CN116249957A - Sensor circuit, pixel circuit and method for controlling pixel circuit - Google Patents

Abstract

Provided are a sensor circuit and a pixel circuit that can reduce a response time of an event sensor and reduce false detections. The sensor circuit has a connection point between the source terminal of the MOS transistor and the light receiving element, and an amplifier having an input terminal connected to the connection point for outputting a voltage depending on In the logarithm of the photocurrent of the light receiving element, wherein the output terminal of the amplifier is connected to the gate terminal of the MOS transistor, and the gate terminal is adjusted to a certain voltage, the voltage depends on The logarithm of the photocurrent. The sensor circuit also includes a capacitor for a negative capacitance generator (NCG).

Description

Sensor circuit, pixel circuit and method for controlling pixel circuit

technical field

The present invention relates to a sensor circuit, a pixel circuit, and a method for controlling the pixel circuit, and more particularly, to a sensor circuit with a logarithmically converted output, a pixel circuit used as an event sensor, and a method for controlling the pixel circuit.

Background technique

Event-based sensors, also called dynamic vision sensors (DVS), have attracted attention as a new type of mobile terminal imaging device. An event-based sensor is a sensor that captures the brightness change of each pixel as an "event" and outputs its information, which has the advantages of low latency, low power consumption, and high dynamic range. Such detection properties are required in feature point extraction for simultaneous localization and mapping (SLAM) technology applied to autonomously traveling robots. In addition, this detection feature is necessary for high-speed image detection of high-speed moving objects, reconstruction of high-resolution images, compensation of motion blur and frame interpolation, etc. However, since the response latency of the event sensor depends on the brightness, the temporal precision of event detection decreases as the brightness decreases.

Contents of the invention

An object of the present invention is to provide a sensor circuit and a pixel circuit that can reduce the response time of an event sensor and reduce erroneous detection of the event sensor.

A first embodiment of the present invention provides a sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element, and a first inverting amplifier having having an input terminal connected to the connection point for outputting a voltage depending on the logarithm of the photocurrent of the light receiving element, wherein the output terminal of the first inverting amplifier is connected to to the gate terminal of the MOS transistor, and the gate terminal is regulated to a voltage that depends on the logarithm of the photocurrent. The sensor circuit also includes a second inverting amplifier having an input terminal connected to the output terminal of the first inverting amplifier, and an output terminal connected to the connection point through a capacitor.

According to the first embodiment, negative capacitance can be formed using an existing inverting amplifier while suppressing circuit scale expansion.

A second embodiment of the present invention provides a sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element, and an amplifier having a sensor connected to the connection point. input terminal, the connection point for outputting a voltage that depends on the logarithm of the photocurrent of the light receiving element, wherein the output terminal of the amplifier is connected to the gate terminal of the MOS transistor, and the The gate terminal is regulated to a voltage that depends on the logarithm of the photocurrent, wherein the amplifier is a differential amplifier having an inverting input terminal connected to the connection point, connected to A non-inverting output terminal of the gate terminal and an inverting output terminal connected to the connection point through a capacitor.

According to the second embodiment, there is no need to add the second inverting amplifier in the first embodiment.

A third embodiment of the present invention provides a pixel circuit including: a sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element, and a first inverting an amplifier, the first inverting amplifier having an input terminal connected to the connection point for outputting a voltage that depends on the logarithm of the photocurrent of the light receiving element, wherein the The output terminal of the first inverting amplifier is connected to the gate terminal of the MOS transistor, and the gate terminal is regulated to a voltage which depends on the logarithm of the photocurrent; and having a first Two inverting amplifiers, connected to the output terminal of the sensor circuit, and used to operate as a sample-and-hold circuit to maintain the input voltage and amplify the variation of the output terminal of the sensor circuit. The pixel circuit further includes a capacitor inserted between the output terminal of the second inverting amplifier and the connection point.

According to the third embodiment, a negative capacitance generation circuit (NCG) may be configured without adding a non-inverting amplifier.

In the third embodiment, a buffer amplifier may be inserted before the second inverting amplifier. A buffer amplifier can, for example, adjust the signal frequency band and operating point.

