Silicon Retina

ELL832-IEC Lab-II

Submitted By,
Nishtha Phour (2023EEN2763)

Submitted to
Prof. Jayadeva

Prof. Kaushik Saha

Department of Electrical Engineering,

Indian Institute of Technology Delhi

May 3, 2024
Contents

1 Introduction 3
1.1 Background of CMOS image sensor . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Imaging System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 CCD Image sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.2 CMOS Image sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Motivation and Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Photo-diode 7
2.1 Principle of Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Modelling of photodiode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Capacitor Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.1 Through diode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3.2 Through Real Capacitor . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4 Final Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Photo-receptor 12
3.1 Principle of operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Modelling of Photo-receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.1 Initial Photo-receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.2 Proposed photo-receptor . . . . . . . . . . . . . . . . . . . . . . . . . 13

4 Laminar Monopolar Circuit 17

4.1 Initial LMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.2 Final LMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

5 Comparator 20
5.1 Utilization of Comparator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

1
 2

5.2 Initial Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

5.3 Final Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

6 Array 32x32 23

7 Results 25
7.1 Photo-diode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
7.2 Photo receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
7.3 LMC circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7.4 Pulse detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
7.5 Array 32x32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
7.5.1 Analog output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
7.5.2 Digital output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
7.6 Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7.6.1 Case 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
7.6.2 Case 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

8 Conclusion 35
Chapter 1

Introduction

Semiconductors, which are widely used worldwide, are essential in bringing modern soci-
ety into the age of electronics. Moore’s law has caused wafer sizes to increase over the
past 50 years to maintain cost competitiveness. The purpose of this project is to explore
the silicon retina. The main task of this project is to focus on one of the functions of
the human retina, which is detecting motion. In other words, this project is to build a
circuit to detect movement, i.e., optical motion sensor, which uses different photo-diodes
to receive and transmit the light intensity.

1.1 Background of CMOS image sensor

The semiconductors used to create digital cameras are called CMOS Image Sensors (CIS).
After detecting information in the form of light or any other electromagnetic radiation,
they turn that information into an image. CMOS image sensors are comprised of inte-
grated circuits that detect data and transform it into a corresponding voltage or current,
which is then transformed into digital information.

Weckler introduced the charge integration operation on a photon-sensing p-n junction

in 1967, and this was considered the foundational idea of a CMOS image sensor [1]. The
CMOS image sensors continue to employ this charge integration technology. Weckler
and Dyck briefly proposed the first passive pixel image sensor in 1968 [2]. The CMOS
active pixel image sensor, which was first described by Peter Noble in 1968, served as
the model for contemporary CMOS image sensors [3]. However, the rapid development
of active pixel image sensors had to wait until the 1990s, when the constraints of CMOS
technology were overcome [4].

3
1.2. IMAGING SYSTEM 4

These days, CMOS image sensors have surpassed CCDs in the majority of fields. As an
integrated technology, a CMOS image sensor provides several functions, quick readout
times, low power consumption, affordability, and several improved qualities. Despite
the superior image capability of CCDs, good performance transistors are challenging to
install using CCD fabrication procedures since these processes are devoted to producing
photo sensor components rather than transistors. Consequently, integrating circuitry
blocks on a CCD chip is quite difficult. But if CMOS imagers can provide imaging
performance comparable to that of the sensor, it is even feasible to integrate all the
necessary functional blocks with the sensor—a camera-on-a-chip—which might greatly
enhance sensor performance and reduce costs. JPL successfully exhibited the first CMOS
image sensor in 1995 [5]. It had circuits for fixed pattern noise reduction, correlated double
sampling, control, and on-chip clocking.
The most advanced CMOS image sensor technology is represented by active pixel sensors,
which are integrated circuits made up of a signal processing unit and a range of pixels.
The next section goes into more depth about them. This article discusses crucial CMOS
image sensor characteristic parameters as well as test procedures for calculating and
assessing them. A CMOS image sensor should ideally have the following characteristics:
no optical cross talk, high frame rate, low noise, linear response, and 100% quantum
efficiency. [6]

1.2 Imaging System

The two types of silicon-based imaging systems are complementary metal oxide semicon-
ductor (CMOS) and charge-coupled device (CCD) image sensors. Every imaging system
has particular advantages and disadvantages of its own according to their application.
Over the past few decades, CMOS image sensors have gained popularity due to their low
power consumption, low cost, and system-on-chip capabilities. They have also seen a
consistent improvement in imaging performance.

