DESCRIPTION

TITLE: Pendulum accelerometric sensor with conditional capacitive detection

The present invention relates to a closed-loop pendular accelerometer sensor with electrostatic control and detection, used for the detection of physical quantities, and a method for controlling such a sensor. The sensor is for example a MEMS technology sensor (English acronym for Micro Electro Mechanical Systems).

An electrostatic pendulum accelerometer comprises a casing and a seismic mass connected to the casing by one or more hinges positioned in such a way that the seismic mass forms a pendulum movable relative to the casing, either in translation or in rotation. The displacements of the seismic mass under the effect of acceleration are generally detected by means of three electrodes.

A first fixed electrode and a second fixed electrode are integral with the housing and connected to an excitation circuit.

The third electrode, mobile, is carried by the pendulum and connected to a detection circuit.

Each fixed electrode forms with the mobile electrode a capacitance whose value depends on their spacing. In the absence of a manufacturing defect and when the sensor is not subjected to an acceleration along its sensitive axis, the pendulum remains in its neutral position in which the two capacitances are equal. On the other hand, when the pendulum is subjected to an acceleration along its sensitive axis, it moves, causing a consecutive decrease in the capacitance formed by the mobile electrode and one of the fixed electrodes, and an increase in the capacitance formed by the mobile electrode and the other fixed electrode.

This capacitance variation also depends on the deformations of the casing and of the pendulum.

In closed-loop operation, the position of the pendulum is controlled in a neutral position, or target position, halfway between the fixed electrodes, by applying to the pendulum an electrostatic force which must compensate for the acceleration applied along the axis sensitive. The electrostatic force is the result of voltages applied to the electrodes to keep the capacitance difference to zero.

The sensor comprises an excitation circuit for each fixed electrode, arranged to supply the electrodes so as to generate said electrostatic force.

The quadratic character of the electrostatic force with respect to the applied voltages complicates the design of the control circuit carrying out the servo-control of the pendulum and the estimation of the acceleration.

To circumvent this difficulty, it is known to control the pendulum in all or nothing using calibrated voltage pulses.

These pulses are applied to one or the other of the electrodes depending on whether it is a question of pulling or pushing the pendulum to bring it back to its set position. The density of pulses intended to push, respectively to pull the pendulum, i.e. the quantity of pulses per time interval, is then an affine function of the acceleration to be measured.

Thus, zero acceleration is compensated by an equal number, on average, of pulses in both directions.

However, the symmetry of the pulses applied to the two electrodes may be imperfect due to a difference between the duration of the pulses applied to the first fixed electrode and the duration of the pulses applied to the second fixed electrode.

In this case, the pulse density is modified by the servo to maintain the pendulum in the set position, which biases the estimate of the acceleration.

In order to improve the performance of this type of sensor, it has been proposed, in document WO 2014/128027, to use a common excitation circuit, thus limiting the problems of manufacturing asymmetry and aging of the electronics of the excitation circuit.

It has also been proposed, in document WO 2017/85142, to implement a fine control phase to send moderate control pulses making it possible to obtain

optimal performance over a reduced measurement range, and an extended operating command phase, in which high amplitude command pulses are sent to extend the measurement range, to load the sensor at full scale, with possibly degraded performance.

Although advantageous in many respects, the aforementioned sensors have a relatively high power consumption.

In view of the foregoing, the object of the invention is to provide an electrostatic pendular accelerometer sensor having reduced electrical consumption, while maintaining improved performance.

Another object of the invention is moreover to propose such a sensor having a simple implementation structure.

The invention thus proposes an accelerometric sensor, comprising a housing, a pendulum fixed to the housing, a mobile electrode carried by the pendulum and connected to a detection circuit, a pre first electrode and a second fixed electrode integral with the casing to form with the mobile electrode two capacitors of variable capacity as a function of a distance between the electrodes, and a control unit configured to carry out detection operations to measure the variable capacities of the capacitors and a control operation of the mobile electrode as a function of the capacities measured by applying a control logic signal from a switch for selective connection of the fixed electrodes to an excitation circuit delivering a control signal to the fixed electrodes to maintain the pendulum in a predetermined position.

The control unit is configured to apply, at each sampling period, a first detection signal to one of said fixed electrodes chosen according to the logic level of the control signal and a second detection signal to the other fixed electrode , the control signal being applied to the electrode to which the second detection signal is applied.

Thus, the reduction of the sensor consumption and the improvement of the performances are obtained by reducing the number of switchings of the switch, by applying two detection signals and a control signal.

According to another characteristic, the first and second detection signals are signals in the form of a slot.

