JP2008187744A - Solid-state imaging device, pixel-signal processing method, analog-signal transferring device, and analog-signal transferring method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve reduction in power consumption, easy layout, and efficient signal transmission, and stable black level control. <P>SOLUTION: A solid-state imaging device comprises an imaging pixel means composed of imaging pixels arranged in row and column directions, a vertical transfer means of outputting pixel signals from imaging elements in respective columns of a selected row to vertical signal lines provided to the respective columns, amplifier means provided for vertical signal lines of the respective columns and receiving pixel signals from imaging pixels in the respective columns, a plurality of horizontal signal lines, and a horizontal transfer means of transferring pixel signals held in the amplifier means in the respective columns into the plurality of horizontal signal lines. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a solid-state imaging device and a pixel signal processing method in the solid-state imaging device, and also relates to an analog signal transfer device and an analog signal transfer method.

Japanese Patent Laid-Open No. 5-207220 JP-A-11-266399

In recent years, CMOS image sensors have attracted attention as image sensors that can replace CCDs. This is because the CCD requires a dedicated process for its manufacture, requires multiple power supply voltages for its operation, and requires a combination of multiple peripheral ICs to operate, making the system very complex. This is because the CMOS image sensor overcomes various problems such as.
The CMOS image sensor can be manufactured by using a manufacturing process similar to that of a general CMOS type integrated circuit produced all over the world, and can be driven by a single power source. Furthermore, since analog circuits and logic circuits using a CMOS process can be mixed in the same chip, it has a plurality of very significant advantages such as the number of peripheral ICs can be reduced.

The CCD output circuit is mainly one-channel output using an FD (Floating Diffusion) amplifier, whereas the CMOS image sensor has an FD amplifier for each pixel, and the output is one line in the pixel array. The column-parallel output type is the mainstream in which these are selected and read out in the column direction at the same time. This is because it is difficult to obtain a sufficient driving capability with an FD amplifier arranged in a pixel, and therefore it is necessary to lower the data rate, and parallel processing is advantageous.
Various signal output circuits of this parallel output type CMOS image sensor have been proposed. For example, there are a pixel output sampled by a switched capacitor and read out, a pixel read out mounted with an amplifier, and an AD converter and a DRAM provided for each column. The present invention mainly relates to a reading method in which an amplifier is provided for each column.

An example of a reading method in which an amplifier is provided for each column is described in Patent Document 1 described above. This will be briefly described with reference to FIGS.
FIG. 15 shows only one column (a circuit system of one vertical signal line VL) corresponding to a certain pixel 102.
The pixel 102 includes a photodiode PD, a reset transistor Trst, an amplifying transistor Tg, and a reading transistor Ts. In this case, the output of the pixel 102 is read using a charge integration amplifier including the common source amplifier 100 and the capacitive elements C1 and C2.
What should be noted here is that the feedback capacitor C2 is precharged using the switch Tr15 and the reference voltage Vref. Thus, a device that can eliminate the offset variation of the common source amplifier 100 can be obtained. Has been made.

FIG. 16 shows an operation timing chart of this circuit. In the T1 period within the horizontal blanking period, the pixel outputs a value in which the signal Vps overlaps the offset voltage Vo. The charge integrating amplifier is in a reset state when the switch Tr13 is turned on by the signal φRC. At this time, similarly, the switch Tr15 is turned on by the signal φRC and the switch Tr14 is turned off by the signal φTC, so that the capacitive element C2 is precharged to the reference voltage Vref.
Next, in the period T2, the switch Tr15 is turned off and the switch Tr14 is turned on by the signals φRC and φTC, so that the precharged reference voltage Vref appears at the output of the charge integrating amplifier. At this time, since the switch Tr13 is turned off simultaneously with the switch Tr15, the reset state is released.
Next, in the T3 period, only the offset voltage Vo is output from the pixel 102, and by integrating it, only the signal component is read out as the gain by using the ratio of the sizes of the capacitive elements C1 and C2 as the output of the charge integrating amplifier. It will be a thing.
The last read signal is sequentially output to the horizontal signal line HL in synchronization with the pulse supplied from the shift register 101.

  As described above, in the readout method in which the amplifier is provided for each column, it is possible to easily remove only the signal component by removing the offset voltage of the pixel, and the readout gain is set by the ratio of the capacitive elements C1 and C2. It can be set arbitrarily. Furthermore, variations in the source follower itself can be removed by precharging with the reference voltage Vref, and there are many excellent points.

Further, Patent Document 2 discloses a different type of method. This will be described with reference to FIGS.
In FIG. 17, arbitrary three columns (vertical signal lines VL1, VLn, VLN) are extracted and shown. In addition, a pixel corresponding to the selection line Vs is shown as the pixel 200 including the photodiode PD and the reading transistor Ts. In this case, when the selection line Vs is selected by the vertical scanning circuit 201, a certain row of pixels is selected, and a signal from each pixel is output to each column (each vertical signal line VL).
In this example, an amplifier 203 is provided for each column (each vertical signal line VL), which is similar to the example of the above-mentioned Patent Document 1, but it should be noted that the standby control signal φP is sent from the horizontal scanning circuit 202 to each amplifier 203. (ΦP1, φPn, φPN) is supplied. As a result, the amplifier 203 can transition between the standby state and the operating state for each column.
The output of each amplifier 203 is selected by the switch 204 and transferred to the output terminal 205 through the horizontal signal line HL. Each switch 204 is ON / OFF controlled by a signal ΦH (φH1, φHn, φHN) from the horizontal scanning circuit 202.

FIG. 18 shows an operation timing chart.
A certain amplifier 203 is selected by the pulse of the signal ΦHn, and the signal is read during that period. Unlike the example of the above-mentioned patent document 1, since the read operation is performed during the period selected by the signal φHn instead of the operation of reading all at once in the horizontal blanking period, the signal φHn is conversely performed. It is not necessary to operate at all during a period not selected by. That is, an unselected amplifier 203 can be in a standby state during that time.
However, in consideration of the fact that it takes a little time to return from standby, the standby control signal φPn rises slightly before signal φHn, but does not consume any current because it is in the standby state during other periods. . Therefore, for example, even if a sensor (imaging pixel array) having 1000 columns is used and the amplifier 203 is also prepared for 1000 columns, only one of them is actually flowing, or the overlap of the signal φPn is considered. Even in the example of FIG. This provides the advantage of very low power consumption.

However, the conventional techniques such as Patent Documents 1 and 2 have the following problems.
First, in the method of Patent Document 1, since an amplifier must be prepared for each column, there is a problem that power consumption inevitably increases. In particular, in this example, the horizontal signal line HL must be driven directly, and the drive time is limited to 1/1000 of the effective period of one horizontal period, for example, in the case of a sensor having 1000 columns. Therefore, it is necessary to operate at a very high speed. This causes an increase in the operating current per amplifier prepared for each column, resulting in an extra increase in power consumption.
The reference voltage Vref effectively determines the black level of the output signal. However, since the fixed voltage is used, the reference voltage Vref does not follow the fluctuation factors such as temperature change and power supply voltage. Therefore, in order to supply a stable black level, a clamp circuit is inevitably required in the subsequent stage, and there is a problem that the circuit scale increases.

  On the other hand, in the case of Patent Document 2, the power consumption is reduced. However, as described above, during the period of the signal φHn, the reading from the pixel signal to the horizontal transfer must be performed all at once. For this reason, even if compared with the example of patent document 1, each amplifier must operate | move at a very fast speed, and the consumption current per amplifier increases. This also increases the layout area of the amplifier, which is very disadvantageous and difficult to design when considering the constraint that the amplifier must be laid out for each vertical signal line from the imaging pixel array, that is, for each column. Is expected.