1.2.1 CCD Image sensor

The underlying working concept of CCD and CMOS image sensors is the same: they
both use the photodiode, a photosensitive element, as their main method of capturing
photons. The technique used to transfer the photonic data to the readout circuit is where
the primary distinctions exist. In CCDs, incoming photons are converted into electrons
1.2. IMAGING SYSTEM 5

by the photodiodes and then successively shifted from one pixel to the next using MOS
capacitors. An amplifier outside the pixel array transforms the electrons into a voltage
that increases linearly with light intensity. The analog signal is subsequently converted
to digital form using an off-chip high-resolution analog-to-digital converter (ADC).
CCD imagers can achieve low noise and full signal maintenance without distortion by
using a specific channel design . Excellent uniformity is promised by the use of a single
electronic conversion chain, which includes the amplifier and ADC, and the passive prin-
ciple of charge transfer. Moreover, a CCD image sensor outperforms a CMOS rival in
terms of homogeneity and fill factor. These benefits have kept CCD sensors at the top of
the high-end imaging market until recently, particularly for applications in science and
medicine.
Nevertheless, the CCD sensors’ maximum speed is limited by the sequential transfer and
readout technique of the electrons in each pixel. A high voltage is applied to the pixel’s
control clock in order to increase the efficiency of charge transfer; this procedure uses
a significant amount of electricity. More importantly, because the CCD image sensor is
incompatible with the CMOS process, system integration is not possible because other
electronics, like an ADC, programmable amplifier, and digital process circuitry, must be
manufactured on a different chip using the CMOS process. It follows that this raises both
the system’s cost and area. The unique charge transfer mechanism increases the mass
production risk that one damaged pixel can ruin the signal capture of the other pixels;
thus the yield of CCD sensor is lower, resulting in the higher cost of CCD sensor.

1.2.2 CMOS Image sensor

About the same time as the CCD’s inception, CMOS sensors were developed. However,
when it was first released into the market, its application was limited due to its initial
low noise performance. Later technological advancements, particularly the creation of the
pinned photodiode (PPD), made the CMOS sensor’s sensitivity and noise performance
on par with that of the CCD sensor. With the use of several approaches, such as the
logarithmic pixel and dynamic range extension techniques, a CMOS image sensor can
achieve an even wider dynamic range than a CCD image sensor.
Most notably, a CMOS image sensor can be implemented on a single chip, allowing for the
integration of pixel arrays with ADCs, peripheral circuits, and digital signal processors
to significantly lower the overall imaging system’s size and cost.
In contrast to the CCD sensor, the CMOS image sensor’s active pixel makes use of
1.3. MOTIVATION AND OBJECTIVE 6

an in-pixel amplifier, which boosts both the randomaccess mechanism and readout speed.
These days, CMOS image sensors are comparable to their CCD cousin in terms of image
quality. They also provide the advantages of low cost, low power consumption, and
system integration.

