JP2019186953A - Photoelectric conversion element, image reading device, and image forming apparatus - Google Patents

Abstract

To reduce a fixed pattern noise while ensuring a necessary response speed.SOLUTION: A photoelectric conversion element includes: a light receiving element that generates a charge according to the amount of received light; a buffer part that buffers a voltage signal according to the charge generated by the light receiving element and outputs the voltage signal; a current control circuit that controls a current flowing in the buffer part to be a predetermined current amount when the buffer part outputs the voltage signal; and an elimination circuit that eliminates a high frequency component of a predetermined bandwidth or higher from the voltage signal output from the buffer part.SELECTED DRAWING: Figure 8

Description

  The present invention relates to a photoelectric conversion element, an image reading apparatus, and an image forming apparatus.

  Conventionally, a CCD is used as a photoelectric conversion element used in a scanner, but a CMOS linear image sensor (CMOS sensor) is attracting attention due to a recent demand for higher speed. The CMOS sensor is the same as the CCD in that incident light is photoelectrically converted by a photodiode (PD). However, the CMOS sensor is different from the CCD in that it performs charge-voltage conversion in the vicinity of the pixel and outputs it to the subsequent stage. In addition, since a CMOS process is used, a CMOS sensor can incorporate a circuit such as an ADC (Analog Digital Converter) and is advantageous over a CCD in terms of high speed.

  In the CMOS linear image sensor, a source follower and a current load for supplying a bias current to the source follower are configured for each pixel, thereby realizing high-speed signal readout. However, there is a problem that noise is worsened by adding a current load, that is, increasing the current flowing to the source follower. In particular, since high frequency noise cannot be removed by CDS (Correlated-Double-Sampling), it causes FPN (fixed pattern noise) and causes vertical stripes on the image.

  For example, in Patent Document 1, an amplification transistor that amplifies a signal output from a photoelectric conversion unit uses only a capacitor as a load, and a write switch unit performs amplification after the capacitor is initialized. An amplifying solid-state imaging device that outputs a signal output from an amplifying transistor to a capacitor and writes to the capacitor during a period in which the transistor shifts from a saturation region operation to a sub-threshold region operation and becomes a metastable state is disclosed. .

  However, the amplification type solid-state imaging device disclosed in Patent Document 1 has a problem in that FPN deteriorates due to a limited signal response speed.

  The present invention has been made in view of the above, and an object of the present invention is to provide a photoelectric conversion element, an image reading apparatus, and an image forming apparatus that can reduce fixed pattern noise while ensuring a necessary response speed. And

  In order to solve the above-described problems and achieve the object, the present invention provides a light receiving element that generates an electric charge according to the amount of received light, and an electric charge that is provided for each of the light receiving elements, and that corresponds to the electric charge generated by the light receiving element. A buffer unit that buffers the voltage signal that has been subjected to voltage conversion and outputs the voltage signal to the subsequent stage, and when the buffer unit outputs the voltage signal, the current flowing through the buffer unit is controlled to have a predetermined amount of current. A current control circuit; and a removal circuit that removes a high-frequency component in a predetermined band or more in a range that does not affect the signal response of the buffer unit from the voltage signal output from the buffer unit.

  According to the present invention, it is possible to reduce the fixed pattern noise while ensuring a necessary response speed.

