Image sensor

Abstract

An image sensor is provided that operates with low power consumption, high speed, and high precision. The CMOS image sensor (10) is configured to include: a pixel unit (1) which is composed of a plurality of pixels (1a) arranged in two dimensions along the row and column directions, each pixel (1a) having a sensor element that detects physical quantities present in nature and converts them into electrical signals; a resistive digital-to-analog converter (8) which is composed of a plurality of unit circuits connected in parallel for generating a ramp wave, wherein a resistor is connected to the output terminal of a CMOS inverter in the unit circuit; and an analog-to-digital converter (5) which has a plurality of integrating analog-to-digital converters (5a) for comparing the signal from the pixel (1a) with the ramp wave and converting it into a digital signal.

Description

Technical Field

[0001] This invention relates to an image sensor. Background Technology

[0002] Previously, a representative image sensor was the CMOS (Complementary Metal Oxide Semiconductor) image sensor. Figure 15 This is a block diagram illustrating the structure of a conventional CMOS image sensor. (For example...) Figure 15 As shown, in a conventional CMOS image sensor 100, pixels 101a are arranged in two dimensions along the horizontal and vertical directions in the pixel section 101. The vertical control circuit 102 selects the pixel 101a of any row by setting one of the access lines 103 to "H".

[0003] Furthermore, the pixels 101a of the selected row simultaneously output a voltage corresponding to the brightness of the pixel. This voltage is input via pixel signal line 104 to each of the integrating A / D converters in A / D conversion unit 105, which is equipped with multiple integrating analog-to-digital converters (hereinafter, analog-to-digital conversion is referred to as A / D conversion, and in the figure, A / D converter is represented as ADC). Then, it is converted into a digital signal in A / D conversion unit 105 and output from the output terminal via level control circuit 106.

[0004] Typically, an integrating A / D converter is used for A / D conversion. This integrating A / D converter uses a ramp wave generated by a current-mode digital-to-analog converter (hereinafter, digital-to-analog conversion is referred to as D / A conversion, and in the diagram, D / A converter is represented as DAC) and a counter to measure the number of clock cycles. Figure 16 This is a circuit diagram showing the basic structure of an integrating A / D converter used in a CMOS image sensor. Figure 17 This is a diagram showing the waveform of the ramp wave input to the A / D converter.

[0005] like Figure 16 As shown, in the case of A / D conversion using an integrating A / D converter, firstly, the switch S between the input and output of comparator 111 is closed. At this time, the input voltage V... in Applying a reference voltage V in_0 Typically, a pixel-side reference voltage is applied to the gate of the source follower within pixel 101a, and the source voltage is usually set to V. in_0 At this time, for the reference voltage V ref Provide the reference output voltage to the D / A converter. In this state, open switch S to allow input from the input voltage V. inThe signal of pixel 101a reflects the brightness.

[0006] Next, as Figure 17 As shown, the D / A converter is controlled to generate a falling ramp wave. Counter 112 is clocked after an initial reset, thus starting the clock count. Then, the input reference voltage is compared with the input voltage V... in The difference is the difference between the reference voltage and the reference voltage V of the D / A converter. ref When the difference signals are consistent, the output of comparator 111 inverts, counter 112 stops, and the count value at this time is output as the A / D conversion value. However, to ensure conversion accuracy for weak signals, a reference voltage V is used. ref Most of the time, the reference voltage of the D / A converter temporarily rises by V. off Then it decreases. This is because the offset time T is subtracted from the conversion time Tc. off The obtained time and input voltage V in It is proportional to the input voltage V, therefore the conversion time Tc can be used to obtain the input voltage V. in The A / D conversion value.

[0007] In addition, the integrating A / D converter used in image sensors requires a ramp wave, but this ramp wave is mostly formed by the D / A converter (for example, see Patent Documents 1-3 and Non-Patent Document 1). Figure 18 This is a circuit diagram illustrating the structure of a current-type D / A converter used in conventional CMOS image sensors. For example... Figure 18 As shown, a number of unit current sources 122 are provided in a typical current-type D / A converter in the past.

