JPS642272B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a solid-state imaging device, and in particular, the present invention relates to a solid-state imaging device, and particularly to a solid-state imaging device in which signals from light-receiving elements formed in a two-dimensional array are transmitted through a MOS type switching element whose opening and closing are controlled by pulses. This invention relates to an improvement in a signal processing circuit for a solid-state imaging device that is read out one after another.

[Background of the invention]

A large number of photodiodes (e.g., 500 x 400) are formed in both vertical and horizontal directions on a semiconductor substrate, and the charge stored in each photodiode is transferred by driving pulses in the vertical and horizontal directions. A solid-state imaging device in which signals are successively read out on a signal output line through MOS type switching elements whose opening and closing are controlled is known, for example, in Japanese Patent Laid-Open No. 54-37427, proposed by the present applicant.

This type of solid-state imaging device has a problem in that various clock pulses leak into the signal output line via stray capacitance when reading signals from each light receiving element, deteriorating the signal-to-noise ratio (S/N). . In particular, the approximately 7MHz clock pulse that drives the horizontal MOS type switching element acts equivalently as a low impedance (several hundred Ω) signal source for the preamplifier connected to the signal output line, and the output of the preamplifier is A noise component synchronized with the clock pulse appears with a magnitude more than 10 times that of the optical signal.

For this reason, conventionally, an amplifier with a sufficiently large dynamic range is used as a preamplifier, the above-mentioned noise component and optical signal component are amplified together, and a filter is connected after the preamplifier to remove the clock pulse noise component. In this way, only the optical signals were separated. According to the sampling theorem, a frequency up to half the horizontal clock frequency is a component that can be effectively used, and this will hereinafter be referred to as the video band. In addition, in a solid-state imaging device, if the accumulated charge from the light receiving element remains unread, the resolution in the horizontal and vertical directions deteriorates, so the preamplifier described above has a low input impedance (several kilohms or less).
In addition, amplifiers with flat frequency characteristics up to around 10MHz have been used.

However, in order to satisfy the above conditions, the power supply for the preamplifier must have a higher voltage value than the power supply for other drive circuits, especially in order to have a large dynamic range, which hinders the miniaturization of the device and the use of a single power supply. It was becoming. Furthermore, in order to lower the input impedance, the feedback resistance cannot be increased, and as a result, there is a problem in that random noise generated from the feedback resistance cannot be reduced.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a signal processing circuit for a solid-state imaging device that does not interfere with reading optical signals from a light receiving element even if the dynamic range of the preamplifier is narrow.

[Summary of the invention]

The present invention uses a preamplifier that amplifies the output signal from the light receiving element, which is equipped with a feedback loop in which the output signal is fed back to the input side via an impedance element, and the open loop frequency characteristic of the preamplifier is The low-pass characteristic has a cutoff frequency f ₁ within the video band, and the cutoff frequency f ₂ determined by the capacitance C ₀ present at the input end of the preamplifier and the input impedance R _io of the preamplifier is also within the video band. The feature is that it is set so that it exists. Furthermore, by setting the cutoff frequency _f3 of the preamplifier's feedback loop within the video band, a signal processing circuit that can read out signals without deteriorating the S/N ratio even if the preamplifier's dynamic range is narrow can be obtained. It will be done.

[Embodiments of the invention]

Hereinafter, details of the present invention will be explained with reference to the drawings.

FIG. 1 is a circuit configuration diagram of a solid-state imaging device and its signal processing circuit, and for convenience of explanation, it consists of one photodiode 1 and MOS switches 3 and 6 necessary to read out optical signals from this diode. This figure shows a typical simplified circuit configuration.

A current proportional to the amount of incident light flows through the photodiode 1, thereby accumulating a signal charge Q ₀ in the parasitic capacitance _CP . One end of the photodiode 1 is controlled by the output pulse of the vertical shift register 2
MOS transistor 3 and horizontal shift register 5
The signal output line 7 is connected to the signal output line 7 via a MOS transistor 6 controlled by the MOS transistor 6 .

Vertical shift register 2 outputs a clock pulse every 1 field period, and MOS transistor 3
conduction. Since the capacitance C ₁ of the vertical signal line 4 is several tens to hundreds of times larger than the parasitic capacitance C _P of the photodiode 1, most of the signal charge Q ₀ is absorbed by the capacitance C when the MOS transistor 3 becomes conductive. It will be transferred to ₁ .

