FR2746234A1 - METHOD AND DEVICE FOR BASIC BAND TRANSFER OF THE DELAY AFFECTING A TRANSPOSE RADIO FREQUENCY SIGNAL - Google Patents

Abstract

The invention relates to a method and a device for delaying a modulated radio frequency signal. For an omega0 pulsation radiofrequency signal modulated by an OMEGA pulsation modulating signal, and to be delayed by a delay value tau, the method and the device consist in applying to the modulating signal a delay tau 'verifying the relationship: tau' = tau (1 + - omega0 | OMEGA) + 2k vpi / OMEGA where k denotes an integer, any delay applied to the modulated radiofrequency signal being removed. Application to digital radio frequency transmission D.A.B, with OFDM modulation.

Description

Method and apparatus for delaying the delay affecting a baseband

  The invention relates to a method and a device for reporting the delay affecting a signal in baseband.

transposed radiofrequency.

  The implementation of current radio technologies

  digital broadcasting, known as English

  Saxon DAB for Digital Audio Broadcast, requires the

  most of the time that the broadcasting of radio programs

  This is done from two modulated signals carrying the scattered information, delayed in relation to each other by a delay value t equal to at most a fraction of the period of the carrier wave. This is particularly the case when broadcasting and broadcasting

  mentioned above are carried out on a single-frequency network of

  sion in which each transmitting station has a

  plurality of banks of radiating sources such as

  Figures in a top view in the figure forming a transmitting antenna. These radiating sources are constituted by radiating dipole banks, emitting in vertical polarization and are placed in the form of a pylon so as to

  obtain an azimuth radiation pattern, orthogonal plane

  vertical pylon, substantially omnidirectional. This is also the case when the actual transmission is not single frequency but OFDM type in which at a single carrier frequency is substituted a predetermined number of discrete frequencies, spaced an identical frequency interval and covering a Band

frequencies of determined value.

  A theoretical study has shown that the conclusions relating to the single-frequency emission, as far as

  the omnidirectionality character of the radiation pattern

  from several sources of a transmitting

  would be maintained in the more specific case of the

OFDM modulation.

  Such a result is justified by the fact that: - the radiated field resulting at great distance by a radiating multisource antenna is the vector sum of the contributions of each source; - the conclusion regarding the radiated field

  result remains valid for each carrier wave

  of the OFDM signal, the radiated power

  overall amount being substantially carried over to each carrier wave

is.

  For a more complete study of the omnidirectional nature of

  of OFDM transmit sources, reference may be made to the French patent application

  No. 91,08089, inventor P.PIOLE, filed on June 28, 1991.

  In the aforementioned cases, as shown in FIG. 1a, the banks of radiating sources are placed, four in number for example, at the top of the pylon and

oriented at 90 from each other.

  To obtain a radiation pattern having the best omnidirectionality character, the four banks of radiating sources are placed in a square whose side has a dimension depending on the frequency of the wave or the carrier waves used. This dimension can vary from 4.0 m in band I to 0.70 m in V band. Two banks of geometrically opposite radiating sources are

  supplied in phase by the modulated carrier wave or waves.

  In order to improve the omnidirectionality character of the radiation pattern of the assembly, a delay or phase shift is introduced between the supply signals of the geometrically radiating source bank groups.

opposed.

  This delay is introduced thanks to a delay line inserted on the supply line of one of the groups of banks of

radiant sources.

  This delay line, for a frequency carrier wave

  fo = 1492 MHz, consists of a coaxial cable

  accordingly, large diameter 7/8 inch,

  1 inch = 2.54 cm, to minimize radio loss

  triques, but whose length reaches nearly 200 meters,

  to produce, at this frequency, a delay x = 550 ls.

  This technique, as represented in FIGS. 1a and 1b, is currently satisfactory but

  many limitations and disadvantages.

  In the first place, although a coaxial cable of large section is used, it is essential, before actual emission of the delayed modulated signal, to ensure a precompensation of the introduced radioelectric losses, because of the insertion of this coaxial cable, in order to obtain a delayed modulated signal of sufficient level and to ensure transmission under similar conditions of amplitude or energy level of the delayed modulated signal

  and non-delayed modulated signal. The leadership character-

  The radiation pattern of the transmitting antenna is indeed optimal when, due to the dual geometrical symmetry of the transmitting antenna in the azimuthal plane, the amplitude levels of the delayed and undelayed signal are identical. . The amplification needed for

  such pre-compensation is of the order of 20 dB.