A fourth embodiment of the present invention provides a method of controlling a pixel circuit, the method including: a sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element; phase amplifier, the first inverting amplifier has an input terminal connected to the connection point for outputting a voltage that depends on the logarithm of the photocurrent of the light receiving element, wherein the The output terminal of the first inverting amplifier is connected to the gate terminal of the MOS transistor, and the gate terminal is regulated to a certain voltage, the voltage depends on the logarithm of the photocurrent; and having A second inverting amplifier connected to an output terminal of the sensor circuit and configured to operate as a sample-and-hold circuit to hold an input voltage and amplify a change amount of the output terminal of the sensor circuit by the pixel An initial program circuit execution of the circuit comprising the steps of: sensing for resetting the holding a reset signal of the held value of the circuit, and sending an event signal to the arbiter circuit; and releasing the reset signal upon completion of communication with the arbiter circuit.

According to the fourth embodiment, the kickback voltage at reset is small, so the input of photocurrent is not affected.

Description of drawings

[FIG. 1A] FIG. 1A is a schematic diagram of a conventional sensor circuit with a logarithmically converted output;

[FIG. 1B] FIG. 1B is a schematic diagram of a conventional sensor circuit with a logarithmically converted output;

[Fig. 2] Fig. 2 is a schematic diagram of the pixel circuit provided by the first embodiment of the present invention;

[FIG. 3A] FIG. 3A is a schematic diagram of an operation sequence of a conventional pixel circuit;

[FIG. 3B] FIG. 3B is a schematic diagram of the operation sequence of the pixel circuit in the first embodiment;

[FIG. 4] FIG. 4 is a pattern diagram of a pixel circuit provided by a second embodiment of the present invention;

[FIG. 5] FIG. 5 is a schematic diagram of a modification of the pixel circuit provided by the second embodiment;

[FIG. 6A] FIG. 6A is a schematic diagram of an inverting amplifier of the pixel circuit provided by the second embodiment;

[FIG. 6B] FIG. 6B is a schematic diagram of the differential amplifier of the pixel circuit provided by the second embodiment;

[FIG. 7] FIG. 7 is a schematic diagram of a pixel circuit provided by a third embodiment of the present invention;

[FIG. 8A] FIG. 8A is a schematic diagram of the capacitance structure of the negative capacitance generation circuit provided by the fourth embodiment of the present invention;

[FIG. 8B] FIG. 8B is a schematic diagram of another structure of the capacitance of the negative capacitance generating circuit provided by the fourth embodiment;

[ Fig. 9] Fig. 9 is a schematic diagram of a variable capacitance structure of a negative capacitance generating circuit provided in a fourth embodiment.

Detailed ways

Embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

Figure 1A shows a conventional sensor circuit with a log-converted output. A source terminal of the MOS transistor Tr is connected to a cathode terminal of a photodiode (PD), which is a light receiving element. The connection point is the logarithmically converted output LOG-OUT, which provides an output that depends on the logarithm of the photocurrent. When the MOS transistor Tr operates in the subthreshold region, the gate-source voltage V _gs is logarithmically proportional to the photocurrent as shown below.

[equation 1]

where _Iph is the drain current corresponding to the photocurrent, _I0 is the saturation current, _Vth is the threshold voltage, kT/q is the thermal voltage, and n is a constant determined by the structure of the MOS transistor. Therefore, as shown in FIG. 1A , when the gate voltage is a constant value, the voltage at the connection point becomes a voltage depending on the logarithm of the photocurrent, which is the logarithmically converted output LOG-OUT.

As shown in Figure 1A, since the photodiode (photodiode, PD) and MOS transistor Tr are configured as a source follower biased by the photocurrent flowing through the photodiode (photodiode, PD), the response time of the sensor circuit is at the source The settling time of the follower dominates and depends on the photocurrent. Here, the setup time of the source follower is determined by the parasitic capacitance (C _g ) between the gate and the source of the MOS transistor Tr and the parasitic capacitance (C _P ) of a photodiode (photodiode, PD) as a load. Therefore, when the illuminance is low, since the photocurrent of the photodiode (PD) is small, the response time is worse.

