Mass deflection self-testing for MEMS accelerometers

Abstract

The invention relates to mass deflection self-testing of MEMS accelerometers, in particular to a method for self-testing a capacitive MEMS accelerometer comprising a first proof mass and a second proof mass, a state machine for implementing the self-testing method, and a MEMS device comprising at least one capacitive MEMS accelerometer, the first proof mass and the second proof mass are arranged in a dual differential configuration. The method applies a bias voltage pattern including alternating bias periods and readout periods. The self-test includes a deflected state and a return state, and a success or failure of each self-test state is determined.

Description

Technical Field

[0001] The present disclosure relates to methods, state machines, and MEMS devices. More particularly, the present disclosure relates to a method for self-testing mass deflection of a capacitive MEMS accelerometer, a state machine for controlling such a method, and a MEMS device including at least one capacitive MEMS accelerometer. Background Art

[0002] MEMS stands for microelectromechanical systems. It refers to a technology in which mechanical elements, sensors, actuators, and electronic devices are integrated on a common silicon substrate through microfabrication techniques. In essence, a MEMS device is a miniature mechanical device integrated with an electronic circuit to create a system with unique functions.

[0003] A MEMS accelerometer is a sensor that measures the acceleration force it experiences. These devices are commonly used in various applications such as automotive systems, consumer electronics, and industrial equipment.

[0004] A MEMS accelerometer includes one or more proof masses. A proof mass is the part of the accelerometer that moves in response to acceleration. It is typically a small suspended mass that undergoes displacement relative to the reference frame to which it is anchored via a suspension system when the device is subjected to an acceleration force. The suspension system is designed to hold the proof mass in place when not subjected to acceleration while allowing the proof mass to move in response to an external force. The suspension provides stability and accurate sensing of acceleration. A typical suspension system includes one or more anchor points and one or more springs that suspend the proof mass to the corresponding anchor points.

[0005] Typical MEMS accelerometers use capacitive sensing technology. One or more sensing capacitors are used to measure the displacement of the proof mass. When the proof mass moves, the distance between the proof mass and the fixed electrode changes and thus the capacitance changes, allowing the measurement of acceleration. The fixed electrode of the sensing capacitor is typically referred to as the stator. Although the movement of the proof mass in a typical MEMS accelerometer is linear rather than rotational, the proof mass can be referred to as a rotor.

[0006] A MEMS accelerometer may also include electronic circuits for processing the signals generated by the sensing capacitors. These circuits amplify, filter, and / or convert the capacitance change into output data, such as a voltage or digital signal corresponding to the measured acceleration.

[0007] Mass Deflection Self-Test (MDST) is a method used in MEMS accelerometers to check the functionality and integrity of the device. The mass deflection self-test technique involves applying a known force or acceleration to the MEMS accelerometer and then measuring the resulting deflection of the sense mass within the device. The application of the known force can be achieved electrically; in this case, it is referred to as biasing. Capacitive electrodes (also referred to herein as sense capacitors) can be used to bias by applying a voltage across the gap between the sense mass and one or more stators, which causes the deflection of the sense mass. By analyzing this deflection, it can be determined whether the output of the MEMS accelerometer matches the expected response. This process is typically automated and can be initiated at regular intervals or in response to specific events to ensure the continued reliability and accuracy of the MEMS accelerometer.

[0008] During the self-test process, the MEMS accelerometer can evaluate whether any degradation or failure has occurred within the device. Any differences between the expected and measured deflection values, failure to achieve a minimum required deflection magnitude, and / or failure to return to the rest position within a return time window before the return period expires and / or after deflection aborts can indicate potential problems such as mechanical fatigue, material degradation, or electronic faults. The mass deflection self-test can detect problems such as insufficient range of motion of the sense mass or situations where the capacitance signal value indicating the position of the sense mass does not return to the initial value prior to the self-test. Such errors in the capacitance signal value are sometimes referred to in the art as offset errors and can be caused by mechanical or electrical problems.

[0009] Early detection of such mechanical or electrical problems can help prevent failures of the accelerometer and ensure the accuracy of the accelerometer's measurements, thereby improving the overall reliability of systems that employ the accelerometer. Overall, the mass deflection self-test plays a crucial role in maintaining the performance and reliability of MEMS accelerometers, enabling them to provide accurate and consistent data for a variety of applications that require precise measurement of acceleration forces.

[0010] Description of Related Art

[0011] Capacitive microelectromechanical sensors have become part of many consumer devices and are also used in various safety-critical applications such as electronic stability control (ESC) in vehicles. Especially in safety-related applications, it is important to identify potential faults in the mechanical or electrical signal paths of capacitive sensors.

[0012] US9383860B2 discloses a circuit system for time-multiplexed detection of output signals from an output transducer of a capacitive MEMS sensor such that one acceleration channel is tested at a time.

[0013] A single differential MEMS accelerometer can also be tested in parallel mode when the clocks of the channels are identical to each other and each channel has an independent selection between the deflection phase and the non-deflection (return) phase during the bias period. However, as will be presented, robust self-testing of a dual differential MEMS accelerometer requires time multiplexing.

[0014] US10323957B2 discloses a circuit system for intermittent deflection bias generation. In this prior art, a bias voltage mode is used to deflect the sense mass, where the intermittent bias phase and the readout phase are repeatedly alternated.

[0015] EP3404422A1 discloses a high-voltage excitation circuit for generating a dual differential excitation circuit for a capacitive transducer (such as a MEMS capacitive transducer). A single high voltage is applied to the transducer as two consecutive stimuli with opposite polarities.

[0016] US20180143218A1 discloses a MEMS system and a method for performing a static friction test on the MEMS system by means of a set of monotonically increasing test voltages. An input multiplexer connects a test voltage across each input electrode to deflect the moving element. An output multiplexer circuit connects one or two moving elements to the input of a detector circuit during the test mode.

[0017] US20100251800A1 discloses a method for MEMS body self-testing. One of the MEMS bodies is urged to move against the movement that the MEMS body would naturally follow in response to an applied physical acceleration, and the other MEMS bodies are urged to move in accordance with the movement that the other MEMS bodies would naturally follow in response to the same applied physical acceleration. SUMMARY OF THE INVENTION

[0018] The following embodiments are exemplary. Although the specification may refer to "one", "a", or "some" embodiments, this does not necessarily mean that each such reference is to the same embodiment, or that the features apply only to a single embodiment. Individual features of different embodiments can be combined to provide additional embodiments.

[0019] It should be understood that the device is configured to perform the corresponding method, but in some cases only the device or only the method is described.

[0020] The aim is to provide methods and devices for solving the problem of self-testing differential MEMS accelerometers, especially dual differential MEMS accelerometers. The aim of the present invention is achieved by a method according to one aspect of the present invention. The aim of the present invention is also achieved by a state machine according to another aspect of the present invention and a MEMS device according to yet another aspect of the present invention.

[0021] Preferred embodiments of the present invention are disclosed in the dependent claims.

[0022] According to a first aspect, there is provided a method for self-testing a capacitive MEMS accelerometer, the capacitive MEMS accelerometer including a first proof mass and a second proof mass. The first proof mass and the second proof mass of the capacitive MEMS accelerometer are arranged in a double-differential configuration. The first proof mass is associated with two sensing capacitors (C1, C2), the sensing capacitors being configured to capacitively detect the position of the first proof mass relative to a first stator and a second stator, and the second proof mass is associated with two sensing capacitors (C3, C4), the two sensing capacitors being configured to capacitively detect the position of the second proof mass relative to a third stator and a fourth stator. The first stator is electrically coupled to the third stator, and the second stator is electrically coupled to the fourth stator. The method includes: deflecting the first proof mass and the second proof mass in mutually opposite first directions by applying a first bias voltage across the sensing capacitor (C1) of the first stator and across the sensing capacitor (C3) of the third stator during a bias period of a bias voltage mode. The bias voltage mode is a repeating mode including intermittent bias periods and readout periods, wherein a deflection bias voltage is selectively applied during the bias periods and no deflection bias voltage is applied during the readout periods, and wherein the bias voltage mode repeats at a frequency above the resonant frequency and / or -3dB bandwidth of the capacitive MEMS accelerometer. The method includes: obtaining first output data representing a proof mass equivalent capacitance signal (C1-C2) of the first proof mass and second output data representing a proof mass equivalent capacitance signal (C3-C4) of the second proof mass during each readout period, wherein the first output data and the second output data are obtained at mutually different times during the readout period.

