This is a U.S. National Phase Application under 35 U.S.C. § 371 of International Patent Application No. PCT/EP2020/088005, filed Dec. 29, 2020, and claims priority to European Patent Application No. 20382162.4, filed Mar. 6, 2020, and European Patent Application No. 20150184.8, filed Jan. 3, 2020, which are incorporated by reference in their entireties. The International Application was published on Jul. 8, 2021, as International Publication No. WO 2021/136792 A1.
The present invention relates, in a first aspect, to a measuring apparatus made to measure and obtain the electromagnetic signals or noise or electromagnetic interference (EMI) (from now on, conducted emissions) generated by an equipment under test (EUT) and the Z or Y or S parameters of the EUT or any other meaningful set of parameters that can be computed from the aforementioned ones or from voltages and currents (from now on, characterization parameters).
The measuring apparatus can be used to design a filter to attenuate the conducted emissions generated by the EUT or a matching network.
A second aspect of the present invention relates to a measuring method adapted to perform methodological steps with the apparatus of the first aspect of the invention.
The measurement of the conducted emissions generated by an EUT and of the characterization parameters of the EUT is performed in the prior art independently, by means of two separated apparatuses, for example by means of a spectrum analyser, for the interference signals, and an impedance analyser or a network analyser, for the characterization parameters of the EUT.
The use of the above mentioned two separate apparatuses is accepted in the prior art as necessary to perform measurements, but it exhibits several problems. First, if the EUT is generating conducted emissions at the measuring ports, its characterization by means of the characterization parameters of the EUT may be very poor (the conducted emissions interfere with the measurements).
Besides, the use of those separate measurement apparatuses provides measurements at different times and under different operating conditions, so that one cannot know which was the conducted emissions generated by the EUT exactly at the same time and under the same operating conditions as the measurements of the characterization parameters were obtained, and vice versa.
This is a great disadvantage, which makes it very difficult and prone to errors for the skilled person, who needs the two types of measurement information (for example, for modelling the EUT). In consequence, these two kinds of information are linked by means of estimations, which is always a source of errors, impeding to meet the strict low error tolerances demanded by some regulations.
Some prior art documents are identified and briefly described below, since they represent some relevant examples of instruments and/or methodologies that can be used to measure the excitation-dependent parameters of an EUT (for instance, the impedance, S parameters, conversion loss of a mixer, conversion efficiency, etc.)
US2002053899A1 describes a test set to measure the S parameters of EUTs with more than two ports (multiport) by means of switching matrices placed between the signal generator, the receivers and the EUT. The instrument includes the signal processing to transform the conventional S parameters to mixed-mode S parameters in one side, and to a time-domain representation on the other side (equivalent to a reflectometer measurement).
GB2466028A describes a high frequency non-linear measurement system for analysing the behaviour of high power and high frequency amplifiers. The measurement system includes multiplexers and demultiplexers formed by filters, directional couplers and splitters, improving previous measurement systems that only used multiplexers and demultiplexers exhibiting poor transmission and reflection characteristics at certain frequencies.
Paper “Measurement of Passive R, L, and C Components Under Nonsinusoidal Conditions: The Solution of Some Case Studies”, Luigi Ferrigno et al., IEEE Transactions on Instrumentation and Measurement, vol. 57, no. 11, November 2008, pp 2513-2521, describes a methodology to find the values of passive R, L and C components under conditions where the measurement signals are non-sinusoidal, based on linear system identification and modal parameter estimation techniques.
WO2013127911A1 describes a method for characterizing, at a given frequency, reflected waves of a frequency translating device (such as a mixer, in phase/quadrature modulators and demodulators, etc.) having at least two ports. The proposed method determines the frequency conversion factor of the EUT (with an integrated LO) by measuring the reflection factor using a one port network analyser, while applying known impedances at the other port of the EUT and a filter for image rejection. The method needs to assume reciprocity between up conversion and down conversion.
U.S. Pat. No. 6,356,852B1 describes an interface that allows to connect a two-port network analyser to a multiport EUT (that is, with more than two test ports). The interface device has at least two levels of switches, and is adapted to be coupled between the test ports of the EUT and a two-port network analyser.