[0008] In this current-type D / A converter, a switch 123, controlled according to the input signal decoded by the decoder 121, switches the direction of current flow to either the load resistor 124 side or the power supply 125 side, thereby controlling the current value flowing through the load resistor 124 to generate a voltage in the load resistor 124. Then, by gradually increasing the current flowing through the load resistor 124 over time, a ramp wave can be obtained as the output.

[0009] Existing technical documents

[0010] Patent documents

[0011] Patent Document 1: International Publication No. 2013 / 122221

[0012] Patent Document 2: Japanese Patent Application Publication No. 2013-239951

[0013] Patent Document 3: Japanese Patent Application Publication No. 2018-148541

[0014] Non-patent literature

[0015] Non-Patent Literature 1: S. Yoshihara, et al., “A 1 / 1.8-inch 6.4MPixel 60frame s / s CMOS Image Sensor With Seamless Mode Change”, IEEE Journal of Solid-State Circuits, December 2006, Vol.41, No.12, pp.2998-3006 Summary of the Invention

[0016] The problem the invention aims to solve

[0017] However, the aforementioned conventional current-mode D / A converters have the following problems. The first problem is power consumption. Figure 19 This is a circuit diagram illustrating the structure of a unit current source in a current-mode D / A converter. (Example:) Figure 19 As shown, the unit current source of the current-type D / A converter includes a transistor M1 that determines the current value, a cascaded transistor M2 that enhances constant current to improve linearity, and transistors M3 and M4 that function as switches to switch the current path.

[0018] In the case of a CMOS image sensor, the output voltage V out voltage amplitude V s The maximum is around 1.2V. Since transistors M1 and M2 need to operate in the saturation region, the voltage V between the drain and source is... DS1 V DS2 The minimum required is 0.3V. Therefore, the power supply voltage V DD A voltage of 1.8V is required. Here, if the load resistance is set to R... L Then the current I flowing in the D / A converter DAC Represented by the following numerical expression 1.

[0019] [Formula 1]

[0020]

[0021] In addition, the power consumption P of the D / A converter DAC Represent it using the following numerical expression 2.

[0022] [Formula 2]

[0023]

[0024] In recent years, the load capacitance has increased due to the increase in the number of pixels and the required frame rate. However, in order to ensure a constant response time constant, it is necessary to reduce the load resistance R. L Therefore, the power consumption of D / A converters tends to increase, making power consumption reduction a major challenge. Furthermore, since image sensors are temperature-sensitive, dark current increases significantly as operating temperature rises; therefore, from the perspective of image quality, it is strongly desired to suppress power consumption as much as possible.

[0025] The second issue is response speed. If the time response characteristics of the D / A converter are insufficient, it will hinder the high-speed operation of the image sensor. Figure 20 This is a diagram showing the waveform of a ramp wave under the condition of multiple scans of a small voltage range. In recent years, a method has been proposed as follows: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] Figure 20 The method involves scanning a small voltage range of approximately 50mV and performing A / D conversions multiple times, averaging the resulting conversion values ​​to reduce conversion noise. However, in this method, the time response characteristics of the D / A converter are insufficient, making it difficult to perform multiple A / D conversions within a certain time frame.

[0026] Figure 21 This is a diagram showing the ideal and actual waveforms of the ramp wave generated using a D / A converter. (Example) Figure 21 As shown, in conventional D / A converters, even generating a 50mV ramp wave at 50ns results in waveform distortion, and linearity can only be ensured within the 40ns and 40mV region. Therefore, conventional D / A converters require generating ramp waves with greater margins, which hinders high-speed operation.

[0027] Therefore, the object of the present invention is to provide an image sensor that performs actions with low power consumption, high speed and high precision.

[0028] Solution for solving the problem

[0029] In order to solve the above-mentioned problems, the inventors studied digital-to-analog converters that generate ramp waves in image sensors and found that resistive digital-to-analog converters obtained by connecting a resistor connected in parallel with the output of an inverter have inherently lower power consumption than conventional current-type digital-to-analog converters that use a current source, thus completing the present invention.

[0030] That is, the image sensor according to the present invention includes: a pixel unit, which is composed of a plurality of pixels arranged in two dimensions along the row and column directions, each of the pixels having a sensor element for detecting physical quantities present in nature and converting them into electrical signals; a resistive digital-to-analog converter, which is composed of a plurality of unit circuits connected in parallel for generating a ramp wave, wherein a resistor is connected to the output terminal of a CMOS inverter in the unit circuit; and an analog-to-digital converter, which includes a plurality of integrating analog-to-digital converters for comparing the signal from the pixels with the ramp wave and converting it into a digital signal.

