FR2911451A1 - Incident signal e.g. radio frequency signal, processing e.g. filtering, method, involves configuring elementary circuit in analog to digital conversion states, and in reference states in which group of capacitors forms capacitive charge - Google Patents

Abstract

The method involves configuring each elementary circuit (CI1) in a filtering state in which a group of switched elementary capacitors is connected to an input capacitive unit (MCEI), where the circuit is configured by successive approximations connected between input and one of outputs (SI1-SI3) of analog/digital conversion and filtering circuit (CFCANI). The circuit is configured in analog/digital conversion states, and in reference states in which the group of capacitors forms a reference capacitive charge, where the output delivers a sample of bits of output digital signal. An independent claim is also included for a device for processing an incident signal, comprising an analog to digital conversion and filtering circuit.

Description

Method and device for processing an incident signal, in particular

  The present invention relates to the processing of an analog signal, and more particularly to the filtering and the analog / digital conversion of such a signal. The invention applies advantageously but not exclusively to the processing of signals, in particular radiofrequencies, undergoing filtering and conversion to a digital signal. This is the case, for example, for terrestrial digital television signals, or more generally for signals conveyed in frequency-division multiplex transmission systems. An incident analogue signal, for example having information located in a desired frequency range (desired channel for example) and frequency components located outside said desired frequency range (channels adjacent to the desired channel), generally undergoes one or more transpositions of frequencies to bring it back to baseband. The incident signal down in baseband is then intended, after filtering, to be converted into digital form to then undergo specific processing, such as for example channel decoding processing including demodulation and source decoding processing. The filtering and the analog / digital conversion are generally performed by external elements, that is to say by non-integrated elements. It is therefore not possible to make the analog part and the digital part of a device for processing a radio frequency signal on a single integrated circuit. In addition, the external elements of the analog part of the signal processing can not follow the same evolution, especially in terms of size, as the digital part made in CMOS technology. It is thus proposed, for example, a device and a method for filtering and analog / digital conversion of a signal while limiting the presence of non-integrated elements.

  In one aspect, there is provided a method of processing an incident signal in which an intermediate signal is derived from the incident signal. Then, the intermediate signal is delivered to a filter and analog-to-digital conversion circuit (CFCANI) comprising at least one current input for receiving the intermediate signal, several outputs (SI1, SI2, SI3), capacitive input means (MCEI) connected to the input, a plurality of successive successive approximation conversion circuits (CI1, C1, CI3) each connected between the input and one of the outputs (SI1, SI2, SI3) and each comprising a group of elementary capacitors switched. Then, each elementary circuit (CI1, Cl2, CI3) is successively and cyclically configured: in a filtering state in which the group of switched elementary capacitors is connected to the capacitive input means, then in several distinct states of analog conversion / numerically, then in a reference state in which the group of switched elementary capacitors has a reference capacitive load, said outputs each successively delivering a sample of several bits of a digital output signal corresponding to the filtered intermediate signal.

  The method uses capacitive filtering means as well as a successive approximation converter using the switched capacitance technique. Thus, it becomes possible on the one hand to integrate the filtering and conversion means with the rest of the signal processing device. On the other hand, the filtering means and the converter are made from capacitors and switches whose size and realization within an integrated circuit evolve at the same time as the technology used. The analog part can then be improved at the same time as the digital part.

  Here, the advantage of mixing the filtering and analog / digital conversion steps within one and the same circuit reduces the cost of design and the noise of the processing chain. It is also advantageous to use the technique of passive switched capacitors, both for filtering and for analog / digital conversion, which makes it possible in particular to adopt a high sampling frequency. The filtering and analog-to-digital conversion circuit (CFCANI) advantageously comprises at least N + 2 outputs (SI1, SI2, SI3) each delivering, at a delivery frequency fd, a sample of N bits of the digital output signal, and N + 2 elementary circuits (CI1, Cl2, CI3) each connected between the input and one of the N + 2 outputs (SI1, SI2, SI3) and each comprising at least one group of N + 1 switched elementary capacitors connected in parallel ( CE11, CE12, CE13, CE14, CE21, CE22, CE23, CE24, CE31, CE32, CE33, CE34) as well as a comparator (COMP1, COMP2, COMP3) connected to the corresponding output (SI1, SI2, SI3). According to this mode of implementation, each elementary circuit (CI1, Cl2, CI3) is then successively and cyclically configured at a control frequency Fs equal to at least the product of the number of elementary circuits by the delivery frequency fd, and with a time shift of 1 / Fs between two consecutive elementary circuits, and: the group of N + 1 elementary capacitors is connected to the capacitive input means in the filtering state, the N + 1 capacitors are connected between the comparator (COMP1, COMP2, COMP3) and a first or a second reference voltage (Vhigh, Vi0) selectively to obtain, successively at the frequency Fs, the N bits of a sample of the digital output signal at the corresponding output ( SI1, SI2, SI3) in N distinct analog-to-digital conversion states, and - the group of N + 1 elementary capacitors has a reference capacitive load in the reference state.

  The ratio of the value of the capacitance of the capacitive input means to the sum of the capacitances of the N + 1 elementary capacitors is advantageously greater than a threshold chosen as a function of the bandwidth of the analog signal. The value of this threshold makes it possible to adjust the parameters of the filtering, and in particular the cut-off frequency fcpr, according to the characteristics of the signal. Those skilled in the art will be able to adjust the value of this threshold according to the application and the desired performances. As an indication, with a relatively high threshold and for a relatively stiff filtering with a single filter capacitor CIIR, the cutoff frequency fcpr of the low-pass filtering is given by the relation:

  C ~ 1 CIIR 2 where C is equal to the sum of the capacitances of the N + 1 elementary capacitors and Fs is the sampling frequency.

  Thus, with a sampling frequency Fs equal to 600 MHz and to obtain a cut-off frequency fcpr of the order of 2.4 MHz, a threshold equal to 40 will be chosen. Similarly, for a signal of bandwidth of 10 MHz and for a sampling frequency Fs of 600 MHz, a threshold equal to 19 will be chosen so as to obtain a cut-off frequency fcpr equal to 5 MHz. The useful band is distributed between -5MHz and + 5MHz and the edge of the band is attenuated by 3dB. The capacitive input means may comprise a single filtering capacitor and, in the filtering state of each elementary circuit, the group of N + 1 elementary capacitors of said elementary circuit is connected to the filtering capacitor. The filter capacitor makes it possible to filter the signal thanks to a large capacity, called memory capacity, compared to the elementary capacitances of the elementary circuits. In addition, Fs the filter capacitor is not reset but remains connected to the intermediate signal continuously. When the intermediate signal, which is a current signal, evolves, the filtering capacitor makes it possible to average its value over time and thus to perform the filtering. More particularly, the average filter capacitor samples that are converted by the elementary circuits. This provides low-pass filtering on the signal. The number of elementary circuits depends on the number of processing steps required for filtering and conversion of the intermediate signal by the device. Thus, for a conversion of an analog signal into an N-bit digital signal, the method comprises according to an embodiment: a filtering step during which the signal is also loaded into the elementary capacitors; for converting the N bits and - a step of resetting (resetting) the elementary capacitors during which they are brought back into a reference state. If we increase the number of steps, the number of elementary circuits must also increase so as to be able to constantly process the intermediate signal. According to one embodiment, the capacitive input means may furthermore comprise at least one additional filtering capacitor, the filtering and analog / digital conversion circuit then comprising a number of additional outputs and additional elementary circuits equal to number of additional filter capacitors; the filtering state of each elementary circuit then comprises different filter sub-states in which the group of N + 1 elementary capacitors is respectively connected to the different filtering capacitors. Then, said elementary circuit is successively placed in each of the sub-states of filtering. A second filtering capacitor makes it possible to carry out an additional average on the values already averaged by the first filtering capacitor. This is followed by a second low-pass filtering. We can therefore obtain a higher or lower order filter according to the number of filter capacitors. In addition, the additional filter capacitors add steps in the signal processing. Thus, for each additional filter capacitor, an elementary circuit is added to the device. After the reference state, it is possible to configure the elementary circuit in a calibration state in which the comparator of the elementary circuit is connected to two equal voltages so as to determine and correct its offset. The purpose of the calibration state is to limit the offset (in English: offset) that may appear at the comparator and introduce noise into the converted signal. This state adds a step in the signal processing and an elementary circuit is added. The current input for receiving the intermediate signal may be a differential input and each elementary circuit then comprises a second group of N + 1 switched elementary capacitors connected in parallel.

  Although the method according to this aspect is compatible with any type of frequency conversion method preceding the filtering and the analog / digital conversion, this aspect advantageously combines with the frequency translation method described in the patent application entitled Method and device processing of an incident signal, in particular, frequency transposition filed in the name of the Applicant on the same day as the present patent application. In other words, in a particular application, the incident signal is delivered to a transconductor stage and a current output of the transconductor stage is connected to the filtering capacitor so as to deliver to the filtering capacitor a current signal and its respectively opposite during the two half periods of each period of a periodic signal and thus obtain, at the filter capacitor, a signal transposed into frequency as an intermediate signal. At the occurrence of each half period of the periodic signal, the voltage of said current output, seen from said filter capacitor, is initialized to a value equal to that of the voltage of said filter capacitor.