FIG. 1 is a diagram illustrating the overall configuration of the photoelectric conversion element. FIG. 2 is a diagram illustrating a configuration of a pixel, a pixel circuit, and a storage portion included in the photoelectric conversion element. FIG. 3 is a diagram illustrating timing at which a signal is written from the pixel circuit to the memory capacity. FIG. 4 is a diagram illustrating a noise spectrum in the photoelectric conversion element. FIG. 5 is a diagram illustrating a configuration example of a pixel unit for noise reduction. FIG. 6 is a diagram illustrating a state of a signal read from the pixel circuit to the storage unit. FIG. 7 is a diagram showing a noise spectrum in the photoelectric conversion element having the configuration shown in FIG. FIG. 8 is a diagram illustrating a configuration example of a pixel portion of the photoelectric conversion element according to the embodiment. FIG. 9 is a diagram illustrating a signal read from the pixel circuit to the storage unit via the band limiting unit. FIG. 10 is a diagram illustrating a noise spectrum of a photoelectric conversion element including the pixel portion illustrated in FIG. FIG. 11 is a diagram illustrating a configuration of a pixel portion in which circuit addition is suppressed with respect to the configuration illustrated in FIG. FIG. 12 is a diagram illustrating signals read from the pixel circuit illustrated in FIG. 11 to the storage unit. FIG. 13 is a diagram illustrating a configuration of a pixel portion in a photoelectric conversion element including a CDS portion. FIG. 14 is a diagram illustrating the reason why high frequency noise cannot be corrected by CDS. FIG. 15 is a diagram illustrating a configuration example of a pixel unit having a function corresponding to the band limiting unit. FIG. 16 is a diagram illustrating an operation of the photoelectric conversion element including the pixel portion illustrated in FIG. FIG. 17 is a diagram illustrating a noise spectrum of a photoelectric conversion element having a pixel portion including a storage portion. FIG. 18 is a diagram illustrating the configuration of the photoelectric conversion element according to the embodiment. FIG. 19 is a flowchart showing a method for adjusting the bandwidth limitation. FIG. 20 is a diagram illustrating a state before adjustment. FIG. 21 is a diagram illustrating a state after adjustment. FIG. 22 is a diagram illustrating an outline of an image forming apparatus including an image reading apparatus having, for example, a photoelectric conversion element.

  First, the background that led to the present invention will be described. FIG. 1 is a diagram illustrating the overall configuration of a CMOS linear image sensor (photoelectric conversion element) 10. Each of the PIX (R) 20, PIX (G) 22, and PIX (B) 24 has about 7000 PDs (photodiodes: light receiving elements), and is configured for each of RGB colors. The PDs are arranged in one direction for each color of light to be received, and generate charges according to the amount of received light. Each of PIX_BLK (R) 21, PIX_BLK (G) 23, and PIX_BLK (B) 25 has about 7000 pixel circuits (PIX_BLK), and is configured for each RGB color. That is, each PD is provided with a pixel circuit (PIX_BLK).

  Each pixel circuit (PIX_BLK) converts the charge accumulated in the PD into a voltage signal, and outputs a signal to the analog memory through the readout line. The PIX_BLK includes a transfer transistor that transfers the charge of PD to the floating diffusion (FD), a reset transistor that resets the FD, and a source follower transistor that buffers the FD voltage and outputs it to the read line. Unlike an area sensor, a linear sensor reads signals independently from each pixel of RGB, so that a readout line exists independently for each pixel.

  The AMEM 26 has, for example, about 7000 analog memories (Cs, which will be described later) for each of RGB colors, holds signals for each pixel, and sequentially outputs image signals in units of columns. When the AMEM 26 holds the signal, a global shutter system is realized in which the operation timing of PIX and PIX_BLK, that is, the exposure timing is the same for RGB.

  The ADC 27 has the same number of AD converters as the number of columns, and sequentially AD-converts image signals in units of columns. The ADC 27 has the same number of AD converters as the number of columns and performs parallel processing, thereby realizing high-speed operation as a photoelectric conversion element while suppressing the operation speed of the AD converter.

  The signal AD-converted by the ADC 27 is held for each pixel by the parallel-serial converter (P / S) 28, and the held signal is sequentially output to the LVDS 29. The photoelectric conversion element 10 processes parallel data processed in parallel for each pixel in the main scanning direction on the upstream side of the P / S 28, but processes serial data for each RGB color on the downstream side of the P / S 28. . The signal output from the P / S 28 is converted into a low-voltage differential serial signal by the LVDS 29 and output to the subsequent stage. The timing control unit (TG) 30 controls each unit constituting the photoelectric conversion element 10.

  FIG. 2 is a diagram illustrating a configuration of the pixel 200, the pixel circuit (PIX_BLK) 210, and the storage unit 261 included in the photoelectric conversion element 10. The pixel 200, the pixel circuit 210, and the storage unit 261 constitute a pixel unit in the photoelectric conversion element 10. The photoelectric conversion element 10 has, for example, about 7000 pixel portions for each color. Specifically, in the photoelectric conversion element 10, for example, PIX (R) 20 includes approximately 7000 pixels 200, PIX_BLK (R) 21 includes approximately 7000 pixel circuits 210, and AMEM 26 includes approximately 7000 pixels. Storage unit 261. The same applies to the other colors (GB).

