Connecting the human body - Models, Connections and Competition

(1)

Institutionen för systemteknik

Department of Electrical Engineering

Examensarbete

Connecting the human body - Models, Connections

and Competition

Examensarbete utfört i Elektroniksystem vid Tekniska högskolan vid Linköpings universitet

av

Kiran Uliveppa Kariyannavar (kirka696@student.liu.se) LiTH-ISY-EX--11/4505--SE

Linköping 2011

Department of Electrical Engineering Linköpings tekniska högskola

Linköpings universitet Linköpings universitet

(2)

(3)

Connecting the human body - Models, Connections

and Competition

Examensarbete utfört i Elektroniksystem

vid Tekniska högskolan i Linköping

av

Kiran Uliveppa Kariyannavar (kirka696@student.liu.se) LiTH-ISY-EX--11/4505--SE

Handledare: J Jacob Wikner

isy, Linköpings universitet

Jan Hederen

Ericsson AB

Examinator: Oscar Gustafsson

isy, Linköpings universitet

(4)

(5)

Avdelning, Institution Division, Department

Division of Electronics Systems Department of Electrical Engineering Linköpings universitet

SE-581 83 Linköping, Sweden

Datum Date 2011-08-18 Språk Language Svenska/Swedish Engelska/English Rapporttyp Report category Licentiatavhandling Examensarbete C-uppsats D-uppsats Övrig rapport

URL för elektronisk version

http://www.es.isy.liu.se

ISBN — ISRN

LiTH-ISY-EX--11/4505--SE Serietitel och serienummer Title of series, numbering

ISSN —

Titel

Title Connecting the human body - Models, Connections and Competition

Författare Author

Kiran Uliveppa Kariyannavar (kirka696@student.liu.se)

Sammanfattning Abstract

Capacitive communication using human body as a electrical channel has at-tracted much attention in the area of personal area networks (PANs) since its introduction by Zimmerman in 1995. The reason being that the personal infor-mation and communication appliances are becoming an integral part of our daily lives. The advancement in technology is also helping a great deal in making them interesting,useful and very much affordable. If we interconnect these body-based devices with capacitive communication approach in a manner appropriate to the power, size, cost and functionality, it lessens the burden of supporting a communi-cation channel by existing wired and wireless technologies. More than that, using body as physical communication channel for a PAN device compared to traditional radio transmission seems to have a lot of inherent advantages in terms of power and security etc. But still a lot of feasibility and reliability issues have to be ad-dressed before it is ready for prime time. This promising technology is recently sub-classified into body area networks (BAN) and is currently under discussion in the IEEE 802.15.6 Task Group for addressing the technical requirements to un-leash its full potential for BANs. This could play a part in Ericsson’s envision of 50 billion connections by 2020. This thesis work is part of the main project to investigate the models, interface and derive requirements on the analog-front-end (AFE) required for the system. Also to suggest a first order model of the AFE that suits this communication system.

In this thesis work the human body is modeled along with interfaces and transceiver to reflect the true condition of the system functioning. Various re-quirements like sensitivity, dynamic range, noise figure and signal-to-noise ratio (SNR) requirements are derived based on the system model. An AFE model based on discrete components is simulated, which was later used for proof of concept. Also a first order AFE model is developed based on the requirements derived. The AFE model is simulated under the assumed interference and noise conditions. The first order requirements for the submodules of the AFE are also derived. Future work and challenges are discussed.

Nyckelord

Keywords Capacitive Communication, Intra-Body Communication, Body Coupled Commu-nication, Analog Front End, Body Area Networks, Near Field Communication

(6)

(7)

Abstract

Capacitive communication using human body as a electrical channel has at-tracted much attention in the area of personal area networks (PANs) since its introduction by Zimmerman in 1995. The reason being that the personal infor-mation and communication appliances are becoming an integral part of our daily lives. The advancement in technology is also helping a great deal in making them interesting,useful and very much affordable. If we interconnect these body-based devices with capacitive communication approach in a manner appropriate to the power, size, cost and functionality, it lessens the burden of supporting a communi-cation channel by existing wired and wireless technologies. More than that, using body as physical communication channel for a PAN device compared to traditional radio transmission seems to have a lot of inherent advantages in terms of power and security etc. But still a lot of feasibility and reliability issues have to be ad-dressed before it is ready for prime time. This promising technology is recently sub-classified into body area networks (BAN) and is currently under discussion in the IEEE 802.15.6 Task Group for addressing the technical requirements to unleash its full potential for BANs. This could play a part in Ericsson’s envision of 50 billion connections by 2020. This thesis work is part of the main project to investigate the models, interface and derive requirements on the analog-front-end (AFE) required for the system. Also to suggest a first order model of the AFE that suits this communication system.

In this thesis work the human body is modeled along with interfaces and transceiver to reflect the true condition of the system functioning. Various re-quirements like sensitivity, dynamic range, noise figure and signal-to-noise ratio (SNR) requirements are derived based on the system model. An AFE model based on discrete components is simulated, which was later used for proof of concept. Also a first order AFE model is developed based on the requirements derived. The AFE model is simulated under the assumed interference and noise conditions. The first order requirements for the submodules of the AFE are also derived. Future work and challenges are discussed.

(8)

(9)

Acknowledgments

Firstly, I would like to thank Jan Hederen and Ericsson AB, for this exciting thesis opportunity.

I would like to thank my supervisor Dr. J Jacob Wikner, Senior Lecturer at the Electronics Systems department, for his thoughtful discussions and support. His critique and guidance has added substantial value to the work. He has always been a source of inspiration.

