Design Methodology for a Low-Power Two-Stage CMOS Operational Amplifier for Optical Receiver Applications

ABSTRACT


INTRODUCTION
Fast progress in CMOS technology has allowed for the development of low-power, affordable, compact, and portable electronics.The degree of integration in many modern technologies, such as Wi-Fi, radio frequency identification (RFID), Bluetooth, and ZigBee, has rapidly developed in recent years [1,2].The growth of optical fiber links is a result of the widespread use of data transmission in various communication systems.The fiber-optic transceiver uses a variety of communication algorithms to protect the digital data packets' spectral purity, which originates from a baseband granting consent at very low frequencies when expanding an integrated optical wireline broadband connection.A highly effective wideband amplifier is required for the receiver on the front end of an optical transceiver [3][4][5].Although very highbandwidth systems are typically implemented using GaAs or Si-bipolar technology, CMOS technology provides a more effective and economical substitute for integrating the complete optical front end in a low-power and compact form factor [6].Therefore, in an optical communication circuit system, a low-power wideband CMOS is essential.For the receiver to function well, the Transimpedance Amplifier (TIA) must show sufficient and stable levels over the whole operating range [7].Various receiver applications and needs are taken into account while adjusting the performance characteristics [8].The reduction in the size of CMOS technology poses constraints for radio frequency devices operating at high frequencies, such as problems with power leakage and parasitic effects listed [9,10].MOS amplifiers are commonly used components in a wide range of analog and mixed-signal circuits and systems.The two-stage CMOS opamp depicted in Figure 1 is commonly utilized due to its uncomplicated design and durability.When building an opamp, many electrical properties such as gain bandwidth, slew rate, common-mode range, output swing, and offset must be considered.Op-amps require frequency adjustment for closedloop stability due to their design for negative feedback connections.To attain the necessary level of stability, as measured by phase margin, other performance aspects are sometimes sacrificed.Creating an op-amp that fulfills all requirements requires a well-thought-out compensation strategy and design process [11,12].As we know today, the miniaturization of electronic circuits and using batteries for power supply have become very important.Low-power CMOS amplifiers play a vital role in design.This is a vibrant and thriving area of study, with discoveries and initiatives taking place continuously; however, it also raises certain basic concerns inevitably.
Reduced Supply Voltage: Some of the advanced processes that prevail today, offer the designers the ability to operate at a low voltage level on the transistors, thus reducing the overall power consumption a lot [8].
Current Mirrors: These circuits are useful for managing currents in a manner that re-creates them without much power loss.Cascode stages enhance gain and decrease power dissipation while maintaining single-stage configurations [9].
Subthreshold Operation: When the transistors are operated at the "weak inversion" level, there can be very low power consumption, as indicated by the above figures, but with the potential for low gain and low speed.FinFETs and the development of 3D transistors also result in enhanced control of leakage currents, reducing power consumption.The Limitations and Challenges of CMOS [10,13]: • Trade-offs: In most cases, operations are deliberately downgraded to lower their power consumption profile.A reduction in the supply voltage of a system may result in a lowering of gain and bandwidth.• Leakage Currents: Thus, even in off-states, there is a leakage current in the case of transistors, and it affects the static power.This is something that can never be fully dealt with, in other words, leakage remains a challenge.• Noise: However, in low-power designs, the signal can be easily contaminated by noise and therefore needs careful design and layout [14].• Material Science: Research is going on the discovery of new materials that could offer enhanced performance at low power supplies.• Circuit-Device Co-Design: Choosing optimum circuits and characteristic combinations to achieve the maximum degree of power savings.Dahigaonkar et al. [15] presented a low-voltage, low-power CMOS op-amp for high-frequency applications.In this design, the total harmonic distortion of a low-power op-amp has a high slew rate.Panda et al. [16] improved the slew rate by utilizing an auxiliary circuit along with adjustable biasing circuitry.The CMOS technology was applied to the op-amp with a 2pF load capacitor.A feedback current and variable-gain voltage CMOS op-amp were also developed.Akbari et al. [17] presented an ultra-low-power CMOS operational amplifier based on flicker noise reduction utilizing a weak inversion technique.The op-amp was created using CMOS technology.Ranjan [18] introduced a low-power op-amp built with 180nm CMOS technology that minimizes power usage to 87% with the application of a power-efficient charge steering mechanism.Raghuveer et al. [19] suggested a low-power opamp that lowers the offset voltage for sensing small analog signals.1.8V of supply voltage was used in the creation of this op-amp.Based on the simulation findings, this op-amp can provide 2μV of offset voltage.Mai et al. [20] introduced a lowpower, low-voltage op-amp with a novel chopping method that eliminates switching transistors from the high-gain circuit.Using a 5V power supply.Faheem et al. [21] a low-voltage opamp with current feedback developed using 180nm technology.With a 1.2V supply voltage, this op-amp operates on both the voltage mode and current mode techniques.Parthipan et al. [22].With a power dissipation of 224.8μW, this op-amp has a common-mode rejection ratio of 92dB, unity-gain bandwidth of 1.45GHz, slew rate of 174.2V/μs, and DC gain of 88dB when operating at a 1.5V power supply.Srivastava and Sharma [23] for low-power and high-gain applications.To enhance the op-amp characteristics, including DC gain, slew rate, unity-gain bandwidth, power consumption, area, and settling time, this design makes use of a variety of technologies.Maia et al. [24] have presented a low-power CMOS operational transconductance amplifier developed using 0.13μm technology.To increase the common-mode rejection ratio, paralleled transistors were added to the input transistors during the amplifier's design.This allowed the amplifier to attain an open-loop gain of 87.34dB.Al-Hashimi and Abugharbieh [25].This operational amplifier is composed of two stages: an operational amplifier with a low supply voltage that uses common-mode feedback techniques to remove current sources and boost output voltage swing, and a pulse width modulator that converts output signal information from the voltage domain to the time domain.Dehkordi and colleagues [26] have presented a low-noise, low-power CMOS op-amp, an active feed-forward network built on the current mirror structure carried out in this design.According to the simulation results, this op-amp has an open loop gain of 54.53dB and a unity-gain frequency of 1.7GHz.Surabhi [27] has given a design and study of low-power, high-gain amplifiers for DAC applications.These designs consist of an operational amplifier, differential amplifier, and common source amplifier with various loads, including resistive, active, and current mirror loads.Techniques for lowering the offset voltage and power consumption of a two-stage folded cascade op-amp were given by Stancu et al. [28].For low power consumption and low offset voltage, this op-amp was built with a two-stage folding cascade.
This study aims to design low-power CMOS amplifiers for important use in optical receiving circuits.In addition to that, the low supply voltage in some advanced processes prevalent today allows us to work at a low voltage level on the transistors, which reduces the overall power consumption and thus improves the power consumed compared to previous studies, which reduces the consumed currents in these circuits.