  The pixel 200 includes a PD (photodiode: light receiving element) that photoelectrically converts incident light. The PD outputs the accumulated charge to the pixel circuit 210. The pixel circuit 210 includes a floating diffusion (FD) that performs charge-voltage conversion, a reset transistor that resets the FD, a transfer transistor that transfers the charge of the PD to the FD, and a source follower (SF) that buffers and outputs a signal in the subsequent stage Have A signal from the SF is read out to the subsequent stage through a read wiring. That is, SF is a buffer unit that buffers and outputs a voltage signal corresponding to the charge generated by the PD. A storage unit 261 is connected to the subsequent stage of the pixel circuit 210.

  The storage unit 261 includes a selection switch (SL) that selects the pixel 200, a current source (Is) that supplies a bias current to the SF, a selection switch (S) that selects the storage unit 261, and a memory capacity (analog memory: Cs). Have The storage unit 261 outputs a signal to the above-described AD converter. The current source (Is) is a current control circuit that controls the current flowing through the buffer unit to have a predetermined amount when the buffer unit outputs a voltage signal.

  The photoelectric conversion element 10 is a global shutter in which the writing operation to the memory capacity (Cs) is performed simultaneously for all the RGB pixels, but after the reading operation from the memory capacity (Cs), the RGB three pixels are sequentially added. Serial processing is read out to the subsequent stage.

  FIG. 3 is a diagram illustrating timing at which a signal is written from the pixel circuit 210 to the memory capacity (Cs). When the photoelectric conversion element 10 reads a signal accumulated in the PD, an SF output, that is, a signal (Vsf) at the readout wiring is output, and the pixel selection switch (SL) and the memory capacity selection switch (S) are turned on. When reading an optical signal, the signal is a signal that decreases from an initial state (reset level) to a signal level corresponding to the optical signal.

  At this time, since the storage unit 261 includes the current source (Is), the pixel circuit 210 can respond to the signal change of Vsf at high speed, and can write the signal to the memory capacity (Cs) at high speed. It can be carried out. This is because the current source (Is) ensures a sufficient current required for charging / discharging (discharging here) of the memory capacity (Cs).

  However, there is a problem that noise is increased if a current source is provided. This is because the current source responds to the signal at a high speed but also allows high-frequency noise to be tracked. For example, when high-frequency noise is generated like Vsf shown in FIG. 3, Vsf at the write period end timing (hold timing) to the memory capacity (Cs) fluctuates, and the memory capacity (Cs) The signal level (Vs) to be written is written with an error (Δ) with respect to the original level (FIG. 3: dotted line).

  At this time, the error Δ (generation of noise) generally differs for each pixel, so that it becomes FPN (Fixed-Pattern-Noise) and a vertical stripe is generated on the image. In FIG. 3, the high-frequency noise is a single frequency for the sake of explanation, but it is actually white noise containing various frequency components. Although not shown in FIG. 3, Vsf is at the reset level while the control signal RS is ON, and is at the signal level when RS is OFF and T is ON.

  As described above, the current source Is makes it possible to read a signal at a high speed, but causes a problem of FPN due to an increase in noise and causes image quality degradation. Further, since the influence of the high frequency noise cannot be removed by CDS (Correlated-Double-Sampling), it is necessary to suppress the noise itself.

  FIG. 4 is a diagram illustrating a noise spectrum in the photoelectric conversion element 10. As shown in FIG. 4A, the photoelectric conversion element 10 has characteristics in which noise exists in all frequency bands and noise intensity decreases as the frequency becomes higher (1 / f noise).

  However, as shown in FIG. 4B, the photoelectric conversion element 10 has an overwhelmingly wider band (frequency width) on the high frequency side as cumulative noise (product of unit frequency and noise amount: energy spectrum). The contribution ratio of noise increases as the frequency increases. Therefore, it is important to suppress high-frequency noise in order to reduce FPN and aliasing noise.

  FIG. 5 is a diagram illustrating a configuration example of a pixel unit for noise reduction. The pixel unit illustrated in FIG. 5 is different from the pixel unit illustrated in FIG. 2 in that the pixel unit includes a storage unit 262 in which the current source Is does not exist.