However, while the instruments or methods disclosed by the prior art documents cited above have only been designed to measure some specific parameters of an EUT (such as the S-parameters or the frequency conversion), none of them have been built to measure both the excitation-independent and excitation-dependent parameters of an EUT at all, much less in a coherent and integrated way.
It is, therefore, necessary to provide an alternative to the state of the art which covers the gaps found therein, by providing a measuring apparatus which allows to perform measurements of both the conducted emissions generated by an EUT and the characterization parameters thereof.
To that end, the present invention relates to a measuring apparatus, comprising:
As stated above, the EUT may have less than N ports (in this case, some of the ports of the EUT would remain unused), or more than N ports, in this case some of the ports of the EUT would remain unmeasured. With this understood, in the present section, the EUT will be supposed a M-port device.
For a preferred embodiment of the measuring apparatus of the first aspect of the present invention, the arbitrary waveform generator is configured and arranged to generate (simultaneously or sequentially) said combination of N test signals from discrete sequences of length L with auto-correlation
where x* represents the complex conjugate and [l+n]L represents a circular shift, with a modulus outside the origin lower or equal than 1/√{square root over (L)} for n≠0, and modulus of the cross-correlation
with a modulus lower or equal than 1/√{square root over (L)}.
According to some embodiments, the measuring unit has N 2N or 3N ports.
For an embodiment, the measuring apparatus of the first aspect of the present invention, the arbitrary waveform generator is configured and arranged to simultaneously generate said combination of N test signals and/or simultaneously inject the generated N test signals to the N ports of the coupling network, and wherein:
For an alternative embodiment, the arbitrary waveform generator is configured and arranged to inject the generated N test signals to the N ports of the coupling network, and wherein:
Depending on the embodiment, the aforementioned N test signals are tones or chirp signals or modulated signals or pulses or impulses or wideband signals covering a frequency range to be measured.
For a preferred implementation of the embodiment for which the N test signals are pulses, they form pseudonoise (PN) sequence signals.
According to a further embodiment, the processing unit comprises processing means to process the received measured electrical signals using correlation techniques with the injected Ntest signals, to separate data representative of the conducted emissions generated by the EUT from data representative of the characterization parameters of the EUT.
For an embodiment, the coupling network contains Line Impedance Stabilization Network (LISN) channels configured and arranged:
According to an embodiment, the processing unit is configured to compute a modal decomposition of data representative of the aforementioned measured electrical signals.
For an embodiment, the processing unit comprises the EMC (Electromagnetic compatibility) detectors (peak, quasi-peak and average detectors) applied directly on the modal decomposition of data representative of the aforementioned measured electrical signals.
For an embodiment, the signal generator is configured to generate and inject N test signals with a period smaller than the switching period of the EUT connected or to be connected thereto, to characterize the variations along time of conducted emissions generated by the EUT and characterization parameters of the EUT, whether because the signal generator is adapted to operate only with EUTs having a known switching period which is always greater than that provided by the signal generator, or, preferably, because the signal generator can be adapted, specifically the period of the test signals, to a plurality of different switching periods of different EUTs.
In this sense, this document discloses in a posterior section how the information required from the EUT is directly obtained from the measurements (b2M and b4M). For instance, if the measured EUT features a switching-mode power supply at its ports, the switching period can be easily extracted from a single measurement of the conducted emissions (the first harmonic in the spectrum of these emissions provides the switching speed). Therefore, the instrument does not need to have preliminary information about the EUT (although this case is also embraced by the present invention, for other embodiments), but to perform a measurement of the conducted emissions, detect the first harmonic of the emissions, and then inject PN sequences (or other kinds of excitations) suitable to measure the changing impedance of that particular switching-mode power supply. The same applies to other kind of switching devices such as AC-AC, AC-DC, DC-AC and DC-DC converters.
It should be emphasized that a measurement is a complex process wherein the instrument may have to interact several times with the EUT in order to fully characterize it. At each iteration the instrument may generate different kind of excitations (Vg) to find features of the EUT that permit a full characterization of SEUT and VN(see description of these parameters in a posterior section in this document), even in time-varying situations, as the described above.
