An Input Amplifier for Body-Channel Communication

Institutionen för systemteknik Department of Electrical Engineering Examensarbete An Input Amplifier for Body-Channel Communication Master thesis performed at the division of Electronics Systems by Md Hasan Maruf LiTH-ISY-EX--12/4634--SE Linköping 2013 2 An Input Amplifier for Body-Channel Communication An Input Amplifier for Body-Channel Communication Examensarbete utfört i Elektroniksystem vid Tekniska högskolan i Linköpings universitet av Md Hasan Maruf [email protected] LiTH-ISY-EX--12/4634--SE Linköping 2013 Handledare: Muhammad Irfan kazim ISY, Linköpings Universitet Examinator: Dr. J Jacob Wikner ISY, Linköpings Universitet 3 An Input Amplifier for Body-Channel Communication Abstract Body-channel communication (BCC) is based on the principle of electrical field data transmission attributable to capacitive coupling through the human body. It is gaining importance now a day in the scenario of human centric communication because it truly offers a natural means of interaction with the human body. Traditionally, near field communication (NFC) considers as a magnetic field coupling based on radio frequency identification (RFID) technology. The RFID technology also limits the definition of NFC and thus reduces the scope of a wide range of applications. In recent years BCC, after its first origin in 1995, regain importance with its valuable application in biomedical systems. Primarily, KAIST and Philips research groups demonstrate BCC in the context of biomedical remote patient health monitoring system. BCC transceiver mainly consists of two parts: one is digital baseband and the other is an analog front end (AFE). In this thesis, an analog front end receiver has presented to support the overall BCC. The receiver (Rx) architecture consists of cascaded preamplifier and Schmitt trigger. When the signals are coming from the human body, they are attenuated around 60 dB and gives weak signals in the range of mV. A high gain preamplifier stage needs to amplify these weak signals and make them as strong signals. The preamplifier single stage needs to cascade for the gain requirement. The single stage preamplifier, which is designed with ST65 nm technology, has an open loop gain of 24.01 dB and close loop gain of 19.43 dB. A flipped voltage follower (FVF) topology is used for designing this preamplifier to support the low supply voltage of 1 V because the topology supports low voltage, low noise and also low power consumption. The input-referred noise is 8.69 nV/sqrt(Hz) and the SNR at the input are 73.26 dB. The Schmitt trigger (comparator with hysteresis) is a bistable positive feedback circuit. It builds around two stage OTA with lead frequency compensation. The DC gain for this OTA is 26.94 dB with 1 V supply voltage. The corner analyzes and eye diagram as a performance matrix for the overall receiver are also included in this thesis work. 4 An Input Amplifier for Body-Channel Communication Acknowledgments Firstly I want to thank Almighty Allah for giving me an opportunity to study at the Linköping University and meet some nice people over here. In my heart I want to thank my teacher and examiner Dr. J Jacob Wikner for his great support and suggestions during the whole thesis work. I would like to thank my supervisor Muhammad Irfan Kazim for his ideas and support during the thesis work. Many thanks to my friend Abdulah Korishe and other BAN group members for the brainstorming sessions and fun times. I will be ignoring my profound duties if I do not mention the unconditional love and support from my mother and other family members during all these years of my masters level studies. 5 An Input Amplifier for Body-Channel Communication Table of Contents 1 Introduction....................................................................................................................................10 1.1 Motivation...............................................................................................................................10 1.2 History....................................................................................................................................10 1.3 System description..................................................................................................................12 1.4 Grouping of entire BCC..........................................................................................................15 1.5 Specifications for the BCC.....................................................................................................16 1.6 Outline....................................................................................................................................17 2 Literature Survey............................................................................................................................18 2.1 Introduction.............................................................................................................................18 2.2 Zimmerman, 1995...................................................................................................................18 2.3 Hachisuka et al., 2003-2005...................................................................................................19 2.4 Shinagawa et al., 2003-2004...................................................................................................20 2.5 Wegmueller et al., 2007..........................................................................................................21 2.6 Song et al., 2006-2007............................................................................................................22 2.7 Lin et al., 2011........................................................................................................................22 2.8 Summary.................................................................................................................................23 3 Body Channel Specification...........................................................................................................25 3.1 Introduction.............................................................................................................................25 3.2 Motivation...............................................................................................................................25 3.3 Applications............................................................................................................................25 3.4 Coupling approaches...............................................................................................................26 3.5 Body model.............................................................................................................................26 3.6 Summary.................................................................................................................................30 4 Analog Receiver Front End Review...............................................................................................31 4.1 Introduction.............................................................................................................................31 4.2 Analog receiver front end topologies......................................................................................31 4.2.1 Resistive feedback amplifier chain.................................................................................31 4.2.2 Capacitive feedback amplifier chain...............................................................................32 4.2.3 Amplifier.........................................................................................................................32 4.3 Receiver chain architecture.....................................................................................................33 4.4 AFE requirement analysis.......................................................................................................34 4.5 Summary.................................................................................................................................35 5 Design of Preamplifier....................................................................................................................36 5.1 Introduction.............................................................................................................................36 5.2 Flipped voltage follower.........................................................................................................36 5.3 Preamplifier performance.......................................................................................................38 5.3.1 Gain analysis of preamplifier..........................................................................................41 5.3.2 Noise analysis of preamplifier........................................................................................42 5.3.3 Results.............................................................................................................................44 5.4 Summary.................................................................................................................................45 6 Design of Schmitt Trigger..............................................................................................................46 6.1 Introduction.............................................................................................................................46 6.2 Compare with comparator......................................................................................................46 6.3 Advantages of CMOS Schmitt trigger....................................................................................46 6.4 Schmitt trigger topologies.......................................................................................................47 6.5 Schmitt trigger performance...................................................................................................48 6.6 Results.....................................................................................................................................50 6.7 Summary.................................................................................................................................51 6 An Input Amplifier for Body-Channel Communication 7 Receiver Results.............................................................................................................................52 7.1 Introduction.............................................................................................................................52 7.2 System level performance of receiver chain...........................................................................52 7.3 Transistor level performance of receiver chain.......................................................................54 7.4 Corner analysis.......................................................................................................................55 7.4.1 Types of combinations....................................................................................................55 7.4.2 Performance....................................................................................................................56 7.4.3 Eye diagram....................................................................................................................57 7.5 Noise analysis.........................................................................................................................59 7.6 Summary.................................................................................................................................61 8 Conclusion and Future Work..........................................................................................................62 8.1 Conclusion..............................................................................................................................62 8.2 Future work.............................................................................................................................63 9 Abbreviation...................................................................................................................................64 10 References....................................................................................................................................69 11 Appendix A...................................................................................................................................72 11.1 VerilogA code........................................................................................................................72 12 Appendix B...................................................................................................................................76 12.1 Characteristics of CMOS065 nm technology.......................................................................76 13 Appendix C...................................................................................................................................78 13.1 Preamplifier design...............................................................................................................78 13.2 Two stage CMOS OTA.........................................................................................................79 13.2.1 OTA gain.......................................................................................................................80 13.2.2 Frequency response.......................................................................................................80 7 An Input Amplifier for Body-Channel Communication Index of Figures Figure 1: Average power consumption vs data rate [1]......................................................................11 Figure 2: BCC top level system description.......................................................................................13 Figure 3: Transceiver chain with the channel outline of the overall system......................................13 Figure 4: System view of the overall BCC.........................................................................................14 Figure 5: Electric fields produced by PAN transmitter T. A small portion of the electric field G reaches the receiver R [3]...................................................................................................................18 Figure 6: Inra-body communication setup [6]....................................................................................20 Figure 7: Inter-body communication setup [6]...................................................................................21 Figure 8: Human body model with capacitive coupling.....................................................................26 Figure 9: Electrical model of the human body...................................................................................27 Figure 10: Arm unit element with the combination of capacitance and resistance............................27 Figure 11: Torso unit element description as a circuit view of the body model.................................28 Figure 12: Groundmesh for body model............................................................................................28 Figure 13: Resistive feedback amplifier chain topology....................................................................32 Figure 14: Capacitive feedback amplifier chain topology..................................................................32 Figure 15: Testbench of the overall AFE by showing the body model and coupling capacitors.......33 Figure 16: Analog receiver front end chain for the BCC...................................................................34 Figure 17: Flipped voltage follower (FVF)........................................................................................36 Figure 18: Open loop analysis of FVF...............................................................................................37 Figure 19: Testbench for preamplifier................................................................................................38 Figure 20: System