the modulation index is always larger than unity, and there-- fore, the energy distribution as given by the expansion of the frequency modulated wave into the Bessel function series shown in Equation 1a, shows that where m is larger than unity, a greater part of the energy will be contained in the first and nth order terms representing the sidebands. In a system according to this invention, m is always smaller than unity, and therefore, the greater part of the energy will remain in the carrier.

In the discussion above no statement was made concerning the relationship between the frequencies F and N. Under ordinary circumstances the frequency F is largev in comparison to frequency N. However, it can be easily shown that if the frequency F modulated with the frequency N should form part of a beat frequency, an expression similar to Equation 3, for the amplitude of the resulting beat frequency B can be obtained in which The resulting beat frequency signal thus has the same characteristics as the original frequency modulated signal and contains the same signal frequencies in the form of sidebands. By proper selections of the values F and F1, a condition can be arrived at in which E is small in comparison to N. s

The invention will be best understood from a consideration of the following detailed description in view of the accompanying drawings forming a part of the specification; nevertheless, it is to be understood that the invention is not confined to the disclosure, being susceptible of such changes and modifications as define no material departure from the salient features of the invention as expressed in the appended claims.

In the drawings:

Figure 1 shows in block form a radio transmitter arranged to produce a frequency-modulated radio frequency wave in accordance with this invention.

Figure 2 shows in block form a receiver for use with the transmitter of Fig. 1.

Figure 3 shows in block form a radio transmitter for the amplitude modulation of the radio wave by the new modulating wave produced in accordance with this invention.

' Figure 4 shows in block form a receiver for use with the transmitter of Fig. 3.

Figure.5 shows in block form a modification of Figure 8 illustrates a detailed arrangement of discriminator unit R in the receivers in Figs. 2 and 4.

Figure 9 illustrates in detail an arrangement for use in filter units G of Figs. 2, 4 and 5.

Figure 10 illustrates the frequency response curve of filter unit G and discriminator unit R.

Figure 11 is a curve showing the distribution of the energy in the sidebands for two cases, with a total frequency deviation in one direction of 100 cycles and 1000 cycles as a function of the signal frequency.

Figure 12 is a similar energy distribution curve showing the energy content in the sidebands if the deviation from the carrier is 20 cycles in one direction.

Figure 13 illustrates the percentage change of the instantaneous values of the unmodulated carrier due to modulation.

Figure 14 shows an exaggerated curve representing one quarter cycle of a frequency modulated wave modulated with a, signal frequency ten times the carrier frequency.

Figure 15 shows the attenuation of the interfering signals originating in channels adjacent to the carrier as a function of carrier separation.

. Referring more more particularly to Fig. 1, A is a source of original signal, which may be a microphone having an original audio frequency input covering the range of 250 to 3000 C. P. S., while B is an audio frequency amplifier of conventional type, which may be any amplifier such as is in common use; C is a modulator unit which might be of the magnetically controlled condenser type, or of the so-called reactance modulator type illustrated by Fig. 6. The object of modulator unit C is to provide a capacitive or inductive variation as a padder tuning element of oscillator D.

The reactance modulator circuit illustrated by Fig. 6 accomplishes modulation electronically and is the preferred form. The circuit shown by Fig. 6 is approximately the same as used in certain frequency modulation systems of COIlVBll-r. tional type, where the'amplified audio frequency signal from the output of amplifier unit B is supplied to the input connections I and 2 and is applied to modulator grid 1 of vacuum tube 9 which might be commercial type 6L7. B is the plate terminal of vacuum tube 9 which is fed through radio frequency choke 5 from a source of voltage II, which is approximately 250 volts.

Plate 6 is coupled to control grid ,8 through avariable compression condenser l8 having a maximum capacity of approximately 100 paf. and resistor l9'whichis approximately 150,000 ohms. 20 is a resistance of .5 megohm and 4 is also a resistance of .5-megohm. I2 is a'cathode resistorof 300 ohms and I3 is a cathode by-pass condenser of .01 ,uf. I4 is a series screen r'esistor of 30,000 ohms which is by-passed by mica condenser l6 of .01 f. and also electrolytic condenser !5 having a capacity of 8 cf. 2| is a variable compression mica condenser having a maximum capacity of 30 i f while 3 is the common ground connection and AZ-23 are connections to the oscillator tuned circuit. The audio frequency signal applied to the modulator grid 7 of the reactance modulator tube 9 will cause a phase shift in the resistance capacity network made 7 up by condensers l82l and resistors l9-20, which will be proportional to the amplitude of the signal applied, and the number of shifts per second will be proportional to the frequency of the signal applied. This exerts a tuning effect on controlled oscillator D. Modulator unit C is arranged to shift the fre-- quency of oscillator D from its mean frequency of, for example, 1,000,000 C. P. S. through a range of 20 C. P. S. plus and 20 C. P. S. minus, a total carrier shift of 40 C. P. S. equal to a maximum carrier frequency of 1,000,020 and a minimum carrier frequency of 999,980 C. P. S. The maximum shift will then correspond to the maximum amplitude of the original signal, and the total number of shifts per second will be equal to the frequency of the original signal, the entire opera tion being identical with that commonly employed in conventional frequency modulation, except that the total carrier shift is extremely small, as compared to the usual shift of '75 kc. plus and '75 kc. minus, or a total shift of 150 kc. The narrow deviation frequency-modulated radio-frequency wave is then amplified by a radio-frequency amplifier I and connected to the antenna K. L is a conventional power supply for the transmitter.

