FR2699762A1 - Low rate digital phase modulator for telecommunications - Google Patents

Abstract

The modulator (1) separates the incoming signal (E) into real and imaginary components (2), and sub-samples them (3). These components are applied to Nyquist filters (4,5) and converted to polar form (amplitude/angle ) (6). The amplitude is split into word lengths (8) and converted to analogue form (9) then passed through a low pass filter (10).The angular component is passed through a phase interpolator (7) with an offset command frequency (CF) and a local oscillator (Fe) applied to it. The digital phase output is compared to a reference (REF) from the magnitude loop, and converted to analogue (12), then passed through a low pass filter. The output S is then produced as a modulated and smoothed signal.

Description

Phase modulation method and digital modulator implementing this method
The present invention relates to a low rate phase modulation method. It also relates to a low bit rate digital modulator implementing this method.

 The sampling techniques used in digital modulators equipping telecommunication systems in particular have the effect of producing a repetition of the spectra of the modulated signals. This repetition can not generally be tolerated and must be eliminated so that only the useful part of the spectrum remains.

 Currently, this repetition is annihilated by filtering after modulation by means of an "agile" band-pass filter in frequency or by oversampling of the signal.

But such a bandpass filter is difficult to achieve and the oversampling technique has limitations due to the calculation time.

 The object of the invention is to overcome these disadvantages by proposing a low rate phase modulation method which does not have a repetition of the modulated spectrum.

 According to the invention, the low rate digital modulation method for modulating input data and producing an intermediate frequency modulated output signal is characterized by comprising an initial step of filtering the input data. followed by a step of processing the filtered data comprising a step of processing the digital phase component including a step of modulating the phase of the processed data.

 Thus, with the method according to the invention, the signal is filtered before being applied to the digitally controlled oscillator and is processed in the phase domain. A pseudo-frequency modulation of the digitally controlled oscillator is then performed in order to obtain a smoothing function and thus to eliminate the repetition of the spectrum. A slight oversampling is then sufficient. This results in a direct intermediate frequency modulation with frequency agility for a wide variety of waveforms.

 According to an advantageous version of the method according to the invention, the initial filtering step comprises a first step of splitting the input data, a second step for coding and oversampling the data resulting from this splitting step and a step for filtering the respective module and phase of the split, coded and oversampled input data.

 According to another aspect of the invention, the low-rate digital modulator, for modulating input data and producing an intermediate frequency output signal, implementing the method according to the invention, comprising control modulation means numerical system, characterized in that it further comprises means for splitting the input data, means for encoding and oversampling the divided data in Cartesian coordinates, means for filtering the split and encoded data, means for Cartesian / polar conversion for converting the filtered data into polar coordinates and delivering digital module and phase components respectively, and means for processing the digital components including, as modulation means, phase modulation means for modulating the digital phase component, the data filtering means being located upstream of the modulator means phase transition.

Other features and advantages of the invention will become apparent in the description below. In the accompanying drawings given by way of non-limiting examples - FIG. 1 is a block diagram of a modulator
according to the invention - Figure 2 is a temporal representation
theoretical of a smoothing function implemented
in a modulator according to the invention ~ - Figure 3 represents the theoretical gain of the filter of
smoothing; FIG. 4 is a temporal representation
experimental smoothing function performed according to
the invention - Figure 5 shows the gain of the smoothing filter
experimentally recorded; FIG. 6 represents a spectrum of a signal of
output without smoothing the phase; FIG. 7 represents a spectrum of a signal of
output with application of the method according to the invention - Figure 8 illustrates an alternative embodiment of a
digital modulator according to the invention - FIG. 9 is a set of S1-S4 chronograms of
signals representative of the modulator of FIG. 8; FIG. 10 represents the spectrum of the output signal
a digital-to-analog converter implemented
in the modulator of FIG. 8 - FIG. 11 represents the spectrum of the output signal
of the modulator of Figure 8.

 We will now describe a particular embodiment of a low-rate digital modulator according to the invention with reference to FIG.

 The low-rate digital modulator 1 comprises a first stage 2 for splitting input data E.

The split data are then coded or addressed in a coding module 3 which generates two slightly oversampled signals applied respectively, via filters 4, 5, notably Nyquist type filters, to the two inputs of a Cartesian coordinate converter 6. in polar coordinates. This converter generates two digital components respectively of module r and phase 8. The module component r is applied to a decimator module 8 by a predetermined integer N. The output of the module 8 is connected to the input of a first digital converter 9, which generates a signal whose module is filtered by a first low-pass filter 10.The digital phase component O, including the frequency control and the modulation, is applied to a phase interpolator 7 having an input CF of frequency control. The output of the interpolator 7 is applied to the input of a digitally controlled oscillator 11 driven by an oscillation frequency Fe. This oscillator 11 generates a modulated digital signal which is inputted to a second digital converter. analog 12 receiving as a voltage reference a REF module reference signal from the first low-pass filter 10. The modulated analog signal from the second digital-to-analog converter 12 is applied to a second low-pass filter 13 which outputs a signal of output S intermediate frequency modulated and smoothed.