level results for preamplifier................................................................................39 Figure 21: Biasing circuit for the preamplifier...................................................................................39 Figure 22: The transistor level implementation of preamplifier.........................................................40 Figure 23: Open loop gain analysis for the preamplifier....................................................................41 Figure 24: Close loop gain analysis for the preamplifier...................................................................42 Figure 25: Input referred noise PSD @ 10 MHz by performing the noise analysis with the DC analysis...............................................................................................................................................43 Figure 26: Noise comparison between thermal noise and flicker noise.............................................43 Figure 27: An OTA in a non-inverting configuration effectively forming a positive feedback Schmitt trigger..................................................................................................................................................47 Figure 28: An OTA in an inverting configuration effectively forming a positive feedback Schmitt trigger..................................................................................................................................................47 Figure 29: Schmitt trigger testbench..................................................................................................48 Figure 30: Schmitt trigger schematic (differential two stage amplifier) for the receiver chain.........49 Figure 31: Gain performance for the Schmitt trigger.........................................................................49 Figure 32: Input referred noise for the Schmitt trigger......................................................................50 Figure 33: Receiver chain presenting with two preamplifier stages and one Schmitt trigger stage for the AFE...............................................................................................................................................52 Figure 34: System level results showing transmitter output, receiver input, preamplifier output and receiver output ..................................................................................................................................53 Figure 35: Transistor level results of the different nodes of the receiver chain.................................54 Figure 36: Gain performance of the overall receiver chain in different frequencies..........................54 Figure 37: Corner analysis of the receiver chain................................................................................56 Figure 38: Eye diagram for the receiver chain...................................................................................57 Figure 39: Eye diagram performance @ 100 ns period......................................................................58 Figure 40: Adding 200 uV noise with supply voltage and Tx,Rx outputs.........................................59 Figure 41: Adding 1 mV noise to the supply voltage and showing output signal of Tx,Rx .............60 Figure 42: Input referred noise of the overall receiver @ 10 MHz....................................................61 8 An Input Amplifier for Body-Channel Communication Figure 43: A conventional amplifier with resistive load.....................................................................78 Figure 44: Two stage OTA with biasing circuit and output buffer.....................................................79 9 An Input Amplifier for Body-Channel Communication Index of Tables Table 1: BAN specification draft [1]..................................................................................................12 Table 2: Grouping of subprojects.......................................................................................................15 Table 3: Specifications for the BCC...................................................................................................16 Table 4: Summary of the literature review based on important topics...............................................23 Table 5: Different capacitive value for body model...........................................................................29 Table 6: Input referred noise at different frequencies.........................................................................44 Table 7: Preamplifier results to support the receiver chain................................................................44 Table 8: Schmitt trigger performance analysis...................................................................................50 Table 9: Measurement table for eye diagram.....................................................................................57 10 An Input Amplifier for Body-Channel Communication Chapter 1 1 Introduction 1.1 Motivation Body-channel communication (BCC) is one of the most interesting topics in the recent years. Communication through the body is attractive to the researchers due to its low power operation. BCC is clearly defined by Body Area Network (BAN). BAN is formally defined by IEEE 802.15 as, “A communication standard optimized for low power devices and operation on, in or around the human body (but not limited to humans) to serve a variety of applications including medical, consumer electronics, personal entertainment and other” [1] [16]. Wearable electronic devices like wrist-type computers, earphones, video eyeglasses, and headmounted display need to communicate and maintain natural connection with the human body. BCC explores this natural connection and uses the human body as a transmission medium. This connection, which is near to the human body, can interact with external work via gateway devices. BCC can operate at MHz frequencies without large couplings because the signals transmit between the conductive human tissue of the body coupled transceiver and the floated ground with a capacitive return path [2]. Low impedance capacitive coupling is consuming most of the power. So, it is possible to reduce the power consumption by using a low impedance capacitive coupling. This communication system has different applications such as under the human health monitoring system cardiac monitoring, blood pressure measuring etc. Now business card handshake, door code unlock etc. are getting useful as different applications with this communication system. More examples are the two hearing aid devices that could communicate between each other, or the head-speakers communicating with the music player/telephone, and so on. 1.2 History Communication, through the human body for different applications, is not a new topic for the researchers. In this regard, Zimmerman was the pioneer who gave his idea about personal area networks (PANs) in 1995 and created lots of interest and hope in the short range wireless field. After his proposed idea, different works had done which form the basis for a new kind of communication system as an alternative to the short-range wireless system. The phrase BAN came from the phrase body sensor network (BSN) which was first formally introduced by Professor Guang-Zhong Yang in his book “Body Sensor Networks” in 2006 [1]. The idea is to pass current in safe limits through the body and thus make it possible for two devices to communicate. This concept applies in recent years to the field of biomedical sciences. They are using this concept for ECG/EEG monitoring, heart beat measuring and so on. They are successful in their results and experiments. Then this concept spread out in different fields of studies. Some of the most common uses of BAN technology are stated below [1]: • Body Sensor Network • Heath Monitoring • Wearable Audio • Mobile Device Centric • Personal Video Devices 11 An Input Amplifier for Body-Channel Communication • Remote Control and I/O Devices In 2006, IEEE 802.15 task group 6 promoted a summary about BAN which set the standard of the BAN. The BAN task group had compared BAN with other short-range communication to set the standard. In this way, the task group had given idea about power consumption over data rate of the different short-range communications. They also provided some draft specifications as a BAN requirement for the initial purpose. 1G 100M 10M (Bit/s) 1M 100K 10K 1K Wireless USB IEEE 802.11 a/b/g Bluetooth ZigBee 2 5 10 20 50 100 200 500 1000 (mW) Figure 1: Average power consumption vs data rate [1] Note: Copied from “Body Area Networking (BAN)” written by Erik Karulf ( A survey paper written under guidance of Prof. Raj Jain) Figure 1 is mainly a comparison between BAN and other communication networks with respect to average power and the data rate. Traditional communication networks provide high data rate and also need more power to supply this high data rate. Wireless USB gives high data rate as a cost of higher power consumption. Bluetooth and Zigbee show potential low data rate applications and have high power consumption. Compared to all these networks, BAN gives better results with respect to power and data rate which demonstrate in figure 1. The specifications in table 1 describe the initial requirement of the BAN. It is given by the IEEE 802.15 task group 6. They set the specifications as a starting point of view of BAN. 12 An Input Amplifier for Body-Channel Communication Table 1: BAN specification draft [1] Distance 2 m standard 5 m special use Network Density 2-4 Nets/m2 Network Size Max:100 devices/network Power Consumption ~1 mW/Mbps Startup Time < 100 us or < 10% of Tx slot Latency (end to end) 10 ms Network setup time < 1 sec (Per device setup time excludes network initialization) Future proof Upgradable, scaleable, backwards compatible Very Low, Low and High duty cycle modes Allows device driven degradation of services In this thesis paper, the main goal is to design an integrated circuit for mixed signal BAN chip. The chip can read data from a small memory which is essentially a smart electronic tag. Different companies like Ericsson are supporting research in this field previously. It is also an interesting topic in many institutes and universities such as Korea Advanced Institute of Science and Technology (KAIST), Japanese Docomo-NTT, IBM, Philips and more. 1.3 System description This section 1.3 just gives a brief outline of the BCC and the components that are used in BCC. Data can pass through the human body by following two different approaches. It can pass across human skin, or it can pass through the surface of human skin. Zimmerman [3] proposed that narrowband low frequency signal could be transferred through human body using an electrostatic coupling. His proposed model had some limitations with the capacitive return path as well as the data rate. Another scheme used electromagnetic waves of 10 MHz. It had also problems with the bandwidth of conventional frequency modulation (FM) as well as frequency shift keying (FSK) [4]. In figure 2, the transceiver of BCC that can communicate through the human body is shown. It has two different parts on both sides such as digital baseband (DBB) and analog front end (AFE). 13 An Input Amplifier for Body-Channel Communication Figure 2: BCC top level system description. Figure 3: Transceiver chain with the channel outline of the overall system. In figure 3, the transceiver concept for the overall BCC illustrate as a general point of view. Computer or Smart-phone can be used as an application layer. Data passes from the application layer through an FPGA board to the transmitter. Smart tag (transmitter) can be implemented either on a low power microcontroller, or on an FPGA board for testing purpose. The actual transceiver (TRx) circuitry also connects through Optocouplers in order to isolate the grounds. In the transceiver, there will be needed an interfacing circuit, for example, DAC or directly transmitted the signal to the human body through driver. Then adapt the channel and transmission speed, different types of filters and drivers are needed for the transmitter path. The human channel will distort the signal with the body resistance and especially the capacitive interface will alter the signal significantly. 14 An Input Amplifier for Body-Channel Communication The receiver needs to amplify the weak signal with low-noise amplifiers (LNA) because it is heavily attenuated by the body attenuation. Filters are needed to remove noise. It is possible to remove the noise by limiting the bandwidth and selecting the right band of frequencies. Eventually, it needs to recover the original data from the weak and noisy data. For this purpose, after filters there will be an ADC or comparator or Schmitt trigger to generate the original signal. The human body can be modeled as a distributed RC model. It has a unit element (UE) which can be divided into arm unit element and torso unit element and can be modeled with the RC links. Figure 4: System view of the overall BCC System view of the overall BCC is showen in figure 4. The total system divides into different parts (also shown in figure 3) such as application layer which will implement in software, a digital baseband in FPGA and the analog front end (AFE) in ASIC. The Tx and Rx both have the digital baseband and AFE part. The digital baseband part mainly implements in FPGA which uses Manchester encoding scheme because of its advantages in simpler circuitry for clock and data recovery. For the Tx part, we need to define an output driver for controlling the slew rate, either through a digital charge-pump like interface or some other means to control the rising and falling edges of the signal. Control of the rising and falling edges need with priority in order to control the shape of the pulses on the other side of the capacitor. In addition to this, the signal may have to be filtered. The difficult and challenging part is Rx which has many things that need to be controlled. When the signals are coming from a human body, it is attenuated around 60 dB and become weak. If Tx pulses are at 4 or 5 V, the Rx pulses are in the lower mV region. Weak signals need to be stronger and also need to be free from noise. For this purpose, we need a preamplifier that can amplify the signal to the desired level of expectation. We are using a differential preamplifier that can give us higher gain, as well as low noise. For getting higher gain, we are cascading the preamplifier. At the analog/digital interface, for getting the correct information, we are using Schmitt trigger. It is also helping the system for getting higher gain, as well as low noise. 