  The signal considered here is a current signal. During frequency translation, the frequency translation device supplies the filter capacitor with the current signal and its opposite during each half-period of the transposition signal. The device first initializes the voltage of the current output, i.e., equalizes the voltage of the current output and the voltage of the capacitor. There is no load movement (parasitic current) due to a voltage difference when the initialised current output and the filtering capacitor are connected, ie when a current occurs. half period of the periodic signal. The gain is not degraded. The transconductor stage may comprise two current outputs respectively delivering two current signals of opposite sign. In such an embodiment, at each first half-period of the periodic signal, the first current output is connected to the filtering capacitor and, at each second half period of the periodic signal, the second current output to the filtering capacitor. . In addition, the initialization of the voltage of the first current output comprises fixing, at least at the end, and preferably during the whole, of each second half period of the periodic signal, the voltage of said first current output. to the value of the voltage of the filter capacitor, and the initialization of the voltage of the second current output comprises fixing, at least at the end, and preferably during the whole, of each first half period of the periodic signal, from the voltage of said second current output to the value of the voltage of the filter capacitor. During the first half-period, the voltage of the filter capacitor changes at the same time as that of the first current output, as a function of the signal, whereas the voltage of the second current output changes independently as a function of the charges which out. Thus, a difference between the voltage of the filter capacitor and the voltage of the second current output appears during the first half-period, that is to say during the half-period during which the second current output is not connected to the filter capacitor. In order to avoid, during the occurrence of the second half-period, the parasitic currents due to this difference in voltage, the voltage of the second current output is initialized before connecting this second current output to the filtering capacitor. This initialization consists here in imposing the voltage of the filtering capacitor at the second current output. In this way, the voltage difference is zero, the charges leaving the second current output are absorbed and parasitic currents are avoided. In other words, preferably, during each half period of the periodic signal, a current output of the transconductor stage is connected to the filter capacitor while the voltage of the filter capacitor is imposed on the other current output. from the transconductor stage. Preferably, the voltage of one of the current outputs is set to the value of the voltage of the filtering capacitor via an amplifier stage mounted as a follower between the filtering capacitor and the current outputs. The amplifier stage mounted follower makes it possible to impose the input voltage at the output, without modifying the input voltage. In the present case, the input voltage of the amplifier stage is the voltage of the filter capacitor and the output voltage of the amplifier stage is the voltage of the current output of the transconductor stage which is not connected to the filter capacitor by the transposition device. Thus, the amplifier stage makes it possible to impose the voltage of the filtering capacitor at the current output which will be connected to said capacitor during the next half-period, without modifying the voltage of the filtering capacitor, that is to say in limiting the degradation of the signal. In this way, when the current output is connected to the filter capacitor at the occurrence of the next half period of the periodic signal, the voltage of the current output and the voltage of the filter capacitor are equal. According to another embodiment, the transconductor stage comprises a single current output delivering a current signal, in which: - each first half period of the periodic signal comprises a first part during which said current output is connected to the capacitor filtering, and - each second half period of the periodic signal comprises a first portion during which said current output is connected to the filter capacitor in the opposite way to that performed during the first part of the first half period of the periodic signal, and in which the initialization of the voltage of the current output comprises: - fixing the voltage of the current output at the opposite value of the voltage of the filtering capacitor at least at the end of the second part of each first half-period of the periodic signal, and - the fixing of the voltage of the current output to the value of the voltage of the filtra capacitor at least at the end of the second part of each second half period of the periodic signal. In this embodiment, the connection between the transposition stage and the filtering capacitor is reversed at each half-period of the periodic signal. However, the output of the transconductor stage may have a parasitic capacitance which is also reversed at each half period. Thus, during the occurrence of each half period, the filter capacitor is connected to the parasitic capacitance whose voltage is inverted with respect to that of the filter capacitor. To then avoid a parasitic current and therefore a degradation of the signal, it imposes the opposite voltage of the filter capacitor at the current output, before the inversion by the frequency transposition stage. In this way, during the occurrence of the next half period, the voltage of the current output seen from the filter capacitor is equal to the voltage of said capacitor.

  Fixing the voltage of the current output to the corresponding value is performed for example throughout the second part of the corresponding half-period of the periodic signal. Here again, preferably, the voltage of the current output is fixed at the value of the voltage of the filter capacitor or the opposite of this value by means of an amplifier stage mounted in a follower between the filter capacitor and the filter capacitor. current output. Whatever the mode of implementation, the current output (s) of the transconductor stage are advantageously but not necessarily differential. In another particular application, the incident signal is a radio frequency signal and the transconductor stage belongs to a low noise amplifier stage.

  Although the foregoing aspects are compatible with any subsequent digital processing of the converted signal, they advantageously combine with the filtering method described in the patent application entitled Method and device for filtering and converting an analog signal from an analog signal deposited at the same time. Applicant's name on the same day as this patent application. In other words, in a particular application, the filtering and analog / digital conversion circuit is able to filter the frequency components of the intermediate signal located outside a desired frequency range; the method then advantageously further comprises: - a reshaping of the digital output signal in a substantially identical form, at least in said desired frequency range, to the form of the intermediate signal and - a final digital filtering of the digital output signal delivered shaped, so as to filter the frequency components located outside said desired frequency range. The desired frequency range is for example the bandwidth of the wanted signal or the width of the desired frequency channel. The filtering circuit and analog / digital conversion performs an analog filtering for the purpose of filtering for example the residues of adjacent channels. In this way, the analog filtering makes it possible on the one hand to limit the dynamics of the analog / digital converter, and on the other hand to adapt the sampling frequency of the analog / digital converter to the bandwidth of the filtered intermediate useful signal. . The analog / digital conversion is thus improved with respect to the useful signal to be converted. However, insofar as the analog filtering is carried out in particular with so-called memory capacitors such as the filtering capacitor, the order of the analog filtering is low, for example less than or equal to 2, and does not make it possible to obtain some applications a good cut of adjacent. The analog filtering also causes a deformation, in the desired frequency range, of the intermediate signal. This deformation results in a loss of information or a degradation of the useful signal. The fitness step notably makes it possible to correct the distortion of the signal caused by the analog filtering. Thus, despite low-order analog filtering, for example order 1, on the one hand to limit the dynamics of the analog / digital converter, and on the other hand to limit the degradation of the useful signal. The fitness step makes it possible to recover, at least over the desired frequency range, a signal that is identical or almost identical to the intermediate signal.

  The final digital filtering then makes it possible to filter the frequencies situated outside the desired frequency range. This final filtering is performed numerically and can therefore have a transfer function elaborated and adapted to the spectrum of the signal to be filtered, without being bulky nor expensive. I1 also makes it possible to effectively filter the adjacent ones without deformation of the useful band of the digital signal reformatted. The filtering step of the intermediate useful signal is thus postponed during the digital processing part. It is indeed easier and less expensive to perform a transfer function developed with a digital filter than with an analog filter. Moreover, thanks in particular to the filtering capacitor, it limits the dynamics of the analog / digital converter during the conversion of the signal, while avoiding the loss of information or the degradation of the useful signal.

  According to one embodiment, the reshaping is performed within a digital fitness filter having a working frequency equal to the control frequency Fs and having a reverse transfer function of the transfer function of the analog filtering performed by the filtering circuit and analog / digital conversion at least in said desired frequency range. Thus, at the output of the fitness filter, in the frequency range, we obtain a signal identical to that which would have been obtained in the absence of analog filtering. The transfer function of the digital fitness filter can be calculated from the transfer function of the analog filtering. Converting an analog transfer function to a digital transfer function is conventional for those skilled in the art. This calculation is further simplified by the fact that the working frequency of the digital filter is equal to the control frequency Fs of the elementary circuit (ie of the analog / digital converter), that is to say is equal to the frequency of 'sampling. According to another embodiment, the reshaping is performed within a digital fitness filter having a different working frequency, preferably less than the control frequency Fs, and having a transfer function of which the coefficients result from an approximation of the inverse of the transfer function of the analog filtering performed by the filtering and analog / digital conversion circuit, at least in said desired frequency range.

  This embodiment is more particularly concerned with the case where the analog / digital converter upsamples the filtered intermediate signal. Over-sampling makes it possible to obtain a digital signal that is more faithful to the filtered intermediate signal. However, it is not necessary to use the same oversampling frequency for digital processing. Thus, it is common to use a working frequency smaller than the sampling frequency. In this case, the calculation of the digital equivalent of the inverse transfer function of the analog filtering is different and can in particular be calculated more easily by approximation in the desired frequency range. More precisely, the determination of the coefficients advantageously comprises a selection of selected reference frequencies in said desired frequency range and, for each of these reference frequencies, a minimization of the difference between the corresponding value of the inverse transfer function of the analog filtering. and the corresponding value of the transfer function of the fitness filter. In other words, the transfer function of the fitness filter is calculated in this example by considering a finite number of frequencies selected in the desired frequency range and for which the difference between the actual value of the function of the transfer of fitness filter and the desired value equal to the inverse of the transfer function of the analog filtering.

  Preferably, the reference frequencies are chosen uniformly in the desired frequency range. The uniform choice of reference frequencies makes it possible to obtain a correct approximation of the inverse transfer function of the analog filtering by the transfer function of the fitness filter over the entire width of the desired frequency range. The transfer function of the fitness filter is for example of the finite impulse response type. The transfer function can then be written in a polynomial form whose only coefficients are to be determined. In particular, deviation minimization methods make it possible to determine the coefficients for this type of filter. Preferably, the reshaping and the final digital filtering are carried out within the same filter whose transfer function, at least in the desired frequency range, is equal to the product of the transfer function of the fitness filter by the transfer function of the final digital filter. This mode of implementation can be applied if the working frequency of the filter is equal to or different from the control frequency Fs. This implementation mode makes it possible to perform the fitness step and the final digital filtering step at the same time. This makes it possible to avoid obtaining a digital signal with a large number of bits between the fitness filter and the final digital filtering, because of the reshaping of the signal even outside the desired frequency range. . To do this, it suffices to multiply the two transfer functions to obtain only one. The signal obtained at the output of the filter remains the same but the realization of the device is simplified. In another aspect, there is provided a device for processing an incident signal, comprising a filter and analog-to-digital conversion circuit (CFCANI) of an intermediate signal derived from the incident signal, comprising: at least one input into current to receive the intermediate signal; -several outputs (SIl, SI2, SI3); capacitive input means (MCEI) connected to the input; a plurality of configurable successive approximation circuits (CI1, C1, CI3) each connected between the input and one of the outputs (SI1, SI2, SI3) and each comprising a group of switched elementary capacitors; each elementary circuit (CI1, Cl2, CI3) having a filter state in which the group of elementary capacitors is connected to the capacitive input means, a plurality of distinct analog / digital conversion states, and a reference state in which the group of capacitors switched elementary capacitors has a reference capacitive load, said outputs being adapted to successively output a sample of several bits of a digital output signal corresponding to the filtered intermediate signal.

