FR2727516A1 - Physical quantity likely to vary considerably measuring device - Google Patents

Abstract

The flow in a conduit is measured by a differential pressure sensor (CPD) (C1,C2) which uses the changing capacitance of an armature moving under a pressure difference. An amplifier of charges (2,A) consisting of an amplifier and capacitance produces an alternating signal (V) whose numeric frequency (7) (SU) represents the quantity being measured.A comparator (3) with upper and lower thresholds supplies a low pass numerical filter (4) whose output provides small positive and negative potentials (+dVdc, -dVdc). An amplifier (6a) with capacitance (6b) produces a signal to correct for changing drift and amplitude and this is fed to the input of the charge amplifier (2).

Description

 The present invention relates to a device intended for measuring a physical quantity whose evolution over time is oscillatory and which has a significant drift and of a random nature and / or a significant variation in its value.

 More specifically, the invention relates to such a device making it possible to compensate for a drift of the signal to be measured sometimes sometimes very clearly exceeding the significant evolution of this signal.

 In a more specific application of the invention, the measurement device preferably comprises capacitive type measurement means which are particularly suitable for being associated with integrated circuits of type C OS whose operation is based on the use of capacitors switched. However, the invention is not limited to the use of this type of measuring means.

 Measuring devices comprising capacitive measuring means of a physical quantity with an oscillatory appearance are already widely used. They generally include a capacitive sensor delivering a signal which is sampled to form a series of charge values which are a function of the physical quantity. These values are processed in a charge amplifier forming part of an interface connected between the sensor and the use circuit intended to process the measurement signal.

The interface also includes means for converting the output of the charge amplifier into a frequency signal which is the measurement signal usable by the operating circuit.

 When the physical quantity to be measured has a large random drift or varies over a wide range of values, it is advisable to provide in the interface compensation means for the disturbances which may result.

Several methods are currently used for this, including, for example, the use of a high-pass filter connected to the output of the charge amplifier.

 However, such a high-pass filter is not suitable for the cases to which the invention applies, since the gain of the charge amplifier must have a minimum value imposed by the minimum signal to be detected. nais if the drift of the quantity to be measured is large, the minimum gain will cause the amplifier to saturate when a significant drift appears, which then blocks any possibility of measurement.

 Another method is to use a compensation loop whose feedback path has an analog filter. This method also has drawbacks, especially if the measurement signal has a low, or even very low, cutoff frequency.

 The object of the present invention is to provide a measurement device of the kind described briefly above and making it possible to accept large quantities of drift and variation in amplitude without resulting in the drawbacks indicated in the above-mentioned solutions. above.

 The subject of the invention is therefore a device for ensuring a physical quantity capable of undergoing a significant drift and / or of varying over a wide range of values, and the evolution of which over time has an oscillatory appearance, this device comprising a Measurement sensor intended to deliver an alternating voltage representative of this oscillatory shape, a comparator intended to compare said alternating voltage with at least a threshold to generate a digital signal whose frequency is representative of said physical quantity, and an adjustment loop for compensate for the effect of the drift on the result of the Measurement, this device being characterized in that said adjustment loop comprises a digital low-pass filter connected to the output of said comparator to generate a control signal, when said digital signal has a characteristic predetermined time representative of a drift variation, and in that said The adjustment loop further includes a compensation voltage generator connected between said low-pass filter and said measurement sensor to apply a compensation voltage to the latter opposing drift and for, each time said signal appears. command, nest the compensation voltage delivered to said sensor by a predetermined amount as a function of the evolution of the drift.

 Thanks to these characteristics of the invention, it is possible not only to compensate for the drift and to accept large variations in the measured quantity, but also to compensate for disturbing effects due to intrinsic faults in the circuit elements used such as those due to leakage currents or variations in the temperature of the components used.

