JP4770577B2 - Solid-state imaging device - Google Patents

Description

  The present invention relates to a solid-state imaging device.

  The operation of the comparator in the solid-state imaging device in which the analog-digital converter (A / D converter) is arranged in parallel in the conventional column is shown in the operation timing chart of FIG.

  The conventional A / D converters arranged in parallel in the column are intended to ensure that the output signal value with respect to the input signal stably has ideal conversion characteristics. For example, as shown in FIG. After the resetting of the comparator is completed, the reference voltage Vsaw is shifted to the negative side to eliminate the use of the nonlinear part of the conversion characteristic, so that a linear conversion characteristic can be obtained (Patent Document) 1). This is characterized in that an A / D conversion is performed by providing an offset to the reference voltage by a function inside the reference voltage generation circuit. In the drawing, RS is a read switch element, RD is a reset switch element, Vsig is an output voltage of the current amplifier, S1 to S3 are switch elements, Vsaw is a reference voltage (for example, reference voltage RAMP), VM is an input voltage on the capacitor side, Vout is an output voltage of the inverter, Vth is a threshold voltage of the inverter, VH is a high level voltage of Vout, and VL is a low level voltage of Vout.

  Further, an example of a comparator that improves the analog-digital conversion characteristics by eliminating the influence of the parasitic capacitance is shown in the circuit configuration diagram of FIG. As shown in FIG. 10, a buffer amplifier A2 is connected in series to the input side of the inverter A1, and switches S3A and S3B are connected in series between the input side of the buffer amplifier A2 and the output side of the inverter A1. A configuration in which the switch S3C is connected between the connection point of the switches S3A and S3B and the ground line is disclosed. In this circuit configuration, when resetting, the switches S3A and S3B are turned on, and when the reference voltage Vsaw changes, the switch S3C is turned on so that the change in the output voltage does not reach the input side (see Patent Document 1). .)

  However, with such a circuit configuration, the circuit scale of the reference voltage generation circuit is increased by providing a function for adding an offset to the reference voltage generation circuit. Further, since the reference voltage Vsaw enters all the comparators of the column parallel A / D converter, a large amount of input capacitance is charged, and it takes time until the reference voltage becomes stable, and linear A / D In order to perform D conversion, A / D conversion time increases.

JP 2000-286706 A

  The problem to be solved is that, in order to provide A / D conversion by providing an offset to the reference voltage, the circuit scale of the reference voltage generation circuit increases as the reference voltage generation circuit has a function of adding an offset. In addition, it takes time for the reference voltage to stabilize, and further, in order to obtain linear conversion characteristics, the time for A / D conversion increases.

  It is an object of the present invention to perform linear A / D conversion at high speed without increasing the circuit scale.

  In the solid-state imaging device of the present invention, the solid-state imaging device includes a plurality of pixels that convert an incident light amount into an electrical signal, and a plurality of analog-digital conversion devices that convert an analog signal obtained from the pixel into a digital signal. The analog-to-digital conversion device includes a comparator that compares a reference signal with an analog signal obtained from a pixel, and the comparator generates an offset between the reference signal component and the analog signal component obtained from the pixel. It is characterized by that.

  In the solid-state imaging device, since the comparator generates an offset between the reference signal component and the analog signal component obtained from the pixel, the offset is forcibly added after the reset of the comparator is released.

  According to the solid-state imaging device of the present invention, since the offset is forcibly added after the reset of the comparator is released, there is an advantage that reliable A / D conversion can be performed. This also makes it possible to compare the reference voltage signal in a period in which the reference voltage signal is linear, and to obtain A / D conversion characteristics having linearity. Further, since it is not necessary to provide a period for charging extra charges, there is an advantage that A / D conversion can be performed at high speed.

  One embodiment of the solid-state imaging device of the present invention will be described with reference to the circuit diagram of FIG. 1 and the block diagram of FIG. FIG. 1 shows a comparator of an analog / digital conversion device in a CMOS image sensor as a solid-state imaging device for explaining the first embodiment of the present invention, and FIG. 2 shows the CMOS image sensor.

  First, as shown in FIG. 2, the solid-state imaging device 1 includes a plurality of unit pixels (hereinafter simply referred to as pixels) 101a, 101b,... Thus, the pixel array 11 is configured. Each of the pixels 101x is composed of, for example, a photodiode and an in-pixel amplifier.