Referring to the receiver block schematic in Fig. 2, K represents the antenna, N the R. F. amplifier, O the convertor, all of which are conventional; P represents the I. F. amplifier which,

for example, is tuned to a resonant frequency of 465,000 C. P. S. N and P are tuned to accept the carrier and both sidebands as in conventional practice. E represents a fixed R. F. oscillator, tuned to a fixed frequency of 464,900 C. P. S., which is designed to have a relatively high amplitude, in comparison to the output of I. F. amplifier P, which it is beating with. With static conditions'at the transmitter of Fig. 1, the mean beat frequency developed in the mixer-amplifier unit F is C. P. S., which is the beat between the fixed oscillator unit E and the I. F. output of unit P. As the transmitter of Fig. 1 is modulated; causing a frequency shift of 40 cycles maximum (with a signal at unit A of maximum amplitude), the I. F. at the output of unit P shifts by a like amount, thus causing a similar shift in the beat note frequency developed in unit F, from a minimum value of 80 to a maximum value of C. P. s., the total shift being dependent on the amplitude of the original signal, and the number of shifts per second being dependent on'the frequency of the original signal, as has been previously shown for the carrier and I. F. waves. This shifting frequency, 80 to 120 C. P. S., is now acting as an audio-frequency carrier. Thisfrequency- 7 modulated audio-frequency carrier is then passed through filter G to attenuate impressed frequencies having a frequency greater than 120 C. P. S., such as harmonics and the other modulation products of the mixer-amplifier F. This audio frequency carrier wave is hereinafter referred to as the subcarrier wave.

Fig. 9 illustrates in detail a parallel resonant filter which I have used in practice as unit G of Fig. 2. Referring more particularly to Fig. 9, 59 and 50 represent the input terminals, from amplifier-mixer unit F; B! is a transformer, whichmight well have 2. turns ratio 1:1 and a secondary, inductance 63, which might have a value of approximately 4 henrys; condenser 65 would then have an approximate value of .25 ,lf., so that the resonant circuit represented by in ductance B3 and capacity 55 would be approximately in resonance at 100 C. P. S. Primary inductance 62 could have a like value of inductance, and might also be tuned to resonance with capacity 64, if so desired, although it is not essential that the primary inductance be in resonance; vacuum tube 12 is used as a coupling element. 68 represents the common ground connection, while 66 and 6'. represent the output terminal connection to frequency discriminator R, or to succeeding stages of filter G. As previously mentioned such a, filter as illustrated by Fig. 9 will attenuate all impressed frequencies having a frequency greater than the pass band characteristic of the tuned resonant circuit to a far reater degree than it will attenuate the high signal frequency modulation component of the 100 cycle carrier Wave, which it is, desired to pass on to frequency discriminator unit R.

The following will clarify the action of a parallel resonant circuit as used in filter G and illustrated by Fig. 9, in response to an impressed frequency modulated wave in which the modulating signal frequency is of higher order than the carrier frequency. Filter unit G is resonant at the carrier frequency, which, in the example given, is 100 cycles. In a frequency modulated wave represented by the equation the value of the unmodulated function Ab cos .Bt will coincide with the value of the modulated wave for any time t at which Nt=21rnt is an integer multiple of 1r. Between those points of coincidence the modulated wave will deviate from the values of the unmodulated wave determined by the magnitude of the modulation index m. If B is small in comparison to N, and m is small, then the frequency modulated wave corresponding to Equation 5 will attain a form as shown in Fig. 14. In this figure a quarter cycle of the carrier frequency B is shown modulated by a frequency N equalling ten times B, where 69 is the unmodulated carrier wave 13 and I is the modulated wave form, H representing fixed points where the modulated and unmodulated waves coincide; the modulation index m is then 0.01. At any point where modulation occurs the slope of the wave will change to a slope corresponding to a higher or lower frequency by an amount determined by the depth of modulation. The maximum deviation can be expected at the point where the slope of the unmodulated wave is a maximum. For a cosine function this will be at a, quarter cycle of the, wave B. If the function in Equation is computed for a given set of values 13, N and m, the deviation of the modulated Wave A=A cos Bt-lsin Ni) from the unmodulated wave in percent of the instantaneous value of the unmodulated-wave can be calculated simply by forming the difference between the two and taking the instantaneous value of the unmodulated wave as a basis of comparison. Figure 13 representing the percentage change of a cycle carrier modulated with a signal frequency of 1000 cycles and with a modulation index of 0.01 was calculated in such a way. It might be seen that in the first period of 0.00025 second the maximum deviation is only minus 0.08 percent against the unmodulated wave. This deviation reaches its maximum value of plus and minus 3.3 percent at; a distance equaling a quarter cycle of the signal frequency before and after the quarter cycle of the carrier frequency.

If the resonant circuit G shall follow the modulation according to Equation 5, the Q factor of the circuit as defined by the Equation 6 must be chosen accordingly.

In (6), Rp denotes the equivalent parallel loss resistance of the circuit in ohms, C the capacity in farads and L the inductivity in henrys. If at any moment the source of energy supply in circuit G should be disconnected, the voltage E across the circuit would not fall to zero immediately, but due to the stored energy in the circuit would drop according to the expression B E: E e 2Q Sin (Bt+) 7 in which e denotes the base of the natural logarithms. If the circuit is connected to a source of energy with the terminal voltage E0 sin Bt, maximum amplitude will not be reached immediately, but again according to the function Obviously, if any changes due to modulation should take place in the instantaneous value of the carrier wave, then the Q factor and the constants of the circuit must be such that if, for instance, the circuit was instantaneously disconnected from its source of voltage, the voltage would change by a higher percentage than that indicated by the modulation index m. For instance, for the values shown in Fig. 13, this would mean that at the point of largest difference between modulated and unmodulated wave form the resonant voltage of the circuit must decay by more than 3.3% of its actual value within a time of 0.00025 second. In a generalized form the condition cos Bt cos Bt-cos (3 sin Ni) P -T Q (9) cycles. This system, is, for inherent reasons and for its practical purposes, limited to low values of modulation index. For instance, if the modulation index should be increased to 0.1, the maximum signal frequency which could be passed with a carrier and a resonant circuit G of 100 cycles and a Q of 8.7 would be limited to 110 cycles.

