WO2007074287A2

WO2007074287A2 - Method for common mode rejection with differential line pairs

Info

Publication number: WO2007074287A2
Application number: PCT/FR2006/051384
Authority: WO
Inventors: Fahd Hassoun; Ahmed Zeddam
Original assignee: France Telecom
Priority date: 2005-12-22
Filing date: 2006-12-19
Publication date: 2007-07-05
Also published as: WO2007074287A3; FR2895628A1

Abstract

The invention concerns a method for transmitting data in differential mode on a two-wired link using a common attenuation mode, corresponding connection device and computer programme. The invention concerns a method for transmitting data in differential mode on a link comprising at least two conductors, to a receiver equipment, for attenuating the effects of the common mode. In accordance with the invention, such a method includes a step of attenuating at least one common mode signal interfering with said data transmission, which consists, on each of said conductors, in: estimating by adaptive filtering of said common mode signal carried on said conductor; determining at least one parameter for adapting said conductor, for reconstructing said common mode signal estimated on said conductor; injecting in phase opposition said common mode signal reconstructed on said conductor.

Description

A method of transmitting differential mode data over a two-wire link implementing a common mode mitigation, connecting device and corresponding computer program.

FIELD OF THE INVENTION The field of the invention is that of the transmission of data in differential mode on a two-wire link, or more generally on a link comprising a plurality of conductors. More specifically, the invention relates to attenuation, or even suppression, of common-mode signals disturbing such differential-mode data transmission. It applies in particular, but not exclusively, to the transmission of high-speed data on a pair of twisted copper or on a power line.

2. Prior Art

High-speed data transmission techniques of the "Digital Subscriber One"("Digital Subscriber One") type have grown considerably in recent years, since they allow transmission rates of several Mbits to be attained. / s, without disrupting the existing infrastructure of fixed telephone networks. Indeed, these techniques allow the joint transmission of voice and multimedia data on a single copper twisted pair, and are therefore much less expensive to implement than fiber optic transmissions. In particular, they allow most users broadband access to the global Internet network, regardless of their distance from the central office. Other high-speed transmission techniques are currently under study _" which include the use of power lines for voice and data broadcast by Current Carrier Online (or CPL). The user accesses the network via a connection device, commonly called a modem, which is connected on the one hand to the two-wire link of the network carrying the data, and on the other hand to the user's equipment (telephone , computer, server _" etc.). Such a connection of the modem to these devices can be carried out in radio-electric form, for example according to the WiFi (registered trademark) standard. 3. Disadvantages of prior art

In these high-speed transmission systems, the useful signal is transmitted in differential mode between the two conductors of the two-wire link. The quality of the transmission is therefore highly dependent on the symmetry of the conductors. Indeed, during transmission, an electromagnetic field is radiated by each of the conductors. If the two conductors are perfectly symmetrical, the fields created by each wire are of neighboring intensities and opposite directions, so that the resulting electromagnetic field is almost zero.

Conversely, an asymmetry of the conductors causes a conversion of the differential mode voltages into common mode voltages. These common mode voltages generate an electromagnetic field that is radiated by the pair of conductors.

While the symmetry of the conductors is generally verified within a given range of operating frequencies, it often degrades as the frequencies increase. In addition, the dissymmetry of the transmission and reception equipment of the communication network may also lead to the appearance of conducted and radiated emissions. Indeed, certain external disturbances, as well as the particular mode of injection of the useful signal on the two-wire link and the nature of the network connection equipment can also lead to a conversion from a differential mode to a common mode, and vice versa. .

For transmission in differential mode on a pair of twisted cables, the existence of emissions conducted in common mode and the existence of radiated electromagnetic fields are therefore directly related.

