FR2652470A1 - Method and device for limiting the rise of secondary lobes in a power transmission installation for a digital single carrier having two or four phase states - Google Patents

Abstract

Method and device for limiting the rise of secondary lobes of the spectrum transmitted by a power transmission installation with a digital single carrier modulated by two or four phase states (BPSK or QPSK). Methods (18) are provided, within the modulator but after the baseband digital filtering (13), for detecting the envelope (R) of the modulating signal (A + iB), and other means (20, 24) are provided for performing phase pre-distortion [(A + iB)<i phi (R)>] able, a priori, to correct the phase distortion due to the non-linear amplitude/phase conversion by the power amplifier of the transmitter.

Description

The present invention relates to a method and to a device for limiting the rise of side lobes of the spectrum emitted by a power transmission installation with a digital single-carrier modulated with two or four phase states. It relates more particularly, but not exclusively, to the field of telecommunications by satellite using new digital transmission techniques.

Emissions in the direction of a satellite involve high powers due to the distance of the satellite. The power amplifiers used to generate these emissions are generally of the PROGRESSIVE WAVE TUBE (TOP) type,
KLYSTRON, or SOLID STATE. Their power at saturation, for obvious economic reasons, is always reduced to a minimum. It is nevertheless, in single-carrier transmission, provided sufficient to allow a satisfactory limitation of the rise of the secondary lobes which occurs on a digital carrier when the power to be transmitted approaches the saturation power, these secondary lobes then running the risk of disturbing the transmissions. on adjacent frequencies.

To give an order of magnitude, the international organization "INTELSAT" which currently holds the majority of international satellite traffic, specifies that the energy density emitted on the spectrum uplifts of a digital single-carrier in QPSK (modulation at four phase states), type IBS / IDR, must remain 26 dB lower than the average energy density of the carrier in its band. In the case of an amplifier with
TOP, this rejection of 26 dB is currently only possible if the output setback, or "Back-Off", of this TOP is in the vicinity of 5 dB, which penalizes the cost of the installation due to the relatively high saturation power that this PT must necessarily present in order to allow sufficient recoil to be obtained.

To overcome this drawback, it has already been proposed means for direct linearization of the microwave power amplifiers used in this type of transmitter. These means consist in reproducing, using an appropriate assembly, a non-linearity which compensates for that created by the power amplifier. Such equipment operates at the frequency of the amplifier, that is to say at microwave frequency, which leads to a non-negligible cost inherent in microwave technology, in particular for small or medium power transmitters.

The invention aims to restrict, without substantial increase in cost, the usable power of these power amplifiers, by limiting the sidelobe rises of the spectrum emitted by a power transmission installation with a digital single-carrier modulated to two or four. phase states (BPSK or QPSK modulation).

It consists, at the level of the modulator but after the digital filtering in baseband, in actually or fictitiously detecting the envelope of the modulating signal normally obtained at the output of this modulator, and in applying to the signal, still at the level of this modulator, a phase modulation, auxiliary and corrective, by this modulated carrier envelope, this phase modulation according to a modulation law opposite to that due to the Amplitude / Phase conversion of the power amplifier of the transmitter.

In any case, the invention will be well understood, and its advantages and other characteristics will emerge from the following description of a few non-limiting examples of embodiment, with reference to the appended schematic drawing in which - Figure 1 is the Amplitude / transfer function. Amplitude
of a PT power amplifier used in a
power transmission installation in the direction of a
satellite; - Figure 2 is the Amplitude / Phase transfer function of
this same TOP amplifier; - Figures 3 to 5 respectively represent the spectra
output obtained, by simulation, for the same
amplifier to TOP and for three input setbacks,
respectively 8 dB, 10 dB, 12 dB, to which
respectively correspond output setbacks of 2.9
dB, 4.4 dB, and 6 dB ;; - Figure 6 shows the same, for a 5 dB input recoil
giving an output recoil of 2.3 dB, the spectrum of
output that would be obtained by completely removing the
phase response of this same amplifier to PT; - Figure 7 shows the spectrum that would be obtained in the
same conditions as figure 6, but keeping only
a quarter of the phase variation due to the Amplitude / Phase response; - Figure 8 is a block diagram of a first device
implementation of the invention; - Figure 9 is a block diagram of another device
implementation of the invention; - Figure 10 is a variant of the device of Figure 9.