TIA ARCHITECTURES
In an optical communication system, the optical receiver front-end is a key component in determining the dynamic range of the overall network.A photodetector, such as a photodiode (PD), and a high-speed TIA are the standard components of an optical receiver system.The current signal from the photodetector is amplified using a high-speed TIA [11,12].To ensure a consistent voltage signal that consistently oscillates at a suitable level for clock and data recovery, specific applications might require the installation of an additional high-gain broadband post-amplifier (PA) and equalizer, see Figure 2. To increase overall efficiency, the TIA measures the receiver's response and velocity.Therefore, to meet the needs of the optical receiver, it is essential to design a small-sized, low-power TIA employing CMOS technology.Here's a breakdown of TIAs: Here's a breakdown of TIA [12,13]: • Function: Current to Voltage Conversion -This is where TIA's apply a measurable amount of current, such as that generated from a photodetector or photodiode, and then amplify it into a voltage output.
This makes it simple to process and provide an interpretation of the same signal.• Implementation: There is no other choice than to use op-amps when designing this type of circuit.The most fundamental configuration incorporates a highvalue feedback resistor (Rf) whose terminal is connected across the op-amp.The current flows through the input corresponding resistor Rf where the voltage drop across the resistor is amplified by the op-amp.• Gain: As can be seen for a TIA, the gain is directly determined by the feedback resistor (Rf).As derived from Ohm's Law, the output voltage Vout = (Iin * Rf) where Iin is the input current and Rf is the load resistance.Due to the inverting amplifier configuration, negativity is used for the gain, making its representation as (-Rf) standard.