According to a theoretical smoothing function obtained by implementing the method according to the invention and illustrated in FIG. 2, between each discrete value of phase 8 (i), O (i + 1), O (i + 2) respectively obtained at instants iT, (i + 1) .T, (i + 2) .T, staircase interpolation is performed with an increment #O (i) / n, with:
n: number of samples per period T,
- # O (i) = O (i + 1) -O (i)
Te: signal sampling period (T = n.Te).

 To this theoretical smoothing function, corresponds a theoretical transfer function, represented in FIG. 3 in a Gain (dB) / reduced frequency (f / F) space, which has a symmetrical gain lobe 30 at zero frequency, with F frequency sampling of the modulating signal.

 The smoothing function actually performed by implementing the method according to the invention, shown in FIG. 4, differs from the theoretical function in that the interpolation function in stair steps is substantially offset with respect to the discrete values 8 (i), O (i + 1), O (i + 2), the offsets at these points corresponding to calculation residuals R (i-1), R (i) and each interpolation increment having the value ( 8 (i) + R (i-1)) / n ', n' being an approximation of the aforementioned magnitude n relative to the theoretical smoothing function.

 n 'is preferentially chosen such that n' = 2.P <n, this choice being dictated by the fact that a division by a power of 2 is very simple to realize on the material plane (shift). The obtained smoothing is an estimate of the theoretical value. The proposed modulator is a variable-rate modulator, the value of n 'to be determined at each change of flow.

 The transfer function actually obtained by implementing the method according to the invention, represented in FIG. 5, has a shape substantially similar to that of the theoretical transfer function, with however a slightly concave plateau for a reduced frequency range of 0. 4. The smoothing function actually carried out with the method according to the invention makes it possible to eliminate any repetition of the modulated spectrum.

The smoothing effect of the modulation phase obtained with the method according to the invention is particularly highlighted with reference to FIGS. 6 and 7 which respectively represent the spectrum of an output signal of a modulator without phase smoothing and the spectrum of an output signal of a modulator implementing the method according to the invention. The following notations are considered below:
fs: symbolic rate of flow of the modulating digital signal,
k: oversampling factor of the modulating signal before smoothing,
fo: frequency of the carrier delivered by the digitally controlled oscillator,
fe: clock frequency of the digitally controlled oscillator and sampling frequency of the intermediate frequency signal.

 The spectrum S (f) without smoothing of the phase has a main line at the frequency fo, secondary lines at the frequencies fo-2kfs, fo-kfs, fo + kfs, fo + 2kfs. There is a decrease D in 1 / f due to the blocking function of the digitally controlled oscillator. Moreover, the secondary lines at frequencies fo + kf s and fo + 2kfs correspond to spurious modulation spectra which must be attenuated by at least 50 dBc. In addition, the spectrum S (f) has an attenuation envelope A in sin (x) / x due to the blocking effect of the digital / analog converter.

 With the method according to the invention, a smoothing of the modulation phase is obtained and the spectrum SL (f) of the output signal always has a main line at the carrier frequency fo but on the other hand has a high attenuation of the parasitic lines PL , in particular with a decay in 1 / f2 provided by the smoothing function and the parasitic spectral lines in fo + kfs are weakened without further oversampling the modulation signal in Cartesian representation.

 Another embodiment of the digital modulator according to the invention can also be envisaged, making it possible to extend the operating range of the digital modulator to higher carrier frequencies by increasing the frequency of the sampling clock of the digital / analog converter. with reference to Figures 8 to 11.

 In this variant embodiment, with reference to the block diagram 80 of FIG. 8 and by way of nonlimiting example, the digital data Dn are sampled at a clock frequency H (input signal S2) generated by a counter digital signal 81 by M receiving as input an input signal S1 derived from a conversion sampling clock He which delivers, via a logic inverter 82, a sampling signal S3 to a digital / analog converter 84. The signal Ss the output of the digital-to-analog converter 84 is applied to the input of a band-pass filter 85.

With reference to FIG. 9, the digital / analog converter 84 takes into account an input signal S4 comprising a periodic sequence of elementary sequences comprising two phases.
a first phase with a null signal as long as the signal S2 (of frequency M times smaller than the input signal S coming from the counter 81 is at a logic high level,
a second phase containing a data item DI present at the input of the module 83 during the first positive edge of the conversion sampling signal S3, occurring during the transition to the low state of the sampling signal S2.

 This results in a decrease in the width of the sampling pulse which has the effect that the attenuation in sin (x) / x, related to the rectangular shape of the pulse, is sensitive only at high frequencies. .