  In view of these problems, an object of the present invention is to provide a more suitable solid-state imaging device and pixel signal processing method. In other words, in a configuration in which a charge integration amplifier is provided for each column, a circuit that has low current consumption and does not increase the layout area and has a black level tracking capability is provided. The present invention also provides a preferred signal transfer technique as an analog signal transfer apparatus and an analog signal transfer method.

  The solid-state imaging device according to the present invention includes an imaging pixel unit in which imaging pixels are arranged in a row direction and a column direction, and pixel signals from the imaging pixels of each column in a selected row for each imaging column unit. A vertical transfer means for outputting to a vertical signal line provided to the vertical signal line; an amplifier means provided for the vertical signal line in each column; Horizontal signal lines, and horizontal transfer means for distributing and transferring the pixel signals held by the amplifier means in each column to the plurality of horizontal signal lines.
  Furthermore, a multiplexer means for forming an output pixel signal as serial data from pixel signals transferred by the plurality of horizontal signal lines by selecting the plurality of horizontal signal lines in order.
  Further, a shield wiring using the reference potential of the multiplexer means may be provided between the plurality of horizontal signal lines.
  The horizontal transfer means overlaps the outputs of the amplifier means passing through the plurality of horizontal signal lines while shifting by one clock time.
  The horizontal transfer means is configured so that each amplifier means can be individually switched between an on state and a standby state by an amplifier control signal supplied to each of the amplifier means. Each amplifier means is individually connected to one of the plurality of horizontal signal lines by an amplifier selection signal that connects each of the amplifier means to a specific horizontal signal line of the plurality of horizontal signal lines. Each of the amplifier means arranged in each column is sequentially turned on for a predetermined period and connected to the horizontal signal line so that the pixel signal output from each amplifier means is Sequentially transferred by a plurality of horizontal signal lines.
  The horizontal transfer means outputs the amplifier control signal and the amplifier selection signal as a common signal. Alternatively, the amplifier control signal may be connected to the horizontal signal line by the amplifier selection signal after being controlled to be turned on by the amplifier control signal for each of the amplifier means without being common. And the amplifier selection signal is generated.
  The horizontal transfer means has a shift register, and shifts a pulse signal having a pulse width set in accordance with the number of horizontal signal lines and the data rate of the pixel signal output from each amplifier means. The amplifier selection signal is generated by inputting to the register.
  The horizontal transfer means takes the logical sum of the amplifier control signal for each of the amplifier means with a pulse as an amplifier selection signal for the corresponding amplifier means and a pulse whose edge timing is advanced from the pulse. To generate. For example, an nth amplifier selection signal for the nth amplifier means and an (n-1) th amplifier selection signal for the (n-1) th amplifier means are ORed to obtain the nth amplifier. The amplifier control signal for the means is generated.
  When the amplifier means is a charge integrating amplifier, the reference potential generating means for generating a reference potential based on the black level of the pixel signal output from the horizontal signal line, and the reading of the pixel signal by the vertical transfer means. And charging means for charging the feedback capacitance of the charge integrating amplifier means to the reference potential generated by the reference potential generating means.
  Alternatively, it further comprises reference potential generating means for generating a reference potential based on the black level of the pixel signal output from the horizontal signal line, so that the reference potential becomes the reference potential of the output amplifier of the multiplexer means. To do.
  In addition, a holding unit is further provided for holding the pixel signal input in the horizontal blanking period, for example, in the amplifier unit even when the amplifier unit is in a standby state.
  When the amplifier means is a charge integration amplifier, the charge integration amplifier means may have a feedback capacitor formed of a variable capacitance element.

  In the pixel signal processing method of the present invention, pixel signals from the imaging pixels of each column in the row selected by the imaging pixel means in which the imaging pixels are arranged in the row direction and the column direction are provided for each column. The pixel signal vertical transfer step for outputting to the vertical signal line, the signal holding step for holding the pixel signal in the amplifier means provided for the vertical signal line in each column, and the pixel signal held in the amplifier means for each column And a horizontal transfer step for distributing and transferring to a plurality of horizontal signal lines.

In the solid-state imaging device and the pixel signal processing method of the present invention described above, the imaging pixel signal of each column in the selected row is obtained from the imaging pixel means (pixel array) in which the imaging pixels are arranged in the row direction and the column direction. This is a method of reading in parallel during the horizontal blanking period. The readout pixel signal of each column is output to the horizontal signal line during the effective period in the horizontal period via the charge integration amplifier means arranged in each column (each vertical signal line), and is horizontally transferred. .
The present invention mainly has the following three features in such column-parallel pixel signal readout.
First, the charge integration amplifier means provided for each column is configured to be in a standby state while holding the pixel signal read during the horizontal blanking period.
Second, in the charge integration amplifier means provided for each column, the feedback capacitance of the amplifier is precharged to a certain reference potential during the read operation, and the reference potential is automatically controlled based on the black level.
Third, the horizontal transfer of pixel signals from each amplifier means is performed in parallel using a plurality of horizontal signal lines, thereby reducing the data rate per horizontal signal line.

The analog signal transfer apparatus according to the present invention corresponds to an analog signal holding means in which a signal holding section for holding an analog signal in the row direction is arranged as one row or a plurality of rows, and each column of the analog signal holding means. And a plurality of horizontal signal lines, and horizontal transfer means for distributing and transferring the outputs of the amplifier means to the plurality of horizontal signal lines. The horizontal transfer means transfers the output of the amplifier means that pass through the plurality of horizontal signal lines so as to overlap while shifting by one clock time.
In the analog signal transfer method of the present invention, the analog signal from the analog signal holding means in which the signal holding unit for holding the analog signal in the row direction is arranged as one row or a plurality of rows is provided for each column. A vertical transfer step for outputting to the vertical signal line, a signal holding step for holding an analog signal in the amplifier means provided in the vertical signal line in each column, and an analog signal held in the amplifier means in each column, A horizontal transfer step of distributing and transferring the analog signals held by the amplifier means to the plurality of horizontal signal lines so that the signals are transferred at overlapping timings while being shifted by one clock time on the plurality of horizontal signal lines. Prepare.
According to such an analog signal transfer device and analog signal transfer method, the horizontal transfer of the analog signal from each amplifier means in the horizontal direction is performed in parallel using a plurality of horizontal signal lines. The data rate per hit can be lowered.

According to the solid-state imaging device and the pixel signal processing method of the present invention, the following effects can be obtained.
First, each charge integrating amplifier means can take a standby state by holding the pixel signal read during the horizontal blanking period by the function of the holding means. As a result, power consumption can be greatly reduced. This is because each charge integrating amplifier means starts only when the timing to actually transfer data in the horizontal direction comes after holding the pixel signal read from the imaging pixel means during the horizontal blanking period. This is because the time may be in a standby state.
In particular, if the horizontal transfer means sequentially turns on each of the charge integration amplifier means arranged in each column for a predetermined period and sequentially outputs the pixel signal to the horizontal signal line, the charge integration amplifier means is necessary. An operation that is turned on only at the time is realized, which is effective in reducing power consumption.

More specifically, a method in which pixel signals for one row are read in parallel in each column in the horizontal blanking period, and then the pixel signals in each column are sequentially transferred horizontally in the effective period (from pixel signal reading to horizontal direction). In a system that does not perform all the data transfer at once, each charge integrating amplifier means needs to hold data, and therefore cannot be set in a standby state. The charge integrating amplifier means can be turned on only when necessary, and can be set in a standby state thereafter. In this case, each charge integration amplifier means does not perform a process from the reading of the pixel signal to the horizontal transfer all at once, so that a high operation speed is not required, and this is also effective in reducing power consumption. This also makes it possible to reduce the layout area of the charge integrating amplifier means, which is convenient for design.
Further, the hold means can be realized with a very simple configuration by adopting a circuit configuration in which the output of the charge integrating amplifier means is fixed to a predetermined hold voltage.