BASIC OP-AMPS EQUATIONS
Both the decrease in mobility caused by the normal field and the limitation in velocity related to MOS devices will be ignored for simplicity.MOSFET equations for drain current and conductivity are based on the parasitic capacitance and the ratio of the width to the length of the conductor between the source and the drain, taking into account the effect of voltage as shown in the Eqs.( 1)-(3) [28]: (1) where,  eff =   −   for nMOS and  eff =   − |  | for PMOS will be used throughout the paper.For bulk MOSFETs at room temperature, strong inversion often necessitates effective voltages (V "eff") above around 200 to 250mV.

SILICON-ON-INSULATOR (SOI)
The conductive substrate is eliminated in SOI chips.Although they have unique pathologies of their own, they can achieve reduced parasitic capacitance and better subthreshold slopes, which can result in lower power and/or higher speed [29].
CMOS technology is being used more and more in ultralow power systems like remote sensors that harvest energy from the environment and implantable medical devices that must operate for years on a tiny battery.At the cost of multiple orders of magnitude decrease in performance [30,31], static CMOS gates functioning in the subthreshold regime can reduce the energy consumed per operation by an order of magnitude.Although SOI technology has long been the focus of research, it only gained commercial traction in 1998 when IBM began using it for PowerPC microprocessors.Although SOI has the potential to have reduced power consumption and higher performance [32], it also has certain odd transistor behavior that makes circuit design more difficult to handle and increases the cost of production.A cross-section of an inverter in an SOI process is shown in Figure 3 [33].
Benefits of SOI in CMOS optical amplifiers [30,34]: • Compactness: System integration on a single chip enables circuits with a reduced size and form factor that are portable.• Lower Cost: From the above lithographic process and possibilities, it can therefore share fabrication processes with CMOS, thus cutting down on the overall manufacturing costs.• Scalability: Integrated optical amplifiers offer versatility in their design for mass production since a CMOS platform is readily scalable.

DYNAMIC CIRCUITS
By substituting a single resistive pull-up for each of the pMOS transistors linked to the inputs, matched circuits lower the input capacitance [30].Rationalized circuits have several disadvantages, including nonzero VOL, static power consumption, sluggish rising transitions, and contention over falling transitions.Dynamic circuits avoid these problems by utilizing a timed pull-up transistor in place of an always-on pMOS.A comparison of (a) static CMOS, (b) pseudo-nMOS, and (c) dynamic inverters is shown in Figure 4.
Since the clock Ǿ is zero during precharge, the clocked PMOS is ON and sets the output Y high.The clock is set to 1 and the clocked pMOS is turned off during evaluation.The output could be low or stay high [35].Dynamic CMOS amplifiers benefit [30]

DOMINO LOGIC
Certain functions, like XOR gates, need inversion by necessity, whereas domino gates are by nature noninverting [29].This problem can be solved in three ways: by employing dual-rail domino logic, by delaying clocks, and by forcing inversions into static logic [31].A traditional static CMOS XOR gate driven by the last domino circuit can be used to construct the required XOR gate at the end of the path in various circuits, including arithmetic logic units (ALUs).But now that it isn't rising monotonically, the XOR output can't trigger additional domino logic directly.To guarantee that the inputs are monotonic during evaluation, a second method involves directly cascading dynamic gates in place of the static CMOS inverter.This involves delaying the clock to the subsequent gates [32].This is frequently implemented in NOR-NOR PLAs and content-addressable memory (CAMs).Figure 5. displays the symbols for the HI-skew inverter, domino AND, and dynamic NAND.

POWER SUPPLY NOISE
Over the course of a big chip, VDD and GND vary.Both are vulnerable to di/dt noise and IR drop-induced power supply noise [34].As seen in Figure 6, IR drops happen between the supply pins and a block, drawing a current I across the resistance R of the power supply grid.When the current fluctuates quickly, di/dt noise develops across the power supply's inductance L. In particular, di/dt noise can be significant for blocks that are idle for multiple cycles [30].There are design and layout techniques to mitigate power supply noise in CMOS amplifiers: • Low Impedance Current Source: We can achieve this by using parallel current mirrors or a wide transistor where the impedance of the bias current source is reduced, and thus the susceptibility to noise injection is reduced.• Power supply decoupling capacitors: Locating CP directly at the power pins of the amplifier filters high-frequency noise to the ground before it enters the circuit.The appropriate capacitor values for your system can be chosen based on the noise frequencies you wish to eliminate.• Careful layout: To reduce induction effects, it is recommended to shorten the power supply traces to the amplifier.It should be remembered that inductance can further exacerbate the noise coupling effect.