According to an embodiment, the processing unit is configured and arranged to process the N test signals and the measured electrical signals, and also to design a filter to attenuate the conducted emissions generated by the EUT.
For a further embodiment, alternative or complementary to the above mentioned embodiment, the processing unit is configured and arranged to process the N test signals and the measured electrical signals, and also to design a matching network for the optimal transference of the conducted emissions generated by the EUT.
The present invention also relates, in a second aspect, to a measuring method, comprising:
Preferably, the method of the second aspect of the present invention comprises using the measuring apparatus of the first aspect of the invention to perform the method steps, wherein:
For another embodiment, the method of the second aspect of the present invention comprises using the measuring apparatus of the first aspect of the invention to perform the method steps, wherein:
According to an embodiment of the method of the second aspect of the present invention, the method comprises:
For an embodiment of the method of the second aspect of the present invention, the step of designing the optimal filter further comprises carrying out an optimization process in order to reduce the number filter components combinations to be virtually connected to and simulated with the built circuital and modal models.
According to an implementation of that embodiment, the optimization algorithm comprises at least one of the following algorithms, or a combination thereof: genetic algorithm, gradient algorithm, conjugated gradient algorithm, and Broyden-Fletcher-Goldfarb-Shannon algorithm.
In the following some preferred embodiments of the invention will be described with reference to the enclosed figures. They are provided only for illustration purposes without however limiting the scope of the invention.
In the present section some working embodiments of the measuring apparatus of the first aspect of the present invention and of the different signals intervening in the operation thereof, will be described with reference to the Figures.
The description below refers to embodiments of the apparatus/method of the present invention to perform sequential measurements (Approach A) of conducted emissions and impedance and also simultaneous measurements (Approach B) thereof.
Measurement Steps to Perform a Sequential or Simultaneous Measurement of Conducted Emissions and Impedance:
The embodiments described above for the measuring apparatus of the first aspect of the present invention, allow the computation of the conducted emissions and characterization parameters of an EUT. These can be combined to obtain a generic equivalent Thevenin/Norton model of the EUT. By a generic Thevenin/Norton equivalent it is understood in this document any characterization of an EUT (
The block diagram of the instrument that can perform these measurements is shown in
The Arbitrary Waveform Generator and the Measuring Unit can work in a base band configuration or include frequency mixers, upconverters, downconverters, etc. The Measuring Unit contains k×N signal measurement devices, which can be actual or equivalent (a multiplexing schema could be used if needed).
The Processing Unit can be embedded into the physical instrument or be hosted in an external PC or the Cloud.
The Coupling Networks can be made in a variety of configurations, none of which refers to a switching matrix. For instance, using power dividers and directional couplers, impedance bridges, circulators, voltage or current probes, etc. This definition means that in such coupling networks all ports are always interconnected (contrary to what can happen in a switching matrix with more inputs than outputs or vice versa, where only those ports placed at the switching position are interconnected).
In order to demonstrate the feasibility of the instrument, it can be modelled as seen in
The N signal generators of the block diagram of
The following analysis has been performed using a very general definition on normalized waves (and, therefore, of S parameters), as seen in
Let it be the following column vectors,
with=1, . . . , 4. If the S parameter matrix of the EUT and Coupling Networks are
with i=1, . . . , N, let it be the diagonal matrices
with i=1, . . . , 4, j=1, . . . , 4. Finally, let it be the diagonal matrices
with j=1, . . . , 4.
Then,
b=SEUTa
a1M=K1MVg
a3M=K3MVn+b
b3M=K3MVn+a
b1M=S11Ma1M+S13Ma3M
b3M=S31Ma1M+S33Ma3M
b2M=S21Ma1M+S23Ma3M
b4M=S41Ma1M+S43Ma3M
From these equations, it follows that
b3M=(IN−S33MSEUT)−1S31Ma1M+(IN−S33MSEUT)−1S33M(IN−SEUT)K3MVna3M=SEUT(IN−S33MSEUT)−1S31Ma1M+(IN+SEUT(IN−S33MSEUT)−1S33M)(IN−SEUT)K3MVn.