Referring first of all to Figures 1 and 2, there is shown - In Figure 1 the Amplitude / Amplitude 1 response curve of a PT power amplifier usable in a power installation, digital and single-carrier, in the direction of a satellite. On this graph, the hindsight
input is plotted on the abscissa, to one dB per square, and
the output recoil is plotted on the ordinate, also at
one dB per square.

- In figure 2 the Amplitude / Phase 2 response curve of this
same amplifier to TOP. Entry recoil is carried in
abscissa, at 5 dB / square, and the phase shift is carried
ordered, at 5 degrees / square.

These two characteristics are non-linear, and curve 1 shows that as the input level increases, there is some compression of the output level, which is not particularly surprising.

On the other hand, curve 2 shows that the phase, which one might expect to see substantially constant, varies considerably with the input level of the amplifier. The signal at the output of the amplifier is therefore finally both amplitude compressed and phase modulated as a function of the envelope of the signal modulating at the input of this amplifier. Now, it is known that a digital single-carrier unit in BPSK or in QPSK (modulation with two or four phase states) has a significant modulation of its envelope which is due to transmission filtering.

These non-linearities have the detrimental effect of causing the secondary lobes to rise in the spectrum emitted at the output of the power amplifier.

By way of illustration, FIGS. 3, 4, and 5 represent the output spectra 3, 4, 5 obtained by simulation for this same PT for three input declines, respectively 8dB, 10 dB, and 12 dB, to which respectively correspond 2.9dB, 4.4dB, and 6dB output rollbacks. On these graphs, the frequencies are conventionally plotted on the abscissa, and the decibels are plotted on the ordinate with 5 dB / square.

In figure 3 (output recoil of 2.9 dB), the sidelobe rises 6.7 exhibit a maximum emitted energy density which is approximately 22 dB lower than the average energy density of the carrier in its band, which is insufficient compared to the standards
INTELSAT which recommend a difference of at least 26 dB.

In the case of FIG. 4 (output retreat of 4.4 dB), the difference is about 24 dB, and is therefore still insufficient.

This difference is on the other hand of the order of 27 dB in the case of FIG. 5 (output decrease of 6 dB), which corresponds well to the desired performance.

From these three curves 3, 4, 5, it appears that to obtain, with this PT amplifier, a rejection of 26 dB on the side lobes 6 and 7, it is necessary to use a PT whose output recoil is around 5 dB.

The invention, which aims to reduce the output recoil necessary for such a power amplifier, is based on the observation made by the inventor, according to which the phase modulation effect, due to the non-linear characteristic Amplitude / phase of this amplifier, is largely predominant for this phenomenon of secondary lobe rise on the amplitude compression effect due to its nonlinear Amplitude / Amplitude characteristic.

By way of illustration of this observation, FIG. 6 shows the spectrum 8 plotted, by simulation and still for this same PT, by (fictitiously) completely suppressing the phase response of FIG. 2, and therefore keeping only the non- characteristic. linear in amplitude of FIG. 1. This curve 8 was plotted for an input retreat of 5 dB giving an output retraction of 2.3 dB, and it is then noted that a rejection of the secondary lobes is obtained. 6 and 7 which is substantially equal to the desired 26 dB value.

It finally appears that it is possible, if we manage to eliminate the phase distortion, to gain appreciably for the same rejection of the secondary lobes, 3 dB on the power necessary for the transmitter: for a given power to be transmitted, it is possible to use an amplifier which, at saturation, emits half the power, and therefore much less expensive.