  Secondly, the introduction of a coaxial cable of large diameter and respectable length poses a problem of infrastructure and congestion and, to a lesser extent, maintenance of this type of installation. In particular, we can mention that

  aging of the components, especially dielectric

  the stability in time of the value of the delay t introduced is not satisfactory and, as a consequence, campaigns for the measurement and control of the directionality character of the antenna radiation must be regularly carried out. conducted on site, by means of a

  hardware embedded on a helicopter for example.

  The object of the present invention is to remedy the aforementioned drawbacks by implementing a method and a device for transferring the delay affecting a transposed radiofrequency signal to the baseband, the delay line consisting of a coaxial cable of large size. length used in the devices of the prior art is thus removed. The device and the method for delaying a determined delay value t of an amplitude-modulated pulse-frequency radiofrequency signal by a modulating signal of

  pulsation n, objects of the present invention, are noted

  in that each allows respectively to apply to the modulating signal a delay t 'satisfying the relation: t' = x (1 0O) + 2kn / the delay t applied to the modulated radio frequency signal being

  deleted, where k denotes an integer.

  In particular, they find application in the implementation of broadcasting and digital broadcasting stations

  of the DAB type, in particular in OFDM modulation.

  They will be better understood by reading the description

  and in the following drawings, in which, in addition to Figs. 1a and 1b relative to the prior art: Fig. 2a shows an illustrative block diagram of the delay baseband method and delay device affecting a transposed radiofrequency signal, in accordance with the object of the present invention; FIG. 2b represents a frequency diagram used in the case of DAB transmission in OFDM modulation; FIG. 3a shows a group delay diagram and a phase diagram in the case where the device according to the invention is implemented in the form of digital filters of the RIF type; FIGS. 3bl to 3b4 represent diagrams

  amplitude and delay time depending on the frequency

  quence for a specific RIF digital filter with a

  number of coefficients equal to 127 for a frequency band

  these modulating the modulating signal between 13568 kHz and 15104 kHz;

  FIG. 4a represents another embodiment of

  in digital form, the device object of the present invention, wherein the modulating signal is converted into a plurality of frequency components by frequency transform;

  FIG. 4b represents an elementary delay law;

  provided to each elementary frequency component in the case of the embodiment of FIG. 4a; FIG. 4c represents a preferred embodiment of a delay module of each elementary frequency component according to the delay law represented in FIG. 4b; FIG. 5 gives a general layout diagram of a device that is the subject of the present invention in an OFDM modulation digital DAB modulator used in a

transmission station.

  A more detailed description of the process and

  device objects of the present invention will now be

  given in connection with Figure 2a.

  The method according to the invention makes it possible to apply

  delay of a determined delay value t of an amplitude-modulated w0 pulse-frequency signal by a pulse-modulating signal n by applying a specific delay of value t 'to the above-mentioned modulating signal, this delay being of course understood function of the delay value t and satisfying the relation: (1) t '= (1) + 2kr / Q In the aforementioned relation, we indicate that k denotes a

positive or negative integer.

  It is also recalled that due to the modulation

  amplitude, the sign affecting the ratio o0 / 0 corresponds to

  actually does it to get the upper sideband,

respectively lower.

  Of course, according to a particular aspect-

  advantage of the method which is the subject of the present invention, the application of the delay r 'of specific value on the modulating signal, that is to say in the baseband, makes it possible to

  remove the application of the delay r on the radiofre-

  Quence modulated to obtain, as shown in Figure 2a, on the one hand, the signal S'2 (t) and of course the signal Sl (t). The signal obtained on the delayed channel by the value I 'satisfies the relation: (2) S'2 (t) e [Q (t-)] The equality of the phases of the signal S2 (t) of the art

  obtained by direct delay of the radiofrequency

  this modulated, as shown in FIG. 1b, and the signal S 2 (t) then makes it possible to obtain the relation (1) previously

  mentioned in the description, giving the value of the delay

  i 'according to the value of the corresponding delay t

  applied to the modulated radio frequency signal.

  Of course, the method which is the subject of the present invention may consist in generating the value delay t 'on the modulating signal from an all-pass filter for example. The phase shift introduced on the modulating signal is

then equal to the value Q x t '.

  A more detailed description of the implementation

  of the aforementioned all-pass filter, in the case where the radiofrequency signal is an OFDM-modulated DAB signal,

  will now be described in connection with Figure 2b.