However, such radiating fields are also likely to disturb the electromagnetic environment of the conductors. For example, high-speed transmission systems deployed on copper pairs, which use high frequencies, can disrupt radio services using the same frequency band. Indeed, the lines behave like antennas HF (high frequency), and the ray fields can harm the operations of some neighboring devices (for example domestic appliances) or certain services, like amateur radio for example. 4. Objectives of the invention

The invention particularly aims to overcome these disadvantages of the prior art. More specifically, an object of the invention is to provide a differential mode data transmission technique on a link comprising at least two conductors, which generates conducted emissions and radiated electromagnetic fields, reduced compared to the prior art. In other words, the invention aims to provide such a differential data transmission technique that reduces the common mode compared to existing solutions.

The invention also aims to provide such a technique which has a reduced sensitivity to possible dissymmetries of the transmission network, and which is capable of operating in a very wide range of transmission frequencies.

Another objective of the invention is to propose such a technique which is particularly well suited to broadband transmission of the xDSL type on copper networks, but also to power line transmission in line, or

CPL, on electrical networks. The invention also aims to provide such a technique that does not affect the useful information to be transmitted, and that is simple and inexpensive implementation.

5. Presentation of the invention

These objectives, as well as others which will appear later, are achieved by a differential mode data transmission method on a link.

∞making at least two drivers, to receiving equipment.

According to the invention, such a method comprises a step of attenuating at least one common mode signal disturbing said data transmission, implementing on each of said conductors; an adaptive filtering estimation of said common mode signal carried on said driver; a determination of at least one adaptation parameter of said conductor, for reconstructing said estimated common mode signal on said conductor; an injection in phase opposition of said reconstructed common mode signal on said conductor.

Thus, the invention is based on a completely new and inventive approach to the transmission of data in differential mode on a two-wire twisted pair link. Indeed, the invention proposes to consider the common mode, generated during transmission by possible dissymmetries of the conductors, as an undesirable disturbance to be minimized, to quantify this common mode, then to suppress it without affecting the useful information flowing in differential mode. To do this, the technique of the invention is based on a clever use of adaptive filtering, combined with an adaptation (source and impedance) of the drivers for injecting the unwanted common mode in phase opposition in each of the conductors, so as to reduce its effects.

This reduces the common mode affecting the transmission, and consequently the radiated emissions that are directly related to it, which preserves the electromagnetic environment of the receiving equipment.

In addition, the reduction of the conducted and radiated emissions obtained by the implementation of the invention makes it possible to amplify the power of the transmitted signals (that is to say useful information), which offers the advantage to increase the range of the system and the electromagnetic immunity with respect to external disturbances. It will be noted that the invention applies to any type of link on which data can be transmitted in differential mode, that is to say to the links which comprise at least two conductors, and therefore in particular to two-wire links and more generally muitîfilaires. For the sake of simplification, this type of bond will be referred to by the term "two-wire link" in the remainder of this document. Similarly, although the invention is particularly advantageous in case of high-speed transmissions, elfe can be applied to any type of transmission in differential mode, regardless of its frequency.

Advantageously, the common-mode signal that is estimated and reinjected in phase opposition is a common mode current. Indeed, it is relatively easy to dispose of the global common-mode current flowing in the receiver equipment, from which we can therefore, by adaptive filtering, determine an estimate of the common-mode disturbing current flowing on each of the conductors.

Thus, during said estimation, an adaptive filtering of said common-mode current measured in said receiver equipment is performed from a current measured on each of said conductors.

Preferably, said data transmission is of xDSL or CPL type. The invention also applies, of course, to any other type of transmission, regardless of its bit rate and its support (electrical network, communication network, etc.). According to an advantageous characteristic of the invention, said determination of said parameters of adaptation implements, for each of said conductors: a determination of an adaptation voltage source for performing said anti-phase injection of said reconstructed common mode current on said conductor; a determination of an impedance of adaptation of said conductor; and said matching impedances are chosen so as not to alter a differential mode voltage between the conductive bits.

Thus _"after estimating the sources that allow injected in opposite phase the estimated common mode on each wire is estimated your matching impedances for imposing the same common mode voltages than those measured before adjustment.