  A column parallel analog / digital converter (hereinafter abbreviated as a column parallel A / D converter) 12 is a reference signal (reference voltage) RAMP generated from a reference signal generator 104 that generates a reference signal (reference voltage). For each row line H0, H1,... (Represented as Hx) and an analog signal obtained from the unit pixel 101x via the column lines V0, V1,... (Represented as Vx). For example, a comparator 102x (102a, 102b,...) Arranged for each column line V0, V1,... And a counter (for example, an up / down counter) that counts the comparison time until the comparison is completed and holds the result. ) 103x (103a, 103b,...) And has an n-bit (n is a natural number) digital signal conversion function.

  The horizontal output line 16 includes an output line of n (n is a natural number) bit width, and n sense circuits and output circuits corresponding to the respective output lines. Further, as a control circuit for sequentially reading out signals from the pixel array 11, a timing control circuit 13 for generating an internal clock, a row scanning circuit 14 for controlling a row address and a row scanning, and a column scanning for controlling a column address and a column scanning. A circuit 15 is arranged.

  The counter 103x preferably has an up / down counter configuration in order to simplify the configuration. Furthermore, it is more preferable to provide the memory means for holding the count result because the column parallel A / D converter 12 can perform comparison and parallel operation of the count and the output.

  Next, the comparator 102 will be described with reference to the block diagram of FIG.

  As shown in FIG. 1, the comparator 102 x compares the analog signal obtained from the pixel and the reference voltage RAMP generated from the reference signal generation unit 104. The comparator 102 has a differential amplifier 109 at least at the first stage. The input side of the differential amplifier 109, that is, the side to which the reference signal RAMP is input is the input element 111 and the side to which the analog signal from the pixel is input. An input element 112 is provided, and a capacitive element 107 is provided between the RAMP input and the input element 111, and a capacitive element 108 is provided between the analog input and the input element 112. The input element 111 includes a switch element (reset switch) 105, the input element 112 includes a switch element (reset switch) 106, and the switch elements 105 and 106 are asymmetrically configured. For example, the switch elements 105 and 106 are formed of different sizes. For example, assuming that the switch element 106 has a gate length L and a gate width W, the switch element 105 is formed to have a gate width N * W which is N times, for example, and the gate length is formed to L.

  That is, the switch elements (reset switches) 105 and 106 have the size of the switch element 106 on the VSLin side of W / L and the size of the switch element 105 of the VREFin side is (N * W) / L, and are obtained from the pixels. The switch element 105 which is a reset switch connected to the element to which the reference voltage VREF is input is larger than the switch element 106 which is a reset switch connected to the element to which the analog signal is input. As a result, a capacitance is added to the switch element 105 on the VREFin side.

  Note that the asymmetric configuration can also be obtained by forming the switch elements 105 and 106 in the same size and adding a capacitor to at least one of the switch elements.

  In the comparator 102x, an analog signal obtained from the pixel 101x is input to VSL and a reference voltage RAMP is input to VREF. Before the first reading is stabilized and the comparison is performed by the comparator 102x, the “L (Low)” level is input to the XRESET signal. By this operation, the operating point of the first-stage differential amplifier 109 used in the comparator 102x is determined, and the difference between the operating point of the first-stage differential amplifier 109 and the respective input voltages is stored in the capacitive element 107 and the capacitive element 108. Sampled (held). This operation has an effect of removing the offset of the differential amplifier 109 of the comparator arranged in parallel with the column and reducing the fixed pattern noise.

  Next, the operation of the solid-state imaging device of the present invention will be described with reference to the operation timing chart of FIG. 3 and the block diagram of FIG.

  As shown in FIGS. 3 and 2, first, a reset signal is input to XRESET of the comparator 102x to reset the comparator 102x so that the potentials on both sides (input / output sides) of the comparator 102x are the same. From there, VREF slopes, performs A / D conversion of the reset level, and then performs A / D conversion of the signal. That is, after the first reading from the unit pixel 12 of any row line Hx (H0, H1,...) To the column lines V0, V1... Is stabilized, the comparator 102x (102a, 102b,. The obtained analog signal from the column line Vx (V0, V1,...) And the reference signal (reference voltage RAMP) generated by the reference signal generation unit 104 are input, and the two signals are compared. When the comparison is performed, the reference voltage RAMP has a stepped waveform that changes with time. At the same time as the reference voltage RAMP changes with time, the up / down counter 103x (102a, 102b,...) Counts down.

