GB2246921A

GB2246921A - Demodulation of amplitude modulated carrier waves

Info

Publication number: GB2246921A
Application number: GB9016376A
Authority: GB
Inventors: Peter Richard Dent
Original assignee: Texas Instruments Ltd
Current assignee: Texas Instruments Ltd
Priority date: 1990-07-25
Filing date: 1990-07-25
Publication date: 1992-02-12
Anticipated expiration: 2010-07-25
Also published as: GB2246921B; GB9016376D0

Abstract

An amplitude modulated carrier wave 1 is demodulated by a process including sampling and digitising it 2 at a rate exceeding the Nyquist rate for the signal, and the digitised samples are used to produce two successions of values, the first 4 at a rate at least twice that of the highest modulation frequency (for a short term energy calculation) and the second 5 at a rate that is less than half the lowest modulation frequency (for a long term energy calculation), each value being proportional to the mean energy of the digitised samples which have occurred since the last value in the same succession was produced. An output signal representing the modulation signal is produced by calculation 6 from the difference between the concurrent values from the two successions. Digital filtering 3 may be applied to the digitised samples before processing to limit the bandwidth of the demodulator to that of the modulated carrier wave. Digital smoothing may be applied after the calculation to reduce noise introduced by the calculation processes. The output signal may be converted to analogue form with a digital to analogue converter 9. <IMAGE>

Description

IMPROVEMENTS IN OR RELATING TO THE DEMODULATION OF AMPLITUDE MODULATED CARRIER WAVES The present invention relates to the demodulation of amplitude modulated carrier waves.

The conventional way in which an amplitude modulated carrier wave is demodulated is to detect the varying peak signal strength using a rectifier and then averaging the rectified signal, the averaging being over a period of sufficiently long duration to smooth out the carrier wave itself. Demodulation in that way is simple and cheap to implement and produces a satisfactory output signal provided that the signal to be demodulated is neither of too small nor of too large amplitude and is not subjected to rapid changes in amplitude additional to the intended modulation, introduced by changes along the transmission path.Radio-frequency and/or intermediate-frequency amplifiers of high gain with automatic gain control can compensate for much of the undesirable variation in amplitude occurring in the signal transmission path, but undesirable distortion of the demodulated signal can occur which is not corrected by automatic gain control.

The demodulation itself makes use of the nonlinearity of the characteristic of a diode or of an amplifying device such as a transistor or a triode, and the presence of the necessary non-linearity also introduces distortion into the modulation component derived by the demodulation. That distortion is clearly undesirable.

It is an object of the present invention to provide an improved method and apparatus for demodulating an amplitude modulated carrier wave.

According to the present invention there is provided apparatus for demodulating an amplitude modulated carrier wave signal including means for deriving from the signal a sequence of digitised samples at a sampling rate exceeding the Nyquist rate for the signal, means for calculating a succession of short term values of the signal from the sequence of digitised samples, each value being calculated from samples occurring in a short time section of the sequence, the reciprocal of the short time being at least twice that of the highest modulation frequency, means for calculating a succession of long term values of the signal from the sequence of digitised samples, each value being calculated from samples occurring in a long time section of the sequence, the reciprocal of the long time being less than half the lowest modulation frequency, and means for forming the ratio between concurrent short term and long term values, and deriving an output signal representing the amplitude of the modulation signal.

The short term and long term values calculated from the sequence of digitised samples may represent the energy of the signal and the derivation of the output signal representing the amplitude of the modulation signal may include calculating the modulation signal amplitude corresponding to the difference between the two energy values.

The apparatus may include means for filtering the sequence of digitised samples before the calculation of the long term and short term values so that the samples are restricted to the bandwidth of the modulation components of the carrier wave signal.

The means for deriving an output signal may include digital to analogue conversion means so that the output signal is an analogue signal. Waveform shaping means may also be provided to smooth and filter the digital signal before conversion to analogue form.

The means for calculating the long term values may comprise a low pass digital filter connected to receive the short term values.

The apparatus may include a suitably programmed digital signal processor to perform some or all of the digital signal processing required.

According to a second aspect of the present invention there is provided a method of demodulating an amplitude modulated carrier wave signal in which the signal is converted into a succession of digitised samples at a sampling rate exceeding the Nyquist rate for the signal, the digitised samples are processed to provide two successions of values dependent on the magnitude of the samples, the two successions of values being derived from samples occurring during time sections of different lengths, the reciprocal of the time section for values of the first succession being at least twice the highest frequency of the modulation and the reciprocal of the time section for values of the second succession being less than half the lowest frequency of the modulation, and an output signal representing the modulation signal is derived from the ratio between concurrent values of the two successions.

Each value of the successions of values dependent on magnitude of the samples may be proportional to the mean energy of the signal over the period since the last value of the particular succession was produced. The derivation of the modulation signal may include calculating the amplitude of that signal from the square root of the energy ratio represented by the ratio between the values of the two successions.

The succession of digitised samples may be subjected to filtering to reduce extraneous components before they are processed to produce the two successions of values.

