JPH0414366A - Picture reader - Google Patents

Abstract

PURPOSE:To attain N-gradation reading for a picture or the like by selecting the increase in the charge to be within 1/(2N)X100% in a complete close contact type nonmagnification sensor. CONSTITUTION:The operation of the read system of a photodetector section of a complete close contact type nonmagnification sensor consists of the storage, charging and reset operation in total three operations. A capacitor Cs of the photodetector section is once charged before picture reading. The photodetector section has a total charge quantity of Qt=Cs.Vmp as a sensor. Suppose that the storage of the charge to the photodetector section is finished and a light radiates to the photodetector section, a discharge current flows from an MSS diode proportional to the exposure of the light. After the storage, when the photodetector section is selected by an analog switch, a charge current flows to compensate the charge Qt lost from the capacitor Cs of the photodetector section. A signal of a differentiating waveform outputted from an operational amplifier is integrated by an integration circuit SIGMA and a differentiating waveform proportional to the charge in the photodetector section is outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image reading device, and more particularly, to an image reading device using a full-contact, same-magnification sensor. For example, it is applied to an image reading device using a two-dimensional equal-magnification sensor.

Prior art Publicly known documents describing the prior art related to the present invention include, for example, Japanese Patent Application Laid-Open No. 62-36961 and Japanese Patent Application Laid-Open No. 64-
There is a publication No. 15970.

First, Japanese Patent Application Laid-Open No. 62-36961 relates to a contact type image sensor that has a size that corresponds one-to-one to the width of the document used for reading, and that reads the document in close contact with the document. The purpose is to form a transparent conductive layer to prevent the operation of signal processing circuits and destruction of photoelectric conversion elements due to electrostatic discharge, and there is no mention of whether multi-gradation is possible. do not have.

Furthermore, Japanese Patent Application Laid-Open No. 15970/1983 relates to an image reading device that reads image information while relatively moving a document for image reading while in close contact with a -dimensional line sensor. a light-shielding layer disposed on the surface of a transparent substrate, an insulating layer disposed on the light-shielding layer, an optical sensor disposed on the insulating layer, and a protective layer disposed on the optical sensor, Image information is read by receiving reflected light from the back surface of the substrate onto the document surface using an optical sensor, and light shielding means is provided on the side surface of the substrate and at least part of the surface and side surface of the protective layer. It is something.

However, although the image reading device proposed in this publication describes specific light shielding means and a method for preventing an increase in sensor output due to backside light, this does not suggest whether multi-gradation is possible. It has not been.

In addition, the previously proposed "image sensor" contributed to an increase in sensor output due to stray light from the edge (mainly when the sensor was dark during backside illumination), but it also ensured multi-gradation. Since noise is not defined for this purpose, there is no suggestion as to whether multi-gradation is possible.

In a fully contact type sensor, for example, the light source, the document, and the motor for feeding the document are used.

It will be placed near the image sensor.

In other words, noise caused by turning on the light source, noise from the motor, and static electricity noise caused by friction when the document is fed can easily enter the image sensor, especially its photoelectric conversion section and drive circuit. It means to end up.

The increase in noise that accompanies such miniaturization raises the possibility that the S/N ratio during image information reading will decrease and that multi-gradation reading will become difficult. Countermeasures against this problem have been proposed, and the above-mentioned "image sensor" is one of them. However, this was not a condition for ensuring N gradation.

■--Niku The present invention was made in view of the above-mentioned circumstances.
The purpose of this invention is to provide an image reading device that is capable of reading multiple gradations of N or more gradations using a full-contact, same-magnification sensor.

