JPH0875777A - Semiconductor acceleration sensor - Google Patents

Abstract

PURPOSE: To provide a semiconductor acceleration sensor which greatly improves detecting sensitivity at self-diagnosis time and has high reliability. CONSTITUTION: A mass part 29 is set in a semiconductor substrate 25. The sensor is provided with an acceleration-detecting part 1 which detects an acceleration from a stress generated as a result of the acceleration impressed to the mass part 29, and a self-diagnosing part which diagnoses whether the acceleration-detecting part 1 operates normally. The self-diagnosing part is equipped with a differential amplifier circuit 3 which differentiates and amplifies a change of potential caused when the mass part 29 is shifted because of the potential applied to the acceleration-detecting part 1, and a signal-holding circuit 5 which holds outputs of the differential amplifier circuit 3.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor acceleration sensor, and more particularly to an integrated semiconductor acceleration sensor having a self-diagnosis circuit.

[0002]

2. Description of the Related Art Conventionally, most automobile airbags operate by detecting the acceleration generated during the collision. Here, the acceleration sensor used in the airbag and the like is desired to be of a semiconductor type for reasons such as high accuracy. Further, it is preferable that the acceleration sensor used in the airbag or the like can constantly check that the sensor is normally operating, in view of safety requirements. That is, it is required to have a self-diagnosis function for checking whether this sensor is operating normally. There are many documents about the semiconductor acceleration sensor having this self-diagnosis function (for example, TECHNICAL DIGEST OF THE 11TH SENSORSYMPOSIUM, 1
992.pp43-46).

Here, the semiconductor acceleration sensor disclosed in the above-mentioned document will be described below with reference to a schematic sectional view of FIG. In this semiconductor acceleration sensor, a heat resistant glass layer 23, a Si layer 25, and a top cap 27 are laminated, a movable mass 29 is provided in the Si layer 25, and is held by a flexible portion 31. Movable mass 2
The thickness of the flexible portion 31 is reduced so that the flexible portion 31 can be easily bent when a force is applied to it. Further, a cavity 41 having a predetermined width is provided on the upper surface of the movable mass 29 and on the joint surface between the Si layer 25 and the top cap 27. The top cap 27 surface side of this cavity 41 and Si
Electrodes 33 and 35 are provided on the surface side of the layer 25, and a predetermined voltage can be applied to the electrodes 33 and 35 by a power source (not shown).

Next, the operation of this semiconductor acceleration sensor will be described. When acceleration occurs in the direction of the paper surface of FIG. 4A, a force is applied to the movable mass 29. As a result, the flexible portion 31 is bent and mechanically deformed, and the space between the hollow portions 41 is narrowed. Therefore, the electrode 3 provided in the cavity 41 is
The capacitance between 3 and 35 changes, and the integrated circuit 37 detects the capacitance change, whereby the acceleration can be detected. This acceleration detection method is called a capacitive acceleration detection method. Further, this acceleration sensor has a small error due to temperature change, and for example, by injecting a vibration absorbing material or the like into the cavity 39 between the heat resistant glass layer 23 and the Si layer 25, the error in detection due to resonance is reduced. Therefore, it can be said that it is suitable for mounting on an automobile or the like.

Next, the self-diagnosis function of the conventional semiconductor acceleration sensor will be described. In this description, a resistance type acceleration detection method will be described in contrast to the capacitive type acceleration detection method described above.

The structure of the semiconductor acceleration sensor having the self-diagnosis function using the resistance type acceleration detection method is almost the same as the structure of the semiconductor acceleration sensor having the self-diagnosis function using the capacitive acceleration detection method. Although the description is omitted, the difference is that the resistance element 43 is provided on the upper surface of the flexible portion 31. The operation of the semiconductor acceleration sensor having the self-diagnosis function by the resistance type acceleration detection method will be described below with reference to FIGS.

When no voltage is applied to the electrodes 33 and 35, the movable mass 29 is arranged at a predetermined place (dotted line portion in FIG. 4B). Here, when a voltage for self-diagnosis is applied between the movable mass 29 and the electrodes 33 and 35 on the top cap 27, an electrostatic attractive force is generated. By applying this electrostatic attraction to the movable mass 29, the movable mass 29 is simulated.
Then, the distance between the movable mass 29 and the top cap 27 changes, the flexible portion 31 bends, and mechanical deformation occurs (solid line portion in FIG. 4B). As a result, the electric resistance of the resistance element 43 provided on the upper surface of the flexible portion 31 changes. The integrated circuit 3
The sensor 7 detects whether the sensor is operating normally, and if the operation is abnormal, various processes such as warning display are performed.