15 An Input Amplifier for Body-Channel Communication 1.4 Grouping of entire BCC As the whole system is a big task for a master thesis, it is divided into several subprojects. Table 2: Grouping of subprojects SL No. Subprojects 1 Transmitter (Tx) 2 Receiver (Rx) 3 Digital baseband coding and modulation 4 Application layer and channel modeling (APPs) Table 2 decribes the grouping of the subprojects that can make the whole BCC system. The main idea of subproject 1 is to perform a detailed analysis and partial implementation of the transmitter parts (Tx) of the AFE. The transmitter must inject voltage pulses on the capacitive plate connected to the human body. Since the attenuation is high in the body link system and due to the desire of low capacitance (small plates) connected to the body, the transmitter driver circuit needs to be carefully designed. On the secondary plate side, the received voltage level will decrease significantly and will be in the mV region. In order to deal with the noise, one could increase the transmitted voltage levels as well as power consumption. The pulse waveforms and other interfere may require some filtering in order to shape the transmitted pulses accordingly. The filter can ensure that high frequency components are not injected in the body. Further on, the filter should probably be equalized to track the body characteristics, both spectrally and temporally. In subproject 2, the main idea is to do a detailed analysis and partial implementation of the receiver parts (Rx) of the AFE. The receiver is sensitive since the attenuation is high in the body-channel communication. At the transmitter side, several volts can be injected on the capacitance plates. On the secondary plate side, the voltage levels decrease significantly. Through the body, the signals are even more attenuated. Eventually on the receiver, they are in the mV region. The pulse waveforms and other interfering might require some filtering in order to shape the received pulses accordingly. At the analog/digital interface, Schmitt trigger or comparator requires to recover the original information. In this thesis, the main focus is on the receiver (Rx) which means subproject 2. The entire thesis discusses about the implementation of Rx, preamplifier to amplify the weak signal and Schmitt trigger to recover the original signal. Subproject 3 is implementing digital baseband coding and modulation. The idea is to do a detailed analysis of the prerequisites with respect to the communication baseband and implement algorithms to communicate over the human body link. Subproject 4 is describing the application layer and channel coding (APPs). Even though, Manchester encoding has selected as the most promising because of its advantages, it still needs to consider the channel coding and channel adaptation. For this purpose, several channel coding approaches need to observe the effect of the BCC. Several channel coding may be forward error correction, interleaving, resending data etc. 16 An Input Amplifier for Body-Channel Communication 1.5 Specifications for the BCC In table 3, the specifications of the overall BCC are described. This is mainly the target specification of the entire system. All values are valid for all corner conditions. Table 3: Specifications for the BCC Item Variable Min Type Max Process node L Unit Comment nm The ST65 nm RF kit is used for this thesis. 1.1 V It can be 1 or 1.2 V depending on the implementation. 12 MHz This frequency has been identified as the most promising range. 5 mW 65 Supply voltage VDD 0.9 Reference data frequency FREF 0.05 10 Output AFE POUT power 1 Input AFE sensitivity SIN 2 Jitter1 TJ 1.00% Cycle-to-cycle jitter2 with respect to the data clock. Long-term jitter3 TLJ 20.00 % Measured over a number of clock periods and this must be correlated with cycle-tocycle jitter, receiver architecture and clock frequency. 1 mW The potentially output driver is omitted. 125 deg From a reuse point of view it makes sense Power PDIS Consumption Temperature TR range -40 mW Sensitivity, in this case, will be measured as the minimum voltage (rms) that should be detected by the receiver. Total chip A 1 sqmm The application layer and some of FPGA area computer I/F circuits omitted. 1 Jitter is the oscillation of signal with respect to the ideal signal. The difference between two adjacent clock cycle over 1,000 clock cycles in the clock period defines as cycle-to-cycle jitter2. Long-term jitter3, is the difference between output clock from the ideal clock position, over several consecutive cycles. 17 An Input Amplifier for Body-Channel Communication 1.6 Outline The section 1.6 describes the outline of the later chapters and briefly mentions the chapters separately. In chapter 2, the main discussion is about the literature review. How the researchers thought about BCC and how they proceed about their concepts, experiments, results etc. are mentioned in this chapter. In chapter 3, it is mainly a brief discussion about the body model. The electrical body model, the unit element, different resistance and capacitance values are described in chapter 3. In chapter 4, the focus is on the AFE receiver. Analog receiver front end topologies and AFE requirement analysis are discussed in this chapter. Chapter 5 mentions the basic of preamplifier with the proposed architecture for this thesis paper. Different analysis are performed to get the results of the preamplifier. The performance results are shown with graphs. Chapter 6 illustrates the basic of Schmitt trigger with the proposed architecture for this thesis paper. Different analysis are performed to get the results of the Schmitt trigger. The performance results are shown with graphs. In chapter 7, the overall receiver performance and results are described. Different analysis, to prove the receiver is working for this communication system, is presented in this chapter with proper graphs. In chapter 8, the conclusion and future works are reported to make improvement in the future on this communication system. 18 An Input Amplifier for Body-Channel Communication Chapter 2 2 Literature Survey 2.1 Introduction In the growing technology world, BCC creates lots of interest to the researchers. In this reason, different researchers come with different ideas. They investigate the sending and receiving data through the human body using different characterized BCC. Their ideas differ in many ways such as data rate, frequency range, coupler size, coupling method and so on. 2.2 Zimmerman, 1995 The development of BCC is firstly introduced in 1995 by Zimmerman [3] with his research in personal area network (PAN). In his paper, he mentions single-ended approach with electrostatic coupling. In the work, the human body considers as a conductor because of its impedance value compare to electrodes value. The human skin is acting as an insulator because of its large impedance. Transmitter and receiver both can model as capacitor plates [3]. As a result, the transmitter can capacitively couple to the receiver through the human body channel. Figure 5 shows the electric field distribution throughout the human body. Figure 5: Electric fields produced by PAN transmitter T. A small portion of the electric field G reaches the receiver R [3]. Note: Copied from “Personal Area Networks (PAN) : Near-Field Intra-Body Communication” written by Zimmerman, 1995, p. 21 In his paper, Zimmerman mentions human body as a node and conclude with a lumped-circuit electrical model. This model experiments with different capacitance values which mention clearly in the work. This model also electrically describes the near field coupling mechanism around the human body. The measuring displacement current is attenuated by -12 dB when the subject is barefoot, and -28 dB when a ground wire connects to the subject's forehead [3]. So from the paper it is clear that feet are the best location for PAN devices [3] because it has less attenuation. For the PAN, it needs to implement the transceiver. Both the receiver and transmitter connect with the 19 An Input Amplifier for Body-Channel Communication human body with the electrode. Different design issues like cost, size, channel capacity are considered when the electrode is designed. The paper is giving a clear idea about the design issue. On the baseband side, the author experiments with two different modulation techniques. In the linear case, it is on-off keying (OOK), and for the nonlinear case, it is direct sequence spread spectrum (DSSS) techniques. Spread spectrum has the highest received signal than the OOK. The receiving signal strength needs to increase. For that reason, resonator circuit has been used in OOK. Still OOK is inferior to spread spectrum by 60%. But there are some problems with the spread spectrum, such as tracking phase, mysterious negative correlation compares with OOK whereas it is easy to implement. This paper shows a carrier frequency of 333 MHz and a data rate of 2.4 kbps [3]. 2.3 Hachisuka et al., 2003-2005 In the publications [4] and [5], the authors are mainly inspired with the personal area network (PAN) which is motivated them to investigate about the galvanic coupling. In the paper, they mention sine wave as an input signal. Input signal frequency varies from 1 to 40 MHz for the intrabody communication to find the best solution. The signal propagation is not only done with the air but also involved with the human body. The experiment shows that the signal propagation through the human body is superior than through the air up to 30 MHz. 10 MHz is the optimized frequency of the intrabody communication because it has a maximum gain of -26 dB [4]. In the paper, two important points mention in this communication system. One is that surroundings have no effect on differential approach. Another one is that the impedance between the electrode and the human body are independent of metal or alloy. They design a human phantom arm which reduces the uncertainty in the results. The characteristics of human phantom are similar to the human body characteristics between 1 MHz to 10 MHz frequencies. In their experiment, they use small, lightweight, energy-saving, wearable FM transmitters and receivers with the carrier frequency of 10.7 MHz. Here, they monitor two different scenarios. If they use a noise source like the mobile phone, microwave etc., it is same as without using any noise source that means signal can be received correctly. After that, they also experiment with digital data using FSK modulation techniques with different hardware. They get 9.6 kbps data rate without any bit error rate. In the paper [5], they mention two types of electrode model scenarios. Firstly, they experiment with four-electrode model (two transmitters and two receivers), which place on the surface of an arm. Secondly, they experiment with two-electrode model (one transmitter and one receiver) that place on the skin of the human body. They calculate gain for both four-electrode model and two-electrode model for different frequencies (from 1 to 1000 kHz) and distances (from 50 to 300 mm). The calculation results and the measurement results give a satisfacting feeling for the two-electrode model. For the four-electrode model, there is a difference between values when the frequency increases. The two-electrode model has 20 dB higher gain [5] over four-electrode model in the KHz frequency range. The authors prefer the two-electrode model by doing the both experiments. For two-electrode model, one electrode can place on foot, and the other can be at arm. The gain also investigates with different arm positions such as arm up, arm horizontal and arm down. 20 An Input Amplifier for Body-Channel Communication 2.4 Shinagawa et al., 2003-2004 In the composition [6] and [7], the authors mainly discuss a near field sensing transceiver for intrabody communication based on the electro-optic (EO) effect. For the coupling, they use capacitive coupling for the data transmission. An electric field sensor with the effect of electro-optic (EO) and laser light have used to the receiver part of the overall transceiver. The advantage of using electro-optic sensor is that it measures without ground contact and has an extremely high input impedance which is useful to detect unstable electric field. A phantom body model is proposed by the authors for their experiment. The phantom body model uses capacitive coupling for Tx and Rx electrodes. The distance between Tx and Rx electrode is 1 m for this experiment. Their experiment states that this kind of setup for the data transmission can support TCP/IP (10 BASE) half duplex communication at 10 Mbps data rate [6]. In their experiment, they also confirm that the data transmission through the human body can be possible with the wearing cloth of the human without direct contact with the skin. They also experiment with intra-body communication and inter-body communication. So, it is mainly two types of data communication through the human body. For intra-body communication, the test person touches the Tx (right side transceiver) and Rx (left side transceiver) electrode. Then data is transferred through the human body by using these two electrodes. For inter-body communication, the test person may be two person. These two test person shake their hands. One test person touches the Tx electrode and another test person touches the Rx electrode. After that data transfer through the two human bodies. In figure 6, it shows intra-body communication setup where one test person requires. In figure 7, it mentions inter-body communication setup where two test person require for the system setup. Figure 6: Inra-body communication setup [6] Note: Copied from “ A near-field-sensing transceiver for intra-body communication based on the electro-optic effect” written by Shingawa et al., 2003, p. 299 21 An Input Amplifier for Body-Channel Communication Figure 7: Inter-body communication setup [6] Note: Copied from “ A near-field-sensing transceiver for intra-body communication based on the electro-optic effect” written by Shingawa et al., 2003, p. 299 2.5 Wegmueller et al., 2007 In this paper [8], the authors mainly try to model the human body as a communication channel. The authors also investigate that the galvanic coupling is suitable for on-body sensor network in biomedical monitoring system. The signal is differentially transmitting through the transmitter electrodes and also receiving the signal differentially through the receiver electrodes. They use the current as 1 mA peak amplitude. This current directly enters into the human body and creates a potential distribution over there. For supporting this current, they use finite-element (FE) models. The purpose of this current is to detect the impedance distribution with the electrical impedance tomography (EIT). The writers divide the human body into different parts like skin, fat, muscle etc. for better understanding. The dependency of joints in the human body, distance between transmitter and receiver, their size and types of electrodes on attenuation factor are also studied in this paper. They make their clinical trial with 20 subjects where the average age is 47.2 years and so on. The trail has approved by the Swiss National Advisory Commission on Biomedical Ethic (NEK-CNE, Bern, Switzerland) [8]. Depending on the transmitter and receiver combinations, the measurements set into four different terminals like along the arm, through the thorax, along the leg and through the entire body. To support their clinical trial, they also compare their coupling characteristics with different coupling strategies. They state their results like when the distance between transmitter and receiver are increasing, the attenuation is also increasing. But the influences of transmitter sizes, are different impacts on the attenuation whereas in the receiver, it is not much variation in different sizes. The larger the joints in the human body make the attenuation higher, and a decrease of the fat resistance leads to a lower attenuation. Comparing the thorax with the legs and arms, it has a low attenuation effect and also there is no attenuation during the motion of the body. 22 An Input Amplifier for Body-Channel Communication 2.6 Song et al., 2006-2007 Energy efficient wideband signaling analog front end (AFE) receiver has proposed in this article [9]. Wideband is more efficient because of its power consumption and data rate for the on-body networks. In this paper, the AFE divides into four different parts : on-chip symmetric bias circuit, wideband preamplifier, Schmitt trigger and inverting buffer. The on-chip symmetric bias circuit configure with a complementary design of two resistances, which give 50 Ω input impedance for high speed operation and DC biasing with no external bias circuit. The preamplifier uses fully complementary folded cascade topology with low supply voltage and fabricates by 0.18 um standard CMOS technology. It gives 60 dB DC voltage gain, 1.5 mW power consumption at 1 V, high slew rate, 937 MHz unity-gain frequency and so on [9]. Finally, the authors get 10 Mb/s data rate with the sensitivity of -27 dBm and 200 MHz bandwidth for the proposed AFE. In another paper [10], the writers mainly design PHY transceiver, which adopts the direct sequence spread spectrum (DSSS) technique for faster code acquisition and interference rejection. The AFE consists of wideband amplification, digital conversion and three level pulse shaping for the pulse position modulation (PPM) signals. The PHY receiver design by depending on the frequency range of the three level PPM and DSSS where the communication is packet based. This packet based communication consists of four components: a synchronization (SYNC) header, a PHY header, a variable length payload and a cyclic redundancy code (CRC). The proposed digital baseband (DBB) drive with 0.9 V supply voltage with realizing as a hierarchical block gating (HBG) architecture. The transceiver fabricates in 0.18 um standard CMOS technology with a maximum data rate of 10 Mb/s. In the same year, the authors come up with another publication [11] where the transceiver is digital based with wideband signaling. For high speed operation of the human body communication (HBC), a direct-coupled interface (DCI) adopts. The propose transceiver consists of a direct digital transmitter and an all digital clock and data recovery (CDR) circuit. As CDR circuit uses to reduce the power consumption, the authors propose a low-voltage digitally controlled oscillator and a quadratic sampling technique. In this paper, the human body considers as a bandpass filter in singleended approach. The mention digital transceiver fabricates in 0.25 um standard CMOS technology with a data rate of 2 Mb/s at a bit error rate of 1.1x 10 -7. It consumes 0.2 mW power with a supply voltage of 1 V. 2.7 Lin et al., 2011 The authors of the paper [12] develop a biomedical system-on-a-chip (SOC) for intrabody communication (IBC) system. They investigate that the human body is more power efficient than the air because of its lower carrier frequency. They propose an IBC system mainly built up with two units: communication unit and system control unit. The communication unit consists of a receiver, a transmitter and an analog-to-digital converter (ADC). The receiver designs with low power selfmixing methodology and uses on-off keying (OOK) modulation technique for transmission. The system control unit is a microcontroller unit (MCU) which control the signal receiving command and transmission of data. The MCU has four important states to operate its command. In the first state, which is idle state, makes MCU in an idle condition. The second state is convert state. It is mainly turning on the ADC and converting the biomedical signal into a digital format. In the next state data has transmitted through the on-off keying transmitter. Finally, the continuous command measure and transmit the converted data. This transmission method is much more power efficient compared to other short-range communication systems. The propose transceiver implements in 0.18 um CMOS process and needs less area for chip fabrication because it does not need any antenna. 