WO2009127695A1

WO2009127695A1 - Method for cable testing with insertion of directional coupling device

Info

Publication number: WO2009127695A1
Application number: PCT/EP2009/054541
Authority: WO
Inventors: Josy Cohen; Nicolas Ravot
Original assignee: Commissariat A L'energie Atomique
Priority date: 2008-04-18
Filing date: 2009-04-16
Publication date: 2009-10-22
Also published as: FR2930348A1; FR2930348B1

Abstract

The invention relates to a method and device for testing network cables in order to detect and locate defects in the network cables. A coupling device (CPL) is placed at the junction point between a main cable section and two secondary sections, whereby said device can allow the passage of test signals with a first attenuation from a first inlet (INa) to a first outlet (OUTa) and the passage of measurement signals with a second attenuation from the second outlet (OUTb) to the first inlet (INa), but not the passage of test or measurement signals between the two outlets. Thus the multiple reflections disturbing the test signal analysis are minimized. The test signals are emitted from the inlet of the main section and are reflected at the impedance discontinuities, which may be due to cable defects.

Description

METHOD OF TESTING CABLES WITH DEVICE INSERTION

DIRECTIONAL COUPLING

The invention relates to a method and a device for testing network cables, for detecting and locating faults in the cables of this network.

The networks concerned may be energy transmission networks or low-speed communication networks, in fixed installations (distribution network, internal or external communication network) or mobile networks (energy or communication network in an airplane , a boat, an automobile, etc.). The cables concerned are low frequency transmission cables (i.e. non-coaxial cables); these are for example two-wire cables, in parallel lines or in twisted pairs, shielded or not, etc., provided that the speed of propagation of the signals in these cables can be known and is sufficient to generate measurable propagation delays. These networks can be organized according to different known topologies: bus, tree, mesh, ring, star, linear, or mixed of these different topologies. The networks concerned by the invention are networks in which the junction points between branches of the network only require, for the normal operation of the network, connections by simple contacts between the conductors (for example soldered or crimped or pressed contacts). Are not concerned networks whose normal operation requires, at the junctions, couplers with controlled input or output impedance (microwave couplers, optical couplers, etc.) establishing controlled attenuations between a main branch and branches. secondary. An important application of the invention relates to the testing of electrical cable networks in vehicles, especially motor vehicles, and more particularly:

- the test of the supply networks in the vehicle,

- and the test of the communication buses operating according to the CAN (Control Area Network) protocol or an equivalent protocol. The CAN protocol is intended for signals whose flow rate is of the order of 1 Megabit / second, corresponding to much lower working frequencies than those which will be used for the test.

The defects concerned by the test method according to the invention are defects that can affect the electrical operation of the circuits of which the cables are part and which may have sometimes very critical consequences (failures of electrical systems in an aircraft for example), or even, in the case of electric cables, faults that can directly cause fire starts (short circuits, arcs in dry environment or in the presence of moisture, etc.). It is important to be able to detect these defects to remedy them in time.

It is understood that the problem of fault detection is all the more important that cable networks are longer and more complex or that they are more difficult to access (buried cables for example). This is why remote detection and localization systems have been devised, operating from one end of the cable. The methods used are so-called OTDR methods, in which a signal injected at one end of a cable propagates in this cable and part of the amplitude of the signal is reflected at the location of the defect, for example due to the impedance discontinuity that the signal encounters at this location. If the speed of propagation of the signals in the cable (related to its characteristic impedance in the case of an electric cable) is known, the measurement of the time which separates the wave emitted from the reflected wave gives an indication of the distance between the end of the cable and the fault.