Other characteristics of the invention will appear in the course of the description which follows, given solely by way of example and made with reference to the appended drawings in which:
- Figure 1 is a block diagram of a measuring device according to the invention, this device being of the capacitive type and intended more specifically to measure the flow of a fluid in a pipeline;
- Figure 2 shows graphically the operating principle of the measuring device shown schematically in Figure 1;
- Figure 3 is a block diagram of a capacitive differential sensor with its interface which is designed according to the prior art;
- Figure 4 is a block diagram of the flow measurement device using the concept of the invention;
- Figure 5 shows the more detailed diagram of a first embodiment of the device according to the invention;
- Figure 6 shows graphiquenent the operating principle of the device shown in Figure 5;
- Figures 7 and 8 are chronogranes illustrating the operation of the device of Figure 5;
- Figure 9 is a partial diagram of a second embodiment of the measuring device according to the invention;
- Figure 10 shows graphically the operation of the second embodiment of the invention;
- Figure 11 is a partial diagram of a third embodiment of the device according to the invention;
- Figure 12 graphically illustrates the embodiment of Figure 11;
- Figure 13 is a partial diagram of a fourth embodiment of the device according to the invention;
- Figure 14 graphically illustrates the embodiment of Figure 13;
- Figure 15 is a partial diagram of a fifth
Embodiment of the device according to the invention;
- Figure 16 graphically illustrates the embodiment of Figure 15;
- Figures 17 and 18 are partial diagrams of the device according to the invention, conforméMent to two alternative embodiments thereof.

In the description which follows, the invention will be illustrated with reference to a particular application of
Measurement of a flow rate of a fluid in a pipeline.

However, it is understood that the concept of the invention can be applied in a very large number of cases, each time a physical quantity with an oscillatory appearance is measured using a sensor delivering an alternating voltage which is confronted with a threshold (or, as we will see later, with several thresholds), in order to generate an impulse signal whose frequency is representative of the magnitude measured.

 The basic concept of the invention applies regardless of the nature of the sensor used. However, in what follows, it will become clear that the advantages which the invention makes possible are particularly noticeable, if this sensor is of the capacitive type and even more specific, if this sensor is of the differential capacitive type.

 In the case of a capacitive sensor, the signal representative of the quantity measured is obtained by sampling by transforming a series of accumulations of electrical charges into an alternating electrical voltage.

This alternating voltage is confronted with a threshold in a comparator which provides at its output a digital signal whose frequency translates the value of the physical quantity to
Measure.

 This being the case, it is known that the measurement of the flow of a fluid, gaseous or liquid, can be carried out using a fluidic oscillator which generates alternating pressure variations whose frequency is directly proportional to the flow rate of the fluid. Such a measuring device is described for example in US Pat. No. 4,694,702.

 The operation of a fluidic oscillator is based on the dynamics of the vortices which are formed in two small cavities provided on both sides of a measuring body placed in the flow of the fluid. This configuration creates instabilities causing a continuous oscillation in the flow between the two sides of the measurement body. FIG. 1 very schematically shows the method based on this measurement principle and used in this example of application of the invention.

FIG. 1 shows a pipe CA in which a measuring body CM is placed, the lateral surfaces of which define the cavities V in which the vortices are formed. Each of these cavities is connected to a small tubing T connecting the measurement point to a differential pressure sensor CPD disposed in the wear body CM and connected to an interface I. This is designed to develop a useful digital signal SU intended to be used in a CE operating circuit, for example a device capable of regulating the flow rate in the pipeline
CA as a function of the measured flow variations.

The CPD differential pressure sensor is preferably of the micro-machined type produced by photolithography techniques on silicon according to a technology developed by the Applicant for capacitive microacceleroreters. A description of these can be found in an article by Y. de Coulon et al. in wUransducersw of 1993, Yokohama, entitled "Design and Test of a Precision
Servoaccelerometer with Digital Output ".

It suffices to point out here that such a sensor comprises a moving mass suspended between two fixed electrodes, the assembly forming two capacities. The mobile mass is subjected respectively on these two faces to the pressures prevailing in the tubes T shown in FIG. 1 of
So that the differential pressure applied to the moving mass can vary the two capacities that it forms with the fixed electrodes.

 FIG. 2 shows as a function of time, in a) an example of the flow rate of the fluid flowing in the pipe CA, in b) the flow rate of the differential pressure aP which results therefrom in the CDP sensor, and in c) the output signal SU which must appear on the output of interface I.

 According to the prior art, the output signal SU is obtained by sampling the charge values accumulated on the capacitors of the sensor, using the interface I shown in FIG. 3 and commonly used to transform the signal of a sensor capacitive differential into a signal of the kind shown in Figure 2c.