In order to achieve the above object, the present invention includes (1) a transparent substrate, a photoelectric conversion element arranged on the transparent substrate, and a transparent protective layer formed on the photoelectric conversion element. The transparent protective layer is brought into close contact with the document, illumination light is irradiated from the back surface of the transparent substrate, and the photoelectric conversion element receives reflected light from the document, and the structure is configured such that the transparent protective layer is brought into close contact with the document, and the light reflected from the document is received by the photoelectric conversion element. In the full contact type 1x sensor, the amount of noise that increases due to the input light from the back side of the transparent substrate compared to when it is completely dark is 1/(2N) with respect to the white output of the 1x sensor.
(2) Consists of a transparent substrate, photoelectric conversion elements arranged on the transparent substrate, and a transparent protective layer formed on the photoelectric conversion element, and the transparent protective layer is A full-contact, life-size sensor that is placed in close contact with an original, irradiates illumination light from the back side of the transparent substrate, and is configured to receive reflected light from the original with the photoelectric conversion element, and is capable of accumulative reading of N or more gradations. The width of the charge amount of the noise in the output of the same-size sensor is within 1/(2N)×100% of the charge amount of the white output of the same-size sensor. Hereinafter, the present invention will be explained based on examples.

First, in the current full-contact, same-magnification sensor, our target for multilevel reading is 16 gradations, and it is necessary to distribute the total charge so that it can at least be divided into 16 equal parts. Then, 6% of the total amount of charge becomes the amount of charge required for one gradation. However, when predicting the curvature of the γ characteristic (the slope of sensor output and exposure amount) and the tolerance of gradation, it is considered appropriate to set it to half of that, 3%.
%, it is also possible to maintain the difference between adjacent gradations as an advantage.

Figure 4 is a diagram showing the γ characteristic curve considering gradation.
FIG. 5 is a γ characteristic curve when white waveform flatness (PRNU) is taken into account.

As shown in Figure 4, when considering the characteristic curve of the sensor output against the exposure amount, in the case of 16 gradations, the leakage charge amount is simply 6% of the total charge amount, but γ When predicting the characteristic curvature and gradation tolerance, it is considered appropriate to set it to half of that, 3%, as described above, and this also makes it easier to distinguish between adjacent gradations. It becomes possible.

Furthermore, it is necessary to consider the influence of white waveform flatness (PRNU) on the γ characteristics, and when this is taken into consideration, the situation shown in FIG. 5 can be considered. This means that the γ characteristic curve also has variations, which are included in the above 3%.

Among these, what is particularly problematic is the increase in sensor output due to backside light in a state of complete contact. This is especially true for
This is mostly due to the fact that the sensor section is not sufficiently shielded from light, and also due to stray light components from the end face of the sensor. In this way, in an image reading device using a light-transmitting substrate, light reflected from the side surface of the substrate or direct light from the light source may be reflected directly or by the upper light-shielding layer, or may be reflected by the sensor b.
The photoelectric conversion element is irradiated by the leakage light between the IT. As a result, in addition to the photocurrent due to light reflected from the document as original image information, the above-mentioned stray light is generated.

A photocurrent is generated due to the leaked light, resulting in a decrease in the S/N ratio. This is shown in Figures 3 (a) and (b).
).

FIGS. 3(a) and 3(b) are diagrams showing the optical paths of stray light and leakage light from the sensor output, in which 21 is the sensor section, 22 is the light shielding section, 23 is the thin glass, 24 is the adhesive, 25 is the substrate, 2
6 is a support, and 27 is a Xe lamp.

Figure 3 (a) is a diagram showing the stray light component from the edge of the sensor board. Considering the width direction of the sensor board, the light from the back side is reflected by the light from the light shielding part directly below the sensor, and the light is irradiated. In this state, stray light components appear on the sensor section.

In this state, thinking power increases. This is particularly a problem when reading a document whose surface is black. In other words, when irradiated with light, the memory power varies within the plane, resulting in variation in γ. Therefore, linearity regarding gradation is impaired.

Figure 3(b) is a diagram showing the inter-bit leakage light components of the sensor board. Considering the sensor board in the longitudinal direction, the light from the back surface is due to the leakage of light other than the light directly below the sensor, so the light is In the irradiated state, it appears as a leaked light component in the sensor section. In this state, the thinking power increases as before. This is particularly a problem when reading a JM manuscript in a black state. In other words, this causes variations in light irradiation, resulting in variations in γ. For this reason, linearity with respect to gradation is impaired.