In Time Domain Reflectometry (TDR) methods, an electromagnetic wave is injected into the cable in the form of a voltage pulse, a voltage step, or the like. The wave reflected at the location of the impedance discontinuity is detected at the injection site and the time difference between the transmitted and received edges is measured. The position of the fault is determined from this difference, and the amplitude and polarity of the reflected pulse give an indication of the type of fault (open circuit, short circuit, resistive fault, or other). There are also Frequency Domain Reflectometry (FDR) methods, which consist in injecting at the input of the cable a sinusoid wobulated in frequency continuously or in steps and in measuring the frequency deviation or phase between the emitted wave and the reflected wave. The published patent application WO 02/068968 describes a frequency domain reflectometry method. In a variant called SWR for "Standing Wave Reflectometry", the nodes and bellies of a stationary wave generated by the combination of an incident wave and the reflected wave are detected. Frequency domain reflectometry methods are effective for analyzing a single cable. They are difficult to use when the cable has taps. The time domain reflectometry methods can be used even with derivations but the analysis of the reflected signals is difficult due to the presence of multiple reflections.

It has also been proposed in the published patent application WO 2004/005947, a method both temporal and frequency of injecting a wobulated signal linearly with a Gaussian amplitude envelope. Spectrum spread reflectometry methods have also been proposed in the article "Spread Spectrum Sensors for Critical Fault on Live Wire Networks" by Cynthia Furse et al. In Journal of Structural Control and Health Monitoring, Volume 12, Issue 3 -4, 2005. A signal is transmitted as a low-level pseudo-random code on a network, even when in use; this signal and its echo generated by the possible defect are correlated with variable time offsets to establish a correlation curve as a function of time. This curve has correlation peaks at time offsets related to the position of the faults and junctions and / or branches of the network. This system is particularly suitable for detecting intermittent faults as it can operate even while the network is in use; however, intermittent faults may only occur when the network is in service and disappear when it is no longer (eg a fault that would occur while a plane is flying but disappears on the ground). This method can be used for cables with taps, but it retains ambiguities: it is not possible to say on which branch a fault is detected. In addition, it requires significant calculation costs.

Finally, a similar method but simply using the signals or the natural noise circulating in the cable, and not a pseudo-random code injected at the entrance of the cable, was proposed in the article by Chet Lo and Cynthia Furse " Noise-Domain Reflectometry for Locating Wiring Faults "published in IEEE Transactions on Electromagnetic Compatibility, Vol 47 No. 1, February 2005. High correlation peaks are detected in a signal correlation process with itself. This method suffers from the same defect as the previous one, that is to say that it does not easily make it possible to remove ambiguities of position when there are several branches and it requires important means of calculation.

The object of the invention is to facilitate the interpretation of the measurement results in these methods, for electrical cables having a T-structure (also called a Y-structure), that is to say having at least one shunt. More specifically, the invention seeks to minimize the multiple reflections generated by the presence of several branches in the network, reflections that make it more difficult to interpret the results of the measurements. For this purpose, the invention proposes a method of testing a cable network comprising at least one main cable section and at the end of this section a junction from which at least a first and a second secondary section depart, the network being such that the junctions require, for the normal operation of the network, only connections by simple contacts between the conductors (in particular welded or crimped or pressed contacts), the method comprising the injection at the input of the main section of test signals capable of propagating in the network and and collecting on this input measurement signals resulting from the reflection of the test signals on the discontinuities present in the network, the method being characterized in that it places the of the junction, to ensure the passage of test signals and measurement signals, a coupling device having at least a first input and two outputs, this device is nt adapted to ensure the passage of test signals with a first non-zero attenuation from the first input to a first output and the passage of measurement signals with a second attenuation from the second output to the first input, but not the passage of test or measurement signals between the two outputs, the first input of the coupling device being connected to the end of the main section and each output being connected to the input of a respective secondary section. The coupling device may be a two-input power divider having the same first attenuation of the input to each of the two outputs and having the same second attenuation from each output to the input. The coupling device may also be an N (N> 2) power divider having a single input and N outputs each connected to a respective one of N, the divider having the same first attenuation of the input to each of N outputs and having the same second attenuation since each output to the input.

In another embodiment, the coupling device comprises at least one unidirectional coupler having an input, an output, and an auxiliary terminal (coupled input / output); it has a first attenuation of the input towards the output, a second attenuation, greater than the first, of the auxiliary terminal towards the input; moreover, it does not allow or almost no passage of measurement signals from the output to the input (while the power divider allows this passage). The coupler input is connected to the end of the main section; the output and the auxiliary terminal are respectively connected to the inputs of the secondary sections. But in practice, for reasons of symmetry, the coupling device comprises two unidirectional couplers; the second is constituted as the first (and preferably has the same attenuations); their two entries are united; the output of a coupler is connected to the auxiliary terminal of the other and vice versa.