[0031] Alternatively, the resistive digital-to-analog converter may include: a high-bit conversion unit, which is obtained by connecting a number of unit circuits in parallel corresponding to the number of high-bits, wherein one end of a resistor is connected to the output of a CMOS inverter and the other end of the resistor is connected to the output of the unit circuit; and a low-bit conversion unit, which is obtained by connecting a number of unit circuits in parallel corresponding to the number of low-bits, wherein one end of a resistor is connected to the output of a CMOS inverter and the other end of the resistor is connected to a resistor between the terminals of the unit circuit.

[0032] In the transistor of the CMOS inverter in the unit circuit, the channel length can be set to 90 nm or less.

[0033] Alternatively, the resistive digital-to-analog converter may be positioned at both ends of the signal line used to supply the ramp wave to the analog-to-digital converter.

[0034] Furthermore, the inventors analyzed the time response when a ramp wave is generated by a digital-to-analog converter and discovered a method for generating ramp waves that avoids waveform distortion by controlling the offset voltage.

[0035] That is, the image sensor of the present invention may also be able to input a constant offset value to the resistive digital-to-analog converter when the voltage of the ramp wave changes over time.

[0036] Alternatively, when the time-varying rate of change of the voltage of the ramp wave changes multiple times with time intervals, the offset value can be made to change according to the change in the time-varying rate.

[0037] Alternatively, the offset value can be calculated based on a first reference voltage, a second reference voltage different from the first reference voltage, a first moment when the voltage of the ramp wave becomes the first reference voltage, and a second moment when the voltage of the ramp wave becomes the second reference voltage.

[0038] The effects of the invention

[0039] According to the present invention, a ramp wave is generated using a resistive digital-to-analog converter, which significantly reduces the average power consumption compared to a current-type D / A converter with the same output resistance, thereby enabling an image sensor to operate with low power, high speed and high precision. Attached Figure Description

[0040] Figure 1 This is a block diagram illustrating the structure of an image sensor according to the first embodiment of the present invention.

[0041] Figure 2 A is shown Figure 1 The circuit diagram shown is an example of the structure of a resistive D / A converter 8. Figure 2 B is a circuit diagram showing the inverter 81 of the resistive D / A converter 8.

[0042] Figure 3 It is used to find Figure 1 The circuit diagram shown illustrates the current consumption and power consumption of the resistive D / A converter 8.

[0043] Figure 4 It shows from Figure 1 The circuit diagram shown is the equivalent circuit of the resistive D / A converter 8 from the output terminal.

[0044] Figure 5 It is a graph showing the current consumption of a current-type D / A converter, the current consumption of a resistive D / A converter, and the average current consumption, corresponding to the output voltage of the D / A converter.

[0045] Figure 6 This is a circuit diagram showing a D / A converter that generates a ramp wave and a distributed RC circuit that becomes the load.

[0046] Figure 7 This is a circuit diagram illustrating a distributed RC circuit in an image sensor according to a first embodiment of the present invention and a D / A converter driving the distributed RC circuit from both sides of the distributed RC circuit.

[0047] Figure 8 This is a circuit diagram showing the equivalent circuit of a D / A converter considering the load.

[0048] Figure 9 This is a block diagram illustrating the structure of a resistive D / A converter for an image sensor according to a second embodiment of the present invention.

[0049] Figure 10 It is shown in Figure 9 The waveforms of the output voltage and the voltage of the load circuit with capacitance shown in the resistive D / A converter 28 are given when a constant offset value is applied as the voltage of the ramp wave changes over time.

[0050] Figure 11 This shows the output voltage V of a resistive D / A converter without applying a correction value when the rate of change of time changes twice. DAC and the voltage V of the load circuit with capacitance. out The waveform diagram.

[0051] Figure 12 This is a waveform diagram showing the output voltage of the D / A converter and the voltage of the load circuit with capacitance when the offset value changes as the time rate of change of the ramp wave changes multiple times with time and the time rate of change of the voltage of the ramp wave changes accordingly.