1.3 Motivation and Objective

The primary objectives of this report can be summarized as follows:
(i) based on the thorough theoretical analysis, to build a behavioral model of the CMOS
image sensor on linearity;
(ii) to verify the behavioral model through cadence virtuoso
(iii) briefly understanding how an pixel works

1.4 Outline
This thesis is comprised of eight chapters. The organization of the thesis is arranged as
follows.
Chapter 2 gives a literature review. Based on a review of the fundamental principle
of photodetection, an overview of the CMOS image sensors is given. Different pixel
structures aimed for different applications are introduced.
Chapter 3 presents a theoretical analysis 3-T photoreceptor (or pixel structure)
In Chapter 4, based on the analysis of Chapter 3, a LMC circuit is made which basically
mimic the layered structure and connectivity of neural circuits in the biological retina.
In chapter 5, a comparator circuit is made to detect every change or movement.
In Chapter 6, a 4x4 array is made using both photoreceptor, LMC and comparator.
Chapter 7, a 32x32 array is tried.
Chapter 8 discusses about results.
Chapter 9 tells where I concluded my report
Chapter 2

Photo-diode

This chapter gives a brief introduction to the Photo-diode and its modelling.

2.1 Principle of Operation

A photodiode is a semiconductor device that converts light energy into electrical cur-
rent. When photons of light strike the photodiode’s semiconductor material, they create
electron-hole pairs within the material. This process generates a flow of current known
as photocurrent, which is directly proportional to the intensity of the incident light. By
applying a reverse bias voltage across the photodiode, the flow of photocurrent can be
controlled and measured. Photodiodes are commonly used in various applications such
as light detection, optical communication, and sensing due to their fast response times
and high sensitivity to light. They are essential components in devices ranging from sim-
ple light sensors to advanced optical communication systems, enabling the detection and
manipulation of light signals with high precision and efficiency.

2.2 Modelling of photodiode

In our report, we model the behavior of a photodiode using an equivalent circuit com-
posed of three key components: a current source, a resistor, and a capacitor.
Firstly, the current source symbolizes the photocurrent generated by the photodiode when
exposed to light. This current is a result of photons exciting electrons within the semi-
conductor material, leading to the creation of electron-hole pairs. The magnitude of this
photocurrent is directly proportional to the intensity of incident light.
Secondly, we incorporate a resistor into the model to represent the dark current or leakage

7
2.3. CAPACITOR MODELLING 8

current of the photodiode. This current flows through the device even in the absence of
light. Although typically minimal, the dark current contributes to the overall behavior
of the photodiode.
Lastly, we include a capacitor in the circuit to account for the junction capacitance of
the photodiode. This capacitance arises due to the formation of a depletion region at the
semiconductor interface. It influences the photodiode’s frequency response and transient
behavior, particularly affecting its response time to changes in incident light.
By employing this simplified equivalent circuit, we aim to understand the photodiode’s
operation, including its response to varying light levels and its interaction with external
electronic circuits. This model serves as a valuable tool for our research endeavors, facili-
tating the design and optimization of photodiode-based systems for applications such as
light sensing, optical communication, and imaging.

2.3 Capacitor Modelling

Capacitor can be modelled as a diode in reverse bias or an actual capacitor

2.3.1 Through diode

Using a diode in reverse bias as a capacitor in the equivalent circuit of a photodiode
provides an alternative approach to modeling its capacitance. When a diode is reverse
biased, it exhibits a capacitance due to the depletion region formed between the p-type
and n-type semiconductor regions. This capacitance behaves similarly to the junction
capacitance of a photodiode and can be utilized in the equivalent circuit.
By incorporating the reverse-biased diode as a capacitor in the model, we effectively
represent the capacitance associated with the photodiode’s depletion region. This ca-
pacitance affects the photodiode’s frequency response and transient behavior, influencing
factors such as its response time to changes in incident light and its ability to filter high-
frequency components in the electrical signal.
This approach offers simplicity and convenience in modeling the photodiode’s behavior,
as it leverages the inherent characteristics of the reverse-biased diode. However, it’s es-
sential to consider the specific capacitance-voltage characteristics of the diode and ensure
its suitability for the desired application. Additionally, the accuracy of the model may
vary depending on factors such as temperature, bias voltage, and frequency of operation.
Nonetheless, incorporating a reverse-biased diode as a capacitor provides a practical and
2.3. CAPACITOR MODELLING 9

effective means of representing the capacitance in the photodiode’s equivalent circuit.