By extension, if there are N pairs of secondary sections connected to the end of the main section, N groups of two unidirectional couplers, N> 1, will be placed for the test of a network comprising N pairs of secondary sections connected to the main stretch. And if there are only 2N-1 secondary sections, one will connect to the auxiliary terminal of one of the couplers a suitable load instead of the non-existent secondary section.

The coupling device may alternatively comprise, in place of two unidirectional couplers, a bidirectional coupler having a first input / output torque and a second input / output torque and being adapted to ensure the passage of the test signals with a first attenuation from each input to the respective output of the same pair but not the passage of test signals of the entry of a couple towards the exit of the other pair; it is also able to ensure the passage of the measurement signals with a second attenuation of the output of each pair towards the input of the other pair but not or almost none of the output to the input of the same pair or the passage of test or measurement signals between the two outputs; the inputs are both connected to the end of the main section and the outputs are respectively connected to each secondary section.

In the case where it has 2N or 2N-1 pairs of secondary sections connected to the junction, with N> 1, N bidirectional couplers will be used. One of the outputs will be connected to a suitable load in the case of an odd number (2N-1) of secondary sections. The two inputs of all the couplers are connected to the end of the main section.

In the case of a cable intended for the transport of energy or of low frequency signals (frequency much lower than the frequency of the test signals, for example at least 10 times lower), a respective low-pass filter can be interposed between the end of the main section and the input of each of the secondary sections, this filter passing energy or signals at low frequency and not passing the microwave test signals. It can also be provided that a bias tee (in English "bias tee") is interposed between the end of the main section and the input of the coupling device, and another bias tee is interposed between a respective output of the device coupling and a respective input of each of the secondary sections, the bias tees ensuring the separation between the high frequency components (the test signals) and the low-frequency components (the signals useful for the operation of the network) of the signals they receive on their input. The coupling device inserted at the junction is a passive device, as are the bias tees and low-pass filters.

Other features and advantages of the invention will appear on reading the detailed description which follows and which is given with reference to the appended drawings in which: - Figure 1 shows a network of shunt cables that it is desired to analyze;

FIG. 2 represents the general principle of the invention;

FIG. 3 represents a first embodiment of the invention;

FIG. 4 represents a second embodiment of the invention;

FIG. 5 represents a third embodiment of the invention; FIG. 6 represents an application in the case of an even number (2N) of secondary sections connected to a bypass;

FIG. 7 represents an application in the case of an odd number (2N-1) of secondary sections;

FIGS. 8 and 9 represent embodiments in which it is desired to retain the possibility of continuing to transmit continuous or low-frequency signals without being obliged to disconnect the coupling devices intended for network testing.

To simplify the explanations, the invention will be described in the case of a time domain reflectometry.

FIG. 1 diagrammatically shows a network with one junction and two branches having three sections T1, T2 and T3. The sections T2 and T3 have input ends E2, E3 connected to a junction point A located at an output end S1 of the section T1. Thus, if we follow the network starting from an input end E1 thereof, we successively meet a cable length L1 of the section T1 to its output end S1, then a junction at A, and, according to whether one follows the section T2 or the section T3, respectively a cable length L2 of the section T2 of the input end E2 to an output end S2, or a length of cable L3 of the end of E3 input of the section T3 to an output end S3. The junction point A connects the ends S1, E2 and E3.

This is a simple example of a T (or Y) network. The sections considered and represented by a line may consist of a sheathed conductor wire or a pair of sheathed son. This network can be used to transporting energy or communication signals from the input E1 to the outputs S2 and S3, or in the opposite direction from an output S2 or S3 to the other or to the input E1. This is why the notions of input and output are used here only for the application of reflectometry and to fix the direction of emission of measurement signals that will be used for the detection of faults: the input E1 will be here by hypothesis the one on which the test signals of the network and the section T1 are applied will be called main section whereas the sections T2 and T3 will be called secondary sections. The outputs of the sections T2 and T3 may be in open circuit or short circuit or adapted to the impedance of the respective cables (if they are adapted they do not cause reflection).