23 An Input Amplifier for Body-Channel Communication 2.8 Summary The above publications prove that the body-channel communication is now creating lots of interests among the researchers with its different advantageous applications. Many publications theoretically measure the values and some publications prove practically their works as well. Most of the authors who study about the electrode size conclude that a larger area in transmitter improves the signal strength. Most of the publications try to prove that less power consume less area. Based on the measurements or values, AFE design for this thesis paper that can use to implement the BCC. This thesis paper tries to improve the basic requirements and also try to support the overall BCC. In this paper, the AFE design with ST65 nm standard CMOS technology. Table 4 mentions the difference among authors' experiments and also differences in parameters that discuss in the literature papers. Table 4: Summary of the literature review based on important topics Zimmerman Hachisuka Shinagawa Model Electrical Not applied Not applied Coupling scheme Capacitive Capacitive Capacitive Freq. range 100 kHz-500 kHz 10 kHz-50 GHz Not reported Modulation OOK FSK Not reported Data rate 2.4 kbps 9.6 kbps 10 Mbps Location on Body Arm, foot, waist, head Arm Hand Electrode size 25x25 mm2 30x30 mm2 70x20 mm2 80x80 mm2 130x40 mm2 900x25 mm2 30x30 mm2 20x20 mm2 Not reported Orientation Vertical Horizontal Single Electrode Rx type Current amplifier followed by a chopper Intrabody FM receiver An Electric field sensor which uses the EO effect and laser light 24 An Input Amplifier for Body-Channel Communication Wegmullar Song Lin Model Electrical Electrical Not reported Coupling scheme Galvanic Capacitive Ag/AgCl Freq. range 1 kHz-10 MHz 1 MHz-200 MHz 50 Mhz-300 MHz Modulation FSK, BPSK PPM OOK Data rate 128,255 kbps 10 Mbps 2 Mbps Location on Body Arm, wrist, leg Not applied Wrist Electrode size 880 mm2 54 cm2 560 cm2 Not applied Not reported Orientation Horizontal Not applied Not reported Rx type Not reported Amplifier followed by Schmitt trigger Self-mixing receiver which consists of LVA, Cascade gain amplifier, STD amplifier, LVM, lowpass filter, Comparator and buffer 25 An Input Amplifier for Body-Channel Communication Chapter 3 3 Body Channel Specification 3.1 Introduction In BCC, the human body acts as a communication channel. The basic principle of this communication describes that a small electric field passes through the human body to propagate a signal between devices that are in the proximity of, or in direct contact with, the human body. The signals are spikes and inject into the human body in two different approaches which are capacitive coupling and galvanic coupling. In this thesis, capacitive coupling is used to establish a body model where the transmitter and receiver are also present. 3.2 Motivation There are many kinds of wireless communication available for different applications. There are also many devices that are using this short-range wireless communication such as NFC, PAN, BAN, BCC etc. This communication system can be divided into two different types. One is near field inductive coupling, and another one is far filed electromagnetic communication. In the short distance communication, inductive coupling is effective for an on-body communication. For example implanted cardiac defibrillators (ICDs), cochlear implants and artificial retinas [20] are the most implementing applications for inductive coupling. But most of the applications use far field communication because of its performance over inductive coupling. Power consumption is one of the major controlling issues for the overall system. For maintaining better power consumption, BCC is proposed by the IEEE 802.15.6 standard group. BCC can transmit data at 10 Mbps and give lower power consumption. So electrical coupling, mainly far field communication, gives support to the BCC for maintaining data rate and power consumption. 3.3 Applications In daily life, human needs some support to prove that he or she is in good health condition. BCC provides support to the human body by continuous monitoring of physiological parameters. So, this medical monitoring and electronic medical records make physicians work easier in the healthcare industry. Sensors, attach to the human body, perform a major role to measure different healthcare issues. For example, the electrocardiogram (ECG), electroencephalography (EEG), body temperature, blood pressure and so on. But these sensors are not wireless; they connect to the body with the wire. To make this complexity easier to monitor the patient condition, there need a wireless connection between the human body and the sensors. This wireless link also provides support to the physicians in more monitoring applications such as during patient's movement, during surgery etc. So, BCC gives lots of support for the health monitoring system. It also gives electronic support to the human life. Data can transfer from one mobile to another mobile through the human body without any wire. It can also give more electronic applications such as hearing aid, door code unlock, transfer visiting card by hand shaking and so on. 26 An Input Amplifier for Body-Channel Communication 3.4 Coupling approaches Two different approaches can be possible to inject data into the human body. One is capacitive coupling, and another one is galvanic coupling. Receiver (Rx) and transmitter (Tx) both need to connect to the coupler for these coupling approaches. Each coupler consists of two electrodes. For capacitive coupling, it needs to connect the electrodes near to the human body. It does not need any direct connection to the human skin. But for galvanic coupling, it needs to connect the electrodes to the human skin directly. The electrodes are designed with two different structures like horizontal and vertical. Capacitive coupling can be possible in both of the structure; but, galvanic coupling can only be possible in horizontal structure. Dielectric material uses in between of the two electrodes. There is an important difference between the capacitive coupling and galvanic coupling. For capacitive coupling, it has influenced by the environment around the human body; whereas for galvanic coupling, it has influenced from the body physical parameters. 3.5 Body model For capacitive coupling, it is not necessary to couple the transmitter and receiver to the human body directly. But it is important to have close proximity of the coupler to the human body. So, a differential pair of electrodes uses to couple the transmitter and receiver. Figure 8: Human body model with capacitive coupling Figure 8 mentions body model with capacitive coupling, which has used in this thesis work. The capacitive coupling needs to the differential pairs of electrode that can place on both sides of the Tx and Rx. It is important to couple the human body capacitively which shows in figure 8. Transmission length of the body channel fully depends upon on the body resistance and coupling capacitor. This means that the signal loss of the receiver increases with the channel length. In this thesis work, the body model defines with unit elements consisting of resistances and capacitance. 27 An Input Amplifier for Body-Channel Communication Figure 9: Electrical model of the human body In figure 9, it defines the electrical body model which uses in this thesis work. The human body model is created by cascading arm unit elements and torso unit elements. The calculation about the body model is followed by studying previous literature. For example, Zimmerman's method [3] uses to calculate the coupling capacitance of the external ground. The unit elements mainly design with the resistance and capacitance. Table 5 describes different resistances and capacitance values. Figure 10: Arm unit element with the combination of capacitance and resistance. 28 An Input Amplifier for Body-Channel Communication Figure 11: Torso unit element description as a circuit view of the body model Figure 12: Groundmesh for body model Figure 10 and 11 represent unit elements of the body model for this thesis work. In figure 10, the arm unit model defines with proper design. It is designed with the resistance and capacitance. Figure 11 shows a torso unit element and it also designs with the resistance and capacitance. So, the total body model cascades with several arm unit elements and torso unit elements. In figure 12, another important element for body model is mentioned which is groundmesh. Finally, it is clear that all the elements are designed with RC links. The different resistance and capacitance values are defined in the table 5 below. 29 An Input Amplifier for Body-Channel Communication Table 5: Different capacitive value for body model Component Value Description ArmCap 48 pF Parallel body capacitance of the arm UE ArmCapGnd 2 pF Coupling capacitance to the external ground from arm UE ArmRes 60 Ω Parallel resistance of the arm UE TorsoCap 310 pF Parallel body capacitance of the torso UE TorsoCapGnd 5 pF Coupling capacitor to external ground from torso UE TorsoRes 7Ω Parallel resistance of the torso UE Ct_txgnd 10 fF Coupling capacitance between Tx ground and ground Ct_rxgnd 10 fF Coupling capacitance between Rx ground and ground FeetCap 10 nF Coupling capacitance to external ground from the feet PlateCap 10 pF Coupling capacitance between Tx/Rx and human body PlateGnd 50 pF Coupling capacitance between Tx/Rx electrodes crossCap1 100 f Coupling capacitance between vgnd_trx1 to vgnd_trx2 crossCap2 100 f Coupling capacitor between vgnd_trx2 to vgnd_trx1 groundCap1 10 f Coupling capacitor between vgnd_trx1 to univarsal gnd! groundCap2 10 f Coupling capacitor between vgnd_trx2 to univarsal gnd! 30 An Input Amplifier for Body-Channel Communication 3.6 Summary The main idea of this chapter is to define the human body model. Different unit elements are also defined by their values in this chapter. There is a difference between lumped electrical model and simplifier RC electrical model. For lumped electrical model, it is difficult to find out the unit elements value with the theoretical analysis. The reason for that it is more complicated than the simplifier RC model. For making the system less complex, the simplifier RC model is designed in this chapter to define the human body. 31 An Input Amplifier for Body-Channel Communication Chapter 4 4 Analog Receiver Front End Review 4.1 Introduction The analog receiver front end is one of the major parts of body-channel communication (BCC). In BCC, after the body the signal attenuates and becomes very weak. For making the signal stronger with the gain, there needs a preamplifier. For amplifying the weak signal, different amplifier topologies present which can give a better solution. If Tx pulses are at 4 to 5 V, the Rx pulses might be in the lower region. The highest frequency component will be at the carrier frequency, and the lowest frequency in half. All the possible sampled frequencies may occur in between of these two frequencies. At the end of the receiver chain, there may be a comparator or Schmitt trigger for shaping the data. With the fast growing wireless communication, sensor networks are giving good response to form body-channel communication. It is a communication between two devices by using the human body as a channel. There are lots of applications that can implement with it; for example touch based authentication, electronic payment service, e-business card, and touch based advertizement service. There are three important parts in BCC such as channel (human body), a receiver and a transmitter. In this chapter, the main focus will be the analog receiver front end. 4.2 Analog receiver front end topologies A receiver is one of the important parts in AFE for working the overall BCC. Signals are attenuated by the human body. After that, these signals reach to the receiving input as weak signals. The receiver needs to amplify the signal and converts the signals to meet the desired output. Based on this concept, two AFE topologies have considered to perform the overall system. 4.2.1 Resistive feedback amplifier chain To amplify the weak signal, it is necessary to maintain high gain in the receiver chain. In the resistive feedback, each amplifier in the receiver chain uses a resistance as feedback to acquire expected close loop gain. That means there is a resistance between input and output to boost up the close loop gain. At the end of the receiver chain, there will be a Schmitt trigger to support the overall system. Because to gather a higher gain and recover the original signal, it needs a bistable circuit. Figure 13 shows resistive feedback amplifier chain topology which maintains several stages to acquire higher gain. The feedback resistances (R 0, R3) and also the input series resistances (R 1, R2) control the gain of the entire circuit mainly the close loop gain. Low frequency close loop gain is given by ACL =(1+ R1 ∣LG∣ ) R0 (1+∣LG∣) (1) Where ACL is the closed loop gain and LG is the loop gain which is determined by the open-loop DC gain of the amplifier. 