  FIG. 2b shows the frequency diagrams of a signal corresponding to OFDM modulation, the radiofrequency signal having for example a width of

  Af0 in which a plurality of discrete frequencies

  whose central frequency is indicated by f0, is

  distributed to form the plurality of carrier waves.

  Typically, the f0 value is equal to 1470 MHz in band 6

for example.

  The modulating signal itself consists of a central frequency signal F0 whose value is shifted for each carrier by a specific value AF equal to

4 kHz for example.

  Under these conditions, and for a broadband signal thus constituting the modulating signal, the delay introduced by passing through a filter is given by the group delay of the set:

(3) di / dn = AT / An.

  In the aforementioned relation, the expression Ai / An represents the ratio of the increase of delay introduced by the insertion of the filter on the modulating signal to the increase of frequency or of the pulsation of this same modulating signal with respect to the central frequency of

  this one F0 o the central pulsation of this one n0 = 2n.F0.

  For di / dn = t '= I' (l + O /), the global delay introduced on the modulating signal is written as follows: (4) I '= I d dn = Jr (l + 0) An = w01 (1 + 1) An In this relation, the ratio n / (0 is less than 1 / 1000. For this reason, the quantity 1 / 0O can be

  neglected in front of the quantity l / n. The integration of the expression

  sion of t 'as a function of n then gives the relation:

(5) I '= 01 Log n / 0.

  In this relation, it is indicated that n0 denotes the central pulsation of the modulating signal, as represented in FIG.

Figure 2b.

  The delay t 'on the modulating signal can then be generated by means of an all-pass filter whose transfer function satisfies the relation: (6) T (jn) = e rg / In general, it is indicated that the delay device making it possible to carry out the baseband delay of the delay affecting a transposed radiofrequency signal, in accordance with the subject of the present invention, can be realized either by an analog type filter, or by a

digital type filter.

  The realization by means of a digital filter will be preferred because of the possibilities of signal processing

  currently available routinely.

  In a first embodiment, as shown in FIG. 3a, the aforementioned all-pass filter may consist of a filter of the RIF or response filter type.

Impulse Finished.

  For the following delay parameter values applied to the modulating signal, delay in which the expression of t 'is given with an approximation as follows: (7) t' = t (1 + fo (1 - F - Fmin)) + K Fmin Fmin o: K = 350 ns x = 550 ns Fmin = 13 568 kHz Fmax = 15 104 kHz AF = 4 kHz fo = 1 492 MHz

At = 18 ns.

  The ratio F / AF = 384. The aforementioned values are then applied to a RIF type filter with an amplitude

  constant and a tpg group delay present

  as a slope 1 / X. FIG. 3a shows the variation of the group tpg propagation time as a function of the modulation frequency of the modulating signal, the above-mentioned group delay time being substantially an affine function of the frequency in the considered frequency band. , Fmax. In FIG. 3a, Fe represents the sampling frequency of the modulating signal, this frequency being

equal to 8,192 MHz.

  The expression of the group delay time in the analog domain is written:

(8) tpg (n) = nP + tpgM.

  In this relationship, P denotes the slope and tpgM denotes the value of the group delay for the value

  average of the modulation frequency of the modula-

  This is the frequency F0 or the pulse W0.

  The corresponding phase is then given by the relation: (9) p (n) = I (P + tpgM) An = n2 p + tpgM O + K The phase of the equivalent digital filter obtained by changing the variable On = nTe verifies the relation (10) (Qn) = On2 P + tpgM Qn + K '2Te2 Te In the aforementioned relation (10), P' designates a slope value obtained following the aforementioned variable change and K '

  the new value of the constant K after this same change

variable.

  The calculation of the coefficients of the RIF type filter then has the following expression: (11) h (m) = 1 / nt T (On) e-jnm An -1 / f ej (Qn2 P '/ 2Te2 + Qn (m + tpgM / Te) + K ') QA

  In FIGS. 3b1, 3b2, 3b3 and 3b4, there is shown

  a preferred embodiment in which a RIF type filter with 200 coefficients has been implemented.

  for the values and frequency parameters

  after: - modulation frequency band of the modulating signal: 13568 kHz 15104 kHz, - amplitude ripple in the band: 0.04 dB,

  - group delay time ripple: 10 ns.