Therefore, as shown in FIG. 1B , feedback control is known for reducing the effective load capacitance (C _P ) through a photodiode (PD). The feedback control amplifier 11 is an inverting amplifier that receives a logarithmic conversion output as a photocurrent input and applies feedback control to the gate terminal of the MOS transistor Tr. When the gain of the amplifier 11 is -A, the voltage amplitude of the photodiode (photodiode, PD) is suppressed to 1/(1+A), so that the effective load capacitance is also suppressed to 1/(1+A).

It is also known that a negative capacitance generating circuit (negative capacitance generator (NCG)) is used as a method of reducing effective load capacitance (for example, see Patent Document 1). NCG is a non-inverting amplifier with positive feedback applied through capacitor C _pF . When the gain of the amplifier is A, the input capacitance C _NEG is expressed by Equation 2. When the gain A>1, the input capacitance becomes a negative capacitance.

[equation 2]

C _NEG = C _p (1-A) (2)

Similar to the active matrix sensor of Patent Document 1, by connecting the NCG to the logarithmic conversion output (photocurrent input) of the pixel circuit, the effective load capacitance can be reduced. This approach shortens the settling time of the source follower.

However, the former feedback control has limitations in reducing the effective load capacitance for the following reasons. First, when the feedback amplifier is configured as a common-source amplifier, the gain is limited to about 50 times. Second, the parasitic capacitance (C _g ) of the gate terminal of the MOS transistor Tr cannot be substantially reduced. Third, the slew rate of the source follower limits the voltage change in the photodiode (PD). Therefore, there is a limit to shortening the response time of event sensors.

Furthermore, when the latter NCG is applied to the pixel circuit shown in FIG. 1, there are problems of increasing the power consumption of the pixel circuit and increasing the circuit scale due to the realization of a non-inverting amplifier.

Therefore, in this embodiment, a new NCG is realized in the pixel circuit.

first embodiment

FIG. 2 shows a pixel circuit provided by the first embodiment of the present invention. A source terminal of the MOS transistor Tr is connected to a cathode terminal of a photodiode (PD), which is a light receiving element. The connection point serves as a logarithmic conversion output LOG-OUT which provides an output depending on the logarithm of the photocurrent of the light receiving element. An amplifier 21 with an input connected to the connection point outputs a voltage (V _log ) that depends on the logarithm of the photocurrent, which varies based on the intensity of incident light into a photodiode (PD). Furthermore, the output terminal of the amplifier 21 is connected to the gate terminal of the MOS transistor Tr, and the gate terminal is adjusted to a certain voltage which depends on the logarithm of the photocurrent.

An inverting amplifier comprising capacitors C1, C2 and amplifier 22 is connected to the output of the sensor circuit. A switch SW shorting the input and output is connected to the amplifier 22 and acts as a sample-and-hold circuit with a capacitor C1 to sample the output of the sensor circuit. One sampling period starts from the time the switch SW is turned off. Since the output of the sensor circuit at the above timing is the reference level, the amplifier 22 amplifies and outputs the amount of change in the output of the sensor circuit. That is, the input at the time when the switch SW is turned off is the reference level, and the inverting amplifier amplifies the voltage change of the logarithmically converted output into a voltage (V _diff ) representing a luminance change.

The output of the inverting amplifier is connected to detectors 23 and 24 . The detector 23 outputs an event detection signal (ON _event ) once the output terminal voltage (V _diff ), which is the difference between the output of the sensor circuit and the hold value of the sample hold circuit, exceeds a given threshold voltage. Similarly, once the output terminal voltage (V _diff ) is lower than a given threshold voltage, the detector 24 outputs an event detection signal (OFF _event ). Accordingly, detectors 23 and 24 can detect positive and negative brightness changes, respectively. Once the initial program circuit 25 detects that the output of the sample-and-hold circuit exceeds a given threshold, the initial program circuit 25 controls the switch SW to be turned on to reset the reference value for detecting brightness changes. Furthermore, when receiving an event detection signal, the initial program circuit 25 outputs an event signal for post-processing of the peripheral circuit.