[0023] According to some embodiments, the method includes: obtaining third output data representing a sum capacitance value (C1+C2) of the first proof mass and fourth output data representing a sum capacitance value (C3+C4) of the second proof mass during each readout period, wherein each of the first output data, the second output data, the third output data, and the fourth output data is obtained during the readout period, and calculating a first normalized differential capacitance value by dividing the first output data by the third output data, and calculating a second normalized differential capacitance value by dividing the second output data by the fourth output data.

[0024] According to some embodiments, the method further includes: when it is determined that the deflection of either the first detection mass block or the second detection mass block has reached a first minimum required deflection amplitude through deflection in the bias voltage mode based on the first output data and the second output data, or based on the first normalized differential capacitance value and the second normalized differential capacitance value, or when it is determined that a first predetermined maximum deflection period has expired, abort the application of the first bias voltage. After the abort, apply the bias voltage mode without applying any deflection bias voltage during the bias period to return the first detection mass block and the second detection mass block towards the rest position.

[0025] According to some embodiments, the method further includes: when applying the bias voltage mode without applying any deflection bias voltage during the bias period, obtain the first output data and the second output data, and optionally obtain the third output data and the fourth output data, and based on the average value of the first output data and the second output data, or optionally based on the average value of the first normalized differential capacitance value and the second normalized differential capacitance value, determine that the first detection mass block and the second detection mass block have reached the rest position, or determine the expiration of a first return period.

[0026] According to some embodiments, the method further includes: when it is determined that the first detection mass block and the second detection mass block have reached the rest position, or when the first return period has expired, during the bias period, apply a second bias voltage across the sensing capacitor (C2) of the second stator and across the sensing capacitor (C4) of the fourth stator to deflect the first detection mass block and the second detection mass block in a mutually opposite second direction opposite to the corresponding first direction, and obtain the first output data and the second output data, and optionally obtain the third output data and the fourth output data, and calculate the first normalized differential capacitance value and the second normalized differential capacitance value. The method further includes: when it is determined that the deflection of either the first detection mass block or the second detection mass block has reached a second minimum required deflection amplitude based on the first output data and the second output data, or optionally based on the first normalized differential capacitance value and the second normalized differential capacitance value, or when a second predetermined maximum deflection period has expired, abort the application of the second bias voltage. After the abort, apply the bias voltage mode without applying any deflection bias voltage during the bias period to return the first detection mass block and the second detection mass block towards the rest position. The method includes: while applying the bias voltage mode without applying any deflection bias voltage during the bias period, obtain the first output data and the second output data, and optionally obtain the third output data and the fourth output data, and based on the average value of the first output data and the second output data, or based on the average value of the first normalized differential capacitance value and the second normalized differential capacitance value, determine that the first detection mass block and the second detection mass block have reached the rest position, or determine the expiration of a second return period.

[0027] According to some embodiments, the self-test includes one or two deflection portions, each deflection portion including deflecting the first detection mass and the second detection mass by applying a respective first bias voltage or second bias voltage. If it is determined that the deflection of the first detection mass or the second detection mass has reached a respective first minimum required deflection amplitude or second minimum required deflection amplitude before the expiration of the respective first predetermined maximum deflection period or second predetermined maximum deflection, and the application of the respective first bias voltage or second bias voltage during the bias period in the bias voltage mode is aborted, the deflection portion of the self-test is considered to have passed. If the application of the respective first bias voltage or second bias voltage is aborted when the predetermined maximum deflection period expires, the deflection portion of the self-test is considered to have failed.

[0028] According to some embodiments, the method includes: one or two return portions, during which the first detection mass and the second detection mass are returned toward the rest position by not applying any deflection bias voltage during the bias period. If the first detection mass and the second detection mass have reached the rest position within a respective first predetermined maximum return period or second predetermined maximum return period after the application of the respective first bias voltage or second bias voltage is aborted, or if the first detection mass and the second detection mass have reached the rest position within a respective first return time window or second return time window, the return portion of the self-test is considered to have passed. If it is determined that the predetermined first maximum return period or second maximum return period has expired and the first detection mass and the second detection mass are not determined to have reached the rest position, or if it is determined that the first detection mass and the second detection mass are not determined to have reached the rest position within the respective first return time window or second return time window, the return portion of the self-test is considered to have failed.

[0029] According to some embodiments, the method further includes: after the capacitive MEMS accelerometer has entered the self-test mode or the channel of the multi-channel capacitive MEMS accelerometer to be self-tested has changed, waiting for a predetermined waiting period, applying a bias voltage mode to the capacitive MEMS accelerometer in the self-test mode, and after determining that the waiting period has expired, applying the bias voltage mode without applying any deflection bias voltage during the bias period, and obtaining a plurality of samples of the first output data and the second output data during the readout period, and optionally also obtaining a plurality of samples of the third output data and the fourth output data. The method further includes: determining an offset value by averaging the plurality of samples of the first output data and the second output data during the readout period and averaging the plurality of first normalized differential capacitance values and the second normalized differential capacitance values, and based on the offset value, determining at least one threshold or threshold window for determining whether the first detection mass and / or the second detection mass has reached the rest position.

[0030] According to some embodiments, if any part of the self-test is considered to have failed, the self-test can be repeated a predetermined number of times.

[0031] According to some embodiments, when repeating the self-test, the self-test is restarted from the step of applying the bias voltage mode without applying the bias voltage during the bias period to re-determine the offset value.

[0032] According to a second aspect, there is provided a state machine configured to control the mass deflection self-test of a capacitive MEMS accelerometer. The self-test is configured to apply a bias voltage pattern that is a repeating pattern including intermittent bias periods and readout periods, where a deflection bias voltage is selectively applied during the bias periods and no deflection bias voltage is applied during the readout periods, and where the bias voltage pattern repeats at a frequency above the resonant frequency and / or -3dB bandwidth of the capacitive MEMS accelerometer. The state machine includes, in execution order: i) a wait state for waiting a predetermined wait period after the capacitive MEMS accelerometer has entered the self-test mode or the channel to be self-tested has changed, ii) an offset determination state for determining an offset value by averaging multiple samples of first output data and second output data during the readout period of the bias voltage pattern when applying the bias voltage pattern without applying any deflection bias voltage during the bias periods, iii) a first deflection state for deflecting a first sense mass and a second sense mass in mutually opposite first directions by applying a first bias voltage across the sense capacitor (C1) of a first stator and across the sense capacitor (C3) of a third stator during the bias period of the bias voltage pattern, where the first stator is electrically coupled to the third stator, iv) a first return state for returning the first sense mass and the second sense mass towards the rest position by not applying any deflection bias voltage across the sense capacitor during the bias period, v) a second deflection state for deflecting the first sense mass and the second sense mass in mutually opposite second directions opposite to the respective first directions by applying a second bias voltage across the sense capacitor (C2) of a second stator and across the sense capacitor (C4) of a fourth stator, where the second stator is electrically coupled to the fourth stator, vi) a second return state for returning the first sense mass and the second sense mass towards the rest position by not applying any deflection bias voltage across the sense capacitor, and vii) an end state for determining whether the results of the first deflection state and the second deflection state and the first return state and the second return state are determined to be a failure or a pass, and if any of these states is a failure, optionally repeating the self-test starting from the offset determination state.

[0033] According to a third aspect, there is provided a MEMS device including at least one capacitive MEMS accelerometer. The MEMS device further includes circuitry implementing the state machine.