The invention also relates to a device for connecting a client terminal to a communication network, iedit network implementing a differential mode of data transmission to said device, on a link comprising at least deox conductors. According to the invention, such a device comprises means for attenuating at least one common mode signal disturbing said data transmission, comprising: adaptive filtering estimation means of said common mode signal carried on each of said conductors; means for determining at least one adaptation parameter of each of said conductors, for reconstructing said estimated common mode signal on said conductor; injection means in phase opposition of said reconstructed common mode signal on each of said conductors.

Preferably, said estimation means comprise at least two adaptive adapters of the RLS (Recursive Least Mean Square) type with vector input each associated with one of said conductors. Of course, we can also consider using other adaptive filters, such as Least Mean Square (LMS) filters or NLMS (Normalized Least Mean Square) algorithms (standardized least squares algorithm). for example.

Advantageously, said device comprising at least two input terminals each connected to one of said conductors, said injection means comprise at least two matching circuits, each connected to one of said input terminals and each comprising: an adaptation voltage source making it possible to carry out said injection in opposition of the phase of said signal; in common mode rebuilt on iedit driver; an impedance of adaptation of said conductor; and the adaptation impedances are chosen so as not to alter a differential mode voltage between said input terminals.

According to an advantageous characteristic of the invention, said device belongs to the group comprising: a CPL type modem; an xDSL type modem.

The invention also relates to a computer program comprising program code instructions for executing the steps of the method as previously described when said program is executed on a computer. 6. List of figures

Other advantages and characteristics of the invention will emerge more clearly on reading the following description of a particular embodiment of the invention, given as a simple illustrative and nonlimiting example, and the appended drawings, among which: : - Figure 1 shows, as a reminder, a synoptic of an adaptive filter; FIG. 2 illustrates the principle of minimizing the square error implemented in the adaptive filter of FIG. 1; FIGS. 3A and 3B respectively show the principle of data transmission in differential mode on the one hand, and in common mode on the other hand, on a two-wire link; FIG. 4 illustrates the relationship between the common-mode current flowing on a two-wire link and the electric field radiated by this link; FIG. 5 schematizes the context in which the invention is inscribed, in the form of the equivalent electrical circuit of a transmission central and of the two-wire link connecting it to a client modem; Figure 6 shows in flowchart form the common mode reduction technique of the invention; FIG. 7 describes the equivalent electrical circuit of FIG. 5 after adaptation of the conductors according to the principle of FIG. 6; FIG. 8 illustrates the results obtained by implementing the invention; FIG. 9 represents a block diagram of a client modem in which a reduction of the common mode is performed according to the principle of the invention. ?. Description of a particular embodiment of the invention

The general principle of the invention is based on the reduction of the common mode affecting differential transmission on a two-wire link by: adaptive filtering quantization of this common mode on each driver; adaptation on each driver to inject in opposite phase

The common mode previously quantified, so as to reduce this undesirable and disruptive mode. Recall first, in relation with Figure 1, the principle of adaptive filtering, which is based on the search for optimal parameters by minimizing a performance criterion. Frequently this minimization is done by searching for the least squares. Adaptive filtering is particularly well suited to the case where the spectra of the wanted signal and the disturbing signal are superimposed, which is the case here for the common mode and the differentiated mode! propagated on the two-wire link.

It is recalled that an adaptive filter is a digital system whose coefficients themselves change as a function of the external signals. It is conventionally constituted of two distinct main parts: a digital filter 10 with adjustable coefficients w _k ; an algorithm 11 for modifying the coefficients w _k based on an optimization criterion, in this case the minimization of the mean squared error σ].

In Figure 1, the adaptive filter is constituted by all the elements located outside the dashed lines. More precisely, in the diagram of FIG. 1, there is shown: the known or measured excitation signal x (n); the output signal y _p (n) unreachable from an unknown process; the measured output signal y (n) has an unknown disturbance e (π); the modeled signal y _w (n) using the parameters w _k of the digital filter 10; the difference signal ε (n) between the model yj _^ n) and the measure y (n). Such an adaptive filter makes it possible, according to its use, to obtain an estimate of the output signal y _p (n) or the unknown disturbance e (n), from the excitation signal x (n).

