GB2213337A

GB2213337A - Removing unwanted modulation

Info

Publication number: GB2213337A
Application number: GB8828173A
Authority: GB
Inventors: Henry William Hawkes
Original assignee: UK Secretary of State for Defence
Current assignee: UK Secretary of State for Defence
Priority date: 1987-12-07
Filing date: 1988-12-02
Publication date: 1989-08-09
Anticipated expiration: 2008-12-02
Also published as: GB8828173D0; GB2213666A; GB8728577D0; GB2213666B; GB8828175D0; GB2213337B

Abstract

In a normaliser circuit, unwanted modulation accompanying a modulated carrier is removed or reduced considerably by detecting the carrier at 11, combining at an adder 13, the unwanted modulation received at y and an integrated version of the detector output, and multiplying at 9 the original carrier with the adder output supplied via a divider 14 (i.e. a basic normaliser) receiving a control voltage v. For removal of a particular modulation frequency band, the detector 11 output is supplied to adder 13 via a band-pass filter 15. Where the modulation band includes unwanted signals, a version of the carrier including the unwanted signals only is detected at 17 and applied to the adder 13 via band-pass filter 16. Further removal of the undesired components is achieved by multiplying at 22 the carrier extracted at 20,21 with the output of adder 13. A preceding normaliser 25 removes other undesired modulations e.g. due to fading. <IMAGE>

Description

Normaliser Circuits This invention relates to normaliser circuits, in particular to normaliser circuits in which an AC signal is divided by a control amplitude to give, as quotient, an AC signal whose amplitude is related, as a multiple or fraction, to the control amplitude. Such a normaliser circuit is described in our copending GB Appln No 8712556 (Publn No 2, 191, 321A), where the control voltage is termed a reference amplitude.

The present invention provides normaliser circuits which additionally can filter-out selected modulating frequencies of an AC carrier signal while normalising others. It has one application in normalising such signals prior to detection of the wanted modulation.

According to the present invention a normaliser circuit comprises: a linear detector per se having an input connection for a modulated carrier input signal including unwanted modulation; an integrator connected to receive the detector output; an adder having one input connected to receive the integrator output and another input connectable to receive unwanted modulation; a divider having its denominator input connected to the adder output and its numerator input connectable to receive a control voltage; and a multiplier having one input connection for said modulated carrier input and its other input connected to the divider output.

Said normaliser circuit may also comprise a band-pass filter for passing unwanted modulation connected between the integrator input and said other input of the adder.

The normaliser circuit may further comprise: a second linear detector per se having a second input connection for a version of said first-mentioned input signal having different proportions of the wanted and unwanted modulation, in the same modulation band, from that of the first-mentioned input signal; a band-pass filter passing said unwanted modulation and connected to receive the output of said second detector, and a connection between the filter output and said other adder input for feeding said unwanted modulation to said adder in substantially the same phase as the unwanted modulation of said first-mentioned input signal and at an amplitude substantially to cancel the unwanted modulation.

The output of said multiplier may be connected via a carrier-passing filter to one input of a second multiplier whose other input is connected to the input of the divider.

The present invention will now be described, by way of example, with reference to the accompanying drawing, which is a block schematic circuit diagram of a normaliser circuit embodying the present invention.

Referring to the drawing and neglecting for the moment normaliser 25, as in Appln No 8712556 the modulated carrier input signal is fed to a multiplier 9 and to a linear detector 10 comprising a detector element 11, eg a diode, plus an integrating circuit 12 whose output is fed via an adder 13 (which may be simply a resistor network) to a divider 14 as denominator. As in the aforesaid Appln, the divider receives as numerator a control voltage V and its output is fed to multiplier 9 constituted in a known manner by eg a balanced square-law mixer. (Although not stated in terms, it is implicit in that Appln that the linear detector includes an integrator which passes slow variations but stops the carrier frequency).

Consider first the case where the unwanted modulation is additional to a wanted modulation and is of relatively low frequency. The time-constant of integrator 12 can be arranged if required to allow passage of such low frequencies. The relatively high-frequency wanted modulation will not pass through circuit 12, and the nett effect is an output from circuit 9 in which very lowfrequency amplitude modulations originally present have been substantially eliminated. The adder 13 is not required in this case.