According to yet another characteristic, the switch comprises a first input terminal at a reference potential provided by the excitation circuit and a second input terminal at zero potential for selectively connecting said fixed electrodes to the excitation circuit or at zero potential.

In one embodiment, the excitation circuit includes a digital-to-analog converter connected to the switch and driven by the control unit.

The detection circuit may comprise an amplifier stage having an input connected to the mobile electrode and an output connected to an analog-to-digital converter having an output connected to the control unit.

For example, the control unit comprises a first pendulum position estimator connected at the input to the output of the detection circuit and an output connected to the negative input of a comparator having an output connected to an input of a corrector having an output connected to a sequencer and to a second estimator having a first output connected to the positive input of the comparator and a second output providing an estimate of the acceleration.

In one embodiment, the control unit is configured to apply the second detection signal and the control signal by means of a common detection and control pulse.

The invention also relates to a method for controlling an accelerometric sensor as defined above, comprising the steps of:

-detection of the variable capacitances of the capacitors by applying, at each sampling period, a first detection signal to one of the fixed electrodes chosen according to the logic level of the control signal and a second detection signal to the other fixed electrode ;

-control of the movable electrode as a function of the capacities measured by applying a control logic signal from a switch for selective connection of the fixed electrodes to an excitation circuit delivering a control signal so as to apply the control signal to the electrode to which the second detection signal is applied.

According to this method, the second detection signal and the control signal are advantageously applied by means of a common detection and control pulse.

Other aims, characteristics and advantages of the invention will appear on reading the following description, given solely by way of non-limiting example, and made with reference to the appended drawings in which:

[Fig 1] is a schematic view of a sensor according to one embodiment of the invention;

[Fig 2]

and

[Fig 3] are timing diagrams showing the application of the first and second detection signals and of the control signal to the fixed electrodes as a function of the logic level of the control signal; and

[Fig 4] is a timing diagram illustrating another embodiment of the invention in which the detection signal applied to the controlled electrode and the control signal are applied by means of a common detection and control pulse .

There is shown in Figure 1 an accelerometric sensor according to the invention, designated by the general reference numeral 1.

The accelerometric sensor here is a microelectromechanical system, also called MEMS, by etching a plate of crystalline or semi-crystalline material, such as silicon.

The sensor comprises a casing 2 to which a solid body 3 is articulated by means of a hinge 4 positioned in such a way that the solid body 3 forms a pendulum movable with respect to the casing 2 according to a pivoting movement.

thesensor 1 comprises a first fixed electrode 5.1 and a second fixed electrode 5.2 which are integral with the casing and which are connected to an excitation circuit designated by the reference 6, and a third electrode 5.3 carried by the solid body 3 is connected to a detection circuit 7. A control unit 8 is connected to the excitation circuit 6 and to the detection circuit 7.

The excitation circuit 6 comprises an output connected to a switch 9 with two positions connected to the first electrode 5.1 and to the second electrode 5.2 to selectively connect them to the excitation circuit 6.

More precisely, the switch 9 comprises a switch II arranged to connect the first electrode 5.1 either to the output of the excitation circuit 6 or to ground and a switch 12 arranged to connect the second electrode 5.2 either to the output of the circuit of excitation 6 is grounded.

Switch 9 is controlled by control unit 8.

The control unit 8 comprises a first estimator 10 connected at the input to the detection circuit 7 and an output connected to a negative input of a comparator 11 having an output connected to an input of a corrector 12 whose output is connected to a sequencer 13.

The control unit 8 further comprises a second estimator 14 having an input connected to the output of the corrector 12, an output connected to the additive input of the comparator 11 and an output providing the estimate of the acceleration i.

Furthermore, the excitation circuit 6 comprises a digital-analog converter 15 connected to the switch 9 and controlled by the control unit 8.

The detection circuit 7 comprises a main amplifier stage 16 comprising a charge amplifier 17 equipped with a loop capacitor 18 of capacitance Cref and a switch U.

The amplifier stage has an input connected to the mobile electrode 5.3 and an output connected to an input of an analog-digital converter 19 having an output connected to the first estimator 10 of the control unit.

This sensor works as follows.

The control unit 8 manages the operation of the sensor and in particular the chronology of the various operations sampled at a frequency FS. The sequencer 13 sequences the operations within the sampling periods Ts, by sequentially and cyclically controlling the digital-analog converter 15 by a command w, the analog switches II, 12 by a command 5, the analog-digital converter 19 by a command c and the analog switch 13 by a command r.

Depending on the logic state of command 5, one of electrodes 5.1,

5.2 is connected to the output V of the digital-analog converter 15 while the other electrode 5.2, 5.1 is simultaneously connected to ground. The electrode connected to the output of the converter is thus positioned at the reference potential supplied by the excitation circuit 6.