In the case of actual operation (for example, when a car or the like collides), a considerable amount of force is applied to the acceleration detecting section. When performing self-diagnosis, a force smaller than that is applied, so that it is possible to perform actual operation detection and self-diagnosis simultaneously with one acceleration sensor.

[0009]

However, in the conventional semiconductor acceleration sensor, the electrostatic attraction generated when a voltage is applied to the electrodes 33 and 35 is small, and the flexible portion 31 does not actually bend so much. There was little change in the electrical resistance of. Therefore, the sensor signal detected by the integrated circuit 37 becomes very small, which makes self-diagnosis difficult.

The above situation can be avoided by providing a high power supply for self-diagnosis separately from a power supply for driving a general circuit, but it is not provided with such a high power supply for a moving body such as an automobile. It is realistic and cannot be said to be an essential problem solving. There is also a drawback that the circuit for self-diagnosis becomes complicated. Further, since the sensor signal is so small, when it is mounted on an automobile or the like, there is a possibility that a malfunction may occur if it is amplified as it is due to noise generated by vibration of the automobile or the like.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a highly reliable semiconductor acceleration sensor that significantly improves the detection sensitivity during self-diagnosis. is there.

[0012]

In order to achieve the above object, a feature of the present invention is to provide a mass portion on a semiconductor substrate, and an acceleration detecting portion for detecting acceleration by a stress generated by applying acceleration to the mass portion. In the semiconductor acceleration sensor including a self-diagnosis unit that diagnoses whether or not the acceleration detection unit is operating normally, the self-diagnosis unit is configured such that the mass unit is controlled by applying a potential to the acceleration detection unit. That is, a differential amplifier circuit that is displaced and differentially amplifies a change in potential generated by this displacement, and a signal holding circuit that holds an output of the differential amplifier circuit are provided.

Here, the differential amplifier circuit includes a first switch which operates at a first clock, and a second switch which is connected in parallel with the first switch and operates at a second clock. A third switch that operates on a first clock; a first inverter whose output and input are short-circuited by the third switch; the switch connected in parallel; and the first inverter. And a capacitor provided therebetween, and more preferably, the second inverter and the buffer are connected in series,
Is connected in series to the inverter. In addition, the first switch, the second switch, and the third switch can be formed on the same substrate.
It is preferably composed of an S (N-channel MOS) transistor. Further, it is preferable that all the above-mentioned inverters and buffers are composed of NMOS inverters or CMOS (Complementary MOS) inverters in that they can be formed on the same substrate. Further, the signal holding circuit is preferably configured by a D flip-flop in that it can hold the signal of the differential amplifier circuit. Further, it is preferable that the differential amplifier circuit and the signal holding circuit are provided on the same substrate as the acceleration detection unit.

[0014]

According to the above configuration, since the differential amplifier circuit is provided in the self-diagnosis section, the potential of the signal output from the acceleration detection section, which is detected when acceleration is applied to the mass section provided on the semiconductor substrate, is detected. The difference between the changes in the voltage is relatively compared and this comparison is performed by subtracting the voltage before and after the voltage change.Therefore, fluctuations in the power supply voltage and common-mode noise mixed in the signal output from the acceleration detection unit are canceled out. Is done.
Therefore, this comparison result differentially amplifies the signal from which the in-phase noise and the like have been subtracted, so that a minute sensor signal can be differentially amplified with high accuracy.

Further, for example, when the differential amplifier circuit detects an abnormality in the acceleration detecting section, the abnormality signal must be held and transmitted to an external device or the like. In this respect, it is meaningful to use a signal holding circuit that holds the output of the differential amplifier circuit.

[0016]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic diagram of a semiconductor acceleration sensor including a self-diagnosis unit according to an embodiment of the present invention. This semiconductor acceleration sensor includes an acceleration detector 1 that detects acceleration, a differential amplifier circuit 3 that differentially amplifies a difference before and after a potential change output by the acceleration detector 1, and the amplified difference. And a signal holding circuit 5 for holding.

First, the acceleration detector 1 according to this embodiment will be described. Here, the one using the above-mentioned resistance type acceleration detection method is used. Here, the variable resistance described in the acceleration detection unit 1 of FIG. 1 corresponds to the resistance element 43 in FIG. 4B. The resistance element 43 is preferably manufactured by thermal diffusion or the like, and among them, a piezo resistance element of a strain sensitive gauge is preferable. Further, the flexible element 31 in FIG. 4 (b) is bent so that the resistive element 43 provided at a portion where a compressive stress is applied to the flexible portion 31 and the resistive element 43 provided at a portion where a tensile stress is applied. If the resistors are connected in series like the variable resistor described in the acceleration detecting unit 1 of FIG. 1 and the potential at the midpoint thereof is used as a reference, the potential change becomes large when self-diagnosis is performed.