[0034] The present invention is based on the concept of deflecting a differentially operated proof mass of a MEMS accelerometer using an intermittent bias voltage across its sensing capacitors at a frequency above the accelerometer's normal operating bandwidth, and reading output data between each two consecutive bias voltage application periods. While the two proof masses are simultaneously deflected by a bias voltage pattern applied to the sensing capacitors coupled to an electrically common stator, the output data of the proof masses are measured independently of each other. The deflection amplitude of the proof masses is controlled in real time based on a determination of the current deflection amount at each readout period to ensure that the maximum allowable deflection amplitude never causes mechanical contact between the proof masses and the stator. The deflection of the proof masses is preferably repeated in two opposing deflection directions corresponding to positive and negative accelerations. A state machine can control the self-test process.

[0035] The present invention has the following advantages: it provides a robust self-test with high tolerance to external acceleration during the self-test. The method also ensures that no mechanical contact between the electrodes of the capacitive MEMS accelerometer occurs due to the self-test, even when the self-test is performed in the presence of external acceleration. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In the following, the present invention will be described in more detail in conjunction with preferred embodiments with reference to the accompanying drawings, in which:

[0037] Figure 1A and Figure 1B A dual differential MEMS accelerometer is shown.

[0038] Figure 2 It is a simplified timing diagram.

[0039] Figure 3 The states of the dual differential MDST state machine are shown.

[0040] Figure 4 is a schematic diagram of an exemplary accelerometer mass deflection self-test system including a mass deflection self-test state machine (MDST STM).

[0041] Figure 5 Shown are simulated outputs representing a three-channel self-test of the three different axes (X, Y, Z) of a three-axis dual-differential MEMS accelerometer.

[0042] Figure 6 The digitized proof mass equivalent capacitance signal for an actual MDST measurement performed on a tri-axial accelerometer is shown.

[0043] Figure 7 Measured MDST data is shown for a case where the measured polarity is considered to fail on both axes. DETAILED DESCRIPTION

[0044] Figure 1A and Figure 1B schematically shows a double-differential MEMS accelerometer 10. Figure 1A schematically shows the key mechanical elements of the double-differential MEMS accelerometer 10. Various mechanical layouts are applicable. In this example, there are two sense masses 101, 102 suspended by springs 110 to anchor points 11, 12. In this example, the springs 110 enable the sense masses 101, 102 to move in a plane in the directions of positive and negative acceleration as indicated by the arrows above and below the MEMS accelerometer. The stators 105A, 105B, 106A, 106B are fixed such that movement in the direction of positive or negative acceleration causes a change in the capacitance C1, C2, C3, C4 of the sense capacitors formed by or attached to the sense masses 101, 102 and the stators 105A, 105B, 106A, 106B. This change in capacitance indicates the magnitude of the acceleration detected in either of the two indicated directions. As is known in the art, bias voltages are fed to the accelerometer during normal operation of the accelerometer for measuring acceleration. However, such bias voltages are fed symmetrically such that they do not cause deflection of the sense masses and can be referred to as non-deflecting bias voltages to distinguish them from deflecting bias voltages, which are used to intentionally deflect the sense masses for testing purposes.

[0045] For a dual-differential MEMS accelerometer 10 having two mechanically separated but electrically coupled sense masses 101, 102, compared with an accelerometer having a single sense mass, channel control during MDST becomes more complex. This is because in a dual-differential MEMS accelerometer, the two sense masses 101, 102 have stators 105A, 105B that are at least electrically coupled to sense capacitors connected to the two sense masses. In an actual MEMS design, stator 105A and 106A and / or stator 105B and 106B may also be mechanically coupled. When a bias voltage is fed to the electrically coupled stators 105A, 106A or stators 105B, 106B to deflect the sense masses 101, 102, both sense masses 101, 102 will be deflected simultaneously. In this context, the electrically coupled stators may also be referred to as common stators. In this context, stator 105A is referred to as the first stator, stator 105B is referred to as the second stator, stator 106A is referred to as the third stator, and stator 106B is referred to as the fourth stator. The first stator 105A and the third stator 106A are electrically coupled to each other, thus forming a first common stator, and the second stator 105B and the fourth stator 106B are electrically coupled to each other, thus forming a second common stator. The self-test deflection motion is different from the motion of the sense mass when it is subjected to acceleration. During the self-test deflection, the sense masses 101, 102 move in an anti-acceleration mode, in which the motion of the sense mass is not visible in the output signal used to detect acceleration in the acceleration detection mode of a fully differential accelerometer. On the other hand, an external acceleration is not visible in the average output signal obtained for detecting the deflection of the dual-differential MEMS accelerometer in the sense mass deflection self-test operation mode. Since the two types of output signals are not correlated with each other, the presence of an external acceleration does not invalidate the self-test.

[0046] The electrical equivalent circuit of the dual-differential MEMS accelerometer 10 is visualized in Figure 1B as can be seen from Figure 1BAs can be seen, the first common stator, in other words, the first stator and the third stator (105A and 106A), are electrically identical, and the second common stator, in other words, the second stator and the fourth stator (105B and 106B), are electrically identical, in other words, at the same potential. Depending on the mechanical construction of the MEMS accelerometer, the electrically coupled stators may also be mechanically coupled and formed of components of the same rigid stator structure. When the dual-differential MEMS accelerometer 10 is subjected to a positive acceleration, the corresponding fully differential capacitive acceleration detection signal is (C3 - C1)+(C2 - C4). When the electrically coupled stators (105A and 106A; 105B and 106B) are set to a high voltage for MDST purposes and all other electrodes are set to ground voltage, the resulting electrostatic force will generate a capacitive signal corresponding to the transfer function (C1 + C3)-(C2 + C4). Thus, the electrostatic deflection of the sensing masses 101, 102 does not generate a signal equivalent to the acceleration detection signal caused by an external acceleration. This difference affects the MDST operation in various ways but can also be exploited.

[0047] To detect the effect of the deflection during MDST, at least two output data need to be detected. If only the normal fully differential capacitive acceleration detection signal (C3 - C1)+(C2 - C4) is detected, this will ignore any external acceleration and offset that occur during MDST. In the worst case, if MDST is operating in the presence of a high external acceleration, the external acceleration and / or offset may cause an unwanted mechanical contact between the sensing capacitors arranged between the sensing masses and the stators.

[0048] The MDST method applies an intermittent reflection bias generation scheme in which a bias voltage pattern is used to deflect the sensing masses, in which the intermittently repeated bias phases alternate with the readout phases. The bias phases and the readout phases are repeated at a rate, for example, between 100 times and 200 times the bandwidth of the sensor element. Each deflection bias voltage pulse gradually changes the position of the sensing masses away from their rest position (unbiased position), and the position or the change in position (also referred to as deflection) can be detected in real time during the readout period. During the readout period, a non-deflection bias voltage is applied for readout. The rest position is determined at the start of MDST before any deflection, and the return of the deflected sensing masses to the rest position is determined by comparison with the determined rest position.

[0049] The biasing phase and the readout phase are repeated at a rate that, for example, extends from 10 times to 1000 times the resonant frequency and / or -3dB bandwidth of the sensor element. The acceleration sensor is typically designed to have a low-pass characteristic frequency response. Thus, -3dB refers to the frequency at which the low-pass gain is reduced by 3dB compared to the gain at ~0Hz. When the proof mass moves significantly during a single biasing phase, a low repetition rate in the range of 10 times only gives limited detection time and / or amplitude resolution, while a high repetition rate in the range of 1000 times or more of the bandwidth may overly increase the data rate. The high repetition rate also leads to the following challenges: the technical challenge of implementing a suitable (very fast and stable) high-voltage source to be used as the bias voltage and the challenge in the dynamic adjustment of the bias voltage. A repetition rate in the range between 100 times and 200 times the resonant frequency of the sensor element and / or above the -3dB bandwidth may be most suitable for this purpose, which has good resolution but avoids becoming computationally unnecessarily burdensome. To prevent external acceleration from adversely affecting the deflection of the proof mass for self-test purposes, the bias voltage is preferably set slightly higher (i.e., 10% to 30% higher) than the estimated voltage required for the actual desired self-test deflection. However, in the case where the bias voltage is applied only during a short biasing period, the inertia of the proof mass causes it to deflect only a part of the maximum deflection in each bias voltage period. In other words, the concept of short, repeated bias voltage pulses is that each bias voltage pulse gradually changes the position of the proof mass away from the rest position (unbiased position) of the proof mass, and the position or change in position of the proof mass can be detected in real time during the readout period between two biasing phases. The change in the position of the proof mass relative to the rest position is referred to as deflection herein.