The accelerometric sensor is controlled by the control unit so as to implement, at each sampling period TS, a detection phase of the variable capacities Ch and Cb formed between the first fixed electrode 5.1 and the movable electrode

5.3 and between the second fixed electrode 5.2 and the mobile electrode 5.3, respectively, and a control phase in which an excitation signal is applied to one of the fixed electrodes so as to return the pendulum to its set position thanks to the electrostatic force applied to the plates of the capacitor whose fixed electrode has been selected by the command s. At the end of the detection phase, the corrector 10 determines the sign of a logic control signal bs in order to determine whether the control signal u must be applied to the fixed electrode 5.1 or to the fixed electrode 5.2

If bs=+l, the voltage is applied to the electrode 5.2 which then pulls the pendulum towards it.

If bs=-l, the voltage is applied to the electrode 5.1 which then pulls the pendulum towards it.

The detection of capacities is carried out conditionally, according to the sign of the control signal bs, the chronological order of the detections being determined by the sign of bs.

The linearized expression of the capacitance of the first so-called “high” fixed electrode Ch and of the second so-called “low” fixed electrode Ch is given by the following relations:

(eq. 1)

(eq. 2)

And the relative position of the pendulum is given by the relation:

(eq. 3)

in which Co denotes the initial capacitance, Ci denotes the active capacitance, z is the position of the pendulum, e the width of the air gap, that is to say the distance between the electrodes 5.1 and 5.3 or between the electrodes 5.2 and 5.3, which are equal at rest, Vref is the reference voltage supplied by the analog-digital converter 15 and applied to the electrodes and Qb and Qh are the charges transferred to the detection circuit 7, which correspond to the charge variations at the terminals of the capacitors variables subject to a voltage rising edge ranging from 0 to Vref.

Thus, for each sampling period, during the detection phase, two capacitive readings are implemented to estimate the position of the pendulum and supply the corrector.

Within a sampling period, the position of the pendulum varies little between the two measurements.

The detection of the variable capacities is a conditional detection, the order of the capacitive detection being conditioned on the sign of the logic control signal bs coming from the corrector 12. A pseudo-random permutation of the order of the detections is therefore performed, the permutation being pseudo-random because of the properties of the control signal bs in a sigma delta type loop which are those of a white noise filtered by a high-pass transfer function determined in particular by the corrector 12.

The first detection is performed on the non-controlled electrode, and the second is implemented on the electrode which is going to be controlled, the control logic signal being available from the start of the real time period because its calculation is launched as soon as the detections made during the previous sampling period are available.

Referring to Figure 2, for example, when bs = +1, a first detection D I is implemented for the upper electrode 5.1 by applying a read pulse, then a second detection D2 is implemented on the bottom electrode by applying a second measurement pulse.

A control pulse is then supplied to the low electrode Vb by controlling the switch 9.

Referring to Figure 3, in the case where bs = -1, the pulse detection order is reversed.

This conditional detection makes it possible to limit the number of switchings of switch 9 and consequently to reduce consumption, the command being applied directly to the electrode to be controlled. Indeed, it is no longer necessary to operate the switch 9 which remains in the same state between the detection phase D2 and the control phase.

It should also be noted that the order of the detections is permuted at high frequency. The biases of electronic origin generated by the command pulses are thus transformed into noise due to the pseudo-random nature of the commands.

Indeed, any detection bias generates noise with a spectral appearance identical to that of the pulsed control signal bs. Thanks to conditional detection, the biases are greatly reduced, by being multiplied by the mean value of the control signal bs , only the noises remaining, for which the tolerances are increased.

Thus, conditional detection becomes pseudo-random both for the parasitic force it exerts and for the position measurement itself.

Furthermore, whereas in the state of the art the detection pulses were periodic and their spectrum consisted of lines, the permutation has a spectral spreading effect, which makes it possible to limit the excitation of high-frequency parasitic modes. and, above all, to regulate this excitation by the command, which becomes permanent and slowly variable or, in other words, less singular in frequency.

Furthermore, with reference to FIG. 4, according to another aspect, the second capacitive detection pulse and the control pulse form a common detection and control pulse.

In other words, the rising edge of the control signal is shifted to coincide with the falling edge of the second sense signal.

This mode of implementation makes it possible to carry out capacitive detection and control by means of the same pulse signal.

The electrostatic force applied to the pendulum, the direction of which is decided by the control signal bs, is constituted by the difference between the force exerted by the detection and control pulse and by the first detection signal applied to the other electrode.