32 An Input Amplifier for Body-Channel Communication Figure 13: Resistive feedback amplifier chain topology 4.2.2 Capacitive feedback amplifier chain In capacitive feedback amplifier chain, the feedback is mainly controlled by the capacitor. The main advantage of this topology over resistive feedback topology is that it gives less noise. For this topology, it also needs a Schmitt trigger at the end of the receiver chain to boost up the gain. This topology mainly reduces the input-referred noise of the amplifier chain compare to the resistive feedback. Figure 14 shows the capacitive feedback amplifier chain topology which cascades with several stages for higher gain. Figure 14: Capacitive feedback amplifier chain topology 4.2.3 Amplifier The amplifier is one of the key factors in the receiver chain to amplify the weak signal. The amplifier has gained in the range of 10 1 to 105. As the amplifier is used in the feedback system, its open loop gain must be considered with its close loop gain. It has the following important parameters: • Close loop Gain • Offset voltage and low Noise • Low power 33 An Input Amplifier for Body-Channel Communication • Maximum output voltage swing • Supply Rejection • Slew rate • Unity Gain Bandwidth • Phase Margin • Power Consumption For maintaining the lower supply voltage and less power consumption, different techniques have been designed to support the system. The most useful and common techniques are folded cascade, OTA, Capacitive Feedback etc. Flipped Voltage Follower (FVF) is also a useful technique to achieve the best results. 4.3 Receiver chain architecture Analog front end (AFE) is one of the most important parts in BCC after considering the digital baseband (DBB). Manchester encoded data, which is coming from DBB, pass through the Tx and come out by Rx through body channel. The overall system defines in figure 15. Figure 15: Testbench of the overall AFE by showing the body model and coupling capacitors. In figure 15, the main idea behind this testbench is to make it a complete transceiver in both of the sides. So, if left side is Tx, then right side is Rx and vice versa. It has also body coupled capacitors and the human body as a channel. The lower mid part shows different capacitors that are used to couple the ground with the transmitter or receiver. It means three important things. First is coupling between the transmitter and receiver grounds, the second is coupling the transmitter ground with the external ground, and a third is coupling the receiver ground with the external ground. It is important to understand because of the interface and noise purpose. In this thesis work, analog receiver front end consists of two building blocks. One is preamplifier, which extends up to two stages for giving a good gain. Another one is Schmitt trigger, which is stably triggered the output of the preamplifier to positive and negative states with two thresholds against the variation of the received pulse swing. 34 An Input Amplifier for Body-Channel Communication Figure 16: Analog receiver front end chain for the BCC. Figure 16 shows an analog receiver front end for this thesis work. First two stages are preamplifier, and the last stage is Schmitt trigger. The preamplifier mainly designs as a negative resistive feedback amplifier. It has feedback resistance and also input series resistance which control the loop gain of the circuit. It has biased input port, which given the proper bias in the internal circuit of the preamplifier. The two stages of preamplifier connect with a capacitance and biasing resistances. The biasing resistance gives proper DC biasing to the positive and negative terminals of the preamplifier. The same type of connection implements for the Schmitt trigger as it also needs biasing support. The Schmitt trigger is mainly a positive resistive feedback amplifier which control the two threshold levels of it. 4.4 AFE requirement analysis AFE requirement analyzes from the power transmitt by the Tx, propagation loss due to the channel and the noise. From the thesis specification, the frequency value is fixed but the capacitor value may be changed due to the loss and noise. 1 P txout = ⋅V 2swing⋅f ⋅C 2 where, P txout is the output power of the transmitter V swing is the Voltage swing f is the reference frequency C is the electrode capacitance 1 P txout = ⋅( 0.5V )2⋅10 MHz⋅100 pF 2 =125 uW =-9.0309 dBm (2) 35 An Input Amplifier for Body-Channel Communication Propagation Loss, PL = 20log (VTXOUT / VRXIN ) Assuming a propagation loss is 60 dB, i.e. if the transmitter out is 1 Vp-p then the receiver input should be 1 mVp-p. Now, the input power of the receiver is Prxin = Ptxout + PL = -9.0309+(-60) = -69.0309 dBm Background noise can be calculated from the bandwidth requirements of a digital data of 10 MHz frequency. Digital data consists of several overtones; if we consider two tone signal, then the transmitter needs to transmit data at 50-60 MHz bandwidth. Background noise = -174 dBm/Hz Background noise @ 10 MHz = -104 dBm Background noise @ 20 MHz = -101 dBm Background noise @ 40 MHz = -98 dBm Background noise @ 60 MHz = -96.5 dBm Background noise @ 80 MHz = -95 dBm The background noise at 10 MHz is -104 dBm from the calculation. A digital data consists of more than two tones; for example, the transmitter needs frequency range 50-60 MHz to transmit their two tone signal. If the bandwidth is 60 MHz, then the background noise is -96.5 dBm which is approximately 27 dBm less than the receiver input power. So, it is possible to recover the signal with this specification. 4.5 Summary The AFE consists of the receiver and transmitter. The receiver part designs with different parameters based on the requirement of the system. Amplification and recover of the original signal are the most two important parts of the receiver design. In receiver chain, different topologies presents to support the AFE. The basic of receiver chain topology as well as the topology that uses to support the overall system are discussed in this chapter. But it is not enough to specify the basic and other things in a chapter. It is just a common trend that discuss in this chapter. Requirement analysis also takes a major role to understand some basic. 36 An Input Amplifier for Body-Channel Communication Chapter 5 5 Design of Preamplifier 5.1 Introduction The preamplifier is an important building block in an RF transceiver design. The preamplifier needs to boost the amplitude of the incoming signal to an appropriate level, which optimally loads the input of the Schmitt trigger. So, it mainly makes the weak signal into a strong signal. The signals are attenuated very badly by the human body and then come to the input of the preamplifier. So it needs high gain and also it needs low noise from the preamplifier performance. The technique flipped voltage follower chooses for this preamplifier. The architecture [15] gives low voltage, low noise and low power. Appendix C in the later part of this thesis work gives basic information about conventional amplifier. It will give a good support to answer the question why we use flipped voltage follower topology. 5.2 Flipped voltage follower For reducing the supply voltage, different techniques propose to meet the requirement in analog and mixed signal circuits like folding, triode-mode and subthreshold operation of metal oxide semiconductor (MOS) transistors, floating gate techniques and current mode processing [13]. To maintain the requirement of the integrated circuit design, flipped voltage follower (FVF) chooses in this thesis work. It is a kind of basic cell, which is suitable for low power and low voltage operation. Compare to other topologies, FVF gives a wide range of frequency band and lower output impedance, which is the main advantage of this topology. It is one kind of voltage follower but the main difference of the traditional voltage follower and FVF is that FVF has low output resistance. In traditional voltage follower for improving its high output resistance, there needs to increase the transconductance gain gm which requires large current biasing and also the large W/L ratio. Figure 17: Flipped voltage follower (FVF) The basic structure of FVF mentions in figure 17. Here M1, M2, and M3 represent PMOS transistors. Vin is the input voltage, Vout is the output voltage, Vbias is the biasing voltage, and Ibias is the biasing current. It is a cascade amplifier with negative feedback. The input voltage connects to the gate terminal of M1 transistor and the output is in source terminal. Due to shunt feedback from transistor M2, it obtains very low output impedance, very low voltage close to 37 An Input Amplifier for Body-Channel Communication transistor threshold (Vth), low static power consumption and high gain bandwidth. From figure 17, it is clear that the drain terminal of transistor M1 provides biasing to the gate terminal of M2. Based on that, the architecture is named as flipped voltage follower (FVF). A practical limitation that can be observed from FVF cell is that it is given very small input and output signal swing. Figure 18 represents an open loop analysis of the FVF. Here Vt is the test voltage source, Vr is the received voltage from node Y, and rb is the resistance in parallel to biasing current. Figure 18: Open loop analysis of FVF The open loop resistance at node Y is approximately ROLY ≈r b∥g m r o1 r o2 1 (3) The dominant pole at node Y is defined as 1 (C Y ROLY ) ω pY = (4) CY = Parasitic capacitance at node Y. The open loop resistance at node X is approximately (1+ ROLX ≈ rb ) r o1 g m1 ∥r o2 (5) A high frequency pole at node X is defined as ω pX = 1 (R OLX C X ) (6) CX = Parasitic capacitance at node X. The open loop gain of FVF is Vr VT =−g m2 ROLY AOL = (7) 38 An Input Amplifier for Body-Channel Communication The gain bandwidth product is drawn as GB = g m2 CY (8) The closed loop resistance at node X is RCLX = ROLX (1+∣AOL∣) r (1+ b ) r o1 ) (r o2∥ g m1 = ( g m2 (r b∥g m1 r o1 r o2)) (9) Depending upon the current source Ib, the approximation of RCLX changes. For simple current mirror, it gives one approximation; and for cascade current mirror, it gives another approximation. But it clears that RCLX is very low resistance. 5.3 Preamplifier performance As the paper is going through the ST65 nm CMOS process, the preamplifier design is also made by a ST65 nm CMOS process. The paper [15] gives great inspiration to develop the preamplifier. The preamplifier needs to boost the amplitude of the incoming signal to an appropriate level, which is optimally loads the input of the Schmitt trigger. So, it is very important to maintain a good gain in preamplifier. Reducing power is also an important factor to develop the body-channel communication. There will be always a linearity issue in the close loop circuit. Noise is another issue for building up a good preamplifier. Figure 19: Testbench for preamplifier 39 An Input Amplifier for Body-Channel Communication In figure 19, the preamplifier testbench defines. Two blocks are used in the testbench; one converts a single signal to differential in the input side, and another is a differential output signal into a single signal. The DC levels are very important thing for the preamplifier measurements. From the virtual ground, it needs to be carried almost equal value of the differential inputs. An sp1switch from analogLib uses to give the proper negative feedback from the output of the preamplifier. For the open loop analysis, the sp1switch needs to open. Figure 20: System level results for preamplifier In figure 20, the results show for system level behavior of the preamplifier. The first graph shows the input signal, which has low amplitude around 200 mV; and the second graph shows the output signal, which indicates that the signal amplifies around 2 V. For getting higher gain, the system level gain of the preamplifier is set to 30 dB. The biasing circuit gives the proper biasing for the internal circuit of the preamplifier in transistor level, which defines in figure 21. Figure 21: Biasing circuit for the preamplifier 40 An Input Amplifier for Body-Channel Communication The biasing circuit mainly gives proper bias voltage to the transistors. Figure 21 is a simple biasing circuit where the current mirror uses to bias the PMOS. It uses PMOS, NMOS and load resistance. The gate biasing is useful for the biasing purpose because it can fix with the other connection. Figure 22: The transistor level implementation of preamplifier In figure 22, the transistor level implementation of preamplifier is mentioned b . The circuit is developed with the concept of flipped voltage follower. Two cascades flipped voltage follower stages connect with a series resistance (R SeriesRes). This resistance is mainly a degeneration resistance. The advantages of using this degeneration resistance is that when the input signal is weak, small RSeriesRes gives high gain and low noise. Neglecting the short-channel effect and body effect, and 2 1 assuming RSeriesRes ≫ [15], the equivalent input transconductance is [15]. ( g m1 r o1 r o2 ) RSeriesRes The gain of the circuit is approximately, AV ≈ R LoadRes RSeriesRes (10) The important thing is that the gain does not depend on the transistor but they depend on the resistance. For this, it gives good linearity and high accuracy performance. The circuit is also suitable for low power like 1 V or less. As the thesis requirement is 1 V, this circuit is chosen for that purpose. 41 An Input Amplifier for Body-Channel Communication 5.3.1 Gain analysis of preamplifier Gain means increment of the signal from the input to the output as a power or amplitude. It is normally a ratio between the output signal of a system to the input signal of the same system. If the gain is more than 1, then it is said that the output signal is amplified from the input signal. Figure 23: Open loop gain analysis for the preamplifier Gain requirement or the overall attenuation of the signal in the receiver depends on multiple factors such as the maximum distance traveled, lower supply voltage, lower value of coupling capacitors due to thicker stratum corneum and the requirement of single signal electrode. The large amount of attenuation due to this factor (typically 60-80 dB) is compensated by ac coupled multistage preamplifier for a gain of 60 dB. The open loop gain is considered when there will be no feedback in the system. In figure 23, it shows the open loop gain of the preamplifier. The value measures at 24.01 dB for the open loop gain. 42 An Input Amplifier for Body-Channel Communication Figure 24: Close loop gain analysis for the preamplifier The close loop gain is another kind of important gain analysis which gives the overall gain performance of the system. The close loop gain mainly performs with feedback circuit. Figure 24 shows the close loop gain result which is 19.34 dB. The unity-gain frequency measures 474 MHz from the close loop analysis. Open loop gain must be greater than the close loop gain. From the results of the both analyzes, it is clear that the open loop gain is higher than the close loop gain. 5.3.2 Noise analysis of preamplifier Noise analysis intends for circuits which have a linear time-invariant operating point. In other word, noise analysis works with circuits that have a DC operating point set of suitable bias circuits. Examples of circuits with DC operating point are operational amplifier (opamp, OTA), LNAs. In an opamp, the DC operating point is established by apposite biasing circuits. The biasing scheme ensures that all transistors are in the desired operating region and sufficient currents flow in different nodes of the opamp. For performing noise analysis, a DC analysis should be run with noise analysis to confirm the operating point of opamp. During noise analysis, the switch opens in figure 19. 