From these, all other waves (and therefore, the voltages and currents) at all the ports of the circuit of
From these equations, several measurement strategies (time-domain, frequency-domain, mixed-domain, or spread-spectrum) can be envisaged.
For instance, two very basic approaches, which can be enriched at several stages, would be the ones described below.
Approach A:
Suppose an EUT emitting stationary interference. First, the effect of Vn is measured when Vg=0 (a1M=0), yielding
b2M0=S23Ma3M=S23M(IN+SEUT(IN−S33MSEUT)−1S33M)(IN−SEUT)K3MVn
b4M0=S43Ma3M=S43M(IN+SEUT(IN−S33MSEUT)−1S33M)(IN−SEUT)K3MVn.
If then adequately timed (synchronized with the interference or with the 50-Hz mains signal, . . . ) measurements are performed with Vg≠0 (a1M≠0), the following waves are measured,
b2M=S21Ma1MS23Ma3M=(S21M+S23MSEUT(IN−S33MSEUT)−1S31M)a1M+b2M0
b4M=S41Ma1MS43Ma3M=(S41M+S43MSEUT(IN−S33MSEUT)−1S31M)a1M+b4M0
If N linearly independent (at all frequencies) (column) vectors a1M, k=1, . . . , N are generated, and its responses measured, the following excitation and response matrices (made up of column vectors) can be constructed,
A=[a1M,1 . . . a1M,N]
B2=[b2M,1−b2M0−b2M,N . . . b2M0]
B4=[b4M,1−b4M0−b4M,N . . . b4M0],
with
B2=(S21m+S23MSEUT(IN+S33MSEUT)−1S31M)A
B4=(S41M+S43MSEUT(IN+S33MSEUT)−1S31M)A.
Since A is invertible, SEDT can be computed from either expression. For instance,
SEUT=(IN+S23M−1(B2A−1−S21M)S31M−1S33M)−1(S23M−1(B2A−1−S21M)S31M−1) Equation 1
Once SEUT is known, Vn can be readily computed.
Example A: Consider the case of a two-port EUT modelled using the characterization of
The Coupling Networks considered for the instrument feature each a CISPR-16 50Ω//50 μH LISN channel, a limiter attenuator and a directional coupler.
The Measurement steps for this case are:
After these five steps, all the information to construct the Thevenin equivalent model of the EUT has been obtained.
Approach B:
Now, if the excitation is a spread-spectrum one, with the signal generators generating highly-uncorrelated sequences, all the above measurements could be performed simultaneously. The system would be simultaneously excited by N pseudonoise (PN) sequences and the response of the EUT recorded. Therefore, by performing N·N correlations of all responses by all PN sequences, response column vectors as those described above would be recovered, one for each exciting sequence (although this kind of measurements, and the associated correlations, are time-domain, as before they are characterized by their frequency-domain counterparts for analysis purposes):
b2M=(S21M+S23MSEUT(IN−S33MSEUT)−1S31M)a1M+b2M0
b4M=(S41M+S43MSEUT(IN−S33MSEUT)−1S31M)a1M+b4M0
b2M0=S23Ma3M=S23M(IN−SEUT(IN−S33MSEUT)−1S33M)IN−SEUT)K3MVn
b4M0=S43Ma3M=S43M(IN−SEUT(IN−S33MSEUT)−1S33M)(IN−SEUT)K3MVn.
In this case, due to the spreading effect of the correlation to signals other than the exciting PN sequence, the terms b2M0 and b4M0 would have a low value and could generally be ignored.