In the pixel circuit of this embodiment, the capacitor _CF for feedback is also inserted between the output terminal of the amplifier 22 and the cathode terminal of the photodiode (photodiode, PD). That is, between the input and output of the two amplifiers 21 and 22 connected in cascade, in other words, positive feedback through the capacitor _CF is applied between the photocurrent input of the amplifier 21 and the brightness change of the amplifier 22 between outputs. This configuration corresponds to connecting the NCG to the log-converted output (photocurrent input) of the pixel circuit. Wherein, when the gain of the amplifier 21 is _A1 , the input capacitance C _NEG of the photocurrent input is expressed by Equation 3.

[equation 3]

Therefore, NCG can be configured without adding non-inverting amplifiers. The negative capacitance of NCG can be lowered in the effective load capacitance. That is, the effective load capacitance in the photodiode (PD) including the parasitic capacitance (C _g ) of the gate terminal of the MOS transistor Tr can be reduced as the effective load capacitance decreases by the feedback control shown in FIG. 1B . Therefore, the response time of the event sensor can be further shortened.

FIG. 3A shows an operation sequence of a conventional pixel circuit. For comparison, a conventional sequence performed by the initial program circuit 25 is shown. Once the output terminal voltage (V _diff ) outputted as a brightness change exceeds a predetermined threshold voltage, it is detected as a "trigger" event, and the pixel circuit outputs an event signal for post-processing of peripheral circuits.

Specifically, as a peripheral circuit, a two-dimensional encoder is connected, which is also called an arbiter circuit. The initial program circuit 25 of the triggered pixel circuit sends a REQ_R signal to the Y-axis address encoder and receives an ACK_R signal in response. In addition, the triggered initial program circuit 25 sends a REQ_C signal to the X-axis address encoder and receives an ACK_C signal in response. Once the initial procedure circuit 25 receives the ACK_C signal from the arbiter circuit, the initial procedure circuit 25 sends an event transfer signal (Evt.Trans.) to indicate that a series of event handshakes with peripheral circuits have been completed.

Furthermore, the initial program circuit 25 sends an event transfer signal (Evt.Trans.) and at the same time sends a reset signal to the sample hold circuit, and turns on the switch SW. Accordingly, the charge held in the capacitor C1 is discharged, the reference level of the input of the inverting amplifier is updated, and the next event detection cycle starts. That is, the hold value of the sample-and-hold circuit is reset, and the next sampling cycle starts. However, when the kickback voltage at reset is large, it may be a factor that a sampling error occurs at the input as an input reference and causes false detection by the detector.

FIG. 3B shows an operation sequence of the pixel circuit of the first embodiment. In the pixel circuit of the present embodiment, as shown in FIG. 2 , a capacitor C _F is inserted between the input and output terminals of the two cascade-connected amplifiers 21 and 22 configured as NCG. In addition, in the pixel circuit of this embodiment, when the output terminal voltage (V _diff ) of the inverting amplifier exceeds a predetermined threshold voltage and is detected as a trigger event, the initial program circuit 25 sends a reset signal to the sample-and-hold circuit for reset. A reference value for detecting brightness changes without waiting for a series of event handshakes to complete. Therefore, the kickback voltage at reset is small and will not affect the photocurrent input. In addition, since the kickback voltage is mitigated during the settling time described below, the detection accuracy of the detector can be improved.

After the trigger of the output terminal voltage (V _diff ) is detected, a reset signal is sent to complete a series of event signal exchanges with the peripheral circuit, and the reset signal is released until the next event detection cycle (called setup time) begins. In this embodiment, compared with the prior art, the setup time can be used as a preparation period for the next event detection, so that the detection accuracy of the detector can be improved.

second embodiment

FIG. 4 shows a pixel circuit provided by the second embodiment of the present invention. The difference from the first embodiment lies in the arrangement of the capacitance C _F of the NCG for feedback. A source terminal of the MOS transistor Tr is connected to a cathode terminal of a photodiode (PD), which is a light receiving element. An amplifier 31 having a connection point as an input outputs a voltage (V _log ) that depends on the logarithm of the photocurrent that varies based on the intensity of incident light into a photodiode (PD). An inverting amplifier comprising capacitors C1, C2 and amplifier 32 is connected to the output of the sensor circuit. The output of amplifier 32 is connected to detectors 33 and 34, which can detect positive and negative brightness changes, respectively. The initial program circuit 25 controls to turn on the switch SW, and outputs an event signal for post-processing of the peripheral circuit. In the second embodiment, the inverting amplifier 36 is connected to the output terminal of the amplifier 31, and the output terminal is connected to the logarithmic conversion output LOG-OUT through the capacitor _CF.

