FR2679329A1 - Method and device for calculating a centre of gravity, the weights of which are pulsed electrical parameters - Google Patents

Abstract

The present invention relates to a device for calculating a centre of mass, the weights of which correspond to pulsed electrical signals appearing on input channels (L1 ... Li ... Ln). This device comprises a distributor (D1 ... Dn) connected to each input channel and supplying first and second identical outputs proportional to its input, first delay means (Ri) connected to the first outputs, second delay means (R'i) connected to the second outputs, first and second adders ( SIGMA , SIGMA 'R) respectively receiving the outputs of the first and second delay means, and means (5) for determining the time difference ( DELTA t) between the instants when the pulses obtained at the outputs of the adders reach the same amplitude (A).

Description

ce qui correspond au calcul du barycentre de quantités ai affectées des coefficients qi. METHOD AND DEVICE FOR CALCULATING A BARYCENTER
OF WHICH THE WEIGHTS ARE IMPULSE ELECTRICAL SIZES
The present invention relates to the processing of output signals from measuring devices requiring barycentric type calculations. In many measuring devices and sensors, we obtain n measurements, qi (i = l ... n) which must undergo a treatment of the type

which corresponds to the calculation of the barycenter of quantities ai assigned coefficients qi.

 The present invention applies in particular to detectors with localization of particles or of radiation.

For example, FIG. 1 represents a localization detector, such as a wire chamber or a microstrip detector comprising numerous electrodes L1, L2 ... Ln on which are likely to arrive impulse electrical charges whose maxima form a load profile of the type designated by curve 1 in FIG. 1. To determine the central position xg of profile 1, it is necessary to do the above operation, with

Detectors of the aforementioned localization type are used in particular for the detection of radiation or of particles in association with particle accelerators or sources of synchrotron radiation. In these applications, arrivals of radiation or particles can follow one another at very rapid rates and it is therefore appropriate that the processing time linked to the calculation of the barycenter is as short as possible,
The most direct method for carrying out this barycenter calculation consists of detecting the charge on each of the electrodes L1 to Lnt, storing the maximum values, carrying out a digitization and then carrying out the necessary calculations in a computer. However, the implementation of this method has two significant drawbacks. On the one hand, it requires numerous sorted electronic circuits for detection and amplification which are difficult to adjust and costly; on the other hand, the data processing time is relatively long.

 This process can be simplified by injecting the electrodes in a resistive or capacitive line to divide the charge on an impedance bridge and by reading the signals at certain locations on the bridge by preamplifiers.

This decreases the number of channels to be processed but the information remains to be processed, which requires coding of the charges on a large number of channels if 1 tone wants to maintain good accuracy during division and if one does not want to deteriorate differential linearity. These operations conventionally occupy several hundred nanoseconds.

Methods have been proposed in which calculations are performed in the time domain rather than in the space domain. For this, each of the channels L1, L2 ... Ln is connected to a tap on a delay line 2, for example of the LC type or of the coaxial type. The delays r1 ... rn ~ l between the taps (generally equal to each other) correspond in the case of the previous example to the intervals between the abscissae xl ... Xn. At outputs 3 and 4 of the delay line, therefore, pulses appear as a function of time. The phase shift At between these pulses corresponds to the value sought

 This system brings a great simplification compared to the digital computation system of T the prior art since it suffices to measure a time difference and then possibly to digitize for storage the value of this difference without it being necessary to carry out additional calculations. However, this delay line system has the disadvantage of being limited in speed, since it is not known how to make good quality delay lines in integral and differential linearity with very small delays between taps.

Currently, good quality delay line systems allow at best detections of T pulses separated by durations of T of the order of a few hundred nanoseconds for a hundred takes.

 An object of the present invention is to provide a device making it possible to increase this speed at least by a factor of 10.

 Another object of the present invention is to provide such a device making it possible to carry out in a simple and rapid manner the processing of information originating from taps of an impedance bridge.

 For this, the present invention provides an apparatus generally allowing the calculation of various barycentric functions, in particular when the weighting coefficients are weighted slopes of signals and, in certain cases, amplitudes of signals.

 To achieve these objects, the present invention provides a device for calculating a barycenter whose weights correspond to values of impulse electrical signals appearing on input channels, this device comprising a distributor connected to each input channel and providing first and second identical outputs proportional to its input, first delay means connected to the first outputs, second delay means connected to the second outputs, first and second summers respectively receiving the outputs of the first and second delay means, and means for determining the time difference between the instant s when the pulses obtained at the outputs of the summers reach the same amplitude.

 According to an embodiment of the present invention, constant fraction discriminators make it possible to determine the above-mentioned instants.