One notable advantage of employing a reverse-biased diode as a capacitor is its inherent
simplicity and integration within the photodiode structure. By leveraging the depletion
region capacitance of the diode, we eliminate the need for an additional discrete capacitor
component in the circuit. This simplifies the overall circuit design, reduces component
count, and potentially lowers manufacturing costs. Additionally, the diode’s capacitance
characteristics are inherently tied to its reverse bias voltage, allowing for direct control
and adjustment of capacitance without the need for additional circuitry.
However, this approach also presents certain limitations and trade-offs compared to using
a real capacitor. One key drawback is the limited tunability and precision of the diode’s
capacitance compared to discrete capacitor components. The capacitance of the reverse-
biased diode is primarily determined by its intrinsic properties and bias voltage, which
may limit fine-tuning capabilities and precise control over capacitance values. Addition-
ally, the capacitance-voltage characteristics of the diode may exhibit nonlinear behavior,
introducing complexities in circuit analysis and design.

2.3.2 Through Real Capacitor

In our thesis investigation, we explore the use of a real capacitor in the modeling of photo-
diodes, contrasting it with alternative approaches such as utilizing a reverse-biased diode
as a capacitor. Incorporating a real capacitor into the equivalent circuit of a photodiode
offers several distinct advantages.
Firstly, using a real capacitor provides precise control over capacitance values, allowing
for fine-tuning of the photodiode’s frequency response and transient behavior. Unlike the
capacitance of a reverse-biased diode, which is primarily determined by intrinsic proper-
ties and bias voltage, a discrete capacitor offers greater flexibility in adjusting capacitance
to meet specific design requirements.
Moreover, a real capacitor exhibits linear capacitance-voltage characteristics, simplifying
circuit analysis and design compared to the nonlinear behavior often associated with
reverse-biased diodes. This characteristic enables more accurate modeling of the photo-
diode’s electrical behavior and facilitates the optimization of circuit performance.
Additionally, discrete capacitors typically offer superior stability and reliability compared
to the capacitance inherent in reverse-biased diodes. They are less susceptible to varia-
tions in temperature, bias voltage, and frequency, providing consistent performance across
a wide range of operating conditions.
2.3. CAPACITOR MODELLING 10

However, it’s essential to acknowledge that incorporating a real capacitor into the pho-
todiode model may introduce complexity and increase circuit size and cost compared to
alternative approaches. Discrete capacitor components require additional board space
and may incur higher manufacturing expenses, particularly in high-volume production
scenarios.
In conclusion, while using a real capacitor in photodiode modeling offers advantages in
terms of precision, linearity, and stability, it also presents trade-offs in complexity and
cost. Understanding these considerations is crucial for informed decision-making in the
design and implementation of photodiode-based systems.
2.4. FINAL MODELLING 11

Figure 2.1: Photodiode Modeling

2.4 Final Modelling

Figure 2.1 represent photodiode modelling
Chapter 3

Photo-receptor

Photoreceptors are specialized cells found in the retina of the eye that play a fundamental
role in the visual process. Understanding the principle of operation of photoreceptors is
essential for comprehending the mechanisms underlying vision. This section aims to
elucidate the fundamental principles governing the operation of photoreceptors, focusing
on their response to light stimuli and the generation of neural signals.