[0052] Figure 13 This is a diagram showing the structure of the calibration circuit.

[0053] Figure 14 It is shown Figure 13 The diagram shows the relationship between the output voltage, reference voltage, and time in the calibration circuit shown.

[0054] Figure 15 This is a block diagram showing the structure of a conventional CMOS image sensor.

[0055] Figure 16 This is a circuit diagram showing the basic structure of an integrating A / D converter used in a CMOS image sensor.

[0056] Figure 17 This is a diagram showing the waveform of the ramp wave input to the integrating A / D converter.

[0057] Figure 18 This is a circuit diagram showing a current-type D / A converter used in conventional CMOS image sensors.

[0058] Figure 19 This is a circuit diagram showing the structure of a unit current source in a current-mode D / A converter.

[0059] Figure 20 This is a diagram showing the waveform of a ramp wave when multiple scans of a small voltage range are performed.

[0060] Figure 21 This is a diagram showing the ideal waveform and the actual waveform of the ramp wave obtained using a D / A converter. Detailed Implementation

[0061] The embodiments for carrying out the present invention will now be described in detail with reference to the accompanying drawings. Furthermore, the present invention is not limited to the embodiments described below.

[0062] (First implementation method)

[0063] First, the image sensor according to the first embodiment of the present invention will be described. Figure 1 This is a block diagram illustrating the structure of the image sensor in this embodiment. (As shown...) Figure 1 As shown, the image sensor 10 of this embodiment includes a pixel unit 1 having multiple pixels 1a, an A / D converter 5 that converts pixel signals into digital signals, and a resistive D / A converter 8 that supplies a ramp wave, which serves as a reference voltage, to the A / D converter 5. That is, in the image sensor 10 of this embodiment, the D / A converter used to generate the ramp wave supplied to the A / D converter 5 is not a current-type D / A converter, but a resistive D / A converter 8.

[0064] Additionally, in the image sensor 10 of this embodiment, it is also possible to use, for example... Figure 15 The CMOS image sensor shown is similarly equipped with a vertical control circuit 2 that controls the row access line 3 connected to pixel 1a, a pixel signal line 4 connected to pixel 1a to send pixel signals to the A / D conversion unit 5, a horizontal control circuit 6 that controls the output of the digital signals generated in the A / D conversion unit 5, an overall control circuit 7, and a clock circuit 9, etc.

[0065] [Pixel Part 1]

[0066] In the pixel unit 1, a plurality of pixels 1a are arranged two-dimensionally along the row and column directions. Each pixel 1a of the pixel unit 1 has a sensor element that detects physical quantities existing in nature and converts them into electrical signals. Here, physical quantities existing in nature refer to visible light, infrared light, ultraviolet light, X-rays, electromagnetic waves, electric fields, magnetic fields, temperature, pressure, etc.

[0067] [A / D Conversion Unit 5]

[0068] The A / D conversion unit 5 converts the pixel signals of each pixel 1a from the pixel unit 1 into digital signals by comparing them with the ramp wave from the resistive D / A converter 8. The A / D conversion unit 5 is composed of multiple integrating A / D converters 5a.

[0069] [Resistive D / A Converter 8].

[0070] Figure 2 A is shown Figure 1 The circuit diagram shown is an example of the structure of a resistive D / A converter 8. Figure 2 B is a circuit diagram showing the inverter 81 of the resistive D / A converter 8. Figure 2In the resistive D / A converter 8 shown in Figure A, a unit circuit with a resistor connected in parallel is connected to the output of the inverter 81. Furthermore, the power supply for the inverter 81 in the resistive D / A converter 8 uses a reference voltage V. REF In this resistive D / A converter 8, the decoder circuit 82 receives the input signal, and the decoded signal is input to each inverter 81.

[0071] Regarding the resistive D / A converter 8, for example, the high 2 bits are constructed using a segmented D / A converter employing thermometer codes, and the low 2 bits are constructed using a binary D / A converter employing an R-2R resistor ladder. The resistive D / A converter 8 operates as a 4-bit D / A converter. The segmented D / A converter constituting the high-bit conversion section is obtained by connecting unit circuits in parallel in a number corresponding to the number of high-bit bits, in which the other end of the resistor is connected to the output terminal of the unit circuit.