Many thanks to the ConnectedUs teammates Dilip and Bibin for the brain-storming and fun times.

Thanks to my peers Vishnu and Dinesh in the lab for making it pleasure to work at the lab.

I am deeply grateful for unconditional love and support by my parents. An extra special thank you goes to my brother who has always been so sup-portive in many respects during my stay away from home.

(10)

(11)

Introduction

1.1 Background

Human body capacitive communication deals with using the human body as a elec-trical channel for communication. Since its introduction by Zimmerman [16] in 1995 it has attracted much attention in the area of personal area networks (PANs) . Due to the advancement in technology, electronic devices are becoming innovative, entertaining and affordable. They have become an integral part of our daily life. If we can interconnect the various body-based devices in a manner appropriate to the power, size and cost then we lessen the burden of supporting a communica-tion channel by existing wired and wireless technologies. If one body-based device is capable of communicating to all other body-based devices and to an external device attached to a wired/wireless network, then all the body-based devices can communicate to the outside world. Using body as physical communication channel compared to traditional radio transmission seems to have a lot of inherent advan-tages in terms of power and security. To address the feasibility, reliability and technical development issues IEEE 802.15.6 Task Group is discussing this commu-nication system under the sub-classification as body area networks (BAN). This could play a major part in Ericsson’s envision of 50 billion connections by 2020.

So Ericsson came up with three thesis topics as shown in table 1.1, which are connected parts of the main project to investigate the opportunity.

Table 1.1. SubProjects. Sl No SubProject(Thesis number)

1 The Connected Me - Proof of concepts (LiTH-ISY-EX--11/4504--SE) 2 Connecting the human body - models, connections, competition

(LiTH-ISY-EX--11/4505--SE)

3 Connected Me : Hardware for high speed BCC (LiTH-ISY-EX--11/4503--SE)

This particular thesis topic is subproject2. Throughout the discussion further 3

(18)

4 Introduction we refer to subproject1 and subproject3 as they are strongly connected subprojects.

1.2 System description

The fundamental idea of the main project "ConnectedMe" is to use the body as a communication link as shown in figure 1.1.

The project development involves designing a digital baseband which takes care of modulation and demodulation of the data. For example initially field programmable gate array (FPGA) can be used for baseband processing with any suitable communication protocol. Then off-the-shelf components can be used for transmitter and receiver which is interfacing the human body channel using some electrode. Prove the concept in hardware. Also we need to model the body suitable to verify the loop of communication and derive requirements for AFE. Later a first order AFE model suitable for this communication need to suggested based on the study of the human body channel.

Figure 1.1. Main project top level system description.

1.3 Purpose and goals of this thesis work

The scope of this particular subproject is to do detailed analysis of the prerequi-sites with respect to the communication baseband, i.e, to model the human body communication system in the form of usage and derive the requirements for AFE based on the system model. Check what kind of software needs to be used so that the software top level loop can be closed easily. Create a AFE model which can be used in the demo model for proof of concept. Reliability of the human body link has to be studied. Finally suggest a first order model of AFE suitable for the system based on the requirements derived.

(19)

1.4 Outline 5

1.4 Outline

The work presented in this report is outlined in this section. In section 2 in-troduction to human body communication is made. The previous work done on modeling the human body and various AFE developed are examined. The base-band modulation and electrode related literature study is done in subproject1 and subproject3. In section 3 we study the whole system modeled in the form of usage. We will see quantitatively how different factors of the system affect the performance. Section 3.5 shows the correlation between initial practical measure-ments and simulated response with tuning the parameters in model accordingly. AFE requirements are derived in section 4. The modeling and simulation results of first order model of the AFE for proof of concept is shown in section 5. Later, in section 6 a prospective AFE model is examined for its usability in the system. We restrict the implementation to first order models to prove the concept. Finally we conclude the thesis report in section 7 discussing future work and challenges along with conclusion of this thesis work.

1.5 Goals that were met

The human body capacitive communication system was modeled in cadence re-flecting the true condition of the system functioning, with human body model, interfaces with the AFE and environment. The strong influence of few factors in the system compared to others were shown. The first order requirements on the AFE were derived in terms of sensitivity, dynamic range , noise figure and signal-to-noise ratio. The best and worst case conditions of the system and in-terference in the band of interest were taken into consideration while deriving the requirements for AFE. The measured results could be correlated with the sim-ulation results after tuning few parameters in the system. The prototype AFE built using off-the-shelf components supporting Manchester encoding was mod-eled and simulated for gaining confidence in building the hardware along with its usage to evaluate the digital modulation/demodulation method. It was simulated to achieve 15 Mbps data rate but due to hardware limitations it could reach 12 Mbps data rate. A prospective AFE working on the principle of using pulse based communication in a narrowband sense was evaluated by using first order models. Two bands were considered for 10 Mbps and 40 Mbps operation and simulated to have a BER of less than 10−5 _{and around 10}−3 _{with noise and interference in the}

(20)

(21)

Chapter 2

Literature study and review

2.1 Human body communication variants

Capacitive coupling and galvanic coupling are two conceptually different approaches to use body for data communications. These are explained in subsections 2.1.1 and 2.1.2.

2.1.1 Capacitive coupling

First introduced by Zimmerman in the context of human body communication [16]. As shown in the figure 2.1, we see the electric fields that exist between the human body ,earth ground, transmitter (Tx) and receiver (Rx) electrode indicated by the arrows. These electric fields exist because of the difference in the potential in all these surface. If we can modulate the fields at the transmitter by varying the potential at the transmitter electrode and detect the variation of potential at the receiver electrode, we can potentially transfer information through the body. So this kind of information transmission by electro-statically (capacitively) coupling currents into the body is called body coupled communication. We will look at the electrical model of this concept later in section 2.3.