The detection is carried out by means of the carrier of the detection and control signal, while the control is carried out in the baseband.

This mode of implementation still makes it possible, theoretically, to reduce consumption by a third, by reducing the switching of switch 9.

Likewise, a one-third reduction in the bias error due to the waveforms, which depends on the number of pulses during the sampling period, is obtained.

Finally, the scale factor error due to waveforms is completely removed.

Indeed, the acceleration equivalent to the applied force is written:

(eq. 4)

with: Ci the active capacitance, e the width of the air gap, m the mass of the pendulum, and ah2 and ab2 the mean square of the voltages applied to the high and low electrodes at each sampling period.

In an implementation with three detection and control pulses, the force applied was written:

in which
°ch corresponds to the bias and acb+°ch corresponds to the scale factor.

By implementing a first detection pulse n and a second detection and control pulse, the force applied is written:

Assuming that the time constants are short compared to the duration of the pulses, any waveform fault oerr2 will have the same effect whether it is on a detection pulse or a control pulse, superimposing the theoretical mean square.

By noting and the asymmetrical part of this defect, we have:

(eq. 7a)

(eq. 7d)

In a three-pulse implementation, we get:

(eq. 8) in which 3 and s 2 err corresponds to the bias and
corresponds to the scale factor.

With a two-pulse implementation, we have:

(eq 9) in which 2 and a rr corresponds to the bias and {oc,th ~ °d,th) corresponds to the scale factor.

It is therefore seen that the error caused by the control pulse is compensated by the error caused by the detection pulse.

The scale factor error is eliminated and should no longer be taken into account during preliminary calibration phases and which could be significant at certain operating temperatures and during sensor aging.

CLAIMS

1. Accelerometric sensor, comprising a housing (2), a pendulum (3) fixed to the housing, a movable electrode (5.3) carried by the pendulum and connected to a detection circuit (7), a first electrode (5.1) and a second fixed electrode (5.2) integral with the casing to form with the movable electrode two capacitors of variable capacitance as a function of a distance between the electrodes, and a control unit (8) configured to carry out detection operations to measure the capacitances variables of the capacitors and a control operation of the movable electrode as a function of the capacitances measured by applying a logic signal ( bs ) for controlling a switch (9) for selective connection of the fixed electrodes to an excitation circuit (6) delivering a control signal ( u ) to the fixed electrodes to maintain the pendulum in a predetermined position, characterized in that the control unit is configured to apply, at each sampling period, a first detection signal at one of said fixed electrodes chosen according to the logic level of the control signal, and a second detection signal at the other fixed electrode, the control signal ( u ) being applied to the electrode to which is applied the second detection signal.

2. Sensor according to claim 1, in which the first and the second detection signals are signals in the form of a square wave.

3. Sensor according to one of claims 1 and 2, wherein the switch (9) comprises a first input terminal to a reference potential provided by the excitation circuit (6), and a second input terminal at a zero potential to selectively connect said electrodes to the excitation circuit or to the zero potential.

4. Sensor according to any one of claims 1 to 3, wherein the excitation circuit (6) comprises a digital-to-analog converter connected to the switch (9) and controlled by the control unit (8).

5. Sensor according to any one of claims 1 to 4, wherein the detection circuit (7) comprises an amplifier stage having an input connected to the movable electrode (5.3) and an output connected to a digital-to-analog converter having a output connected to the control unit.

6. Sensor according to any one of claims 1 to 5, wherein the control unit (8) comprises a first estimator

(10) pendulum position input connected to the output of the detection circuit and an output connected to the negative input of a comparator

(11) having an output connected to an input of a corrector (12) having an output connected to a sequencer (13) and to a second estimator (14) having a first output connected to the positive input of the comparator and a second output providing an estimate of the acceleration.

7. Sensor according to one of claims 1 to 6, wherein the control unit (8) is configured to apply the second detection signal and the control signal by means of a common detection and control pulse.

8. Method for controlling an accelerometric sensor according to any one of claims 1 to 7, characterized in that it comprises the steps of:

-detection of the variable capacitances of the capacitors by applying to each sampling period a first detection signal to one of the fixed electrodes (5.1, 5.2) chosen according to the logic level of the control signal and a second detection signal to the other fixed electrode;

- control of the movable electrode (5.1) as a function of the capacities measured by applying a logic signal for controlling a switch (9) for selective connection of the fixed electrodes to an excitation circuit (6) delivering a control signal of so as to apply the control signal ( u ) to the electrode to which the second detection logic signal is applied.

9. Method according to claim 8, in which the second detection signal and the control signal are applied by means of a common detection and control pulse.