[0050] To avoid the problems depicted above, it is preferable that after each deflection, the output data from each proof mass is read separately. In other words, the capacitance at the proof mass 101 and the capacitance at the proof mass 102 are read separately. The capacitance at the proof mass 101 generates a first proof mass equivalent capacitance value (C1 - C2), and the capacitance at the proof mass 102 generates a second proof mass equivalent capacitance value (C3 - C4). During the entire deflection process, these two readings are performed between every two consecutive deflection periods. This process is shown in the simplified timing diagram as Figure 2 shown in Figure 2An example is shown in which the stators 105A, 106A are excited using repeated high voltage bias periods. During the readout period, the capacitances at the two sense masses 101, 102 are detected, one at a time. During the readout period, the readout signal, shown as a short pulse, is fed into one sense mass at a time. This enables the capacitance of the respective single sense mass into which the readout signal is fed to be detected. The read pulses are fed symmetrically into the sense masses and thus do not deflect. By separately reading the sense mass equivalent capacitance values during the readout period, the differential signal levels of each of the sense masses 101, 102 can be tracked and mechanical contact caused by MDST deflection can be prevented; when it is determined that the desired maximum deflection has been reached, the deflection process can be controlled by stopping any further MDST deflection.

[0051] When the fully differential capacitance output signal is (C3 - C1)+(C2 - C4), the acceleration in fully differential mode is determined as the difference between the two sense mass equivalent capacitance values (C1 - C2) and (C3 - C4). In MDST mode, in order to detect the deflection, the MDST mode capacitance output signal is determined as the sum: (C1 - C2)+(C3 - C4), which may be referred to as the sum mode capacitance value. Additionally, the sum capacitance values (C1 + C2) and (C3 + C4) may be detected for normalization purposes. According to some embodiments, obtaining the two sum capacitance values may require adding two additional read instances during the readout period between each bias period to enable each capacitance value to be read at a different time from any other capacitance value read, but according to some embodiments, the sum capacitance values may be obtained while obtaining at least one of the sense mass equivalent capacitance values. The sum capacitance values are generally not affected by any movement of the sense masses, unless there is some non-linearity or asymmetry between the sense masses therein. The sum capacitance values can be used to calculate the normalized differential capacitance values in post-processing. The normalized differential capacitance values are determined by the functions (C1 - C2) / (C1 + C2) and (C3 - C4) / (C3 + C4).

[0052] Since normalization allows for higher linearity, the normalized differential capacitance values are particularly useful for so-called gap modulation capacitive transducers. As is known in the art, a gap modulation capacitive transducer refers to a plate capacitor in which the gap between the plates is changed. In an exemplary gap modulation capacitive transducer, the capacitance (C1 - C2) is inherently non-linear, but it can be linearized (normalized) by determining the normalized differential capacitance value (C1 - C2) / (C1 + C2). If the repetition rate of the deflection is kept constant, adding additional readouts to the readout period reduces the time available for each individual readout.

[0053] When measuring only a single sense mass at a time, the sensitivity of the measurement is half compared to the fully differential capacitance output signal (C3 - C1)+(C2 - C4). The same applies to the measurement of the sum capacitances (C1 + C2) and (C3 + C4) in the MDST deflection, which also have a sensitivity that is half of the sum capacitance (C1 + C2 + C3 + C4) that would be in normal operating mode.

[0054] Many MEMS accelerometer devices include more than one accelerometer, for example, for measuring accelerations in different directions. A typical MEMS device includes, for example, an accelerometer for detecting accelerations along two mutually orthogonal sense axes, or an accelerometer for detecting accelerations along three mutually orthogonal sense axes (e.g., the x-axis, y-axis, and z-axis). One or more of these accelerometers can be fully differential.

[0055] The clocking scheme for the MDST of a fully differential MEMS accelerometer using intermittent deflection phases and two readout phases can also be used for the MDST of a single differential MEMS accelerometer. In such a case, the same single sense mass equivalent capacitance value (C1 - C2) can be detected at the two readout phases, or the other readout phase can be set to zero, in other words, the other readout phase can be omitted.

[0056] Each read sample is digitized, and the MDST is controlled based on the measurement data by means of a digital state machine. Figure 2 The timing scheme shown is also valid for the return phase, i.e., when the sense mass is allowed to stabilize to an un - deflected position, in other words, to a rest position where the sense mass is not intentionally deflected by the deflection bias voltage. In this case, the deflection bias voltage applied to the stators 105A, 106A is set to ground during the HV bias phase, or more specifically, the deflection bias voltage across the sense capacitors C1, C2, C3, and C4 is zeroed, such that the deflection MDST bias across the capacitors is removed and the bias across the sense capacitors returns to the non - deflected low - voltage symmetric bias.

[0057] Mass Deflection Self - Test State Machine

[0058] A state machine (STM) for MDST (MDST STM) monitors the deflection amplitude caused by the repetitive biasing phase. The deflection amplitude is detected during the readout phase, and when the predefined target deflection amplitude is reached, in other words, the minimum required deflection amplitude is reached, MDST stops detecting the mass block deflection process. The repetitive biasing voltage pattern makes the test easier than applying a single preset test voltage because the value of the applied test voltage does not have to be precise. As is known in the art, the pull-in voltage refers to the biasing voltage sufficient to apply an electrostatic force to the detection mass block, which exceeds the mechanical force generated by the spring suspending the detection mass block, such that if the pull-in voltage is allowed to affect the detection mass block long enough, the detection mass block will be pulled to the maximum deflection, which typically causes the detection mass block or its movable electrode to come into contact with a so-called buffer (in other words, a structure that mechanically limits the maximum deflection amplitude). By applying a biasing voltage exceeding the pull-in voltage but only for a short period of time, the deflection of the detection mass block can be controlled such that it does not exceed the maximum value that would cause the detection mass block to inadvertently come into contact with another structure.

[0059] The state machine for MDST in a dual-differential MEMS accelerometer is different compared to the conventional single-differential MDST disclosed in US10323957 B2. The state machine can be implemented by a processor, a field-programmable gate array (FPGA) or implemented as an ASIC.

[0060] It can be assumed that in any Nyquist-type or oversampling-type ADC, for example, by using a successive approximation analog-to-digital converter (SAR ADC) or a Σ-Δ ADC and decimation filtering, the equivalent capacitance values of the detection mass block (C1 - C2) and (C3 - C4) are digitized. The first digitized detection mass block equivalent capacitance signal and the second digitized detection mass block equivalent capacitance signal are labeled herein as A1_DEC = ~C1 - C2 and A2_DEC = ~C3 - C4. The polarity of the digitized detection mass block equivalent capacitance signal is always controllable. For example, the first digitized detection mass block equivalent capacitance signal A1_DEC can be configured as -(C1 - C2) to adjust the deflection data from different detection mass blocks to have the same polarity with each other.

[0061] Figure 3Shows the main states of a dual-differential MDST state machine (STM) according to some embodiments. The STM shows the MDST test for a single channel. The STM is always updated when new data is available. When testing more than one channel, at least some of the tests can run in parallel, in which case each channel has its own MDST STM. If the tests are run serially, a single MDST STM may be sufficient to control the self-test process itself. If more channels are to be tested, at least one upper-level STM may be required to ensure the proper functioning of the self-test, for example if one or more channels are disabled or one or more channels fail the self-test. The MDST STM implements a relatively simple implementation for determining when to interrupt the self-test or its phases, for generating self-test status information from each self-test phase, and for measuring the time period for each self-test phase.