  FIG. 6 is a diagram illustrating a state of a signal read from the pixel circuit 210 to the storage unit 262. When the signal accumulated in the PD is read, the SF output (Vsf) is the same as in FIG. 3 in that the signal level is lowered from the initial state (reset level) to the signal level corresponding to the optical signal. However, since the current source (Is) is not provided in the storage unit 262, the SF operates with almost no current (sub-threshold region operation). For this reason, the response of the Vsf signal is limited, and the signal writing speed to the memory capacity (Cs) is also limited accordingly. This is because the current required for charging / discharging (discharging in this case) of the memory capacity (Cs) is small, so that the response takes time.

  Note that the initial voltage at Vs shown in FIG. 3 is set to a low level, but in the configuration shown in FIG. 2, the discharge path of the memory capacity (Cs) does not exist, so the Vs potential is only in the voltage decreasing direction. Because it cannot change.

  In the case of the sub-threshold region operation described above, the response speed of Vsf, that is, the signal band is limited. Therefore, the high frequency noise shown in FIG. 3 is suppressed, and an error of Vs due to the noise does not occur. However, since the signal response speed is limited at the same time, it takes time for Vsf to reach a desired level. This is not particularly problematic when the operation speed of the photoelectric conversion element 10 is slow, but when operating at high speed, Vs is determined (held) before reaching a desired signal level. Therefore, as in the case shown in FIG. 3, there is an error (Δ) with respect to the originally written signal level (dotted line), and the FPN due to the response being limited instead of suppressing the FPN caused by high frequency noise. Will occur.

  As described above, when the photoelectric conversion element 10 operates in the subthreshold region without the current source Is, noise itself can be suppressed, but FPN is newly generated due to limited signal response.

  FIG. 7 is a diagram showing a noise spectrum in the photoelectric conversion element 10 having the configuration shown in FIG. Noise is present in all frequency bands, and the characteristic that noise intensity decreases as the frequency becomes higher is the same as the example shown in FIG. However, in the configuration shown in FIG. 5, since the response (bandwidth) of the signal is greatly limited, noise is reduced as a whole.

  Note that the band up to the frequency fa in FIG. 7 indicates the band necessary for the signal response of Vsf. This is equivalent to the signal band on the low frequency side, and the fact that this band is limited means that the response of the signal is limited.

  FIG. 8 is a diagram illustrating a configuration example of a pixel portion of the photoelectric conversion element according to the embodiment. In order to reduce FPN, it is necessary to limit the noise band while ensuring the response of the necessary signals. In the pixel unit of the photoelectric conversion element according to the embodiment, the storage unit 261 includes a current source (Is), and includes a band limiting unit 400 that limits the noise band independently of the source follower of the pixel circuit 210. Specifically, as illustrated in FIG. 8, the pixel unit of the photoelectric conversion element according to the embodiment is provided with a band limiting unit (LIM) 400 in the subsequent stage of the SF of the pixel circuit 210.

  The band limiting unit 400 is configured by a switch (VR), for example, and is arranged in series between the pixel circuit 210 and the storage unit 261. The band limiting unit 400 can easily limit the band of the SF output signal by configuring a filter with the ON resistance of the switch (VR) and the memory capacity (Cs). That is, the band limiting unit 400 is a removal circuit that removes high frequency components of a predetermined band or higher from the voltage signal output from the buffer unit. The voltage by the switch (VR) is a DC voltage, and when reading a signal from the pixel circuit 210, the switch (VR) is always turned on.

  Here, the band formed by the band limiting unit 400 and the memory capacity (Cs) is set so as to limit the noise band while ensuring the necessary signal response (in a range that does not affect the signal followability). The ON resistance of the switch (VR) can be easily set to an arbitrary value by setting the switch size and control signal voltage.

  FIG. 9 is a diagram illustrating signals read from the pixel circuit 210 to the storage unit 261 via the band limiting unit 400. In the example shown in FIG. 9, when the signal accumulated in the PD is read, the SF output (Vsf) is reduced by the signal level from the initial state (reset level), and high-frequency noise is superimposed on Vsf. It is the same as the example shown in.

  However, since the band limiting unit 400 limits the band of the high frequency noise superimposed by Vsf, the high frequency noise is suppressed at Vlim. In the configuration shown in FIG. 8, since the band limiting unit 400 is configured independently of the source follower, the source follower itself can limit the band of the SF output signal while maintaining high speed operation. Optimization that limits the bandwidth so as not to affect the responsiveness of the signal (writing period to the memory capacity (Cs)) is possible. Therefore, in the configuration shown in FIG. 8, there is no shortage of response as shown in FIG. 6, and the signal Vlim in which noise is suppressed is held in the memory capacity (Cs).