Next, the structural requirements of the differential amplifier circuit 3 according to this embodiment will be described. A switch 9a that operates on the clock Φ ₁ and a switch 9b that operates on the Φ ₂ are connected in parallel, and a signal from the acceleration detection unit 1 is input when one of them is on. Note that these switches use NMOS transistors. The capacitor 13 is provided to store the potential difference between the contact A and the contact B and to prevent common-mode noise. The inverter 15 has a switch 11 that can short-circuit its output and input. The switch 11 is designed to operate with the clock Φ ₁ . In addition,
In this embodiment, the inverter 15 is a CMOS inverter. In addition, the inverter 17 and the buffer 1
9 is arranged mainly for the purpose of logical matching and waveform correction.

As described above, the differential amplifier circuit 3 includes the switches 9a and 9b connected in parallel, the capacitor 13, the inverter 15, the inverter 17, the buffer 19, and the like.
Are connected in series and have a switch 11 for short-circuiting the output and input of the inverter 15.

Next, the signal holding circuit 5 according to this embodiment will be described. The signal holding circuit 5 includes a differential amplifier circuit 3
Holds the output from and plays the role of the external interface to output to the outside. In this embodiment, a D flip-flop will be used. The determination circuit 3 and the signal holding circuit 5 described above are preferably formed on the same substrate.

Next, the operation of the semiconductor acceleration sensor according to this embodiment will be described. (1) FIG. 2A is a diagram for explaining a self-diagnosis electrode that generates electrostatic attraction for self-diagnosis. As shown in this figure, the voltage V is applied to the self-diagnosis electrode 21 when the inverted signal of the clock Φ ₁ is high voltage (ON state) (clock Φ ₁ is low voltage (OFF state)). There is. Here, in this embodiment, the self-diagnosis electrode 21 is
The electrodes 33 and 35 in FIG. Therefore,
When the clock Φ ₁ has a low voltage (off state), the electrode 33
And 35 are energized.

FIG. 2B is a diagram showing timings of the clock Φ ₁ and the clock Φ ₂ . Clock Φ
As shown in the inverted signal and the clock [Phi ₂ signals this figure _1, always inverted signal of the clock [Phi ₁ when the clock [Phi ₂ is turned on are timed so on. Further, the sensor output when no voltage is applied to the self-diagnostic electrode 21 (in the case of a differential output, the output on one side thereof) is V _REF , and the sensor output when a voltage is applied to the self-diagnostic electrode is V _SEN. And

Here, the clock Φ ₁ is ₁ from 100 [Hz].
000 [Hz] is preferable, more preferably 300 [Hz]
To 600 [Hz]. This is because when the frequency is too high, the movable mass 33 cannot follow the vibration, and when the frequency is low, a leak current is generated from the capacitor 13 when the signal changes.

(2) Next, description will be made with reference to FIG. Generate the clocks Φ ₁ and Φ ₂ from the pulse generation circuit 7,
The clock Φ ₁ is input to the switches 9a and 11, and the clock Φ ₂ is input to the switch 9b.

First, when the clock Φ ₁ has a high potential, the switch 11 is turned on and the inverter 1
The output and the input of 5 are short-circuited. As a result, the contact B is biased to the logic threshold value V _th1 of the inverter 15 and is in a reset state (input and output are at the same potential). Further, the switch 9a is turned on, and the acceleration detection unit 1
However, since the self-diagnosis voltage V is not applied to the electrodes 33 and 35, the electrostatic attraction force is not applied to the sensor. Therefore, contact A is charged to V _REF .

Next, when the clock Φ ₁ has a low potential, the switches 9a and 11 are turned off. As shown in FIG. 2B, when the clock Φ ₂ becomes high in potential after a predetermined time has passed since the clock Φ ₁ became low in potential, the switch 9b is turned on. Also, the electrode 33
Since the self-diagnosis voltage V is applied to and 35, the contact A is charged to V _SEN .

Here, the reason why the clock Φ ₂ is set to the high potential after a predetermined time has passed since the clock Φ ₁ is set to the low potential is to turn on the switch 9b after the switch 11 is surely turned off. This is because

The potential drop of the capacitor 13 (V _REF -V
_{Since th1} ) is stored, the potential of the contact B changes from V _{REF by the} amount of change in the potential of the contact A (V _SEN −V _REF ).