[0062] The first state is the initial stable waiting state 300. The STM remains in this waiting state 300 for a predefined stable waiting period. For example, a stable waiting period may be required if the system clock changes when the channel under test changes or when entering the MDST mode from the normal operation mode. In the waiting state 300, a non-deflection bias voltage may be applied, but no deflection bias voltage is applied, and no measurement data is obtained, or any obtained measurement data is discarded because it is useless.

[0063] In this context, the term "self-testing a channel" refers to a self-test performed by deflecting (one or more) sense mass blocks by using a pair of electrically coupled stators (also referred to as common stators), where the sense mass blocks are configured to measure acceleration about a single axis. In a dual-differential accelerometer, biasing a pair of electrically coupled stators causes the deflection of two sense mass blocks because the two sense mass blocks are deflected by these electrically coupled stators. For each self-test, as will be described, at least two types of output data are detected. Output data refers to the detected output capacitance values that are ultimately digitized. Examples of output data are the first sense mass block equivalent capacitance signal and the second sense mass block equivalent capacitance signal, or the first digitized sense mass block equivalent capacitance signal and the second digitized sense mass block equivalent capacitance signal after digitization. In the following examples, the first sense mass block equivalent capacitance signal and the second sense mass block equivalent capacitance signal are used as output data. Depending on the type of transducer used, the sense mass block equivalent capacitance signal can be, for example, a current or voltage signal, while the digitized signal can be any desired type of digital signal.

[0064] When the predefined stable waiting period expires, the STM moves to the offset determination state 302. According to a preferred embodiment, the offset data is obtained by averaging the output data. The offset data can be obtained by averaging the digitized detection mass equivalent capacitance signal. The averaging is performed within a predefined period. The predefined period can be determined as, for example, a number of clock cycles, or can be determined as a predefined or minimum number of samples obtained before any detection mass deflection is performed. For example, if the MEMS time constant τ of the acceleration sensor = 1 / w(-3db), the averaging period is preferably between 1*τ and 100*τ, and most preferably around 10*τ. Since averaging effectively removes the influence of external interference, the longer the averaging period, the more robust the offset value. On the other hand, an overly long averaging period increases the time required to perform a test run.

[0065] The purpose of the offset data is to detect whether the detection mass returns to its expected rest position after deflection; after deflection and return, the offset should be the same as before its deflection, or at least within a predefined offset window relative to the offset determined when the MDST deflection is initiated. In this context, the offset refers to the output data values received from the channel under test before the deflection voltage is applied and after the return period. The offset indicates that the detection mass is not completely at its expected and / or assumed rest position, and / or that there is some source of offset at any point in the signal path even when the detection mass is considered at rest, and / or that crosstalk occurs between adjacent samples, and / or that crosstalk occurs between different MEMS sensors, between MEMS sensor package structures, between bond wires, or between channels included in the MEMS device. For example, if there is no digital trimming of the output signal received from the MEMS accelerometer, the data received from the ADC or decimation filter may include parasitic offsets that do not depend on the position of the detection mass. In the normal acceleration measurement operation mode, any source of parasitic offset in the MEMS accelerometer is typically compensated such that the digital output signal received at 0g acceleration is equal to 0 LSB. The MDST mode significantly changes the operation of the MEMS accelerometer relative to the normal acceleration detection mode, and thus it is not recommended to apply the digital offset compensation determined for the normal acceleration detection mode during MDST. For example, in the MDST mode, the output data rate changes, the accelerometer output is detected at different times on different channels, and the clock scheme is different from the normal acceleration detection mode. Any change in the analog clock and MEMS bias may result in different parasitic offsets compared to the normal accelerometer operation mode.

[0066] The calculated offset is as follows:

[0067] OFFSET = average(A1_DEC_all_collected_data + A2_DEC_all_collected_data), where "all_collected_data" refers to all samples of output data such as digitized sense mass equivalent capacitance signals obtained during the period for determining the offset to be used for MDST before applying the bias voltage. The offset is determined based on the average of the output data obtained from the two sense masses.

[0068] When the offset determination state 302 is completed, the STM moves to the first deflection state 304, during which the sense masses are deflected in the positive or negative deflection direction by applying an intermittent repetitive bias voltage to one of the common stators of the differential arrangement of sense masses and obtaining at least two desired types of output data. The deflection bias voltage applied during the bias period of the first deflection state 304 is referred to as the first bias voltage. In this example, there are two types of output data, namely A1_DEC and A2_DEC, in other words, the first digitized sense mass equivalent capacitance signal and the second digitized sense mass equivalent capacitance signal.

[0069] The duration of the first deflection state 304 depends on two different conditions. If either of these conditions is considered to be met, the STM moves (305) to the next state, i.e., the first return state 306, in which neither the first bias voltage nor any other deflection bias voltage is applied during the bias period. The first end condition is that the desired deflection threshold of A1_DEC or A2_DEC is reached within the maximum allowable deflection period, in which case the first deflection test result is "pass". Regarding the first deflection state 304, the deflection threshold can be referred to as the first predetermined minimum deflection. The second end condition is that a timeout occurs, which results in the first deflection test result being considered "failed" because the sense mass may not have deflected the desired amount during the maximum allowable deflection period.

[0070] The deflection threshold can be determined using the absolute value, i.e., absolute(A1_DEC) or absolute(A2_DEC). The absolute value depends on which sense mass is being tested. Alternatively, signed A1_DEC or A2_DEC data can be used to perform the deflection threshold comparison. In the case of using signed A1_DEC or A2_DEC data for threshold comparison, it must be ensured that the obtained deflection data has the same polarity in the digitized sense mass equivalent capacitance signals A1_DEC and A2_DEC, and the deflection threshold polarity should be set according to the deflection polarities of A1_DEC and A2_DEC so that the signal path polarity can be verified.

[0071] When the first deflection state 304 is completed, the STM proceeds to the first return state 306. The STM remains in the first return state 306 until a return state condition is met. The return state condition can be, for example: the average of the digitized sense mass equivalent capacitance signals A1_DEC and A2_DEC of a single sample, i.e., (A1_DEC + A2_DEC) / 2, or the average of the digitized sense mass equivalent capacitance signals A1_DEC and A2_DEC calculated from multiple consecutive samples on a sliding window, i.e., (A1_DEC + A2_DEC) / 2, is determined to fall within a predetermined return threshold window relative to a previously determined offset value, in which case the first return test result is "pass". The number of samples can be only one sample for each digitized sense mass equivalent capacitance signal, or multiple consecutive samples for each digitized sense mass equivalent capacitance signal, where the number of consecutive samples is between 2 and 200. Also, the long sampling period increases the total length of the MDST process. If the STM remains in the first return state 306 until the maximum return period expires, i.e., the first return timeout elapses, then the first return test result is "fail" because the sense mass has not returned to the rest position within the maximum predetermined first return period determined by the first return timeout. Instead of the first return timeout, a first return time window can be determined for the first return state 306 such that if the first return state condition is achieved within the first return time window, the first return test is considered "pass", and if the first return state condition is achieved outside the first return time window, the first return test is considered "fail". Preferably, then the end time of the first return time window will coincide with the end of the first return period. By determining the return time window instead of only observing the end point of the maximum return period, an error where the (one or more) sense masses return to the rest position too quickly after deflection abort can be detected.

[0072] Preferably, the return threshold window is determined such that it includes the offset value and can allow a small deviation from the offset value. In the first deflection state 304, the timeout refers to the expiration of the first deflection period starting when entering the first deflection state 304.