Figure 1 is a diagram showing the output status of the sensor in the dark state due to light irradiation from the back side of the board, and shows the increase in memory power in relation to the white output, indicating that the output increases by irradiating with a lamp. I understand. The phenomenon of variation in γ cannot be completely eliminated even if various countermeasures are taken. However, if the conditions for ensuring N gradation properties are satisfied, multi-gradation reading becomes possible.

Here, the output shown in FIG. 1 is described by the amount of charge in the storage type sensor. In other words, for the charge amount Qt of white output,
The amount of increase in thinking power is Qm, and the following conditions are required to obtain N gradation.

Qt 2N As shown in FIG. 2, this is the linearity of the exposure amount and the sensor output with respect to the increase in the thinking power. In particular, linearity deteriorates on the low exposure side.

In order to avoid this, it is necessary to suppress the increase in output on the low exposure side. On the other hand, this increase in output also affects white output, but it can be ignored depending on its absolute value.

FIG. 6 is a diagram showing the output waveform of the complete contact type sensor. Normally, if the noise of the memory output has N gradations with respect to the white output of the sensor, then it is l/ (2N)
It is necessary to be within ×100%. This is because these outputs can be converted into charge amounts, and the charge amount Qn due to noise with respect to the charge amount Qt of the white output of the sensor is required to satisfy the following relationship.

Qt 2N A fully contact type 1x sensor that satisfies this condition.

It becomes possible to read signals of N gradation signals. The amount of charge converted into the width of this noise is generated regardless of the white output state or the memory output state, and it is necessary to suppress these charges within the linear region of the γ characteristic curve.

Note that the range of this noise is white output, memory output, and 1b
There is a temporal change in the unit of it, and there is also a change due to in-plane uniformity. This is shown by Qnl, which represents the 1-bit temporal change in thinking ability, and Qn2, which represents the variation within the area of thinking ability.
, the temporal change of 1 bit of white output is Qn3. If the in-plane variation of white output is Qn4, it is necessary to satisfy the following.

FIG. 7 shows this. In FIG. 7, the exposure amount and sensor output are taken, and their characteristic curves are drawn.

In this way, no matter how good the linearity of γ is, if the factor of variation due to noise enters into it, this linearity will be distorted and multi-gradation property will be impaired. At this time, it goes without saying that the limit at which this multi-gradation property is impaired is the above-mentioned condition for a fully contact type 1x sensor with N gradations.

FIG. 8 is a cross-sectional configuration diagram of the MSS sensor.

31 is the manuscript, 32 is 5iON (1, cim), 33 is 5iON (5000 people), 34 is P1 type a-S i
(amorphous silicon): OH (400 people), 35 is a-8i: OH (200 people), 36 is a-8i: H (
5000 people), 37 is An (1 μm), 38 is Cr (3
00 people), 39 is ITO (indium tin oxide, 7
5o people), 4o is Cr (1000 people), and 41 is a lighting window. In addition, FIG. 9 shows T F T (Thin Fil
5 in the figure.
1 is gate electrode polysilicon (3500 people), 52 is gate oxide film (850 people), 53 is polysilicon (750 people)
54 is AQ (1μm), 55 is 5iON (1μm)
m), 56ha 5iON (5000 people), 57 is 510
2 (1500 people), 58 is the Nch and Pch diffusion layer.

Figures 10 (a) to (k) are diagrams showing the process of forming a 1-size sensor.
In figure (b), the active layer of NchtTFT is B”
(boron) to dope the channel with 5E12 (5
X1012) 1/ca+2. And 102
A gate oxide film (
850 people). Next, as shown in FIG. 3(c), polysilicon (3000 layers) to be used as a gate electrode is deposited by the CVD method. After that, in Figure (d), a gate electrode is formed by dry etching, and an N
Select the ch active layer and make the P4' type 4E15 (4X10
1s) Inject 1/cm2.