43 An Input Amplifier for Body-Channel Communication Figure 25: Input referred noise PSD @ 10 MHz by performing the noise analysis with the DC analysis. Noise normally specifics as a spectral density in rms volts per root Hertz, V/sqrt(Hz). Figure 25 shows the variation of input referred noise of the receiver with different frequencies. The inputreferred noise PSD @ 10 MHz measures 8.69 nV/sqrt(Hz) by performing the noise analysis. From the value, it is clear that noise performance at reference frequency is good for the overall receiver chain. Figure 26: Noise comparison between thermal noise and flicker noise 44 An Input Amplifier for Body-Channel Communication Different types of noise present in an integrated circuit. Among them, thermal noise and flicker noise are the most dominant one. Figure 26 presents the comparison between thermal noise and flicker noise. Thermal noise sometimes knows as Johnson noise. It is mainly thermal excitation of the resistors and transistors. As flicker noise is known as 1/f noise, so its power spectrum varies as 1/f. Table 6 lists different input referred noise value for different frequencies. Table 6: Input referred noise at different frequencies Frequency (Hz) Input Referred Noise (nV/sqrt(Hz)) 100 K 60 1M 19 10 M 8.69 Input SNR Calculation is mentioned below: Assuming 100 MHz bandwidth by considering the higher harmonics. So the noise voltage would be 86.9 uV Input SNR = 20log (400 mV/86.9 uV) = 73.26 dB 5.3.3 Results Preamplifier results are listed below: Table 7: Preamplifier results to support the receiver chain Supply Voltage 1V Open Loop Gain 24.01 dB Close Loop Gain 19.63 dB Unity-Gain Frequency 474 MHz Input referred noise PSD @10 MHz 8.69 nV/sqrt(Hz) Phase Margin 59.87 Power Consumption 2.3 mW Biasing Current 83.931 uA As the preamplifier is an important building block of this thesis work, so it is needed to satisfy its result for the overall system requirement. The supply voltage chooses 1 V from the thesis specification. In this thesis work, we want to perform low supply voltage, low noise and low power consumption for operating the BCC. The gain is an important issue with the analyze part. As the system is a low voltage system, so the gain need to be high to amplify the input signal. The two types of gain analyze are performed in this thesis work. These are the open loop and close loop gain analysis. From the analyzes, it observes that the open loop gain is 24.01 dB and close loop gain is 19.34 dB. Unity-gain frequency is the frequency where the gain of the system drops to unity. The 45 An Input Amplifier for Body-Channel Communication unity gain frequency measures 474 MHz after performing the close loop gain analysis of the circuit. Gain and noise both relate to each other. If the circuit gives high gain, then its noise is also high. So for the low noise, gain can not be too high for a circuit. For maintaining the low noise requirement of the circuit, noise analyze gives good result which is 8.69 nV/sqrt(Hz). It is mainly the input referred noise of PSD which measures at the frequency of 10 MHz. From the stability point of view, phase margin is a clever technique to satisfy the overall system. The phase margin is 59.87 for the preamplifier. Power consumption of the preamplifier measures as 2.3 mW and gives support to the requirement of low power. 5.4 Summary From the individual analysis, the preamplifier is working fine. For the body-channel communication, it needs to maintain a good gain for the weak signal. Receiver input is coming from the human body, and it attenuates. As the preamplifier is the first building blocks for the receiver of this thesis work, it needs to be good enough to support the overall system and recover the weak signal. In the gain analyze, it gives supportive result to the system; as well from the noise analyze, it maintains the low noise requirement. These two results are important for satisfying this thesis specification. The power consumption of the preamplifier is a little bit high for this communication. By doing proper biasing, it is possible to achieve better power consumption over this preamplifier. If the biasing current keeps low, then the power consumption will give good result. 46 An Input Amplifier for Body-Channel Communication Chapter 6 6 Design of Schmitt Trigger 6.1 Introduction Schmitt trigger is a kinds of bistable and positive feedback circuit. It coverts a continuous signal into a stable logical signal which gives 1 or 0. It has two levels; one is a high threshold voltage (V thhigh), and another one is a low threshold voltage (V th-low). When the input voltage goes above the V thhigh, it gives a logical signal of 1. When the output is below the V th-low, then it gives the logical 0. If the input voltage remains in between of these two levels, then there is no change in the output signal. 6.2 Compare with comparator The comparator is also an important building block for the analog front end. The comparator is mainly maintaining one reference signal. If the output signal is above the reference signal, then it gives high signal; and if the output signal is below the reference signal, then it gives low signal. From the function, it is clear that comparator has one threshold level whether Schmitt trigger maintain two threshold levels. The gain of the Schmitt trigger should be linear to support the feedback configuration; whether the gain of comparator usually not linear, for that reason, it does not support the feedback configuration. Noise is an important parameter for overall performance. Schmitt trigger reduces the sensitivity to noise and disturbances compare to comparator. 6.3 Advantages of CMOS Schmitt trigger Schmitt trigger uses in many applications both in analog and digital circuit. Most of the applications use the Schmitt trigger to either simplify the design or increase the performance of the design. It has also other advantages which are mentioned below: • It has high input impedance which is typically 1012 Ω; • It has balanced input and output characteristics; • It has two thresholds which are normally symmetrical to half the supply voltage; • Thresholds show low variations by changing the temperature; • It has low power consumption; • High noise immunity. Schmitt trigger also uses in the integrated circuit, where slow edge transition needs to convert the signal into digital form. So, it clears that Schmitt trigger can transfer analog signal to the digital signal. 47 An Input Amplifier for Body-Channel Communication 6.4 Schmitt trigger topologies There are many approaches to implement a Schmitt trigger. Operational transconductance amplifier (OTA) can be used as a Schmitt trigger. Normally OTA is a differential structure, and it has two inputs. Output can be differential or single depending upon the circuit criteria. If there is a connection between positive input and output, then the OTA can be acted as a Schmitt trigger. Figure 27: An OTA in a non-inverting configuration effectively forming a positive feedback Schmitt trigger Figure 27 shows an OTA in a non-inverting configuration effectively forming a Schmitt trigger. Here, Rs and Rf are the series, and feedback resistance respectively, Vin is the input voltage and Vo is the output voltage. For non-inverting configuration, the input connects to the series resistance which attaches with the positive terminal of OTA. So, the circuit acts as a positive feedback Schmitt trigger. The output voltage has the same sign that the input voltage is as it is non inverting. That means when the input is above the high threshold or below the low threshold the output gives the same sign. So, it is possible to use non inverting OTA as a Schmitt trigger by applying the threshold levels. In a non-inverting Schmitt trigger configuration, two threshold levels can be possible. One is positive threshold, and another is negative threshold. When the non-inverting terminal is above the inverting terminal, it gives positive threshold; because it switches to the positive supply voltage. When the inverting terminal is above the non-inverting terminal, it gives negative thresholds because; it switches to the negative supply voltage. V threshold = V supply Rs (R f + Rs ) Figure 28: An OTA in an inverting configuration effectively forming a positive feedback Schmitt trigger (11) 48 An Input Amplifier for Body-Channel Communication Figure 28 describes an OTA in an inverting configuration effectively forming a positive feedback Schmitt trigger. Here, Rs and Rf are the series, and feedback resistance respectively, Vin is the input voltage and Vo is the output voltage. For the inverting configuration, it is just the opposite of non inverting OTA. In this case, the input connects in the negative terminal of OTA, but, the feedback connection is in the positive terminal. The output gives the opposite sign of the input signal and it is also valid for the high and low threshold levels. It can also be used as a Schmitt trigger. 6.5 Schmitt trigger performance Schmitt trigger is another important building block for the receiver chain. To recover the original signal, there needs a Schmitt trigger at the end of the receiver chain. Schmitt trigger testbench for this thesis work mentions in figure 29. In testbench, there are differential inputs but the output is single. There is a positive feedback connected by a resistance to maintain the threshold levels. Figure 29: Schmitt trigger testbench The Schmitt trigger has two threshold levels for triggering the output voltage. This two threshold levels are given values against the variation of the received pulse swing. The Schmitt trigger consists of three resistors (R6, R7, R8). These resistors gives positive and negative threshold values for triggering the output voltage (VTH, VTL). Thus VTH and VTL are expressed below: R6 ((R7∥R8)+R6) (12) (R6∥R8) ((R6∥R8)+R7) (13) V ThresholdHigh (TH) = V ThresholdLow (TL) = 49 An Input Amplifier for Body-Channel Communication Figure 30: Schmitt trigger schematic (differential two stage amplifier) for the receiver chain In figure 30, it mentions Schmitt trigger schematic for this thesis work. It is mainly a two stages OTA architecture. Two stage mainly refer to the gain stages in the amplifier. The first stage consists of a differential pair, which drives the output stages. The first stage has a n-channel differential input pair with a p-channel current mirror active load. To provide stability compensation, a feedback path in the form of a resistor (R1) and capacitor (C0) is inserted. Capacitance C0 is mainly a Miller capacitance, which gives a high gain between the input and output stages. Appendix C mentions clearly the two-stage OTA gain equation and frequency response. In figure 31 and 32, the performances of Schmitt trigger are drawn. For the overall receiver, it also needs a higher gain in the Schmitt trigger stages. Schmitt trigger gives 29.64 dB gain which is good to support the receiver chain. The input referred noise PSD @10 MHz is 6.65 nV/sqrt(Hz). Figure 31: Gain performance for the Schmitt trigger 50 An Input Amplifier for Body-Channel Communication Figure 32: Input referred noise for the Schmitt trigger 6.6 Results The Schmitt trigger performances are also very important for the overall receiver chain. The gain and the noise are two important measurements for the Schmitt trigger. The performances of Schmitt trigger are mentioned in table 8. Table 8: Schmitt trigger performance analysis Parameters Results Supply Voltage 1V DC Gain 29.64 dB Input-referred noise @ 10 MHz 6.65 nV/sqrt(Hz) Power Consumption 563.6 uW As it is the thesis specification that the supply voltage should be 1 V, Schmitt trigger performs with low supply voltage of 1 V. It is important to maintain DC biasing levels in the Schmitt trigger. Gain is the main parameters for the Schmitt trigger to recover the signal. It has given a DC gain of 29.64 dB. So, it can successfully recover the original signal. Another important parameter is noise. For finding the noise value, noise analyze performs. The input-referred noise at 10 MHz measures as 6.65 nV/sqrt(Hz). So, it gives a satisfactory result to the noise point of view. The power consumption is also low which is 563.6 uW. 51 An Input Amplifier for Body-Channel Communication 6.7 Summary The Schmitt trigger has many advantages which is already mentioned in this chapter. A wide threshold range, low power consumption, high noise immunity, and low board space are the most renown advantages which makes it separate from another circuit. The schematic, that is used in this thesis to design the Schmitt trigger, is a two stage OTA architecture. Two stage CMOS OTA is a classic circuit used in many integrated circuits. It can easily suit to the low power applications. For the overall system, it needs better performance from the Schmitt trigger. The performance of Schmitt trigger is good in this thesis work. So, it can give better support to the overall BCC system. 52 An Input Amplifier for Body-Channel Communication Chapter 7 7 Receiver Results 7.1 Introduction The overall receiver chain consists of three stages. First two stage design as preamplifier and last stage designs as Schmitt trigger. For getting the higher gain from the preamplifier, it is cascaded into two stages. The overall gain performance is very important due to recover the original signal. 7.2 System level performance of receiver chain For recovering the weak signal from the body channel, it is important to get a good gain from the receiver chain. At the system level, the receiver chain is designed by using verilogA which is available in Appendix A. The receiver chain for AFE is showen in figure 33. The receiver chain consists of three parts. First two parts cascade with two preamplifier and the last part is Schmitt trigger. As it is already mention that the receiver needs good gain, so two preamplifier connect with a capacitor and biasing resistors. Schmitt trigger, followed by preamplifier, also connects with a capacitor and biasing resistors. The biasing resistors are very important to maintain DC biasing level in the both inputs. There is another biasing input present in each part of the circuit to give a proper gate voltage to the internal transistors. Figure 33: Receiver chain presenting with two preamplifier stages and one Schmitt trigger stage for the AFE 53 An Input Amplifier for Body-Channel Communication Figure 34: System level results showing transmitter output, receiver input, preamplifier output and receiver output In figure 34, it is clear that the data recover in system level is perfectly done in receiver output. The first result indicates the transmitter output which is mainly XOR of the clock and the input signal. This output then passes through the human body and gives spikes due to the attenuation caused by the body channel and power loss of the body. So, this spikes are the receiver input which is clear in second result. Third result shows the preamplifier output after amplification of the spikes. By setting the Schmitt trigger's two threshold values, the last result gives the receiver output. So, from the figure it is clear that the transmitted signal recovers perfectly in the receiver output. 54 An Input Amplifier for Body-Channel Communication 7.3 Transistor level performance of receiver chain Figure 35: Transistor level results of the different nodes of the receiver chain Figure 36: Gain performance of the overall receiver chain in different frequencies 55 An Input Amplifier for Body-Channel Communication The transistor level results do not make so difference to the system level design. Figure 35 shows the transistor level behavior of different nodes of the receiver chain. The results are transmitter output, receiver input, preamplifier output, and Schmitt trigger output respectively. As it is already mention that for the high gain, the preamplifier extends in two stages, so the preamplifier output result has been shown after the second stage. The preamplifier stage convert 30 mV input signal to a voltage swing of 500 mV. The Schmitt trigger then converts this signal to the 1 V rail to rail which gives the digital data. The power consumption of the receiver measures 863.4 uW due to the proper biasing. In figure 36, the overall receiver gain is showing. The gain is above 50 dB. For the receiver chain, it needs more than 50 dB gain to recover the signal. So, the proposed receiver is working fine as a gain point of view. 7.4 Corner analysis Corner analysis is a kind of analysis where worst-case simulation presents. It is mainly varying different process nodes with different supply voltages and temperatures. For finding the worst-case performance in the integrated circuit, corner analysis gives a better view of the variations. For example, when the transistors are changing their states from one logic state to another, then their speed may change. To capture the variation of the transistors, corner analysis gives effective results. 7.4.1 Types of combinations Three types of situation or state are mainly possible in the process nodes: Fast (F), Slow (S) and Typical (T). But at a time two combinations are possible. For this two combination, the first letter indicates the N-Channel and the second letter indicates the P-Channel. For example, if FS corner analysis is performed then N-Channel is fast and P-Channel is slow. In this thesis 9 process combinations are used to support the overall system. The combinations are as follows: • Fast-Fast (FF) • Slow-Slow (SS) • Typical-Typical (TT) • Fast-Slow (FS) • Fast-Typical (FT) • Slow-Fast (SF) • Slow-Typical (ST) • Typical-Fast (TF) • Typical-Slow (TS) In this thesis work, all of these combinations are performing. The performance of corner analysis is mentioned in next section 7.4.2. 