According Ie principle of the recursive least square algorithm (RLMS) which is préférentieϋemβnt used in the context of the invention, the filter w _k of the optimal parameters are obtained by reaching Ie minimum mean square error. As illustrated by FIG. 2, which represents the quadratic error ε ² (n) at instant n and its derivative with respect to the coefficient w n), this solution can be reached recursively by correcting the values of the coefficients w _k in every instant n in the opposite direction to the evolution of the quadratic error with respect to the vector of the coefficients of the filter W (n):

where y is a gradient weighting factor. As the quadratic error at time n is:

= (y (n) -X (n) ^τ w) ² ₍₂₎

He comes:

We deduce that the search for the optimum can be done with the following recursive algorithm:

W (n) = W (n - 1) + 2γε (n) X (n) (4) which is designated by the name of LMS (Least Mean Square) algorithm. The quantities that must be available are:

the vector of the p coefficients of the adaptive filter at instant / 1-1:

W (n-1) = [WOCn-I) ₁ WiCn-I), -, wp-1 (n-1) f (5)

the last p values of the input signal:

X (n) = [x (n), x (n-1), -, x (n-p + 1) f (6) - the value of the output signal y (n) to calculate the difference to instant n: ε (n) = y (n) -] r "w _iX (ii-f) ₍₇₎

the adaptation gain y of the dream algorithm (generally much less than 1). The value of the adaptation gain is difficult to fix: if it is chosen too weak, the convergence towards the optimal value is very high, if it is chosen too strong, the convergence is done by oscillating for a long time around the value of i'optfmum and finally _"if the adaptation gain is too éievé the optimization process diverges. The adaptive filtering algorithm has the advantage of being particularly simple to program, and to require few calculations. On the other hand, it has the disadvantage of converging slowly, with the risk of diverging when the adaptation gain is too great. A variant of this adaptive filtering algorithm, called the NLMS algorithm, uses a gain normalized by the power σ _x ^" of the signal x (n), Such an algorithm can also be used as an adaptive filtering algorithm in the context of the invention. The principle of this NLMS algorithm, which is well known to those skilled in the art, will not be discussed in greater detail.The principle of the transmission of data over a two-wire link is now recalled in connection with FIGS. , on which the signal can propagate in two ways: in differential mode, as shown in Figure 3A, in common mode, as shown in Figure 3B, propagation is done in differential mode when the signal is transmitted to a single The differential mode current I _MD propagates on the "out" conductor 30, flows through the receiving equipment 31 and returns via the "return" conductor 32.

Conversely, the propagation is done in common mode when the disturbance is transmitted to all active drivers. The common mode current I _MC propagates in part on each of the two conductors 30, 32 in the same direction (i.e. a first common mode current I _MC propagates on the first conductor 30 and a second current common mode I _MC2 propagates in the same direction on the second conductor 32). The common mode current I _M c then loops back to the ground 33 through the parasitic capacitances 34 of the receiving equipment 31.

The current in common mode! _M c may be connected, as a first approximation, to the electric field E radiated by the two-wire link 30, 32 by equation (12); _j _ HiCi T + ^τ ï \ 1C2. 17 __ 2 f __ * z TL_ * J I_M £ _

2 0.8 * r ( ¹² > With:

E in μV / m; f the frequency in MHz of data transmission on the two-wire link 30, 32; r (in m) the distance from the cable 30, 32; and where L represents the length (in m) of the two-wire link 30, 32.

FIG. 5 illustrates this relation linking the common-mode current I _MC to the radiated electric field E, in a particular case of transmission at a frequency f = 1 MHz, at a distance r = 1 m from the cable 30, 32, and for a length of two-wire link L = 20m.

