EP0165144A1 - Electronic high-resolution chronometric system - Google Patents

Abstract

System to increase the accuracy of the measurement by correcting the non-linearity of the vernier ramp and buffer expansion circuits (3-4) used for fine measurement. A main counter (2) counts roughly the whole number of clock periods (SH, 1) and the verniers (3-4) the remaining fine measurements between a start time (S1) and a stop time (S2). Compensation is obtained by a programmable delay generator (6) controlled by a processor circuit (5) to produce locally (S10, S20) during a calibration phase at least one series of measurements reproducing start and end times. stop with the same constant relative time phase but by varying the starting phase with respect to the clock to cover the range of variation and allow the calculation of the average deviations of non-linearity of each ramp.

Description

The present invention relates to an electronic chronometry system including the time measurement method and the corresponding chronometer apparatus. The invention is addressed, in this technical field, to measurement systems having a high resolution, better than 100 picoseconds.

Electronic chronometers for non-repetitive phenomena, which measure the time interval between a start pulse and a stop pulse, very often proceed by counting the periods of a clock with a well-known frequency. In general, this time base circuit is formed using a quartz oscillator of high stability compensated in temperature. The time T to be measured is then equal to N. T to within + T / 2, _T being the clock period, N being the number present in the counter which is started by the start pulse and stopped by the pulse stop.

When a measurement resolution of the order of a few hundred picoseconds or less is desired, the temporal resolution of the electronic counters is no longer sufficient, and the technique of time-amplitude conversion is generally used. The starting pulse causes the start of a sawtooth or ramp which is expressed by a voltage of the form V = kT where k is a constant. The stop pulse causes this linear variation to block.

The quantification of time can be done in several possible ways. One of the most used is the multiplication of time t by a factor K, the time Kt being measured by the method already mentioned of clock counting.

To obtain this multiplication coefficient, the ramp is made by charging a capacitor C with a constant current 1 (V = It / C, voltage across the capacitors). This is then discharged by an equally constant current with a well-defined value i given by i = I / (K-1), which gives an overall duration Kt for the charge plus the discharge. An expansion by K of the charging time for the measurement is thus produced, the resolution then being equal to τ / K.

• L'impulsion de départ provoque le démarrage d'une tension en forme de rampe V (t) qui est arrêtée, au bout d'un temps TI, par la première impulsion d'horloge qui suit.

It is certain that the relative precision of the stopwatch with ramp, for important times, is lower than that of stopwatches with counting. So, when we have to measure long times with a quantification of the order of a few hundred picoseconds or less, we proceed by combining a counting of clock periods, called main counting, and verniers with ramp. This technique is described in particular in the article by Ronald NUTT entitled "Digital Time Intervalometer", published in The Review of Scientific Instruments volume 39, Number 9, of September 1968, pages 1342-1345. The process which is briefly recalled is as follows:

• The starting pulse causes the start of a ramp-shaped voltage V (t) which is stopped, after a time TI, by the first clock pulse which follows.

The phase between the starting pulse and the clock being a priori arbitrary, the time T1 will be between 0 and _T.

The voltage V (Tl) is then converted into the form of an expanded time as indicated above and is digitized (time expansion and analog-digital conversion).

_T étant la période d'horloge, Tl^* et T2 ^* étant alors les valeurs quantifiées de Tl et T2. Dans le cas du vernier à rampe avec expansion de temps et utilisation de l'horloge principale, le quantum est égal à _T/K.The stop pulse in turn causes a ramp to start; like the start pulse, it is stopped by the first clock pulse which follows after a time T2. The stop pulse only blocks the main counter after taking this same clock pulse into account. The counter indicates N. The measured time is then given by:

_T being the clock period, Tl ^* and T2 ^* then being the quantized values of Tl and T2. In the case of the ramp vernier with time expansion and use of the main clock, the quantum is equal to _{T /} K.

