US2315173A - Arrangement for noise elimination at the receiver - Google Patents

Description

March 30, 1943. M. WALD 2,315,173

ARRANGEMENT FOR NOISE ELIMINATION AT THE RECEIVER Filed April 4, 1.940 2 Sheets-Sheet 1 F'IG.3.' F |G.2.

Ann

lAlAlAl n I INVENT OR.

M20462 WW ATTORNEY March 30, 1943. M. WALD 2,315,173

ARRANGEMENT FOR NOISE ELIMINATION AT THE RECEIVER Filed April 4, 1940 2 Sheets-Sheet 2 INVENTOR.

ATTORNEY Patented Mar. 3%, 19 13 ARRANGEMENT FOR NOISE ELIll/[INATION AT THE RECEIVER Martin Wald, Bucharest, Rumania; vested in the Alien Property Custodian Application April 4, 1940, Serial No. 327,896

Claims. .(Cl. 250-) This invention relates to a method for noiseelimination based on the principle of compensation. This principle is as follows:

Two closely neighbouring frequency-ranges ar selected from the antenna output by similar tuned filters. The signal-filter receives the signal and the accompanying noises, while the other auxiliary-filter receives only the noises. It is clear that in connecting the two filters in opposition to each other, the noise-oscillations will not be eliminated because two voltages of different frequencies cannot neutralize each other, instead they will give rise to beats. To get real compensation the oscillations, which have to cancel each other, ought to be identical in frequency and in direct opposed phase. None of the methods hitherto known have fulfilled this condition. The difficulty arises in the practical realisation of the phase-compensation.

The object of the invention is a method in which the phase and frequency of the noiseoscillations We have to neutralize are exactly considered, allowing a correct phase-compensation. According to the invention the noise-oscillations occurring in the auxiliary-filter or filters will be first transformed by means of non-linear circuits in such a way, that they become identical in frequency and in direct opposed phase to the noiseoscillations existing in the signal-filter. The operation indicated we will further designate as phase-compensation.

The means for carrying this out in practice according to the invention are illustrated by way of examples in the accompanying drawings.

Fig. 1 illustrates the transmitting and receiving of a so-called ideal-shock-voltage.

Fig. 2 shows the time-curve of an ideal-shockvoltage.

Fig. 3 shows an equivalent circuit diagram on the receiving end.

Fig. 4 shows a diagram of an arrangement for phase-compensation by means of frequency-shift in the local oscillator.

Fig. 5 shows a diagram of an arrangement for producing the necessary impulse voltage.

Fig. 6 shows the time-curve of the voltages occurring in different points of the arrangement in Fig. 5.

Fig. 7 shows a diagram of a Kipp amplifier.

Fig. 8 shows the diagram of a receiver designed according to the method of phase-compensation by means of frequency-leap in the local oscillator.

Before passing to a description ofthe methods according to the invention, the time curve of the noise oscillations caused by a so-called idealshock-voltage will be examined. We suppose that the key Mon the left-hand part of Fig. 1 is closed suddenly at the time i=0. Across the pure ohmic resistance R the voltage will rise suddenly from zero to thevalue of the batteryvoltage EB. The current through R will also suddenly rise to the value EB/R. This time-curve is shown in Fig. 2 and the curve of the same kind will be designated as an ideal-shock-voltage. This wave-form can really be achieved only approximately, since in every electrical circuit there is also inductance and capacitance. The inductance flattens the rise of the current while the capacitance flattens that of the voltage.