In this state, when the MOS transistor 6 is turned on by a clock pulse (for example, 7 MHz) from the horizontal shift register 5, the charge accumulated in the capacitor _C1 flows to the signal output line 7. A preamplifier 9 with a feedback loop 8 of impedance Z _f is connected to the signal output line 7.
is connected, and the combined capacitance of the parasitic capacitance of the input stage of the preamplifier 9 and the capacitance of the output line of the solid-state image sensor is
Assuming C ₀ , the signal readout circuit can be approximated as shown in FIG.

Here, assuming that there is a signal current source in parallel to the capacitor C ₀ as shown in FIG. 2, calculate how this power source flows to the preamplifier side. The voltage at the input end 7 of the preamplifier is V _i , and the voltage at the output end 10 is V ₀
Then, the following formula holds true. Note that s is a Laplace operator.

I _s =V _i sC ₀ +(V _i −V ₀ )1/Z _f (1) V ₀ =−GV _i (2) Here, it is assumed that the gain G of the preamplifier is given by the following equation.

G=G ₀ /1+jf/f ₁ (3) G ₀ is the DC gain, and f ₁ is the cutoff frequency of the amplifier. Furthermore, the impedance Z _f of the feedback loop is determined by the resistance R _f and the capacitance C _f as shown in Figure 2.
If it is parallel, it is given as follows.

Z _f =R _f /1+jf/f ₃ ...(4) Here, f ₃ =1/2πC _f R _f ...(5).

Now, when the transfer function of the preamplifier output V ₀ to the signal current I _s is calculated from equations (1) and (2), it is as follows.

V ₀ /I _s = -GZ _f /1+jf2πC ₀ Z _f +G Z _f /1+jf2πC ₀ Z _f /G...(6) Here, the gain G of the preamplifier is about 1000, so 1/G is is ignored. G
has a cutoff frequency f ₁ given by equation (3), 1/G can be ignored if f ₁ is 1/100 or more of the highest frequency of the video. In addition, there is no deterioration of the f characteristic of the preamplifier, and the feedback impedance Z _f
If there is no capacitance C _f , then (6) becomes as follows.

V ₀ /I _s =−R _f /1+jf2πC ₀ R _f /G ₀ =R _f /1+jf/f ₂ …(
7) Here, f ₂ is given by the following equation.

f ₂ = G ₀ /2πC ₀ R _f …(8) If the above f ₂ is within the video band, the frequency characteristics will deteriorate, so in order to reduce the deterioration of the frequency characteristics, set R _f and G ₀ appropriately. choose. For example, C ₀ =
When 30pF, G ₀ = 1000, and R _f = 500KΩ,
f ₂ =10.6MHz, and deterioration of frequency characteristics within the video band can be almost ignored.

Next, the problem of clock noise, which is the subject of the present invention, will be described. The problem with the MOS type image pickup device is that the clock pulse that drives the horizontal switch 6 leaks into the signal line 7 via the parasitic capacitance. Since the output impedance of this noise is low, at a few kilohms or less, there is almost no feedback from the preamplifier regarding this clock noise.
Therefore, clock noise multiplied by the open loop gain G of the amplifier is output from the preamplifier. Since this noise is more than 10 times larger than the signal, conventionally the final stage of the preamplifier required a device with an extremely large dynamic range.
For example, if the signal current is 1 μA and the feedback resistor R _f is 500KΩ, the preamplifier output will be 0.5V _PP , but clock noise will occur over 10V _PP . For this reason, the preamplifier requires a power supply voltage of about ±10V, which requires a separate power supply from that of other circuits.

In the present invention, the influence of clock noise is reduced by giving the open loop gain G of the preamplifier a special frequency characteristic. For example, if the cutoff frequency f ₁ of the preamplifier is 0.7MHz, the gain at 7MHz will be about 1/10, so a dynamic range of about 1V _PP will be sufficient for the final stage of the preamplifier.

However, simply lowering the cutoff frequency of the preamplifier causes the following problems. That is, in equation (6), if Z _f =R _f and G = G ₀ /1 + jf/f ₁ , the transfer function is calculated as follows.