[0072] On the other hand, the low-bit conversion section, which performs the conversion of low-order bits, can be constructed using a binary D / A converter employing an R-2R resistor ladder. The binary D / A converter is obtained by connecting a unit circuit in parallel with a number corresponding to the number of low-order bits. In this unit circuit, one end of a resistor is connected to the output of a CMOS inverter, and the other end of the resistor is connected to a resistor disposed between the terminals of the unit circuit. In this case, by setting the resistor value to... Figure 2 The ratio shown in A allows for an accurate output voltage. The bit allocation between the high-order and low-order bits can be appropriately set according to the application or specifications, but from the viewpoint of accuracy and area, it is preferable to set both to approximately equal bit numbers.

[0073] As an inverter 81, it can be used with, for example, having Figure 2 The image shows NMOS and PMOS CMOS inverters, as shown in B. Traditional current-mode D / A converters did not use core transistors with tiny gates in their internal logic. Because a minimum voltage rating of 1.8V was required, I / O transistors with a voltage rating of around 3.3V were used. Therefore, current-mode D / A converters not only had a large area but also large capacitance, making high-speed operation difficult and resulting in high power consumption.

[0074] In contrast, the resistive D / A converter 8 used in the image sensor 10 of this embodiment only requires transistors with a withstand voltage of approximately 1.0V to 1.2V. Therefore, it is possible to use, for example, tiny core transistors with a channel length of less than 90nm, minimizing the channel length. This allows the resistive D / A converter to achieve sufficiently low on-resistance even with small transistors, thus enabling good linearity. Consequently, by using a resistive D / A converter, the footprint and power consumption of the D / A converter can be reduced, and processing speed can be increased. Furthermore, by configuring the D / A converter in this way, power consumption can be significantly reduced compared to conventional current-type D / A converters.

[0075] Figure 3 This is a circuit diagram used to determine the current consumption and power consumption of a resistive D / A converter 8. Figure 4 This is a circuit diagram showing the equivalent circuit viewed from the output of the resistive D / A converter 8. (Example) Figure 3 As shown, the resistive D / A converter 8 can convert the output voltage relative to the reference voltage V. REF The conductance on the set side relative to the output terminal is denoted as Gx, and the conductance on the set side relative to the output terminal is denoted as G(1-x). At this time, the output resistance R... L Expressed by the following formula 3, and with output resistance R L Constant.

[0076] [Formula 3]

[0077] (where 0 < x < 1)

[0078] In addition, the output voltage V out It can be expressed by the following formula 4, which is proportional to x.

[0079] [Formula 4]

[0080] V out =V REF ·x

[0081] Therefore, such as Figure 4 As shown, it is possible to output resistance R L Keep the output voltage V constant. out Changes. Furthermore, the current I flowing through the voltage source is represented by the following equation 5.

[0082] [Formula 5]

[0083]

[0084] Therefore, power consumption P D Represent this with the following expression, 6.

[0085] [Formula 6]

[0086]

[0087] As a result, the current flowing in the resistive D / A converter 8 is maximum when x = 0.5, and this maximum current is 1 / 4 of the current flowing in the current-type D / A converter represented by Equation 1 above. This is because the D / A converter used in the image sensor generates a voltage from 0 to the reference voltage V. REF The slope wave between them, therefore the average current I is obtained by the following equation 7. AVE .

[0088] [Formula 7]

[0089]

[0090] Therefore, the current consumption of a resistive D / A converter, which essentially generates a ramp wave, is as small as 1 / 6 of that of a current-mode D / A converter. Furthermore, due to the power supply voltage V in the case of a current-mode D / A converter... DD V REF It is about 0.6V higher, so if the reference voltage V is used... REF Set it to 1.2V, and set the power supply voltage V. DD If we set it to 1.8V, then we can calculate the power consumption ratio using the following formula 8.

[0091] [Formula 8]

[0092]

[0093] Figure 5 This is a graph showing the current consumption of a current-type D / A converter, the current consumption of a resistive D / A converter, and the average current consumption, corresponding to the output voltage of the D / A converter. For example... Figure 5 As shown in Equation 8, if a resistive D / A converter is used, even with the same output resistance, the power consumption can be reduced to 1 / 9 of that of a current-type D / A converter.