For the sake of simplicity, at first we will treat the noise elimination in the case that the noise occurs in the form of an ideal-shock-voltage. As will be shown hereafter the results can be generalized to the case where the noise-oscillations have any wave-form. An ideal-shock-voltage can be considered as the sum of sinusoidal oscillations according to the well known theorem of Fourier. A continuous frequency-spectrum will be found with a bandwidth from zero to infinity. All these components of different frequencies will be radiated from the aerial A1 in Fig. '1 and afterwards they will energize the receiving-aerial A2. The wave-form of the noiseoscillations occurring in the receiving-aerial A2 will be different from that of an ideal-shockvoltage since the components of different frequency will be damped in different measure during their propagation from A1 to A2. To the aerial A2 is coupled a filter F1 having a small pass-band. For the frequencies lying within this small pass-band the space-damping during the propagation from A1 to A2 can be considered as constant and therefore the oscillations across the output terminals of filter F1 will be the same as if an ideal-shock-voltage would be impressed directly on the receiving-aerial A2. We have the equivalent scheme in Fig. 3. The ideal shockvoltage is produced by closing the key M which suddenly impresses the battery-voltage across the primary circuit of the band filter. As is well known, the voltage occurring across the condenser of the secondary circuit has the wave-form of damped fluctuating-oscillations as follows:

wherein 7 denotes the natural frequency of the uncoupled circuit, f+p and f-q are the two frequencies of the double humps oi. the resonantwhich equation is applicableat any time t .0,if A

i= denotes the instant at which the key M was closed. As is shown in Equations '1 and 2; the

amplitudes of the noise-oscillations occurring-in the filters F1 and F2 are equal to each other,-but

they have different frequencies.

The method according to the invention is shown diagrammatically in Fig 4 and maybe called phase compe'nsation by means I of frequency-shift in thewlocal oscillator. In Fig. 4

are'two filters F .and F2 coupled to the antenna A. The filter Fiis tuned to the signal-frequency while the filter F1 is tuned to a closely neighbour injg" frequency (for instance kc. higher or lower). The filter F1 feeds as usual-- -the grid G4 of the tube V1', whi1e the filter Faisconnected to the'gridG4 of thetube V2; To the anodes of both tubes and filter F1 is connected in pushpull arrangement and is tuned to the intermediate frequency and feeds into the I. F. part of the receiver, which is as usual. The oscillatormodulator tube V produces-as usualthe local oscillator-frequeney fo=f+f1, which is higher by the .I. F. frequency 1 upper than thesignal-frequency. During the time at which there is no mutual conductance g of the tube.

grid and oscillator-grid, calculating the I. F. component of this product and multiplying by the Thus we obtain:

cos (f.-+q)

The Equation 4 which represents the I. F. voltage delivered by the tube V1 is built up in the same way as Equation 1 and therefore the waveform of it is the same as that of the noise-oscilla- V .tions in filter F1, yet here is also the phase angle 95 being equal to that of the oscillator-voltage which existed at the instant i=0 (the instant at which the ideal shock-voltage was occurring).

Now we will consider the phenomena occurring in the tube V2.

noise-voltage in the aerial, the filter Faremaining samelas in V1 that is ,f-l fi while th phase of the oscillator voltage in ,V2 is identical to the phase of .the oscillator voltage in -V1. This accordingly is accomplished by a synchronising voltage originating from the oscillator circuit in V1 and used forcontrolling the oscillator voltage in V. In Fig. ithe synchronisingdevice is designated-with The oscillator frequency in V2 can be also influenced bythe device K in such a manner thatif across the output terminals of K there suddenly occurs an impulse voltage, this causes a sudden increase of the frequency in V2 by the amount M, which is the tuning-difference between the filters F1 and F During the time there is no noise -voltage and only the .7 signalvoltagegacts on the aerial, this signal will be transposed by'the tubeV1 inthe normal way to the intermediate frequency and fed into the I. F. stage V1. The tube V; has no influence on this operation since the signal-frequency can not pass through the filter Fz and thus the grid G4 bf valve V2 is not excited]. Now we will consider The voltage on the modulatorgrid' is'given by the noise-oscillations in filter F2 which are represented by Equation 2. Until the instant t=0 at which the shock-voltage appears, the oscillator-voltage in V2 has the same frequency and phase angle as that in tube V1. The noise voltages beginning at the .time i=0 in filter F2 will be amplified and rectified by the arrangement K to be described further and the produced negative voltage will be fed into th grid of tube 'VF causing the capacitance between the grid G4 and cathode to be decreased and thus there results a sudden increase of the oscillator-frequency by the amount Af. At any time t 0 the oscillator-voltage may be represented as follows:

Wherein the phase angle o is the same as in Equation 3 in the expression for the oscillatorvoltage of the tube V1 since until t=0 both oscillator-voltages were in phase-accordance. Multiplying the voltages acting on the modulator and oscillator-grid, that is the expressions in Equations 5 and 2 together, and considering only the I. F. component of this product, we obtain for the I. F. voltage delivered from the tube V2:

V V l e- (f.-+q) which expression is identical with that for the I. F. voltage delivered from V1 according to Equation 4. As Fig. 4 shows, the two tubes V1 and V2 work in opposite phase with respect to the I. F. filter F: and therefore in the latter the noiseoscillations originated from the tubes Vrand V2 cancel each other and only the signal-voltage rea mains. After. the noise-oscillations in filter F2 originating from the shock voltage have decayed, the grid-voltage of tube VF becom zero and the grid-cathode. capacitance as well as the frequency of the oscillator-voltage in V2, again take normal values. By virtue of the synchronising voltage delivered from thedevice S, the phase-accordance between the oscillator-voltages in both tubes thecase that at time i=0 an ideal-shock-voltage V 7 occursin the aerial. Across the filters F1 andFz noise-oscillations will arise, the time-curve of which may be represented by the Equations 1 and 2. The frequency of'the oscillator-voltage or in the valve V1 is f+ fl and supposing that the phase angle of it at .the instant in which'the shock-voltage occurs has the value we obtain:

The ,IQF. voltage delivered frorn the tube V1 can be obtained'as known-by forming the prod- V1 and V2 will be established-again. and there.- fore the whole arrangement is able to compensate another noise-impulse. '1

We supposed tillnow, that the rectified impulse voltage resulting from the arrangement K occurs at time t=0 without any delay. Reallyit will take a finite time depending on the construction till the necessary value of the impulse voltage will be achieved to produce thefrequency-leap in tubeV This delay causes a fault of phase-in the compensation and therefore it must be kept as small as possible. Fig. 5 shows the diagram of an arrangement for producing 01 the necessary impulse voltage. The filter F2 which feeds the tube V2 in Fig. 4 is also connected to the grid of the high frequency amplifier V3 which operates as an aperiodic amplifier. Aperiodic amplifying is chosen since this operates without any delay while amplification with tuned anodecircuits requires a finite time till the amplitude of the amplified oscillations will achieve its final value. The amplification factor of the tube V3 determines the lower limit of the noise-oscillations at which the arrangement for noise-eliminating will operate. A smaller noise-voltage produces an impulse voltage which is not able to cause the frequency-shift in the oscillator-modulator tube V2. This lower limit of response of the noise-eliminating arrangement ought to be directly proportional to the amplitude of the signal voltage. For this purpose the A. V. C. of the receiver acts also on the tube V: as is indicated by the arrow in Fig. 5. In the case of a strong signal the lower limit of response will be larger, while with a weak signal the lower limit will be lower. The oscillations amplifier by the tube V3 will be rectified in the diode D1. In Fig. 6 the upper curve a shows the shape of the oscillations according to Equation 2. The curve b shows the shape of the detected voltage. The dotted line corresponds to the envelope of curve a. The condenser C1 is so chosen that the time-consonant of charging it through the diode D1 is so small that the voltage across it at any time follows the increasing voltage curve. When the voltage according to curve a is decreasing (dotted line) the condenser will be discharged through the resistance R1, the value of which is so chosen, that the exponential curve of discharging is a little slower than the decay of the oscillations according to curve a. The time-curve of the voltage across the condenser C1 which is impressed on the grid of tube V4 is shown in the curve b in Fig. 6. For the anode-load of V4 is chosen a pure inductance, the value of which is, at the frequencies in question, much smaller than the plate-resistance of .V4. Under this condition the anode-current has the same wave-form as the grid-voltage and the voltage drop across the anode coil is where L is the inductance and the time derivative of the anode-current. The shape of this voltage is shown by the curve c in Fig. 6. As it is seen the voltage arises suddenly, yet this rise will be a little flattened by the unavoidable capacitances (these of the tubes etc.) connected parallel to the load coil. The voltage according to the curve 6c coming from the tube V4 is impressed on an arrangement which will be hereinafter called a Kipp-amplifier. The Kipp-amplifier according to the invention is a back-coupled amplifier in which the feed-back voltage acts through a rectifier and therefore only in one direction. Fig. '7 shows a diagram of a Kipp-amplifier according to the invention. The tube V5 is connected with condenser-resistance coupling to the tube Vs which serves as a phaseinverting tube to make possible the back-coupling in correct phase. To the anode of tube V6 is connected the diode D2 in series with the condenser C2 and resistance R2 and to the common-point P of diode D2 and condenser C2, the grid of the tube V5 connected. If a positive grid-Voltage is suddenly impressed on the tube V5, the back-coupling does not act since the cathode of the diode D2 becomes more positive thanits anode and therefore no current will flow throughit, but if the voltage at the grid of tube V5 is negative, the cathode of the diode D2 becomes more negative than its anode and a strong current will start through it causing the negative voltage of the grid G1 to increase rapidly until the anode current in the tube V5 will be suppressed and the condenser C2 will be negatively charged so that the anode of the diode D2 becomes as negative as its cathode. After this the condenser C2 will be discharged through the resistance R2. The time-curve of charging and discharging of the condenser C2 is shown by the curve d inFig. 6. g It consists of two exponential curves. The rising part has the time-constant wherein A denotes the amplification factor of V5, S the transconductance of V6 and C2 the capacitance. The falling part has the time constant C2R2 since no back-coupled voltage is acting, Where R2 is so chosen that the discharging will be a little slower than the decay of the oscillations in the filter F2 in Fig. 4. Supposing the diode D2 to have an ideal rectifier curve, every negative voltage-impulse of any small finite value will cause voltage-variation at the grid G1 of valve V5 according to the curve 6d. In reality the diode D2 has for very small positive voltages a very high resistance so that it may be considered as an open circuit and therefore it will set a lower limit of response for the negative impulse acting on the grid G1 of tube V5. 7