V ₀ /I _s =-R _f /1+if2πC ₀ R _f 1+jf/f ₁ /G ₀ =-R _f /1+jf/f ₂・(1+jf/f ₁ ) =-R _f /1+jf/f ₂ +(jf/ f ₂ ) (jf/f ₁ ) …(9) Here, f ₁ and f ₂ are positive and f ₁ < 4f ₂ , so the transfer characteristic given by equation (9) shows the characteristics of a second-order system, f′=
It has a characteristic with a large peak at the frequency of √ _{1 2} . For example, if f ₁ = 0.7 MHz and f ₂ = 10.6 MHz, then f' = 2.7 MHz, and in this case, |V ₀ /I _s (f = f') | ≒ 3.9R _f (10), so the DC gain This results in a gain of about 4 times. This characteristic is shown in FIG. In order to make this a flat frequency characteristic within the necessary video band, a complicated correction circuit is required. Therefore, in the present invention, in order to prevent the above-mentioned peak from occurring in the gain characteristics, it is considered that the impedance Z _f of the feedback loop has a frequency characteristic.

In other words, as Z _f , the characteristic given by (4) Z _f = R _f /1 + jf / f ₃ ... (4) However, assuming f ₃ = 1/2πC _f R _f and substituting it into (6), V ₀ /I _s =−R _f (1+jf/f ₃ )/1+jf・2πC ₀ 1+jf/f
₁ /G ₀ ×R _f /1+if/f ₃ =−R _f /1+jf/f ₃ +jf/f ₂ (1+jf/f ₁ ) =−R _f /1+jf (1/f ₃ +1/f ₂ )+(jf /f ₂ ) (jf
/f ₁ ) = −f ₁ f ₂・R _f / (jf) ² + (jf) (f ₁ f ₂ /f ₃ +f ₁ ) + f
₁ f ₂ …(11). In this equation, when K=(f ₁ f ₂ /f ₃ +f ₁ ) ² −4f ₁ f ₂ ≧0 (12), the denominator can be factorized. At this time By setting f ₊ , you can set it outside the video band. For example, in order to satisfy equation (12) when f ₁ =0.7 MHz and f ₂ =10.6 MHz, f ₃ may be set to 1.56 MHz or less. Furthermore, for example, if f ₃ = 1MHz, then f ₊ =
7.07MHz, f _- = 1.05MHz. Therefore, in this case, since the frequency of f ₊ is sufficiently higher than the video band, f _- =
This means that only high-frequency compensation needs to be performed for frequencies above 1.05MHz.

Next, the gain G ₀ of the preamplifier and the capacitance at the input end
The cutoff frequency f ₂ determined from C ₀ and feedback resistance R _f will be described. As mentioned above, C ₀ = 30pF,
When G ₀ = 1000, if R _f = about 500KΩ, f ₂ =
10.6MHz, and _f2 can be set outside the video band. However, since the noise current generated from the preamplifier is given by the following equation,
In order to improve the S/N ratio, it is necessary to make the value of R _f as large as possible. _o ² = 4kTf ₀ (1/R _f + 4/3π ² C ₀ ² f ₀ ² r _eq ) …(15) Here, K is Boltzmann's constant, T is absolute temperature, f ₀
is the highest frequency, and r _eq is the equivalent noise resistance of the FET in the first stage of the preamplifier.

However, when R _f is increased, the above-mentioned cutoff frequency f ₂ becomes lower and falls within the video band. To prevent this, it is necessary to increase the gain G of the preamplifier, but increasing the gain over 1000 times is undesirable from the viewpoint of circuit oscillation and stability.

The present invention is characterized in that the cutoff frequency _f2 is also set within the video band. That is, cutoff frequencies f ₁ , f ₂ , and f ₃ are set within the video band, and K given by equation (12) is selected to be positive. In this way, the transfer function given by equation (11) is given in the form of two stages of first-order delay elements. If this cutoff frequency is f ₊ and f _- , it becomes possible to select the frequency of f ₊ outside the video band. For example, even if f ₁ = 0.7MHz and f ₂ = 2MHz, if f ₃ = 0.2MHz, f ₊ =
7.5MHz, f _- = 0.186MHz. Therefore, it can be seen that in the downstream stage of the preamplifier, it is only necessary to perform high-frequency compensation for the f _- cutoff frequency.
The transfer characteristics in this case are shown in FIG. The dotted line indicates the frequency characteristics of the high-frequency compensation circuit to be corrected in the subsequent stage of the preamplifier.

By setting f ₊ outside the video band as described above, it is possible to design a preamplifier that is not affected by clock noise and has a simple correction circuit.