The output of filter G is connected to the

input

24 and 25 of the discriminator unit R, whichis a conventional demodulator circuit commonly called a frequency discriminator, tuned to a mean frequency of 100 C. P. S. The values of the circuit elements required for the proper functioning of the demodulator in the desired frequency range are disclosed in connection with the detailed description of Fig. 8, where two iron core transformers 26 and 21, with the

primaries

28 and 29 connected in series by connecting link 30 have their secondaries 3! and 32 tuned to the

frequencies

80 and 120 cycles by means of condensers 33 having an approximate value of l ,uf., and 34, having an approximate value of .15 ,uf., provided secondaries 3! and 32 have an approximate like value of inductance of 4.5 h. The common junction 35 of the two tuned secondaries is linked to a ground connection 36. The potential ends of the tuned secondaries are connected to

grids

39 and 40 of two triode amplifier tubes 44 and with a common cathode 4! connected to ground through the common cathode bias resistor 31. An outlet 38 is provided to furnish control voltage for the automatic frequency control unit C1. The plates 42 and 43 of the tubes 44 and 45 are connected through condensers 5| and 52 to the primary 53 of an audio output transformer 54, the secondary 55 of which is feeding potentiometer 56. The balance of the discriminator circuit is obtained by adjusting the slider 49 of the potentiometer 4B, which might have a resistance of 25,000 ohms, in series with the two equal resistances 46 and 4'! of approximately 100,000 ohms. Plate voltage is supplied to the tubes 44 and 45 through the connection 50. Under conditions of balance, the audio frequency carrier voltage becomes 0, between plates 42 and 43, only the recovered signal voltage appearing at these points. The variable output taken off from the potentiometer 56 across the terminals 5'! and 58 is fed either to a correcting network NET or across the high pass filter S to the audio frequency amplifier T. Typical working characteristics of such a circuit are shown in Fig. 7, which represents an idealized input-frequency output voltage characteristic such as possessed by a discriminator. of the type shown in Fig. 8. Input frequencies are represented by the abscissae, and output voltages are represented by ordinates on an arbitrary scale of +1 to -1. With an input carrier frequency of 100 C. P. S., which corresponds to a point of zero amplitude of the modulating signal wave,

the characteristic shows an output voltage of zero. As the signal Wave approaches maximum amplitude in the negative direction, the carrier frequency increases toward 120 C. P. S., and the output voltage approaches a value of -1. As the modulating signal wave varies from negative maximum to zero, completing the half cycle, the frequency decreases from 120 C. P. S. to 100 C. P. S. and the output voltage varies from 1 to 0. On the next half cycle of the modulating signal wave the frequency decreases to 80 C. P. S.

and returns to 100 C. P. S., yielding a like half cycle (0, +1, 0) of recovered voltage at the receiver. It is thus obvious that the recovered audio voltage is dependent upon the amount that the 10 carrier frequency deviates from its mean frequency, and that the audio frequencies recovered depend upon the rate at which the carrier frequency deviates from the mean frequency.

In agreement with the mathematical discussion above, Fig. 10 shows average physical characteristics of units G and R. In this figure the line A represents the attenuation of the resonant filter G as a function of frequency, taking the carrier resonant frequency as zero level and applies to frequencies not associated with the desired frequency-modulated subcarrier. 'Curve A was arrived at by impressing a constant amplitude variable frequency across input terminals 59 and 60, and measuring filter response across

output terminals

66 and 67. Referring to the same zero level, curve B indicates the frequency response of the filter G to sidebands associated with the frequency-modulated subcarrier and of the recovered audio signal across the output terminals 5i and 58 of the discriminator unit R.

Figure 10 shows the practical results obtained with this invention in recovering the original modulating signal, the frequency of which is higher than that of the carrier frequency.

Figs. 11 and 12 indicate to what extent the energy can be concentrated in the carrier, if the modulation index is chosen properly. Curve A shows the total energy content represented in all sidebands as a function of the signal frequency for a case in which the maximum frequency deviation from the main carrier frequency is selected as cycles. The total energy content of all sidebands i less than 10 percent, in comparison to the large percentage of the energy present in the sidebands if the frequency deviation is increased to 1000 cycles, as indicated in curve B of Fig. 11. A similar curve showing the energy content of all sidebands for the practical case where the frequency deviation is 20 cycles is shown in Fig. 12, which indicates that the maximum energy content of all sidebands is always less than 2'% for modulating signals above 250 cycles.

In the example given above, the signal is moved through a value 20% higher than resonance and 20% lower than resonance. This is many times greater than is possib1e in a radio frequency circuit. For example, if the I. F. was 4.3 megacycles and, the total carrier shift in a, frequency modulated system was 100 kc., the discriminator circuit would be forced to operate on a relatively small departure from resonance. If K denotes the total frequency shift for maximum amplitude of the signal, then K is defined as the modulation factor, indicating the percentage change of the mean carrier frequency. In the present case, the extremes above and below resonance are a high percentage of the mean frequency, and, therefore, it is-possible to develop good efficiency in recovering the audio frequency energy represented by the frequency swing between 80 and cycles. of the subcarrier frequency always exactly at the same value, a conventional automatic frequency control circuit C1 might be provided which is controlled by the discriminator unit R. Unit C1 is essentially the same as unit C in Fig. 1. Any deviation from the correct Value of the subcarrier frequency will disturb the balance of the discriminator unit R, thus furnishing a control voltage to unit C1, of such polarity and magnitude that the correct value of the subcarrier frequency is restored. This might be important for the reason that the automatic frequency control In order to maintain the frequency 11 unit is compensating forall fluctuations in frequency due to instability of the oscillators either on thetransmitting or on the receiving end, thus maintaining a constant subcarrier frequency.

Automatic volume control, in its conventional form, would also be desirable, in order to maintain'a relatively constant amplitude at the output of I. F. amplifier P. No circuits are shown forautomatic volume control, as this would be arranged in accordance with such circuits as commonly used. The output of R is passed through :unit S, which is a conventional high-pass filter with a cut-off frequency of 120 cycles. This removes any remaining frequencies below 120 cycles, which, up to thi point have acted as an audio frequency carrier for the original signal, also any additional undesired low frequencies making the overall system substantially noise free, by a combination'of filters. The output of unit S is coupled into a conventional audio frequency amplifier T, and reproduced on a conventional reproducer U. These units may be any type desired and are'not illustrated in detail. The system, as described above, with the example given, would not reproduce original audio frequencies below 120 cycles, and might discriminate against frequencies just above this pass band, but would reproduce all audio frequencies lying between 250 and 3,000 without discrimination, in well designed circuits.