  In the present invention, the charge integration amplifier means provided for each column precharges the feedback capacitance of the amplifier to the reference potential generated based on the black level during the read operation. As a result, it is not affected by variations in the threshold voltage of the charge integrating amplifier means, and the black level of the pixel signal can be stably maintained following various fluctuation factors.

According to the present invention, the pixel signal from each amplifier means (for example, charge integrating amplifier means) is transferred in the horizontal direction in parallel using a plurality of horizontal signal lines, so that the data rate per horizontal signal line is obtained. Lower. As a result, the horizontal signal lines can be directly driven even by the amplifier means having a low driving capability laid out for each column. In other words, it is possible to eliminate the need to prepare an amplifier with high driving capability and increase the layout area.
In addition, pixel signals transferred in parallel by a plurality of horizontal signal lines can be efficiently converted into serial data by providing multiplexer means.

  Further, providing a shield wiring using the reference potential of the multiplexer means between the plurality of horizontal signal lines has the advantage that the problem of coupling between the plurality of horizontal signal lines is eliminated. Also, when driving a long horizontal signal line with an amplifier with a low output impedance provided for each column, a noise jump from the parasitic capacitance, which is a problem, should be provided with a shield line and connected to the reference voltage of the multiplexer. Thus, the capacitive element for sampling the potential of the horizontal signal line and the parasitic capacitance between the horizontal signal line and the shield line exist on the same reference potential, and jump from the reference potential through the parasitic capacitance. There is an advantage that it is not necessary to worry about such noise at all.

  In addition, by making the reference potential generated based on the black level of the pixel signal output from the horizontal signal line as the reference potential of the output amplifier of the multiplexer means, it is possible to control the black level at the multiplexer portion. Become. For this reason, there is no time or number of restrictions on the black level control that could normally be performed only once in one horizontal period, and stable black level control with high followability is possible.

Further, in the present invention, the horizontal transfer means outputs an amplifier control signal for controlling the amplifier means to stand-by / on and an amplifier selection signal for connecting the amplifier means to the horizontal signal line as a common signal. When the time required for the amplifier means to return from the standby state to the on state is sufficiently short, the circuit structure of the horizontal transfer means is simplified by combining the amplifier control signal and the amplifier selection signal. Since the number of wirings in the column direction is reduced, the layout for each column can be simplified.
The horizontal transfer means inputs a pulse signal having a pulse width set according to the number of the horizontal signal lines and the data rate of the pixel signal output from each amplifier means to the shift register. By generating the amplifier selection signal, the horizontal transfer means can be configured with a very simple configuration.
Further, when the amplifier control signal and the amplifier selection signal are not a common signal, the horizontal transfer means, the amplifier control signal for each of the amplifier means, a pulse as an amplifier selection signal for the corresponding amplifier means, By generating a logical sum with a pulse whose edge timing is advanced from the pulse, it can be realized with a simple configuration.

  Further, by forming the feedback capacitance of the charge integrating amplifier means with a variable capacitance element, a variable gain amplifier can be obtained. This means that the charge integration amplifier means can have a PGA (Programmable Gain Amp) function that is absolutely necessary in a normal camera system, which can simplify the system at the later stage and also at a pixel stage at an early stage. Since the gain of the signal can be increased, there is an advantage that a sufficient signal amplitude can be secured even when the signal is small, and noise resistance is improved.

  Further, in the analog signal transfer device and the analog signal transfer method of the present invention, the horizontal transfer of the analog signal from each amplifier means is performed in parallel using a plurality of horizontal signal lines. This is suitable for high-rate signal transmission, and it is possible to directly drive the horizontal signal line even with amplifier means having a low driving capability laid out for each column.

  Hereinafter, a solid-state imaging device and a pixel signal processing method according to the present invention will be described with reference to first to sixth embodiments.

<First Embodiment>
FIG. 1 shows the configuration of the first embodiment of the present invention. This shows a circuit system that outputs an imaging pixel signal obtained by the pixel array 1 as a captured image signal from the output unit 10 in the solid-state imaging device. Illustration and description of a lens system that forms an incident optical path for the pixel array 1 and a signal processing system subsequent to the output terminal 10 are omitted.

The pixel array 1 is formed by arranging a large number of imaging pixels 2 in the row direction and the column direction. The imaging pixel 2 is configured by connecting a photodiode PD, a transfer transistor Ttr, a reset transistor Trst, an amplification transistor Tg, and a selection transistor Ts as illustrated.
The selection transistors Ts of the imaging pixels 2 arranged in the column direction in the pixel array 1 are each connected to a common vertical signal line VL (VL1, VL2,... VLn).
Further, a set of the reset control line Lrst, the transfer control line Ltr, and the selection control line Ls corresponds to each row of the pixel array 1 so that common readout control is performed for the plurality of imaging pixels 2 arranged in the row direction. Arranged. A control pulse is given to the reset control line Lrst, the transfer control line Ltr, and the selection control line Ls by the vertical scanning circuit 3 at a predetermined timing.
The gate of the reset transistor Trst of each imaging pixel 2 is connected to the reset control line Lrst, the gate of the transfer transistor Ttr of each imaging pixel 2 is connected to the transfer control line Ltr, and the gate of the selection transistor Ts of each imaging pixel 2 is selected. The configuration is connected to the control line Ls.
The vertical scanning circuit 3 sequentially selects each row in the pixel array 1 and causes each vertical signal line VL to read out the pixel signal. Although described in detail later, the vertical scanning circuit 3 executes an operation of applying a reset level to the vertical signal line VL by applying a pulse to the reset control line Lrst and the selection control line Ls, and also applies a pulse to the transfer control line Ltr and the selection control line Ls. An operation of applying a pixel signal corresponding to the electric charge accumulated in the photodiode PD by application to the vertical signal line VL is executed. The vertical scanning circuit 3 sequentially performs such a reading operation on the imaging pixels 2 in each row.
Reading of pixel signals from a selected row of imaging pixels 2 is performed in a horizontal blanking period within one horizontal period. That is, in the horizontal blanking period, pixel signals from each imaging pixel 2 in one row selected by the vertical scanning circuit 3 are output in parallel to the vertical signal lines VL1, VL2,. become.

A charge integrating amplifier 9 is provided for each column, that is, for each vertical signal line VL1, VL2,.
The charge integrating amplifier 9 includes capacitive elements Ccp and Cs, switch elements Ssh, Srt, Spc, Spx, Shd, and an amplifier A.
Capacitance elements Ccp (Ccp1 to Ccpn) are arranged in series in the input system of each amplifier A (A1 to An) from each vertical signal line VL (VL1 to VLn).
The capacitive element Cs (Cs1 to Csn) is a feedback capacitance of each amplifier A (A1 to An).
The amplifier A (A1 to An) is an inverting amplifier.
In this case, the amplifier A is a capacitive feedback amplifier, and a signal is output with a gain corresponding to the capacitance ratio of the capacitive elements Ccp and Cs. The pixel signal charge read from the imaging pixel 2 is accumulated in the capacitive element Cs, so that the charge integrating amplifier 9 has a function of holding the pixel signal charge.

An amplifier drive circuit 4 for driving the charge integration amplifier 9 is provided. The amplifier drive circuit 4 supplies signals φSH, φRT, φPC, and φXPC to the charge integration amplifiers 9 at predetermined timings.
The switch elements Ssh (Ssh1 to Sshn) in the charge integrating amplifier 9 turn on / off the input system of the amplifier A from the vertical signal line VL according to the signal φSH.
The switch elements Srt (Srt1 to Srtn) are switches that short-circuit the input and output of the amplifier A, and are turned on / off according to the signal φRT.
Switch elements Spc (Spc1 to Spcn) are switches that connect a charge system to the capacitive element Cs, and are turned on / off in response to a signal φPC.
The switch elements Spx (Spx1 to Spxn) are turned off during the charge system for the capacitive element Cs, and are turned on / off by a signal φXPC which is an inversion pulse of the signal φPC.
The switch elements Shd (Shd1 to Shdn) are switches that are turned on when the output of the amplifier A is fixed to the hold voltage Vhd.