[0094] In CMOS image sensors, to improve the image quality in dark scenes, the following processing is sometimes performed: such as Figure 20 As shown, only signals around 0mV to 50mV are converted. Multiple ramp scans are performed as needed to reduce noise, and the average value is taken after multiple conversions. In this case, if the amplitude is set to β times the full scale, the current consumption of the resistive D / A converter is expressed by the following equation 9.

[0095] [Formula 9]

[0096]

[0097] In Equation 9 above, for example, if β is set to 0.05, then β / 2 is 0.025, and the average current is 0.15 times the current when the full scale is scanned as represented by Equation 7 above, which becomes a very small current consumption.

[0098] On the other hand, in the case of a current-type D / A converter, the current is constant regardless of the scan level, so it is impossible to achieve such a reduction in current consumption. Therefore, using a resistive D / A converter in a CMOS image sensor is very beneficial in terms of reducing power consumption. Depending on the situation, in addition to setting up a D / A converter for full-scale scanning, sometimes a resistive D / A converter is also set up for full-scale scanning. Figure 20 The D / A converter shown is a partial voltage scan, but even with such a D / A converter, the increase in power consumption can be suppressed extremely effectively by using a resistive D / A converter.

[0099] like Figure 15 As shown, in a CMOS image sensor, the output of the D / A converter is supplied to multiple comparators distributed in space. Figure 6 This is a circuit diagram showing a D / A converter that generates a ramp wave and a distributed RC circuit that becomes the load. The load circuit is precisely converted into resistors and capacitors as shown. Figure 6 The circuit exhibits an RC distributed constant as shown. Therefore, at the drive and open terminals of the D / A converter, the signal is delayed, resulting in a reduction in amplitude. Here, if the resistance per unit length is set to R... u Set the capacitor to C. u If the length is set to L, then the reference time constant τ of the RC distributed constant circuit is expressed by the following formula 10.

[0100] [Formula 10]

[0101]

[0102] The longer the reference time constant τ is, the greater the impact. Figure 7 This is a circuit diagram showing the distributed RC circuit in the image sensor of this embodiment and the D / A converter driving the distributed RC circuit from both sides. In the image sensor 10 of this embodiment, it is also possible to use... Figure 7 As shown, ramp waves are supplied to the comparators from both sides. Therefore, the length L of the RC distributed constant circuit corresponding to each D / A converter is effectively halved, thus shortening the time constant to 1 / 4 and reducing its impact. As a result, higher accuracy and faster speed A / D conversion can be achieved. Furthermore, the load capacitance of each D / A converter is effectively halved, and since the output resistance can also be doubled, there is almost no increase in overall power consumption.

[0103] As described in detail above, the image sensor of this embodiment includes a resistive D / A converter and an A / D conversion unit. The resistive D / A converter is formed by connecting a unit circuit with a resistor connected in parallel to the output terminal of a CMOS inverter to generate a ramp wave. The A / D conversion unit is composed of multiple A / D converters that compare the signal from the pixel with the ramp wave and convert it into a digital value. Therefore, compared with conventional CMOS image sensors, power consumption can be significantly reduced.

[0104] (Second implementation method)

[0105] Next, the image sensor according to the second embodiment of the present invention will be described. Figure 8 This is a circuit diagram showing the equivalent circuit of a D / A converter considering the load. One challenge in D / A converters implemented in CMOS image sensors is generating high-speed ramp waves. For example... Figure 21 As shown, in D / A converters, waveform distortion limits the usable time range, making high-speed conversion difficult. This is because, as... Figure 8 As shown, a capacitor C exists in the circuit. L The response of such a circuit when a ramp wave is input is represented by the following equation 11.

[0106] [Formula 11]

[0107] V DAC (t)=kt

[0108] Furthermore, if the voltage V in the load circuit that takes the waveform represented by the above equation 11 as input... out Performing a Laplace transformation, it becomes the following expression 12.

[0109] [Formula 12]

[0110] (where τ=R) L C L )

[0111] Then, if the time response is obtained by applying the inverse Laplace transform to the above equation 12, it becomes the following equation 13.

[0112] [Formula 13]

[0113]

[0114] In Equation 13 above, the first term represents the ideal ramp wave, and the second term represents the voltage error. Since this voltage error represents the response of the step wave, the offset voltage V... off Represent this using the following numerical expression, 14.