  When the reference voltage RAMP, which is a stepped waveform that changes with time, and the analog signal Vx obtained from an arbitrary pixel become equal, the output of the comparator 102x is inverted, and the down-counting of the up / down counter 103x is stopped, A count corresponding to the comparison period is held. At the time of the first reading, the reset component ΔV of the unit pixel 12 is read, and in the example of the present invention, a count period of 7 bits is provided to compare ΔV.

  In the second reading, in addition to ΔV, a signal component corresponding to the amount of incident light for each unit pixel 12 is read, and the same operation as the first reading is performed. That is, after the second reading from the unit pixel 12 of any row Hx to the column line Vx is stabilized, the analog signal Vx obtained from the pixel and the reference signal generated by the reference signal generation unit 104 are sent to the comparator 102x. The voltage RAMP is input and the above two signals are compared.

  When the comparison is performed, the reference voltage RAMP has a stepped waveform that changes with time. At the same time as the reference voltage RAMP changes with time, the up / down counter 103x counts up. That is, the A / D conversion period at the signal line level is up-counted. When the reference voltage RAMP, which is a stepped waveform that changes with time, and the analog signal Vx obtained from an arbitrary pixel become equal, the output of the comparator 102x is inverted and the up-counting of the up / down counter 103x is stopped. At the same time, the count value of the difference between the count values corresponding to the second comparison period is obtained from the count value corresponding to the first comparison period. At this point, correlated double sampling (CDS) has already been performed. Then, A / D conversion excluding the reset signal is performed. The data remaining in the counter 103x is only the counter data of the signal line.

  After the end of the A / D conversion period, the column scanning circuit 15 outputs the difference count value to the outside as a pixel signal digital output value, and then repeats the same operation for each row in order to obtain a two-dimensional image. .

  Next, the operation of the comparator 102x in the above-described AD conversion operation will be described with reference to the operation timing chart inside the comparator of FIG.

  As shown in FIG. 4, the operation of the comparator 102 x represents VREFin and VSLin during the period when XRESET is “L (Low)” level. In the comparator 102x, VSL (analog signals V0, V1,... From the pixels) and VREF (reference signal RAMP) are input to the input unit. First, the comparator 102x is reset once. The inputs VSL and VREF are connected by the reset signal XRESET, and VSLin and VREFin are determined at the operating point inside the comparator 102x (this is determined during reset). From this state, the “H (High)” level is input to the XRESET signal, the reference voltage VREF is temporally changed, and the comparison is performed. If the first-stage differential amplifier 109 has symmetry, The input unit remains at the same potential even after reset is released. As a result, there is no point where VREFin and VSLin are at the same potential in a period in which the reference voltage VREF is changed with time and the comparison is performed.

  Therefore, as in the prior art, after performing the comparison, after resetting the comparator, the reference signal generation unit 104 gives an offset to the reference voltage, and VREFin to which the signal component of the reference voltage VREF is transmitted has a potential higher than VSLin. Thus, a mechanism in which a period in which VREFin and VSLin are always at the same potential occurs in the comparison period, and the reference voltage VREF can be compared in a period in which the reference voltage VREF is linear is provided. Therefore, if a function is provided in the reference signal generation unit, the area becomes large. Further, by giving an offset to the reference voltage VREF after the reset signal XRESET becomes “H” level, the reference voltage VREF is simultaneously input to a large number of column parallel comparators. This is necessary and increases the A / D conversion time. That is, as shown in FIG. 5, if an offset component (dotted line in the drawing) is added to the reference voltage RAMP in the region A, it takes time to charge up to the offset voltage value, so the comparison is made in the regions B and C where the RAMP is non-linear. Then, it is necessary to take a long analog-digital conversion period. Further, as shown in FIG. 6, when the input of the first-stage differential amplifier 109 has an offset and the reference voltage RAMP is linear, the analog-digital conversion characteristic is also linear. On the other hand, when there is no offset in the input of the first-stage differential amplifier 109, the analog-digital conversion characteristics are also nonlinear.