(3) Here, if the sensor operates normally, V _SEN and V _REF should be different. Here, in one inverter, as shown in FIG. 3, the logic threshold V _th1
Rather than changing vertically at a certain inclination (generally 50 ~
It is 100, and here, the absolute value is taken to be 0 or more. ), It changes from 50 to 50 for slight changes in input.
The output change when amplified 100 times will be obtained. Therefore, by adopting the configuration of this embodiment,
Amplification can be obtained with a simple configuration. Further, since it does not require any adjustment such as offset, it can be easily used.

Now, _assuming that V _SEN is greater than V _REF , this amplified output is input to the inverter 17,
Further, when input to the buffer 19, the output to the D flip-flop becomes logic 1. On the other hand, when the output of the sensor does not change, that is, when V _SEN and V _REF are the same, the output of the inverter 15 becomes logic 1, and the output of the inverter 17 and the buffer 19 from the signal holding circuit 5 to D The output to the flip-flop becomes logic 0. Therefore,
The output to the D flip-flop of the signal holding circuit 3 is logical 0
In the case of, the sensor recognizes that there is an abnormality, and the output of the D flip-flop holds logic 0 (sensor abnormality signal).

Here, in the case of a sensor abnormality, the output of the inverter 15 becomes V _th1 , but in this case, as shown in FIG. It is desirable to design the logical threshold value V _th2 of the above to be smaller than V _th1 . As a result, even when the inverter 17 inputs V _th1 , the output can be brought close to 0, and can be made almost 0 by the buffer 19.

The threshold value of the inverter 17 is V _th2
May be designed to be smaller than V _th1 by about 10 mV in consideration of variations between elements. this is,
For example, the channel width of the pull-up PMOS of the inverter 17 may be made smaller than that of the inverter 15 by several percent. The same effect can be obtained by changing other parameters (channel length, etc.). When V _SEN is smaller than V _REF , V _th2 may be designed to be larger than V _th1 by about 10 mV.

From the above, if the potential V _{REF of the} contact A changes, even if the change amount is small, the change amount is detected and changed (logic 1) or unchanged (logic 0) signal. Can be output to a D flip-flop which is a signal holding circuit. Therefore, the semiconductor acceleration sensor according to the present embodiment can reliably detect even a minute sensor signal. In addition, this self-diagnosis unit
All circuits may be on a single power supply (eg 5V). Further, the capacitor 13 prevents any common-mode noise (noise), and does not require offset adjustment or temperature drift adjustment.

Further, in this embodiment, since the potential change is relatively detected by the potential difference between the original value (V _{REF in} this embodiment) and the changed value (V _{SEN in} this embodiment), the power supply is changed. It is possible to avoid malfunctions due to voltage fluctuations and common mode noise.

[0035]

As described above, according to the present invention, the potential change is relatively detected by comparing the difference before and after the potential change, so that the fluctuation of the power supply voltage and the malfunction due to the common mode noise can be avoided. You can Therefore, the detection sensitivity during self-diagnosis can be significantly improved.
The reliability of the acceleration sensor can be improved.

[Brief description of drawings]

FIG. 1 is a schematic diagram of a semiconductor acceleration sensor according to the present invention.

FIG. 2 is a diagram for explaining the operation of the semiconductor acceleration sensor according to the present invention.

FIG. 3 is a diagram showing input / output characteristics of an inverter.

FIG. 4A is a diagram schematically showing a semiconductor acceleration sensor, and FIG. 4B is a diagram for explaining the operation thereof.

[Explanation of symbols]

 1 Acceleration Detection Section 3 Differential Amplifier Circuit 5 Signal Holding Circuit 7 Pulse Generation Circuit 9a Switch 9b Switch 11 Switch 13 Capacitor 15 Inverter 17 Inverter 19 Buffer 21 Self-Diagnosis Electrode 23 Heat-Resistant Glass Layer 25 Si Layer 27 Heat-Resistant Glass Layer (Top Cap) 29 Movable Mass 31 Flexible Part 33 Electrode 35 Electrode 37 Integrated Circuit 39 Cavity 41 Cavity 43 Resistance Element

Claims (1)

[Claims] 1. A mass part is provided on a semiconductor substrate, and an acceleration detecting part for detecting acceleration by a stress generated by applying acceleration to the mass part, and a diagnosis as to whether or not the acceleration detecting part is operating normally. In the semiconductor acceleration sensor including a self-diagnosis unit for performing, the self-diagnosis unit is configured to apply a potential to the acceleration detection unit to displace the mass unit and differentially amplify a change in potential generated by the displacement. A semiconductor acceleration sensor, comprising: an amplifier circuit; and a signal holding circuit that holds the output of the differential amplifier circuit.