  As described above, the band limiting unit 400 is provided between the storage unit 261 having the current source Is and the pixel circuit 210, so that the writing period to the memory capacity (Cs) is not affected. Therefore, FPN due to high frequency noise or insufficient signal response can be suppressed.

  FIG. 10 is a diagram illustrating a noise spectrum of a photoelectric conversion element including the pixel portion illustrated in FIG. 10 is the same as FIG. 4A in that noise is present in all frequency bands and the noise intensity decreases as the frequency becomes higher. However, as shown in FIG. 10, in the configuration of the pixel portion shown in FIG. 8, noise on the high frequency side is reduced with respect to the characteristic (dotted line) shown in FIG.

  Here, the noise on the low frequency side is not reduced in order to ensure the response of the signal. However, as described with reference to FIG. 4, the influence of the noise on the high frequency side is extremely small as the circuit noise in the photoelectric conversion element. Note that, as described with reference to FIG. 7, the band up to the frequency fa indicates a band necessary for the signal response of Vsf, and the pixel unit illustrated in FIG. 8 has a Vsf response by optimizing the band. Is secured.

  FIG. 11 is a diagram illustrating a configuration of a pixel portion in which circuit addition is suppressed with respect to the configuration illustrated in FIG. In the configuration shown in FIG. 8, the band limiting unit 400 is provided independently of the source follower (SF). However, as shown in FIG. 11, the band limiting unit 400 may be band limited by using a switch included in the storage unit 263. Is possible. That is, the storage unit 263 has a configuration in which the pixel selection switch (SL) also functions as a band limiting function. When the pixel unit includes the storage unit 263, the effect of limiting the band is not different from the configuration shown in FIG.

  FIG. 12 is a diagram illustrating signals read out from the pixel circuit 210 illustrated in FIG. 11 to the storage unit 263. As shown in FIG. 12, when a signal is written to the memory capacity (Cs), the control signal of the pixel switch SL and the memory selection switch S is turned on by setting it to a high level, but the storage unit 263 controls the pixel switch SL. The band is limited by setting the high level of the signal low.

  The storage unit 263 utilizes the fact that the ON resistance of the MOS transistor is changed by changing the gate voltage. The storage unit 263 performs band limitation by increasing the ON resistance when the High level for the pixel switch SL is set lower than that of the storage unit 261. The signal input to the gate of the pixel switch SL is a control signal for switching the ON / OFF state of the pixel switch SL. This ON / OFF switching is performed in the same manner as in FIG. Since the pixel switch SL (MOS switch) of the storage unit 263 is an NMOS, the value of the high level is changed. However, if the pixel switch SL is configured of the PMOS, the value of the low level may be changed. Further, the gate of the pixel switch SL is a node that can be applied with an arbitrary voltage inside the photoelectric conversion element 10, but is a node that can be applied with an arbitrary voltage from the outside via a terminal. May be.

  The high level of the control signal SL input to the pixel switch SL of the storage unit 263 can be changed. As a result, even when the high frequency noise superimposed on Vsf varies among PDs, it is possible to set the band limit appropriately. As described above, when band limiting is performed by the pixel switch SL (MOS switch) included in the storage unit 263, the FPN caused by the individual difference of PD can be reduced by changing the limiting band by changing the amplitude of the control signal SL. Can be reduced.

  The storage unit 263 has the pixel selection switch (SL) having the function of the band limiting unit 400 described above, but the memory capacity selection switch (S) may be configured to have the function of the band limiting unit 400. Good. In addition, the storage unit 263 can change the ON resistance of the MOS switch so that the band can be changed. However, the storage unit 263 may be configured so that the memory capacity value can be changed.

  Next, the configuration of the pixel portion in the photoelectric conversion element including the CDS portion that performs correlated double sampling will be described. FIG. 13 is a diagram illustrating a configuration of a pixel portion in a photoelectric conversion element including a CDS portion. Some photoelectric conversion elements realize not only the signal level in the memory capacity (Cs) of the storage unit 264 but also the CDS by holding the reset level in the memory capacity (Cr). CDS is a technique for correcting the FPN by extracting only the net signal level by subtracting the reset level from the signal level of the pixel, but CDS cannot remove the influence of high frequency noise.