It is obvious that the receiver of Fig. 2 may be turned to resonance with the desired carrier in the ordinary manner, .by'tuning the radio frequency circuits of R. F. amplifier-N and the oscillator circuits of converter this is true even though the next adjacent carriers were relatively close to the desired carrier. For example, if the transmitter of Fig. 1 is transmitting at the frequency of 1000 kc. and this signal is frequency modulated plus and minus cycles, the R. F. and convertor circuits in the receiver of Fig. 2 must also be tuned to 1000 kc. in order to produce the 100 cycle subcarrier beat note at the input to filter G; if the next adjacent carrier was separated by 1 k0,, this carrier would also be amplified by the R. F. amplifier N and L. F. amplifier P and would therefore appear in the mixer unit F and also beat with fixed oscillator E, which has a frequency of 464,900 C. P. 8.; this is so because'the pass band of N and P must be wide enough to pass all side-bands oi the desired carrier; the undesired carriers, being separated from the desired carrier by 1000 C. P. S. would produce a subcarrier beat frequency, at the input tofilter G, of 900 or 1100 C. P. 5., depending upon whether-the undesired carrier was above or "below the desired carrier. The two carriers, separated by 1000 cycles will not beat with themselves to produce a spurious beat frequency ,of 1000 cycles, provided the input characteristics of mixer unit F are linear; the design of mixer circuits having such characteristics is well known in the art. Thedesired carrier, having a subcarrier beat note of 1000. P. S., is accepted and passed by 'filter'G and the signal recovered by discriminator R, as both of these circuits are resonant at 100 C. P. 3.; the undesired carriers, having subcarrier beat notes of 900 or 1100 cycles would be rejected by both filter G and discriminator unit R, and therefore, no interference would be produced with the desired carrier. It is obvious that .the nextadjacent carriers, separated from :the desiredcarrierby 2..kc. above or below, would alsoxproduce subcarrienbeat .notes of 2100 and 12 1900 cycles, which would also be rejected "by filter G. 'It should be noted that filter G is shown as a single stage in Fig. 9, but it is obvious that any number of such stages could be used in cascade, preferably separated by coupling tubes. In the example given, the carriers are separated by 1000 C. P. S. the receiver of- Fig. 2 must have suflicient selectivity in filter circuits G and discriminator R to accept the desired carrier and reject the undesired adjacent carriers; it has previously been shown that the sidebands of these undesired carriers are of such low amplitude that they cannot interfere with the desired carrier, even though the carriers were much closer than 1000 C. P. S. given in this example; the necessary carrier separation is strictly a function of the selectivity of circuits G and R; this selectivity could be increased by inserting a standard band rejection filter between the mixer unit F and filter unit G, at the points marked ER in Fig. 2; such a band rejection filter would be conventional in design, and would be tuned to reject the undesired carrier; such a filter would need have a rejection band only 40 cycles wide, as the undesired carrier must always have its main energy within the narrow band through which it is modulated; if it is desired to reject more than one interfering carrier, several band rejection filters might be connected in cascade, each tuned to a different frequency, as is conventional practice.

Referring now to Fig. '3, units A. B. C and D are identical with the corresponding units of Fig. 1 and perform the same functions, producing at the input to mixer-amplifier F a signal frequency frequency-modulated wave of 40 C. P. S. frequency swing. This is mixed with a constant frequency wave from the oscillator E to produce a frequency modulated low-frequency wave having a mean frequency of cycles, which is passed through low pass filter J, the purpose of which is to pass all frequencies below the highest signal frequency, for instance-3000 C. P.'S., and to re- .move all superfluous combination frequencies and harmonics generated in the mixer circuit F. The frequency-modulated low-frequency wave is then amplified by audio-frequency amplifier H and caused to modulate the amplitude of a radiofrequency wave supplied to the modulator M by the constant frequency oscillator E2. The amplitude-modulated wave is amplified by a'high frequency amplifier I and fed tothe antenna K. As before, L is a power'supply for the radio transmitter. The final signal may be On any carrier frequency desired and unit E2 may be a crystal controlled oscillator or any other type desired for a particular class of service. The wave radiated from antenna K normally will have a band-width equal'to twice the highest frequency of the frequency-modulated low-frequency carrier passed by the filter J. However. this can be reduced by approximately one-half by the use of filters between modulator M and the amplifier I to effect single-sideband transmission in the conventional manner.

Referring to the receiver block diagram, as illustrated in Fig. 4, as the receiving unit for the radio transmiter of'Fig. 3, K represents the antenna; N a radio frequency amplifier; P an intermediate frequency amplifier; O the convertor; Q the demodulator or detector circuit. All of these circuits are conventional and maybe of any type desired, such as those in common use. No detailed schematics are shown for these units.

.As before, however, the pass band for the radio- ,frequency and inter-mediate frequency ampli- 13 fiers should accept the carrier and first order sidebands.

The output of I. F. amplifier P is connected to the input of amplifier-mixer unit F, where it is mixed with the constant frequency wave fromoscillator E which is adjusted to the same frequency as the output of I. F. amplifier. P, which, in the example given, is 465 kc., so that the resultant beat note is zero; inasmuch as the output of I. F. amplifier P was amplitude modulated by the original 100 cycle subcarrier, which in turn had been frequency modulated by the original signal, the effect of beating the output of I. F. amplifier P with its own frequency is essentially the same as demodulation, and only the 100 cycle subcarrier frequency with its associated frequency modulated sideband will remain, the carrier having been eliminated by heating with its own frequency. Unit C1 is a reactance tunin unit arranged for automatic frequency control in a manner similar to the arrangement shown for Fig. 2, so that oscillator E may be held at a constant frequency, thus resulting in a continuous zero beat, even though the frequeency at the output of unit P might vary. In the example of Fig. 2 reactance tuning unit C1 received its control voltage from discriminator unit R; in the present instance the control voltage is derived directly from mixer unit F; it is wel1 known that in any mixer stage where two frequencies are present, the DC. plate current will rise sharply as zero beat is approached, reaching a maximum at zero beat; by utilizing the voltage drop across a series plate resistor in the mixer tube of unit F. a control voltage is developed which will cause reactance' tuning unit C1 to maintain the frequency of oscillator E so that Zero beat always results. This method of obtaining automatic frequency control is well known in the art. The output of amplifier-mixer unit F is connected to the input of filter G, and from this point on the operation of the circuit is the same as shown for Fig. 2; as previously shown, additional frequencies equal to the carrier separation will be present: these will be attenuated in the manner previously described in detail; as in the case of Fig. 2 a band rejection filter tuned to the interfering carrier frequencies might be inserted at the points marked BR, if this was desirable or necessary.