The network is a network in which the junctions consist of simple electrical contacts between a conductor of a section and a conductor of another section (contacts without electrical losses). The network is therefore not a network that would require for its normal operation links via controlled impedance couplers and attenuations, as is found for example in coaxial cable television networks or in optical networks. The contacts are contacts by welding (soldered contacts) or by pressure (crimped contacts, removable contacts inserted into a connector).

In the case of conventional time domain reflectometry, a test pulse is applied from the input E1, and a signal pattern called a "reflectogram" is collected on this same input; the reflectogram is the plot of a curve representing the evolution of a voltage amplitude recorded at the input E1 over time. The very short duration of the pulse is such that it is possible to distinguish the signal emitted from the reflected signal, taking into account the propagation speed (assumed to be known) and the minimum distance that is to be detected for the localization. faults.

FIG. 2 represents the principle of the test method according to the invention. The test is done with signals whose bandwidth includes frequencies much higher than the normal operating frequencies of the network in operation. Typically, for a power transmission network, normal usage frequencies can range from DC to a few tens of Hertz or a few hundred Hertz. For a communication network between sensors in a vehicle car (network type CAN, Control Area Network), the useful bandwidth is a few megahertz, allowing rates of the order of 1 megabit / second. For the test signals, the bandwidth will be several tens of megahertz or more in order to be able to determine the positions of the defects by detecting the time positions of the return signals reflected by these defects. The bandwidth of the test signals is therefore at least ten times greater than the bandwidth of the signals of normal use of the network.

A coupling device CPL is placed at the junction point which interposes - with regard to the test signals - between the output end S1 of the main section and the input ends E2, E3 of the sections secondary.

The coupling device has the following minimum properties: it passes the test signals transmitted on the input E1 of the main section both to the section T2 and to the section T3, with a first attenuation; it can pass the measurement signals in the opposite direction with a second attenuation which is not necessarily the same as the first one, but it does not let signals pass (or it allows them to pass insignificantly for the measurement , that is to say with a very strong attenuation) between the two outputs, so as to avoid multiple reflections. Finally, means are provided for passing the signals (voltages and electric currents) from the normal operation of the circuit, whether these signals are communication signals in a communication network, or that they are supply voltages and currents. energy in a power grid. Note that the means for passing the signals of normal operation are only useful if the coupling devices are permanently present at the junctions; they would not be so if the coupling devices were set up only for the test and removed in normal operation. In a first embodiment, represented in FIG. 3, the coupling device consists of a set of two unidirectional microwave couplers CpA and CpB each having an input and two outputs. The coupler CpA has a main input INa, an output OUTa and an auxiliary terminal (also called coupled input / output) AUXa. The coupler CpB has a main input INb, an output OUTb and a auxiliary terminal AUXb. The auxiliary terminal of one coupler is connected to the output of the other coupler and vice versa.

For each coupler, the properties are as follows: the test signal applied to the input goes to the output with a first attenuation which is small, for example between 0 and 3dB, but this signal does not pass towards the auxiliary terminal ( almost infinite attenuation). A feedback measurement signal applied to the auxiliary terminal passes to the input with a second attenuation greater than the first, for example of the order of 20 dB but does not pass to the output. A measurement signal applied to the output does not pass to the auxiliary terminal and may pass to the input but with very high attenuation. In addition, the ports preferably have impedances adapted to the characteristic impedance of the line to which they are connected, for example a standardized impedance of 50 ohms at the working frequency of the test signals or more generally in the bandwidth of these signals.

The inputs INa and INb of the two unidirectional couplers are joined and connected to the output S1 of the secondary section T1. The output OUTa of the first unidirectional coupler is connected with the auxiliary terminal AUXb of the second, and it is connected to the input E2 of a secondary section T2. Symmetrically, the output OUTB of the second coupler is connected with the auxiliary terminal AUXa of the first and is connected to the input E3 of the other secondary section T3.