  The device preferably further comprises control means capable of successively and cyclically placing each elementary circuit in the filtering state, then in the successive states of analog / digital conversion, and then in the reference state. Preferably, the outputs (SI1, SI2, SI3) are capable of delivering each, at a delivery frequency fd, a sample of several bits of the digital output signal, and the control means are able to place, at a frequency of control Fs equal at least to the product of the number of elementary circuits by the delivery frequency fd and with a time shift of 1 / Fs between two consecutive elementary circuits, successively each elementary circuit (CI1, Cl2, CI3) in the filtering state , then in the analog / digital conversion states in order to obtain, successively at the frequency Fs, the bits of a sample of the digital output signal at the corresponding output (SI1, SI2, SI3), then in the state of reference. Preferably, the filtering and analog / digital conversion circuit (CFCANI) comprises: at least N + 2 outputs (SI1, SI2, SI3) each capable of delivering, at the delivery frequency fd, an N-bit sample of digital output signal, - at least N + 2 elementary circuits each connected between the input and one of the N + 2 outputs (SI1, SI2, SI3) and each comprising at least one group of N + 1 elementary capacitors (CE11, CE12 , CE13, CE14, CE21, CE22, CE23, CE24, CE31, CE32, CE33, CE34) connected in parallel and N + 1 configurable switches (Ill, I12, I13, I14, I21, I22, I23, I24, I31 , I32, I33, I34) and a comparator (COMP1, COMP2, CMOP3) connected to the corresponding output (SI1, SI2, SI3), the group of N + 1 elementary capacitors of each elementary circuit: - being connected to the means capacitive input to the filter state of said elementary circuit, being connected between the comparator and a first or a the second reference voltage (Vhigh, Vi0) selectively in the N distinct states of analog / digital conversion of said elementary circuit, and -presentant a reference capacitive load in the reference state of said elementary circuit. The ratio of the value of the capacitance of the capacitive input means to the sum of the capacitances of the N + 1 elementary capacitors is preferably greater than a threshold chosen as a function of the bandwidth of the analog signal. The values of the N + 1 elementary capacitors of each elementary circuit are for example C / 2, C / 4,..., C / 2N, C / 2N. The capacitive input means comprise a filtering capacitor and, for each elementary circuit, the switches of the N elementary capacitors of values C / 2, C / 4,..., C / 2N-, C / 2N have three states. switching and the switch of the (N + 1) 1st capacitor C / 2N value has two switching states. The control means are able to configure the switches of an elementary circuit in the same switching state during the filtering state of the elementary circuit, and in the other switching states during the N distinct states of analog / digital conversion and the reference state. Preferably, the capacitive input means further comprise at least one additional filtering capacitor. The device comprises a number of additional outputs and additional elementary circuits equal to the number of additional filter capacitors, and for each elementary circuit, each switch further has an additional number of switching states equal to the number of additional filter capacitors. The filtering state of each elementary circuit comprises different filter sub-states in which the group of N + 1 elementary capacitors is respectively connected to the different filtering capacitors and the control means are able to configure the switches of the elementary circuit so as to to successively place it in each of the sub-states of filtering. In the reference state of an elementary circuit, the switch of the elementary capacitor of value C / 2 of said elementary circuit is configured to connect said elementary capacitor of value C / 2 to one of the two reference voltages and the switches other N elementary capacitors are configured to connect said N elementary capacitors to each other of the two reference voltages.

  Preferably, the filtering and analog / digital conversion circuit comprises an additional output and an additional elementary circuit and each elementary circuit also has a calibration state in which the comparator is connected to two equal voltages in order to determine and correct its offset. . The control means are furthermore capable of configuring the switches of each elementary circuit so as to place it in its calibration state after its reference state. The device may also have a differential architecture in which each elementary circuit comprises a second group of N + 1 switched elementary capacitors connected in parallel and N + 1 configurable switches. In a particular application, there is provided a device for processing an incident signal comprising: a transconductor stage comprising an input for receiving the incident signal; frequency translation means connected between the transconductor stage and the filtering capacitor comprising: generation means capable of delivering a periodic signal; a set of switches configurable in response to a control signal; control means able to deliver the control signal from the periodic signal to configure the set of switches so as to connect a current output of the transconductor stage to the filter capacitor and deliver to the filter capacitor a current signal and its opposite respectively during the two half periods of each period of the periodic signal and thus obtain, at the of the filtering capacitor, a signal transposed into frequency as an intermediate signal re. The device further comprises initialization means able to initialize, at the occurrence of each half period of the periodic signal, the voltage of the current output, seen from said filtering capacitor, to a value equal to that of the voltage said filter capacitor. According to one embodiment of the device, the transconductor stage comprises two current outputs capable of respectively delivering two signals of opposite sign current. The set of switches has a first state in which it connects the first current output to the filter capacitor and a second state in which it connects the second current output to the filter capacitor. The control means are adapted to deliver the control signal so as to configure said set of switches in the first state during each first half period of the periodic signal and in the second state during each second half period of the periodic signal. In addition, the initialization means are capable of: setting the voltage of the first current output to the value of the voltage of the filtering capacitor at least at the end of each second half-period of the periodic signal; from the second current output to the value of the voltage of the filter capacitor at least at the end of each first half period of the periodic signal. Preferably, the initialization means comprise an additional set of switches configurable in response to an additional control signal as well as additional control means capable of delivering the additional control signal from the periodic signal. The additional set of switches has a first state in which it connects the filter capacitor to the second current output and a second state in which it connects the filter capacitor to the first current output. The additional control means are capable of delivering the additional control signal so as to configure said additional set of switches in the first state at least at the end of each first half period of the periodic signal and in the second state at least at the end. each second half period of the periodic signal. The additional control means are, for example, capable of delivering the additional control signal so as to configure the said additional set of switches in the first state and the second state respectively during the entire first half period and the second half period of the periodic signal, respectively. .

  According to another embodiment, the transconductor stage comprises a single current output able to deliver the current signal. The set of configurable switches has a first state in which it connects the current output to the filter capacitor, a second state in which it connects, in a manner opposite to that performed in the first state, the current output to the filtering capacitor and a third state in which it electrically disconnects the current output of the filter capacitor. The control means are able to deliver the control signal so as to configure said set of switches in the first state during the first part of each first half period of the periodic signal, in the second state during the first part of each second half period. of the periodic signal and in the third state during the second part of each first half period and second half period of the periodic signal. The initialization means are able to set the voltage of the current output to the opposite value of the voltage of the filter capacitor at least at the end of the second part of each first half period of the periodic signal. from the current output to the value of the voltage of the filter capacitor at least at the end of the second part of each second half period of the periodic signal. Preferably, the initialization means comprise an additional set of switches configurable in response to an additional control signal as well as additional control means capable of delivering the additional control signal from the periodic signal. The additional set of switches has a first state in which it connects the filter capacitor to the current output, a second state in which it connects, in a manner opposite to that performed in the first state, the filter capacitor at the output. current and a third state in which it electrically disconnects the filter capacitor from the current output. The additional control means are able to deliver the additional control signal so as to configure said additional set of switches in the first state at least at the end of the second part of each first half period of the periodic signal, in the second state in the second state. less at the end of the second part of each second half period of the periodic signal and in the third state during the first part of each first and second half period of the periodic signal. The additional control means are, for example, capable of delivering the additional control signal so as to configure the said additional set of switches in the first state and the second state respectively during the second part of the first half period and the second half period respectively. of the periodic signal. Preferably, the initialization means furthermore comprise an amplifier stage mounted as a follower between the filtering capacitor and the current output (s) via the additional set of switches. Preferably, the device has a differential architecture.

  In another particular application, there is provided a device in which the filtering circuit and analog / digital conversion is able to filter the intermediate signal so as to filter the frequency components located outside a desired frequency range. The device further comprises reshaping means capable of reshaping the digital output signal in a form substantially identical to the initial shape of the intermediate signal, at least in the desired frequency range, and additional filtering means suitable for performing a final digital filtering of the digital signal reformatted so as to filter the frequency components located outside said desired frequency range. According to one embodiment, the reshaping means comprise a digital fitness filter having a working frequency equal to the control frequency Fs and having a reverse transfer function of the transfer function of the analog filtering effected by the filter circuit and converting analog / digital at least in said desired frequency range. According to another embodiment, the reshaping means comprise a digital fitness filter having a different working frequency and preferably less than the control frequency Fs and having a transfer function whose coefficients result from a approximation of the inverse of the transfer function of the analog filtering performed by the filtering circuit and of the analog / digital conversion at least in the desired frequency range. The coefficients are for example determined so that for each frequency taken in a reference frequency group selected in said frequency range, the difference between the corresponding value of the inverse transfer function of the analog filtering and the corresponding value of the function of Transfer of the fitness filter is minimal. The fitness filter is for example of the finite impulse response type.

  Preferably, the reshaping means and the additional digital filtering means comprise a same filter whose transfer function is equal to the product of the transfer function of the fitness filter by the transfer function of the final digital filter. The device can be realized in the form of an integrated circuit. In another aspect, there is provided a radio frequency signal receiver comprising a device for processing a radio frequency signal as defined above. Other advantages and characteristics will appear on examining the detailed description of embodiments and embodiments, in no way limiting, and the accompanying drawings, in which: FIG. 1 is a block diagram of the internal structure of a embodiment of a device; FIG. 2 represents an embodiment of a filtering and analog / digital conversion circuit; FIGS. 3 to 9 represent different configuration states of the filtering and analog / digital conversion circuit of FIG. 2; and - Figures 10 to 24 schematically illustrate other embodiments and implementation. In FIG. 1, the reference REC denotes a receiver of a radio frequency signal, for example a cellular mobile telephone. The REC receiver comprises an ANT antenna picking up the incident signal, an ETA analog signal processing stage and a digital signal processing stage ETN. The ETA analogue processing stage and the digital processing ETN stage are separated by a CFCANI and CFCANQ filtering and conversion circuit.

  It is considered that the incident signal picked up by the antenna ANT comprises, for example, several frequency channels each having a central frequency specific to each channel and an identical width which will be called frequency range.