The object of the method of the invention is therefore to introduce very upstream and more precisely at the level of the modulator and after digital filtering, a phase predistortion making it possible to compensate for the effect of the phase non-linearity, which is normally obtained at the output of the power amplifier.

Furthermore, the inventor observed that the correction thus obtained had the advantage of not being appreciable. By way of illustration, FIG. 7 shows the spectrum 9 obtained by simulation, again for this same
TOP, keeping only a quarter of the phase variation due to its non-linear Amplitude / Phase characteristic: the result is very little different from the previous one. This shows that the distortion due to the Amplitude / phase response, although being the main cause of distortion, is not very critical as to the precision to be provided to improve the spectrum increases. This remark is very interesting because, the phase correction to be made not being very sensitive, and the transmitters all having substantially the same phase responses, it is possible to implement the present invention using pre-correcting devices suitable for both amplifiers of the same type.

The method of the invention consists, at the level of the modulator but after the digital baseband filtering, in actually or fictitiously detecting the envelope of the modulating signal normally obtained at the output of this modulator, and in applying to the signal, always at the level of this modulator, an auxiliary and corrective phase modulation, by this modulated carrier envelope, this auxiliary phase modulation according to a law opposite to that due to the Amplitude / Phase conversion of the power amplifier of the transmitter.

A first device (entirely digital) for implementing this method is shown diagrammatically in FIG. 8, relating to a modulation with four phase states, or
QPSK, and which represents the modulator, with pre-correction device 10 according to the invention.

Conventionally, the digital baseband train is separated into two independent trains, P and Q, each of half frequency, which are applied to the two inputs 11 and 12 of the digital filtering device 13.

The two filtered trains A and B at the output 14 and 15 of the digital filter 13 respectively represent the real term A and the imaginary term B of the analytical signal
A + iB representative of the QPSK modulation which will be obtained later, at the output of the modulator.

de ce signal analytique est l'information représentative du signal modulant.The module

of this analytical signal is the information representative of the modulating signal.

To fictitiously detect this envelope, these channels A and B are applied, at 16 and 17, to a circuit 18 for digital calculation of this value of R by application of the formula
R2 = A2 + B2
This value R thus calculated is applied at 19 to a calculation circuit 20 in which have been entered beforehand, via the access 21, the values of the phase shifts (R) which give the variation curve opposite to the Amplitude / Phase response (figure 2) of the installation power amplifier.

Each value of R applied at 19 thus corresponds in circuit 20 to a phase shift value 9 (R), and this same calculation circuit supplies directly on its outputs 22 and 23 the value of the sine and cosine of this angle 9.

These values sin (p (R) and cos (R) at 22 and 23 are applied to another digital calculation circuit 24, which also receives the aforementioned A and B trains, and which performs the desired corrective phase rotation of the signal vector complex "A + iB" by product calculation
(A + iB) i-ç (R)
In practice, this phase rotation is simply reduced to carrying out the following two numerical operations
X = A. cos (R) - B.sin (R)
Y = A. sin (R) + B.cos 9 (R) where X and Y are the two digital trains “pre-corrected in phase” at outputs 25 and 26 of calculation circuit 24.

It is obvious that the MODEM of FIG. 8 being a digital MODEM, the phase pre-correction circuit 10 is in fact totally integrated into this MODEM, and is therefore only carried out by programming additions (or "SOFTWARE "), without the addition of any physical electronic element itself.

The two trains X and Y are then applied to the remaining part of the modulator, which very conventionally comprises a converter on each channel.
Digital / Analog 27, 28, an analog filter, 29,30, and a respective ring modulator 31, 32 each receiving the intermediate frequency from the local oscillator 33, one through a 90 degree phase shifter 34, and the other directly, in order to obtain the two desired quadrature tracks at 35 and 36, these two tracks then being conventionally added at 37 to obtain at 38 the train finally modulated with four phase states, this train being applied to the remaining part of the transmission chain comprising the power amplifier, the phase rotation of which is thus pre-corrected.