Thus, the test signals from the main section enter the unidirectional couplers and pass on the outputs, for example with a 3dB attenuation, but do not pass to the auxiliary terminals. The test signals reflected on the impedance discontinuities of the secondary section T2 return to the output OUTa and to the auxiliary terminal AUXb. On the output OUTa they are strongly attenuated (for example 8OdB) and do not come to form a significant return signal towards the main section. On the AUXb input, they are attenuated with an average attenuation, for example of 20 dB, and can be exploited at the input of the main section to determine the positions of impedance discontinuities.

Symmetrically, the test signals reflected on the impedance discontinuities of the other secondary section T3 return to the output OUTb and to the auxiliary terminal AUXa. On the output OUTb they are strongly attenuated (for example 8OdB) and do not come to form a significant return signal towards the main section. On the input AUXa, they are attenuated with an average attenuation, for example of 20 dB, and can be exploited at the input of the main section in superposition with the signals from the section T2.

The multiple reflections of the test signal at the ends of the secondary sections are eliminated because of the very strong isolation existing between an auxiliary terminal and the output of the same coupler, for example between the terminal AUXa and the output OUTa (in both directions). spread).

In another embodiment shown in FIG. 4, the coupling device consists of a bidirectional coupler having two inputs INa and INb and a respective output OUTa, OUTb associated with each input. The two inputs are joined and connected to the end S1 of the main section. The output OUTa is connected to the input E2 of the first secondary section and the output OUTb is connected to the input E3 of the second secondary section.

The bidirectional coupler preferably has inputs and outputs having impedances adapted to the characteristic impedances of the sections to which they are connected. The transmission and coupling characteristics of the bidirectional coupler are as follows: between an input such as INa and the associated output OUTa, the test signals can pass, with a slight attenuation, for example between 0 and 3dB. In the other direction, from OUTa to INa or from OUTb to INb, the attenuation is very strong, for example 8OdB so that the signals propagate in this direction only insignificantly. Between an output, for example OUTa, and the input which is not associated with it, in this case INb, the signals (which are the measurement signals after reflection on impedance discontinuities of the secondary sections) can pass with attenuation. average, for example 2OdB. These signals, although attenuated, will return to the main section and may be exploited on the input E1. Between an input, for example INa and the output which is not associated with it, for example OUTb, the signals can not pass. Finally, between two inputs or two outputs, signals can not pass. Such unidirectional or bidirectional couplers are very conventional microwave components and are commonly commercially available.

Multiple reflections in the secondary sections are again eliminated.

In yet another embodiment, the microwave coupling device is a simple divider / power combiner. A divider / combiner in its simplest version has one input and two outputs. It divides the signal applied to the input into two half energy signals (attenuation of 3dB) that it transmits to the losses near on each of the outputs; and it has a reciprocal combiner function, i.e., the signals applied to the outputs pass to the input without attenuation (near losses). The inputs and outputs preferably have impedances adapted to the characteristic impedance of the main and secondary sections at the frequency used for the test signals. The insulation between the two outputs is total. Such passive divider / combiner components are common microwave.

Fig. 5 shows a solution using a divider / combiner by two designated by the reference DIV. The IN1 input of the combiner divider is connected to the output S1 of the main section. Each output OUT1, OUT2 of the divider is connected to the input E2 or E3 of a respective secondary section.

The test signals reflected by the section T2 to the output OUT2 do not propagate towards the section T3, avoiding multiple reflections which make it more difficult to interpret the measurement signals collected on the input E1 of the main section.

The preceding embodiments relate to the case where there are only two secondary sections connected bypass on the main section. In the case where there is a number of sections greater than 2, several similar coupling devices or a multiple output coupling device will be used.

If the number of secondary sections is an even number 2N with

N greater than 1, for example, the configuration of FIG.

N bidirectional couplers. If the number of secondary sections is an odd number, then N bidirectional couplers will also be used by connecting to the unused output of the last coupler a T or charge termination. impedance adapted to that of this output, in accordance with the diagram of FIG. 7.