The output signal may be in analogue form obtained by digital to analogue conversion of digital values derived from the difference between the values of the two successions. Some waveform shaping may be applied to the digital values to smooth and filter them before the digital to analogue conversion.

The long term and short term values may be derived from the squares of the samples, as when they represent the energy of the signal, or they may be derived from the samples without squaring but with rectification of the polarity. The processing to derive the output signal from the difference between the long term and short term values will depend on whether squaring or rectification is used to produce the values.

The long term values may be derived directly from the samples or indirectly by low pass filtering the short rerm values.

Among the advantages of a demodulator according to the invention is that the modulation can be derived from an input signal without substantial error and distortion.

In particular, the use of digital calculation can ensure the linearity of the demodulation process.

Examples of the method of demodulating an amplitude modulated carrier wave and apparatus for carrying out the method will now be described with reference to the single figure of the accompanying drawing which is a flow diagram of an example of the method.

Referring to the drawing, an input signal in the form of an amplitude modulated carrier wave at radio frequency (RF) or at intermediate frequency (IF) enters at 1 and is applied to an analogue to digital converter 2. The sampling rate of the analogue to digital converter 2 must exceed the Nyquist rate for the input signal. The digital outputs of the converter may be in serial or parallel form to suit the subsequent digital processing. Conveniently the output is in 8-bit or 16-bit words.

Digitised samples produced by the converter 2 are then subjected to a digital filtering process in the block 3 for the purpose of centring the response bandwidth of the demodulator on the carrier wave. This filtering can be arranged to reduce the effect of spurious pulses and other noise signals on the demodulated output signal.

The digitised samples after filtering are used separately in the calculation of short term energy values in block 4 and in the calculation of long term energy values in block 5. The short term energy values are obtained by squaring the digitised samples occurring during short time sections and summing the squares or calculating the average value of the squares over the period since the last time the value was worked out, each value being derived during a respective short time section, the reciprocal of which is at least twice the highest modulation frequency that the demodulator is required to produce as output. The long term energy values are derived in the same way as the short term energy values, but during long time sections, the reciprocal of one such section being lower than half the minimum modulation frequency that the demodulator is required to produce as output.If, for example, the required range of modulation frequencies is from 20 to 20,000 Hz, then the long term energy values are derived at a repetition rate which is lower than 10 Hz, and the short term energy values are derived at a repetition rate which is higher than 40,000 Hz.

In block 6 the ratio between the two energy values existing at the time is worked out, either by simple division if the energy values represent the averages of the values of the squares, or, if the energy values are the sums of the squares, by dividing the long term energy values and/or multiplying the short term energy values to compensate for the different numbers of samples used to form the values and then dividing one value from the other. The ratio between the values should be calculated at the same repetition rate as the derivation of the short term energy values.

The energy ratio values from block 6 are used in block 7 for the calculation of the amplitude deviations corresponding to the energy differences.

If the amplitude modulated carrier signal is A (1 + M.cos wmt).cos wct where A is a constant M is the modulation index wm is the modulation frequency and wc is the carrier frequency; the energy of the signal is (by squaring) B (1 + M.cos wmt)2.cos2wct where B is a constant.

The long term average energy EL is B.cos2wct, because (1 + M.cos wmt) averages out to the value 1.

In the short term the averaging period is less than the period of the modulating waveform so that the modulation factor M.cos wmt can be regarded as a constant X for a particular period. The short term average energy E5 becomes B(l+X)2cos2wct.

It follows that the short term average energy Es equals (1+X)2 times the long term average energy EL. From that relationship we obtain

which represents the calculation carried out in the block 7.

If the modulation index M is small, say 0.1, then a simpler calculation can be used to derive the modulation ES - EL 1 ES X = = 2 '- - 1 2EL 2 EL at the cost of some distortion of the modulation signal obtained.

In block 8 some digital processing is carried out to smooth and filter the digital values obtained from the amplitude deviation calculation performed in block 7.

This processing should be designed to eliminate any noise introduced into the digital values by the calculations performed in blocks 4 to 7.

Finally, the digital values from block 8 are converted to analogue form by a digital to analogue converter 9 and are produced as an output signal at 10.

As described above, the long term energy values are derived directly from the digitised samples. Instead of doing that the long term energy values could be derived from the short term energy values by low pass filtering.

In place of the energy calculations in blocks 4 and 5 and the calculation of the amplitude deviation in block 7, the demodulator could derive long term and short term averages of rectified values by converting all negative digital values obtained from the converter 2 to positive digital values and then averaging all the digital values over the long term and short term periods as described above. Again, the long term values could either be derived directly from the digitised samples or indirectly by low pass filtering the short term values. Using the notation employed above: the short term average rectified value RS = (1+X) RL where RL is the long term average rectified value.

From which RS X = - 1 RL which represents the calculation to be performed in the block 7, the block 6 forming the ratio between the values from the blocks 4 and 5.