It should be noted that the signals (P, Q), (A, B), (X, Y) are defined (at the points where they are represented) in digital form (the change to analog being made by the digital / analog encoders 27 , 28).

In other words, all the preceding operations are numerical calculation operations. To create this phase modulation, the spectrum of which will occupy approximately twice the initial spectrum and taking into account the filtering necessary to remove the harmonics from the sampling frequency (to constitute the analog channels to be transmitted), it is necessary to have a minimum of 4 samples per symbol.

The embodiment of FIG. 8 is entirely digital, and consequently has the advantage of being able to be carried out without any addition of material, by simple addition of “SOFTWARE”, ie programming. With reference to FIG. 9, an analog embodiment of this phase rotation pre-correction will now be described.

In this FIG. 9, the reference 39 designates the QPSK modulator proper, which receives on its input 40 the modulating digital train. There emerges at 41 the intermediate frequency carrier, modulated with four phase states in this example, which is applied to the phase pre-correction circuit 10 of the invention.

The signal at 41 has its envelope detected in a detection circuit 42, and this detected envelope signal is applied to an analog response circuit 43 representative of the “Amplitude / Phase” curve inverse to that of the power amplifier ( FIG. 2) and capable of supplying at its output 44 a modulation slope coefficient, in degrees / dB, capable of controlling, in the opposite direction to the phase modulation created later by the envelope of the modulating signal in the amplifier of power, an auxiliary phase modulator 45 which receives the signal at 41. The train thus pre-corrected in phase leaves the pre-corrector 10 at 46 to be applied to the remaining part of the transmission chain.

Finally, FIG. 10 shows an alternative embodiment of the corrector 10 of FIG. 9, which is in fact a digital-analog hybrid solution because it comprises some of the elements of FIG. 8.

du signal analytique
A + iB c'est à dire de détection de l'enveloppe du signal modulant obtenue après filtrage numérique en bande de base.According to this embodiment, the channels A and
B at the output of the digital baseband filtering device 13 are, as in the case of the embodiment according to FIG. 8, applied to a digital calculation circuit 18 of the module

of the analytical signal
A + iB, ie detection of the envelope of the modulating signal obtained after digital baseband filtering.

On the other hand, the analogy with the device of FIG. 8 stops there, because the remaining part of the phase corrector operates in a manner rather similar to the embodiment according to FIG. 9.

Consequently, the numerical value of the envelope
R at the output 19 of the calculation unit 18 is converted into an analog signal in a digital-analog converter 47, and this envelope signal is applied, like the device according to FIG. 9, to an analog response circuit 43 representative of the "Amplitude / Phase" curve inverse to that of the power amplifier located downstream of the chain, and capable of supplying at its output 44, a modulation slope coefficient opposite to that relating to the phase modulation which will be created by the modulating signal envelope in the power amplifier.

The signal thus created at 44 controls an auxiliary corrective phase modulator 45, which receives at 41 the signal coming from the remaining part 48 (after filtering by 13) of the modulator, and the signal finally pre-distorted in phase comes out as above from this. auxiliary modulator 45 to 46 to be applied to the remaining part of the transmission chain.

The signal at the output of the transmitter is finally completely corrected for phase distortion, so that the corrected spectra obtained, as the applicant has been able to verify for input setbacks of less than 10 dB, are practically identical to the curves according to the FIG. 6 (total phase correction, according to a law with constant slope) and according to FIG. 7 (partial phase correction of three quarters).

As goes without saying, the invention is not limited to the embodiments which have just been described.

It is thus that the possibility of adding an amplitude pre-correction to the phase precorrection is not excluded from the scope of the present invention, and that other embodiments would be imaginable.