3.1 Principle of operation

Operation of Photoreceptor Cells: The operation of photoreceptor cells can be summa-
rized in several key steps:
1. Absorption of Photons: Photoreceptor cells contain specialized light-sensitive pigments
known as photopigments, which are embedded in the membrane discs of their outer seg-
ments. When photons of light enter the eye and strike the photoreceptor cells, they are
absorbed by these photopigments, causing a conformational change in their molecular
structure.
2. Generation of Photocurrent: The conformational change in the photopigments trig-
gers a cascade of biochemical reactions within the photoreceptor cell, ultimately leading
to the activation of cyclic nucleotide-gated (CNG) ion channels in the cell membrane.
Opening of these channels allows the influx of positively charged ions, primarily sodium
and calcium, into the photoreceptor cell, resulting in the generation of a depolarizing
photocurrent.
3. Hyperpolarization: In response to the depolarizing photocurrent, the photoreceptor
cell undergoes a transient depolarization followed by a sustained hyperpolarization. This
hyperpolarization occurs due to the closure of CNG channels and the activation of potas-

12
3.2. MODELLING OF PHOTO-RECEPTOR 13

sium channels, leading to an efflux of potassium ions from the cell. As a result, the
photoreceptor cell’s membrane potential becomes more negative than its resting state.
4. Release of Neurotransmitter: The hyperpolarization of the photoreceptor cell decreases
the release of neurotransmitter molecules, primarily glutamate, at its synaptic terminals.
This modulation of neurotransmitter release alters the signaling to downstream neurons
in the retina, including bipolar cells and horizontal cells, initiating the processing of vi-
sual information.
The operation of photoreceptor cells involves the conversion of light stimuli into electrical
signals through a series of molecular and cellular processes. This electrical signaling is
fundamental to the transmission of visual information within the retina and forms the
basis of visual perception. Further research into the electrical properties of photoreceptor
cells holds promise for advancing our understanding of vision and developing innovative
approaches for restoring sight in individuals with visual impairments.

3.2 Modelling of Photo-receptor

In this section, the photoreceptor circuit that I took from the reference paper and how I
modified it.

3.2.1 Initial Photo-receptor

The first processing stage is based on the active logarithmic photoreceptor circuit that
transduces the photocurrent into a voltage signal. The output codes the logarithm if is in
subthreshold. The change is linearly proportional to the relative change in the intensity.
Feedback from the amplifier speeds up the response by creating a virtual ground at the
photodiode and allows the photoreceptor to drive the next stage even at low intensities.
Fig 3.1 represent initial photoreceptor

3.2.2 Proposed photo-receptor

In the initial photoreceptor, M3 is used only to bias the circuit. I have used a row selector
just like SRAM kind of technology. Now, I want minimal Vds across M4 to pass the signal
without any distortion.
Fig 3.2 represent final photoreceptor
3.2. MODELLING OF PHOTO-RECEPTOR 14

Figure 3.1: Initial Photo receptor

3.2. MODELLING OF PHOTO-RECEPTOR 15

Figure 3.2: Proposed Photo receptor

3.2. MODELLING OF PHOTO-RECEPTOR 16
Chapter 4

Laminar Monopolar Circuit

Laminar monopolar circuits (LMCs) are essential components in neuroscience research,

particularly in electrophysiology experiments aimed at studying the electrical activity of
the brain. These circuits play a crucial role in recording neural signals from multiple
depths or layers within brain tissue, providing insights into the functional organization
and information-processing mechanisms of the brain. This report delves into the electrical
principle of operation of LMCs and their significance in neuroscientific investigations.

4.1 Initial LMC

The second processing stage is functionally similar to the laminar monopolar cell (LMC)
in the fly visual system. This circuit removes the DC of the photoreceptor signal and
amplifies transient changes. It outputs an amplified and inverted version of the photore-
ceptor output . The circuit consists of an inverting differentiator with an adaptive element
between the output and the input. The transient gain in this circuit is set by the ratio of
capacitors(C1/C2 = 26.6). The DC level is set bythe current in the bias transistor . The
corner frequency of the amplifier is set by the capacitances and the conductance of M6.
This conductance is set by a source follower circuit biased by Vresh. The output of LMC
are then accessed by a comparator which detects the changes in input(Chapter 4). Am-
plifying and high-pass filtering the photoreceptor output signal prevents DC mismatch in
the photoreceptor output from affecting the motion computation. This high-pass filtering
is used as we want to detect every smallest change without troubling the eye.
Fig 4.1 represent initial LMC circut.