2.1.2 Galvanic coupling

In comparison to capacitive coupling applying static charged electrode, galvanic coupling provides alternating currents over multiple electrodes [14]. Coupling of alternating current into the human body is direct. It uses differential coupler electrodes at the transmitter to establish a modulated electric field which is sensed by the detector using differential electrodes. So basically transmission between the coupler and detector units happens by coupling signal currents galvanically into the human body. Also the frequency range of operation is quite low (10 kHz to 1 MHz) [14].

So throughout the discussion we refer to capacitive coupling system or mecha-nism by the name human body communication (HBC) , intra body communication

(22)

8 Literature study and review

Tx _Rx

Earth ground

Figure 2.1. Near field communication electric field.

(IBC) or body coupled communication (BCC) etc.

2.2 Motivation to go for human body

communi-cation

The major motivation to use human body as channel compared to existing wireless communication systems is its inherent security and low power aspect. Inherent security comes from the fact that we use electrostatic fields and that too near field regions for operation whose strength drastically falls off with distance cubed. Even if we use larger electrodes to create stronger fields they would be still be inefficient in radiating the signal as they would be very small compared to wavelength of the signal. This inefficient radiation also means that we are not polluting the congested electromagnetic spectra [16].

Low power aspect in comparison to traditional radio frequency (RF) transmis-sion is due to the fact that we need not drive a very small impedance which is a major source of power consumption. The small impedance in RF transmission is due to the requirement of matching the impedance to free space impedance for maximum power transmission. In human body capacitive communication we have large load of the body and to maximize the coupling larger electrode plates are used [16].

(23)

2.3 Human body model for capacitive communication 9 is capable of communicating to all other body-based devices and to an external device attached to a wired/wireless network, then all the body-based devices can communicate to the outside world reducing the burden of supporting a commu-nication channel by existing wired and wireless technologies. Depending on the channel capacity supported these interconnected body-based devices can share computational resources, either performing distributed computation or relying on neighbors with more specialized and higher capacity processing power to perform functions too intensive for the resident processor [16].

2.3 Human body model for capacitive

communi-cation

Some of the existing electrical models used for human body capacitive commu-nication are examined in this section. Zimmerman in his paper treated body as a simple node and came up with a lumped-circuit electrical model as shown in 2.2 [16]. The various capacitance modeled are explained in table 2.1.This electri-cally describes the near-field coupling mechanism around the human body.

T R Cb Cc Ca Cd Cf Ce Ch tb te rb re Body node Cg

Figure 2.2. Zimmerman HBC model.

Since frequency range of kHz is used, impedance of the return path is very large compared to that of the body channel and this model is quite appropriate.

In 2007 Cho, et al. [1] developed a distributed resistor-capacitor (RC) model to analyze the signal transmission through the human body at higher frequencies up-to 150 MHz and 1.2-m length along the body. They claim that as the transmission length of the body channel increases, both the resistance of the body and the coupling capacitance to the external ground increase and these elements cause signal loss at the receiver, and its amount depends on the channel length. The model consists of the cascaded unit blocks with an RC parallel network and shunt

(24)

Table 2.1. Description of various coupling factors in Zimmerman’s lumped-circuit

elec-trical model.

Model component Description

Cd Transmitter ground electrode te to ex-ternal ground coupling capacitance Cg Receiver ground electrode te to external

ground coupling capacitance

Cf Receiver body electrode to body cou-pling capacitance

Cc Transmitter body electrode to body coupling capacitance

Ch Coupling capacitance between receiver electrodes

Ca Coupling capacitance between trans-mitter electrodes

Ce Coupling capacitance between body and earth ground

Cb Coupling capacitance between the transmitter environment electrode te and the body

capacitors, which model the electrical couplings across the human body. This seemed to a more logical model developed so far. It could be used for various lengths across the body and also in the system for designing the AFE. We shall see details in section 3.1 as we use this model in our work. The radiation effect limitations in terms of frequency and transmission power has been discussed in the same paper [1].

In the same year as previous paper Song, et al. [11] showed that human body behaves as a bandpass filter with a bandwidth of about 100 MHz and shows approximately 6-dB attenuation. But the measurement was based on 15 cm across the arm which is one small condition in which we would want the system to work. Recently finite-element method (FEM) is utilized to investigate the Electric-field intra-body communication channel [15]. The FEM investigation finds that the capacitive return path is critical to the characteristics of the EF-IBC (Electric Field Intra-body-communication ) channel but the same conclusion was drawn from the other previous works.

2.4 Analog front end architectures

There have been a few AFE architectures developed for body coupled communi-cation. In other words different AFE’s based on several modulation schemes have been tried for this communication.

The HBC transceiver firstly introduced by Zimmerman [16] has a very low data rate of 2.4-kb/s. It uses a single carrier frequency of 330-kHz. It was anyway good

(25)

2.4 Analog front end architectures 11 enough to introduce the concept. Later on, HBC receiver developed by Shinagawa, et al. [9] used a special electro-optic sensor to achieve 10-Mb/s data rate. But it consumes too high power to apply for human body applications. The usage of two electrodes for signal and ground might be very inconvenient. Nevertheless it proved intra-body communication at 10 Mbps for a distance of 150 cm between hands. Also the scenario of two persons handshake and intra-body communication through clothes were confirmed.