FIG. 5 shows an embodiment of the preamplifier 9 according to the present invention. 11 is FET for first stage amplification, 12 is this
This is the load resistance R _L of the FET. If the forward transfer admittance of the FET 11 is g _n , then the voltage gain of the first stage is expressed as g _n R _L . 13 is the second stage amplifier,
14 is a resistance (R ₁₄ ), and 15 is a capacitance (C ₁₅ ). These elements are for determining open loop characteristics and are set so that f ₁ =1/2πC ₁₅ R ₁₄ . 16 is a final stage amplifier, and the open loop gain G of this preamplifier is expressed as follows.

G=g _n R _L・G1・1/1+jf/f ₁ G2 =G ₀ 1/1+jf/f ₁ (16) Here, G ₀ =g _n R _L・G1・G2. Feedback impedance Z _f is expressed as Z _f =R _f /1+jf/f ₃ when configured by a parallel circuit of R _f and C _f . Here, f ₃ =1/2πC _f R _f .

Now, for example, f ₁ = 0.7MHz, f ₂ = 2MHz, f ₃ =
To set it to 0.2MHz, the constants of each circuit element can be selected as follows.

R ₁₄ = 1KΩ C ₁₅ = 1/2π・R ₁₄・f ₁ = 227pF R _f = 2MHz C _f = 1/2πR _f・f ₃ = 0.4pF Also, for example, if the capacitance at the input end of the preamplifier is
In the case of 30 pF, the gain G ₀ may be selected as G ₀ =2πC ₀ R _f ×f ₂ =754. With this choice, the transfer function is expressed as _V0 / _Is = _-Rf /(1+jf/f ₊ )(1+jf/ _f- ). where f ₊ = 7.5MHz, f _- = 0.186MHz
Therefore, high-frequency compensation can be performed by a circuit having a transfer function of H(jf)=1+jf/f _- at the subsequent stage of the preamplifier. Note that this characteristic can be realized by a simple emitter peaking circuit as shown in FIG.

〔Effect of the invention〕

According to the present invention, it is possible to easily design a circuit that is less affected by clock noise in a preamplifier for a MOS image sensor. As a result, it is not necessary to make the power supply voltage of the preamplifier extremely high, and the noise caused by the feedback resistance of the preamplifier can be significantly reduced.

[Brief explanation of drawings]

Figure 1 is a schematic diagram of a circuit for reading out signals from one pixel of a MOS image sensor, Figure 2 is a diagram showing the configuration of a preamplifier, and Figure 3 is a diagram for reading out signals when the open-loop characteristics of the preamplifier are restricted. An example of the frequency characteristics shown by the transfer function of the circuit, FIG. 4 is a frequency characteristic of a circuit according to an embodiment of the present invention, FIG. 5 is a circuit diagram of an embodiment of the present invention, and FIG. FIG. 2 is a diagram showing an example of a correction circuit according to the present invention.

Claims (1)

[Claims] 1. Opening/closing signals of a plurality of light receiving elements arranged in the horizontal and vertical directions are controlled by pulses.
In a solid-state imaging device that sequentially reads out a signal to a signal processing circuit including a preamplifier via a MOS type switching element, the preamplifier is provided with a loop that feeds back the output signal to the input side via an impedance element, The loop frequency characteristic has a first cutoff frequency within the video band.
It has a low-pass characteristic with f ₁ and a second frequency determined by the capacitance C ₀ present at the input end of the preamplifier and the input impedance R _io of the preamplifier.
The cutoff frequency f ₂ of the feedback loop exists within the video band, and the cutoff frequency f ₃ of the feedback loop also exists within the video band, and the three cutoff frequencies f ₁ , f ₂ , f ₃ A solid-state imaging device characterized in that the relationship K=(f ₁ f ₂ /f ₃ +f ₁ ) ² −4f ₁ f ₂ 0. 2. Fourth cutoff frequency f ₄ calculated from the three cutoff frequencies f ₁ , f ₂ , f ₃ 2. The solid-state imaging device according to claim 1, wherein the frequencies f ₁ , f ₂ , and f _{3 are set so that the frequencies f 1 , f 2 , and f 3} are higher than the video band. 3. Fifth cutoff frequency f ₅ calculated from the three cutoff frequencies f ₁ , f ₂ , f ₃ 2. The solid-state imaging device according to claim 1, further comprising a correction circuit for amplifying signal components in a higher frequency region by a circuit after a preamplifier.