 It can be seen that the CPD sensor forms two variable capacities C1 and C2 which share a reinforcement cor 1 submitted to the differential pressure AP resulting from the pressures P1 and P2 prevailing in the respective pipes T.

The communal armature is thus subjected to deviations which cause corresponding variations in the capacitors C1 and C2 and consequently variations in voltage AVX and -AVX on their respective armatures. The armature horn a 1 is connected to the input of a charge amplifier 2 composed of an amplifier A and of a capacitor C mounted between the input and the output of the latter. The signal appearing on this output will be of the form: CV = AV (C1 - C2), in which C, C1 and C2 are the corresponding capacitance values. The voltage V appearing on the output of amplifier 2 is thus a voltage alternative, the frequency of which is representative of the physical quantity to be measured, in this case the flow rate in the pipeline CA of FIG. 1. This voltage V is applied to a coiparator 3, at the output of which the signal SU of the form is generated shown in Figure 2c.

 In the case of a measurement carried out using the device which has just been described, the problem which the invention proposes to solve is particularly felt. Indeed, the useful frequency range is 0.1 to 45 Hz and at the lower frequency of this dondaine, the variation in AC capacitance is only 10 fF (festofarads). However, the drift of the CPD sensor, due for example to temperature variations (sensor leakage currents, for example) or to absolute pressure variations, can reach up to 1 pF, which is 100 times higher than the minimum variation that the device must still be able to measure.

 Figure 4 is a block diagram of a measuring device into which the basic features of the invention are incorporated. It can be seen that this device comprises a capacitive differential sensor CPD and a charge amplifier 2 identical to those already described in connection with FIG. 3. The output of the charge amplifier 2 is connected to a comparator 3 which in the embodiment preferred described here is double threshold.

 According to the invention, the outputs 3a and 3b of the comparator 3 are connected to the respective inputs 4a and 4b of a digital filter 4 of the low-pass type. The outputs 4c and 4d of this filter 4 are connected together to a compensation voltage generator 5, generating on the one hand voltage fractions + vX and -AVs and comprising on the other hand an integrator 6 composed of an amplifier 6a and a capacitor 6b connecting the input and the output thereof to each other.

 The output of the integrator 6 is connected to the input of the charge amplifier 2 by means of a capacitor C. It can therefore be seen that the mounting of the filter 4 and of the generator 5 constitutes an adjustment loop allowing compensate for the effects of drift and other disturbances on the signal delivered at the output of comparator 3.

 The digital filter 4 is intended to transmit to the integrator 6 fractions of compensation voltage only if it detects in the output of the comparator 3 elements representative of the drift and / or other disturbances introduced by the components of the circuit himself. Correct in particular the drift is an essentially slower phenomenon than the flow variations that one wishes to measure. The digital filter is a low-pass filter which reacts only to slow and persistent variations of the useful signal in order to draw the signal to compensate for this drift.

 Referring now to Figure 5, a preferred embodiment of the invention will be described. This embodiment is implemented using the technology of connoted capacities, well known to specialists.

 The fixed armatures of capacitors C1 and C2 are connected respectively to the supply voltage V and to the supply voltage V via switches la and lb. They are also connected to the potential of the trap V, by means of respective switches 2a and 2b. The capacitor C connected in parallel to the amplifier A is also placed in parallel with a switch lc. The capacitor C is connected to the trap via a switch 2c and to the output of the amplifier 6a through a switch ld.

 In the embodiment described here, the comparator 3 has two outputs 3a and 3b on which the signals V and Vo appear respectively, the levels of which change when the signal present at its input respectively exceeds up a threshold sl or down a threshold s2.

 The outputs 3a and 3b of the comparator 3 are respectively connected to the inputs S and R of a flip-flop RS 7, the output Q of which outputs the output signal SU from the measurement device on a conductor 8.

 In FIG. 5, the elements making up the low-pass filter 4 and the compensation voltage generator 5 have been surrounded by six lines. The input D of a first flip-flop 9 of type D is connected to the output 3a of the comparator 3. Its input Ck is connected to a clock circuit 10 which delivers several secondary clock signals derived from the general clock signal H of the measuring device (See curve e in FIG. 7). In this case, the input Ck of the flip-flop 9 is connected to the output p4 of the clock 10 and the signal of this output is represented by the curve c in FIG. 7. The output Q of the flip-flop 9 is connected to one of the inputs of an AND gate 11, the other input of which is connected to the output 3a of the comparator 3.