Korea Advanced Institute of Science and Technology (KAIST) has been onto the research very much. An energy-efficient wideband signaling receiver AFE exploiting wideband symmetric triggering technique was presented for the hu-man body as a data transmission medium [10]. The AFE implemented in 0.18-Î¼m MOS process delivers 10 Mbps data rate with -27 dBm sensitivity, occupies 0.04 mm2 and consumes 4.8 mW power. But all measurements were conducted only between the wrist and the fingertip that corresponds to the distance of about 25 cm and also it does not consider interference effect at all.

For the first time an AFE which considered interference was proposed by Cho, et al. [2]. It was also scalable from 60-kb/s to 10-Mb/s operation. The body antenna effect which interferes with signals in the human body channel was exam-ined. In order to overcome the body antenna effect, a 4-channel adaptive frequency hopping scheme using the 30-120 MHz band was introduced to the body channel transceiver. Measurements are done for 1.8 m operating distance with better sen-sitivity of -65 dBm and bit-error-rate (BER) less than 10−5_{. This is not suitable}

for simultaneous multi-channel transmission and also is quite complex in architec-ture. For measurements transceiver modules are attached to left and right hands with Ag/AgCl electrodes and the two arms are stretched out. The electrodes are not spaced away from direct contact to the body, which is a general scenario of usage. The transmitter module is powered by a battery to make external ground coupling. The receiver module is connected to a laptop through a serial cable. The usage scenario of both transmitter and receiver being battery powered is not tested.

Another low-power wideband signaling(WBS) digital transceiver for data trans-mission through a human body was developed by Song, et al. [11] with a 0.25 um standard CMOS technology. Although it has 2 Mbps data rate at a BER of less than 10−7_{. Again interference is not considered and coupling condition to the}

body is not adequate.

Jin, et al. [5] proposed another low power wearable transceiver for human body communication. It claims to acquire high data rate up to 15-Mb/s in 1-2 m nominal communication range. It was implemented in 0.18 um consuming no more than 5 mW from a 1.8-V supply. The total sensitivity as a single-chip is -30 dBm (f0=10 MHz). Even though it is much low power and high speed than Cho, et al. [2], again this scheme does not consider interference. The measurement scenarios are also not shown.

(26)

2.5 Conclusion

We saw that most of the AFE architectures are built and tested in one condition of operation. This leaves a doubt on their operation in other condition. Also most of them do not consider interference while designing the AFE. So there is a need to study the channel characteristics systematically and develop the AFE from conceptual level considering all typical scenarios of operation and interference. Also most of the research work done does not clearly show the complete system setup in which the simulation is done.

(27)

Chapter 3

Modelling and simulation of

human body for BCC

The modeling of human body communication system in the form of usage is de-scribed and the simulation results for different practical conditions are shown in this section. The impact of various factors in the system affecting the channel response is shown. Both the frequency domain and time domain response are analyzed.

3.1 The distributed RC human body model

From the literature study on the human body models we found that Cho, et al. [1] model was most appropriate in comparison with other models to model the internal structure of the human body electrically. So in our work we model the human body similarly as distributed RC along arm and torso as shown in figure 3.1 [1]. There is a very small addition to this model, i.e, the explicit feet capacitance. This feet capacitance to external ground can be high due to larger surface area and normally we are not in direct contact with earth ground. So we included this and the impact of this is analyzed later. The blocks named as ’A’ and ’T’ stand for arm and torso blocks respectively. They are cylindrical approximation to arm and torso segments. ’Xa’ and ’Ya’ are height and diameter of each arm segment. Similarly ’Xt’ and ’Yt’ are diameter and height of each torso segment. ’A’ and ’T’ blocks consists of parallel resistance and capacitance and coupling to external ground from that node of the body, as shown in Cadence model in figure 3.2. pPar in the figure 3.2 means its parameter value is inherited from parent schematic. This is useful because these blocks are instantiated many times and they can be assigned to a new variable, for example Ca and Ct for parallel capacitance in arm and torso blocks respectively. In figure 3.2 Vin and Vout would represent input and output nodes of a particular segment ,and Gnd is external ground node. The height (h) and arm-to-arm (w) distance is quite average and arbitrary .The values of the capacitance and resistance are given and explained in section 3.2.

(28)

14 Modelling and simulation of human body for BCC

Figure 3.1. Human body model.

(29)

3.2 Simulation model including transceiver and ground coupling 15

3.2 Simulation model including transceiver and

ground coupling

Figure 3.3 shows the model of the human body communication system in the form of usage. The whole model is implemented using standard components from Cadence libraries. "bodyLinkModel_Distributed_RC" block is modeled as in fig-ure 3.1. The transmitter and receiver are modeled as ports "Port1" and "Port2" respectively, which would be helpful further in analysis of the system in terms of S-parameters and Z-parameters. The various capacitance and resistance involved in the model are listed along with the values and description in table 3.1.

Figure 3.3. Human body model including external conditions.

The electrode to body coupling capacitance (Rx_HB_Cc and Tx_HB_Cc) and transciever ground electrode to external ground coupling capacitance (Rx_Gnd _Cc and Tx_Gnd_Cc ) values are taken from the measurements done by Zim-merman [16]. The coupling capacitance between transmitter/receiver electrodes (Rx_Self_Cc and Tx_Self_Cc ) and resistance ( R_Rx and R_Tx ) model the bandwidth of the transmitter and receiver. The explicit feet cap ( Explicit_Feet _Cap ) is given a nominal value of 50 pF which is in between the parallel body capacitance value of the arm and torso units, which is later varied to analyze its impact. Receiver and transmitter ground to external ground coupling capaci-tances (Rx_Gnd_Cc and Tx_Gnd_Cc) are again typical values taken from Zim-merman’s work [16] and transceiver coupling capacitance Tx_Rx_Cc is given a typical value of 100f as seen in Cho, et al. [1].