The tuning method just described is for use when two or more adjacent channels are operated close to each other, all channels using a system similar to that illustrated by the transmitter of Fig. 3 and the receiver of Fig. 4. However, such a system might be operated as a single channel between two other carriers of a different type, such as conventional amplitude modulated carriers, so that this type of tuning, that is. the zero beat method achieved by the use of oscillator E, was unnecessary. In this case the operation can be simplified by entirely eliminating units E. F

and C1, and connecting the output of I. F. amplifier P directly to a conventional demodulator unit Q, as illustrated by the dotted lines, so that the original subcarrier frequency of 100 cycles is re covered by rectification; the output of demodulator Q would then be connected directly to the input of filter G, and from this point on the operation of the receiver is as shown above. As shown for the receiver of Fig. 2, automatic volume control might be added in the conventional manher, which would be desirable for reasons previously shown.

If desired, or necessary, an equalizing network,

designed to correct the signal frequency response might be inserted at the points market NET, in

audio-frequency signal.

2 would recover the original signal.

both Figs. 2 and 4. In well designed circuits this equalizing network should not be necessary but should such a network be desired for any reason. the design would follow conventional practice. The fidelity of either system would be the maximum possible with the circuits used.

If the system is to be operated as a wire line system, as in telephone service, the audio input from the originating telephone is fed into the audio-frequency amplifier B, of Fig. 3. The output of audio-frequency amplifier H, which is the low-frequency frequency-modulated carrier wave of to 120 C. P. S., is fed into the wire line. At the receiving point the line can be connected directly to the input of a demodulator R, or if it is desired to eliminate noise or signals lying in other frequency ranges on the line, the line willbe connected to the input of filter G. Following the filter will be a demodulator R and such conventional audio-frequency networks as are required to convey the recovered audio-frequency wave to the receiving telephone.

As applied to carrier current telephony, the frequency-modulated low-frequency carrier of 80 to 120 C. P. S. can be used in lieu of the original audio-frequencies to modulate a plurality of carrier frequencies each separated from the next by an amount sufiicient to permit selection of the individual 40 C. P. S. carrier channels by proper filters at the receiving point. This process and the equipment necessary for carryin it outform the subject of a companion application and will not be described in detail herein.

The methods described of modulating an audio frequency subcarrier wave, in themselves provide a new means of secret communication, as it is selfevident that the frequency-modulated cycle audio-frequency carrier wave or the radio wave amplitude-modulated thereby would not provide an intelligent signal on any existing equipment, and that the subcarrier wave in itself would pro- Vide an efficient masking signal, effectively concealing the intelligence contained in the frequency modulation of this subcarrier wave; only a special receiver of the type illustrated in block schematic form by Fig. 4 would recover the. original It is also obvious that existing equipment would not recover the intelligence transmitted by the system of Fig. 1, due to the small modulation index; only a receiver of the type shown in block schematic form by Fig. It should be further noted that any existing system of privacy in communication, such as the many forms of frequency inversion, might be combined with the methods herein described to provide additional new methods of secret communication.

A simplification of the method is illustrated by the block diagram of Fig. 5. As in Figs. 1 and 3, A is a microphone, B is an audio-frequency amplifier, and C is a modulator. Controlled oscillator D has a mean frequency of 100 cycles which is varied between the limits of 80 and C. P. S. to produce directly the low-frequency frequency-modulated wave that. is produced by the beat method in Fig. 3. Since a greater tuning effect is required to vary the frequency of a 100 C. P. S. oscillator between the limits of 80 and 120 C. P. S. than to vary the frequency of a high frequency oscillator through a range of 40 C. P. S., a magnetic condenser might be more effective than the modulator of Fig. 6. Alternatively, modulator unit C might consist of astandard varistor as commonly used in telephone communication, so designed as to provide a resistive variation to the frequency determining resistors of a resistance tuned oscillator. The so called varistor generally consists of a plurality of copper-oxide rectifier discs arranged in the form of a Wheatstone bridge and so poled that the passage of a pulsating current through two of the opposite junctures will result in a variation in the resistance presented to a circuit connected to the two conjugate junctures. The end results secured by this or other alternate methods of tuning local oscillator D would be the same as those already described for modulator unit C and illustrated herein by Fig. 6. The output of oscillator D passes through a filter G and amplifier H to the wire line or radio modulator as above explained in connection with Fig. 3.

It should be noted that no accepted definition exists for the depth of modulation in a frequency modulated system. It has been variously defined as (a) the modulation index is 1 when the total shift in carrier frequency is equal to the frequency of the modulating signal, and (b) an ar bitrary factor where the total shift in carrier frequency is considered 100% modulation, regardless of the actual value of this carrier shift. It will appear logical that a more reasonable measure of modulation factor in a frequency modulated wave would be the relation of the total frequency deviation to the mean carrier frequency; this is so because in any resonant circuit of given decrement a very definite frequency deviation from the resonant frequency would be necessary in order to create the maximum amplitude change possible, which in a frequency discriminator circuit such as unit R of Figs. 2 and 4, would correspond tomaximum efficiency. For the purposes of this description the latter definition is used as a measure of modulation factor of a frequency modulated wave, which I have previously denoted as K, wherein K represents the percentage change of the carrier frequency.