Secondly, consider the case where it is desired to remove a particular modulation frequency band from the input signal. This is achieved by connecting a bandpass filter 15 between terminals X and Y, whose pass band is that of the modulation frequency it is desired to remove. This by-passing of the usual path from linear detector to divider allows parts of the modulation spectrum to be substantially removed. The integrator 12 can be removed or retained as the content of the total input signal requires. It will usually be retained, in which case the adder 13 combines its output with that of filter 15.

Thirdly, consider the case where it is desired to remove a particular modulation frequency band but only in respect of unwanted signals in that band. This is achieved by removing filter 15 and connecting to terminal Y a similar filter 16 connected to a detector element 17 similar to element 11. The input to element 17 is a version of the input to circuit 9 which preferably only contains unwanted modulation in the pass band; it is frequently possible to provide such a version in which the unwanted signal is alone present. The amplitude of this input is adjusted by the amplifier/attenuator 18 to obtain the optimum cancellation of the unwanted signal, ie to equal that appearing at terminal X which in effect it replaces.Since the phase of the unwanted signal modulation added at Y should be as close as possible to that of the original signal to multiplier 9, a variable delay circuit may be required at 19 or at the input to the normaliser, in order to match the timing of the two signals.

Although preferred, it is not strictly essential that the unwanted signal should predominate in the added version.

The essential requirement is that the ratio of the wanted/ unwanted signals should be different to the two signals.

However, if the wanted signal predominates in the added version, care must be taken that the resulting output from circuit 9 will not be over-modulated by the wanted signal, and the added version must be selected accordingly. Again the integrator 12 and adder 13 can be omitted if not required.

It will now be convenient to summarise the foregoing and express the results in mathematical terms.

The original signal input to the modified normaliser can be expressed as Signal - S(1 + A + B + C)(1 + D) sin w t - S(1 + A + B + C + D + AD + BD + CD)sin w t in additive form where w t - RF carrier phase A e unwanted modulation not in the band of interest and usually, but not necessarily, of higher frequency B = wanted modulation defining the band of interest C e unwanted modulation, in the band of interest, of a separate version of the original signal preferably, but not essentially, containing substantially only C D = unwanted very low frequency fading which modulates the whole signal including other modulations, ie is multiplicative modulation S = general amplitude level excluding any modulation as above The various modes of operation of the modified normaliser for AVC and additive modulation removal are then as follows: (1) The basic mode (corresponding to the normaliser disclosed in copending Application No 8712556) The time-constant of integrator 12 is arranged to pass all the detected modulation S(1 + A + B + C)(1 + D) and the normaliser output I sinu t (x a constant P dependent on the numerator V and the characteristics of the multiplier).

Henceforth this constant P will be implicit in the final output of the normaliser.

(2) A limiting mode The mode (1) above provides an amplitude-limiter for preventing the RF signal exceeding a predetermined amplitude without merely clipping the signal at that amplitude and thus introducing distortion.

(3) An AVC mode The time-constant of the integrator 12 is arranged to pass only the slow modulation D.

Then output - S(1 4 A + B + C)(1 + D) sin xt S(1 + D) - (1 + A + B + C) sin wt Thus the effect of eg fading has been eliminated giving an output signal of amplitude (1 + A + B + C). This modification to the normaliser of copending Appln No 8712556 provides the normaliser 25, which is connected, if needed, to the input of the normaliser shown in detail ie preceding and in series therewith, to eliminate modulation due, in particular, to fading.

Modes (1), (2) and (3) are cited for interest but do not use embodiments of the present invention.

(4) Mode for eliminating A (C - 0) The time-constant of circuit 12 is arranged to pass only DC and filter 15 is connected to pass A. Then Output of adder 13 - S(1 + A) whence output of multiplier 9 - S(1 + A + B) sin Xt 5(1 + A) B - [ 1 + - ] sin Ut l+A which, after filtering and subsequent detection, becomes B l+A The contamination of B by l/(l + A) can be removed by multiplying by S(l + A). The output of multiplier 9 is detected by a linear detector 20, followed by a high-pass filter 21 to remove the DC component but pass the carrier frequency, and a multiplier 22 similar to multiplier 9. The 5(1 + A) input to multiplier 22 is obtained from the input to divider 14 as shown.