As shown in FIG. 5, the reduction, or even the suppression, of the common mode current read thus makes it possible to reduce or even eliminate the electric field E radiated by the cable 30, 32 and the receiver equipment 31, and therefore mitigate the disturbances that they are likely to cause for their environment (other neighboring equipment, radio services operating in the same frequency range, etc.).

The problem that the invention proposes to solve is that of the suppression or reduction of the common-mode current I _MC in the context of FIG. 6, on which is represented, by way of example: a central telephone (or more generally a communication center) 50, modeled by a voltage source E; a client 51, having an impedance receiver equipment Z; a two-wire link (twisted pair, cable, etc.) comprising two conductors

30 and 32 connecting the receiving equipment of citation 51 to the central 50. As shown in FIG. 8, the data transmission from the central unit 50 to the customer 51 is in a differential mode, which is defined by the differential mode voltages V ₁ and V ₂ at the respective terminals of each of the conductors 30 and 32. The current flowing in each of the conductors 30 and 32 is respectively denoted I ₁ and fe.

The disturbance generated for example by the dissymmetry of the conductors 30, 32 is represented schematically by a common mode current 1 _MC which loops in parallel with the receiving equipment.

FIG. 6 shows a flowchart of the main steps implemented by the invention for reducing, or even completely eliminating, the common-disturbing current I _M0 of FIG. 5, namely: a step 61 of adaptive filtering, preferably by implementation of a recursive algorithm of the RLMS type; a step 63 for calculating adaptation parameters of the conductors 30, 32 constituting the two-wire link.

These two steps 61, 63 are implemented in two specific modules (namely a LIVIS algorithm quantization module and a transmission line adaptation module) implemented in the client modem 51.

In the particular embodiment of the invention described below, the adaptive filtering 61 is of vector-input RLMS type. More precisely, two filters RLMS1 and RLMS2, each associated with one of the conductors 30, 32, and operating independently, are preferably used. The input data 60 feeding the adaptive filtering module 61 are as follows: the common mode current I _M c measured in the client equipment 51; the current I ₁ measured on the first conductor 30; the current k measured on the second conductor 32. Referring to the block diagram of FIG. 1, each adaptive filter

RLMS1 _» RLMS2 thus functions as follows: it has as input an excitation signal x (n) consisting of a frequential vector representative of the common mode current y _c> measured on the client side 51 _" which allows to have a complete picture of this mode. H also has a measured output signal y (n) with a disturbance e (n), which is constituted by a frequency variation of the current I ₁ , 1 ₂ measured on the associated conductor 30, 32.

For the two adaptive filters RLMS1 and RLMS2 to the filtering module 81, we therefore have:

[xi (n) ^≈ I _M c, yl (n> * h \ => RLMS1 / x ₂ (m) = h _m y2 (n) ^≈ hl ^≈ > RLMS2 (13) At the output 62 of the adaptive filtering module 61, an estimate is then obtained of the common-mode currents flowing on each of the conductors 30 and 32 of the two-wire link. Indeed, the first adaptive filter RLMS1 makes it possible to determine an estimate II _MCSSI of the common mode current conveyed by the first conductor 30, from the excitation signal

which can be seen as the sum of a useful differential mode current I1 _MD and a disturbing common-mode current 11 _MC - Similarly, the second adaptive filter RLMS2 makes it possible to determine an estimate l2 _M Cest of the common-mode current conveyed by the second conductor 32, from the excitation signal x (n) = \ _U c and the output signal y ₂ (n) = \ _2, corresponding to the sum of a differential mode current I2 useful _MD and a disturbing common mode current I2 _MC -

The common mode currents estimated at _{M C} is and 2 _MCest , delivered at the output of the adaptive filtering module 61, supply the adaptation module 63, the function of which is to determine the adaptation parameters of the conductors 30 and 32 of the link two-wire, taking into account the common-mode voltages V ₁ and V ₂ , so as to be able to reconstruct and reinject in each of the conductors 30, 32, but in phase opposition, the estimated common-mode currents H _M _C is and l2y _C es _t -

The operating constraint of this adaptation step 63 is the maintenance of the useful information intact, that is to say that the differential mode voltage VMD = V ₁ -V ₂ must not be altered by this adaptation 63 , so as not to disturb the data transmission from the central office 50 to the customer 51.