17
4.1. INITIAL LMC 18

Figure 4.1: Initial LMC

4.2. FINAL LMC 19

Figure 4.2: proposed LMC

4.2 Final LMC

After having a photoreceptor we will have a pulse having some rise time and fall time
which depends on the frame rate. The output of LMC should give a positive impulse
pulse with a width of rising edge and at the negative edge, there will be an impulse in
the negative direction with respect to DC. So for this, I will need fast switching. So I
modified it and used a CMOS inverter with a high pass filter to amplify the transient
changes in output of photoreceptor I have used again an invertor to amplify the output
of LMC.
Fig 4.2 represent proposed LMC
Chapter 5

Comparator

Image sensors are vital components in digital imaging devices, converting light signals into
digital data for processing and storage. Comparators, electronic circuits that compare
two analog voltages and produce a digital output based on their relative magnitudes,
play a significant role in the functionality of image sensors. This report examines the
various ways in which comparators are utilized in image sensors and their importance in
achieving accurate and efficient image capture and processing.

5.1 Utilization of Comparator

1. Analog-to-Digital Conversion: Image sensors often employ pixel-level analog-to-digital
conversion to convert the analog voltage signals generated by photodetectors into digital
data. Comparators are integral to this process, as they compare the analog pixel voltages
with a reference voltage or threshold.
2. Thresholding: Comparators are also utilized in image sensors for thresholding oper-
ations, which enhance contrast or detect specific features in captured images. By com-
paring pixel voltages with predetermined threshold levels, comparators classify pixels as
"bright" or "active" if their voltages exceed the thresholds, and "dark" or "inactive" oth-
erwise. This thresholding process aids in segmenting objects from the background and
simplifying subsequent image processing tasks.
3. Signal Conditioning: Image sensors often require dynamic adjustment of gain or
amplification to optimize performance under varying lighting conditions. Comparators
facilitate this by comparing pixel voltages with reference levels and dynamically adjusting
signal gain. This ensures that captured images maintain adequate contrast and detail
across a wide range of brightness levels, contributing to high-quality imaging capabilities.

20
5.2. INITIAL DESIGN 21

Figure 5.1: Initial ADC

5.2 Initial Design

So initially, a comparator from AHDL library is taken to detect the impulses at the out-
put of the LMC circuit. One of the comparators detects the positive rise and gives output
1 at the rising edge and 0 at other conditions, and the second one detects the negative
edge. So I have used an XOR gate to detect every (i.e. rising and falling) edge.
Fig 5.1 represent initial ADC

5.3 Final Design

The issue in earlier detection (Section 5.2) is, generally an image sensor should store the
change and release it when that input/change is removed. Hence to store this change it
will need a storage element, i.e. JK flip flop. It should work in such a way that at positive
edge output of detected pulse should go to 1 (SET) and then when input is removed it
should go back to 0 (RESET).
Hence our output will follow our input, which we will further discuss in Section 8.3.
Fig 5.2 represent final ADC
5.3. FINAL DESIGN 22

Figure 5.2: Final ADC

Chapter 6

Array 32x32

A 32x32 array of image sensors represents a grid of 1024 individual image sensors ar-
ranged in a 32x32 matrix configuration. Each image sensor within the array functions
as a discrete photodetector, capable of converting incident light into electrical signals for
subsequent processing to LMC and then conversion of analog output from LMC to Dig-
ital output. The use of an array configuration enables simultaneous capture of multiple
images or increased spatial resolution compared to a single sensor.
Following is the modeling of 32x32 array of the image sensor:-

23
 24

Figure 6.1: Circuit Modelling

Chapter 7

Results

In this chapter, we’ll look into the results that we have got in previous chapters.