From the definition of modulation factor given immediately above, it is self-evident that there is no relation between percentage frequency deviation produced by the frequency modulation in the transmitters of Figs. 1 and 3, and the carrier frequency, nor is there any relation between this percentage frequency deviation and the final L. F. frequency in the receiver of Figs. 2 and 4. The carrier and L. F. frequencies might be anything desired, without affecting the final modulation factor or the overall efiiciency of this system.

The only basically important percentage relationship between frequency deviation and the mean resonant frequency exists at the input to frequency discriminator unit R of Figs. 2 and 4.

In the system of Figs. 1 and 2 this relationship is under control in the receiver of Fig. 2. This is so because the resonant frequency, that is, the mean beat frequency produced between the output of I. F. amplifier unit P and fixed oscillator E may be made any value desired. In the example given the mean resonant beat frequency is 100 cycles and the total frequency deviation produced by frequency-modulation at the transmitter is from 80 cycles to 120 cycles.

The foregoing description clearly indicates that in the system of Figs. 1 and 2 the depth of modulation or modulation factor is always under control in the design of the receiver of Fig. 2, andv that this modulation factor may always be made any desired value, so that in well designed cirsuits, the frequency ratio between maximum fr quency deviation and the mean resonant beat frequency is such that maximum amplitude is developed in the discriminator unit R.

' Figure 15 represents the D. B. attenuation be low carrier level, taken at the input to low-frequency discriminator unit R and shows the maximum possible amplitude which the sidebands may attain, in terms of interference with an actjacent channel, Where the two channels are separated by a mean frequency of 250 cycles and both carriers have the same amplitude. Figure 15 was prepared from data contained in the energy distribution curve of Fig. 12 and the filter unit G attenuation curve of Fig. 10. In actual physical practice, the carrier separation would probably exceed 250 cycles and would be limited only by the ability of the receiver to select the desired carrier frequency. For any carrier separation of 250 cycles or greater, interference from the sideband frequencies would be negligible.

Two alternate methods of producing a radio Wave with the energy concentrated in the carrier wave have been described in detail, the method of Figs. 1 and 2 being preferred for radio application. The method of Figs. 3 and 4, however, While being more complex for radio purposes, illustrates better the basic method which results in a frequency-modulated wave of lower mean frequency than the lowest modulating signal frequency, with concentration of energy in the carrier, which permits an economy of frequency space for both wire and radio transmission, This is so because the sideband frequencies do not have the power to interfere with closely adjacent carrier frequencies, even though all sideband frequencies are present. Therefore, an economy of frequency space required is realized by placing adjacent carrier frequencies closer together than the bandwidth occupied by their sideband frequencies, and allowing such sideband frequencies to overlap these closely adjacent carrier frequencies. Both methods are inherently noise free to a high degree, due to the action of filter unit G in attenuating all amplitude modulation not present in the original frequency modulation, while at the same time the original frequency modulation is passed with only small insertion loss, thus rejecting, to a high degree, such amplitude modulation as may have been produced by natural or man made static, etc. Inherently, a high order of frequency stability is obtained, due to the very small frequency deviation caused by modulation; it is well known in the art that carrier frequency stability decreases as the carrier frequency deviation increases. A high order of eificiency in signal recovery is obtained, by the use of a very low subcarrier frequency, so that the small frequency deviation caused by modulation is a high percentage of the frequency of this low frequency subcarrier; therefore, a high modulation factor is always obtained in the subcarrier wave prior to demodulation.

The above examples are for the purpose of illustrating some of the methods and means by which the broad purposes of the invention may be carried out and are not to be deemed as restrictive in any manner. Other modifications and alternatives will occur to those skilled in the art without departing from the scope of the invention as defined by the following claims.

I claim:

1. The method of signaling, which comprises the steps of frequency modulating at a modulation index smaller than unity, a high frequency carrier with a signal containing the intelligence to be transmitted, transmitting said frequency modulated high frequency carrier to a receiving point, beating the received frequency modulated :17 carrier to produce a frequency modulated audio frequency subcarrier, the mean frequency of which shall be lower than the lowest signal frequency, so selecting a band offrequencies including the subcarrier that the subcarrier and its associated sidebands are substantially unattenuated while frequencies lying within the frequency band occupied by the sidebands and not.

associated with the subcarrier are attenuated below a desired interference level, and demodulating the frequency modulated audio frequency subcarrier to recover the original signal.

' 2. The method of signaling, which comprises the steps of frequency modulating 'at a modulation index smaller than unity, a high frequency carrier with a signal containing the intelligence to be transmitted, mixing said frequency modulated carrier with a constant frequency to produce a new frequency modulated carrier wave of audio frequency, the mean frequency of which shall be lower than the lowest signal frequency, transmitting said audio frequency frequencymodulated-subcarrierwave to a receiving point, so selecting a band of frequencies including the subcarrier that the subcarrier and its associated sidebands are substantially unattenuated while frequencies lying within the frequency band occupied by the sidebands and not associated with the subcarrier are attenuated below a desired interference level, and demodulating the subcarrier to recover the original signal.

3. The method of signaling, which comprises the steps of frequency modulating at a modulation index smaller than unity, a high frequency carrier with a signal containing the intelligence to be transmitted, mixing said frequency modulated carrier with a constant frequency to produce a new frequency modulated subcarrier wave of audio frequency, the mean frequency of which shall be lower than the lowest signal frequency, impressing said frequency modulated audio frequency subcarrier as a modulating signal on a second high frequency carrier by amplitude 'modulation, transmitting said second high frequency carrier as an amplitude modulated high frequency signal modulated with the audio frequency frequency-modulated subcarrier to a receiving point, selecting the desired carrier frequency and rejecting all undesired carrier frequencies, demodulating the received amplitude modulated high frequency signal to recover the desired frequency modulated audio frequency subcarrier and demodulating said frequency modulated audio frequency subcarrier to recover the original signal.