[0073] During the first return state 306, when the differential sensor is fully symmetric, the average value is independent of the external acceleration. The threshold can be in the form of a value window, and if the average value of the digitized sense mass equivalent capacitance signal returns with a value within the value window before reaching the timeout, the first return test is considered to be "passed". In the first return state 306, the timeout refers to the expiration of the first maximum return period starting when entering the first return state 306. After the first return state 306 has been completed (passed or failed), the STM proceeds to the second deflection state 308, which is substantially similar to the first deflection state 304, except that the deflection is caused in the opposite direction: if the first deflection state 304 is used to deflect the sense mass in the positive acceleration direction along the axis, the second deflection state 308 is used to deflect the sense mass in the negative acceleration direction along the same axis, or vice versa. The deflection bias voltage applied during the bias period of the second deflection state 308 is referred to as the second bias voltage. To cause a deflection in the direction opposite to the deflection in the first deflection state, the second bias voltage is applied to a capacitor of the mutually coupled stator that is different from the first bias voltage. For example, if capacitors C1 and C3 are used for the deflection in the first deflection state 304, capacitors C2 and C4 are used for the deflection in the second deflection state 308. Since the deflection direction is changed, it is necessary to change the polarity of the deflection threshold, unless the absolute value is used to determine the amount of deflection, in which case the polarity of the deflection threshold can be omitted, but on the other hand, the information about the correctness of the test polarity will also be omitted accordingly.

[0074] Similar to the first deflection state 304, the second deflection state 308 can end when one of two end conditions is reached. These are preferably the same end conditions as those used in the first deflection state 304. The first end condition is that A1_DEC or A2_DEC reaches the desired deflection threshold, in which case the second deflection test result is "passed". Regarding the second deflection state 308, the deflection threshold can be referred to as the second predetermined minimum deflection. The second end condition is that a timeout occurs, which results in the second deflection test result being considered "failed" because the sense mass cannot deflect the desired amount during the maximum allotted time period (although a deflection of the desired amount is expected), and the timeout occurs. In the second deflection state 308, the timeout refers to the expiration of the second deflection period starting when entering the second deflection state 308. Since a circuit system mainly used for detecting acceleration is used to measure the deflection, the first end condition can be determined as an equivalent acceleration value, although the self-test deflection is not actually caused by acceleration. In an accelerometer designed for use in a ground vehicle, an exemplary minimum deflection of the equivalent acceleration value can be 6g or 8g, but depending on the application for which the accelerometer is designed, the minimum deflection can be any value between 0.1g and 1000g.

[0075] According to some embodiments, the first predetermined minimum deflection and the second predetermined minimum deflection are equal, but are determined in opposite directions. The first predetermined minimum deflection and the second predetermined minimum deflection may also have mutually different values.

[0076] According to some embodiments, the first deflection period and the second deflection period are equal, but may also have mutually different values.

[0077] When the first end condition used in determining the first deflection state 304 and / or the second deflection state 308 is satisfied when one of the channels A1_DEC or A2_DEC reaches a desired deflection threshold, the following additional condition (not shown) may be determined: Determine the deflection of the other channel that did not result in the first deflection being satisfied to confirm that the other sense mass is also deflected. When the desired deflection threshold is reached by the first channel of the two channels, the threshold for the "pass" result for the deflection at the other channel may be set lower than the first end condition set for the first deflection state 304 and / or for the second deflection state 308. For example, the threshold for the second deflection may be 1 g to 8 g less than the first end condition.

[0078] When one of the end conditions of the second deflection state 308 is reached, the STM proceeds to the second return state 310, which corresponds to the first return state 306. During the bias period of the second return state 310, no second bias voltage or any other deflection bias voltage is applied to the sense mass. The STM remains in the second return state 310 until the return state condition is satisfied. Preferably, the return state condition in the second return state is the same as the return state condition in the first return state. The return state condition can be to determine that the average value of, for example, the digitized sense mass equivalent capacitance signals A1_DEC and A2_DEC, i.e., (A1_DEC + A2_DEC) / 2, falls within a predetermined return threshold window relative to a previously calculated offset value. In this case, the second return test result is "pass". Alternatively, the STM remains in the second return state 310 until the second return timeout passes. In this case, the second return test result is "fail". In the second return state 310, the second return timeout refers to the expiration of the second maximum return period starting when entering the second return state 310. As an alternative implementation of the second return timeout, a second return time window can be determined for the second return state 310 such that if the second return state condition is achieved within the second return time window, the second return test is considered "pass", and otherwise the second return test is considered "fail". Preferably, then the end time of the second return time window will coincide with the end of the second return period. For a symmetric design, the second return time window can be equal to the first return time window, but the first time window and the second time window can also be different from each other. The average value is again independent of the external acceleration. Preferably, the return threshold window is determined such that it includes the offset value and can allow a small deviation from the offset value.

[0079] According to some embodiments, the first return period and the second return period are equal, but they can also have different values from each other.

[0080] When a predefined return threshold or timeout is reached in the second return state 310, the STM proceeds to the end stage 312 of the STM. If it has been determined that states 304, 306, 308, and 310 end in a "pass" state, the self-test ends and a "pass" state is given to the channel under test. According to some embodiments, if there is one or more failure states, the overall self-test ends in a "fail" state. According to some embodiments, a predefined number of retests are allowed before the overall self-test is considered to have failed. As shown by the dashed arrow 320, the retest can start again from the offset determination state 302, or as shown by the dashed arrow 321, from the first deflection state 304. Restarting from the offset determination state 302 is beneficial because it allows the initial offset to be re-measured in case the root cause of the self-test failure lies in the determined offset, for example due to external interference rather than due to the operation of the dual differential MEMS accelerometer itself (i.e., its proof mass).

[0081] For simplicity, the MDST has been described above for a single channel of a single-axis accelerometer (i.e., an accelerometer that measures acceleration along a single axis). When the same principle is applied to an accelerometer with two or three detection axes, some additional considerations are required in the design of one or more MDST state machines, depending on the design of the MEMS device. The MEMS device can include several accelerometers, but can also include other types of MEMS sensors, such as gyroscopes and / or pressure sensors. If the MEMS device provides output signals in parallel, the MDST test can also be run in parallel on different channels. However, the clocks for all channels should remain in the MDST mode (as Figure 2as shown in , until the self - tests of all channels are completed. The MDST mode refers to an arrangement in which the clock scheme is specifically configured for MDST operations, which include but are not limited to detecting the equivalent capacitance values of the sense mass (C1 - C2) and (C3 - C4), and optionally (C1 + C2) and (C3 + C3). The MDST mode requires a different clock from the normal double - differential acceleration measurement mode, in which the clock is typically configured to obtain the corresponding fully - differential capacitance output signal (C3 - C1)+(C2 - C4). If the MEMS device provides a time - multiplexed output signal, the output data from different channels are provided in a time - multiplexed manner from a single output. If the MEMS device is time - multiplexed, the time - division multiplexed readout channels can be optionally operated in a single - channel mode such that each channel is tested separately using the MDST test scheme described above. If the test of any single channel fails, the test can be repeated a predetermined number of times to avoid false positives, for example, due to some unexpected external interference. The MDST self - test results can be provided as a combined result, which shows failure if any of the tested channels fails, and / or the MDST self - test results can be provided separately for each channel when the channels are not related to each other. In the case of mutually - dependent channels, all tests must pass to achieve an overall MDST self - test pass state.

[0082] According to some embodiments, time - multiplexing can also be implemented such that the sense mass is also deflected in a time - multiplexed manner. In this case, each channel is biased and detected during each full clock cycle. If the MEMS accelerometer has several channels, this arrangement may pose challenges due to the slower deflection repetition rate, but if the MEMS bandwidth is very low compared to the MDST clock rate used for biasing and readout, this option can provide a viable solution for achieving short self - test periods.

[0083] Figure 4An exemplary accelerometer mass deflection self-test system including an MDST STM is schematically shown. A front-end circuit system (FE) 41 includes circuitry for converting the capacitance values C1, C2, C3, and C4 of the MEMS accelerometer 10 into an analog output voltage Vout, such as one or more transducers, amplifiers, and filters. The FE 41 also includes circuitry for providing a bias voltage during MDST. The analog output voltage Vout is digitized by an analog-to-digital converter (ADC) 42, which produces a digital output signal “ACC”. In a normal operation mode, the digital output signal ACC represents the detected acceleration. In the MDST mode, ACC represents the desired digital input value for the MDST STM 43. The MDST STM provides control of one or more signals for the FE 41 to control the operation of the FE 41 during MDST. This control determines when the FE 41 biases the capacitors C1, C2, C3, and / or C4, when the FE causes a reset, etc. Optionally, as shown by the dashed arrow, the MDST STM 43 can also control the operation of the ADC 42. The MDST STM outputs MDST data, which represents the result of the self-test, such as timing and / or status data from the state of the MDST STM.