 According to an embodiment of the present invention, each of the distributors assigns the input signals by a multiplicative coefficient.

 According to an embodiment of the present invention, the differences between the delays of the first and second delay means of successive distributors are in arithmetic progression.

 The present invention also provides a method for calculating the barycenter of quantities affected by weight corresponding to slopes of electrical pulses appearing on input terminals, this method consisting in connecting each line to a first and to a second delay means, summing the respective outputs of the first and second delay means, and measuring the time difference between the instants when the sum signals reach the same predetermined amplitude.

These objects, characteristics and advantages as well as others of the present invention will be explained in detail in the following description of particular embodiments made in relation to the attached figures, among which
Figures 1 and 2 described above were intended to show the state of the art and the problem to be solved by the present invention
FIG. 3 represents a device according to the present invention
FIG. 4 represents the shape of the growth of a signal pulse on an analyzed channel
FIG. 5 represents the appearance of signals obtained by the device according to the present invention; and
Figures 6 and 7 show examples of systems to which T applies the present invention.

 As shown in Figure 3, the device according to the present invention receives as input channel signals L1, L2 ... Li ... Ln. These signals are supplied to analog distributors D1, D2 ... Di ... Dn which provide on their two outputs equal signals proportional to the signals on the channels L1 ... Ln. These distributors can each have a multiplying coefficient ki. Two outputs of each of these distributors are connected to two delay means, respectively delay means R1, R2 ... Ri ... Rn to the first outputs and delay means R'1, R'2 ... .R '1 ... .R'n on second outings.

The outputs of the first delay means are summed in a first analog summer #R and the outputs of the second delay means are summed in a second analog summator L'R TR
Then, as in the device of the prior art, and as will be explained in more detail below, a circuit 5 is provided for supplying a digital value At corresponding to the time difference between the output signals of the summers ER and 2 'TR, this difference being taken for a determined level of amplitude of these signals. We call U (t) the output of the summator ER and W (t) the output of the summator 'R
If on each of the channels L1 ... Ln appear pulses
Vi (t), and if the distributors and summers are linear

 To simplify the description of the operation of the device according to the present invention, it has first been assumed that the pulses Vi (t) applied to each of the channels L1 ... Ln have the form illustrated in FIG. 4, namely that each pulse increases linearly between an instant you and an instant you + Ti to reach a value Mi. Then, the impulses can be extended by increasing or decreasing in any way or by remaining constant.

 We call t0 the maximum of the values tOi, Tmin the minimum of the values tOi + Ti and we assume that Tmin-tO is greater than 0. In addition, we call eM the maximum value of thou + Ri, OTM the maximum value of you + R'i, 0m the minimum value of you + Ti + Ri, and O'm the minimum value of t0i + Ti + R'i. It is assumed in what follows that em is greater than eM and that O'm is greater than 5'M. This means that the differences between the instants of appearance of the different pulses are small compared to the rise times of these pulses. Similarly, the differences between the delays Ri-Rj and R 'iR' j are less than the rise time Ti whatever i and j or, in other words, the delays Ri and R'i placed on the various channels are less than the shortest rise time.

et pour t compris entre O'M et i'm

Then, we can show that the pulse U (t) between the instants eM and em will have the same slope as the pulse W (t) between the instants O'M and O'm, that is to say that the
m, i.e. the curves
W (t) and U (t) have parallel linear portions offset by #t as shown in Figure 5. Under these conditions, we can write, for t between 6M and em

and for t between O'M and i'm

d'où une quantité indépendante des toi

If we consider the respective instants of the signals
W (t) and U (t) where these have the same amplitude A, in the zone where these s signals are parallel, this amplitude corresponding to an instant t1 for the curve U (t) and to an instant t2 for the curve W (t), we can write by equating equations (3) and (4):

hence a quantity independent of you

Instants tl and t2 can be detected from threshold electronic devices (see for example IEEE
Transactions on Nuclear Science, vol. NS-32, 1985, P. 1242 to 1249) and their difference by known time coding devices (see for example IEEE Transactions on Nuclear Science, vol. NS-20, 1973, P. 36 to 51).

 If we examine formula (6), we see that the value At corresponds to the barycenter of the quantities Ri-R'i weighted by weights kiMi / Ti. If ki is constant, that is to say that the distributors Di all have the same gain, At is an affected barycenter of weight Mi / Ti, that is to say a barycenter weighted by the slopes of the pulses Vi (t) arriving on various routes.

 It is a first characteristic of the device according to the present invention to provide a means of measuring a barycenter weighted by slopes of pulses.