  The analog processing stage ETA comprises a LNAI low noise amplifier stage LNAQ connected to the antenna ANT. This LNAI amplifier stage, LNAQ is followed by a frequency transposition stage MTRI, MTRQ. Each frequency transposition stage MTRI, MTRQ receives the signal from the amplifier stage LNAI, LNAQ and here delivers a baseband transposed signal, which will be called an intermediate signal, from a transposition signal. The transposition signals are mutually out of phase by 90 between the two transposition stages MTRI, MTRQ. The transposition signal has a frequency equal to the center frequency of the desired channel. As a result, the intermediate signal is found with a center frequency equal to or approximately equal to the zero frequency. In other words, at the output of the two transposition stage MTRI and MTRQ, on the two processing channels I and Q, two analog signals in quadrature, in baseband, that is to say presenting the signal, are obtained. useful centered around zero frequency. In the rest of the text, we will now describe only one of the two processing channels, for example the voice I, it being understood that the channel Q has a similar structure. At the output of the MTRI mixer, the CFCANI filter and analog / digital conversion circuit is arranged. The CFCANI circuit makes it possible, on the one hand, to filter the channels adjacent to the wanted signal, that is to say the channels located beyond the frequency range of the desired channel and centered on the zero frequency, and on the other hand to convert the filtered intermediate signal into an output digital signal. The digital output signal is then processed in a digital stage ETN of conventional structure and known per se and including a BTDS block for processing and demodulation of the signal. The BTDS block also receives a signal from the Q channel and can then perform the standard channel decoding, including demodulation, and source processing. The BTDS block can be realized in a software way within a processor.

  FIG. 2 represents an embodiment of the filtering circuit and digital analog conversion CFCANI, in differential architecture. It goes without saying that this embodiment is not limited to a differential architecture, and the person skilled in the art will easily adapt the description which follows to a non-differential architecture. In this embodiment, the CFCANI circuit comprises capacitive MCEI input means comprising two filter capacitors CIIRi and CJJR2. The filter capacitors CIIRi and CJJR2 are so-called capacitor capacitors whose capacitance is high and is not reset. It also comprises elementary circuits making it possible, on the one hand, to filter the intermediate signal with the MCEI input capacitive means and, on the other hand, to successively convert the filtered intermediate signal so as to deliver N samples. bits of a digital output signal. Considering the number (2) of filter capacitors and the fact that the CFCANI circuit also has a calibration state, optional (as will be seen in more detail below), the CFCANI circuit has N + 4 elementary circuits.

  In the example described here, N is equal to 3. The CFCANI circuit thus comprises 7 elementary circuits, four of which have not been represented for the sake of simplification. The elementary circuits CI1, Cl2, CI3 each comprise a comparator COMP1, COMP2, COMP3 each having a positive terminal and a negative terminal.

  The elementary circuits CI1, C1, CI3 also include an output SI1, SI2, SI3 each delivering, at the frequency fd, 3-bit samples of the digital output signal. Two Vhigh and VioW reference voltages are also provided. The Vhigh and VioW voltages are preferably chosen so that their average (Vhigh + VioW) / 2 is the working voltage around which it is desired that the MCEI capacitive input means operate and that their difference (Vhigh-Vlow) corresponds approximately to the amplitude of variation of the intermediate signal.

  The elementary circuit CI1 comprises two groups of 4 switches Ill, I12, I13, I14, Ill ', I12', I13 ', I14' and of 4 elementary capacitors CE11, CE12, CE13, CE14, CE11 ', CE12', CE13 ' , CE14 'of respective values C / 2, C / 4, C / 8, C / 8, C / 2, C / 4, C / 8, C / 8. A first group of elementary capacitors comprises the elementary capacitors CE11, CE12, CE13, CE14 and the switches Ill, I12, I13, I14. Each elementary capacitor CE11, CE12, CE13, CE14 comprises a first terminal which is connected to the positive terminal of the comparator COMP1, and a second terminal which is respectively connected to a switch Ill, I12, I13, I14. A second group of elementary capacitors comprises the elementary capacitors CE11 ', CE12', CE13 ', CE14' and the switches Ill ', I12', I13 ', I14'. Each elementary capacitor CE11 ', CE12', CE13 ', CE14' comprises a first terminal which is connected to the negative terminal of the comparator COMP1, and a second terminal which is respectively connected to a switch Ill ', I12', I13 ', I14 . The switches Ill, I12, I13, Ill ', I12', I13 'comprise four switching states and the switches I14 and I14' comprise three switching states. The circuit CI1 also comprises an additional switch Il5 connecting the inputs of the comparator COMP1 to a ground voltage VM. The elementary circuit Cl2 comprises two groups of 4 switches I21, I22, I23, I24, I21 ', I22', I23 ', I24' and four elementary capacitors CE21, CE22, CE23, CE24, CE21 ', CE22', CE23 '. , CE24 'of respective values C / 2, C / 4, C / 8, C / 8, C / 2, C / 4, C / 8, C / 8. A first group of elementary capacitors comprises the elementary capacitors CE21, CE22, CE23, CE24 and the switches I21, I22, I23, I24. Each elementary capacitor CE21, CE22, CE23, CE24 comprises a first terminal which is connected to the positive terminal of the comparator COMP2, and a second terminal which is respectively connected to a switch I21, I22, I23, I24. A second group of elementary capacitors comprises the elementary capacitors CE21 ', CE22', CE23 ', CE24' and the switches I21 ', I22', I23 ', I24'. Each elementary capacitor CE21 ', CE22', CE23 ', CE24' comprises a first terminal which is connected to the negative terminal of the comparator COMP2, and a second terminal which is respectively connected to a switch I21 ', I22', I23 ', I24 . The switches I21, I22, I23, I21 ', I22', I23 'comprise four switching states and the switches I24 and I24' comprise three switching states. The circuit Cl2 also comprises an additional switch I25 connecting the inputs of the comparator COMP2 to a ground voltage VM. The elementary circuit CI3 comprises two groups of 4 switches I31, I32, I33, I34, I31 ', I32', I33 ', I34' and of 4 elementary capacitors CE31, CE32, CE33, CE34, CE31 ', CE32', CE33 ' , CE34 'of respective values C / 2, C / 4, C / 8, C / 8, C / 2, C / 4, C / 8, C / 8. A first group of elementary capacitors comprises the elementary capacitors CE31, CE32, CE33, CE34 and the switches I31, I32, I33, I34. Each elementary capacitor CE31, CE32, CE33, CE34 comprises a first terminal which is connected to the positive terminal of the comparator COMP3, and a second terminal which is respectively connected to a switch I31, I32, I33, I34. A second group of elementary capacitors comprises the elementary capacitors CE31 ', CE32', CE33 ', CE34' and the switches I31 ', I32', I33 ', I34'. Each elementary capacitor CE31 ', CE32', CE33 ', CE34' comprises a first terminal which is connected to the negative terminal of the comparator COMP3, and a second terminal which is respectively connected to a switch I31 ', I32', I33 ', I34 . The switches I31, I32, I33, I31 ', I32', I33 'comprise four switching states and the switches I34 and I34' comprise three switching states. The circuit CI3 also comprises an additional switch I35 connecting the inputs of the comparator COMP3 to a ground voltage VM.

  Before the first configuration state of each elementary circuit, it is considered that it is in its common mode, that is to say that the potential of the comparator terminals is equal to VM, and that the potential of the second The terminal of each elementary capacitor is equal to (Vh, gh + Vlow) / 2.

  FIG. 3 represents the first configuration state of the elementary circuit CI1 which marks the beginning of a filtering and conversion cycle of a sample of the intermediate signal. In this first configuration state, the switches Ill, I12, I13, I14 of the first group of elementary capacitors are configured in the same first switching state in which they connect the second terminal of each elementary capacitor CE11, CE12, CE13, CE14 to the current input of the CFCANI circuit and the filter capacitor CIIR1 connected to the current input. The configuration of the second group of elementary capacitors is similar to that of the first group. In addition, the switch I15 connects the two terminals of the comparator COMP1, and therefore the first terminal of each elementary capacitor of the first and second group of elementary capacitors, to the ground voltage VM. The elementary circuit CI1 is then placed in its first filter sub-state in which the intermediate signal is integrated and sampled in the first and the second group of elementary capacitors for a duration 1 / Fs. The control frequency Fs of the CFCANI circuit therefore defines the sampling frequency.

  This first sub-state filtering performs on the one hand anti-aliasing filtering (in English: anti-alias) through the integration of the signal in the elementary capacitors for a duration 1 / Fs. On the other hand, this first filter sub-state also performs an infinite impulse response filter IIR (in English: infinite impulse response) filter filter capacitor CIIR1 which averages the intermediate signal. The filter capacitor CIIR1 makes it possible to average the intermediate signal over a duration greater than 1 / Fs because the capacitor is not reset. A transfer function low pass filtering is then obtained: Cü (C + CHR1 ûCzzxi z) C FIG. 4 represents the second filter sub-state of the elementary circuit CI1. In this second sub-state of filtering, the switches Ill, I12, I13, I14 of the first group of elementary capacitors are configured in the same second switching state in which they connect the second terminal of each elementary capacitor CE11, CE12, CE13, CE14 to the additional filter capacitor CIIR2. The configuration of the second group of elementary capacitors is similar to that of the first group. In addition, the switch I15 always connects the two terminals of the comparator COMP1, and therefore the first terminal of each elementary capacitor of the first and second group of elementary capacitors, to the ground voltage VM. This second filter sub-state performs a second infinite impulse response filtering command IIR with the additional filtering capacitor CIIR2 which averages the signal previously averaged by the filtering capacitor CIIR1. We then obtain a total low-pass transfer function filtering: C2 (C + Czzxi -CHR1.z). (C + CHR2 -CHR2.z) During the two filtering sub-states of the elementary circuit CI1, the potential of The first terminals of the elementary capacitors remain equal to VM, while the potential of the second terminals of the elementary capacitors vary around (Vh, gh + Vlow) / 2 as a function of the filtered intermediate signal. Thus, if the filtered intermediate signal is positive, then the potential of the second terminals of the first group of elementary capacitors becomes greater than (Vh, gh + V1ow) / 2, for example ((Vhigh + Vlow) / 2) + dV, while that the potential of the second terminals of the second group of elementary capacitors becomes less than (Vhigh + Vlow) / 2, for example ((Vhigh + Vlow) / 2) -dV. At the same time, the second elementary circuit C1 2 is placed in its first filter sub-state in order to process the intermediate signal sample arriving during the second filter sub-state of the elementary circuit CI1. FIG. 5 represents the first state of analog / digital conversion of the elementary circuit CI1. In this first conversion state, the switch I15 is open; for the first group of elementary capacitors, the switch 11 connects the second terminal of the elementary capacitor CE11 (capacitance C / 2) to the reference voltage Vhigh and the switches I2, I3 and I4 connect the second terminal of the elementary capacitors CE12, CE13 and CE14 (of capacity: C / 4 + C / 8 + C / 8 = C / 2) at the reference voltage VIoW; for the second group of elementary capacitors, the switch Il 'connects the second terminal of the elementary capacitor CE11' (capacitance C / 2) to the reference voltage VIoW and the switches I2 ', I3' and I4 'connect the second terminal elementary capacitors CE12 ', CE13' and CE14 '(capacitance: C / 4 + C / 8 + C / 8 = C / 2) at the reference voltage Vhigh. In this way, each group of elementary capacitors behaves as if the second terminals of the elementary capacitors had all been connected to a voltage equal to (Vh, gh + Vlow) / 2.