  FIG. 3bl represents the amplitude diagram in the considered frequency band of the all-pass filter thus produced; FIG. 3b2 represents an enlarged view in the frequency domain considered of FIG. 3bl; FIG. 3b3 represents the delay law of the group delay time as a function of the frequency n of the signal

  modulation, the y-axis being graduated in percent

  of the sampling period Te, and the axis of

  frequency abscissa for the frequency band considered

  3b4 represents, for this same band of

  Frequency, time delay error of sensitive group-

  constant and not more than 10 ns.

  A second embodiment in digital form of a device according to the subject of the present invention will now be described with reference to FIG.

following figures.

  As shown in the above-mentioned figure, the device according to the invention, in this embodiment, comprises an analog-digital converter 1 for digitizing the modulating signal, this converter delivering successive digitized samples of the modulating signal. The sampling frequency Fe of this modulating signal can for example be taken equal to the value previously

indicated in the description.

  The device according to the invention also comprises a direct frequency transform module 2 enabling, from the successive digitized samples delivered by the analog-digital converter 1, to deliver transform coefficients representative of N elementary components of the frequency spectrum of the modulating signal. . By way of non-limiting example, it is pointed out that the direct frequency transform module can be constituted by a

  built-in fast Fourier transform module.

  The direct frequency transform module 2 is then followed by an elementary delay module of each elementary component of the signal frequency spectrum.

  modulating, this module bearing the reference 3 in Figure 4a.

  The elementary delay module 3 introduces an elementary delay on a frequency component of the modulating signal, each elementary delay being proportional to a base delay and to the rank inversely proportional to the

delayed frequency.

  Module 3 is itself followed by a module 4 of

  inverse frequency transform allowing reconstruction

  to kill, from the delayed elementary components,

  delayed digitized samples relating to the modulating signal.

  Finally, a digital-analog conversion module 5

  that renders, from the delayed digitized samples, a delayed modulating signal of the value t 'previously indicated. Of course, the inverse frequency transform module is constituted by an inverse fast Fourier transform module when the module 2 is itself formed by a direct fast Fourier transform module. These two modules may consist of specialized integrated circuits normally available in

trade.

  With regard to the delay module 3, it is indicated that this can be constituted by a plurality of elementary delay modules, each elementary delay module being formed by a shift register providing on each elementary component of rank n a delay. verifying the relation: (12) t'n = TO (1 + h) n In this relation, h denotes a coefficient proportional to the number N of elementary components satisfying the relation: (13) h = N Fo p.Fe F0 denoting the central frequency of the modulation signal, F0 = n0 / 2n, where Fe designates the sampling frequency of the analog-digital converter and numerical analogue, p designating a weighting parameter related to the type of weighting window applied to the digitized samples

  prior to their frequency transformation.

  By way of non-limiting example, it is indicated that for a rectangular window applied to these samples, p may take the value 2, whereas for a window of

Hamming p can take the value 4.

  Of course, when the delay module 3 is produced by virtue of the implementation of shift registers each comprising a correspondingly adapted number of stages in order to ensure the adequate delay value for each frequency component considered, it is necessary to to adjust the value of the samples thus subjected to the corresponding delay by appropriate circuits. These types of circuits are conventional and known to those skilled in the art and,

  as such, will not be described in detail.

  It is easy to understand that, depending on the

  the frequency of the modulating signal, each elementary delay is inversely proportional as a function of the modulation frequency of the modulating signal. FIG. 4b shows a diagram giving the linearized law of each elementary delay i'n expressed in ns as a function of the frequency F expressed in kHz for the frequency band

  of the modulating signal previously mentioned in the descrip-

  tion. It is then understood that the implementation of successive elementary shift modules makes it possible to realize a

  offset for the corresponding frequency component

  dante. However, when the delay module 3 is implemented by means of shift circuits, such an implementation

  requires some adaptation of the digital samples

  generated by the frequency transform module 2 into signals directly applicable to these shift registers, and then reciprocally, of transformation of the offset signals into directly applicable digital samples

  to the inverse frequency transform module 4.

  For this reason, another preferred embodiment of the delay module 3 will now be described in

connection with Figure 4c.

  As shown in FIG. 4c, it is indicated that the delay module 3 and in particular the elementary delay modules constituting it, are advantageously constituted by a read-write controlled RAM, the memory receiving on an input port the digital samples delivered by the frequency transform module 2 for storing them. Of course, the RAM 30 is controlled read-write. It comprises a plurality of elementary storage cells each containing a value of

elementary component.