[0084] In Figure 5 is shown an analog example of a self-test “pass” case test waveform, where all states end in a “pass” state. Figure 5 The example in shows an analog self-test of three channels, each representing an MDST performed using the deflection of a double-differential detection mass, which is performed using a pair of common stators associated with a detection mass configured to detect acceleration along one of the three axes (X, Y, Z) of a triaxial double-differential MEMS accelerometer. In combination with Figure 3The described STM states are shown in the "AX" graph in the upper left part, but the same STM states can be easily recognized from all graphs. The upper trace shows the digitized sense mass equivalent capacitance signals A1_DEC and A2_DEC on three different channels. After each self-test is completed, the channel under test is changed, so that both the stable wait state 300 and the offset determination state 302 are applied at the start of the self-test for each new channel. In this simulation example, the sinusoidal external acceleration is visible in the A1_DEC and A2_DEC data on the Y-axis and Z-axis, resulting in an obvious difference between the obtained A1_DEC value and A2_DEC value. As expected, the external acceleration is not visible in the sum data (A1_DEC + A2_DEC) labeled SUM_X, SUM_Y, SUM_Z (shown in the lower part) for different channels. Some offsets can be seen during the stable wait state 300 and the offset determination state 302, approximately 100 mV at the X-axis and -100 mV at the Z-axis. The peak sum data depends on the external acceleration because the deflection state is stopped using a single differential data (A1_DEC or A2_DEC) as the deflection threshold. In addition to the effect on the deflection threshold, the external acceleration also causes some additional changes in the timing of the states.

[0085] Figure 6 The digitized sense mass equivalent capacitance signals A1_DEC and A2_DEC of the actual MDST measurements performed on the X-axis, Y-axis, and Z-axis of a triaxial accelerometer are shown. The Z-axis is affected by gravity, but no additional external acceleration is applied during this test. In this example, the least significant bit (LSB) of the ADC that converts the capacitance value into a digital value is used as the scale to represent the output data.

[0086] Figure 7 The measured MDST data of the case where the measured deflection direction is considered a failure is shown. The submitted self-test is repeated once, after which the self-test proceeds to test the next channel. The self-test sequence is similar to the previous example, in the order of 1) X-axis, 2) Y-axis, 3) Z-axis. In this example, the gravity is in the X-axis direction and no other external acceleration is applied. In this test, the deflection fails in the X-axis test and the Y-axis test, so both the X-channel test and the Y-channel test are considered "failed", and only the Z-axis test is considered "passed". This type of failure may be caused by an error in the variable that controls the polarity of the read pulse or the polarity of the output signal of the channel, or the inversion of the bit value in the memory or flip-flop, resulting in the inversion of the configuration bit.

[0087] It is obvious to those skilled in the art that with the progress of technology, the basic concept of the present invention can be implemented in various ways. Therefore, the present invention and its embodiments are not limited to the above examples, but they can vary within the scope of the claims.

Claims

1. A method for self-testing a capacitive MEMS accelerometer, the capacitive MEMS accelerometer including a first detection mass and a second detection mass, the first detection mass and the second detection mass being arranged in a double-differential configuration, wherein: - The first detection mass is associated with two sensing capacitors (C1, C2), the two sensing capacitors (C1, C2) being configured to capacitively detect the position of the first detection mass relative to a first stator and a second stator, and - The second detection mass is associated with two sensing capacitors (C3, C4), the two sensing capacitors (C3, C4) being configured to capacitively detect the position of the second detection mass relative to a third stator and a fourth stator, and - wherein the first stator is electrically coupled to the third stator and the second stator is electrically coupled to the fourth stator, The method includes: - Deflecting the first detection mass and the second detection mass in mutually opposite first directions by applying a first bias voltage to the first stator and the third stator during a bias period of a bias voltage mode, wherein the bias voltage mode is a repetitive mode including intermittent bias periods and readout periods, wherein a deflection bias voltage is selectively applied during the bias period and no deflection bias voltage is applied during the readout period, and wherein the bias voltage mode repeats at a frequency above the resonant frequency and / or -3dB bandwidth of the capacitive MEMS accelerometer, Characterized in that the method includes, during each readout period: - Feeding a first readout signal to the first detection mass and obtaining first output data representing the detected mass equivalent capacitance signal (C1-C2) of the first detection mass, and - Feeding a second readout signal to the second detection mass and obtaining second output data representing the detected mass equivalent capacitance signal (C3-C4) of the second detection mass, wherein the first readout signal and the second readout signal are fed to the first detection mass and the second detection mass at mutually different times during the readout period for obtaining the first output data and the second output data respectively at mutually different times during the readout period.

2. The method according to claim 1, further comprising: - During each readout period, obtaining third output data representing the sum capacitance value (C1+C2) of the first detection mass and fourth output data representing the sum capacitance value (C3+C4) of the second detection mass, wherein each of the first output data, the second output data, the third output data, and the fourth output data is all obtained during the readout period, and - Calculating a first normalized differential capacitance value by dividing the first output data by the third output data and calculating a second normalized differential capacitance value by dividing the second output data by the fourth output data.

3. The method according to claim 1, further comprising: - When it is determined that any one of the first detection mass and the second detection mass has reached a first minimum required deflection amplitude through deflection by the bias voltage mode based on the first output data and the second output data, or when it is determined that a first predetermined maximum deflection period has expired, abort the application of the first bias voltage, and - After the abort, apply the bias voltage mode without applying any deflection bias voltage during the bias period to cause the first detection mass and the second detection mass to return toward the rest position.

4. The method according to claim 3, further comprising: - Obtain the first output data and the second output data when applying the bias voltage mode without applying any deflection bias voltage during the bias period, - Determine that the first detection mass and the second detection mass have reached the rest position or determine the expiration of a first return period based on the average value of the first output data and the second output data.

5. The method according to claim 4, further comprising: - When it is determined that the first detection mass and the second detection mass have reached the rest position, or when the first return period has expired, apply a second bias voltage to the second stator and the fourth stator during the bias period to cause the first detection mass and the second detection mass to deflect in a second opposite direction opposite to the corresponding first opposite direction, - Obtain the first output data and the second output data, - When it is determined that the deflection of any one of the first detection mass and the second detection mass has reached a second minimum required deflection amplitude based on the first output data and the second output data, or when a second predetermined maximum deflection period has expired, abort the application of the second bias voltage, - After the abort, apply the bias voltage mode without applying any deflection bias voltage during the bias period to cause the first detection mass and the second detection mass to return toward the rest position, - Obtain the first output data and the second output data when applying the bias voltage mode without applying any deflection bias voltage during the bias period, and - Determine that the first detection mass and the second detection mass have reached the rest position or determine the expiration of a second return period based on the average value of the first output data and the second output data.

6. The method according to claim 2, further comprising: - When it is determined that any one of the first detection mass and the second detection mass has reached a first minimum required deflection amplitude through deflection by the bias voltage mode based on the first normalized differential capacitance value and the second normalized differential capacitance value, or when it is determined that a first predetermined maximum deflection period has expired, abort the application of the first bias voltage, and - After the abort, apply the bias voltage mode without applying any deflection bias voltage during the bias period to cause the first detection mass and the second detection mass to return toward the rest position.

7. The method according to claim 6, further comprising: - Obtaining the first output data, the second output data, the third output data, and the fourth output data, and calculating the first normalized differential capacitance value and the second normalized differential capacitance value when applying the bias voltage pattern without applying any deflection bias voltage during the bias period, - Determining that the first detection mass and the second detection mass have reached the rest position, or determining the expiration of a first return period, based on an average value of the first normalized differential capacitance value and the second normalized differential capacitance value.