  The potential variations around (Vh, gh + Vlow) / 2 that appeared at the second terminals of the elementary capacitors during the filter sub-states then become potential variations, of opposite sign, around VM at the level of the second terminals of the elementary capacitors during the filter sub-states. first terminals of the elementary capacitors, that is to say at the terminals of the comparator COMP1. Thus, if the filtered intermediate signal is positive, then the potential of the positive (respectively negative) terminal of the comparator COMP1 becomes equal to VMûdV (respectively VM + dV), and the output SI1 of the comparator COMP1 is negative.

  The most significant bit of the opposite of the digitized signal is then obtained at the output SI1 of the comparator COMP1. If, for example, the output SI1 of the comparator COMP1 is positive, then the most significant bit is 1. If the output SI1 is negative, the most significant bit is 0. It is considered, for the rest of the description, that the bit of most significant at the output SI1 is: 1. At the same time, the second elementary circuit Cl2 is placed in its second sub-state of filtering and the third elementary circuit CI3 is placed in its first sub-state of filtering in order to process the intermediate signal sample arriving during the second filter sub-state of the elementary circuit Cl2 and the first conversion state of the elementary circuit CI1. FIG. 6 represents the second state of analog / digital conversion of the elementary circuit CI1. In this second conversion state, the switches of the elementary circuit CI1 remain in the same switching state as during the first conversion state, except for the switches I12 and I12 'and possibly III and III' as a function of the value of the weight bit. strong.

  Indeed, when the second terminal of an elementary capacitor of value C / 2N of the first group (respectively second group) of elementary capacitors is connected to the reference voltage Vhigh (respectively VioW), the voltage (Vhigh-VIoW) / 2N is subtracted (respectively added) at the level of the voltage of the positive (respectively negative) terminal of the comparator COMP1. The value of the bit then obtained indicates whether the converter dynamics is centered around zero to determine the next bit (in which case the bit obtained is 0), or else subtracting an additional voltage to determine the next bit (in which case the bit obtained is 1).

  Thus, if the most significant bit is equal to 0, it is necessary to add (respectively subtract) (Vhigh-ViOo) / 4 to the positive (respectively negative) terminal of the comparator COMP1 in order to refocus the dynamics of the converter. This is done by connecting the second terminal of the capacitor CE11 to the reference voltage VI ,, and the second terminal of the capacitor of lower capacitance, in this case the elementary capacitor CE12 of value C / 4, to the reference voltage Vhigh. The circuit is then in a configuration for determining the next bit.

  In the present case, the most significant bit is equal to 1, therefore the switches Ill and Ill 'do not change the switching state and the switch' I12 (respectively I12 ') connects the second terminal of the elementary capacitor CE12 ( respectively CE12 ') to the reference voltage Vhigh (respectively VI), so as to subtract (respectively add) the voltage (Vhigh-VIoW) / 4 at the positive (respectively negative) of the comparator COMP1. The second bit of the signal to be digitized is then obtained at the output SI1 of the comparator COMP1. If the output SI1 of the comparator COMP1 is positive, then the second bit is 1. If the output is negative, then the second bit is 0. It is considered, for the rest of the description, that the second bit at the output SI1 is: 0. At the same time, the second elementary circuit C1 2 is placed in its first conversion state, the third elementary circuit CI3 is placed in its second sub-state of filtering and the fourth elementary circuit is placed in its first sub-state of filtering. FIG. 7 represents the third state of analog / digital conversion of the elementary circuit CI1. In this third conversion state, the switches of the elementary circuit CI1 remain in the same switching state as during the second conversion state, except for the switches I13 and I13 'and possibly I12 and I12' as a function of the second bit. When the second bit is equal to 1, the switches I12 and I12 'do not change the switching state. In the present case, the second bit is 0, so the dynamics of the converter is centered around zero by adding (respectively subtracting) a voltage (VhighûVIoW) / 8 to the positive (respectively negative) terminal of the comparator COMP1. This is done by changing the switching state of the switches I12 and I12 'so as to connect the second terminal of the capacitor CE12 (respectively CE12') to the reference voltage VIoW (respectively Vhigh), and by changing the switching state switches I13 and I13 'so as to connect the second terminal of the capacitor CE13 (respectively CE13') to the reference voltage Vhigh (respectively Vi0). The third bit of the signal to be digitized is then obtained at the output SI1 of the comparator COMP1. If the output SI1 of the comparator COMP1 is positive, then the third bit is 1. If the output is negative, then the third bit is 0. It is considered, for the rest of the description, that the third bit at the output SI1 is: 1. At the same time, the second elementary circuit Cl2 is placed in its second conversion state, the third elementary circuit CI3 is placed in its first conversion state, the fourth elementary circuit is placed in its second sub-state of filtering and the fifth elementary circuit is placed in its first filter sub-state. FIG. 8 represents the reference state of the elementary circuit CI1. Indeed, the three bits of the output digital signal sample were determined by the elementary circuit CI1. The circuit CI1 is then reset (reset to a reference state) so that the next signal sample can be processed. In this reference state, the switch I15 connects each terminal of the comparator COMP1 to the ground voltage VM. In addition, the switches Il, I4 ', I3' and I2 'connect the second terminal of the elementary capacitors CE 11, CE14', CE13 'and CE12' to the reference voltage Vh, gh and the switches I2, I3, I4 and II. connect the second terminal of the elementary capacitors CE12, CE13, CE14 and CE11 'to the reference voltage Vi0 so that each group of elementary capacitors behaves as if the second terminals of the elementary capacitors have all been connected to an equal voltage at (Vh, gh + Vlow) / 2. The reference state makes it possible to place the elementary circuit CI1 in a common mode. At the same time, the second elementary circuit Cl2 is placed in its third conversion state, the third elementary circuit CI3 is placed in its second conversion state, the fourth elementary circuit is placed in its first conversion state, the fifth elementary circuit. is placed in its second filter sub-state and the sixth elementary circuit is placed in its first filter sub-state. FIG. 9 represents a calibration state of the elementary circuit CI1. In this calibration state, the switches are in the same switching state as in the reference state except the switch I15 which no longer connects each terminal of the comparator COMP1 to the ground voltage VM. The reference state makes it possible to determine the offset of the comparator COMP1 by making it compare the middle voltage of the first and second group of elementary capacitors which are identical and equal to VM, since the elementary circuit CI1 was placed in the reference state. just before. Thus, the sign of the output SI1 of the comparator COMP1 makes it possible to know the sign of its offset and thus to correct it.

  At the same time, the second elementary circuit Cl2 is placed in its reference state, the third elementary circuit CI3 is placed in its third conversion state, the fourth elementary circuit is placed in its second conversion state, the fifth elementary circuit is placed in its first conversion state, the sixth elementary circuit is placed in its second filter sub-state and the seventh elementary circuit is placed in its first filter sub-state. In the next step, the first elementary circuit CI1 is relocated to the first sub-state of filtering and so on. Thus, all the intermediate signal samples could be processed by the CFCANI circuit. Reference will now be made more particularly to FIGS. 10 to 19 to illustrate another embodiment which, in combination with that illustrated in FIGS. 1 to 9, further improves the performance of the reception chain, in particular with regard to the frequency transposition means MTRI, MTRQ. FIG. 10 thus represents a first embodiment of the frequency transposition stage MTRI, in differential architecture. It goes without saying that this first embodiment is not limited to a differential architecture, and the person skilled in the art will easily adapt the description which follows to a non-differential architecture. In this first embodiment, the LNAI amplifier stage comprises two differential transconductance stages LNAI1 and LNAI2 with differential architecture. The transconductor stage LNAI1 has for example a transconductance equal to Gm, and the stage LNAI2 has for example a transconductance equal to ùGm. The two current differential outputs of the LNAI amplifier stage are thus provided with two signals of opposite sign current. The signals of opposite sign current are then transmitted first to the frequency transposition stage MTRI, then to the filter capacitor CIIR1.

  The transposition stage MTRI comprises a signal generator GEN and control means MC 1. The generator GEN delivers a periodic signal SP (FIG. 13) to the control means MC1 which deliver from said periodic signal SP a command signal SC1 .

  The control signal SC1 makes it possible to control a set of configurable switches. The set of switches connects on the one hand the differential output of the amplifier LNAI1 to the two terminals of the filter capacitor CIIR1 through two transistors T1, T2, and on the other hand the differential output of the amplifier LNAI2 to two terminals of the filter capacitor CIIR1 via two transistors T3, T4. The transistors T1, T2, T3, T4, for example MOS transistors, are used as switches and are respectively controlled by their gate voltage G1, G2, G3, G4. The control signal SC1 is produced from the periodic signal SP so as to alternately connect, at the rhythm of the half-periods of the signal SP, the output of the amplifier LNAI1 and the output of the amplifier LNAI2 to the filtering capacitor CIIR1. The signal SC1 here designates, for purposes of simplification, generally the control signal of the transistors T1 to T4, it being specified of course that this signal SC1 actually consists of four elementary signals respectively controlling the gates of the transistors. The means MC1 are for example made from logic gates.

  The generator GEN also delivers the periodic signal SP to additional control means MC2 which deliver from said periodic signal SP an additional control signal SC2. The MC2 means are also made from logic gates.

  The additional control signal SC2 makes it possible to control an additional set of switches. The additional set of switches connects the two terminals of the capacitor CIIR1 to the differential output of the amplifier LNAI1 via two transistors T5, T6 and a differential architecture follower amplifier block comprising two amplifiers mounted as followers S1, S2. . The additional set of switches also connects the two terminals of the filter capacitor CIIR1 to the differential output of the amplifier LNAI2 via two transistors T7, T8 and the two amplifiers S1, S2. The transistors T5, T6, T7, T8 are used as switches and are respectively controlled by their gate voltage G5, G6, G7, G8 from four logic signals forming generally the control signal SC2. The amplifier block S1, S2 is mounted as followers and is therefore able to impose at its output, its input potential. The control signal SC2, which controls the gates G5, G6, G7, G8, thus makes it possible to impose the voltage U of the capacitor CIIR1 at the output of the amplifier LNAI1 or of the amplifier LNAI2, that is to say say to impose the value of the voltage U as a value of the voltage V1 or V2. Fig. 11 shows the first configuration state of the MTRI transposition stage. In this first state, the control means MC1 blocks the transistors T3 and T4 via the gate voltages G3, G4, while the transistors T1 and T2 are turned on.