  In addition, the dual-port random access memory 30 is associated with an addressing module 31, this addressing module delivering to each elementary cell constituting the dual-port random access memory 30 a read addressing signal delayed by a corresponding elementary delay. . The dual-port random access memory 30 then delivers from each elementary storage cell a delayed elementary component on the output port as shown in FIG.

Figure 4c.

  It is furthermore indicated that the read control module 31 may consist, for example, of an EPROM type memory controlled by a clock signal CLK and by a write-read discrimination signal E / L, the EPROM memory 31 containing in particular the elementary delay values, as represented in FIG. 4b as a function of frequency, to enable the delayed addressing of each storage cell of the previously mentioned dual-port RAM 30. It is thus understood that according to the equipment intended to be equipped with a device according to the object

  of the present invention, and in particular, characteristics

  radio frequency and geometry of the transmission antenna of the latter, the EPROM can easily be programmed according to specific delay values for frequency domains adapted to these equipment, the delays thus brought to the signal modulating by the intermedi-

  elementary delays thus being an introductory

  very flexible for each type of material.

  Finally, it is pointed out that the method and the device that are the subject of the present invention can be applied in order to carry out the baseband transfer of any delay that affects a transposed radiofrequency signal. This is particularly the case with regard to television signals, especially digital television signals, as well as the radio frequency signals used to provide DAB type transmission. In the latter case and in particular when the DAB signal is generated in OFDM type modulation as shown in FIG. 5, the device according to the present invention, introducing the delay t 'on the second signal channel, can, as represented in this figure, be introduced between the part D and the G part referenced in Figure 5 above, these parts being related to the digital modulation and MDP4 linearity correction respectively. In FIG. 5, the letters referenced A to M designate the various constituents of a DAB modulator.

  in which the modulation is performed in modula-

  OFDM. These parts and the definition thereof are given as a reminder but will not be described because they correspond to elements known from the state of the art, except as regards the delay element constituted by the device according to the invention, introduced at the E.

Claims (7)

  1. Method for Delaying a Determined Delay Value x of a Modulated 0O Pulse Radio Frequency Signal   in amplitude by a pulse modulating signal Q, which is   in that it consists in applying to said modulating signal a delay t 'which is a function of the delay value T satisfying the relation: I' = (1 0) + 2kr / n o k denotes an integer.   2. Method according to claim 1, characterized in that said delay i 'is generated from a transfer function all-pass filter T (jn) = eo0rL gQ / o said all-pass filter introducing on said signal   modulating a delay r 'corresponding to a phase shift Ot'.   3. Device for delaying a determined value of an amplitude-modulated pulse wave 0 by a pulse-modulating signal Q, characterized in that it comprises a delay all-pass filter applied to the modulating signal , said all-pass filter having a transfer function T (jQ) eJL = / 4. Device according to claim 3, characterized in that said all-pass filter comprises: - an analog digital converter for digitizing said modulating signal, said converter delivering successive digitized samples of the modulating signal; a direct frequency transform module, making it possible, from the successive digitized samples, to deliver transform coefficients representative of N elementary components of the frequency spectrum of the modulating signal; elementary delay means of each elementary component of the frequency spectrum of the modulating signal, each elementary delay means introducing a   delay proportional to a basic and inverse   proportional to the frequency of the elementary component   corresponding space and delivering a delayed elementary component; inverse frequency transforming means reconstituting, from the delayed elementary components, delayed digitized samples; - digital conversion means restoring, from the delayed digitized samples, a   delayed modulating signal of the value i '.   5. Device according to claim 4, characterized   in that said elementary delay means is constituted by   each killed by a shift register providing on each elementary component of rank n a delay t'n = t0 (1 + h) nok denotes a coefficient proportional to the number N of elementary components satisfying the relation: h = N Fo p.Fe Fo designating the basic frequency of the modulating signal Fo = n0 / 2i, Fe denotes the sampling frequency of the analog digital and digital analog converter, p denotes a weighting parameter related to the type of window   weighting applied to pre-digitized samples   their frequency transform.   6. Device according to one of claims 4 or 5,   characterized in that said frequent transform module   direct and inverse system is formed by a module of   Forward fast Fourier transform respectively inverse.   7. Device according to claim 4 or 6, characterized in that said elementary delay means are constituted by: a dual-port RAM, controlled read-write, said random access memory comprising a plurality of elementary cells each containing a value of elementary component; read addressing means of said dual port RAM, said addressing means delivering to each elementary cell a read addressing signal delayed by a corresponding elementary delay, said dual port RAM delivering from each   elementary cell a delayed elementary component.