 Such a center of gravity measurement weighted by pulse slopes can be used in various applications. For example, in an RC system as shown diagrammatically in FIG. 6, consisting of a line which can be considered as a succession of RC cells, if a pulse arrives at a point i of the line, we obtain at the two ends of the line pulses of substantially equal amplitudes but of different slopes and the position of the injection point of the pulse corresponds to the barycenter of the delays weighted by these slopes.

Thus, the system according to the present invention can be connected to the two outputs S1 and S2 of the line RC to determine the position of point i.

 It is another feature of the present invention to provide a means for measuring a barycenter weighted by pulse amplitudes.

 For example, as shown in FIG. 7, the inputs L1 ... Li ... Ln of a device according to the invention can, in a particular application, correspond to the outputs of an impedance bridge (resistances or capacities ). This makes it possible to locate, by division of charge or of electric current, the center of the distribution of charges on electrodes connected to the impedance bridge.

 Returning to the example of FIG. 1, in the particular case where pulses arriving on the various channels LI to Ln have the same rise time, which is generally the case when these pulses result from the same radiation, the Ti value is constant. Then, At is the barycenter of the delays (or positions) affected by weight corresponding to the amplitudes Mi. Thus, the system according to the present invention can advantageously be used in place of the system of T the prior art described in relation to the figure 2. Since the delay circuits Ri and R'i are discrete elementary circuits, they can be adjusted to low values with much more precision than taps on a delay line and the determination of the barycenter can be carried out much faster and with better differential linearity.

A means 5 for determining the deviation At is known in the prior art and may include two discriminators with constant fraction for the detection of times t1 and t2 and the measurement of the difference At is provided by time counters or by a conversion time / amplitude followed by digital coding of the amplitude.

All these techniques are very precise and very stable and the precision on the barycenter will depend on the signal to noise ratio of the pulses and on the ratio between the maximum delay between times t1 and t2 where the signals U (t) and
W (t) reach the same amplitude A and the mean rise time of the pulses.

 Above, the present invention has been described in the case where the pulses had, at least on part of their rise, the linear shape illustrated in FIG. 4. In fact, the present invention T simply extends to pulses whose rise is not linear from the moment when part of this climb can be approached by a straight line as obtained from a first order Taylor expansion.

 In practice, it has been determined that with pulses having rise times of 30 to 50 ns, it is possible to carry out the determination of At in 100 to 150 nanoseconds with delays of only a few nanoseconds between the various channels. What limits speed is no longer taking information but coding time.

If we want to obtain a system having the same function as that illustrated in the prior art in relation to FIG. 2, we will choose the values of Ri and R'i so that Ri-R'i = (i + k) R k being a negative or zero positive integer and i varying from 1 to n. However, it should be noted that the present invention makes it possible to perform more complex barycentric calculations inaccessible with normal delay lines, in which it is possible in particular to provide zero delays between certain takes in certain directions only. For example, by choosing R1 = 0, R'l = 0, R2 = 0, R'2 = R, R3 = 0, R'3 = 2R, R4 = 0,
R'4 = 3R, there will be delays between successive points of 0, -1, -2, -3. Thus, #t will always be positive, t1 always preceding t2, which is not the case with a delay line. This simplifies the system for determining the time difference
Likewise by varying the quantities ki, the weights can be modulated at will in the calculation of the barycenter.

 Of course, the present invention is susceptible of numerous variants of application which will appear to those skilled in the art.

Claims (6)

 1. Device for calculating a barycenter whose weights correspond to values of impulse electrical signals appearing on input channels ...... L1 ... .Ln), characterized in that it comprises  a distributor (Dl ... Dn) connected to each input channel and providing first and second identical outputs proportional to its input,  first delay means (Ri) connected to the first outputs,  second delay means (R'i) connected to the second outputs,  first and second summers (ER, 'R) respectively receiving the outputs of the first and second delay means, and  means (5) for determining the time difference (At) between the instants when the pulses obtained at the outputs of the summers reach the same amplitude (A).  2. Device according to claim 1, characterized in that it comprises discriminators with a constant fraction for determining said instants.  3. Device according to claim 1, characterized in that each of the distributors affects the input signals with a multiplicative coefficient (ki).  4. Device according to claim 1, characterized in that the differences between the delays of the first and second delay means of successive distributors (Ri, R'i) are in arithmetic progression.  5. Method for calculating the barycenter of quantities affected by weight corresponding to slopes of electrical pulses appearing on input terminals, characterized in that it consists in connecting each line to a first and to a second delay means, to summing the respective outputs of the first and second delay means, and measuring the temporal difference between the instants when the summed signals reach the same predetermined amplitude.  6. A barycenter calculation method according to claim 5, characterized in that said amplitude corresponds to a predetermined fraction of the maximum amplitude of the linear portion of the summed signals.