(30)

Table 3.1. Component values used in the simulation. Component Value Description

Ca 48 pF Parallel body capacitance value of the arm unit block

Ra 60 Ω Parallel body resistance value of the arm unit block

Ct 310 pF Parallel body capacitance value of the torso unit block

Rt 7 Ω Parallel body resistance value of the torso unit block

Cct 5 pF Coupling capacitance to external ground from torso unit block

Cca 2 pF Coupling capacitance to external ground from torso unit block

Explicit_Feet 50 pF Coupling capacitance to external ground from

_Cap feet

Tx_Rx_Cc 100 fF Transceiver coupling capacitance

Rx_Gnd_Cc 10 fF Receiver ground to external ground coupling ca-pacitance

Tx_Gnd_Cc 10 fF Transmitter ground to external ground coupling capacitance

Rx_HB_Cc 10 pF Receiver electrode to body coupling capacitance Tx_HB_Cc 10 pF Transmitter electrode to body coupling

capaci-tance

Rx_Self_Cc 50 pF Coupling capacitance between receiver electrodes Tx_Self_Cc 50 pF Coupling capacitance between transmitter

elec-trodes

R_Rx 50 Ω Receiver series resistance R_Tx 50 Ω Transmitter series resistance

The resistance (Ra and Rt) and capacitance (Ca and Ct) values of the unit blocks used in human body model are based on the experimental results of Gabriel’s work on electrical properties of the human body tissues [3]. The conductivity (σ) and relative permittivity (ε) varies from 0.1-1 S/m and 50-100 respectively.The arm and torso block are assumed to be isotropic and having equipotential cross sections [1]. The resistance and capacitance values are obtained from the simple equations 3.1 and 3.2, where d and L are the diameter (Ya,Xt) and height (Xa,Yt) of a unit block as shown in fig 3.1.The symbol ’’ is the well known, permittivity of free space.

R= (4 · L)/(σ · π · d2) (3.1) C= () · (ε) · π · d2/(4 · L) (3.2)

(31)

3.3 Frequency domain behaviour 17

3.3 Frequency domain behaviour

3.3.1 Nominal frequency response : 120 cm arm-arm

Throughout this section we use scattering parameters or S-parameters (Sparam) to describe the electrical behavior of the human body communication model. Fig-ure 3.4 shows the magnitude of gain (S21) in dB (Decibel) scale from port1 to port2 from Sparam analysis done in Cadence.The term "nominal" is used to high-light the fact that this could be a nominal practical case of usage compared to another case called "grounded" coming up later in section 3.3.5, where we have the transciever electrodes grounded. The grounded scenario is also the best case scenario for the system.

10

5

10

6

10

7

10

8

10

9

−110

−100

−90

−80

−70

−60

−50

Frequency [hz]

S21 [dB]

S21 results for arm−arm (120cm) simulations

Figure 3.4. S21 results : 120cm , nominal arm-arm.

The S21 result shows high pass characteristics due to capacitive return path and series capacitance introduced by electrode-body interface . Frequencies above 5 MHz seems to be suitable for capacitive communications. The later bending in the curve is because of the bandlimited transceivers. It shows a slight deviation from what is observed in Cho, et al. [1] which might be because of the model difference in terms of the setup external to the human body model. It is not quite clear how

(32)

18 Modelling and simulation of human body for BCC exactly the human body model is connected to external transceivers in Cho, et al. [1]. But the simulation result follows the same trend and similar numbers. We simulate till 100 MHz as the body would tend to radiate the signal more beyond this frequency beating the sole purpose of going for this communication channel [1].

3.3.2 Impact of return path variations

Figure 3.5 shows the variation in gain due to variation in reverse coupling capaci-tance betweeen the transmitter and receiver electrode. The coupling capacicapaci-tance is varied from 10 fF to 100 pF. The lower value of capacitance is from Cho, et al. [1] and higher value is just a typical value to account for better reverse cou-pling condition. The huge variation of 50 dB shows the that reverse coucou-pling in the system is probably the most important factor to consider.Unfortunately the only way to have a better reverse coupling is larger electrodes which might not be suitable for some body based applications.

105 106 107 108 109 −140 −120 −100 −80 −60 −40 −20 Frequency [hz] S21 [dB]

S21 results for variations in the transceiver ground coupling

Tx−Rx−Cc=10f Tx−Rx−Cc=100f Tx−Rx−Cc=1p Tx−Rx−Cc=10p Tx−Rx−Cc=100p

(33)

3.3.3 Electrode to body coupling variations

The next most important factor in the system is the coupling between the trans-mitter and receiver body electrode to body.This capacitance is the total of series capacitance formed due to the air and skin as dielectric. So the weaker capacitance rules, which is the capacitance formed by air as dielectric. So whenever we touch the transceiver electrode we would definitely have better capacitance and better response. This is the case with most of the measurements done in the existing literature. Figure 3.6 shows around 40 dB variations in gain for the typical varia-tions in the coupling capacitor. In this figure we vary the capacitance on both the transmitter and receiver side equally. We also see that varying the capacitance at one end alone would produce similar result. This is confirmed by the figure 3.7 and 3.8 which show same variations. So even if there is a weak link on one side it would affect the system performance.