(5) Mode for eliminating C (A - 0) The arrangement is the same as for mode (4) above, except that filter 15 is omitted and a version of the input containing preferably, but not essentially, only C is fed in at terminal Y as already described. Detector 20 et seq can be used to eliminate the 1/(1 + C) factor in this case as in mode (4).

If a version containing zero B cannot be obtained but one of 5(1 + C + kB)sin wt is available, where S corresponds to S in this version, then provided k is not unity and not large, operation is still possible as follows: The input into terminal Y is now 5(1 + C + kB), whence the output of multiplier 9 becomes S(1 + C + B) ---- . sin #t, which, provided Kb is small kB S (l + C)(1 + - ) l+C

After processing by detector 10 et seq, this becomes

e S ( ~ kB + B) provided B is small S Thus provided k is not unit, B can still be obtained even for non-zero values of k.

(6) Mode for eliminating A and C The arrangement is the same as for modes (4) and (5) except that filter 15 is retained in addition to the aforesaid version of the input being fed in at terminal Y.

The adder output is now (1 + A + C) and the output of the normaliser is S(1 + A + B + C) sinwt S(1 + A + C) = B 1 + ----------- sin w t l+A+C After detection and processing by detector 20 et seq the final output becomes B, as required.

In modes (4) - (6) above it is assumed that the contaminated modulation, eg B/(1 + A + C + D) above, is of such level as not to cause overmodulation, ie B(1 + A + C) < 1. However, the carrier term 1 x sin t could in principle be zero. In that case a known no-carrier double-sideband form of demodulator would be required to replace the simple detector 20 in order to detect B (1 + A + C).

In the above description it has been assumed that the unwanted modulation C has substantially the same carrier frequency as the wanted modulation B. If these two carrier frequencies are different, a modulation of the composite waveform occurs at the beat frequency. This modulation is unwanted and must be removed. If the beat frequency is substantially different from the wanted modulation band of B, it can be removed by the technique described for the removal of A, ie using the normaliser in mode (4) above. If it is in the same band as B, it can only be removed if a separate version, not containing the wanted modulation, or containing a different ratio of wanted/unwanted modulation can be obtained, as in mode (5) above. The foregoing assumes that the unwanted carrier signal is much weaker than the wanted carrier signal.

It was stated earlier that the constant P is dependent on the numerator V and on the characteristics of the multiplier 9. It can be shown that, with available multipliers, the latter characteristics cause the amplitudes of the normalised wanted output signal B, and of any wanted modulation sidebands not subjected to the normalising process, to be reduced to a level equal to or determined by the modulation depth of any unwanted modulation. The wanted signals can be amplified subsequently if required.

Claims

1. A normaliser circuit comprising: a linear detector per se having an input connection for a modulated carrier input signal including unwanted modulation; an integrator connected to receive the detector output; an adder having one input connected to receive the integrator output and another input connectable to receive unwanted modulation; a divider having its denominator input connected to the adder output and its numerator input connectable to receive a control voltage; and a multiplier having one input connection for said modulated carrier input and its other input connected to the divider output.

2. A circuit as claimed in claim 1 comprising a band-pass filter for passing unwanted modulation connected between the integrator input and said other input of the adder.

3. A circuit as claimed in claim 1 or claim 2 comprising: a second linear detector per se having a second input connection for a version of said first-mentioned input signal having different proportions of the wanted and unwanted modulation, in the same modulation band, from that of the first-mentioned input signal; a band-pass filter passing said unwanted modulation and connected to receive the output of said second detector, and a connection between the filter output and said other adder input for feeding said unwanted modulation to said adder in substantially the same phase as the unwanted modulation of said first-mentioned input signal and at an amplitude substantially to cancel the unwanted modulation.

4. A circuit as claimed in claim 2 or claim 3 wherein the output of said multiplier is connected via a carrier-passing filter to one input of a second multiplier whose other input is connected to the input of the divider.

5. A normaliser circuit substantially as hereinbefore described with reference to the accompanying drawing.