It then becomes possible to select a carrier whose frequency is greater than the sampling frequency, and with a conserved quantization signal-to-noise ratio, thanks to band-pass filtering 85.

The frequency spectrum Ss (f) of the output signal of the converter 84 thus has a set of periodic lines around sampling clock frequencies fH, 2fH, 3fH and an attenuation envelope 100 canceling at the frequency of FHe conversion sampling. After bandpass filtering with an appropriate transfer function T, the digital modulator may output an output signal having a frequency spectrum S6 (f) having no repetition. The embodiment which has just been described thus makes it possible to modulate an intermediate frequency signal by means of a complementary fast logic (counter, inverter).

 The low bit rate digital modulator which has just been described can advantageously be implemented in the form of an integrated circuit according to now conventional design and manufacturing techniques.

 Of course, the invention is not limited to the examples that have just been described and many adjustments can be made to these examples without departing from the scope of the invention. Thus, the various modules constituting a digital modulator according to the invention can be designed in multiple ways depending on the available technologies, the frequencies involved and the operating conditions and environment of this modulator.

Claims (13)

 A method of low bit rate digital modulation for modulating input data (E) and producing an intermediate frequency modulated output signal (S), characterized in that it comprises an initial step of filtering said data of input (E), followed by a step of processing said filtered data including a step of processing the digital phase component (0) including a step of modulating the phase of said processed data.  2. Method according to claim 1, characterized in that the initial filtering step comprises a first step of splitting the input data (E), a second step for coding and oversampling the data resulting from said splitting step. and a step of filtering the respective module (r) and phase (8) of the split, coded and oversampled input data.  3. Method according to claim 2, characterized in that the processing step further comprises a step of converting Cartesian coordinates into polar coordinates applied to the filtered data to generate respective digital phase (8) and module (r) components. ).  4. Method according to claim 3, characterized in that the processing step further comprises a step of processing the digital module component (r) comprising a digital / analog conversion step followed by a pass filtering step. down to generate a reference signal (REF).  5. Method according to claim 4, characterized in that the step of processing the digital module component (r) further comprises a step of dividing said module digital component (r) by a predetermined number (N), performed before the digital / analog conversion step.  6. Method according to claims 4 or 5, characterized in that the step of processing the digital phase component (#) comprises a phase interpolation step, and a digitally controlled oscillation step according to a frequency of predetermined modulation (Fe), for generating a modulated digital signal.  The method according to claim 6, characterized in that the processing step further comprises a digital-to-analog conversion step for converting said modulated digital signal into a modulated analog signal with, for conversion reference, the reference signal ( REF) generated at the end of the step of processing the module component (r), followed by a low-pass filtering step for generating said smoothed output signal (S).  8. A low rate digital modulator (1) for modulating input data (E) and producing an intermediate frequency output signal (S), implementing the method according to any one of the preceding claims, comprising digitally controlled modulation means (11), characterized in that it further comprises means (2) for splitting said input data (E), means (3) for coding and oversampling said split data into Cartesian coordinates, means (4, 5) for filtering said fractionated, encoded and oversampled data, Cartesian / polar conversion means (6) for converting said filtered data into polar coordinates (r, @) and outputting digital components respectively of module (r) and phase (8), and means (7-13) for processing said digital components including, as said modulation means, phase modulation means (11) for modulating said digital phase component (8), said data filter means (4, 5) being located upstream of said phase modulation means (11).  9. Modulator (1) according to claim 8, characterized in that the digital component processing means (7-13) further comprises means (8-10) for processing the digital module component (r) of said converted data. and generating in return a module reference signal (REF) having digital-to-analog conversion means (9) for converting the digital module component (r) and low-pass filtering means (10).  10. Modulator (1) according to claim 9, characterized in that the digital component processing means (7-13) further comprises means (7, 11) for processing the digital phase component (8) comprising means phase interpolation method (7) and, under said digitally controlled modulation means, oscillator means (11) at a predetermined frequency (Fe) for generating a modulated digital signal.  11. Modulator (1) according to claim 10, characterized in that it further comprises digital / analog conversion means (12) for converting said modulated digital signal into a modulated analog signal according to, for conversion reference, said module reference signal (REF), and filtering means (13), for outputting an intermediate frequency output signal (S).  12. Modulator according to claim 11, characterized in that said digital / analog conversion means (84) are controlled by a clock signal (S3) derived from a conversion sampling clock (He), and in that that the oscillator means (83) are controlled by a clock signal (S2) from counter means by a predetermined number (M) whose count input is connected to said sampling clock (He), said means of filtering comprising a bandpass filter (85).  13. Modulator (1) according to any one of claims 8 to 12, characterized in that it comprises, under said data filtering means, two filter modules (4, 5), including type "Nyquist" , each placed on an input of said Cartesian / polar conversion means (6) and whose respective inputs are connected to the respective outputs of said coding means (3).