  In FIG. 1, n columns are shown as the charge integration amplifier 9, and the symbols (A, Ccp, Ssh, etc.) of each element constituting the charge integration amplifier 9 are “1” to “n” depending on the column. Although not shown, a required number of columns (vertical signal lines VL) are formed after the (n + 1) column, which is not shown in the figure, and a charge integrating amplifier 9 is provided for each column.

The output terminals of the charge integrating amplifiers 9 (the output terminals of the amplifiers A1 to An...) Are connected to the horizontal signal line HL via the switch elements Ssl (Ssl1 to Ssln, Ssl (n + 1). Connected.
As shown in the drawing, n horizontal signal lines HL are provided as horizontal signal lines HL1 to HLn.
The amplifier A1 in the first column is connected to the horizontal signal line HL1 via the switch element Ssl1. The amplifier A2 in the second column is connected to the horizontal signal line HL2 via the switch element Ssl2. Similarly, the amplifier An in the nth column is connected to the horizontal signal line HLn via the switch element Ssln.
Further, the amplifier An + 1 (not shown) in the (n + 1) th column is connected to the same horizontal signal line HL1 as that in the first column and the amplifier A1 via the switch element Ssl (n + 1). Each amplifier (not shown) in the n + 2th column and thereafter is also connected to the horizontal signal lines HL2, HL3,.
That is, assuming that the number of columns configured in the pixel array 1 is M columns, the charge integration amplifiers 9 of the M columns are sequentially distributed to n horizontal signal lines HL1 to HLn and connected. It becomes.

The horizontal scanning circuit 5 sequentially selects the charge integration amplifiers 9 and performs control to transfer pixel signal outputs from the charge integration amplifiers 9 through the connected horizontal signal lines HL. Therefore, the horizontal scanning circuit 5 outputs an amplifier selection signal φS (φS1 to φSn...) To the switch elements Ssl (Ssl1 to Ssln, Ssl (n + 1)...). The charge integration amplifier 9 corresponding to the switch element Ssl turned on by the amplifier selection signal φS outputs a pixel signal to a predetermined horizontal signal line HL.
The horizontal scanning circuit 5 outputs an amplifier control signal φP (φP1 to φPn...) For controlling the standby state / ON state to each of the amplifiers A1 to An. As a result, each amplifier A can be individually controlled to the on state / standby state. That is, if the pulse as the amplifier control signal φP is in the H state, the amplifier A operates normally, and if it is in the L state, it enters the standby state and does not consume current.

The horizontal scanning circuit 5 performs horizontal transfer control by the amplifier selection signal φS and the amplifier control signal φP. Specifically, for example, when the pixel signal from the charge integration amplifier 9 in the first column is horizontally transferred, the horizontal scanning circuit 5 first returns the amplifier A1 from the standby state to the on state by the amplifier control signal φP1. Subsequently, the switch element Ssl1 is turned on by the amplifier selection signal φs1.
The amplifier control signal φP is also supplied to the switch element Shd via the inverter IV, whereby the switch element Shd is on / off controlled. Therefore, the switch element Shd is turned on when the amplifier A is on standby, and the switch element Shd is turned off when the amplifier A is turned on.

  In this example, since a plurality of horizontal signal lines HL are provided, the output of the charge integrating amplifier 9 is transferred in parallel. Therefore, a multiplexer 8 is provided on the output side so that outputs can be combined into one. That is, the pixel signals of each column transferred by the horizontal signal lines HL1 to HLn are sequentially selected by the multiplexer, and are converted into pixel signals as serial data as in the case of one horizontal signal line HL. The signal is output from the output terminal 10 to the subsequent circuit.

  Further, a black level control circuit 6 that detects the level of the output pixel signal and sets it to an appropriate black level is provided, and is connected to one end of the switch Spc in each charge integrating amplifier 9 via the clamp voltage output amplifier 7. It is configured to provide feedback.

The operation of the first embodiment having such a configuration will be described with reference to the timing chart of FIG.
The readout of the pixel signal is basically performed within the horizontal blanking period. In the horizontal blanking period, an operation as a reset period and an operation as a reading period are performed as shown in the figure.
First, when the reset period is started, all the amplifier control signals φP (φP1, φP2,...) Are set to the H state by the horizontal scanning circuit 5, and all the amplifiers A in each column are returned from the standby state to the on state. .
For the pixel array 1, a row of imaging pixels 2 is selected by the vertical scanning circuit 3. That is, the selection transistor Ts is turned on for a certain column. Next, the reset transistor Trst is turned on, and the reset level Voff appears on the vertical signal line VL through the amplification transistor Tg.
Since the reset level Voff includes variations in reset noise of the imaging pixel 2 and threshold voltage of the amplification transistor Tg, the reset level Voff takes a different value for each imaging pixel 2.

At this time, on the charge integrating amplifier 9 side, the switch element Srt is turned on by the signal φRT, so that the input / output of the amplifier A is short-circuited and feedback is applied, and the threshold voltage Vat of the amplifier A appears in the output Vout.
At the same time, the switch element Spc is turned on by the signal φPC and the switch element Spx is turned off by the signal φXPC, so that one end of the capacitive element Cs is fixed to the clamp voltage Vcp by the clamp voltage output amplifier 7.
At this time, the capacitive element Cs holds the potential of the clamp voltage Vcp with reference to the threshold voltage Vat.
Also, the capacitive element Ccp holds the reset voltage Voff of the imaging pixel 2 with the threshold voltage Vat as a reference.

Next, when the signals φRT and φPC are lowered, the switch elements Srt and Spc are sequentially turned off, and then the signal φXPC is raised and the switch element Spx is turned on.
Since the amplifier A is fed back via the capacitive element Cs when the switch element Spx is turned on, the input of the amplifier A holds the same threshold voltage Vat as before. Accordingly, the clamp potential Vcp appears at the output Vout of the amplifier A due to the charge held in the capacitive element Cs.

Subsequently, a reading period is started.
First, the transfer transistor Ttr in the imaging pixel 2 in the row selected by the vertical scanning circuit 3 is turned on, and charges are transmitted from the photodiode PD, and the signal level Vsig is output from the imaging pixel 2 to the vertical signal line VL.
Since this signal appears overlapping with the previous reset level Voff, it can be expressed as Voff-Vsig.
Since the charge integrating amplifier 9 stores the value of the reset level Voff in the capacitive element Ccp, only the amount of change from the reset level Voff is integrated and output. At this time, the read gain is determined by the capacitance ratio. Expressed by the equation, the output voltage of the charge integrating amplifier 9 is Vout,
Vout = (Ccp / Cs) Vsig + Vcp (Formula 1)
It becomes. As can be seen from Equation 1, the output Vout does not depend on the reset level Voff of the imaging pixel 2 or the threshold voltage Vat of the amplifier A.

  The reading operation so far is in the order of reading the reset level Voff from the imaging pixel 2 and then reading the signal level Vsig, but conversely, the reset level Voff may be read after reading the signal level Vsig. Good.