To avoid uncertainties resulting from accidental coincidences of the start and stop times with the clock, and to avoid the corresponding cases of ambiguity, the ramp stops are produced on the second pulse following (or on the second edge of given direction, said active front, by a clock signal formed by pulses of a certain width, the verniers work in a time domain between _T and 2τ, the principle of measurement remains unchanged.

When one wants to obtain very fine temporal resolutions, the character of linearity of the sawtooth, and of the associated digitization, takes a great importance and it is hardly possible to go below a hundred picoseconds .

An object of the invention is to overcome these limitations by using a method which makes it possible to compensate for the defects resulting from the non-linearity and, in so doing, to correct the measurement; the resolution reached is less than 50 picoseconds.

• - l'erreur de mesure due à la non-linéarité des rampes porte sur le terme (T1-T2) dans l'expression de T, corespondant à la mesure fine des verniers. Ce terme varie dans la page 0 à T d'une période d'horloge (au delà, il constitue un increment qui est pris en compte par le comptage gros) et représente la phase temporelle de la durée T par rapport à l'horloge. Pour une durée T donnée, la valeur de cette phase temporelle va varier en fonction de celle de l'instant de départ tl étant donné, a priori, que l'écart de rampe varie d'un point de fonctionnement à un autre ;
• - en conséquence, si l'on fait varier durant un cycle d'étalonnage l'instant de départ tl dans sa plage de variation égale à la période _T d'horloge, tout en gardant constant l'intervalle de temps entre cet instant et celui d'arrêt t2, on peut produire une série de mesures entachées chacune de l'erreur de mesure liée à la phase de départ tl (la précision de l'étalonnage est fonction du nombre de . valeurs sélectionnées dans la plage considérée). On peut ainsi dresser un tableau donnant l'erreur de mesure en fonction du paramètre Tl mesuré par la rampe. Pour s'affranchir des variations liées au deuxième paramètre T2 dans l'expression (Tl - T2), on effectuera de préférence plusieurs séries de mesures en faisant varier la durée T en sorte que sa phase parcourt également la plage 0 - T . En relevant ensuite les résultats de mesure en fonction de Tl et de (Tl -T2) on peut ainsi dresser un tableau à deux entrées et corriger toute mesure de T du terme d'erreur qui lui correspond.

According to an object of the invention, this object could be achieved by basing the method on the following remarks:

• - the measurement error due to the non-linearity of the ramps relates to the term (T1-T2) in the expression of T, corresponding to the fine measurement of the verniers. This term varies in page 0 to T of a clock period (beyond, it constitutes an increment which is taken into account by the large counting) and represents the time phase of the duration T with respect to the clock. For a given duration T, the value of this time phase will vary as a function of that of the starting instant tl given, a priori, that the ramp deviation varies from one operating point to another;
• - consequently, if the starting time tl is varied during a calibration cycle within its range of variation equal to the clock period _T , while keeping constant the time interval between this instant and that of stopping t2, one can produce a series of measurements each marred by the measurement error linked to the starting phase tl (the precision of the calibration is a function of the number of values selected in the range considered). It is thus possible to draw up a table giving the measurement error as a function of the parameter Tl measured by the ramp. To overcome variations linked to the second parameter T2 in the expression (Tl - T2), several series of measurements will preferably be carried out by varying the duration T so that its phase also covers the range 0 - T. By then reading the measurement results as a function of Tl and of (Tl -T2) we can thus draw up a table with two inputs and correct any measurement of T by the error term which corresponds to it.

In accordance with the present invention, it is proposed to produce an electronic timekeeping system using, for measuring a duration T between a starting instant tl and a stopping instant t2, fine counting means of the vernier ramp type with time expansion to measure the duration Tl between the instant tl and a subsequent edge of a clock signal and the duration T2 between the instant t2 and a subsequent edge of the clock, and large counting means for counting the number N of clock periods of duration _T between said edges. The system is characterized in that it further comprises means for compensating for errors of non-linearity of the ramp in order to determine, in magnitude and in sign, whatever the duration T to be measured, the corrective term to be applied for obtaining the corrected measurement, said corrective term being determined during a calibration cycle as a function of the parameters Tl and (Tl - T2) measured.