4. The method of producing an amplitude modulated carrier wave,'having the main energy confined to a narrower band of frequencies than that-occupied by the signal containing the intelligence to be transmitted, which comprises the steps of producing a signal of given bandwidth, frequency modulating a firsthigh frequency carrier with said signal, at a modulation index smaller than unity, beating the so modulated first carrier wave with a constant frequency wave, thus producing a frequency modulated audio frequency subcarrier wave accompanied with sideband frequencies corresponding to the signal frequencies, whereby the amplitudes of the sideband frequencies of said subcarrier wave are small as compared to the amplitudes of the audio frequency subcarrier itself, and amplitude modulating a second high frequency carrier wave with said frequency modulated audio-frequency subcarrier wave.

5. The method of receiving an amplitude modulated wave modulated by a frequency modulated subcarrier wave as defined in claim 5, which comprises the steps of receiving the amplitude modulated wave, demodulating the desired amplitude modulated wave to recoverithe frequency modulated subcarrier wave, so selecting a band of frequencies including the subcarrier that the subcarrier and its associated sidebands are substantially unattenuated while frequencies lying withing the frequency band occupied by the sidebands and not associated with the subcarrierare attenuated "below a desired interference level, and demodulating the recovered frequency modulated subcarrierwave to recover the original signal.

6. The method of signaling, which comprises the steps of frequency modulating at a modulation index smaller than unity, a high frequency carrier with a signal containing the intelligence to be transmitted, transmitting said frequency modulated high frequency carrier to a receiving point, beating the received frequency modulated carrier to produce a frequency modulated beat frequency, the mean frequency of which shall be lower than the lowest signal frequency, to act as an audio frequency frequency-modulated subcarrier, rejecting all undesired subcarrier frequencies produced from adjacent channels and selecting only the desired subcarrier frequency and demodulating said audio frequency frequency-modulated subcarrier to recover the original signal.

'7, The method of signaling, which comprises the steps of frequency modulating a high frequency carrier with a signal containing the intelligence to be transmitted, at a modulation index smaller than unity, mixing said frequency modulated carrier with a constant frequency wave-to produce a new frequency modulated carrier wave of audio frequency, the mean frequency of which shall be lower than the lowest signal frequency, transmitting the frequency modulated audio frequency subcarrier to a receiving point, rejecting all undesired carriers and selecting only the desired carrier and demodulating the frequency modulated audio frequency subcarrier, to recover the original signal.

8. The method of signaling, which comprises the steps of frequency modulating a low freouency carrier with a signal containing the intelligence to be transmitted, at a modulation index smaller than unity, where the low frequency carrier has a mean frequency lower than the lowest signal frequency, transmitting said low frequency frequency-modulated carrier to'a receiving point, selecting the desired low frequency' carrier and rejecting all others, and demodulating the frequency modulated low frequency carrierto recover the original signal.

9. The method of signaling, which comprises the steps of frequency modulating a low frequency carrier with a signal containing the intel ligence to be transmitted, at a modulation index smaller than unity, where the low frequency carrierhas a mean frequency lower than the lowest signal frequency, impressing said frequency modulated low frequency carrier as a modulating signal on a high frequency carrier by amplitude modulation, transmitting said'amplitude modulated high frequency carrier to a receiving point, reiecting all undesired carriers and selecting only the des red carrier, demodulating the'received amplitude modulated high frequency carrier by rectification to recover the frequency modulated low frequency carrier, and demodulating the fre- 19 quency modulated low frequency carrier to recover the original signal.

10. The method of receiving a frequency modulated high frequency carrier, modulated with a signal containing the intelligence to be transmitted, at a modulation index smaller than unity, which comprises the steps of beating the received frequency modulated carrier to produce a frequency-modulated beat frequency, the mean frequency of which shall be lower than the lowest signal frequency, to act as a frequency modulated audio frequency subcarrier, filtering the frequency modulated subcarrier to attenuate all amplitude modulation products not originally present in the frequency modulated carrier wave, which have a frequency greater than the highest frequency reached by the frequency modulated subcarrier wave clue to modulation by the original signal, also to attenuate all other subcarrier frequencies having a frequency differing from the resonant frequency of the filter, demodulating the filtered frequency modulated subcarrier wave to recover the original signal, filtering the recovered signal to attenuate all residual components with a frequency lower than the highest frequency reached by the frequency modulated subcarrier due to frequency modulation by the original signal and reproducing the filtered signal.

11. A method of privacy in communication, which comprises the steps of frequency modulating a first high frequency carrier with the intelligence to be transmitted, at a modulation index smaller than unity, beating the frequency modulated first high frequency carrier with a constant frequency wave, to produce a frequency modulated beat frequency which shall act as an audio frequency subcarrier, in which the amplitude of the subcarrier at its'mean frequency is large as compared to the amplitude of the sidebands,

amplitude modulating a second high frequency carrier with this frequency modulated audio frequency subcarrier, transmitting the amplitude modulated second high frequency carrier to a receiving point, during which transmission the frequency modulated audio frequency subcarrier acts as a masking signal to prevent recovery of the original signal by unauthorized persons, demodulating the amplitude modulated high frequency carrier to recover the frequency modulated audio frequency subcarrier and demodulating the recovered frequency modulated subcarrier wave to recover the original signal.

12. The method of privacy in communication, which comprises the steps of frequency modulating a high frequency carrier with the intelligence to be transmitted at a modulation index smaller than unity so that the frequency modulated high frequency carrier will appear as a constant frequency, transmitting this frequency modulated carrier to a receiving point, beating the received carrier to produce a frequency modulated audio frequency subcarrier, the mean frequency of which shall be lower than the lowest signal frequency, so selecting a band of frequencies including the subcarrier that the subcarrierand its associated sidebands are substantially unattenuated while frequencies lying within the frequency band occupied by the sidebands and not associated with the subcarrier are attenuated below a desired interference level, and demodulating the frequency modulated audio frequency subcarrier so that authorized persons may "recover the original signal.