105 106 107 108 109 −140 −130 −120 −110 −100 −90 −80 −70 −60 −50 Frequency [hz] S21 [dB]

S21 results for same variations in coupling of Tx and Rx with HB

Tx−Rx−HB−Cc=10f Tx−Rx−HB−Cc=100f Tx−Rx−HB−Cc=1p Tx−Rx−HB−Cc=10p Tx−Rx−HB−Cc=100p

(34)

20 Modelling and simulation of human body for BCC 105 106 107 108 109 −130 −120 −110 −100 −90 −80 −70 −60 −50 Frequency [hz]

S21 [dB]

S21 results for variations in the Tx to HB coupling

Tx−HB−Cc=10f Tx−HB−Cc=100f Tx−HB−Cc=1p Tx−HB−Cc=10p Tx−HB−Cc=100p

Figure 3.7. S21 results : Variations in the Tx to HB coupling.

105 106 107 108 109 −130 −120 −110 −100 −90 −80 −70 −60 −50 Frequency [hz]

S21 [dB]

S21 results for variations in the Rx to HB coupling

Rx−HB−Cc=10f Rx−HB−Cc=100f Rx−HB−Cc=1p Rx−HB−Cc=10p Rx−HB−Cc=100p

(35)

3.3.4 Impedance seen from transmitter and receiver

Figure3.9 shows the variation of impedance looking into the body channel from transmitter and receiver. At low frequencies load impedance is much higher than the source impedance .This could be better for voltage mode operation of the system maximizing voltage transfer. As we go higher in frequency above 10 MHz it becomes comparable. Z11 and Z22 in the figure stands for magnitude of load impedance seen by port1 and port2. They are equal because of the symmetry.

105 106 107 108 109 0 5 10 15 20 25 30 Frequency [hz]

Zparam value [Kohms]

Z11 Z22

Figure 3.9. Impedance variation across the frequency band.

3.3.5 The grounded scenario

This scenario is close to the measurement scenario with instruments which are earth-grounded. The transceiver grounds will be shorted in this case as shown in figure 3.10. This is the best reverse coupling or connectivity that can be achieved in the system. Also this is quite an ideal case and not possible practically to have such a condition. The magnitude of gain (S21) in the frequency band of interest for this scenario is shown in figure 3.11. The signal path loss is very less compared to the nominal case as in figure 3.4 as expected. In the later subsections some results are compared with this and nominal scenario . Also this case is used as the best case scenario for AFE requirement analysis later.

(36)

Figure 3.10. Human body model TB : Grounded scenario.

105 106 107 108 109 −90 −80 −70 −60 −50 −40 −30 Frequency [hz]

S21 [dB]

S21 results : arm−arm (120cm) : grounded scenario

(37)

3.3.6 Body internal impedance variations effect

In the section we see the effect of body internal impedance variations on the system response. The resistance and capacitance of the distributed human model are var-ied based on the variation of the conductivity and resistivity of the human tissues [3].The conductivity and resistivity vary from 0.1-1 S/m and 50-100 respectively. We see from figure 3.12 that the variation is around maximum 8 dB in grounded scenario. Figure 3.13 shows merely any variations. So we can conclude that body variations do not play much part when return path is not strong which generally would be the case of usage. This also shows that deeper modeling of the human body based on skin,fat,muscles,bone and different parts etc would not add much value in terms of AFE requirement analysis. The histogram of the body RC values is shown in figure 3.14.

10

5

10

6

10

7

10

8

10

9

−90

−80

−70

−60

−50

−40

−30

−20

Frequency [hz]

S21 [dB]

S21 results for variations in body RC : Grounded

(38)

24 Modelling and simulation of human body for BCC 105 106 107 108 109 −110 −100 −90 −80 −70 −60 −50 Frequency [hz]

S21 [dB]

S21 results for variations in body RC : Nominal

Figure 3.13. S21 results : Body model RC variations : Nominal scenario.

0 5 10 15 0 10 20 Uniformly Distributed Rt1 0 50 100 150 0 5 10 15

Uniformly Distributed Ra1

0 0.5 1 1.5 x 10−9 0 10 20 Uniformly Distributed Ct1 0 0.5 1 1.5 x 10−10 0 5 10 15

Uniformly Distributed Ca1

0 0.2 0.4 0.6 0.8 1 x 10−11 0 5 10 15 Uniformly Distributed Cct1 0 1 2 3 4 x 10−12 0 5 10 15

Uniformly Distributed Cca1

(39)

3.3.7 Impact of feet capacitance variation

Figure 3.15 shows how the feet capacitance influences the system performance. As feet capacitance increases reaching a stage where the feet is completely grounded we see a loss of around 20 dB at 1 MHz and increasing as we move higher in frequen-cies. But in nominal case as in figure 3.16 it has almost no effect as weak reverse coupling path has already done the damage. So leakage through human body via feet has major effect only when we have strong coupling between transceiver ground electrodes. 105 106 107 108 109 −140 −120 −100 −80 −60 −40 −20 Frequency [hz] S21 [dB]

S21 results : feet capacitance variations : grounded scenario

feet−cap=1p feet−cap=1u no−feet−cap

Figure 3.15. S21 results : Feet capacitance variations : Grounded scenario.

3.3.8 Frequency response over different lengths across the

arm

The response of the system varies across the arm in grounded scenario as shown in figure 3.17. It has a better response for shorter distance as expected. But again in the normal scenario we don’t find any difference because of the weak reverse coupling path overriding the impact of variations for different distance.