• - la figure 1, un diagramme général d'un système de chronométrie électronique conforme à l'invention ;
• - la figure 2, des formes d'ondes relatives au fonctionnement du système selon la figure 1 ;
• - les figures 3 à 8 des courbes de variation illustrant le procédé mis en oeuvre pour compenser les erreurs de mesure résultant de la non linéarité des circuits verniers à rampe ;
• - la figure 9, les relevés d'talonnage effectués conformément à l'invention pour déterminer un tableau des valeurs de compensation ; et
• - la figure 10 , un schéma d'un exemple de réalisation d'un système de chronométrie selon l'invention.

The features and advantages of the present invention will appear in the description which follows; examples of embodiments and applications are given using the appended figures which represent:

• - Figure 1, a general diagram of an electronic chronometry system according to the invention;
• - Figure 2, waveforms relating to the operation of the system according to Figure 1;
• - Figures 3 to 8 of the variation curves illustrating the process used to compensate for measurement errors resulting from the non-linearity of the vernier ramp circuits;
• - Figure 9, the calibration readings made in accordance with the invention to determine a table of compensation values; and
• - Figure 10, a diagram of an exemplary embodiment of a chronometry system according to the invention.

The main means constituting the system appear on the functional diagram of FIG. 1. They already include, according to the aforementioned prior art, a time base circuit called clock 1 to produce a clock signal SH, a main counter circuit 2 for perform the rough measurement, and ramp circuits 3 and 4 to perform the fine measurement.

FIG. 2 shows the corresponding essential signals: a clock signal SH of a determined stable period _T , the pulses SI and S2 which represent the start and stop times of the duration T to be measured, and the ramps SR1 and SR2 of duration Tl and T2 respectively. The duration T is given by N _T + (T1-T2), N being the large count and Tl and T2 the fine values obtained with temporal expansion. In the example shown, the falling edge of the clock SH is the active edge.

In accordance with the invention, the values N, T1 and T2 obtained are transmitted in digital form to a management and calculation processor 5 which may consist of a microprocessor with associated read and write memories and interface circuits. Circuit 5 calculates the time phase AT of duration T with respect to the clock signal, this phase being constituted by the value (Tl - T2) representing the fine measurement which exceeds the integer number N of clock periods.

The other circuits shown consist of a programmable delay generator 6 and a switching circuit 7 and are used to perform the calibration.

To this end, the processor circuit 5 controls the generator 6 to produce local signals S10 and 520, and the switch 7 to transmit these signals to the verniers 3 and 4 in place of the actual measurement signals S1 and S2. The programming of circuit 5 is done to control at least one series of measurements with a constant delay (t2-tl) between signals S10 and S20 and by varying the start time each time, i.e. the phase time of S10, with respect to the clock SH. The constant delay, with very high precision, is produced by the circuit 6 by means, for example, of an assembly of temperature compensated delay lines. A complete calibration cycle will comprise several series of measurements so as to also traverse the range of variation τ of the delay by modifying its value from one series of measurements to the next.

The measurement method implemented by this arrangement will now be demonstrated with the aid of FIGS. 3 to 9. In FIG. 3, we wanted to represent the ramp deviation from an ideal linear variation. At the instant tl + Tl where the charge in the start vernier circuit stops, here considered, we deviate from the value V _M which would be obtained for a linear response, by a positive quantity dV in the case considered VB = VM + dV. The value dV generally varies from one point to another, it can be positive or negative (for example in N), the variation is all the smaller the closer the points are. The response curve shown is given by way of example.

FIG. 4 is a diagram corresponding to the previous one but transposed to the time T measured by the vernier as a function of the real time T _R. The difference in charge dV variable as a function of the operating point and therefore of the parameter Tl which corresponds to the time phase of the instant t _l , is replaced there by the time difference on the measurement of Tl, (and of T2 for the other vernier). The weather measured is of the form Tm = T _R + dt or dt with a sign corresponding to that of dV and an amplitude proportional to that of dV. The shape of the variation in Tm is similar to that of the ramp. We note that _T being the range of variation of Tl (and T2), the difference dt vanishes for T _R = 0 and T _R = _T , extreme points or Tm = T _R. The curve Tm therefore repeats for time to be measured modulo T, that is to say of period τ.