[0115] [Formula 14]

[0116] V off =kτ

[0117] Therefore, it can be seen that if the offset voltage V represented by Equation 14 above is applied when the output voltage of the D / A converter changes, off This can offset the changes in output voltage. Figure 9 This is a block diagram illustrating the structure of the resistive D / A converter in the image sensor of this embodiment. Therefore, in the image sensor of this embodiment, when generating a ramp wave in the D / A converter, instead of performing clock addition or subtraction operations on the set initial value, it is as follows: Figure 9 As shown, when an offset voltage V, as represented by Equation 14 above, is applied... off After applying a corresponding correction value (offset value), clock addition or subtraction is performed. This solves the problem of generating high-speed ramp waves.

[0118] Figure 10 It is shown in Figure 9 The output voltage V of the resistive D / A converter 28 shown is given a constant offset value when the time rate of change of the ramp wave voltage changes. DAC and the voltage V of the load circuit with capacitance. out The waveform diagram. For example... Figure 10 As shown, it can be seen that by applying an offset voltage V when the time rate of change of the voltage of the ramp wave changes... off A suitable correction value can generate an accurate ramp wave.

[0119] Furthermore, even when the rate of change of the ramp wave varies multiple times with time, an accurate ramp wave can be generated by changing the offset value accordingly. Figure 11 This shows the output voltage V of a resistive D / A converter without applying a correction value when the rate of change of time changes twice. DAC and the voltage V of the load circuit with capacitance. out The waveform diagram. For example... Figure 11 As shown, if the time when the ramp wave is generated is set to 0s, then the time rate of change will change, resulting in an error V. error Waveform distortion occurs. Increasing the rate of change of time by a factor of four within 50 ns produces a large error V. error This results in significant waveform distortion.

[0120] Figure 12 This shows the output voltage V of the D / A converter when the offset value changes as the time rate of change of the ramp wave varies multiple times with time, and thus the time rate of change of the ramp wave voltage varies accordingly.DAC and the voltage V of the load circuit with capacitance. out The waveform diagram. For example... Figure 12 As shown, if the time when the ramp wave is generated is set to 0s, then the change in the rate of change of time at this time will result in an error V. error The error V is 0. When the rate of change of time is increased fourfold within 50 ns, the error V increases by changing the correction value. error When the value becomes 0, no waveform distortion occurs, indicating that this method can effectively suppress waveform distortion.

[0121] Thus, in CMOS image sensors using ramp waves whose rate of change varies multiple times with each passing moment, if the signal strength increases to some extent, the rate of change of the ramp wave can be increased, thereby shortening the conversion time and increasing the frame rate. This offers the advantage of achieving high speed or low power consumption due to the shortened conversion time.

[0122] On the other hand, when applying the above method to actual image sensors, it is difficult to calculate the time constant of the load in advance, so a calibration circuit is required. Figure 13 This is a diagram showing the structure of the calibration circuit. Figure 14 This is a graph showing the relationship between the output voltage, reference voltage, and time of the calibration circuit. (Example) Figure 13 As shown, the calibration circuit consists of the output voltage of the resistive D / A converter 28, which is represented in the load circuit, two reference voltages (a first reference voltage and a second reference voltage), a calibration A / D converter 23, and a calibration logic circuit 24.

[0123] exist Figure 14 In the diagram, an ideal ramp wave is represented by a dashed line. If the voltage at time 0 is set to 0, then as shown in... Figure 14 As shown by solid lines, the actual response deviates according to the RC time constant. The voltage is positive, and the dashed lines represent auxiliary lines for the negative portions. In this state, if the voltage at time T1 is set as V1 and the voltage at time T2 as V2, the offset voltage V can be calculated using the following equation 15. off .

[0124] [Formula 15]

[0125]

[0126] Therefore, in the image sensor of this embodiment, the offset voltage V calculated by the above-described formula 15 is used. off It can be applied as a correction value. Specifically, such as Figure 13As shown, the output of the resistive D / A converter 28 is compared with reference voltages V1 and V2. The time-domain correction A / D converter 23, which consists of a comparator 21 and a counter 22 that outputs the time information at this time, is used to calculate the times T1 and T2 based on the count value. Then, the correction value is calculated according to the above formula 15, and the necessary offset voltage (offset value) is output from the correction logic circuit 24.