56 An Input Amplifier for Body-Channel Communication 7.4.2 Performance In this thesis, supply voltage varies from 0.9 V to 1.1 V which is the minimum and the maximum supply voltage for the system. The corner analyzes result is shown in figure 37. The total corner points in this simulation are 9*3*3=81 points, where 9 for the 9 process corner, 3 for the supply voltage variation, and 3 for the temperature variation. Figure 37: Corner analysis of the receiver chain The ones or zeros are decoded in the baseband by sampling the obtained data from the AFE. The sampling frequency is eight times of clock frequency. From this, it is clear that one can detect when the length is more than 0.0125 us. The length of ones and zeros are respectively 0.05 us and 0.05 us. From the corner analyzes, it is clear that the minimum length of ones is 0.025 us, which can be easily detected by the digital baseband. 57 An Input Amplifier for Body-Channel Communication 7.4.3 Eye diagram The eye diagram is a kind of graphical representation that can give a set of information about the high speed digital data transmission. The digital waveform folds when the eye diagram is performing with the individual bit and continuously repeating the samples of the waveform. So, the resultant waveform gives a statistical waveform which is looking like a human eye. It can give information about the noise, jitter, rise time, fall time etc. Figure 38: Eye diagram for the receiver chain In figure 38, the eye diagram is for 60 ns period and figure 39 is for 100 ns. From the graph, it is clear that in different time period the eye diagram gives different values. The measured value of the figure 38 is listed in the following table 8: Table 9: Measurement table for eye diagram Parameters Values Horizontal eye opening 19.5 ns Vertical eye opening 0.94 V Timing variation at zero crossing 4 ns Noise margin 0.47 V Eye level zero 7 mV Eye level one 1 mV Rise time 3 ns Fall time 4.5 ns 58 An Input Amplifier for Body-Channel Communication Horizontal eye opening is eye width. It can be measured between the difference of the two crossing points. The vertical eye opening is the eye height. Normally eye height is the eye amplitude but for the noise the eye tries to close. So, it needs to measure the vertical eye opening. From the table, the value of horizontal eye opening is 19.5 ns and the vertical eye opening is 0.94 V. Timing variation at zero crossing means the amount of distortion in where the zero crossing occur. The value of timing variation of zero crossing measures as 4 ns. Noise margin is the minimum tolerance level of the proper operation of the circuit; and it measures as 0.47 V for the receiver chain. Rise time and fall time are the two important parameters for the integrated circuit. The rise time is 3 ns and fall time is 4.5 ns. Figure 39: Eye diagram performance @ 100 ns period Jitter is an important factor in digital data transmission. It is mainly the deviation of time from the ideal timing of a data-bit event. It can be calculated by the difference between the deviation of rise time and fall time. Rise time is the upward slope of the eye diagram and fall time is the downward slope of the eye diagram. Another way, jitter is the phase shift in a signal caused by a noise. The slope of the eye is a very important thing. It determines how sensitive the signal is with respect to timing errors. As a result, if the slope of the eye is less, then it allows the eye to be opened more, and it will be less sensitive with respect to the timing errors. From the jitter point of view, if the timing error is less sensitive, then it gives less noise. 59 An Input Amplifier for Body-Channel Communication 7.5 Noise analysis Noise is the most important analysis because it distorts the signal and makes it difficult to detect the signal. Noise is not a constant result. It always gives random value. It is not possible to detect the instantaneous value of noise at any time. Noise can be happened, either internally in the circuit, or externally caused by the external source. Different noise factors like thermal noise, flicker noise etc. are responsible for making interference of the signal. Figure 40: Adding 200 uV noise with supply voltage and Tx,Rx outputs 60 An Input Amplifier for Body-Channel Communication Figure 41: Adding 1 mV noise to the supply voltage and showing output signal of Tx,Rx In figure 40 and 41, they present the supply noise performance. In figure 40, 200 uV supply noise is added to the supply voltage. The first graph of figure 40 indicates adding noise to the supply voltage. After that input and output results are showing. In figure 41, 1 mV supply noise is added to the supply voltage. The first graph in figure 41 shows the adding noise to the supply voltage. The rest of the two graphs show input and output results respectively. From the graphs, It is clear that the signal recovers successfully even though the adding of noise. But if the 2 mV supply noise is added to the supply voltage, the signal can not be recovered. So, more than 2 mV is harmful to the signal as it can not be recovered. 61 An Input Amplifier for Body-Channel Communication Figure 42: Input referred noise of the overall receiver @ 10 MHz In figure 42, the input referred noise performance is presented. The value is 41.9 nV/sqrt(Hz). From the value it is clear that the input referred noise is high for the receiver. The biasing transistors of the preamplifier and also the Schmitt trigger are considered in these results. And the biasing circuit is also considered. If we exclude them from the noise analysis then the result will give a better hope for the noise point of view. 7.6 Summary In this chapter, the overall receiver performance is mentioned. Two types of important analysis are performed to observe the receiver performance. One is corner analysis and another is eye diagram. For getting the proper results, these two analyzes give good support to the system performance. From the gain analyze, it is said to be good for the overall system that it gives high gain. The noise performance is also satisfactory at this point although it is high because of proper biasing. The adding noise to the input signal gives good support to the AFE receiver. So, by the overall performance it can be said that this receiver can be used as a BCC for data transfer. 62 An Input Amplifier for Body-Channel Communication Chapter 8 8 Conclusion and Future Work 8.1 Conclusion In a different time, different transceivers present in the body-channel communication (BCC). As it is important that the receive signal carry all the information, it needs to be careful about the receiver sensitivity and also the noise performance. It is also very important to maintain high gain because the signal is attenuated badly by the human body. So, low-power and low-voltage are playing important role in this kind of communication. In this paper, the propose receiver processes in ST65 nm CMOS standard. The components that are used for the designing purpose at the transistor level are from standard cell. Manchester data encoding has been used as suggested by the AFE requirement analysis and previous works done. The proposed receiver is maintaining all the basic requirements that are important for this kind of communication. The finite gain of the preamplifier is most important for the noise at low frequency. It removes high frequency noise. In this paper, the preamplifier close loop gain is 19.34 dB and the Schmitt trigger gives 29.34 dB. As it is necessary to maintain a good gain (approximately 60 dB) in receiver chain, preamplifier and Schmitt trigger follow that requirement. In receiver chain, there are three stages. These three stages divide into two parts; one is cascade preamplifier and another is Schmitt trigger. So from the gain point of view, it gives more than 50 dB as it requires to recover the original signal. To amplify the weak signal and makes it a strong signal, there will need a high gain preamplifier. The designed preamplifier is giving a good gain such as it gives 24.01 dB open loop gain and 19.43 dB close loop gain which also mention above. Flipped voltage follower topology is used to design the preamplifier to support the low supply voltage of 1 V, low noise, and also low power consumption. The input referred noise is 8.69 nV/sqrt(Hz) and the input SNR is 73.26 dB. These results indicate that this preamplifier is suitable for the overall BCC system. The Schmitt trigger, which is mainly a bistable and positive feedback circuit, can convert the signal into logical 1 or 0 or original data. For the overall receiver, it needs good gain performance for the Schmitt trigger, and it gives 29.64 dB DC gain with the same supply voltage. Now the receiver gain is above 50 dB, which perform very strongly for this kind of communication. Corner analysis and Eye diagram are also giving good support to prove the performance of the receiver. 63 An Input Amplifier for Body-Channel Communication 8.2 Future work body-channel communication is creating its faster depend as the communication world gives us powerful new innovation day by day. Several researchers in this field to create a lot of hope for the future. Some future works are listed below which will move this thesis further. • A Strong transmitter driver that can deal with highly varying load impedance of the human body; • Duplex communication of the AFE to maintain channel estimation; • Design a useful biasing circuit for the receiver AFE that can give low power consumption; • Design a single stage amplifier to observe the performance that can help body-channel communication; • Design a programmable gain amplifier (PGA) or variable gain amplifier (VGA) to compare with the developed amplifier for body-channel communication; • Design a low-resolution ADC for recovering the original signal and compare it with the Schmitt trigger and comparator that how it performs in the receiver AFE; • Implement power down concept (PD) to maintain a better view about the receiver AFE when the circuit is not being used; • For the test purpose, develop a single switch operation to control all the tests from outside; • In this thesis work, voltage-to-voltage amplifier topology is used for designing the preamplifier. Instead of using voltage-to-voltage amplifier, we can use current-to-current amplifier. Because it has two good advantages: good dynamic range and the bandwidth are independent of the close loop gain; • Pseudo random binary sequence (PRBS) in Tx for testing. There might be many future works left for this kind of communication. The human body communication can be an interested topic for the future. The recent applications with the bodychannel communication attract the engineers for making the challenging world. Hopefully the future generation will satisfy all the requirements and future tasks to improve the body-channel communication . 64 An Input Amplifier for Body-Channel Communication Abbreviation 9 Abbreviation A Word Meaning Used Page Number Definition AFE Analog Front End 3,14,16,21,23,31,33,35 Consists of receiver, ,62,63 transmitter and antenna sub-systems. ASIC Application Specific Integrated Circuit 14 Standard Cell Design. ADC Analog to digital converter 14,22,63 Converts analog system into digital system. BAN Body Area Network 10,11,12,13,14,15,16,1 Communication is 8,22,25,31,33 entirely within, on and in proximity of a human body. BCC Body-Channel Communication 3,10,12,19,25,34,62,63 Communication where human body is considered as a channel. BSN Body Sensor Network 10 Sensing network through human body. CMOS Complementary Metal Oxide Semiconductor 21,22,38,52,63 Kind of technology to build up the integrated circuit. CRC Cyclic Redundancy Code 21 Error detecting code used in digital system. CDR All digital Clock and Data Recovery 22 Data recovery system to get the original signal. B C 65 An Input Amplifier for Body-Channel Communication D DBB Digital Baseband 22 Kind of data transmission to send the digital data. DCI Direct-Coupled Interface 22 A direct connection provides the coupling between stages. DSSS Direct Sequence Spread Spectrum 19,22 A modulation technique where transmitted signal take more bandwidth than information signal. DAC Digital-to-Analog Converter 13 Convert digital system into analog system. ECG Electrocardiography 10,25 Heart measurement machine. EEG Electroencephalography 10 Electrical measurement along with the human body. EO Electro-Optic 20,23 Optical modification by electrical field. EIT Electrical Impedance Tomography 21 A medical imaging technique. FPGA Field Programmable Gate Array 13,14,16 Manufacturing design unit. FM Frequency Modulation 12,19,23 Modulation technique where information is conveyed by varying the frequency. FSK Frequency shift keying 12,19,23,24 Modulation technique where digital information is transmitted through discrete frequency changes of a carrier wave. FE Finite Model 21 FVF Flipped Voltage Follower 33,36,37 E F One kind of amplification 66 An Input Amplifier for Body-Channel Communication technique. H HBG Hierarchical Block Gating 22 HBC Human Body Communication 22 Communication through human body. I IBC Intrabody Communication 22 Communication through human body. IEEE Institute of Electrical and Electronics 10,11,25 Engineers Standard organization. IBM International Business Machines (Corp.) 12 Organization. ICDs Implanted Cardiac Defibrillators 25 Heartbeat device. K KAIST measuring Korea Advanced Institute of Science 3,12 and Technology University. L LNA Low Noise Amplifier 14,42 Amplification low noise. with LVA Low Voltage Amplifier 24 Amplification low voltage. with LVM Low Voltage Multiplier 24 Multiplication low voltage. with M MIT Massachusetts Institute of Technology 13 University. MCU Microcontroller Unit 22 Embedded design MHz Mega Hertz 10,12,16,19,22,24,34,3 Unit (Frequency). 5,42,43,44,45,49 Mbps Mega bit per second 12,20,23,,24,25 Unit (speed). MOS Metal Oxide Semiconductor 36 Consist of three layer. N NFC Near Field Communication 3,25 To establish Radio Communication with two devices. 67 O OOK An Input Amplifier for Body-Channel Communication On-Off Keying 19,22,23,24 Modulation technique. OTA Operational Transconductance Amplifier 33,42,46,47,8,49,51 Amplification depending upon transconductance. OpAmp Operational Amplifier 42 DC-coupled high gain amplification. P PAN Personal Area Network 10,18,19,25 Establish a computer network around human body. PPM Pulse Position Modulation 22,24 Modulation technique. PGA Programmable Gain Amplifier 63 Amplification by varying the gain as a program. PSD Power Spectral Density 43,44,45,49 Signal distribution over different frequencies PRBS Pseudo Random Binary Sequence 63 R Rx Receiver 3,14,15,20,23,26,29,31 Get the signal or ,33,60 information. RFID Radio Frequency Identification 3 Wireless tags that can transfer data via electromagnetic fields. S SOC System-on-a-chip 22 Combined different component into a single chip. SNR Signal-to-Noise Ratio 3,44,62 Ratio between signal and noise. STD Single-ended to differential 24 One input and two outputs. T Tx Transmitter 12,14,15,20,26,29,31,3 Send the signal or 3,34,59,60,63 information. TRx Transceiver 13,29 Send and receive the signal together. 68 An Input Amplifier for Body-Channel Communication U UE Unit Element 14,29 Identity element. USB Universal Serial Bus 11 Standardization of different port connections. V VGA Variable Gain Amplifier 63 Amplification by varying the gain. 69 An Input Amplifier for Body-Channel Communication References 10 References 1. Erik Karulf, [email protected] “Body Area Network (BAN)” ( A survey paper written under guidance of Prof. Raj Jain). 2. Hoi-Jun Yoo, and N. Cho, “ Body Channel Communication for Low Energy BSN/BAN,” in IEEE Asia Pacific Conference on Circuits and Systems, 2008, pp. 7-11. 