Next, the horizontal blanking period shifts to the effective period. At this time, all the amplifier control signals φP (φP1, φP2,...) Are once in the L state. For this reason, the amplifiers A in each column are in a standby state and no current flows.
By the way, at this time, the output of the amplifier A becomes indefinite if no measures are taken. The signal remains stored in the capacitive element Cs, but if the output becomes indefinite, the signal stored in the capacitive element Cs is not always held as it is.
That is, if the switch element Ssh is formed of an NMOS transistor, the off state is created by setting the gate voltage to the GND level in the off state, but the output of the amplifier A is indefinite. In some cases, the output Vout of the amplifier A may drop to the GND level due to the influence of the leakage current or the like. As a result, the input side of the amplifier becomes a negative potential via the capacitive element Cs, so that the source voltage of the NMOS transistor becomes lower than the gate voltage, the switch element Ssh is turned on, and the signal charge flows out from the capacitive element Cs. Will end up. Of course, if the signal charge of the capacitive element Cs flows out in this way, a correct pixel signal cannot be applied to the horizontal signal line HL even if the amplifier A is subsequently turned on.
Therefore, in this example, in order to prevent such a state, the switch element Shd is controlled by using the amplifier control signal φP as an inversion pulse by the inverter IV. That is, when the amplifier A enters the standby state, the switching element Shd is turned on by this inversion pulse, and the output Vout of the amplifier A is fixed to the predetermined hold voltage Vhd.
In this way, even when the amplifier A is set in the standby state, the output Vout is prevented from becoming indefinite, and the loss of data stored in the capacitive element Cs is prevented.

At this time, the voltage Vin on the input side of the amplifier A is determined by the amount of charge stored in the capacitive element Cs.
Vin = Vhd − ((Ccp / Cs) Vsig + Vcp) −Vat) (Equation 2)
It is expressed.
The hold voltage Vhd is preferably set so that the input voltage Vin does not become a negative potential in consideration of the size of the capacitive elements Ccp and Cs, the maximum amount of the signal charge Vsig, and the like. In particular, when the off-leakage of the NMOS transistor of the process being used is large, the value of the input voltage Vin is not so low that the NMOS transistor is always back-biased, and consideration must be given to reducing the leak. come.
The above is an example when an NMOS transistor is used as the switch element Ssh. However, if a PMOS transistor is used, consideration must be given so that the source voltage does not become higher than the gate voltage. It becomes.

As described above, by the action of the switch element Vhd and the fixed hold potential Vhd, the signal read in the horizontal blanking period can be held in the capacitor element Cs as it is even when the amplifier A is in the standby state.
Therefore, after that, it is only necessary to restore the amplifier A from the standby state again when transferring in the horizontal direction, so the time during which the current is actually consumed becomes very short, and the power consumption is lower than in the conventional example. be able to.

For horizontal transfer, the amplifier A mounted for each column is driven in the horizontal direction.
When the signal of the imaging pixel 2 is read in the vertical direction, it can be read slowly using the horizontal blanking period, but it must be transferred in the horizontal direction at a determined data rate. Therefore, this time, high-speed operation is required, but the horizontal signal line HL is long and a large number of switches are connected, so that there is a large parasitic capacitance, and driving it at a high speed requires a certain amount of drive capability. . It is difficult to drive it with an amplifier A that must be laid out for each column, leading to an increase in layout area.
Therefore, in this example, as described above, a plurality of horizontal signal lines HL1 to HLn are prepared and transferred at a reduced data rate by parallel transfer. The number of horizontal signal lines HL (n) is not limited, but the more the number of horizontal signal lines HL, the lower the data rate per line. In general, it is efficient and preferable to prepare the number of driving lines that has the same driving speed as the reading operation in the horizontal blanking period.

  In the effective period of FIG. 2, after all the amplifier control signals φP are in the L state as described above, the horizontal scanning circuit 5 sequentially supplies the amplifier control signals φP to the H state for a predetermined period from the first column. Then, the amplifiers A in each column are sequentially activated. At the same time, the switch elements Ssl are sequentially turned on by the amplifier selection signal φS, and a signal is transferred through the horizontal signal line HL.

In other words, in the effective period, first, the amplifier control signal φP1 is set to the H state and the amplifier A1 is activated from the standby state for the horizontal transfer of the pixel signals in the first column. Immediately thereafter, the amplifier selection signal φS1 is set to the H state, and the output of the amplifier A1 is transferred by the horizontal signal line HL1.
Subsequently, the pixel control signal φP2 is set to the H state and the amplifier A2 is activated from the standby state for horizontal transfer of the pixel signals in the second column with a shift of the transfer timing for one pixel. Immediately thereafter, the amplifier selection signal φS2 is set to the H state, and the output of the amplifier A2 is transferred by the horizontal signal line HL2.
In this way, horizontal transfer is sequentially performed for each column.

Therefore, each of the amplifiers A1, A2,... Sequentially operates one column at a time shifted by exactly one pixel time, and the switch elements Ssl1, Ssl2, etc. are shifted one by one with respect to the horizontal signal line HL. Connected periodically.
The amplifier selection signal φS is a wide pulse whose width is determined by the number of horizontal signal lines HL, and shifts with a shift of one pixel time, so that it has an overlap period as shown in FIG. In the example of FIG. 2, the length is just written when there are five horizontal signal lines HL.
Since the pixel signal held in each amplifier A is transferred by the connected horizontal signal line while the amplifier selection signal φS is in the H state, for example, in FIGS. 7A, 7B, and 7C. As shown, the outputs of the amplifiers A are transferred so as to overlap each other with a predetermined horizontal signal line HL connected thereto while shifting by one clock time.
Since the signal is horizontally transferred while the amplifier selection signal φS is in the H state, the amplifier control signal φP may be in the H state only during that time to activate the amplifier. In this example, the amplifier A Assuming that it takes some time to start, the amplifier control signal φP is raised prior to the amplifier selection signal φS. That is, the amplifiers A1, A2,... Are connected to the horizontal signal line by the amplifier selection signal φS after being turned on by the amplifier control signal φP.

The horizontal scanning circuit 5 for generating the amplifier control signal φP and the amplifier selection signal φS can be realized by a simple circuit as shown in FIG.
The horizontal scanning circuit 5 shown in FIG. 3 includes a shift register including flip-flops FF (FF1, FF2 to FFn...) And OR gates OR1, OR2 to ORn. FIG. 4 shows the operation timing.
Each flip-flop FF is supplied with the clock Hclk shown in FIG. 4 at its clock input terminal, and latches the D input at the timing of this clock Hclk. This clock Hclk is a clock having the same rate as the data rate of the pixel signal.
The data pulse Hdata shown in FIG. 4 is supplied as the D input to the first flip-flop FF1. The data pulse Hdata is a pulse that is wide with respect to the clock rate as shown in FIG.
The Q outputs (latch outputs) of the flip-flops FF1, FF2, FF3... Become amplifier selection signals φS1, φS2, φS3.
The amplifier control signal φP (φP1, φP2, φP3...) Is obtained as an output of the OR gate OR (OR1, OR2, OR3...).
The OR gate OR1 takes the logical sum of the supplied data pulse Hdata and the Q output of the flip-flop FF1 as an amplifier control signal φP1. The OR gate OR2 takes the logical sum of the Q output of the flip-flop FF2 and the Q output of the flip-flop FF1 to obtain an amplifier control signal φP2. The OR gate OR3 takes the logical sum of the Q output of the flip-flop FF3 and the Q output of the flip-flop FF2 to obtain an amplifier control signal φP3.
That is, the amplifier control signal φPn for the nth amplifier An is generated by ORing the amplifier selection signal φSn for the nth amplifier An and the amplifier selection signal φSn− for the n−1th amplifier An−1. It is set as the structure.

  In this example, since horizontal transfer is performed using a plurality of horizontal signal lines HL, the amplifier selection signal φS needs to have a pulse width of the pixel signal data rate × the number of horizontal signal lines. By inputting a data pulse Hdata having a pulse width of the pixel signal data rate × the number of signal lines to a simple horizontal scanning circuit 5 using a simple shift register circuit as shown in FIG. , And a pulse that shifts in accordance with the data rate of the pixel signal, that is, amplifier selection signals φS1, φS2,... Can be easily formed.