The diagrams of FIGS. 6 and 7 are transposable in the case where a set of two unidirectional couplers is used (in accordance with the diagram of FIG. 3) in place of a bidirectional coupler. In this case, for 2N-1 or 2N bypass secondary sections, there are N groups of two unidirectional couplers, the outputs of each group being connected to a group of two respective sections. In the case of an odd number of secondary sections, a coupler output remains unused; a termination or appropriate load is placed on this unused output (OUTb) which is recalled that it is connected to the auxiliary input AUXa of the other coupler of the same group of 2.

N-splitters / combiners may also be used when there are N secondary sections. The divider has one input and N outputs. The test signal transmitted on each output is a fraction 1 / N of the test signal received at the input (with losses); the signal transmission between outputs is zero; the measurement signal transmission from an output to the input is done without attenuation (with losses).

In the general case, the network of electrical cables is intended to transmit electrical signals or energy, these signals being in a lower frequency band than the band of the test signals. But couplers or dividers are generally designed to operate in a frequency band that is limited and does not include these low frequencies or the continuous. In this case, if it is desired that the coupling device can remain present at the junction during the normal operation of the cable network, it is necessary to provide means for ensuring the passage of useful signals, low frequency or continuous, in the secondary sections. without these signals being stopped by the coupling device. Thus, it is not necessary to disconnect the coupling device for normal use of the network, and moreover one can even perform the test without interrupting the normal use of the cable network.

According to a first solution, one or more polarization tees are added to the CPL coupling device, the function of which is, in a sense, to separate the signal received on its input into a signal. continuous component (or low frequency) it directs to a first output and a high frequency component that directs it to a second output, and in the other direction to recombine the DC and high frequency components. Such bias tees are common microwave components and are commercially available.

The assembly is then that of FIG. 8 (for a junction with two secondary sections), in the case where the coupling device is a bidirectional coupler such as that of FIG. 4. A first upstream bias tee is inserted upstream. of the coupling device, between the end S1 of the main section and the inputs of the coupling device. Two other bias tees are placed downstream, a respective tee placed at the output of each of the outputs of the coupler, between this output and a respective input, E2 or E3, secondary sections. The tees are designated by the letter T. The upstream polarization tee has an input connected to the end S1 of the main section and can receive both low frequency or continuous (DC) signals and high frequency test signals. (RF). It directs the first to a respective input of each of two other downstream polarization tees; and directs the high frequency test signals to the input of the bidirectional coupler. In reverse, it recombines the high-frequency and low-frequency or continuous signals.

The downstream bias tees are mounted in the opposite direction of the upstream tee with respect to the direction of propagation of the test signals: they recombine the signals present on their two inputs (test signals and useful low-frequency signals) to direct them towards the secondary sections. But in the sense of the return of the measurement signals from the secondary sections, they separate the DC or low-frequency component of the high frequency component.

Each downstream tee receives on its second input (the first receiving the continuous or low-frequency signals from the upstream tee) a respective output of the bidirectional coupler. It recombines in a respective secondary section the continuous or low-frequency signal that it receives from the upstream tee and the high frequency signals that it receives from an output of the bidirectional coupler. According to a second solution shown in FIG. 9, the high and low frequency components are separated by inserting two low-pass filters in parallel on the coupling device, one between the end S1 of the main section and the input E2 of the main section. secondary section, the other between the end S1 and the input E3 of the second secondary section. These filters do not pass high-frequency (RF) test signals and allow continuous or low-frequency (DC) signals to pass. The filters are designated FPB in Figure 9.

Another possibility of a device to be inserted at the junction between a main line section and two secondary sections is shown in FIG. 10; it is inspired by the same principles of isolation or attenuation between a forward path and a return path. The CPL coupling device of FIG. 10 comprises two circulators (CIRC1, CIRC2), each associated with an attenuator (ATT1, ATT2) and an insulator (ISOL1, ISOL2). The attenuator is placed upstream of the inlet of the respective circulator. The isolator is placed between the return output of each circulator and the input of the corresponding attenuator. Output S1 of the main line is connected to both attenuators at the same time. The main output of the circulator is connected to the input E2, E3 of a respective secondary line section. The circulators make it possible to direct the test signal towards a respective secondary section and to redirect the measurement signal after reflection towards the main line. The insulators make it possible to keep the direction of the measurement signals in the circulators for a better analysis of these received signals. Attenuators are used to weaken measurement signals from one secondary section to another.