  Thus, in the first state, the capacitor CIIR1 is connected to the differential output of the amplifier LNAI1 which delivers a current proportional to the incident signal, and the voltage U of the capacitor CIIR1 is equal to the voltage V1.

  In addition, still for the first state, the additional control means MC2 block the transistors T5 and T6, while the transistors T7 and T8 are turned on. Thus, the amplifiers mounted as follower S1, S2 can impose the value of the voltage U as a value of the voltage V2 of the differential output of the amplifier LNAI2. Figure 12 shows the second configuration state of the MTRI transposition stage. In this second state, the control means MC1 blocks the transistors T1 and T2 via the gate voltages G1, G2, while the transistors T3 and T4 are turned on.

  Thus, in the second state, the capacitor CIIR1 is connected to the differential output of the amplifier LNAI2 which delivers a current proportional to the incident signal and of opposite sign, and the voltage U of the capacitor CIIR1 is equal to the voltage V2. In addition, still for the second state, the additional control means MC2 block the transistors T7 and T8, while the transistors T5 and T6 are turned on. Thus, the amplifiers mounted as follower S1, S2 can impose the value of the voltage U as a value of the voltage V1 of the differential output of the amplifier LNAI1.

  FIG. 13 represents an example of periodic signal SP delivered by generator GEN to control means MC1 and to additional control means MC2. The periodic signal SP has a period Tc corresponding to the transposition frequency, in this case the center frequency of the channel to be processed. During the first half period of the periodic signal SP, the control means MC1 and the additional control means MC2 will configure the MTRI transposition stage in the first state, while during the second half period of the periodic signal SP, the means of MC1 command and the additional MC2 control means will configure the MTRI transposition stage in the second state. Thus, the capacitor CIIR1 receives a current signal whose sign is inverted every half period of the periodic signal. The signal across the capacitor CIIR1, which is the intermediate signal, is therefore a signal transposed into frequency. In addition, in order to avoid parasitic currents at each change of state of the frequency transposition stage MTRI, the value of the voltage U of the capacitor CIIR1 is imposed as the value of the voltage of the differential output of the amplifier LNAI1 or LNAI2 which is not connected to the capacitor CIIR1. FIG. 14 represents a second embodiment of the frequency transposition stage MTRI ', in differential architecture. It goes without saying that this second embodiment is not limited to a differential architecture, and the person skilled in the art will easily adapt the description which follows to a non-differential architecture. In this second embodiment, the LNAI amplifier stage comprises a single transconductor stage having for example a transconductance equal to Gm and therefore a single differential current output. The current signal is transmitted first to the frequency transposition stage MTRI 'and then to the filter capacitor CIIR1.

  The transposition stage MTRI 'comprises a signal generator GEN' and control means MC1 '. The generator GEN 'delivers a periodic signal SP' (FIG. 19) to the control means MC1 'which deliver from said periodic signal SP' a control signal SC1 '.

  The control signal SC1 'makes it possible to control a set of configurable switches. The set of switches connects the differential output of the amplifier LNAI, to the two terminals of the output capacitor CIIR1, either directly via two transistors T1 ', T2', or inverted via two transistors T3. 'and T4'. The transistors T1 ', T2', T3 ', T4' are used as switches and are controlled by their gate voltage, from four logic signals forming generally the signal SC1 '.

  The generator GEN 'also delivers the periodic signal SP' to additional control means MC2 'which deliver from said periodic signal SP' an additional control signal SC2 '. The additional control signal SC2 'makes it possible to control an additional set of configurable switches. The additional set of switches connects the terminals of the capacitor CIIR1 to the differential output of the amplifier LNAI through transistors T5 ', T6', T7 ', T8', and two amplifiers mounted as followers S1, S2. The transistors T5 ', T6', T7 ', T8' are used as switches and are controlled by their gate voltage, from four logic signals forming generally the signal SC2 '. In the present case, the follower differential amplifier block S1, S2 makes it possible to impose the voltage U of the capacitor CIIR1 or its opposite to the output of the amplifier LNAI, that is to say to impose the value of the voltage U or its opposite (-U) as a value of the voltage V. FIG. 15 represents the first configuration state of the transposition stage MTRI '. In this first state, the control means MC1 'block the transistors T3' and T4 ', while the transistors T1' and T2 'are turned on. Thus, in the first state, the capacitor CIIR1 is directly connected to the differential output of the amplifier LNAI and receives a current proportional to the incident signal. The value of the voltage V of the differential output of the amplifier LNAI is the same as the value of the voltage U of the capacitor CIIR1. In addition, still for the first state, the additional control means MC2 'block the transistors T5', T6 ', T7', T8 '.

  Figure 16 shows the second configuration state of the transposition stage MTRI '. In this second state, the control means MC1 'block the transistors T1', T2 ', T3', T4 'so that the capacitor CIIR1 is not electrically connected to the differential output of the amplifier LNAI. On the other hand, the additional control means MC2 'pass transistors T5' and T6 'while transistors T7' and T8 'are off. Thus, the amplifiers mounted as follower S1, S2 can impose the opposite of the value of the voltage U, as a value of the voltage V of the differential output of the amplifier LNAI, before the inversion of the connections between the LNAI amplifier. and the capacitor CIIR1. Fig. 17 shows the third configuration state of the MTRI transposition stage. In this third state, the control means MC1 'block the transistors T1' and T2 ', while the transistors T3' and T4 'are turned on. Thus, in the third state, the capacitor CIIR1 is connected in reverse to the differential output of the amplifier LNAI and receives a current proportional to the incident signal but of opposite sign. The value of the output voltage V of the amplifier LNAI is equal but of opposite sign to the voltage U of the capacitor CIIR1. In addition, for the third state as for the first state, the additional control means MC2 'block the transistors T5', T6 ', T7', T8 '. Fig. 18 shows the fourth configuration state of the MTRI transposition stage. In this fourth state, the control means MC1 'blocks the transistors T1', T2 ', T3', T4 'so that the capacitor CIIR1 is not connected to the differential output of the amplifier LNAI. On the other hand, the additional control means MC2 'pass transistors T7' and T8 ', while transistors T5' and T6 'are blocked. Thus, the amplifiers mounted as follower S1, S2 can impose the value of the voltage U as a value of the voltage V of the differential output of the amplifier LNAI, before the inversion of the connections between the amplifier LNAI and the capacitor CIIR1.

  FIG. 19 represents an example of a periodic signal SP 'delivered by the generator GEN' to the control means MC1 'and to the additional control means MC2'. The periodic signal SP 'has a period Tc corresponding to the transposition frequency, in this case the central frequency of the channel to be processed. During the first half period of the periodic signal SP ', the control means MC1' and the additional control means MC2 'will configure the transposition stage MTRI' in the first state and the second state, while during the second half period of the periodic signal SP ', the control means MC1' and the additional control means MC2 'will configure the transposition stage MTRI' in the third state and the fourth state. During a first part P1 of the first half-period, the signal SP 'configures, via the control means MC1 and additional control means MC2, the transposition stage in the first state. During this first state, the amplifier stage LNAI delivers to the capacitor CIIR1 a signal in proportional and non-inverted current of the incident signal. The output voltage V of the amplifier LNAI is of the same value as the voltage U of the capacitor CIIR1. Then during the second part P2 of the first half-period, the periodic signal SP 'configures the transposition stage in the second state until the end of the first half-period. In the second state, the differential output of the amplifier stage LNAI is no longer connected to the capacitor CIIR1, but the opposite value of the voltage U of said capacitor CIIR1 is imposed as a value of the voltage V of the differential output of the amplifier LNAI. We therefore have a voltage value V equal to (ùU).

  During a first part P1 of the second half period of the periodic signal SP ', the transposition stage is configured in the third state in which the differential outputs of the amplifier stage LNAI are connected to the capacitor CIIR1 inversely. At the beginning of the second half-period, the voltage U of the capacitor is therefore equal to the opposite (-V) of the voltage of the differential output of the amplifier LNAI seen from said capacitor. During this third state, the amplifier stage LNAI delivers to the capacitor CIIR1 a proportional and inverse current signal of the incident signal.

  The output voltage V of the amplifier LNAI is of the same value as the opposite (-U) of the voltage of the capacitor CIIR1. Then during the second part P2 of the second half-period, the periodic signal SP 'configures the transposition stage in the fourth state until the end of the second half-period. In the fourth state, the differential output of the amplifier stage LNAI is no longer connected to the capacitor CIIR1, but the value of the voltage U of said capacitor CIIR1 is imposed as a value of the voltage V of the differential output of the amplifier LNAI . There is therefore a voltage value V equal to the voltage U of the capacitor CIIR1.

  Thus, the capacitor CIIR1 receives, during a portion P1 of each half period, a current signal whose sign is inverted every half period of the periodic signal. The signal at the terminals of the capacitor CIIR1 is therefore a signal transposed into a frequency which constitutes the intermediate signal. Moreover, in order to avoid parasitic currents at each half-period of the periodic signal SP ', the value of the voltage U or its opposite (-U) is imposed as the value of the voltage V of the differential output of the LNAI amplifier. during a P2 game of the second half of each half-period. Finally, the first part P1 of the first half-period and the first part P1 of the second half-period may be chosen so as to have an equal or different duration. It is the same for the second part P2 of each of the half periods. In addition, the relative duration of the first part P1 and the second part P2 of the same half-period will be adapted according to the case, so as to allow optimum operation of this first embodiment. Reference will now be made more particularly to FIGS. 20 and following to illustrate another embodiment which, in combination with that illustrated in FIGS. 1 to 9 or with that illustrated in FIGS. 10 to 19, further improves the performance of FIG. the reception chain. In this example, the filtering performed by the CFCANI circuit is a low order analog filtering. In the particular case of a low-pass filtering of order 1 and of cutoff frequency fo, this means that a signal at a frequency equal to 10.fo will be attenuated by 20 decibels. If one wishes to have a greater attenuation at the frequency 10.fo, or one chooses a filtering of higher order (a filtering of order 2 will give a attenuation of 40 decibels with 10.fo), one chooses a frequency of cut f'o smaller than fo (because the attenuation increases between 10.f'o and 10.fo). Since the order of the filtering is fixed by the number of capacitors for filtering capacitive input means MCEI, efficient filtering of the adjacent channels is done by choosing a cut-off frequency situated in the desired frequency range, that is to say in the desired channel. However, this has the effect of also filtering the useful signal. Thus, for the frequencies of the desired frequency range that are larger than the cutoff frequency, the attenuation is at least 3 decibels (this corresponds to the attenuation of the cutoff frequency). The filtering of the CFCANI circuit therefore distorts part of the signal. In return, the adjacent channels are filtered despite the low order of the filtering. Thus, the analog / digital conversion of the filtered intermediate signal will require a lower dynamic, i.e. a smaller number of bits.