  FIG. 14 is a diagram illustrating the reason why high frequency noise cannot be corrected by CDS. When CDS is performed using the pixel portion shown in FIG. 13, the memory capacity selection switch R is turned on while the pixel selection switch SL is turned on, and the reset level is first written into the memory capacity. Next, the memory capacity selection switch (S) is turned on, and the signal level is written.

  FIG. 14A shows the CDS operation when high frequency noise is superimposed. A signal is held and Vr is determined at the end of the reset period writing period. Since the level of Vsf at the end of writing is the same as the ideal level (dotted line) without noise, Vr is written with the ideal level. Next, Vs is determined at the end of the signal level writing period, but the level of Vsf at the end of Vs writing is shifted from the ideal level due to the influence of noise. Level is written. As a result, Vs−Vr subtracted by CDS has an error of Δ from the ideal level, and CDS cannot correct the influence of high frequency noise.

  On the other hand, FIG. 14B shows the CDS operation when low frequency noise is superimposed. The process for determining Vr and Vs is the same, except that Vr is written with an error (Δr) due to noise, and Vs is written with almost the same error (Δs) as Δr. This is because, in the case of low-frequency noise, the level changes for a long time. Therefore, if the CDS sampling period (interval between reset level writing and signal level writing) is shorter than the noise period, This is because there is almost no difference in signal changes.

  Therefore, Vs−Vr subtracted by CDS has a small deviation from the ideal level, and CDS can correct the influence of low frequency noise. Note that the effect of CDS on low-frequency noise is determined by the noise period and the CDS sampling period, and it is usually possible to eliminate the influence if the noise has a period of about twice or more the CDS period.

  As described above, CDS can remove the influence of low-frequency noise, but cannot remove high-frequency noise. In FIG. 14, for the sake of simplicity, an example of an output state in the dark (reset level≈signal level) is shown.

  FIG. 15 is a diagram illustrating a configuration example of a pixel unit having a storage unit 265 having a function corresponding to the band limiting unit 400, a memory capacity (Cs), and a memory capacity (Cr). Although CDS has been described above as being unable to remove high frequency noise, conversely, the effect of low frequency noise can be removed. Further, as shown in FIG. 10, the configuration shown in FIG. 8 ensures responsiveness, and thus low frequency noise cannot be limited.

  The storage unit 265 is set as a band limiting unit that limits a band in which the pixel selection switch (SL) cannot be corrected by the CDS, thereby minimizing the influence on the band to be limited, that is, the responsiveness. As described above, the photoelectric conversion element having the pixel unit including the storage unit 265 suppresses the high frequency noise by the storage unit 265 and enables the low frequency noise to be corrected by CDS. It is possible to suppress with.

  FIG. 16 is a diagram illustrating an operation of the photoelectric conversion element including the pixel portion illustrated in FIG. FIG. 16A shows a CDS operation in the case where high-frequency noise is superimposed in the pixel portion shown in FIG. The reset level / signal level writing is the same as in FIG. However, in FIG. 16A, the high-frequency noise superimposed on Vsf at Vlim is suppressed by the storage unit 265. As a result, a level with no error is written in both Vr / Vs. Therefore, no error occurs in Vs−Vr subtracted by CDS.

On the other hand, FIG. 16B shows a CDS operation when low frequency noise is superimposed.
Since the storage unit 265 cannot limit low-frequency noise, the noise superimposed on Vsf is also superimposed on Vlim in the same manner. The subsequent operation is the same as that in FIG. 14B, and since it is low-frequency noise, the deviation in Vs−Vr subtracted by CDS is small and the influence can be corrected.

  As described above, the photoelectric conversion element having the pixel portion including the storage unit 265 suppresses high-frequency noise by the storage unit 265, and enables low-frequency noise to be corrected by CDS. Can be suppressed.

  FIG. 17 is a diagram illustrating a noise spectrum of a photoelectric conversion element having a pixel portion provided with a storage portion 265. Noise is present in all frequency bands, and the characteristic that noise intensity decreases with increasing frequency is the same as in FIG. However, in FIG. 17, the noise on the high frequency side is reduced compared to the example (dotted line) shown in FIG. Here, in order to ensure the response of the signal, the noise on the low frequency side is not reduced, and the noise band that can be corrected by the CDS is not limited, which is different from FIG. In this case, since the band that can be corrected by CDS is on the higher frequency side than the band necessary for response, the influence on the responsiveness can be minimized by not limiting the CDS band.