7.1 Photo-diode
The output of photodiode should be in such a way that it should provide current which
will be converted from current to voltage through photo receptor circuit. Photo-diode
will work in reverse bias. There will be a current which will flow to show that there
is an input applied or there is change in intensity. A resistor to allow the transistor to
discharge and a capacitor to model the junction caps of the photodiode.
Figure 7.1 represents photodiode output

25
7.2. PHOTO RECEPTOR 26

Figure 7.1: Photodiode output

Figure 7.2: Photoreceptor output

7.2 Photo receptor

It should convert current from photodiode to voltage and whenever the row selector is
on it will be available at output.
Figure 7.2 represents photoreceptor output
7.3. LMC CIRCUIT 27

Figure 7.3: LMC output

7.3 LMC circuit

An LMC circuit is used for amplifying the transient changes, which is then amplified by
the sense amplifier.
Figure 7.3 represents LMC circuit’s output where Vph is the output of the photoreceptor
circuit, Vlmc is the output of LMC circuit and vout is output of sense amplifier.
7.4. PULSE DETECTOR 28

Figure 7.4: final output

7.4 Pulse detector

The output of the pulse detector should be one when the input is given and when it is
removed it should reset to zero again.
Figure 7.4 represents final output in which it flowsfrom photodiode (I5) -> photoreceptor
(Vph) -> LMC circuit (Vlmc) -> sense amplifier (vout) -> pulse detector (fout).
7.5. ARRAY 32X32 29

Figure 7.5: Output of 32x32 part1

7.5 Array 32x32

There will be 2 types of input one is analog and other is actual image.

7.5.1 Analog output

Input is given by current having on-time of 5us and pulse period of 15us. This is done
to remove the clock feed through, whereas my row is always on for every 15us. I have
removed some of my clock feed through problems as initially my current and row getting
on was coinciding hence charging and discharging of Cgs Cds will be there. So I put my
input in such a way that both will not coincide.
My output is when every block has an input.
Figure 7.5 and 7.6 represents analog output of 32x32
7.5. ARRAY 32X32 30

Figure 7.6: Output of 32x32 part2

7.5.2 Digital output

I further remove my clock feed through the pulse detector as it is providing output 1 when
my output exceeds a particular level which will help us to negate the clock feedthrough.
In addition to this, my clock of JK flop is applied in such a way that it only senses the
positive edge detector when it is one which sets the output to one and then on next clock
our clock edges it senses the negative edge through K and reset it again to 0.
Figure 7.7 and 7.8 represents digital output of 32x32
7.5. ARRAY 32X32 31

Figure 7.7: Digital output of 32x32 part1

Figure 7.8: Digital output of 32x32 part2

7.6. IMAGE 32

7.6 Image
By giving my photoreceptor input, I have simulated the circuit. Then I have exported
my csv file from the waveform. From that csv file, with the help of people working on
image sensor, I am successful in generating a image. Input is present that means output
is white whereas if input is not present then that means light is not present, i.e. absence
of light. So it will represent black colour.

7.6.1 Case 1
In this case, I have given input current to alternative rows and columns.
Figure 7.9 is output image for given input

7.6.2 Case 2
In this case, I have given input current to alternative columns.
Figure 7.10 is output image for given input
7.6. IMAGE 33