Then, in Figure (e), apply B+ 2E15 (2X
101'') 1/cm' is implanted to form a Pch diffusion layer. At the same time, the resistance of the polysilicon gate is also lowered. Next, in Figure (f), a Si
, 1500 films were deposited by low pressure CVD method, and 5in.
In order to increase the density of the two films and promote planarization, annealing is performed at 90° C. in an 02 atmosphere. This also serves to activate the diffusion layer and the polysilicon gate. In figure (g), 1000 people deposited Cr, which will become the lower electrode of the light receiving part.
In figure (h), the sensor part that becomes the light receiving part is a-8i:I
-((5000 people), a-5i: OH (200 people)
, and deposit P+ type a-8i:OH (400λ). In Figure (i), there is an IT
Form an O film (750 people) and a Cr film (300 people). In Figure (j), in order to enable metal wiring from the light receiving part and TPT, a plasma CVD method is used to form an insulating film between the eyebrows.
JSiONlli (5000 people) was formed, and A1 electrode (
lum). Next, in figure (k), from above,
A 5iON film (1 μm), which serves as passivation for the device, is deposited by plasma CVD. After that, adhesive and thin film glass (glass for wear-resistant layer) are attached to the top. As described above, the sensor substrate is formed.

Regarding the formation of the sensor unit, it is completed by mounting it on a printed circuit board and mounting the Xe lamp.

FIG. 11 is a circuit diagram of a full-contact type 1-magnification sensor, showing a light receiving section, a driving circuit section, and a reading circuit system. One pixel is MSS (Metal-Sem1c
inductor-Sem1insulator)
It consists of a diode and its capacitance (Cs).

The drive circuit section is an analog switch (ASW) using TPT.
) and a dynamic shift register (DS/R). Each of them has one bit arranged for one pixel dot. Dynamic shift register is clocked CM
Consists of O8 and CMO8 for data output,
Data retention is the capacitance between the gate and source on the CMO8 side (
This is made possible by accumulating charges in Cgs).

The data transfer (D in) is controlled by an external two-phase clock (ψ, ψ).

The reading circuit system uses an operational amplifier to read the charging current from the diode, sends the output to the subsequent integral output, and uses a peak hold circuit to gradate the output. Thereafter, the integrated output is canceled by the reset signal ψR. Also,
Noise inherent in the read output is reduced by detecting the offset level of the previous bit and canceling it out.

Next, the operation of the reading system of the light receiving section will be explained. The operation of the reading system of the light receiving part of the full contact type 1x sensor is as follows (
1) W type operation, (2) Charging operation.

(3) It consists of three reset operations. FIG. 12 shows potential changes in each part. Each operation of the light receiving section and the reading system will be explained with attention to one bit according to FIGS. 9 and 10.

(1) Accumulation operation First, before reading an image, the capacitance Cs of the light receiving section is charged once. As a result, all the light receiving parts are Qt=Cs-Va
It has the total amount of charge as a sensor of mp. It is now assumed that the accumulation of charge in the light receiving section has ended, and when the light receiving section is subsequently irradiated with light, the MSS diode causes a discharge current to flow in proportion to the amount of light exposure. As a result, the amount of charge Qt of the capacitance Cs of the light receiving section decreases. The amount of decrease ΔQt is as shown in the following formula. Here, I
p is the current value of the MSS diode during light irradiation, Tsave
represents the accumulation time from once a light receiving section is selected until the next time that light receiving section is selected.

ΔQt=I p 1Tsave ΔVsense=ΔQt/C5 (2) Charging operation After the above accumulation operation, when the light receiving section is selected again by the analog switch, the charge ΔQt lost by the capacitance Cs of the light receiving section
A charging current flows to compensate for this.

The operational amplifier OP shown in Fig. 11 detects this and the current -
Perform voltage conversion. This output voltage is determined by the feedback resistor Rf
Determined by eed. The output voltage Vop is expressed by the following formula. Here, I 5ense is a charging current flowing through the light receiving section.