Then, the matrices
A=[a1M,1 . . . a1M,N]
B2=[b2M,1−b2M0 . . . b2M,N−b2M0]≈[b2M,1 . . . b2M,N]
B4=[b4M,1−b4M0 . . . b4M,N−b4M0]≈[b4M,1 . . . b4M,N],
could be constructed (the A matrix is also constructed by appropriately recording the N N correlation of the input PN sequences, and is, basically, a diagonal matrix at each measurement frequency), and the S-parameters matrix of the EUT could be obtained by
SEUT=(IN+S23M−1(B2A−1−S21M)S31M−1S33M)−1(S23M−1(B2A−1−S21M)S31M−1)
Once SEUT is known, the interference vectors b2M0 and b4M0 can be recovered from
b2M0−b2m−(S21M+S23MSEUT(IN−S33MSEUT)−1S31M)a1M
b4M0−b4m−(S41m+S43MSEUT(IN−S33MSEUT)−1S31M)a1M,
this time using the PN excitations and their responses directly to perform the computations. From b2M0 and b4M0 the interference vector
VN=(K3M(IN−SEUT)·(IN+SEUT−1(IN−S33MSEUT)S33M)S23M)−1b2M0,
can be obtained.
This schema of measurement has been presented only as an example to demonstrate that simultaneous measurements of all the parameters of a (generalized) Thevenin equivalent can be performed. As in the case of the more conventional measurement schemas described above, other measurement steps could be performed to arrive at the same result. For instance, the interference levels might be recovered first, and then the S-parameters of the circuit, or the generators could generate a superposition of PN sequences to achieve code-diversity in the measurements, or the measurement of interferences and S-parameters could be performed sequentially, among others. As before, this basic measurement schema can be enriched with algorithms and techniques which improve the numerical accuracy of the results (interpolations, multiple measurements, . . . ).
Example B: Consider the case of an EUT modelled using the characterization of
Again, all Coupling Networks considered for the instrument feature a CISPR-16 50Ω//50 μH LISN channel, an attenuator (transient limiter) and a directional coupler.
The Measurement steps for this case are:
The two approaches described above are only presented as non-limiting examples of possible measurement strategies. The present invention embraces at least any measurement strategy including the generation and injection of the N test signals described in a previous section of the present document, at least those with the auto-correlation RXX and cross-correlation RXY described above.
Considering the definition given in the previous section of this document for the term Coupling Network, and taking into account the same port numeration shown in
Specifically,
The measuring apparatus of the first aspect of the present invention is more complex and complete than those known in the prior art, with a performance not available by any of them. It not only adds the possibility to simultaneously (or sequentially) measure the Z or Y or S parameters or any other meaningful set of parameters that can be computed from the aforementioned ones or from voltages and currents, and the electromagnetic signals or noise or electromagnetic interference generated by an EUT (or what is the same, its conducted emissions), but it also builds, for some embodiments, the Thevenin or Norton equivalent model and, as a last resort, finds the optimal power-line filter to mitigate the conducted emissions. This apparatus aims to accelerate the design and implementation of electronic EUTs, decreasing their design cost, optimizing its implementation and accelerating their time-to-market.
A person skilled in the art could introduce changes and modifications in the embodiments described without departing from the scope of the invention as it is defined in the attached claims. For example, substituting the above described LISNs internal to the Coupling Networks by one or more LISNs external thereto.
Number | Date | Country | Kind |
---|---|---|---|
20150184 | Jan 2020 | EP | regional |
20382162 | Mar 2020 | EP | regional |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2020/088005 | 12/29/2020 | WO |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2021/136792 | 7/8/2021 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
6356852 | Ke | Mar 2002 | B1 |
20020053899 | Adamian et al. | May 2002 | A1 |
20060017428 | Lin | Jan 2006 | A1 |
20060043979 | Wu | Mar 2006 | A1 |
20140361787 | Paoletti | Dec 2014 | A1 |
20150097575 | Hiraga | Apr 2015 | A1 |
20190285681 | Yin | Sep 2019 | A1 |
20210325438 | Nakamura | Oct 2021 | A1 |
Number | Date | Country |
---|---|---|
2466028 | Jun 2010 | GB |
2013127911 | Sep 2013 | WO |
Entry |
---|
Ferrigno, Luigi, Marco Laracca, and Antonio Pietrosanto. “Measurement of Passive R, L, and C Components Under Nonsinusoidal Conditions: The Solution of Some Case Studies.” IEEE Transactions on Instrumentation and Measurement 57.11 (2008): 2513-2521. |
Number | Date | Country | |
---|---|---|---|
20230027767 A1 | Jan 2023 | US |