During the calibration, a series of measurements is produced with (t ₂ -t ₁ ) equal to a constant value of R and by varying the phase t ₁ to traverse the range 0 -T in a uniform manner. For this, we can consider a determined and sufficient number of samples of values regularly distributed in the range 0 - T. Preferably, we consider the range 0 - τ cut into P slices, of width _T / P each and each comprising several samples as represented in FIG. 5 for a slice Trj of any order j. The number of samples per slice is equal, or substantially, and the average value Tmj of these samples is determined which will characterize this slice. This gives a distribution of P average values Tml to TmP for the P slices Trl to TrP as shown in FIG. 6, each of them distant from the theoretical linear response value by a quantity dt ₁ to dt corresponding _p equal to the average value of the dt deviations for the tranche in question.

est égale, ou sensiblement, à la valeur réelle Δ R (figure 8), compte-tenu que les écarts dmj sont faibles, de signes les uns positifs, les autres négatifs, et d'amplitude variable de sorte que leur valeur moyenne est, sinon nulle, du moins très faible. La différence entre cette valeur moyenne globale Δ Rm calculée et chaque valeur moyenne Δ Rmj de tranche représente ainsi l'écart moyen dmj de la tranche considérée.Average values Tm l to Tmp above are calculated for the measured Tl parameter. In parallel, for each value Tlm given by the start vernier, the stop vernier provides a measured T2 value, similarly called T2m. The fine count value (T _lm - T _2m ) therefore corresponds to the theoretical value (Tl - T2) affected by the measurement error dm. By setting T _1m = T ₁ + dt ₁ and T2m = T2 + dt2, the measurement error dm is equal to dtl - dt2. Considering that the series of calibration measurements is made at T = R constant, the relation: Tm = N _T + (T ₁ - T ₂ ) + dm = R + dm shows that: (Tl - T2) is constant and equal to RN _T = Δ R (N being the large count for the value R). Thus for each measurement, the processor 5 calculates the value (T _lm - T _2m ) = ΔRm = ΔR + dm and for each slice the average value ΔR mj which is equal to the constant value Δ R increased by the average dmj of the section considered (Figure 7). If we now consider the P average values ΔRm calculated for the P slices, we can consider that the average value Δ Rm of these defined by:

is equal, or substantially, to the real value Δ R (Figure 8), given that the deviations dmj are small, of signs some positive, others negative, and of variable amplitude so that their mean value is, if not zero, at least very weak. The difference between this global average value Δ Rm calculated and each average value Δ Rmj of slice thus represents the average difference dmj of the slice considered.

One thus obtains, by examining the results translated by FIGS. 6 and 8 on the one hand, P average value Tmj of the parameter Tl covering the range 0 - _T in P slices of amplitude τ / P and on the other hand, P values dmj means giving the corresponding corrective term to apply to the measurement. Consequently, for a measurement of a time T the value Tml measured by the starting vernier, defines in which slice one is located and a memorized table giving dmj as a function of Tmj makes it possible to extract the corrective term dmj to be applied.

It is clear that this single series of measurements applies well if the duration T to be measured is equal to or close to the calibration value R. The more the difference between T to be measured and R increases, the more the calculated deviation values dm may no longer correspond to the true deviation values to be applied. To overcome these limitations caused by the variation of t ₂ -t _l and therefore of the phase A T = TI-T2 of T in the range 0 to _T , several series of calibration measurements are carried out identical to those mentioned above. but by changing each time the value R to cover the range 0 - _T and thus have the corrective term dm, to be applied, whatever ΔT and therefore the duration T to be measured.