Vop=I 5ense 3Rfeed Also, the waveform of the charging current is a differential waveform because it is a charging current to the capacitor Cs. Here, when the differential waveform is expressed in the formula, it becomes as follows. Here, I is the maximum value of the differential current, and t is the time. 6 (3) Reset operation The differential waveform output from the operational amplifier is then integrated by the integrating circuit Σ, and the differential waveform is proportional to the charge charged in the light receiving section. It is output in the following format. The output is read by a peak hold circuit. Thereafter, the charge stored in the capacitor Csigma of the integrating circuit is discharged by turning on the switch Msigma, and the integrated output is reset. This causes the integrated output potential to drop to ground potential. Thereafter, the output of each light-receiving bit can be read accurately. The reset timing of ψR is generated before the rising edge of the control signal clock.

Although the operation for 1 bit has been described above, it is actually formed of a 1760-bit light receiving section, and the read output is detected every half cycle of the clock.

Next, the operation of the drive circuit section will be explained.

The drive circuit is composed of a dynamic shift register, and charge can be held by accumulating the charge in the capacitance between the gate and source of 0MO8.

The operation of this drive circuit section is as follows: (1) reading data, (
It consists of three parts: 2) data retention, and (3) data output. The circuit diagram and potential changes of each part are shown in FIGS. 11 and 13. Each operation of the active circuit will be explained according to these. Here, ■1, ■2, v3, v4, and v5 indicate the voltages at the input and output terminals of each 0MO8.

(1) Reading data First, data is input from the Din side in negative logic, and the clocked CMOs 8 on both sides are turned on so that the clocked CMO 5(a) becomes active. In this state, the data is positive logic and clocked CMOS
(a), and its output is the subsequent stage 0MO8 (b
) is transmitted to the next clock CMO5(c). In this state, the next stage clock CMOS (c) is not in an active state because it receives a clock signal of the opposite phase. In other words, this indicates that the control signal transfers data bit by bit.

(2) Holding data To hold data, the charge output from the clocked CMO 8 (a) is accumulated in the capacitor Cgsl of the next stage 0 MO 8 (b).

Then, the output of 0MO8(b) is injected into the input gate of the next clock CMO5(c) within the discharge time of the capacitor Cgsl, and this state is maintained while the clock signal is inverted. In this state, the ability to retain data is particularly limited by the clock distortions τΦrise and τΦfa.
It depends on ll. This is shown in Figure 13, where in this signal displacement section, CMO5 (b
) is in a state where the charge of Cgsl is easily discharged. However, the clock distortion τΦrise.

By optimizing the value of τΦfall with Ion of TPT, a state where discharge is likely to occur can be avoided.

(3) Data output The data output from the clocked CMO8(a) and 0MO8(b) has the function of not only data transfer but also activating the analog switch.

In this case, the data output from CMO5(b) that constitutes the shift register is transferred to CMO3(b) installed above it.
It is transmitted to the input gate of the buffer formed by two series of e), (f), and (g). This buffer has the function of shaping the distorted output waveform output from the shift register, and improves the rise and fall characteristics of the analog switch ASW. This allows the analog switch ASW
When the shift register is selected, the output direction can be automatically transmitted to the light receiving section.

Regarding the increase in evaporation power due to backside light, see the 14th section.
This can be verified by the methods shown in Figures (a) and (b).

Figure (a) shows the case of complete darkness.

Figure (b) shows the case of downward Xe irradiation. The evaporation power of a fully contact type sensor is measured in complete darkness, and then the increase in noise is measured by irradiating light from below. Regarding the measurement, the amount of charge is converted back into voltage based on the output value of the integrating circuit as shown in FIG. 11, and the measurement is performed.

Further, FIG. 15 is a cross-sectional view of the same-magnification sensor according to the present invention, in which l is a thin plate glass, 2 is an adhesive, 3 is a quartz substrate,
4 is pansivation, 5 is TPT, 6 is lighting window, 7 is optical sensor, 8 is primary electrode, 9 is AQ electrode, 10 is sensor substrate, 11 is support body, 12 is lamp, 13 is presser roller, 14 is It is a manuscript.