In the particular embodiment of the invention described here, the adaptation step 63 comprises the following substeps.

_Firstly , it is estimated that the voltage sources E _is1 and E _{βst2 are} used to inject in opposite phase the estimated common mode H _MCest , I2 _M e _t on each 30, 32, knowing the equivalent impedance Z the client modem 51,

In a second step, the adaptation impedances Z _684I and Z _βsf2 of each conductor 30 _" 32 are estimated, these adaptation impedances making it possible to guarantee that the common-mode voltages (V ₁₁ V ₂ ) across each of the conductors 30, 32 remain identical to the common mode voltages measured before adaptation, this which makes it possible to maintain the differential mode voltage V _MD = Vi - V ₂ .

Moreover, these adaptation impedances (Z _is i, Z _esi2 ) make it possible to improve the symmetry of the transmission on the two-wire link 30, 32.

In other words, during the adaptation step 63, it is determined from the current H _{MC is} delivered by the adaptive filter RLMS1, and taking into account the common mode voltage V ₁ across the first conductor 30, the adaptation parameters (E _βst1 , Z _e5H ) of this conductor 30, making it possible to reinject in opposite phase into the conductor 30 the estimated common disturbance mode current H _M c _e s _t - Similarly, it is determined from FIG. current 12 _M c _is delivered by the adaptive filter RLMS2, and taking into account the common mode voltage V ₂ across the second conductor 32, the adaptation parameters (E _est2 , Ze _s c) of this conductor 32, allowing to reinject in phase opposition in the conductor 32 the estimated disturbing common-mode current 12 _M is

FIG. 7 shows the diagram of FIG. 5 after adaptive filtering and adaptation of the conductors 30, 32. In practice, the physical adaptation therefore consists of intervening at the input terminals of the client receiving modem 51, whose impedance is denoted Z. An adaptation circuit 70, comprising the voltage source E _is i and the impedance Z _βst i, is connected to the input terminal of the client modem connected to the conductor 30. Likewise, an adaptation circuit 72, comprising the voltage source E _est2 and the impedance Z _βst2 , is connected to the input terminal of the client modem connected to the conductor 32. These adaptation circuits 70, 72 can be integrated into the client modem 51, and reconfigured on a regular basis _" for each data transmission, according to its frequency and the state of the network. For example _"can be chosen to repeat the adjustment 63 for each new transmission of data by the central 50, or on every change of transmission frequency.

Thanks to the matching circuits 70, 72, the current flowing in the first conductor 30 is now equal to hh _MCe s _t (that is, to the estimation error, equal to the useful differential mode current 11 _MD ) and the current flowing in the second conductor 32 is now ia-I2 _MG e * t (that is, in the estimation error near I2 _MD ) - It will be noted that by modem means here the device allowing the customer 51 to connect via the data transmission network to the central station 50. It is clear that, in the case of a digital transmission network of the ADSL type for example, such a connection device does not, strictly speaking, carry out modulation / demodulation, but is however commonly called modem. The technique of the invention therefore consists, in order to reduce the common mode, to intervene on the client side 51, at the level of the modem, and not at the level of the central office 50 or at any other point in the transmission network. The advantage of this choice is that the invention thus makes it possible to act on the total common mode at the end of the line. In fact, it is thus possible to take into account the various linear parameters of the cable (resistance, inductance, capacitance, ...), the total common mode, the length and the symmetry of the twisted pair, the differential mode, which makes it possible to determine the appropriate adaptation for each pair to adapt.