 A second flip-flop 12 of type D has its input connected to the output 3b of the comparator 3, while its input Ck is connected to the output p4 of the clock 10. Its output Q is connected to one of the inputs of a AND gate 13, the other input of which is connected to the output 3b of the comparator 3.

The outputs of AND gates 11 and 13 are connected respectively to the inputs of an NI gate 14.

 The output of the NI gate 14 is connected to one of the inputs of an OR gate 15, the output of which is connected to a reset reset input of a counter 16, the counting input Ck of which is connected to the clock output p4 10.

The counting outputs 16a to 16f of the counter 16 are connected in parallel to as many inputs 17a to 17f of a comparator 17 in which a predetermined number is stored (chosen equal to 5 in the example described here).

When the number counted by the counter 16 is equal to that recorded in the comparator 17, the latter delivers an output signal to the input D of a flip-flop 18 of type D of which 1 the clock input is connected to the output p4 of the clock 10. The output Q of the flip-flop 18 is connected to the other input of the OR gate 15.

The filter 5 also includes two AND gates 19 and 20 with three inputs. A first input of these gates is connected to the output p5 of the clock 10, a second input is connected to the output of the comparator 17, while a third input of the AND gates is connected to the output 8 of the circuit, being specified that the third door entrance
AND 20 is inverted. These AND gates provide their outputs with control signals p5d and p5s respectively.

 The amplifier 6a is of the operational type and peddles a non-inverting input connected to the basket and an inverting input which is connected to a generator 21 of fractions Avs or -Avs of the compensation voltage V. This generator comprises two capacities C3a and C3b which can be selectively connected, on the one hand to the inverting input of the amplifier 6a and, on the other hand to the supply voltages VX and V,. For this purpose, switches 4a to 4d are connected in series with the respective capacitors C3a and C3b and between the respective voltages VX and VM and ground. Furthermore, they are connected in series with respective switches 5da, 5db and 5sa, 5sb between ground and the inverting input of the operational amplifier 6a.

 FIG. 6 illustrates the operating principle of the measurement device represented in FIG. 5. An example of the evolution of the signal sN to the meter is seen, in this example subject to increasing drift D. The figure also shows how the drift is thought, the graph representing the sum V of the signal to be measured and the compensation voltage which is modified, each time the drift is modified. It can be seen that, according to the present embodiment, compensation takes place when the signal V remains longer above the threshold sl established by the comparator 3, than the length of a predetermined tempo T, which in reality amounts to a low-pass filtering function.

Of course, if the drift is of opposite sign, an inverse compensation will be brought to the signal V by the feedback loop.

 Referring again to FIG. 5, it will be noted that the switches la to ld, 2a, 2b and 2c, 4a to 4d, 5da, 5db, and 5sa, 5sb are controlled respectively by the clock signals pl, p2, p4 and p5 according to the chronogranes which will now be described with reference to FIGS. 7 and 8.

 It will be noted that these chronograms are not plotted on the same scale, FIG. 7 showing a little more than half a period of the signal p5, while FIG. 8 shows over a certain number of periods of this signal p5, conventional 'performs drift compensation. Furthermore, by way of example, the conarde signals are assumed to make the respective switches passable, when the signal considered is at the high level.

FIG. 7 illustrates a measurement cycle of the device according to the invention which takes place, in this example, over 30 periods of the clock signal H, periods which correspond to a cycle of the signal p5. It should be noted that during almost all of this period, the elements of the measurement device are not active, the sampling of the measurement values and the compensation for the drift taking place each time only over part of The interval during which the signal p5 is at the low level, and over an even smaller part of the time interval during which this signal p5 is at the high level. It is clear that this property of the measuring device according to the invention makes it possible to save energy. In a practical embodiment, the
The Applicant was able to reduce the active period of the device to only 5% of the total of the period of signal p5.

 Thus, the interval tl in FIG. 7 corresponds to a time of inactivity of the device which corresponds to working only when the signal pl goes to the high level to make the switches 1a to 1d passable (FIG. 5).