Optical sensor (photoelectric conversion unit) 7 on quartz substrate (transparent substrate) 3
are arranged, and a thin plate glass (transparent protective layer) l is provided on the optical sensor 7 via an adhesive 2. The thin plate glass 1 is brought into close contact with the original 14, illumination light is irradiated from the lamp 12 onto the original 14 from the back side of the quartz substrate 3 through the lighting window 6, and the reflected light from the original 14 is received by the optical sensor 7.

As is clear from the above explanation, since the increase in charge is within 1/(2N) x 100% in a fully contact type 1x sensor, it is possible to read N gradations of photographs, etc. .

[Brief explanation of drawings]

Fig. 1 is a diagram showing the output status of the sensor in the dark state due to light irradiation from the back side of the substrate in the image reading device according to the present invention, Fig. 2 is a diagram showing the γ characteristic curve, and Fig. 3 is a diagram showing the sensor output state. Figure 4 shows the optical path of stray light and leakage light. Figure 4 shows the γ characteristic curve considering gradation. Figure 5 shows the γ characteristic curve considering white waveform flatness (PRNU). 6 is a diagram showing the sensor output of the present invention, and FIG. 7 is a diagram showing the sensor output of the present invention.
A diagram showing the γ characteristic curve in the case of Fig. 6, and Fig. 8 shows the MS
The cross-sectional view of the S sensor, Figure 9, is the cross-sectional view of the TPT, Figure 10.
The figure shows the formation process of the equal-size sensor, the surface 11 is a circuit configuration diagram of the equal-size sensor, FIG. 12 is a diagram showing the potential of each part of the circuit of the equal-size sensor, and FIG. Figure 14 is a diagram showing the potential changes of various parts in the drive circuit of the double sensor, and Figure 15 is a diagram showing the experimental environment for investigating the leakage current dependence of the white output.
The figure is a sectional view of a 1-magnification sensor according to the present invention. 1... Thin glass (transparent protective layer), 3... Quartz substrate (transparent substrate), 5... TPT (thin film transistor)
, 6... Lighting window, 7... Optical sensor (photoelectric conversion element)
, 12... lamp, 14... manuscript. Patent applicant Rico Co., Ltd. (1 idiot) Figure 1 Figure 3 (a) (b) Figure 2 Figure 4 Figure 5 Exposure amount n⇠I Exposure amount chart Exposure amount chart +a1 Activity Layer formation poly-sr (b) ch, Boron injection cd) S, O type tL -1 (N-chan) Phosphorus (P) injection (e) S, D formation 2 (P chan) Pon ( Boron) resident (h) o-si, - Etching to large electrode form tJ secondary electrode formation (k passivation formation Fig. l Fig. I Body diagram Fig. 16 Read and hold

Claims (1)

[Claims] 1. Consisting of a transparent substrate, photoelectric conversion elements arranged on the transparent substrate, and a transparent protective layer formed on the photoelectric conversion elements, the transparent protective layer is brought into close contact with the original. , in a fully contact type 1x sensor capable of accumulative reading of N or more gradations, the sensor is configured to emit illumination light from the back surface of the transparent substrate and receive reflected light from the original at the photoelectric conversion element; Due to the input light from the back side of the board, the amount of noise that increases compared to when it is completely dark is 1 compared to the white output of the same size sensor.
An image reading device characterized in that the value is within /(2N)×100%. 2. Consisting of a transparent substrate, photoelectric conversion elements arranged on the transparent substrate, and a transparent protective layer formed on the photoelectric conversion element, the transparent protective layer is brought into close contact with the document, and the back side of the transparent substrate is In a full-contact type 1x sensor that is configured to emit illumination light from the source and receive reflected light from the original at the photoelectric conversion element, and that can perform cumulative reading of N or more gradations, noise in the 1x sensor output can be reduced. An image reading device characterized in that the width of the charge amount is within 1/(2N)×100% of the charge amount of the white output of the same-size sensor.