In all the embodiments, the transmission chain (between the modulator and the input of the transmitter) must have a sufficient band (in the sense of the absence of linear phase distortion) to allow the spectrum widened by this modulation to pass. pre-correction phase.

Claims (9)

1 - A method of limiting the rise of side lobes (6, 7) of the spectrum emitted by a power transmission installation with a digital single-carrier modulated with two or four phase states, characterized in that it consists, at the level of the modulator (13, 27 to 37, 39, 13, 48) but after digital filtering in baseband (13), to detect actually (42) or fictitiously (18) the envelope (R) of the modulating signal (A + iB) normally obtained at the output of this modulator, and to apply to the signal, always at the level of this modulator, a phase modulation, auxiliary and corrective (24, 25), by this envelope (R) of the modulated carrier, this modulation phase according to a modulation law opposite to that due to the Amplitude / phase conversion (2) of the power amplifier of the transmitter. 2 - Method according to claim 1, characterized in that the detection (42) of the envelope (R) of the modulating signal as well as the corrective phase modulation (45) are performed at the output (41) of the modulator (39) and by analog means (42, 43, 45) ,. 3 - Method according to claim 1, characterized in that the detection (18) of the envelope (R) of the modulating signal is carried out fictitiously, immediately at the output of the digital baseband filtering (13), by digital calculation of the module (R) of the analytical signal (A + iB) representative of the modulation. 4 - Method according to claim 3, characterized in that the corrective phase modulation is performed digitally on the signal (A, B) at the output of the digital filtering (13), by a phase rotation [(A + iB) i < p (R>] corresponding to the corrective phase [(p (R)] relating to the value of the calculated envelope (R). 5 - Method according to claim 3, characterized in that the corrective phase modulation is performed, as a function of the value of the envelope (R) thus calculated, at the output (41) of the modulator (13, 48) and by analogue means (43, 45). 6 - Device for implementing the method according to claim 1, characterized in that it comprises, immediately at the output (41) of the modulator (39), means (42) for actually detecting the envelope of the signal on this output (41), and to apply this detected envelope signal to an analog response circuit (43) representative of the "Amplitude / Phase" curve inverse to that (2) of the power amplifier of the transmitter and suitable for supply on its output (44) a modulation slope coefficient capable of controlling, in the opposite direction to the modulation created further downstream by the envelope (R) of the signal modulating in this power amplifier, an auxiliary phase modulator (45 ) which receives the output signal (41) from the modulator (39). 7 - Device for implementing the method according to claim 1, characterized in that it comprises, immediately at the output of the device (13) for digital baseband filtering fitted to the modulator, a device (18) for digital calculation of the modulus (R =

of the analytical signal (A + iB) representative of the modulation, and of application of the result (19) of this calculation to a digital calculation circuit (20) in which the values of the phase shifts [(p (R) have been entered beforehand ] which give the variation curve opposite to the Amplitude / Phase response (2) of the power amplifier of the transmitter, this circuit (20) providing on its output (22, 23) the digital values [cos (R) and sin (p (R)] necessary for the realization, by another digital calculation circuit (24) of the desired corrective phase rotation [(A + iB) i <p (R)] 8 - Device for implementing the method according to claim 1, characterized in that it comprises, immediately at the output of the device (13) for digital baseband filtering fitted to the modulator, a device (18) for digital calculation of the module

of the analytical signal (A + IB) representative of the modulation, and of application of the analog value of this module to an analog response circuit (43) representative of the Amplitude / Phase curve inverse to that (2) of the amplifier transmitter power and capable of supplying at its output (44) a modulation slope coefficient capable of controlling, in the opposite direction to the modulation created further downstream by the envelope (R) of the modulating signal in this amplifier of power, an auxiliary phase modulator which receives the output signal (41) from the modulator (48, 45). 9 - Device according to one of claims 6 to 8, characterized in that it is used in a satellite telecommunications installation.