Let L be the number of measurement series; we denote by R1, R2, ... R _k , ... R _L the L values of R used. To have a uniform distribution, we will consider the range 0 - _T regularly divided into L slices which we call "channels" (to differentiate them from the "slices" relating to Tl, each of width _T / L, each value R _k being such that Δ R _k is in the middle of the corresponding range going from (k-1) _T / L to k _T / L, that is to say ΔR _k = (k-1) τ / L + τ / 2L substantially. this, the delay generator 6 can be equipped with delay devices connected in series to give successive steps τ / L. The table in figure 9 shows the values finally memorized in the random access memories of processor 5.

We are well aware that the value Tlm measured by the starting vernier 3 indicates the section j to be assigned, to which now corresponds either 1 but L values dm _1j to dm _Lj as a function of the phase ΔT of the time T to be measured. The corresponding calculated value Tlm-T2m defines the channel k to be assigned and it is therefore possible to extract the corrective term dm _kj to be applied for the measurement and to obtain the corrected quantity which corresponds very substantially to the real quantity of T.

As a practical example, with a clock of period _T = 10 ns and verniers of expansion coefficient K = 400, the quantum of fine measurement is given by _T / K = 25 ps, constituting the minimum time possible between samples during the calibration cycle. Under these same conditions, the range 0 - _T will be covered by a maximum of 400 distinct values of Tl and therefore of the variable phase t _l . Considering the range 0 - τ divided into 20 slices of 500 ps, i.e. 20 distinct measurable values per slice, we can decide to perform, for example, 800 measurements per channel (series of measurements at constant R) to produce with a distribution substantially uniform 40 values per range., that is a 2/1 probability of producing the different measurable values. With 10 channels spaced Ins from one to the next, the complete calibration cycle will include 8000 measurements for the case considered. To obtain the uniform distribution of the samples in the slices, we preferably carry out a random triggering of these measurements in order to regularly cover the range of variation and record an almost continuous spectrum of the variation of T _1m as a function of Tl. Naturally, the number of slices will be quantitatively determined, depending on whether the 'we are able to carry out a greater or lesser number of measurements and depending on the fineness of the correction to which we want to achieve. Random triggering of the measurement can be produced in various ways; one of them consists in producing at the microprocessor a second local clock of frequency different from that very stable SH delivered by the circuit 1, the frequencies being chosen in an irrational ratio, so that the phase presented by the front active of this local clock with respect to that of reference SH is arbitrary, practically changing value each time. This local clock thus gives successive values T1 varying randomly.

It will be noted in the case of a random triggering that the processor circuit 5 will have to temporarily store the values T1 and T2 measured by the verniers before proceeding with a classification in ascending order of the values T1 measured to then determine the means T1. slice by slice. Care should be taken to ensure that the Tl and T2 values of the same measurement are followed during these operations so as to find in each slice (Figure 7) the values (TI-T2), called Δ R, measured and corresponding to the Tl values of this section so that the determination of the mean deviation dmj retains its full meaning.

The chronometer apparatus proposed puts into practice the process which has just been described using the processor circuit 5 programmed to perform the various calculations and to control during the calibration the switching of the switches 7 to connect the outputs S10 and S20 of generator 6 on the vernier circuits in place of the inputs SI and S2; it also controls the generator circuit 6 to produce the desired series of measurements. Circuit 6 produces a start pulse S 10 and a stop pulse S2 ₀ whose delay; by compared to the start pulse, is low noise (that is to say practically without fluctuations) and is programmable over a space of time substantially equal to τ.

Referring to FIG. 10, there appears a diagram of the system which shows in more detail a ramp vernier circuit and the processor. The vernier circuit 3 comprises a threshold comparator 31 which produces reshaping of the input pulse S1 or S10; the following circuit 32 is a bistable flip-flop, the change of state of which will control, through a gate circuit 33 and a diode 34, the linear charge of the capacitor 35. The clock signal SH then controls, via the circuit 36 consisting of rocker circuits and via the gate circuit 37 followed by the diode 38, the discharge of the capacitor 35. The circuits 39 and 40 represent amplifiers. The start of the charge and the end of the discharge are respectively determined to obtain the desired expansion coefficient, for example 400 Tl, thanks to the threshold comparator 41 at the output which causes the circuit 32 to switch back to the initial position. The counter 42 performs the measurement of the total duration of charge and discharge and this information, measured in number of clock periods SH is transferred to the processor 5 which calculates the corresponding duration T1. The stop vernier 4 is similarly constructed to allow T2 to be calculated.