The results obtained by simulation are very convincing, as illustrated in FIG. 8, in which: the curves of the two left columns illustrate the voltages and currents obtained without adaptation, ie without implementation of the invention; the curves of the two columns on the right illustrate the voltages and currents obtained after quantization of the common mode and adaptation of the conductors according to the principle of the invention. More precisely, the upper curves represent the voltages V _MD and the differential mode currents I _MD , at each end of a two-wire link of length L (V _MD0 and V _MDL , I _MDO and I _MDL ). Likewise, the median curves represent common mode voltages V _MC and currents IM _C at each of the 0 and L ends of the line. Finally, the lower curves illustrate the symmetry of the two-wire link. As can be seen, the technique of the invention therefore makes it possible to change the symmetry of the installation from -28.5dB without adaptation to a mean value centered around -55dB,

Finally, in connection with FIG. 9, a block diagram of a client modem (that is to say a connection device to the transmission network considered) is presented in which the reduction of the common mode described is implemented. previously. Such a modem comprises, besides the means specific to the network connection functions, a central unit P ₁ equipped for example with a microprocessor, a memory M, for example of the RAM ("Read Access Memory") type and a software module Pg At initialization, the instructions of the software Pg are for example loaded from the memory M to be executed by the microprocessor of the central unit P.

In a new data transmission, the modem receives a disturbed differential mode signal from a common mode transmission, MD + MC. The central unit P executes the instructions of the program code Pg, to perform an adaptive filtering to quantify the common mode MC, according to the method described previously in this document, then perform an adaptation of the drivers to reinject the common mode estimated in phase opposition in the drivers. In this way, at the output of the central unit P, the only differential mode MD is recovered, which constitutes the useful information.

Claims

A method for transmitting data in differential mode on a link comprising at least two conductors, to a receiver equipment, characterized in that it comprises a step of attenuating at least one common-mode signal disturbing said data transmission. , implementing on each of said conductors: an adaptive filtering estimate of said common mode signal carried on said conductor; a determination of at least one adaptation parameter of said conductor, for reconstructing said estimated common mode signal on said conductor; an anti-phase injection of said reconstructed common mode signal on said conductor.

2. Transmission method according to claim 1, characterized in that said common mode signal is a common mode current.

3. A transmission method according to claim 2, characterized in that, in said estimation, an adaptive filtering of said common mode current measured in said receiving equipment, is carried out from a current measured on each of said conductors.

4. Transmission method according to any one of claims 1 to 3, characterized in that said data transmission is of type xDSL or CPL.

5, A method of transmission according to any one of claims 2 to 4, characterized in that said determination of said adaptation parameters implements _" for each of said conductors: - a determination of a voltage source of adaptation allowing performing said phase opposition injection of said reconstructed common mode current on said conductor; determining an impedance of adaptation of said conductor; and in that said adaptation impedances are chosen so as not to alter a differential mode voltage between said conductors.

6. Device for connecting a client terminal to a communication network, said network implementing a differential mode data transmission to said device, on a link comprising at least two conductors, characterized in that it comprises means attenuation of at least one common-mode signal disturbing said data transmission, comprising: adaptive filtering means for estimating said common mode signal carried on each of said conductors; means for determining at least one adaptation parameter of each of said conductors, for reconstructing said estimated common mode signal on said conductor; injection means in phase opposition of said reconstructed common mode signal on each of said conductors.

7. Connection device according to claim 6, characterized in that said estimating means comprise at least two adaptive filters of RLMS vector input type each associated with one of said conductors.

8. Connection device according to any one of claims 6 and 7, characterized in that, said device comprising at least two input terminals each connected to one of said conductors, said injection means comprise at least two circuits. adapter means, each connected to one of said input terminals and each comprising: an adaptation voltage source for performing said phase opposition injection of said reconstructed common mode signal on said conductor; an impedance of adaptation of said conductor; and in that said adaptation impedances are chosen so as not to alter differential mode voltage between said input terminals.

9. Connection device according to any one of claims 6 to 8, characterized in that said device belongs to the group comprising: a type CPL modem; - an xDSL type modem.

A computer program comprising program code instructions for performing the steps of the method according to any one of claims 1 to 5 when said program is executed on a computer.