During the interval t2, the capacitors C1, C2 and CX are charged, while the capacitor C is discharged. The capacity Cs then receives the quantity of charge necessary, if necessary, to compensate for the drift.

 During the interval t3, the charges accumulated on the capacitors C1, C2 and C are transferred to the capacitor C of the charge amplifier 2. The interval t4 is used to compare the signal V resulting from the quantity of charges accumulated, with the comparator thresholds 4. The interval p5 makes it possible to charge the capacities C3a and C3b of the generator 21 of fractions of the compensation voltage, while a very small portion of the interval t6 is used to transfer the charges capacities C3a or C3b on the capacity 6b of the integrator 6, provided that the comparator 3 and the filter 4 have found that a modification of the compensation for the drift is necessary.

 In accordance with the embodiment of FIG. 5, the digital filter 4 is arranged so that the counter 16 counts the number of successive sampling cycles during which the result of the comparison carried out in the comparator 3 shows that the voltage V is either above the upper threshold sl, or below the lower threshold s2 fixed in this comparator. On the other hand, when the outputs Va and Va of the comparator 3 change, or when the voltage V is between the two thresholds, the counter 16 is reset to zero.

 FIG. 8 represents, by way of example, several sampling cycles of the Measuring device according to the invention. During cycles 1, 2 and 3, the voltage V changes so that the counter 16 is not able to fill up completely. On the other hand, during cycles 4 to 8 inclusive, we see that the voltage V remains above the upper threshold sl established by the comparator, which means that the counter 16 was able to fill up and satisfy the comparison made in the comparator 17 which can thus deliver an output signal to flip-flop 18.

Simultaneously, the AND gate 19 is opened to deliver a signal to the switches 5da and 5db which connect the capacity
C3a at the input of the integrator 6. I1 results in an increase of a determined fraction 22 (FIG. 8) of the compensation voltage V at the output of this integrator 6. During the following measurement cycles, this quantity continues to be subtracted from the voltage V, until again, for a determined number of consecutive cycles, the voltage V is situated above or below the thresholds sl, s2 fixed by the comparator 3. The value of the load applied to the capacitor C is then corrected again accordingly. Of course, when the voltage V is below the threshold s2, it is the gate 20 which is open and it is the switches 5sa and 5sb which are turned on to operate an inverse correction of the compensation voltage V.

It should be noted that in reality, the predetermined number of cycles set in the comparator 17 must be significantly greater than five (as in the example described here), the
DeManderesse having used a number equal to 640 in a practical realization. It has been found that this predetermined number (or in other words the predetermined time T which must elapse before a drift compensation is carried out) corresponds to a half-period of the signal with the weakest measurer frequency.

 FIG. 9 represents a second embodiment of the device according to the invention in which a comparator 3A is used, opposing only one threshold s to the voltage V.

Furthermore, this embodiment uses another low-pass filter 4A which differs from filter 4 of the first embodiment of FIG. 5 by the omission of the flip-flop 12 of type D.

 In addition, the output Q of the flip-flop 9 is connected to one of the inputs of an EXCLUSIVE gate 23, the other input of which is connected to the output of the comparator 3A. The output of gate 23 is connected to one of the inputs of gate 15. FIG. 10 graphically shows that in the case of FIG. 9, a modification of the compensation voltage is carried out, when the signal V remains the same side of the threshold s during a predetermined time T.

 Figures 11 and 12 illustrate a third embodiment of the invention which differs from the previous two with regard to the arrangement of the low-pass filter.

 In this case, the low-pass filter 4B is similar to that of FIG. 9, however that the second input of the OU-EXCLUSIVE gate 23 is connected to the output Q of the flip-flop RS 7 (which does not exist in the Embodiment of FIG. 9) via a conductor 24. It will be noted that in this third embodiment, a comparator 3 with two thresholds sl and s2, used in the first embodiment, is used.

 The measurement device according to FIGS. 11 and 12 works out a modification of the compensation voltage Voe when the output signal SU has not changed level during a predetermined time T. In other words, we examine the duty cycle . This solution has the advantage of eliminating the risk that the voltage V will never cross one of the thresholds sl or s2, which would have the consequence that the output signal would never change level. It should also be noted that the first and third embodiments of the invention make it possible to obtain a certain hysteresis of the output signal SU.