  Note that, as described in FIG. 7, the band up to the frequency fa indicates a band necessary for the signal response of Vsf, and the band up to fb indicates a noise band that can be corrected by the CDS. In addition, although the example shown in FIG. 17 shows a state in which the band is limited from outside the band that can be corrected by CDS, that is, the band that cannot be corrected, the storage unit 265 may limit the band from the band that can be corrected by CDS. .

  FIG. 18 is a diagram illustrating the configuration of the photoelectric conversion element 10a according to the embodiment. Note that, among the components of the photoelectric conversion element 10a illustrated in FIG. 18, the same components as those illustrated in the photoelectric conversion element 10 (FIG. 1) are denoted by the same reference numerals. The AMEM 26 a has a band limiting unit row 40. The band limiting unit row 40 includes, for example, about 7000 band limiting units 400 for each color, and suppresses high frequency noise. The AMEM 26a has a memory capacity (Cs) and a memory capacity (Cr). The AMEM 26a may be configured to have about 7000 storage units 265 for each color instead of the band limiting unit row 40.

  Then, the signal whose noise is suppressed by the AMEM 26a is read out to each memory capacity in the AMEM 26a and the signal is held for each pixel, and the held signal is sequentially read out to the ADC in RGB. By holding the signal in the AMEM 26a, a global shutter system is realized in which the operation timing of the pixel 200 and the pixel circuit 210, that is, the exposure timing is simultaneous with RGB.

  The ADC 27 has the same number of AD converters as the number of columns, and sequentially AD-converts image signals in units of columns. The DCDS (digital CDS) 31 performs CDS using the reset level / signal level output from the ADC 27. The timing control unit (TG) 30a controls each unit constituting the photoelectric conversion element 10a. The band limiting unit 400 may be configured to be included in the pixel circuit 210.

  The photoelectric conversion element 10a can suppress vertical stripes due to FPN. In the case of an area sensor, SPN deterioration occurs because FPN occurs for each two-dimensionally arranged pixel, but the image quality is not as fatal as vertical stripes.

  Next, a method for adjusting the bandwidth limitation will be described. FIG. 19 is a flowchart showing a method for adjusting the bandwidth limitation. In the pixel unit shown in FIG. 11, it is possible to cope with individual differences in PD noise by making it possible to change the band limit in the band limit unit 400, but if the band is adjusted while detecting the FPN level, It becomes possible to set the optimum band for each individual.

  As shown in FIG. 19, when the adjustment is started, the user first acquires FPN data (S100). FPN data can be easily acquired by acquiring image data in a dark state. The user determines whether the level acquired by “FPN data acquisition” is equal to or less than a threshold value (S102). If the user is below the threshold, the adjustment is completed (S102: Yes), but if the threshold is exceeded (S102: No), the limited bandwidth is changed by “limited bandwidth adjustment” (S104).

  In the process of S104, as shown in FIG. 11, the band is changed by changing the VR voltage. In this case, the band is changed in a direction of further limiting in order to reduce the FPN level. Then, the user obtains FPN data again (S100), and determines whether it is equal to or less than a threshold value (S102). As described above, an optimum band can be set for each individual by adjusting the limited band. Note that the above shows a basic adjustment method, in which an upper limit is set for the number of loop processes for threshold determination and band adjustment, or a limited band is calculated according to the value of the FPN level in order to reduce the number of loop processes. Etc.

  Next, the operation of the photoelectric conversion element at the time of band limitation adjustment will be described. FIG. 20 shows a state before adjustment. The upper diagram shows main scanning output data in a dark state, and the lower diagram shows a noise spectrum. Here, the difference between the maximum value and the minimum value of the output level distribution in the output data in the main scanning direction is defined as the FPN level.

  In the noise spectrum, the frequency fb is the upper limit of the noise frequency that can be corrected by CDS, and fc indicates the cutoff frequency of the band limited by the band limiting unit. Before the band adjustment shown in FIG. 20, fc is higher than fb in terms of the noise spectrum, and therefore the band that cannot be corrected by CDS cannot be completely limited. Therefore, as shown by the main scanning level distribution, the FPN level is generated to some extent.

  FIG. 21 shows a state after adjustment. In the band adjustment, the band to be limited is changed, and after the adjustment, fc becomes lower than fb as shown in FIG. Then, the noise band that cannot be corrected by the CDS that could not be limited in FIG. 20 is limited, and the FPN level is reduced as shown in the main scanning level distribution.