[0127] Furthermore, the correction value calculated by the correction logic circuit 24 is supplied to the addition / subtraction arithmetic unit 20, which is used to calculate the input value of the output resistive D / A converter 28. The output of the resistive D / A converter 28 is then compared with the reference voltages V1 and V2. The correction A / D converter 23 is used to calculate the times T1 and T2 based on the count value, thus asymptotically approaching the ideal value more accurately. Alternatively, the correction value can be calculated based on a single voltage and time, but in this case, errors caused by the comparator's offset voltage or delay are more likely to occur. Therefore, the method using two voltages and two times is more accurate.

[0128] Thus, the resistive D / A converter in the image sensor of this embodiment reduces waveform distortion of the ramp wave by imparting a constant offset value when the voltage of the ramp wave changes over time, thereby achieving high-precision and high-speed A / D conversion. Furthermore, the structure and effects of the image sensor in this embodiment other than those described above are the same as those in the first embodiment described above.

[0129] Furthermore, while the first and second embodiments described above used a CMOS image sensor as an example, the present invention is not limited to this and can also be applied to two-dimensional image sensors for other purposes. Moreover, the image sensor of the present invention includes infrared sensors, terahertz sensors, magnetic sensors, and pressure sensors, etc.

[0130] Explanation of reference numerals in the attached figures

[0131] 1. 101: Pixel unit; 1a, 101a: Pixel; 2. 102: Vertical control circuit; 3. 103: Horizontal access line; 4. 104: Pixel signal line; 5. 105: A / D conversion unit; 5a: Integrating A / D converter; 6. 106: Horizontal control circuit; 7: Overall control circuit; 8. 28: Resistive D / A converter; 9: Clock circuit; 10. 100: CMOS image sensor; 20: Adder / subtractor; 21. 111: Comparator; 22. 112: Counter; 23: A / D converter for calibration; 24: Calibration logic circuit; 81: Inverter; 82: Decoder circuit; 121: Decoder; 122: Current source; 123: Switch; 124: Load resistor; 125: Power supply.

Claims

1. An image sensor, comprising: The pixel unit is a two-dimensional arrangement of multiple pixels along the row and column directions, and each pixel has a sensor element that detects physical quantities present in nature and converts them into electrical signals. A resistive digital-to-analog converter, which consists of multiple unit circuits connected in parallel to generate a ramp wave, wherein a resistor is connected to the output of a CMOS inverter in each unit circuit; and The analog-to-digital converter includes multiple integrating analog-to-digital converters for comparing the signal from the pixel with the ramp wave and converting it into a digital signal. in, As the time rate of change of the voltage of the ramp wave changes, an offset value proportional to the time rate of change of the voltage is input to the resistive digital-to-analog converter.

2. The image sensor according to claim 1, characterized in that, The offset value is equivalent to the product of the voltage's rate of change over time and the response time constant.

3. The image sensor according to claim 1, characterized in that, The resistive digital-to-analog converter has the following features: The high-order bit conversion section is obtained by connecting a number of unit circuits in parallel, corresponding to the number of high-order bits. In this unit circuit, one end of a resistor is connected to the output of a CMOS inverter, and the other end of the resistor is connected to the output of the unit circuit. The low-order bit conversion section is obtained by connecting a number of unit circuits in parallel, corresponding to the number of low-order bits. In this unit circuit, one end of a resistor is connected to the output of the CMOS inverter, and the other end of the resistor is connected to a resistor between the terminals of the unit circuit.

4. The image sensor according to claim 1, characterized in that, The resistive digital-to-analog converter is positioned at both ends of the signal line used to supply the ramp wave to the analog-to-digital converter.

5. The image sensor according to claim 1, characterized in that, The rate of change of the voltage of the ramp wave changes multiple times with time, and the offset value also changes proportionally to the change in the rate of change of time.

6. The image sensor according to claim 1, characterized in that, The offset value is calculated based on a first reference voltage, a second reference voltage different from the first reference voltage, a first moment when the voltage of the ramp wave becomes the first reference voltage, and a second moment when the voltage of the ramp wave becomes the second reference voltage.

Images