The processor circuit 5 is represented in a conventional structure with a microprocessor 51, input 52 and output interface circuits 53, read-only memories 54 and read-only memories 55 and the control buses C, of addressing A and of data D. In the organization of living memories 55 we considered an organization. corresponding to that of FIG. 9 at L addressing lines according to the channel and P addressing columns according to the section, to store the different measurement deviations dm _kj .

The programming of processor 5 is made to accomplish the various successive phases of the method which has been previously described. This technique responds to known, relatively simple measurements, which do not require the software to be reported here in more detail. The result of the measurement after correction is transmitted to an annex 10 operating unit.

Claims (8)

1. Electronic chronometry system using, for measuring a duration T between a starting instant t ₁ and a stopping instant t ₂ , fine counting means of the vernier ramp type with temporal expansion to measure the duration Tl between instant t and a subsequent edge of a clock signal and the duration T2 between instant t ₂ and a subsequent edge of the clock, and large counting means for counting the number N of clock periods of duration _T between said edges, the system being characterized in that it further comprises means for compensating for errors of non-linearity of the ramp to determine, in magnitude and in sign, whatever the duration T to be measured, the corrective term (dm) to be applied to obtain the corrected value (N τ + Tl - T2 + dm), said corrective term being determined during a calibration cycle as a function of the parameters Tl and (Tl - T2) measured. 2. System according to claim 1, characterized in that the compensation means comprise a processor and management processor circuit (5) and a programmable delay generator circuit (6) for producing local signals (S10 - S20) corresponding to said start and stop times and to vary the parameters Tl and Tl-T2 during the calibration so as to calculate the corresponding corrective term (dm) each time. 3. System according to claim 2, characterized in that the management and calculation means are programmed so as to control the delay generator to perform the calibration according to at least one series of measurements with a duration (R) constant between said local signals (SIO, S20) and by modifying each time the start time (t ₁ ) to browse the range of variation of the parameter Tl according to a regular distribution of distinct values. 4. System according to claim 3, characterized in that the range of variation of Tl is divided into P slices of duration τ / P, and for each of them one calculates on the one hand, the average value (Tlmj) of the values Tl measured falling in this range and on the other hand, the mean value (Δ Rmj) of the corresponding values (T1-T2), the calculation then comprising that of the mean value ( ΔR _m ) P mean values ( ΔR _ml to ΔR _mR ) of the parameter (TI-T2) and the corresponding respective deviation (dmj = Δp R _mj -ΔR) corresponding to the corrective term to be applied, section by section, depending on Tl. 5. System according to claim 4, characterized in that the calibration comprises several series of measurements, in number L, to determine L channels regularly covering the range _T of variation of the parameter (T1-T2) using L successive values of the duration T = R constant between said local signals, so as to produce an increment τ / L each time and determine the corrective term to be applied (dm _kj ) as a function of both the value of Tl and of (T1-T2 ). 6. System according to claim 5, characterized in that the management and calculation means (5) include storage means (55) for storing the different values of the corrective term to be applied (dm _kj ) according to a double entry table , on the one hand according to the parameter Tl distributed according to P slices, on the other hand, according to the parameter (T1-T2) distributed according to L channels. 7. System according to any one of claims 2 to 6 characterized in that the compensation means further comprise switching circuits (7) for connecting the inputs of the vernier circuits (3, 4) to the delay generator ( 6) during calibration, to transmit the two local signals (S10, S20) to the verniers. 8. System according to any one of the preceding claims, characterized in that the compensation means are determined to modify, during the calibration, the phase (T1) of the starting instant randomly with respect to the signal reference clock (SH).

Images