 Figures 13 and 14 show a fourth embodiment of the invention in which the measuring device according to the invention comprises a low-pass filter 4C working on the output SU of the flip-flop RS 7, glued in the previous embodiment .

 This low-pass filter 4C peddles two AND gates 25 and 26, the first inputs of which are connected to the output p4 of the clock 10. The second input of the AND gate 25 is inverted and it is connected, in common with the second entrance of door 26, exit of door RS 7.

 The outputs of the AND gates 25 and 26 are respectively connected to the decowpting and counting control inputs 27a and 27b of a co-decoder / decoder 27. The outputs 27c at 27h of the co / decoster 27 are connected respectively to as many inputs two coiparators 28 and 29 in which are stored the positive value and the negative value of a predetermined number N intended to be compared with the content of the counter / countdown 27. When they find equality with their respective tenorised values, the comparators 28 and 29 deliver a signal to the respective inputs of an OR gate 30 the output of which is in turn connected to the input D of a D type flip-flop 31. This receives the signal p4 from the clock 10 on its input Ck, while its output Q is connected to the reset input reset of the up / down counter 27.

 The outputs of the comparators 28 and 29 are also connected to the respective inputs of the AND gates 19 and 20.

 FIG. 14 graphically shows the operation of the measurement device according to the embodiment of FIG. 13. It can be seen that the up / down counter 27 counts the pulses of the signal p4 upwards, when the signal SU is at the high level, and towards low, when this signal is low. Consequently, the compensation voltage generator 6 V receives a control signal if the counter 27 has completely folded back, or in other words if a certain time T has elapsed between two transitions of the output signal SU. It is understood that this operation also applies when the comparator has only one threshold.

 In the embodiment of FIG. 15, a 4D low-pass filter is used which differs from the embodiment of FIG. 13 in that the up / down counter 27 is controlled by the transitions of the outputs Vsh and Vsb of comparator 3 rather than to be by those of the output signal SU. In this case, the AND gate 25, one of the inputs of which is inverting, is replaced by a siMple AND gate 32.

The operation of this assembly appears on figure 16.

 We will now refer to FIG. 17 which represents a variant of the device according to the invention also relating specifically to the digital filter designated by 4E. In this case, the design of the filter is that of FIG. 5, while the output of the NI gate 14 is connected directly to the reset input of the counter 16 so that its reset is not carried out. only when comparator 3 changes output.

Furthermore, the counter 16 is connected in parallel to several comparators 17-1 to 17n which each contain a different fixed number to compare with the content of the counter 16.

The fixed numbers of comparators are increasing, but are increasingly closer to each other. For example, it can be chosen at 640 for comparator 17-1, at 900 for comparator 17-2, at 1000 for comparator 17-3, at 1020 for comparator 17-4, at 1025 for comparator 17 -not.

 Furthermore, the outputs of the comparators are connected to the respective inputs of an OR gate 23 the output of which is connected to the second inputs of the AND gates 19 and 20. Thanks to this circuit, compensation for drift can be performed more and more often, if the results of comparisons made by comparator 3 persist in the same direction.

 Of course, we gave here a wired version of the digital filter, but it is possible to obtain the same result, even a more detailed analysis of the evolution of the drift as well as a more elaborate compensation adapted to this evolution, in using a microprocessor said prograMme.

 According to yet another variant of the invention, the voltage V could be confronted with more than two thresholds, whereby an adequate algorithm, executed by a microprocessor for example, could effect the modification of the compensation according to external criteria or to different times according to the result of the confrontation of the voltage V with these thresholds.

 Finally, according to yet another variant of the invention shown in FIG. 10, it is possible to use a generator of compensation voltage fractions (designated by 21A in FIG. 18) comprising pairs of capacitors of different values. Thus, provision is made for each direction of compensation for several capacities C3a-1 to C3a-n, respectively C3b-1 to C3b-n, each connected by respective switches to the potentials V, VM and of the ground. These switches are controlled in the same way as those in FIG. 5, it being understood that for a specific drift, a couple of capacitors will be chosen which are adapted to the desired degree of compensation for the drift. In this way, it will be possible to adapt the compensation for the drift to the evolution of several parameters by controlling the capacity switches so as to selectively insert them into the generator circuit. This can be used in particular for example for the commissioning of the measurement device where the drift can be very significant at the start. The compensation steps such as 22 (figure 8) can then be important at the start to quickly integrate the initial drift then gradually decreasing, when the system reaches a stabilized operating regime.