  Ideally, sampling preserves all information when the sampling frequency Fs is at least twice the highest frequency of the signal to be sampled. Thanks to the cut-off frequency of the filtering performed by the voluntarily low chosen circuit CFCANI, the sampling frequency Fs can also be chosen low. In addition, since the analog filtering has been important, the amount of information of the filtered intermediate signal, including outside the useful signal, is reduced. The filtered intermediate signal can then be digitized with a smaller number of bits.

  We will now describe in more detail the internal structure of the digital processing stage ETN used in this embodiment. The ETN stage includes a FRFI fitness filter. The filter FRFI is a digital filter, for example a finite impulse response filter whose transfer function is chosen equal to or substantially equal to the inverse of the filter transfer function performed by the CFCANI circuit over at least the desired frequency range. that is, in the desired signal channel.

  This FRFI filter makes it possible to reconstruct the incident signal both inside and outside the desired frequency range. Thus, at the output of the FRFI filter, there is a signal close to the intermediate signal coming out of the mixer MTRI. Digital filtering is then performed again, for example low-pass or bandpass, in order to filter the adjacent channels reshaped by the FRFI filter again. The digital processing stage ETN thus comprises a digital FPBNI filter making it possible to filter the adjacent channels. However, the FPBNI digital filter can now have a high order, since it is easier, less cumbersome and less expensive to make a high order digital filter than a high order analog filter. It is therefore possible to produce an FPBNI filter capable of filtering the adjacent channels by little or no deformation of the useful signal.

  Thus, thanks to the device as described, it is possible to filter the useful signal with a digital filter without the need for an analog-digital converter having a high dynamic. The FRFI filter and the FPBNI filter can be made within the same FNI digital filter. To this end, it suffices to multiply the respective transfer functions of the two filters FRFI and FPBNI and to digitally produce the resulting INF filter. The realization of a single filter performing both the signal shaping and high-order low-pass filtering makes it possible to simplify the production of the device and to improve the useful signal obtained. The final signal obtained on channel I is then sent to the BTDS block.

  FIG. 21 represents a diagram on which the various steps of an implementation of the method corresponding to the architecture of FIG. 20 are mentioned. Thus, there is provided a first step 1 during which an analog filtering and a conversion are performed. analog / digital intermediate signal. Analogue filtering is deliberately low here, that is to say that at least a part of the useful signal is also filtered in order to limit the frequency width of the filtered intermediate signal. The digital output signal thus requires a smaller number of bits. In a second step 2, the digital output signal is reshaped in order to compensate for the deformation of the filtered intermediate useful signal. This reshaping can be carried out in particular by using a filter whose transfer function is equal to the inverse of that of the analog filtering of the CFCANI circuit. Once the output digital signal has been reformatted, the adjacent channels are filtered again because they have also been reshaped during the second step 2. Thus, the third step 3 comprises a digital filtering of the digital output signal reformatted in order to keep only the useful signal. In particular, the final digital filtering can be performed with a much higher order than the analog filtering, which allows the filtering of the adjacent channels without the deformation of the wanted signal. We will now describe in more detail the second step 2. It is considered in a first embodiment that the FRFI fitness filter is clocked at a working frequency equal to the sampling frequency Fs of the analog / digital converter. . The FRFI filter then has a transfer function equal to the inverse of the transfer function of the analog filtering performed by the CFCANI circuit. For example, by taking the expression G from the transfer function of the analog filtering performed by the CFCANI circuit and writing it in the equivalent form F in digital, we have C2 (C + Czzxi ùCzzx1 • e-2inf / Fs). The FRFI filter is then chosen to have a transfer function H equal to: (C + CHR1 ùCzzxi • e-2inf / Fsv. (C + CIIR2 ùCzzx2) e-2inf / Fs) H = F i = C2 Those skilled in the art will easily perform this transfer function H in digital.

  In another embodiment, the case in which the filter FRFI is clocked at a working frequency Fs / Ndi, different from that of sampling Fs of the analog / digital converter, is considered in another embodiment. The H function can no longer be determined as before and the transfer function of the FRFI fitness filter is then a

  Happrox approximation of the inverse H of the transfer function of the analog filtering performed by the CFCANI circuit. This Happrox approximation is the closest to the H function in the signal band that we want to reformat.

  Fig. 22 shows a determination diagram of a

  The function of approximate transfer of the function H. Consider the case in which the filter FRFI is a finite impulse response filter of order Nfir. The Happrox transfer function of the FRFI filter is written as: f 2f 2 1.'Ndii • 'F, 2E + TN {ivF Hap, prax 0 [1 ~ ~ ts ...} ë? 14fir 30 The determination of the Happrox approximated transfer function

  comprises determining the coefficients [ao; ai; ...; aNfir-1] F = 5 10 L i IMir-1 '.f Ndiv F minimizing the difference between the transfer function H and the Happrox function over the width of band of the useful signal. In a first step 10, therefore, the form of the Happrox transfer function is chosen, and more particularly the order of the Happrox transfer function which provides the number of coefficients [ao; ai; ...; aNfir_i] to be determined. In a second step 20, a finite number of reference frequencies f2,..., FNfrq are chosen for which the difference between the two transfer functions H and HapproX will be minimized. The choice of these frequencies and their number will directly determine the precision of the approximation, as well as the complexity of the calculation of HapproX. Thus, reference frequencies are preferably uniformly distributed over the desired frequency range. A vector Fref is then constructed whose coordinates are the reference frequencies determined previously, then, in a third step 30, the vector Href is calculated whose coordinates are the values of the transfer function H for each of the reference frequencies. So we have: I Fre {= f and: Href "" H1. Fref'I fi H (f1) H (f2) Nfrq. H (f_Nfrq) Finally, in a fourth and last step 40, the coefficients [ao; ai; ...; aNfir_1] are computed in a matrix manner, so as to minimize the difference between the desired transfer function H and the Happrox approximated function. The calculation uses a first matrix E defined by: fir-1 f Nfir - 1 wl E = exp 2in f Nfrq t'Nfrq f Nfir --- 1 f Nfrq from which the Happrox coefficients are calculated with the following formula FIGS. 23 and 24 graphically illustrate an example of an approximation of a transfer function H by a FIR filter of order 3 (Nf, r = 3), clocked at a frequency 6 times smaller ( Nd, v = 6) and Happrox transfer function. FIG. 23 represents the module of the two transfer functions and FIG. 24 represents the phase shift. It can be seen that over the frequency range PFD [-10MHz; 10MHz] representing a frequency channel, the curves have similar values. The difficulty of designing filters and the analog-to-digital converter to the digital part is thus postponed, whatever the working frequency of the digital filters and the sampling frequency Fs of the converter. 30

Claims (24)