  As described above, the amplifier control signal φP used to return the amplifier A from standby is required to have its rise before the amplifier selection signal φS as described above. Is used to generate an amplifier control signal φPn by taking the logical sum of the pulses of a certain amplifier selection signal φSn and the preceding amplifier selection signal φSn−1, thereby generating an amplifier control signal φPn. Thus, the amplifier control signal φPn can be easily formed as a pulse that rises and falls simultaneously with the amplifier selection signal φSn.

  That is, the horizontal scanning circuit 5 uses the pulses as the amplifier selection signal φS (x) for the corresponding amplifier A (x) as the amplifier control signals φP1, φP2,. In this case, it may be generated by taking a logical sum with a pulse (amplifier selection signal φS (y) for the amplifier A (y) selected earlier) whose edge timing is ahead of the pulse.

  3 and 4, the amplifier control signal .phi.Pn rises one clock earlier than the amplifier selection signal .phi.Sn. If it is desired to rise two clocks earlier, the amplifier selection signal .phi.Sn and the amplifier selection signal are selected. It is only necessary to calculate the logical sum of the signal φSn-2 to obtain the amplifier control signal φPn, and it is also possible to start up a plurality of clocks in the same manner. However, in the configuration of FIG. 3, the first amplifier control signal φP1 cannot be activated more than two clocks earlier than the amplifier selection signal φS1, and in this case, an extra flip-flop is provided in front of the flip-flop FF1 as a shift register configuration. It is sufficient to adopt a configuration that prepares and feeds in the air.

For example, with such a horizontal scanning circuit 5, the switch element Shd is turned off simultaneously with the rise of the amplifier control signal φP, and the output of the amplifier A again takes the value expressed by the above equation 1.
When the horizontal transfer ends and the amplifier selection signal φS falls, the amplifier control signal φP simultaneously falls to the standby state.

  When the amplifier driving method as described above is used, the amplifiers A provided for the number of columns are in the activated state for the number of horizontal signal lines HL, and the others are in the standby state, so that no current is consumed. Low power consumption. Further, since the data rate per horizontal signal line HL can be lowered, the amplifier A does not require such a high driving capability and can be sufficiently laid out for each column.

5A to 5D show specific circuit examples of the amplifier A provided for each column. Both form an inverting amplifier. FIG. 5A shows a common source grounding amplifier, and a transistor T10 is added for standby. Here, the signal xφP is an inverted pulse of the signal φP.
FIG. 5B is a differential amplifier type, and has an advantage that the threshold voltage Vat when feedback is applied can be controlled by Vref. A transistor T10 is added for standby.
FIG. 5C shows a regulated cascode circuit configuration in which an auxiliary amplifier a2 is provided to increase the gain of the common source amplifier a1. Regulated Cascode is detailed in IEEE Journal of Solid-State Circuit Vol.25, No1, February 1990. In this case also, a standby transistor T10 is added.
FIG. 5D shows a change in the arrangement of the standby transistor T10, which is resistant to variations because the transistors are not in series. Although the circuit of FIG. 5D is written as an application example of FIG. 5A, the same standby transistor connection method can be applied to the circuits of FIGS. 5B and 5C.

As described above, the horizontal transfer is performed in parallel by the plurality of horizontal signal lines HL, which are sequentially selected and output by the multiplexer 8.
A circuit example of the multiplexer 8 is shown in FIG.
The multiplexer 8 includes sampling circuits 51-1 to 51-n corresponding to the horizontal signal lines HL1 to HLn, and an output amplifier 50, respectively.
The sampling circuits 51-1 to 51-n are each composed of four switch elements Sm and capacitive elements Cm.

For example, the sampling circuit 51-1 includes switch elements Sm11, Sm21, Sm31, Sm41 and a capacitive element Cm1. Each switch element Sm11, Sm21, Sm31, Sm41 is turned on / off by signals φm11, φm21, φm31, φm41, respectively. The signals φm11, φm21, φm31, and φm41 are generated, for example, by a timing generator (not shown) in the multiplexer 8 as a pulse that becomes an H state at a predetermined timing with reference to the horizontal synchronization signal.
The switch element Sm11 is a switch that connects the horizontal signal line HL1 to the capacitive element Cm1. The switch element Sm21 is a switch that connects the reference of the capacitive element Cm1 to the reference voltage Vref1. The switch elements Sm31 and Sm41 are switches for transferring the charge of the capacitive element Cm1 to the output amplifier 50. That is, when the switch elements Sm31 and Sm41 are turned on, the capacitive element Cm1 is inserted into the feedback path of the output amplifier 50.
The other sampling circuits 51-2 to 51-n have the same configuration.
A reference voltage Vref2 is connected to the non-inverting input V + of the output amplifier 50.

FIG. 7 shows an operation timing chart of such a multiplexer 8.
As described with reference to FIG. 2 above, the amplifiers A in each column that drive the horizontal signal line HL operate at a timing shifted by one pixel time, so that the output is also sequentially output at a timing shifted by one pixel time. Has been.
As a result of the horizontal transfer operation described above, the pixel signals held in each amplifier A are transmitted on a plurality of horizontal signal lines HL as shown as signals D1, D2, and D3 in FIGS. 7A, 7B, and 7C, for example. The data is transferred so as to overlap while being shifted by one clock time (one pixel time).
With reference to FIGS. 7A and 7D, the horizontal signal line HL1 and the sampling circuit 51-1 will be described.
Initially, the signals φm11 and φm21 are set to the H state from the time t0, so that both the switch elements Sm11 and Sm21 are in the on state. Therefore, the horizontal signal line HL1 charges its output to the capacitive element Cm1 with reference to the reference voltage Vref1. Will be.
Next, the signal φm11 is in the L state at the time t1, the signal φm21 is in the L state at the time t2, and the switch elements Sm11 and Sm21 are sequentially turned off, so that the voltage of the horizontal signal line HL1 is applied to the capacitive element Cm1. It will be sampled.
Finally, at time t3, the signals φm31 and φm41 are simultaneously set to the H state, so that the switch elements Sm31 and Sm41 are simultaneously turned on, and the capacitive element Cm1 is connected to the output amplifier 50.
Since the capacitive element Cm1 is inserted between the output terminal Vout and the negative input terminal V− of the output amplifier 50, virtual grounding works, and the output amplifier 50 has a potential difference stored in the capacitive element Cm1 with reference to the reference voltage Vref2. Is the output Vout. At this time, if the output of the horizontal signal line HL1 is Vhl1, and the output voltage of the amplifier is Vout,
Vout = Vref2 + (Vhl1-Vref1) (Formula 3)
Will be output.

If reference voltage Vref1 = reference voltage Vref2 is assumed here, Equation 3 is
Vout = Vhl1 (Formula 4)
Thus, the output of the horizontal signal line HL1, that is, the output of the amplifier A expressed by the above-described expression 1 is output as it is. Unless there is a particular reason, it is desirable to use the reference voltage Vref1 = Vref2.

The above has focused on only the horizontal signal line HL1 and the sampling circuit 51-1, but as can be understood with reference to FIGS. 7B, 7C, 7E, 7F, and 5G, the horizontal signal line For HL2 to HLn, the same operation is repeated by the sampling circuits 51-2 to 51-n at the timing shifted by one pixel time, and as a result, data is continuously transferred to the output amplifier. It is like that.
Thereby, for example, as shown in FIGS. 7A, 7B, and 7C, signals D1, D2, and D3 appearing on the horizontal signal lines HL1, HL2, and HL3 are converted into serial data at the output Vout in FIG. 7H. The signals D1, D2, and D3 are output.

  By using such a circuit as the multiplexer 8, the data transferred in parallel using the plurality of horizontal signal lines HL1 to HLn can be efficiently converted into one serial data. In this example, only one output amplifier 50 is used. However, it is also easy to prepare a plurality of output amplifiers for a plurality of channels. In that case, it can be realized only by changing the connection destination and the operation timing of the switch elements Sm3 * and Sm4 *.