Claims (12)

 l.-Device for measuring a physical quantity liable to undergo a significant drift and / or to vary over a wide range of values, and the evolution of which in tempo has an oscillatory appearance, this device comprising a measurement sensor ( CPD) intended to deliver an alternating voltage (V) representative of this oscillatory shape, a comparator (3) intended to compare said alternating voltage (V) with at least a threshold (s, sl, s2) to generate a digital signal (SU ) whose frequency is representative of said physical quantity, and an adjustment loop to compensate for the effect of the drift on the result of the Measurement, this device being characterized in that said adjustment loop comprises a digital low-pass filter (5, 5A to 5E) connected to the output of said comparator (3) to generate a control signal (p5d, p5s), when said digital signal (SU) has a predetermined time characteristic representative of a drift variation, and in that said adjustment loop also includes a compensation voltage generator (5) (Vo.q) connected between said low-pass filter (4, 4A to 4E) and said measurement sensor (CPD ) to apply a compensation voltage to the latter opposing the drift and for, on each appearance of said control signal (p5d, p5s), modify by a predetermined amount (22) the compensation voltage delivered to said sensor by function of the evolution of the drift.  2.- Measuring device according to claim 1, characterized in that said measurement sensor (CPD) is a differential capacitive sensor which is connected to said comparator (3) via a charge amplifier (2), and in that said sensor (CPD) and said amplifier (2) include story capacitors (C1, C2, C).  3.- Measuring device according to claim 2, characterized in that said condensing voltage generator (5) comprises an integrator (6) selectively coupled to said differential capacitive sensor (CPD) by means of a told coupling capacity (C) making it possible to transmit said compensation voltage (V) to said sensor, and in that said compensation voltage generator (5) also includes a circuit (21; 21A) with told capacities (C3a, C3b; C3a-1 to C3a-n, C3b-1 to C3b-n) intended to apply to said integrator (6) coMpensation voltage fractions (+ # V @@, -bV,), at each appearance of said control signal (p5d, p5s).  4.- Measuring device according to any one of claims 1 to 3, characterized in that said low-pass digital filter (4; 4D) comprises generating means (9, 11 to 20; 19, 20, 26 to 32 ) to generate said control signal (p5d, p5s), when said control voltage (V) has a value having with respect to said threshold (s) a predetermined relationship during a predetermined time interval (T).  5.- Device according to any one of claims 1 to 3, characterized in that said low-pass nuMeric filter (4B; 4C) comprises generating means (9, 15 to 20, 23; 9, 15 to 20, 23 , 24) to generate said conarde signal (p5d, p5s), when said digital signal (SU) has a predetermined duty cycle.  6.- Device according to any one of claims 4 and 5, characterized in that said means generating the conarde signal include counting means (16 to 20; 19, 20, 25 to 32) to define said interval of teMps predetermined (T), these counting cores being triggered at each transition of the output signal (Va, Vab) of said comparator (3) or of said output signal (SU), and supplying said control signal when their content reaches a position of predetermined count.  7.- Measuring device according to claim 6, characterized in that said counting cores comprise an upward counting counter (16).  8. A measuring device according to claim 6, characterized in that said counting means include an up / down counter (27).  9.- Measuring device according to claim 6, characterized in that said counting means peddle means (17-1 to 17n) to allow the production of said conarde signal (p5d, p5s) for a plurality of counting positions.  10.- Device according to any one of claims 1 to 9, characterized in that said compensation voltage generator (5) comprises means for varying said predetermined amount of voltage (22).  11.- Measuring device according to any one of the preceding claims, characterized in that said measurement sensor (CPD) is a differential pressure sensor.  12.- Measuring device according to claim 11 of the flow of a fluid flowing in a pipe (CA), characterized in that said differential pressure sensor (CPD) is mounted in a fluidic oscillator (V, T) placed in the flow of the fluid.