Figure 7.9: Output Image

7.6. IMAGE 34

Figure 7.10: Output Image

Chapter 8

Conclusion

In wrapping up our exploration of image sensors, we’ve delved into a realm where pho-
todiodes, photoreceptors, and intricate signal processing, like the Local Maxima Count
(LMC) algorithm, converge to capture and interpret the visual world around us. Our
journey has been one of discovery, uncovering the nuanced interplay between these com-
ponents and their pivotal roles in converting light into meaningful electronic signals.
As we reflect on our findings, it’s clear that image sensors are not mere technological
artifacts; they’re the gatekeepers of vision in our digital age. From the photodiodes at
the heart of each pixel, tirelessly converting photons into electrons, to the photoreceptors
meticulously capturing the nuances of light and shadow, every component plays a crucial
part in the image acquisition process.
But it doesn’t stop there. The journey of light through the image sensor is only the
beginning. With the aid of sophisticated algorithms like the LMC, these raw electronic
signals are transformed into actionable data, enabling tasks as diverse as face recognition,
gesture detection, and pulse monitoring. It’s through this fusion of hardware and soft-
ware that image sensors transcend their physical form, becoming catalysts for innovation
and progress.
Yet, our exploration has also revealed challenges on the horizon. As we push the bound-
aries of sensor technology, issues such as pixel scaling limitations and thermal management
loom large, reminding us of the need for continued research and innovation. But with
each challenge comes opportunity, and the potential to unlock new frontiers in fields as
diverse as healthcare, transportation, and entertainment.
In conclusion, image sensors stand as testament to the ingenuity of human creativity,
bridging the gap between the analog world of light and the digital realm of data. As we
look to the future, let us embrace the possibilities that lie ahead, leveraging the power

35
 36

of image sensors to illuminate our world in ways we’ve yet to imagine. After all, in the
dance of photons and electrons, the possibilities are endless, and the journey has only
just begun.
 37

References
[1] G. P. Weckler, "Operation of p-n junction photodetectors in a photon flux integrating
mode," IEEE Journal of Solid-State Circuits, vol. 2, pp. 65-73, 1967.
[2] G. Weckler and R. Dyck, "Integrated arrays of silicon photodetectors for image sens-
ing," IEEE Trans. Electron Devices, Vols. ED-15, pp. 196-201, 1968.
[3] P. Noble, "Self-scanned silicon image detector arrays," IEEE Trans. Electron Devices,
Vols. ED-15, pp. 202-209, 1968
[4] D. Renshaw, P. B. Denyer, G. Wang and M. Lu, "ASIC image sensors," IEEE Inter-
national Symposium on Circuits and Systems, vol. 4, pp. 3038-3041, 1990
[5] R. Nixon, "128x128 CMOS Photodiode-Type Active Pixel Sensor with On-Chip Tim-
ing, Control and Signal Chain Electronics," Proceeding of SPIE, vol. 2415, pp. 117-123,
1995.
[6] A. Krymski, D. V. Blerkom, A. Andersson, N. Block, B. Mansoorian and E. R. Fos-
sum, "“A high speed, 500 frames/s, 1024 × 1024 CMOS active pixel sensor," Symp. VLSI
Circuits, p. 137–138, 1999
[7] Liu, Shih-Chii, MinHao Yang, Andreas Steiner, Rico Möckel, and Tobi Delbruck. "1
kHz 2D Visual Motion Sensor Using 20 X 20 Silicon Retina Optical Sensor and DSP Mi-
crocontroller." IEEE Transactions on Biomedical Circuits and Systems 9, no. 2 (2015):
207-216.
[8] Mead, C. (n.d.). Analog VLSI and Neural Systems. Addison Wesley Publishing Com-
pany.
[9] Shih-Chi Liu, "A neuromorphic aVLSI model of global motion processing in the fly,"
in IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing,
vol. 47, no. 12, pp. 1458-1467, Dec. 2000, doi: 10.1109/82.899640.
[10] T. Delbruck and C. A. Mead, "Adaptive photoreceptor with wide dynamic range,"
Proceedings of IEEE International Symposium on Circuits and Systems - ISCAS ’94,
London, UK, 1994, pp. 339-342 vol.4, doi: 10.1109/ISCAS.1994.409266.
[11] S. Mehta and R. Etienne-Cummings, "A simplified normal optical flow measurement
CMOS camera," in IEEE Transactions on Circuits and Systems I: Regular Papers, vol.
53, no. 6, pp. 1223-1234, June 2006, doi: 10.1109/TCSI.2006.875169.