[equation 3]

C _NEG ＝C _F (1-A ₁ A ₃ )(4)

According to the second embodiment, by using an existing inverting amplifier, negative capacitance can be formed while suppressing circuit scale expansion. Therefore, the effective load capacitance in the photodiode (photodiode, PD) including the parasitic capacitance (C _g ) of the gate terminal of the MOS transistor Tr can be further reduced.

Fig. 5 shows a modification of the pixel circuit provided by the second embodiment. The configuration in which the amplifier 31 and the inverting amplifier 36 in the second embodiment described above are realized by one differential amplifier. A source terminal of the MOS transistor Tr is connected to a cathode terminal of a photodiode (PD). The logarithmic conversion output LOG-OUT as a connection point is connected to the inverting input terminal of the differential amplifier 41, and the non-inverting output terminal is connected to the gate terminal of the MOS transistor Tr and the capacitor C1. A reference potential is applied to a non-inverting input terminal of a differential amplifier 41 serving as an inverting amplifier for feedback control. The non-inverting output terminal of differential amplifier 41 is connected to a sample-and-hold circuit comprising an inverting amplifier and amplifier 42 configured as capacitors C1 and C2. The output of the amplifier 42 is connected to detectors 43 and 44 that output event signals from which an initial program circuit 45 outputs event signals for post-processing by peripheral circuits.

In a modification, the inverting output terminal of the differential amplifier 41 is connected to the logarithmic conversion output LOG-OUT through the capacitor _CF. Since the differential amplifier 41 is used as an inverting amplifier having a gain of −A1, it is not necessary to add an inverting amplifier in the modification.

FIG. 6A shows a circuit diagram of the inverting amplifier 36 of the pixel circuit provided by the second embodiment. In addition, FIG. 6B shows a circuit diagram of the differential amplifier 41 of the pixel circuit provided by the second embodiment. A differential amplifier is shown in parallel with an inverting amplifier, which is commonly used in pixel circuits.

third embodiment

FIG. 7 shows a pixel circuit provided by the third embodiment of the present invention. The difference from the first embodiment is that a buffer amplifier 56 is inserted between the output terminal of the amplifier 51 and the capacitor C1. Buffer amplifier 56 may, for example, adjust the signal frequency band and operating point.

Fourth embodiment

In the fourth embodiment, for example, a feedback capacitance C _F of a negative capacitance generation circuit (NCG) used in a pixel circuit is described in detail.

FIG. 8A shows the structure of the capacitor _CF of the NCG provided by the fourth embodiment of the present invention. Interlayer capacitors are provided in the four metal wiring layers 61 to 64 of the substrate configured as pixel circuits. Electrodes 65a and 65b of capacitor _CF are arranged in layers 62 and 63, and metal films 66a to 66f for shielding are arranged around them. According to this configuration, the capacitor _CF of the NGC can be provided in the substrate configured as a pixel circuit.

FIG. 8B shows another structure of the capacitance of the NCG in the fourth embodiment. In-layer capacitors are provided in three metal wiring layers 71 to 73 of a substrate configured as a pixel circuit. Electrodes 74a and 74b of capacitor _CF are arranged in layer 72, and metal films 75a to 75d for shielding are arranged around them. Also according to this configuration, the NCG capacitor _CF can also be provided in the substrate configured as a pixel circuit.

FIG. 9 shows the structure of the variable capacitance of the NCG in the fourth embodiment. In the variable capacitance circuit, a plurality of capacitance taps 81 ₀ to 81 _n in which capacitance elements CAP_0 to CAP_N, fuses FUSE_0 to FUSE_n, and test switches TEST_0 to TEST_n are connected in parallel are connected in series. Correspondingly, the capacitor tap 81 ₀ is connected to the power supply and ground through the surge switch ROW_SEL, and the capacitor tap 81 _n is connected to the power supply and ground through the surge switch COL_SEL.