As described above, according to the first embodiment of the present invention, it is possible to efficiently read out a pixel signal using the charge integrating amplifier 9, and the black level control circuit 6 and the clamp voltage output amplifier. 7, the precharge voltage of the capacitive element Cs is generated, so that the black level can be controlled without being affected by the variation in the threshold voltage of the amplifier A.
Further, in the effective period, the charge integrating amplifier 9 can enter a standby state while holding the signal read in the horizontal blanking period. For this reason, each charge integrating amplifier 9 is activated only when the timing for transferring data in the horizontal direction is reached, and the power consumption can be greatly reduced by entering the standby state at other times.
Further, by performing horizontal transfer in parallel using a plurality of horizontal signal lines HL1 to HLn, the data rate per horizontal signal line HL can be lowered, and an amplifier having a low driving capability laid out for each column. Even in A, the horizontal signal line HL can be directly driven. For this reason, it is not necessary to increase the layout area by preparing an amplifier with high driving capability. Further, by providing the multiplexer 8, the data transferred in parallel by a plurality of horizontal signal lines HL1 to HLn can be efficiently converted into serial data.

<Second Embodiment>
FIG. 8 shows the configuration of the second embodiment. In the following embodiments, the same parts as those in FIG.
In the case of FIG. 8, the signal φS (φS1, φS2,...) Output from the horizontal scanning circuit 5 is an amplifier control signal and an amplifier selection signal, which is almost the same as that of the first embodiment shown in FIG. The difference is that it is shared.
That is, the signal φS is supplied from the horizontal scanning circuit 5 for standby control of the amplifier A and control of the switch element Shd.
In the case of the first embodiment, the amplifier control signal φP is raised slightly earlier than the amplifier selection signal φS as described in FIG. 2 in consideration of the recovery time of the amplifier A from standby. In the case where A does not take much time to return to standby, the signal φP can be shared by the signal φS as shown in FIG.
In this way, the horizontal scanning circuit 5 can be simplified, and the number of vertical wirings (rows in the column direction) can be reduced, thereby simplifying the layout.

When the amplifier control signal φP for returning the amplifier A from the standby state and the amplifier selection signal φS for connecting the amplifier A to the horizontal scanning line HL are shared as described above, the horizontal scanning circuit 5 can be realized by a circuit as shown in FIG.
The configuration of FIG. 9 is a configuration that is simpler than that of FIG. 3 by simply deleting the OR gate OR for forming the amplifier control signal φP from the configuration of FIG. 3 described above. The operation timing is shown in FIG. 10. The operation timing is the same as that in FIG. 4 except that the amplifier control signal φP is eliminated (the amplifier selection signal φS also serves as the amplifier control signal φP).

<Third Embodiment>
FIG. 11 shows a third embodiment. In this case, the feedback capacitance Csv (Csv1, Csv2,...) Of the amplifier A in each charge integrating amplifier 9 is constituted by a variable capacitance element.
As a result, the term (Ccp / Cs) in the above equation 1 becomes variable, so that a variable gain amplifier can be formed.
It can be easily assumed that the variable capacitor Cvs can be formed by a method in which a plurality of capacitors are arranged and selected by a switch. In this case, a pulse for selecting a capacitance value is generated from the amplifier drive circuit 4.
If such a third embodiment is used, the charge integrating amplifier 9 can be provided with the function of a PGA (Programmable Gain Amp) that is necessarily required in the camera system. Since the gain can be increased at an early stage, there is an advantage that a sufficient signal amplitude can be secured even when the pixel signal level is small, and noise resistance is improved.

<Fourth embodiment>
FIG. 12 shows a fourth embodiment.
In this case, the fixed voltage VrefPC is used as the precharge voltage for the feedback capacitive element Cs in the charge integrating amplifier 9.
The clamp voltage generated by the black level control circuit 6 and the clamp voltage output amplifier 7 is fed back to the reference voltage Vref2 (see FIG. 6) of the multiplexer 8.
Then, the reference voltage Vref2 of the above expression 3, that is, the reference of the output amplifier 50 in the multiplexer 8 is set to the clamp voltage corresponding to the black level, and thereby the black level is appropriately controlled.
In the case of the system as shown in the first embodiment, since the reference of the pixel signal to be read out is controlled according to the black level, the black level control can be performed only in the horizontal blanking period, that is, once in one horizontal period. However, if clamping is performed by the multiplexer 8 as shown in FIG. 12, there is no time limit on the clamping operation, and the black level can be controlled at any time.
Therefore, there is no time or frequency limitation on the black level control, and it is possible to form a clamp circuit with high followability and stability.

<Fifth embodiment>
FIG. 13 shows a fifth embodiment. As in FIG. 12, the clamp voltage generated by the black level control circuit 6 and the clamp voltage output amplifier 7 is fed back to the reference voltage Vref2 of the multiplexer 8. In the case of such a configuration, feedback is performed. Since the black level control by the precharge with respect to the capacitive element Cs becomes unnecessary, the switch elements Spc and Spx and the reference potential VrefPC are omitted from the configuration of FIG. In some cases, such a circuit configuration is also conceivable.
In the case of such a configuration, the same driving can be performed except that the signals φPC and φXPC are deleted from the timing chart of FIG.
According to this embodiment, there is an advantage that the circuit configuration is simplified, but the output Vout expressed by the above equation 1 is
Vout = (Ccp / Cs) Vsig + Vat (Formula 5)
Thus, the motor moves with reference to the threshold voltage Vat of the amplifier. Since this threshold voltage is expected to vary for each amplifier A in each column, care must be taken in designing.

<Sixth Embodiment>
FIG. 14 shows a sixth embodiment.
As described above, in each embodiment, since a plurality of horizontal signal lines HL are prepared, coupling between the horizontal signal lines HL may be a problem depending on the layout method. In addition, since the long horizontal signal line HL must be driven by the amplifier A having a low driving capability provided for each column, it is assumed that the impedance is high and noise is likely to be mixed through a parasitic capacitance or the like.
Therefore, in the sixth embodiment, the above problem is solved by arranging the shield line SiL between the plurality of horizontal signal lines HL1 to HLn and coupling it with the reference voltage Vref1 of the multiplexer 8. Yes.
That is, coupling is prevented by passing the shield lines SiL one by one between the horizontal signal lines HL. In addition, the horizontal signal line HL and the shield line SiL are coupled by a parasitic capacitance Cx, and the shield line is connected to the reference voltage Vref1, thereby sampling the potential of the horizontal signal line HL in FIG. It exists on the same basis as the capacitive element Cm. Then, it becomes possible to prevent the mixing of noise through the parasitic capacitance Cx.

<Modification>
While the embodiments of the present invention have been described above, various modifications can be considered as the present invention.
Especially in the present invention,
First, in the charge integration amplifier provided for each column, the pixel signal read during the horizontal blanking period is held and the standby state is set.
Second, in a charge integrating amplifier provided for each column, automatically controlling a reference potential to be precharged to a feedback capacitor of the amplifier during a read operation based on a black level;
Third, the pixel signals from the charge integrating amplifiers are transferred in the horizontal direction in parallel using a plurality of horizontal signal lines HL.
However, a configuration including at least one of these three elements is within the scope of the present invention.
For example, in the configuration of FIG. 1, if the configuration is such that one horizontal signal line HL is provided and the multiplexer 8 is not provided, the configuration having the first and second characteristics is obtained.
Further, in the configuration of FIG. 12 or FIG. 13, if the configuration is such that one horizontal signal line HL is provided and the multiplexer 8 is not provided, the configuration having the first feature is obtained.
Further, in the configuration of FIG. 1, if the precharge voltage is supplied at a fixed potential without providing the black level control circuit 6 and the clamp voltage output amplifier 7, the configuration has the first and third characteristics.
The present invention can be realized as various modifications as described above.