Claims

1. A method of testing a cable network comprising at least one main cable section (T1) and at the end (S1) of this section a junction from which at least a first and a second secondary section (T2) , T3), the network being such that the junctions require, for the normal operation of the network, only connections by simple contacts between the conductors (in particular soldered contacts or crimped or pressed), the method comprising the injection at the input of the main section of test signals capable of propagating in the network and the collection on this input of measurement signals resulting from the reflection of the test signals on the discontinuities present in the network, characterized in that the location of the junction, to ensure the passage of test signals and measurement signals, a coupling device (CPL) having at least a first input (INa) and two outputs (OUTa, OUTb), this device ant adapted to ensure the passage of the test signals with a first attenuation from the first input (INa) to a first output (OUTa) and the passage of measurement signals with a second attenuation from the second output (OUTb) to the first input (INa), but not the passage of test or measurement signals between the two outputs, the first input of the coupling device being connected to the end (E1) of the main section and each output being connected to the input (E2 , E3) of a respective secondary section.

2. Method according to claim 1, characterized in that the coupling device is a power divider by two (DIV) having a single input (IN1) and having the same first attenuation of the input to each of the two outputs (OUT2 , OUT3) and having the same second attenuation from each output to the input.

3. Method according to claim 1, characterized in that the coupling device is a power divider by N (N> 2) having a single input and N outputs each connected to a respective secondary section of N, the divider having a same first attenuation of the input to each of the N outputs and having the same second attenuation from each output to the input.

4. Method according to claim 1, characterized in that the coupling device comprises at least one unidirectional coupler (CpA,

CpB) having an input, an output and an auxiliary terminal, this coupler having a first attenuation between the input and the output and a second attenuation, greater than the first, between the auxiliary terminal and the input, the coupler not permitting little or no passage of signals from the output to the input, the input of the coupler being connected to the output of the main section, the output and the auxiliary terminal being respectively connected to the inputs of the secondary sections.

5. Method according to claim 4, characterized in that the coupling device comprises a second unidirectional coupler constituted as the first and having its input connected to the input of the first, the output of a unidirectional coupler being connected to the auxiliary terminal from each other and vice versa.

6. Method according to claim 5, characterized in that the coupling device comprises N groups of two unidirectional couplers, N> 1, for testing an array comprising N pairs of secondary sections connected to the main section.

7. Method according to claim 1, characterized in that the coupling device comprises a bidirectional coupler having a first input / output torque (INa, OUTa) and a second input / output torque (INb, OUTb) and being adapted to ensure the passage of test signals with a first attenuation from each input to the respective output of the same pair but not the passage of test signals from the input of one pair to the output of the other pair, and being capable of ensuring the passage of measurement signals with a second attenuation of the output of each pair towards the input of the other pair but not or almost no output to the input of the same pair or the passage of signals test or measurement between the two outputs, the inputs (INa, INb) being connected to the end (S1) of the main section and the outputs (OUTa, OUTb) being respectively connected to each secondary section.

8. Method according to claim 7, characterized in that the coupling device comprises N bidirectional couplers, N> 1, for testing an array comprising N pairs of secondary sections connected to the main section, the inputs of all the couplers being connected to the end of the main section.

9. Test method according to one of claims 1 to 8, characterized in that the test signals are microwave electrical signals, the network is a power supply network or a communication network with a much lower bandwidth. to that of the test signals, and means are provided for passing through the junction the normal operating voltages and currents of the network.

10. A test method according to claim 9, characterized in that a respective low-pass filter (FPB) is interposed between the end of the main section and the inlet of each of the secondary sections, the filter passing energy or the low frequency signals and not allowing the microwave test signals to pass.

11. A test method according to claim 9, characterized in that a bias tee (T) is interposed between the end of the main section and the input of the coupling device, and another bias tee is interposed between a respective output of the coupling device and a respective input of each of the secondary sections, the bias tee ensuring the separation between the high frequency components and the low-frequency components of the signals they receive on their input.