  1. A method of processing an incident signal in which an intermediate signal is produced from the incident signal, the intermediate signal is delivered to a filtering and analog-to-digital conversion circuit (CFCANI) comprising at least one current input for receiving the intermediate signal, several outputs (SI1, SI2, SI3), input capacitive means (MCEI) connected to the input, a plurality of configurable successive conversion circuits (CI1, Cl2, CI3) each connected between the input input and one of the outputs (SI1, SI2, SI3) and each comprising a group of switched elementary capacitors, a process in which each elementary circuit (CI1, Cl2, CI3) is successively and cyclically configured: - in a filtering state in which one connects the group of switched elementary capacitors to the capacitive input means, then - in several distinct states of analog / digital conversion, then - in a reference state in which the group of switched elementary capacitors 20 has a reference capacitive load, said outputs each successively delivering a sample of several bits of a digital output signal corresponding to the filtered intermediate signal. 25   2. Method according to claim 1, in which the filtering and analog / digital conversion circuit (CFCANI) comprises at least N + 2 outputs (SI1, SI2, SI3) each delivering, at a delivery frequency fd, a sample of N bits of the digital output signal, and N + 2 elementary circuits (CI1, Cl2, CI3) each connected between the input and one of the N + 2 outputs (SI1, SI2, SI3) and each comprising at least one group of N + 1 switched elementary capacitors connected in parallel (CE11, CE12, CE13, CE14, CE21, CE22, CE23, CE24, CE31, CE32, CE33, CE34) as well as a comparator (COMP1, COMP2, COMP3) connected to the corresponding output (SI1, SI2, SI3), a method in which each elementary circuit (CI1, Cl2, CI3) is successively and cyclically configured at a control frequency Fs equal to at least the product of the number of elementary circuits by the delivery frequency fd, and with a time offset of 1 / Fs between two elemental circuits consecutive ires, and wherein: - the group of N + 1 elementary capacitors is connected to the capacitive input means in the filtering state, - the N + 1 capacitors are connected between the comparator (COMP1, COMP2, COMP3) and a first or a second reference voltage (Vhigh, Vi0) selectively in order to obtain, successively at the frequency Fs, the N bits of a sample of the digital output signal at the corresponding output (SI1, SI2, SI3) in N separate analog-to-digital conversion states, and - the group of N + 1 elementary capacitors has a reference capacitive load in the reference state.   3. The method as claimed in claim 2, in which the ratio of the capacitance value of the capacitive input means (MCEI) to the sum of the capacitances of the N + 1 elementary capacitors is greater than a threshold chosen as a function of the bandwidth. analog signal.   4. Method according to claim 2 or 3 wherein the capacitive input means (MCEI) comprise a filter capacitor (CIIR1) and wherein, in the filtering state of each elementary circuit (CI1, Cl2, CI3), the group of N + 1 elementary capacitors of said elementary circuit is connected to the filtering capacitor (CIIRI).   5. Method according to one of claims 2 to 4 wherein: - the capacitive input means (MCEI) further comprise at least one additional filter capacitor (CIIR2), - the filter circuit and analog conversion / digital (CFCANI) comprises a number of additional outputs and additional elementary circuits equal to the number of additional filtering capacitors, and- the filtering state of each elementary circuit (CI1, Cl2, CI3) comprises different filter sub-states in wherein the group of N + 1 elementary capacitors is respectively connected to the different filter capacitors (CIIR1, CIIR2) in which said elementary circuit (CI1, Cl2, CI3) is successively placed in each of the filter sub-states.   6. Method according to one of claims 2 to 5 wherein, after the reference state, the elementary circuit (CI1, C1, C13) is configured in a calibration state in which the comparator (COMP1, COMP2, COMP3) of the elementary circuit (CI1, Cl2, CI3) at two equal voltages so as to determine and correct its offset.   7. Method according to one of claims 2 to 6 wherein the current input for receiving the intermediate signal is a differential input and wherein each elementary circuit (CI1, Cl2, CI3) comprises a second group of N + 1 capacitors switched elementary elements connected in parallel (CE11 ', CE12', CE13 ', CE14', CE21 ', CE22', CE23 ', CE24', CE31 ', CE32', CE33 ', CE34').   8. Method according to one of claims 2 to 7 in combination with claim 4, wherein the incident signal is delivered to a transconductor stage (LNAI), a current output of the transconductor stage (LNAI) is connected to the capacitor. filter (CIIR1) so as to deliver to the filtering capacitor (CIIR1) a current signal and its opposite respectively during the two half periods of each period of a periodic signal (SP, SP ') and thus obtain, at the level of filter capacitor (CIIR1), a signal transposed into frequency as an intermediate signal, and in which, at the occurrence of each half period of the periodic signal (SP, SP '), the voltage of said current output, view of said filter capacitor (CIIR1), is initialized to a value equal to that of the voltage (U) of said filter capacitor (CIIR1).   9. Method according to one of claims 2 to 8 in combination with claim 4 wherein the filtering circuit and analog / digital conversion (CFCANI) is able to filter the frequency components of the intermediate signal located outside a frequency range desired, and the method further comprises - reshaping (2) the digital output signal in a substantially identical form, at least in said desired frequency range, to the form of the intermediate signal and - a final digital filtering (3 ) of the output digital signal reshaped so as to filter the frequency components located outside said desired frequency range.   An incident signal processing apparatus comprising a filter and analog-to-digital conversion circuit (CFCANI) of an intermediate signal derived from the incident signal, comprising: - at least one current input for receiving the intermediate signal ; - several outputs (SI1, SI2, SI3); Capacitive input means (MCEI) connected to the input; a plurality of configurable successive approximation circuits (CI1, C1, CI3) each connected between the input and one of the outputs (SI1, SI2, SI3) and each comprising a group of switched elementary capacitors; each elementary circuit (CI1, Cl2, CI3) having a filter state in which the group of elementary capacitors is connected to the capacitive input means, a plurality of distinct analog / digital conversion states, and a reference state in which the group switched elementary capacitors has a reference capacitive load, said outputs being adapted to successively output a sample of several bits of a digital output signal corresponding to the filtered intermediate signal.   11. Device according to claim 10 further comprising control means capable of successively and cyclically placing each elementary circuit in the filtering state, then in the successive states of analog / digital conversion, then in the reference state. .   12. Device according to claim 11 wherein the outputs (SI1, SI2, SI3) are each capable of delivering, at a delivery frequency fd, a sample of several bits of the digital output signal, and wherein the control means are capable of placing, at a control frequency Fs equal to at least the product of the number of elementary circuits by the delivery frequency fd and with a time shift of 1 / Fs between two consecutive elementary circuits, successively each elementary circuit (CI1, Cl2, CI3 ) in the filtering state, then in the analog / digital conversion states in order to obtain, successively at the frequency Fs, the bits of a sample of the digital output signal at the corresponding output (SI1, SI2, SI3) , then in the reference state.   13. Device according to claim 12 wherein the filtering circuit and analog / digital conversion (CFCANI) comprises: - at least N + 2 outputs (SI1, SI2, SI3) each capable of delivering at the fd delivery frequency, a sample of N bits of the digital output signal, - at least N + 2 elementary circuits each connected between the input and one of the N + 2 outputs (SI1, SI2, SI3) and each comprising at least one group of N + 1 switched capacitors (CE11, CE12, CE13, CE14, CE21, CE22, CE23, CE24, CE31, CE32, CE33, CE34) connected in parallel and N + 1 configurable switches (Ill, I12, I13, I14, I21, I22 , I23, I24, I31, I32, I33, I34) as well as a comparator (COMP1, COMP2, CMOP3) connected to the corresponding output (SI1, SI2, SI3), the group of N + 1 elementary capacitors of each elementary circuit : - being connected to the capacitive input means in the filtering state of said elementary circuit, - being connected between the comparator and a first or a second reference voltage (Vhigh, Vi0) selectively in the N distinct states of analog / digital conversion of said elementary circuit, and - having a reference capacitive load in the reference state of said elementary circuit.   14. Device according to claim 13 wherein the ratio of the value of the capacitance of the input capacitive means to the capacitance sum of the N + 1 elementary capacitors is greater than a threshold chosen as a function of the bandwidth of the analog signal.   15. Device according to claim 13 or 14 wherein the values of the N + 1 elementary capacitors of each elementary circuit are C / 2, C / 4, ..., C / 2N, C / 2N.   16. Device according to claim 15 wherein the capacitive input means (MCEI) comprise a filter capacitor (CIIRI) and wherein, for each elementary circuit (CI1, Cl2, CI3), the switches (Ill, I12, I13 , I21, I22, I23, I31, I32, I33) of the N elementary capacitors of values C / 2, C / 4, ..., C / 2N- ', C / 2N have three switching states and the switch ( I14, I24, I34) of the (N + 1) 1st C / 2N value capacitor has two switching states; the control means being able to configure the switches of an elementary circuit in the same switching state during the filtering state, in the other switching states during the N distinct states of analog / digital conversion and in the state of reference.   17. The device according to claim 16, wherein: the capacitive input means (MCEI) furthermore comprise at least one additional filtering capacitor (CIIR2); the device comprises a number of additional outputs and additional elementary circuits equal to to the number of additional filter capacitors, and for each elementary circuit (CI1, Cl2, CI3), each switch further has a further number of switching states equal to the number of additional filter capacitors; the filtering state of each elementary circuit comprising different filter sub-states in which the group of N + 1 elementary capacitors is respectively connected to the different filtering capacitors (CIIR1, CJJR2); the control means being able to configure the switches of an elementary circuit so as to successively place it in each of the sub-states of filtering.   18. Device according to one of claims 13 to 17 wherein, in the reference state of an elementary circuit (CI1, Cl2, CI3), the switch (III, I21, I31) of the elementary capacitor (CE 1 1, CE21, CE31) of C / 2 value of said elementary circuit is configured to connect said capacitor elementary value C / 2 to one of two reference voltages (Vhigh) and the switches (I12, I13, I14, I22, I23, I24, I32, I33, I34) of the N other elementary capacitors (CE12, CE13, CE14, CE22, CE23, CE24, CE32, CE33, CE34) are configured to connect said N elementary capacitors to each other. reference voltages (Vi0).   19. Device according to one of claims 13 to 18 wherein the filtering circuit and analog / digital conversion (CFCANI) comprises an additional output and an additional elementary circuit and wherein each elementary circuit (CI1, Cl2, CI3) has in addition, a calibration state in which the comparator (COMP1, COMP2, COMP3) is connected to two equal voltages so as to determine and correct its offset; the control means being further adapted to configure the switches of each elementary circuit so as to place it in its calibration state after its reference state.   20. Device according to one of claims 13 to 19 having a differential architecture wherein each elementary circuit (CI1, Cl2, CI3) comprises a second group of N + 1 elementary capacitors (CE11 ', CE12', CE13 ', CE14'). , CE21 ', CE22', CE23 ', CE24', CE31 ', CE32', CE33 ', CE34') connected in parallel and N + 1 configurable switches (Ill ', I12', I13 ', I14', I21 ', I22', I23 ', I24', I31 ', I32', I33 ', I34').   21. Device according to one of claims 10 to 20 in combination with 16 further comprising: - a transconductor stage (LNAI) comprising an input for receiving the incident signal, - frequency transposition means (MTRI) connected between the transconductor stage (LNAI) and the filtering capacitor (CIIRi), comprising: - generation means (GEN, GEN ') capable of delivering a periodic signal (SP, SP'), - a set of configurable switches (T1, T2 , T3, T4, T1 ', T2', T3 ', T4') in response to a control signal (SC1, SC1 '), - control means (MC1, MC1') adapted to output the control signal ( SC1, SC1 ') from the periodic signal (SP, SP') for configuring the set of switches (T1, T2, T3, T4, T1 ', T2', T3 ', T4') so as to connect an output to current from the transconductor stage (LNAI) to the filter capacitor (CIIR1) and deliver to the filtering capacitor (CIIR1) a current signal and its opposite respectively during the two half periods of each period of the periodic signal (SP, SP ') and thus obtain, at the level of the filtering capacitor (CIIR1), a signal transposed in frequency as an intermediate signal, the device further comprising suitable initialization means to initialize, at the occurrence of each half period of the periodic signal (SP, SP '), the voltage of the current output, seen from said filtering capacitor (CIIR1), to a value equal to that of the voltage (U ) of said filter capacitor (CIIR1).   22. Device according to one of claims 10 to 21 in combination with 16 wherein the filtering circuit and analog / digital conversion (CFCANI) is able to filter the intermediate signal so as to filter the frequency components located outside of a desired frequency range, the device further comprising: - fitness means (FRFI) capable of reshaping the digital output signal in a form substantially identical to the initial form of the intermediate signal, at least in the frequency range desired, and - additional filtering means (FPBNI) capable of performing a final digital filtering of the digital signal reformatted so as to filter the frequency components located outside said desired frequency range.   23. Device according to one of claims 10 to 22, made in the form of an integrated circuit.   24. A radio frequency signal receiver (REC) comprising a device for processing a radio frequency signal according to one of claims 10 to 23.