  Further, although the charge integration amplifier is used in the column reading for each vertical signal line VL, it is not limited to the charge integration amplifier.

Also, a method of transferring analog signals, such as those used in the present invention, in which a plurality of signal lines (horizontal signal lines) are prepared and their drive timings are overlapped while shifting by one clock for each channel, The present invention is not limited to the solid-state imaging device as in the above embodiment, and can be applied to other fields.
For example, when a plurality of analog data such as an analog memory is present in a row direction or an array of rows and columns and it is desired to read the data accurately and at high speed, the parallel transfer as introduced in the present invention is used. Thus, reading with high accuracy, high speed, and low power consumption is possible.
That is, in an analog memory device or the like, when an analog signal holding unit is formed by arranging a signal holding unit that holds an analog signal in the row direction as one row or a plurality of rows, an analog signal from the analog signal holding unit is formed. Are output to the vertical signal line provided for each column, and the signal value of each signal holding unit is read by the amplifier provided corresponding to each column.
Then, the analog signal held in each amplifier is distributed and transferred to a plurality of horizontal signal lines.

It is explanatory drawing of the structure of the 1st Embodiment of this invention. It is explanatory drawing of the pixel signal read-out operation | movement timing of embodiment. It is explanatory drawing of the circuit example of the horizontal scanning circuit of embodiment. It is explanatory drawing of the operation timing of the horizontal scanning circuit of embodiment. It is explanatory drawing of the circuit example of the charge integration amplifier of embodiment. It is explanatory drawing of the structure of the multiplexer of embodiment. It is explanatory drawing of the operation timing of the multiplexer of embodiment. It is explanatory drawing of the structure of 2nd Embodiment. It is explanatory drawing of the circuit example of the horizontal scanning circuit of 2nd Embodiment. It is explanatory drawing of the operation timing of the horizontal scanning circuit of 2nd Embodiment. It is explanatory drawing of a structure of 3rd Embodiment. It is explanatory drawing of a structure of 4th Embodiment. It is explanatory drawing of the structure of 5th Embodiment. It is explanatory drawing of the structure of 6th Embodiment. It is explanatory drawing of the structure of a prior art example. It is explanatory drawing of the operation timing of a prior art example. It is explanatory drawing of the structure of another prior art example. It is explanatory drawing of the operation timing of another prior art example.

Explanation of symbols

  1 pixel array, 2 imaging pixels, 3 vertical scanning circuit, 4 amplifier driving circuit, 5 horizontal scanning circuit, 6 black level control circuit, 7 clamp voltage output amplifier, 8 multiplexer, 9 charge integration amplifier

Claims (16)

Imaging pixel means comprising imaging pixels arranged in a row direction and a column direction;
For the imaging pixel means, vertical transfer means for outputting a pixel signal from the imaging pixels of each column in the selected row to a vertical signal line provided for each column;
Amplifier means provided for the vertical signal lines in each column, to which pixel signals from the imaging pixels in each column are respectively input;
Multiple horizontal signal lines;
Horizontal transfer means for distributing and transferring the pixel signals held in the amplifier means of each column to the plurality of horizontal signal lines;
A solid-state imaging device comprising: And a multiplexer means for forming an output pixel signal as serial data from the pixel signals transferred by the plurality of horizontal signal lines by selecting the plurality of horizontal signal lines in order. The solid-state imaging device according to claim 1 . 3. The solid-state imaging device according to claim 2 , wherein a shielding wiring using a reference potential of the multiplexer means is provided between the plurality of horizontal signal lines. 2. The solid-state imaging device according to claim 1 , wherein the horizontal transfer means transfers the outputs of the amplifier means that pass through the plurality of horizontal signal lines so as to overlap while shifting by one clock time. The horizontal transfer means is configured to be able to individually switch the amplifier means to an on state and a standby state by an amplifier control signal supplied to each of the amplifier means. Each of the amplifier means is individually connected to one of the plurality of horizontal signal lines by an amplifier selection signal that connects each of the means to a specific horizontal signal line of the plurality of horizontal signal lines. Configured to be able to
Each of the amplifier means arranged in each column is sequentially turned on for a predetermined period and connected to the horizontal signal line so that the pixel signals output from the amplifier means are sequentially transferred through a plurality of horizontal signal lines. The solid-state imaging device according to claim 1 , wherein: 6. The solid-state imaging device according to claim 5 , wherein the horizontal transfer means outputs the amplifier control signal and the amplifier selection signal as a common signal. The horizontal transfer means has a shift register,
By inputting a pulse signal having a pulse width set according to the number of the horizontal signal lines and the data rate of the pixel signal output from each amplifier means to the shift register, the amplifier selection signal is The solid-state imaging device according to claim 5 , wherein the solid-state imaging device is configured to be generated. Said horizontal transfer means, for each of said respective amplifier means, after controlling the on state by the amplifier control signal, according to claim 5, characterized in that connected to the horizontal signal line by the amplifier selection signal Solid-state imaging device. The horizontal transfer means obtains the amplifier control signal for each of the amplifier means by a logical sum of a pulse as an amplifier selection signal for the corresponding amplifier means and a pulse whose edge timing is advanced from the pulse. The solid-state imaging device according to claim 8 , wherein the solid-state imaging device is generated. Reference potential generating means for generating a reference potential based on the black level of the pixel signal output by the horizontal signal line,
3. The solid-state imaging device according to claim 2 , wherein the reference potential is a reference potential of an output amplifier of the multiplexer means. The amplifier means is a charge integrating amplifier,
Reference potential generating means for generating a reference potential based on the black level of the pixel signal output by the horizontal signal line;
Prior to reading out the pixel signal by the vertical transfer means, charging means for charging the feedback capacitance of the charge integrating amplifier means to the reference potential generated by the reference potential generating means;
The solid-state imaging device according to claim 1 , further comprising: Even in a period in which said amplifier means is in a standby state, the solid-state imaging according to claim 1 in which the input pixel signal, characterized in that the hold means to be held within said amplifier means, further comprising apparatus. The amplifier means is a charge integrating amplifier,
2. The solid-state imaging device according to claim 1 , wherein the charge integrating amplifier means has a feedback capacitance formed of a variable capacitance element. Pixel signal vertical that outputs pixel signals from the imaging pixels of each column in the selected row in the imaging pixel means in which the imaging pixels are arranged in the row direction and the column direction to the vertical signal line provided for each column A transfer step;
A signal holding step for holding a pixel signal in the amplifier means provided in the vertical signal line of each column;
A horizontal transfer step for distributing and transferring the pixel signals held in the amplifier means of each column to a plurality of horizontal signal lines;
A pixel signal processing method comprising: Analog signal holding means in which a signal holding unit for holding an analog signal in the row direction is arranged as one row or a plurality of rows;
Amplifier means provided corresponding to each column of the analog signal holding means, and reading the signal value of the corresponding signal holding section;
Multiple horizontal signal lines;
Horizontal transfer means for distributing and transferring the output of the amplifier means to the plurality of horizontal signal lines;
With
The analog signal transfer apparatus according to claim 1, wherein the horizontal transfer means transfers the outputs of the amplifier means that pass through the plurality of horizontal signal lines so as to overlap each other while shifting by one clock time. Vertical transfer for outputting an analog signal from an analog signal holding means, in which a signal holding unit for holding an analog signal in the row direction is arranged as one row or a plurality of rows, to a vertical signal line provided for each column Steps,
A signal holding step for holding an analog signal in the amplifier means provided in the vertical signal line of each column;
A plurality of analog signals held by the amplifier means are transferred so that the analog signals held by the amplifier means of each column are transferred at a timing overlapping on the plurality of horizontal signal lines while shifting by one clock time. A horizontal transfer step for distributing and transferring to the horizontal signal lines;
An analog signal transfer method comprising:

Images