(40)

26 Modelling and simulation of human body for BCC 105 106 107 108 109 −110 −100 −90 −80 −70 −60 −50 Frequency [hz] S21 [dB]

S21 results : feet capacitance variations

feet−cap=1u feet−cap=1n feet−cap=1p feet−cap=1f

Figure 3.16. S21 results : Feet capacitance variations : Nominal scenario.

105 106 107 108 109 −110 −100 −90 −80 −70 −60 −50 −40 −30 −20 −10 Frequency [hz] S21 [dB]

S21 results : different distance across the arm 10cm_nom 40cm_nom 120cm_nom 10cm_gnd 40cm_gnd 120cm_gnd

Figure 3.17. S21 results : Different distance across the arm : Nominal and grounded

(41)

3.3.9 Some practical cases simulation

Figure 3.18 shows simulation results of few practical cases of transmitter and receiver locations .Hand to pocket and pocket to feet locations could be for sure a scenario to use this system. We found that hand to pocket (HPgnd) has better response than pocket to feet (PFgnd) but only in grounded scenario.

105 106 107 108 109 −110 −100 −90 −80 −70 −60 −50 −40 −30 Frequency [hz]

S21 [dB]

S21 results : some practical cases HPnom

PFnom HPgnd PFgnd

Figure 3.18. S21 results : Some practical cases.

3.3.10 Transceiver self capacitance effects

The effect of transmitter and receiver bandwidth due to the self capacitance be-tween the body and ground electrodes is shown in figure 3.19. This is the one which gives the bandpass characteristics to the curve for the system.

3.3.11 Transceiver resistance effects

If we are looking at power mode transmission in the system the following fig-ures 3.20 and 3.21 show the reflection at port1 and port2.Though it does not show major reflection, 50-100 Ω seems be better choice for 10-100 MHz band

(42)

28 Modelling and simulation of human body for BCC 105 106 107 108 109 −110 −100 −90 −80 −70 −60 −50 −40 Frequency [hz]

S21 [dB]

S21 results : Transceiver self capacitance variations

RX−Tx−Self−Cc=1p RX−Tx−Self−Cc=10p RX−Tx−Self−Cc=25p RX−Tx−Self−Cc=35p RX−Tx−Self−Cc=50p

Figure 3.19. S21 results : Transceiver self capacitance variations.

105 106 107 108 109 −20 −15 −10 −5 0 Frequency [hz]

S11 [x10e−4 dB]

S11 results : Transmitter resistance variations

Tx−R=50Ω

Tx−R=75Ω

Tx−R=100Ω

Tx−R=150Ω

Tx−R=200Ω

(43)

3.3 Frequency domain behaviour 29 105 106 107 108 109 −20 −15 −10 −5 0 Frequency [hz]

S22 [x10e−4 dB]

S22 results : Receiver resistance variations

Rx−R=50Ω

Rx−R=75Ω

Rx−R=100Ω

Rx−R=150Ω

Rx−R=200Ω

(44)

3.4 Time domain behaviour

The time domain response of the system at various nodes in the model is analyzed in this section. Also the safety limits in terms of current into the body is examined.

3.4.1 Transient response : Nominal scenario : 120 cm

The time domain response at body input and output in the system for 10 MHz, 2 Vpp input pulse is shown in figure 3.22. The simulation is done in nominal scenario. The narrow received pulse of amplitude less than 4 mV at the receiver input is due to the phase and magnitude response of the system for different frequencies as shown in figure 3.23 and 3.4 .

2.34 2.36 2.38 2.4 2.42 2.44 2.46 2.48 2.5 −1

0 1

Transmitter voltage (Tx_Vout)

[V]

2.34 2.36 2.38 2.4 2.42 2.44 2.46 2.48 2.5 −20

0 20

Voltage at body channel model input (Vin_body)

[mV]

2.34 2.36 2.38 2.4 2.42 2.44 2.46 2.48 2.5 −20

0 20

Voltage at body channel model output (Vout_body)

[mV] 2.34 2.36 2.38 2.4 2.42 2.44 2.46 2.48 2.5 −5 0 5 Time [µsec] [mV]

Voltage at the receiver (Rx_Vin)

Figure 3.22. Transient pulse response : Nominal arm-arm (120cm) .

The rise time at the input is varied and the received pulse width is shown in figure 3.24. This shows that increasing the rise time could help in getting more time to resolve the received signal but at the cost of decreased strength.

(45)

3.4 Time domain behaviour 31 105 106 107 108 109 −50 0 50 100 Frequency [Hz]

S21 phase [deg]

Figure 3.23. S21 phase response : Arm-arm (120cm) : Nominal scenario.

0 5 10 15 20

0 10 20 30

Received pulse Width [ns]

Input pulse rise time (ns)

Reveived pulse width and amplitude Vs Input pulse rise time

0 5 10 15 200

0.5 1 1.5

Received pulse Amplitude [mV]

(46)

3.4.2 Current into the body

Figure 3.25 shows the maximum current entering the body as per the nominal simulation setup.The current is well within the limits of maximum current of 20 mA [13]. This also shows potential for low power applications compared to existing wireless technologies.The simulation setup is similar to figure 3.3 with Port1 being a sinusoidal signal source with 1 V AC amplitude.

106 107 108 0.2 0.4 0.6 0.8 1 1.2 Frequency [Hz]

Max current into body [

µ

A]

Figure 3.25. Maximum current through body versus frequency : Nominal scenario.

3.5 Initial measurements and system parameters

tuning

The correlation between initial measurements in terms of frequency and pulse response of the human body channel and system response based on tuning the model parameters is shown in this section.