3. T. Zimmerman, “ Personal area networks (PAN): Near-field intra-body communication,” Master's thesis, MIT, 1995. 4. K. Hachisuka, A. Nakata, T. Takeda, Y. Terauchi, K. Shiba, K. Sasaki, H. Hosaka and K. Itao, “ Development and performance analysis of an intra-body communication device,” in 12th international conference on Solid State Sensors, Actuators and Microsystems, Boston, June 8-12, 2003. 5. K. Hachisuka, Y. Terauchi, Y. Kishi, T. Hirota, K. Sasaki, H. Hosaka and K. Ito, “ Simplified Circuit Modeling and Fabrication of Intrabody Communication Devices” in 13th international conference on Solid State Sensors, Actuators and Microsystems, Seoul, Korea, June 5-9, 2005. 6. M. Shinagawa, M. Fukumoto, K. Ochiai, and H. Kyuragi, “ A near-field-sensing transceiver for intra-body communication based on the electro-optic effect,” in Proceedings of the 20th IEEE Instrumentation and Measurement Technology Conference 2003, vol. 1, May 2003, pp. 296-301. 7. M. Fukumoto, M. Shinagawa, K. Ochiai, and H. Kyuragi, “ A near-field-sensing transceiver for intrabody communication based on the electrooptic effect,” in IEEE Transactions on Instrumentation and Measurement, Vol 53, December 2004, pp. 1533-1538. 8. M. Wegmueller, A. Kuhn, J. Froehlich, M. Oberle, N. Felber, N. Kuster, and W. Fichter, “ An Attempt to Model the Human Body as a Communication Channel,” IEEE Transactions on Biomedical Engineering, Vol. 54, 10, October 2007. 9. S. Song, N. Cho, S. Kim, J. Yoo, and H. Yoo, “ A 4.8mW 10Mbps Wideband Signaling Receiver Analog Front-End Human Body Communications,” in proceedings of the 32nd European Solid State Circuits Conference 2006, September 2006, pp. 488-491. 70 An Input Amplifier for Body-Channel Communication 10. S. Song, N. Cho, S. Kim, J. Yoo, S. Choi, and H. Yoo, “ A 0.9V 2.6mW Body-Coupled Scalable PHY Transceiver for Body Sensor Applications,” in Digest of Technical Papers. IEEE International Solid State Circuits Conference 2007, February 2007, pp. 366-609. 11. S. Song, N.Cho, and H. Yoo, “ A 0.2mW 2Mbps Digital Transceiver Based on Wideband Signaling for Human Body Communications,” in IEEE Journal of Solid State Circuits, Vol. 42, No. 9, September 2007, pp. 2021-2033. 12. Yu. Lin, Yo. Lin, C. Chen, H. Chen, Y. Yang, and S. Lu, “ A 0.5V Biomedical System-ona-Chip for Intrabody Communication System,” in IEEE Transaction on Industrial Electronics, Vol. 58, No. 2, February 2011, pp. 690-699. 13. R.G. Carvajal, J. Ramirez-Angulo, A.J. Lopez-Martin, A. Torralba, J.A.G. Galan, A. Carlosena, and F.M. Chavero, “ The Flipped Voltage Follower: A Useful Cell for Low-Voltage Low-Power Circuit Design,” in IEEE Transaction on Circuits and Systems—I:Regular Paper, Vol. 52, No. 7, July 2005, pp. 1276-1291. 14. http://www.ieee802.org/15/pub/TG6.html 15. H. Libin, L. Zhiqun, “ A Low-Voltage CMOS Programmable Gain Amplifier for WSN Applification,” Wireless Communications and Signal Processing (WCSP), 2011, pp. 1- 4. 16. http://www.who.int/pehemf/ about/en/what_is_emf_swedish.pdf 17. http://www.holstcentre.com/Home/PartneringinResearch/SharedPrograms/TechnologyInt egration/BodyAreaNetworks.aspx 18. T. Schenk, N. S. Mazloum, L. Tan, and P. Rutten, “ Experimental Characterization of the body-coupled communications channel,” in ISWCS 2008, October 2008. 19. S. Gabriel, R. W. Lau, and C. Gabriel, “ The dielectric properties of biological tissues: II. Measurements in the frequency range 10 Hz to 20 GHz,” Phys. Med. Biol., p.p 2251-2269, Nov. 1996. 20. Hoi-Jun Yoo, C. van Hoof, “Bio-Medical CMOS ICs”. 21. P. E. Allen, and D. R. Holberg, “CMOS Analog Circuit Design”, 2nd ed. Oxford University Press, 2009. 71 22. An Input Amplifier for Body-Channel Communication B. Razavi, “Design of Analog CMOS Integrated Circuits”, Tata McGraw-Hill, 2002. 23. David A. Johns, Ken Martin, “Analog Integrated Circuit Design”, John Willy & Sons, Inc, 1997. 72 An Input Amplifier for Body-Channel Communication Appendix 11 Appendix A For the system level design, it needs verilogA code to design the overall receiver. Because verilogA is an industry define modeling language for analog circuits. As the overall receiver is cascaded with the preamplifier and Schmitt trigger, we need verilogA code for designing the preamplifier and Schmitt trigger at the system level. Section 11.1 and 11.2 in the below define verilogA code for the preamplifier and Schmitt trigger which are used in this thesis work. 11.1 VerilogA code for preamplifier The verilogA code for preamplifier is stated below which is important for the system level performance. // VerilogA for bodyLinkRxTrack3, opamp_verilog3, veriloga // VerilogA for customLib, modelopamp, veriloga // This is a single-poll roll-off model of op-amp. // rin: opamp input resistance // rout: opamp output resistance // ugf: opamp transit frequency 'ft'/unity gain frequency/gain-bandwidth product // gain: open-loop dc gain of the op-amp `include "constants.vams" `include "disciplines.vams" `define dB2dec(x) pow(10,x/20) module OPAMP(Vinp,Vinn,Preampoutp,Preampoutn,vpwr,vgnd,Ibias); parameter real gain=30; parameter real ugf=30M; parameter real rin=1M; parameter real rout=80; //parameter real bw=20K; input Vinp,Vinn,Ibias; electrical Vinp,Vinn,Ibias; output Preampoutn,Preampoutp; electrical Preampoutp,Preampoutn; inout vpwr,vgnd; electrical vpwr,vgnd; analog begin I(Vinp,Vinn) <+ V(Vinp,Vinn)/rin; V(Preampoutp) <+ rout*I(Preampoutp)+ laplace_nd (`dB2dec(gain)*V(Vinp,Vinn),{1}, 73 An Input Amplifier for Body-Channel Communication {1,`dB2dec(gain)/(2*3.141592653*ugf)}); V(Preampoutn) <+ rout*I(Preampoutn)- laplace_nd (`dB2dec(gain)*V(Vinp,Vinn),{1}, {1,`dB2dec(gain)/(2*3.141592653*ugf)}); end 11.2 VerilogA code for Schmitt trigger As the Schmitt trigger depends on two threshold levels to detect the actual signal, so there are two different verilogA is used two verify the whole system. VerilogA(1) // VerilogA for bodyLinkTx, bodyLinkTxOPAMP_withsat, veriloga `include "constants.vams" `include "disciplines.vams" module bodyLinkRx_OPAMP_SchmittTrigger(vpwr, Vinn, Vinp, Triggerout, vgnd,Ibias); inout vpwr,vgnd; electrical vpwr,vgnd; input Vinn,Vinp,Ibias; electrical Vinn,Vinp,Ibias; output Triggerout; electrical Triggerout; parameter real gain=1000; parameter real trise = 1p from (0:inf); parameter real tfall = 1p from (0:inf); real temp; analog begin temp = gain*V(Vinp, Vinn); if(temp > V(vpwr)) temp = V(vpwr); if(temp < V(vgnd)) temp = V(vgnd); V(Triggerout) <+ temp; end endmodule 74 An Input Amplifier for Body-Channel Communication The second verilogA code for Schmitt trigger is stated below: VerilogA(2) // VerilogA for bodyLinkRx, bodyLinkRxAdc, veriloga `include "constants.vams" `include "disciplines.vams" module bodyLinkRxAdc(OUT, TESTOUT, CLK, NPD, RESET, TESTIN, TESTMODE, vInP, vgnd, vinN, vpwr); output OUT; electrical OUT; output [7:0] TESTOUT; electrical [7:0] TESTOUT; input CLK; electrical CLK; input NPD; electrical NPD; input RESET; electrical RESET; input [7:0] TESTIN; electrical [7:0] TESTIN; input [7:0] TESTMODE; electrical [7:0] TESTMODE; input vInP; electrical vInP; input vgnd; electrical vgnd; input vinN; electrical vinN; input vpwr; electrical vpwr; parameter integer hyst_state_init=1; parameter real sigout_high = 1; parameter real sigout_low = 0; parameter real sigtrig_high = 1 ; parameter real sigtrig_low = 0; parameter real tdel = 0 from [0:inf); 75 parameter real trise = 1p from (0:inf); parameter real tfall = 1p from (0:inf); integer hyst_state; real sigout_val; analog begin @ ( initial_step ) begin hyst_state = hyst_state_init; end @ (cross (V(vInP) - sigtrig_high,1) ) if (hyst_state == 0) hyst_state = 1; @ (cross (V(vinN) - sigtrig_low,0) ) if (hyst_state == 1) hyst_state = 0; if (hyst_state == 1) begin sigout_val = sigout_high; end else begin sigout_val = sigout_low; end V(OUT) <+ transition (sigout_val, tdel, trise, tfall); end endmodule An Input Amplifier for Body-Channel Communication 76 An Input Amplifier for Body-Channel Communication 12 Appendix B Appendix B defines the characteristics of ST65 nm technology. There are different types of transistor and resistor models available in the ST65 nm standard. For designing the overall receiver, we use LVT_LP transistor model and P+ Poly resistor model in this thesis work. Section 12.1 describes some important information about transistor model and resistor model. It also defines simulation conditions and output parameter definition for the transistor model. 12.1 Characteristics of CMOS065 nm technology (a)Transistor Models In the CMOS065 technology, different transistors are used for different purpose. In this paper, the main focus on LVT_LP transistors models for low voltage and low power. There are some important information for this kind of models stated below: • Supply voltage (Vdd) is 1.2 V; • Drawn gate length varies from 60 nm to 10 um; • Drawn transistor width varies from 0.12 um to 10 um; • Devices temperature varies from -40 °C to 150 °C; • Vgs, Vds, Vbs vary from 0 to 1.32 V (i.e. Vdd +10%). (b)Conditions of Simulation Simulations were done with Bench v3.3.1dev using Eldo simulator v6.8_3.1. If not explicitly mentioned elsewhere, temperature is set to 25 °C and Vbs to 0 V. Extra global parameters used: • lvtlp_dev = 0 • mismatch_crolles = 0 • gflag__rgateswitch__all__cmos065 = 0 • gflag_rgateswitch_all_cmos065 = 0 • gflag__noisedev__all__cmos065 = 0 • agesimulation = 0 (c)Output parameter definition In what follows, M, W and L (all default to 1) designate the number of devices in parallel (i.e. multiplication factor), the total drawn gate width and the drawn gate length, respectively. • Vt_lin: Threshold voltage defined as Vgs value for which drain current is 40e9 A/sq*M*W/L at Vds = 0.05 V; • Vt_sat: Threshold voltage defined as Vgs value for which drain current is 40e9 A/sq*M*W/L at Vds = 1.2 V; 77 An Input Amplifier for Body-Channel Communication • llin: Drain current at Vgs = 1.2 V, Vds = 0.05 V; • Isat: Drain current at Vgs = 1.2 V, Vds = 1.2 V; • Ioffsat: Drain current at Vgs = 0 V, Vds = 1.2 V; • Ig_on: Gate current at Vds = 0 V and Vgs = 1.2 V; • Ig_off: Gate current at Vds = 1.2 V and Vgs = 0 V; • Cgg_inv: Total gate capacitance at Vgs = 1.2 V, Vds = 0 V, f = 100 kHz; • Cggmean: Average total gate capacitance for Vgs values between 0 V and 1.2 V, Vds = 0 V, f = 100 kHz; • Cgd_0 V: Gate-to-Drain capacitance at Vgs = 0 V, Vds = 0 V, f = 100 kHz; • Cbd_off: Bulk-to-Drain capacitance at Vgs = 0 V, Vds = 0 V, f = 100 kHz; • Gm_c: Drain transconductance at Vgs = Vt_lin + 0.2 V, Vds = 0.6 V, f = 100 kHz; • Gd_c: Drain conductance at Vgs = Vt_lin + 0.2 V, Vds = 0.6 V, f = 100 kHz; • Gain_c: Voltage gain defined as Gm_c/Gd_c. (d)Resistor Models: In the CMOS065 standard, resistors are modeled with the concern of different functionnalities like intrinsic resistor value, bias/temperatures dependencies, noise, parasitic capacitances, matching etc. So, by the concern of these functionnalities different models are presented which are stated below: • Unsilicided P+ Active Resistor (RPODRPO) • Unsilicided P+ Poly Resistor (RPPORPO) • Unsilicided N+ Poly Resistor (RNPORPO) • Silicided N+ Poly Resistor (RNPO) • Unsilicided High Value Poly Resistor (RHIPORPO) In this paper Unisilicided P+ Poly Resistor (RPPORPO) is used for the simulation. 78 13 An Input Amplifier for Body-Channel Communication Appendix C Appendix C mainly focus on defining the conventional amplifier and two stage CMOS OTA. Section 13.1 describes the disadvantage of using the conventional amplifier in the lower supply voltage system. Section 13.2 defines two stage CMOS OTA structure and also calculates the gain of this structure. It also defines the frequency response of two stage OTA. 13.1 Preamplifier design There are many design technique present that can use as a preamplifier. For example, differential pair is the most conventional open loop topology. In the differential pair, different load technique can be possible such as diode connected, resistive load, capacitive load etc. Diode connected load can not be used in the low-supply voltage system. So, resistive load or capacitive load can give better results with low supply voltage system. For improving the linearity issue, degenerative resister can give better solution. It is also important to look at the output voltage swing of the circuit because there is a trade off between output swing and linearity. Figure 43: A conventional amplifier with resistive load Figure 43 represents a conventional amplifier with resistive load. It can also possible with the capacitive load. Here Rload is the load resistance and Rs is the degenerative resistance. The approximate gain of this circuit is Av =− Rload 1 ( +R s) g m1 (14) As the transconductance of transistor M1 is dependent upon on the current I1, so g m1 can be calculated as follows W g m1 =μ C ox (V GS1 −V THN ) (15) L 79 An Input Amplifier for Body-Channel Communication From the equation 15, it is clear that when the input signal change, then the gain will increase. But in that case linearity performance will go down. To improve the linearity, it needs enhancing gm1 which also means increase the current through transistor M1 and M2. As the current increase, the output swing will also decrease. Therefore, there is a trade off between power, output swing and linearity. So, it is difficult to get acceptable performance for the low supply voltage system. 13.2 Two stage CMOS OTA Figure 44: Two stage OTA with biasing circuit and output buffer Figure 44 indicates two stage OTA which is also attached with a biasing circuit in input side and a buffer circuit in the output side. The first stage is the differential input stage which ended with a output stage. The second stage is a common-source gain that has a active load. The active load is an output buffer. The first stage uses p-channel differential input pair with an n-channel current-mirror active load. Capacitor Cc is used to as a feedback to ensure the stability. Two stage OTA has many advantages over single stage OTA. Two stage OTA reduces interaction between gain and output range. So, it has higher gain and improved output voltage swing. It has also reduced input capacitance Cin potentially. It has also some disadvantages. It has needed compensation for stability. It has reduced speed as well. Different compensation techniques can give solution for this problem. For example, introducing a dominate pole can give easy solution for this problem as well as it will improve the phase margin to maintain the standard of linearity. 80 13.2.1 An Input Amplifier for Body-Channel Communication OTA gain In this section 13.2.1, two stage OTA gain is calculated. The two stage OTA consists of two different parts: differential first stage, output second stage and output buffer. The gain of all three parts are also defined in this section. The first stage gain is Av1 = g m1 ( r ds2∥r ds4 ) (16) To explain more in equation 16, gm1 is given as follows √ W ) I L 1 D1 W I bias = 2 μ p C ox ( ) L 1 2 g m1= 2 μ p C ox ( Output impedance rdsi is defined by √ r dsi = α Li √ V DGi+V ti I Di In equation 18, α is a technology-dependent parameter of around 5X106 sqrt(V/m). Gain at the second stage is Av2 =− g m7 (r ds6∥r ds7 ) (17) (18) (19) Gain of the output buffer is Av3 = g m8 (G L +g m8 +g ds8 + g ds9 ) (20) Equation 20 defines output buffer where GL is the load conductance. The total gain of the two stage OTA is AV = Av1 A v2 Av3 13.2.2 (21) Frequency response By using Miller's theorem, equivalent load capacitance can be written as C eq = C c (1+ Av2 ) ≈C c Av2 By doing the small-signal model for the first stage of the opamp gain, it can be written as V1 Av1 = V input =−g m1 Z out1 To explain more on equation 23, Zout1 can be written as 1 Z out1 = r ds2∥r ds4∥ sC eq (22) (23) (24) 81 An Input Amplifier for Body-Channel Communication For midband frequencies, the impedance of Ceq dominates, so it can be written as 1 Z out1 ≡ sC eq 1 ≡ ( sC c Av2 ) (25) So the overall gain can be written as V out V input ≡ Av1 Av2 Av3 g m1 (26) ≡ Av2 Av3 ( sC c Av2 ) g = m1 sC c To find the unity-gain frequency, it is needed to set ∣Av ( j ω ta )∣= 1 and the unity-gain frequency can be written as g ωta = m1 (27) Cc Av ( s)≡ From the equation it is cleared that unity-gain frequency(ωta) is directly proportional to gm1 and inversely proportional to Cc.