  In order to solve this problem, in the present invention, as described with reference to FIG. 1, two pairs of a switch element (reset switch) 105 and a switch element (reset switch) 106 connected to reset the comparator are provided. When the reset signal XRESET is at “H” level, the input of the differential amplifier 110 is forcibly given an offset. By giving an offset, VSLin and VREFin are crossed (inverted) when sloped from reset.

  That is, the size of the switch element 105 that is a reset switch connected to an element to which the reference voltage VREF is input, compared to the switch element 106 that is a reset switch connected to an element to which an analog signal obtained from a pixel is input. Is large, the switch element 105 on the VREFin side has a capacitance. With this configuration, when the reset signal XRESET becomes “H”, the input side of the differential amplifier 110 to which the reference voltage VREF is input is raised to a high potential due to the large overlap capacitance of the switch element 105. .

  That is, at the moment when the reset is turned off, the potential of VREFin is raised by being pulled by the capacitor attached to the switch element 105. Since the amount of potential raised here is determined by the capacitance attached to the switch element 105, the VREFin side is lifted when the reset is turned off. Then, even if the VREFin is not lifted, the parasitic capacitance attached to the switch element (reset switch) 105 is smaller than the capacitance seen from the VREFin attached to the column. Can be offset.

  In short, this makes it possible to easily give a difference between the two signals, and it is not necessary to add an offset function to the reference signal generation unit 104, and the area can be reduced. Further, the period for charging the offset component is short, the A / D conversion period is shortened, and the speed can be increased.

  Even if an offset is added, the reset level is A / D converted, the signal level is A / D converted, and the difference is obtained, so the offset disappears. Therefore, adding an offset in the comparator 102x is not a problem.

  Further, as shown in FIG. 7, the RAMP signal is flattened in time because the wiring resistances, input capacitances, etc. of the comparators 102a, 102b,. Further, as shown in FIG. 8, the RAMP signal is non-linear due to the parasitic resistance attached to the signal, the parasitic capacitance, and the filter characteristics of the comparator capacitance element (input capacitance) immediately after the reference voltage VREF changes with time. Become. In the prior art, when the RAMP signal is in a non-linear state and VSLin and VREFin are inverted, A / D conversion is performed in a non-linear state. On the other hand, in the solid-state imaging device P of the present invention, the signal can be offset inside the comparator 102x simply by changing the size of the switch elements (reset switches) 105 and 106 in the comparator 102x. However, it is possible to perform A / D conversion in the linear region of the RAMP signal without using the nonlinear portion of the RAMP signal.

It is the circuit diagram which showed one Embodiment which concerns on the solid-state imaging device of this invention. It is the block diagram which showed one Embodiment which concerns on the solid-state imaging device of this invention. It is the operation | movement timing diagram which showed operation | movement of the solid-state imaging device of this invention. It is an operation | movement timing diagram inside a comparator. It is a RAMP waveform figure at the time of attaching offset to the reference voltage PAMP. It is an analog-digital conversion characteristic figure by the presence or absence of an offset. It is a block diagram explaining the parasitic capacitance of a comparator. It is a filter characteristic diagram after ty section where RAMP due to parasitic capacitance changes with time. It is the operation | movement timing diagram which showed the operation | movement of the analog-digital converter of the conventional solid-state imaging device. It is the circuit diagram which showed the circuit structure of the conventional comparator.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Solid-state imaging device, 12 ... Analog-digital converter, 101x ... Pixel, 102x ... Comparator

Claims (2)

In a solid-state imaging device having a plurality of pixels that convert an incident light amount into an electrical signal and having a plurality of analog-digital conversion devices that convert an analog signal obtained from the pixel into a digital signal,
The analog-to-digital converter is
A comparator that compares the reference signal with an analog signal obtained from the pixel;
The comparator generates an offset between a reference signal component and an analog signal component obtained from a pixel. The comparator has at least a differential amplifier;
A capacitive element connected in series between the input of the reference signal, the input of an analog signal obtained from a pixel, and the two inputs of the differential amplifier;
A switching element connected to each of two input elements provided in the differential amplifier,
The solid-state imaging device according to claim 1, wherein the two switch elements are configured asymmetrically.

Images