Figure 6. Power supply IR drops
A transmission gate is driven by a dynamic inverter, as seen in Figure 7. Assume that the output is floating high due to the precharge of the dynamic gate.Assume further that Y = 0, and the transmission gate is off [29].Charge sharing between X and Y will occur if the transmission gate is activated, disrupting the dynamic output.

NETWORK FOR ON-CHIP POWER DISTRIBUTION
Power and ground wires inside the cells, as well as additional wires joining the cells together, make up the on-chip power distribution network [29].Most cells have internal ground and power buses that are routed via either metal (1) or metal (2).To provide superior electro-migration immunity and reduced resistance, these wires are usually wider than the minimum.For instance, eight metal (1), power/ground buses are used by the cells on the inner front cover.Usually, an abutment connects these wires between neighboring cells.Rows of cells sharing common ground and power connections can be used in both standard cell designs and data paths [30].
These rows can be fastened together with even broader vertical metal wires in a compact, low-power form.An abstract diagram of this strapping is shown in Figure 8 (a).

DESIGN PROPOSED
The design parameters of the op-amp in Figure 1 are obtained from the procedure proposed in Table 1 and meet all specifications, unlike the op-amp designed with different parameters.This study demonstrates that by utilizing this method, the compensation capacitor   using Eq. ( 4) can be significantly reduced compared to other methods.The designer has more latitude to optimize the op-amp in terms of noise and power by picking from a wider variety of options.
Eq. ( 5) can be used to control the current.By adjusting various parameters, the designer can achieve greater control over the op-amp's power and noise characteristics.
Channel length for CMOS to enhance delay of switching and current, Eq. ( 6) refers to channel length.
The current can be managed with the use of Eq. ( 7).The power and noise characteristics of the op-amp can be more precisely controlled by the designer by varying the W/L of several factors.

RESULTS AND DISCUSSION
The circuit was created to satisfy the requirements listed in Tables 2-4, which show that for positive slews, M1 moves into the cut-off region as Vin+ rises.This indicates that M2 receives all of the current from M0.This transistor will enter the deep linear zone if the gate of M2 is subjected to an extremely high voltage.As a result, M2's drain-source voltage will significantly decrease, leaving M5's gate voltage extremely low.This results in an extremely low M5 gate voltage, which indicates a high overdrive voltage for this transistor.As a result, M5 operates within a deep linear range, thereby raising the output voltage to approximately the same level.Table 2 represents the calculation of Mosfet size with drain current and power consumption when the load capacitor CL equals 5pF.The Mosfet size calculation with drain current and power consumption, when the load capacitor CL equals 1 pF, is shown in Table 3.The table indicates that lowering the load capacitance will lower the power consumed, currents, and other significant variables.
Table 4 displays the computation of the Mosfet size with drain current and power consumption when the load capacitor CL = 0.5 pF.According to the table, reducing the load capacitance will result in a decrease in power consumption, currents, and other important factors.
Figure 9 show the response (transfer function frequency response and phase response) for CMOS, the magnitude equals 110 dB when 1 kHz).These figures explain the behavior of the two-stage CMOS op-amp response.The process of calculating the numbers in Tables 2-4 is a simulation process using MATLAB, where the equations of the programmatically designed system were applied and the values were calculated through simulations built on the basis of the equations.

CONCLUSIONS
The improvements in CMOS technology in this study aim to reduce power and noise levels as well as currents in a twostage low-noise CMOS op-amp circuit.Reducing power consumption and currents results in lower losses as well as higher-quality signals in applications such as precision instruments, communication receivers, and professional audio equipment.Achieving low performance often means increasing the rate of packets to be sent.The overall performance should be ideal for low-noise amplifiers, where the noise level, power consumed, and low currents are the most important features from an application perspective.Therefore, improving the performance of operational amplifiers requires a sound design strategy.
Suggestions for future studies include increasing the number of stages, reducing the width to length, and also using improved feedback for the design.

Figure 2 .
Figure 2. TIA is a standard front-end block diagram for an optical receiver

Figure 7 .
Figure 7. On a dynamic gate-driven pass transistor charge

Figure 9 .
Figure 9. Two-stage CMOS op-amps (transfer function frequency response and phase response)

Table 1 .
Parameter setting for the system proposed

Table 2 .
The results system for load capacitor CL equals 5pF

Table 3 .
The results system for load capacitor CL equals 1pF

Table 4 .
The results system for load capacitor CL equals 0.5pF