Once the required capacitance value of the feedback capacitance C _F of the pixel circuit is determined, the capacitance taps are selected so that the total capacitance value of the selected capacitive elements meets the required capacitance value. The test switches for the selected capacitor taps are opened, and the test switches for the non-selected capacitor taps are shorted. Next, the inrush switches ROW_SEL and COL_SEL are shorted and the fuse for the selected capacitor tap is disconnected. A capacitive element of the required capacitance value for capacitance C _F is connected in series between terminals 82 and 83 once all test switches and surge switches have been opened.

According to the present embodiment, by arranging the feedback capacitance C _F in the metal wiring layer, the NCG can be configured without externally connecting the capacitance element to the conventional pixel circuit.

Claims (9)

1. A sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element, and having a first inverting amplifier having an input terminal connected to the connection point for outputting a voltage that depends on a logarithm of photocurrent of the light receiving element, wherein an output terminal of the first inverting amplifier is connected to a gate terminal of the MOS transistor and the gate terminal is adjusted to a voltage that depends on the logarithm of photocurrent, the sensor circuit comprising:

a second inverting amplifier having an input connected to the output of the first inverting amplifier and an output connected to the connection point through a capacitor.

2. A sensor circuit characterized by having a connection point between a source terminal of a MOS transistor and a light receiving element, and having an amplifier having an input terminal connected to the connection point, the connection point being for outputting a voltage that depends on a logarithm of a photocurrent of the light receiving element, wherein an output terminal of the amplifier is connected to a gate terminal of the MOS transistor, and the gate terminal is regulated to a voltage that depends on the logarithm of the photocurrent, wherein

The amplifier is a differential amplifier having an inverting input terminal connected to the connection point, a non-inverting output terminal connected to the gate terminal, and an inverting output terminal connected to the connection point through a capacitor.

3. The sensor circuit according to claim 1 or 2, wherein the capacitor is formed in a metal wiring layer.

4. A pixel circuit, comprising:

the sensor circuit of claim 1 or 2;

a sample-and-hold circuit connected to an output of the sensor circuit;

a reset circuit for resetting the hold value of the sample-and-hold circuit in case a difference between an output of the sensor circuit and the hold value of the sample-and-hold circuit exceeds a given threshold voltage.

5. A pixel circuit, comprising: a sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element, and having a first inverting amplifier having an input terminal connected to the connection point for outputting a voltage that depends on a logarithm of a photocurrent of the light receiving element, wherein an output terminal of the first inverting amplifier is connected to a gate terminal of the MOS transistor, and the gate terminal is adjusted to a voltage that depends on the logarithm of the photocurrent; and has a second inverting amplifier connected to an output terminal of the sensor circuit and configured to operate as a sample-and-hold circuit to hold an input voltage and amplify a variation amount of the output terminal of the sensor circuit, the pixel circuit including:

a capacitance interposed between the output of the second inverting amplifier and the connection point.

6. The pixel circuit of claim 5, further comprising:

a reset circuit for resetting the hold value of the sample-and-hold circuit in case a difference between an output of the sensor circuit and the hold value of the sample-and-hold circuit exceeds a given threshold voltage.

7. The pixel circuit according to any one of claims 4, 5, and 6, further comprising:

a buffer amplifier inserted before the second inverting amplifier.

8. A pixel circuit according to any one of claims 4 to 7, wherein the capacitance is a capacitance formed in a metal wiring layer.

9. A method of controlling a pixel circuit, comprising: a sensor circuit having a connection point between a source terminal of a MOS transistor and a light receiving element, and having a first inverting amplifier having an input terminal connected to the connection point for outputting a voltage that depends on a logarithm of a photocurrent of the light receiving element, wherein an output terminal of the first inverting amplifier is connected to a gate terminal of the MOS transistor, and the gate terminal is adjusted to a voltage that depends on the logarithm of the photocurrent; and having a second inverting amplifier connected to an output of the sensor circuit and operative as a sample-and-hold circuit to hold an input voltage and amplify a variation of the output of the sensor circuit, the method being performed by an initial program circuit of the pixel circuit, comprising the steps of:

sensing a reset signal for resetting the hold value of the sample-and-hold circuit and sending an event signal to an arbiter circuit in case the difference between the output of the sensor circuit and the hold value of the sample-and-hold circuit exceeds a given threshold voltage;

the reset signal is released upon completion of communication with the arbiter circuit.

Images