Fig. 8 shows by way of example the diagram of a receiver containing the noise-eliminating arrangement described above. The input filter acting on the octode V0 is deliberately damped so that its'bandwidth (about 20 kc.) is twice more than the signal frequency range. The octode V0 transforms the input signal in a high I. F., J (about 1600 kc.) the signal-filter F1 is tuned to this. To a closely neighbouring frequency (cca. 10 kc. higher or lower) the auxiliary filter F2 is tuned. The filter F1 is connected to the octode V1 which transforms the signal-oscillations of the frequency f to those of the second I. F. ii (about 460 kc.) the filter F1 is tuned to this. The auxiliary filter F2 is connected to the grid G4 of the octode V2 and feeds in addition the Kipp amplifier of Fig. '7 through the high fre quency amplifier shown in Fig. 5. The Kipp amplifier acts on the first grid of the octode VF, the fourth grid of which is parallel-connected to the oscillator circuit of the octode V2. If there is no noise. voltage in the aerial, the incoming signal will be changed in frequency by the octode V6 and passing through the filter F1 will feed the second frequency changer octode V1. The signal thus transformed passes the filter F1 tuned to the second I. F. frequency f1 which feeds the I. F. amplifier V1 and the following audio frequency part A of the receiver built up as usual. The octode V2 is not operating since the signal cannot pass the filter F2 and thus its input grid will be not excited. If a noise-voltage occurs in the aerial the damped fluctuating oscillations occurring in the filter F2 will feed the valves V2 and V3. This latter acts on the Kipp amplifier and causes a negative voltage according to Fig. 6d on the grid G1 of the octode VF and the electroncurrent in it will be suppressed suddenly. This causes a decreasing of the 'c'apacitanoegrid-Q cathodeof Vr'and thus the necessaryfrequencyshift in the oscillator frequency-of the'octode V2. The I. F. voltages delivered from the tubes V1 and V2' in the filter Fl during the noise-impulse correspond to the Equations 4 and :6 and thus cancel each other. Only the signalvoltagewill reach the I. F. amplifier V1. After the noise-oscillations in Fz'have decayed (c021. /3000 sec. in usual circuits) the negative voltageon the first grid o'f VF will disappear and the oscillator frequency in V2 will becomeagain the same value 'asthatof tube V1. The phase of the oscillator-voltage in V2 will be also corrected by the synchronizing voltage due to the coil S connected to the anode of V1 and coupled to the oscillator circuit of V2 and thus both the frequency andphase become again the same as in the octode V1. The whole arrangement is again ready to eliminate a new noise-impulse.