Although a demodulating apparatus could be constructed from digital components respectively dedicated to the calculations and other operations performed in the blocks 2 to 9, the apparatus is preferably constructed from a digital signal processor programmed to perform those calculations and other operations with an input analogue to digital converter and an output digital to analogue converter. The apparatus may alternatively include dedicated digital circuits for some of the calculations and/or operations with the remainder being performed by a suitably programmed digital signal processor.

A model of the apparatus has been used to demodulate an amplitude modulated radio-frequency signal of 198 kHz with a bandwidth of 9 kHz. The analogue to digital conversion in block 2 used a sampling rate of 1 MHz. The block 3 filtered the digital sample stream to an 18 kHz band of data centred on 198 kHz. The short term energy values were calculated every 25 As to produce 40,000 values per second. The long term energy values were calculated every 50 ms to produce 20 values per second. The energy difference values and the amplitude deviation values were produced every 25 ps, and after filtering the digital values were converted to analogue form.

Claims

CLAIMS:

1. Apparatus for demodulating an amplitude modulated carrier wave signal including means for deriving from the signal a sequence of digitised samples at a sampling rate exceeding the Nyquist rate for the signal, means for calculating a succession of short term values of the signal from the sequence of digitised samples, each value being calculated from samples occurring in a short time section of the sequence, the reciprocal of the short time being at least twice that of the highest modulation frequency, means for calculating a succession of long term values of the signal from the sequence of digitised samples, each value being calculated from samples occurring in a long time section of the sequence, the reciprocal of the long time being less than half the lowest modulation frequency, and means for forming the ratio between concurrent short term and long term values, and deriving an output signal representing the amplitude of the modulation signal.

2. Apparatus according to claim 1, wherein the short term and the long term values calculated from the sequence of digitised samples are proportional to the energy of the signal over the short term period and the long term period respectively and the means for deriving the output signal includes means for calculating the modulation signal amplitude corresponding to the ratio between the two energy values.

3. Apparatus according to claim 2 wherein the modulation signal amplitude is calculated as the square root of the quotient of the two energy values, minus unity.

4. Apparatus according to claim 1, wherein the short term and the long term values calculated from the sequence of digitised samples are proportional to the average values of the digitised samples occurring during the short term periods and the long term periods respectively after rectification of their polarities.

5. Apparatus according to any preceding claim including means for filtering the digitised samples before calculation of the long term and short term values so that the samples are restricted to the bandwidth of the modulation component of the carrier wave signal.

6. Apparatus according to any preceding claim further including digital to analogue conversion means for converting the output signal to analogue form.

7. Apparatus according to any preceding claim including waveform shaping means connected to receive the digital signals representing the amplitude of the modulation signal and effective on those digital signals to reduce the noise in the output signal resulting from the calculations used to form the digital signals representing the amplitude of the modulation signal from the digital samples.

8. Apparatus according to any one of the preceding claims, wherein the means for calculating the long term values comprises a low pass digital filter connected to receive the short term values.

9. Apparatus according to any preceding claim including a suitably programmed digital signal processor to perform some or all of the digital signal processing required.

10. Apparatus for demodulating an amplitude modulated carrier wave signal substantially as herein described and as illustrated by the single figure of the accompanying drawing..

11. A method of demodulating an amplitude modulated carrier wave signal in which the signal is converted into a succession of digitised samples at a sampling rate exceeding the Nyquist rate for the signal, the digitised samples are processed to provide two successions of values dependent on the magnitude of the samples, the two successions of values being derived from samples occurring during time sections of different lengths, the reciprocal of the time section for values of the first succession being at least twice the highest frequency of the modulation and the reciprocal of the time section for values of the second succession being less than half the lowest frequency of the modulation, and an output signal representing the modulation signal is derived from the ratio between concurrent values of the two successions.

12. A method according to claim 11, wherein the values derived from the digitised samples are proportional to the mean energy of the signal over the period since the last values of the respective successions were derived, and the derivation of the output signal includes calculating the amplitude of the modulation signal represented by the ratio between the values of the two successions.

13. A method according to claim 10, wherein the modulation signal amplitude is calculated as the square root of the quotient of the two energy values, minus one.

14. A method according to claim 11, wherein the values derived from the digitised samples are proportional to the averages of the digitised samples occurring since the last values of the respective successions were derived, after rectification of the polarities of the digitised samples.

15. A method according to any one of claims 11 to 14, wherein, before the successions of values are derived therefrom, the digitised samples are subjected to filtering to reduce extraneous components in the samples.

16. A method according to any one of claims 11 to 15, including filtering the output signal representing the modulation signal to reduce the noise introduced by the processing of the digitised samples.

17. A method according to any one of claims 11 to 16, wherein the processing of the digitised samples to form the values of the second succession comprises low pass filtering the values of the first succession.

18. A method according to any one of claims 11 to 17, further including converting the output signal to analogue form.

19. A method of demodulating an amplitude modulated carrier wave signal substantially as herein described and as illustrated by the single figure of the accompanying drawing.