8. The method according to claim 7, further comprising: - When determining that the first detection mass and the second detection mass have reached the rest position, or when the first return period expires, applying a second bias voltage to the second stator and the fourth stator during the bias period to deflect the first detection mass and the second detection mass in a second opposite direction opposite to the corresponding first opposite direction, - Obtaining the first output data and the second output data and obtaining the third output data and the fourth output data, and calculating the first normalized differential capacitance value and the second normalized differential capacitance value, - Aborting the application of the second bias voltage when determining that the deflection of either the first detection mass or the second detection mass has reached a second minimum required deflection amplitude, or when a second predetermined maximum deflection period expires, based on the first normalized differential capacitance value and the second normalized differential capacitance value, - After the aborting, applying the bias voltage pattern without applying any deflection bias voltage during the bias period to return the first detection mass and the second detection mass toward the rest position, - Obtaining the first output data, the second output data, the third output data, and the fourth output data when applying the bias voltage pattern without applying any deflection bias voltage during the bias period, and - Determining that the first detection mass and the second detection mass have reached the rest position, or determining the expiration of a second return period, based on an average value of the first normalized differential capacitance value and the second normalized differential capacitance value.

9. The method according to any one of claims 3 to 8, wherein The self-test includes one or two deflection portions, each deflection portion including deflecting the first detection mass and the second detection mass by applying a corresponding first bias voltage or second bias voltage, wherein: - If, when determining that the deflection of the first detection mass or the second detection mass has reached a corresponding first minimum required deflection amplitude or second minimum required deflection amplitude before the expiration of a corresponding first predetermined maximum deflection period or second predetermined maximum deflection period, the application of the corresponding first bias voltage or second bias voltage during the bias period of the bias voltage pattern is aborted, then the deflection portion of the self-test is considered to have passed, and - If the application of the corresponding first bias voltage or second bias voltage is aborted when the predetermined maximum deflection period expires, the deflection part of the self-test is considered to have failed.

10. The method according to any one of claims 4 to 8, wherein, The method includes one or two return parts, during which the first detection mass and the second detection mass are returned towards the rest position by not applying any deflection bias voltage during the bias period, wherein: - If the first detection mass and the second detection mass reach the rest position within the corresponding first predetermined maximum return period or second predetermined maximum return period after the application of the corresponding first bias voltage or second bias voltage is aborted, or if the first detection mass and the second detection mass reach the rest position within the corresponding first return time window or second return time window, the return part of the self-test is considered to have passed, and - If it is determined that the predetermined first maximum return period or second maximum return period has expired and the first detection mass and the second detection mass are not determined to have reached the rest position, or if it is determined that the first detection mass and the second detection mass are not determined to have reached the rest position within the corresponding first return time window or second return time window, the return part of the self-test is considered to have failed.

11. The method according to any one of claims 1 or 3 to 5, further comprising: - After the capacitive MEMS accelerometer has entered the self-test mode or the channel of the multi-channel capacitive MEMS accelerometer to be self-tested has changed, waiting for a predetermined waiting period, during which the bias voltage pattern is applied to the capacitive MEMS accelerometer in the self-test mode, and - After the predetermined waiting period has expired, applying the bias voltage pattern without applying any deflection bias voltage during the bias period and obtaining a plurality of samples of the first output data and the second output data during the readout period, - Determining an offset value by averaging the plurality of samples of the first output data and the second output data during the readout period, and - Based on the offset value, determining at least one threshold or threshold window for determining whether the first detection mass and / or the second detection mass have reached the rest position.

12. The method according to any one of claims 2 or 6 to 8, further comprising: - After the capacitive MEMS accelerometer has entered the self-test mode or the channel of the multi-channel capacitive MEMS accelerometer to be self-tested has changed, waiting for a predetermined waiting period, during which the bias voltage pattern is applied to the capacitive MEMS accelerometer in the self-test mode, and - After the predetermined waiting period has expired, applying the bias voltage pattern without applying any deflection bias voltage during the bias period and obtaining a plurality of samples of the first output data and the second output data, the third output data and the fourth output data during the readout period, - Determining an offset value by averaging a plurality of first normalized differential capacitance values and second normalized differential capacitance values, and - Based on the offset value, determining at least one threshold or threshold window for determining whether the first detection mass and / or the second detection mass has reached the rest position.

13. The method according to any one of claims 1 to 8, wherein If any part of the self-test is considered to have failed, the self-test can be repeated a predetermined number of times.

14. The method according to claim 11, wherein, If any part of the self-test is considered to have failed, the self-test can be repeated a predetermined number of times, and when the self-test is repeated, the self-test restarts from the step of applying the bias voltage pattern without applying the bias voltage during the bias period to re-determine the offset value.

15. The method according to claim 12, wherein If any part of the self-test is considered to have failed, the self-test can be repeated a predetermined number of times, and when the self-test is repeated, the self-test restarts from the step of applying the bias voltage pattern without applying the bias voltage during the bias period to re-determine the offset value.

16. A state machine configured to control a mass block deflection self-test of a capacitive microelectromechanical MEMS accelerometer, the capacitive microelectromechanical MEMS accelerometer including a first detection mass and a second detection mass, the first detection mass and the second detection mass being arranged in a double-differential configuration, wherein: - The first detection mass is associated with two sensing capacitors configured to capacitively detect the position of the first detection mass relative to a first stator and a second stator, and - The second detection mass is associated with two sensing capacitors configured to capacitively detect the position of the second detection mass relative to a third stator and a fourth stator, and - Wherein, the first stator is electrically coupled to the third stator and the second stator is electrically coupled to the fourth stator, Wherein, the self-test is configured to apply a bias voltage pattern to the capacitive MEMS accelerometer, the bias voltage pattern being a repeating pattern including intermittent bias periods and readout periods, wherein a deflection bias voltage is selectively applied during the intermittent bias periods and no deflection bias voltage is applied during the readout periods, and wherein the bias voltage pattern repeats at a frequency above the resonant frequency and / or -3dB bandwidth of the capacitive MEMS accelerometer, wherein the state machine includes, in execution order: - A wait state for waiting for a predetermined wait period after the capacitive MEMS accelerometer has entered the self-test mode or the channel to be self-tested has changed, - An offset determination state for determining an offset value by averaging a plurality of samples of first output data and second output data during the readout period of the bias voltage pattern when the bias voltage pattern is applied without applying any deflection bias voltage during the intermittent bias period, - A first deflection state, for deflecting the first detection mass and the second detection mass in mutually opposite first directions by applying a first deflection bias voltage to the first stator and the third stator during a bias period of the bias voltage mode, wherein the first stator is electrically coupled to the third stator, - A first return state, for returning the first detection mass and the second detection mass towards a rest position by not applying any deflection bias voltage across the sense capacitor during the bias period, - A second deflection state, for deflecting the first detection mass and the second detection mass in mutually opposite second directions opposite to the mutually opposite first directions by applying a second deflection bias voltage to the second stator and the fourth stator, wherein the second stator is electrically coupled to the fourth stator, - A second return state, for returning the first detection mass and the second detection mass towards the rest position by not applying any deflection bias voltage across the sense capacitor, and - An end state, for determining whether the results of the first deflection state and the second deflection state and the first return state and the second return state are determined to be a failure or determined to be a pass, and optionally repeating the self-test starting from the offset determination state if any of these states is a failure, characterized in that during each readout period: - A first readout signal is fed to the first detection mass, and first output data representing a detection mass equivalent capacitance signal of the first detection mass is obtained, - A second readout signal is fed to the second detection mass, and second output data representing a detection mass equivalent capacitance signal of the second detection mass is obtained, wherein the first readout signal and the second readout signal are fed to the first detection mass and the second detection mass at mutually different times during the readout period to obtain the first output data and the second output data respectively at mutually different times during the readout period.

17. A MEMS device includes at least one capacitive MEMS accelerometer, wherein, The MEMS device further includes circuitry for implementing the state machine according to claim 16.

Images