Having now particularly described and ascertained the nature of my said invention and in what manner the same is to be performed, I declare that what I claim is:

1. In a radio receiver, an antenna, a first filter for selecting the transmitted signal from said antenna, and a second filter for se ecting from said antenna an auxiliary frequency range close to said transmitted signal, oscillators associated with each of said filters for producing an intermediate frequency signal in each filter, and means operative upon the receipt of a disturbance for causing the frequency of the'oscillator associated with said auxiliary frequency range to change by an amount equal to the tuning 'difierence between the two'filters and means for combining the disturbance signal outputs of said first and second filter in opposed relation for balancing out disturbances.

2. In a radio receiver; an antenna, a first filter for selecting the transmitted signal from said antenna, and a second'filter 'for selecting from said antenna an auxiliary frequency range close to said transmitted signal, oscillators associated with each of said filters for producing'an intermediate frequency signal in each filter, means operative upon the receipt of a disturbance for causing the frequency of the oscillator associated with said auxiliary frequency range to change by an amount equal to the tuning difference between the two filters, means for combining the disturbance signal outputs of said first and second'filter in opposed relation for balancing out disturbances, and means for synchronizing the voltages of said oscillators when no disturbances are being received.

3. In a radio receiver, an antenna, a first filter for selecting the transmitted signal from said antenna, and a second filter for selecting from said antenna an auxiliary frequency range close to said transmitted'signal, oscillators associated with each of said filters for producing an intermediate frequencysignal in each filter, means operative upon the receipt of a disturbance for causing the frequency of the oscillator associated with said auxiliary frequency range to change by an amount equal to the tuning difference between the two filters, means for combining the disturbance signal outputs of said first and second filter in opposed relation for balancing out disturbances, and means for synchronizing the voltages of said oscillators when no disturbances are being received, said last mentioned means comprising a coupling coil for interconnecting said oscillators.

4. In a radio receiver, an antenna, a first filter for selecting the transmitted signal from said antenna, and a second filter for selecting from said antenna an auxiliary frequency range close to said transmitted signal, oscillators associated with each of said filters for producing an intermediate frequency signal in each filter, means operative upon the receipt of a disturbance for causing the frequen y of the oscillator associated with said auxiliary frequency range to change by an amount equal to the tuning diiference between the two filters, said means comprising a multigrid tube having one grid joined to said lastmentioned oscillator and another grid joined to said second filter, means for combining the disturbance signal outputs of said first and second filter in opposed relation for balancing out disturbances, and means whereby the eifective capacitance of said tube is varied when a negative voltage is impressed upon said first grid.

5. In a radio receiver, an antenna, a first filter for selecting the transmitted signal from said antenna, and a second filter for selecting from said antenna an auxiliary frequency range close to said transmitted signal, oscillators associated with each of said filters for producing an intermediate frequency signal in each filter, means operative upon the receipt of a disturbance for causing the frequency of the oscillator associated with said auxiliary frequency range to change by an amount equal to the tuning difference between the two filters, said means comprising a multigrid tube having one grid joined to said lastmentioned oscillator and another grid joined to said second filter, means for combining the disturbance signal outputs of said first and second filter in opposed relation for balancing out disturbances, a negative feedbackamplifier in circuit with said multi-grid tube, said feedback being supplied through a half wave rectifier, circuit means for impressing the output of said amplifier upon said multi-grid tube for varying the eifective capacitance thereof.

MARTIN WALD.

Images