' 13. The method of signalin by modulated carrier currents in which the carrier frequencies are separated .by a, narrower frequency band than the bandwidth occupied by the sideband frequencies of any adjacent modulated carriers, which comprises the steps of producing a frequency modulated carrier modulated with a signal frequency containing the intelligence to be transmitted at a modulation index smaller than unity, and associated with sidebands the energy content of which is small in comparison to the energy content of the carrier; and in which the amplitude of any frequency component in the sideband becomes smaller the farther said component is separated from its associated carrier, transmitting said. frequency modulated carrier to a receiving point, beating the received frequency modulated carrier to produce a frequency modulated beat frequency, the mean frequency of which shall be lower than the lowest signal frequency, to act as a frequency modulated audio frequency subcarrier, impressing the same on an audio frequency frequency-discriminator circuit to recover the original signal.

14. The method of separating a desired frequency modulated carrier from a number of similar modulated carriers received in a common channel, all modulated at a modulation index smaller than unit, in which the frequency between any tWo adjacent carriers is smaller than the maximum bandwidth of the side bands associated with any one of the frequency modulated carriers, which comprises beating the received frequency modulated carriers to produce a plurality of frequency modulated audio frequency subcarrier waves, one of which having a mean frequency lower than the lowest signal frequency, passing this last-mentioned subcarrier wave through a filter circuit resonant at the mean frequency of the subcarrier and frequency discriminating said filtered frequency modulated audio frequency subcarrier wave to recover the original signal.

l5pIn a receiver for the reception of a frequency modulated carrier modulated with a signal frequency containing the intelligence transmitted, at a modulation, index smaller than unity, including a constant frequency oscillator producing a final frequency modulated beat frequency subcarrier having a mean frequency lower than the lowest signal frequency, a parallel resonant circuit, the resonant frequency of which is equal to said frequency modulated subcarrier mean frequency, the Q factor of said circuit satisfying the condition I wave amplitude modulated by a frequency modulated subcarrier having a modulation index less than unity, a resonant filter circuit tuned to the mean frequency of the subcarrier and having a Q factor which satisfies the condition cos F t for any given signal frequency F5 and subcarrier mean frequency Fe, a frequency discriminator circuit following the resonant filter circuit and including resonant circuits tuned to slightly different frequencies F1 and F2, above and below the subcarrier mean frequency F and satisfying the condition for their respective Q factors, when F1 or F2 are substituted for Fe.

17. In a transmitting circuit for the transmission of modulated carrier currents, a high frequency oscillator, a reactance modulator modulating said oscillator, a constant frequency oscillator separated from the mean frequency of the first oscillator by a frequency which is less than the lowest modulating frequency, a mixer circuit associated with both the aforesaid oscillators, a low pass filter circuit following said mixer circuit to pass all the difference frequencies produced in the mixer circuit.

18. In a transmitting circuit for the transmission of modulated carrier currents, a reactance modulator, a high frequency oscillator frequency modulated by said reactance modulator, a constant frequency oscillator separated from the mean frequency of the first oscillator by a frequency which is less than the.lowest modulating frequency, a mixer circuit associated with both aforesaid oscillators, a low pass filter to pass all difierence frequencies produced in said mixer circuit and a high frequency oscillator amplitude modulated with the output of said mixer circuit through said low pass filter.

19. The method of receiving an amplitude modulated wave modulated by a frequency modulated audio frequency subcarrier wave as defined in claim 5, which comprises the steps of receiving the amplitude modulated wave, beating the received amplitude modulated carrier with a constant frequency wave the frequency of which is equal to the received carrier frequency, thus producing a direct current component corresponding to zero beat frequency and sideband components corresponding to the frequency modulated audio frequency subcarrier, so selecting a band of frequencies including the subcarrier that the subcarrier and its associated sidebands are substantially unattenuated while frequencies lying within the frequency band occupied by the sidebands and not associated with the subcarrier are attenuated below a desired interference level, and demodulating the frequency modulated audio frequency subcarrier to recover the original signal.

20. The method of signaling, which comprises the steps of frequency modulating a low frequency carrier with a signal containing the intelligence to be transmitted, at a modulation index smaller than unity, the low fraquency carrier having a mean frequency lower than the lowest signal frequency, impressing said frequency modulated low frequency carrier as a modulating subcarrier on. a high frequency carrier by amplitude modulation, transmitting said amplitude modulated high frequency carrier to a receiving point, beating the received amplitude modulated carrier with a constant frequency wave the frequency of which is equal to the received carrier frequency, thus producing a direct current component corresponding to zero beat frequency and sideband components corresponding to the fre quency modulated low frequency subcarrier, so selecting a band of frequencies including the subcarrier that the subcarrier and its associated sidebands are substantially unattenuated while frequencies lying within the frequency band occupied by the sidebands and not associated with the subcarrier are attenuated below a desired interference level, and demodulating the frequency modulated audio frequency subcarrier to recover the original signal.

21. The method of receiving an amplitude modulated high frequency wave, modulated by a frequency modulated subcarrier wave having a mean frequency lower than the lowest signal frequency and frequency modulated at a modulation index lower than unity, which comprises beating the received amplitude modulated carrier with a constant frequency wave the frequency of which is equal to the received carrier frequency, thus producing a direct current component corresponding to zero beat frequency and sideband components corresponding to the frequency modulated low frequency subcarrier, so selecting a band of frequencies including the subcarrier that the subcarrier and its associated sidebands are substantially unattenuated While frequencies lying within the frequency band occupied by the sidebands and not associated with the subcarrier are attenuated below a desired interference level, and demodulating the frequency modulated audio frequency subcarrier to recover the original signal.

22. The method of signaling, which comprises the steps of frequency modulating at a modulation index smaller than unity, a high frequency carrier with a signal containing the intelligence to be transmitted, transmitting said frequency modulated high frequency carrier to a receiving point, beating the received frequency modulated carrier to produce a frequency modulated beat frequency, the mean frequency of which shall be lower than the lowest signal frequency, to act as an audio frequency frequency-modulated subcarrier, and demodulating said audio frequency frequency-modulated subcarrier to recover the original signal.

STANLEY D. EILENBERGER.