  Next, an image forming apparatus including an image reading apparatus having the photoelectric conversion element 10a according to the embodiment will be described. FIG. 22 is a diagram illustrating an outline of an image forming apparatus 50 including an image reading device 60 having, for example, a photoelectric conversion element 10a. The image forming apparatus 50 is, for example, a copying machine or an MFP (Multifunction Peripheral) having an image reading device 60 and an image forming unit 70.

  The image reading device 60 includes, for example, a photoelectric conversion element 10a, an LED driver (LED_DRV) 600, and an LED 602. The LED driver 600 drives the LED 602 in synchronization with a line synchronization signal output from the timing control unit (TG) 30a. The LED 602 irradiates the original with light. The photoelectric conversion element 10a receives reflected light from a document in synchronization with a line synchronization signal or the like, and a plurality of PDs (not shown) generate charges and start accumulation. The photoelectric conversion element 10a outputs image data to the image forming unit 70 after performing AD conversion, parallel serial conversion, and the like.

  The image forming unit 70 includes a processing unit 80 and a printer engine 82, and the processing unit 80 and the printer engine 82 are connected via an interface (I / F) 84.

  The processing unit 80 includes an LVDS 800, an image processing unit 802, and a CPU 804. The CPU 804 controls each part of the image forming apparatus 50 such as the photoelectric conversion element 10a. In addition, the CPU 804 (or the timing control unit 30) controls each PD to start generating charges according to the amount of received light substantially simultaneously.

  The photoelectric conversion element 10a outputs, for example, image data of an image read by the image reading device 60, a line synchronization signal, a transmission clock, and the like to the LVDS 800. The LVDS 800 converts received image data, a line synchronization signal, a transmission clock, and the like into parallel 10-bit data. The image processing unit 802 performs image processing using the converted 10-bit data, and outputs image data and the like to the printer engine 82. The printer engine 82 performs printing using the received image data.

10, 10a Photoelectric conversion element 20 PIX (R)
21 PIX_BLK (R)
22 PIX (G)
23 PIX_BLK (G)
24 PIX (B)
25 PIX_BLK (B)
26 AMEM
27 ADC
28 P / S
29 LVDS
30, 30a TG
31 DCDS
40 Band Limiting Unit Row 50 Image Forming Device 60 Image Reading Device 70 Image Forming Unit 200 Pixel 210 Pixel Circuits 261 to 265 Storage Unit 400 Band Limiting Unit

JP 2010-178117 A

Claims (8)

A light receiving element that generates an electric charge according to the amount of light received;
A buffer unit provided for each of the light receiving elements, for buffering a voltage signal obtained by performing voltage conversion according to the electric charge generated by the light receiving element, and outputting the buffered signal to a subsequent stage;
A current control circuit that controls the current flowing through the buffer unit to have a predetermined current amount when the buffer unit outputs the voltage signal;
A removal circuit that removes a high-frequency component of a predetermined band or more in a range that does not affect the signal response of the buffer unit from the voltage signal output by the buffer unit;
A photoelectric conversion element comprising: The removal circuit includes:
Having a MOS transistor,
The photoelectric conversion element according to claim 1, wherein a band of a high-frequency component to be removed is determined in advance according to a magnitude of an on-resistance of the MOS transistor. The removal circuit includes:
The photoelectric conversion element according to claim 2, further comprising a node to which an arbitrary voltage can be applied so as to change a value of an on-resistance of the MOS transistor. A CDS unit that performs correlated double sampling on the voltage signal output from the buffer unit;
The removal circuit includes:
4. The photoelectric device according to claim 1, wherein a high frequency component in a band higher than a frequency component band that can be removed by the CDS unit is removed from the voltage signal output from the buffer unit. 5. Conversion element. The light receiving element is
The photoelectric conversion element according to any one of claims 1 to 4, wherein the photoelectric conversion elements are arranged in one direction for each color of light to be received. The removal circuit includes:
The photoelectric conversion element according to claim 5, wherein a high-frequency component is removed so that fixed pattern noise generated in the light receiving element is equal to or less than a predetermined threshold value. An image reading apparatus comprising the photoelectric conversion element according to claim 1. An image reading apparatus according to claim 7;
An image forming apparatus comprising: an image forming unit that forms an image based on an output of the image reading apparatus.

Images