3.5.1 Pulse peak-to-peak amplitude correlation

Based on the initial measurements using a ESD (electrostatic discharge) wrist strap the model is tuned to see if we can achieve similar results. The ESD strap is used on both hands stretched. The measurement is done using earth grounded pulse generator with different frequencies and measuring the amplitude in the earth grounded oscilloscope.This is similar to the grounded scenario as we saw

(47)

3.5 Initial measurements and system parameters tuning 33 previously in figure 3.10. Figure 3.26 shows the correlation between the measured and tuned simulation results.

10 20 30 40 50 12 14 16 18 20 22 24 26 28 Frequency [Mhz]

Peak−Peak Output Voltage [mV]

Peak−to−Peak recieved pulse amplitude Versus Frequency Measured Peak−Peak recieved Voltage Simulated Peak−Peak recieved Voltage

Figure 3.26. Peak to peak recieved pulse amplitude vs frequency : Measurements correlation.

3.5.2 Frequency domain behaviour

Figure 3.27 shows the magnitude response in terms of S21 parameters after tuning the simulation model. The approximated/tuned transmitter-to-body capacitance is around 100 pF. These values are later used as best case values for the dynamic range analysis.

(48)

34 Modelling and simulation of human body for BCC 105 106 107 108 109 −55 −50 −45 −40 −35 −30 −25 −20 Frequency [hz]

S21 [dB]

S21 results : Measurement with sinusoidal signals and grounded scenario

Figure 3.27. S21 results : Measurement with sinusoidal signals : Grounded scenario.

3.6 Conclusion

The human body communication system is modeled based on previous work but in a complete usage form. The various factors affecting the communication system is studied quantitatively in a setup close to the practical scenario of usage. The analysis is done in both practical scenario and ideal scenario of operation which would show how much various factors can affect the system in both scenarios and also which factor we need to consider more than others. Analysis show that return path plays a major role in the performance of the system compared to other factors. The next major impact is due to the body electrode and body coupling capacitance. Variations in body and different practical conditions are very meek when the return path is weak. So considering these facts its better to take care of the extreme cases while designing the AFE for system. Safety requirement in terms of current into the body in the frequency band of interest seem to be well within limits.The simulation model is tuned for transmitter-to-body capacitance based on initial measurements. The tuned value for capacitance is used later in the case of grounded scenario to derive requirements for AFE where this is considered the best case. The rest of the parameters are later tuned based on the electrodes used, after the printed-circuit-board (PCB) design for proof of concept and can be found in subproject3.

(49)

Chapter 4

AFE requirement analysis

The first order requirements on the AFE in terms of sensitivity, dynamic range, noise-figure and SNR are derived in this section . Results from previous work such as limits on transmitted power and interference have been used along with the system model developed. Based on the derivation an appropriate communication methodology is suggested.

4.1 Transmitter power spectral density

consider-ation

The maximum transmitted signal power to be used in the band of interest is shown in figure 4.1. This is based on the previous research to measure maximum power spectral density of the transmitted signal that does not violate the FCC regulations [1]. None of the previous works on the AFE has taken this into con-sideration explicitly. The curve shows an opposite trend compared to S21 results in figure 3.4.This is further used to calculate the sensitivity requirement in the next section 4.2.

4.2 Reciever sensitivity versus frequency

Sensitivity is one of the main specifications of any receiver . It is normally taken as the minimum input signal required to produce a specified output signal having a specified signal-to-noise (S/N) ratio. The sensitivity requirement on the AFE in the frequency band of interest is shown in figure 4.2. Based on the maximum transmitted signal power to be used in the band of interest as shown in figure 4.1 and S21 results in the band of interest as shown in figure 3.4 , the received signal strength is calculated and it could be taken as the sensitivity requirement on the receiver. It can be seen that frequencies below 10 MHz have really stringent requirement on the sensitivity. Above 10 MHz its relatively constant .

(50)

36 AFE requirement analysis 100 101 102 −10 −8 −6 −4 −2 0 2 4 Frequency [MHz] Maximum Transmitter PSD [dBm/MHz]

Maximum Transmitter PSD vs. Frequency

Figure 4.1. Maximum transmitter power spectral density versus frequency.

100 101 102 −95 −90 −85 −80 −75 −70 Frequency [MHz] Receiver Sensitivity [dBm]

Receiver Sensitivity vs. Frequency

(51)

4.3 Dynamic range versus frequency 37

4.3 Dynamic range versus frequency

The human body communication system has a lot of influential parameters which can lead to huge variations in received signal strength. So the overall dynamic range of the receiver, which is essentially the range of signal levels over which the receiver can operate needs to be evaluated. The low end of the range is governed by its sensitivity whilst at the high end it is governed by its overload or strong signal handling performance. The dynamic range needed for different frequency of operation is shown in figure 4.3. The best and worst case for measurement of the dynamic range (DR) is taken from initial measurements simulation model tuning and worser nominal values of the variation factors in the system respectively.

100 101 102 30 35 40 45 50 55 60 65 Frequency [MHz] Required receiver DR [dB] Receiver DR vs. Frequency

Figure 4.3. Dynamic range versus frequency.

4.4 Noise figure versus frequency

The requirements on noise figure (NF) of the AFE for different frequencies within the band of interest is shown in figure 4.4. The graph is plotted based on equa-tion 4.1 , where B is bandwidth in Hz , Sensitivity is known from secequa-tion 4.2 and

(52)

38 AFE requirement analysis -174 dBm is the noise floor of a receiver for 1 Hz bandwidth.

Sensitivity= −174dBm + 10logB + NF + SNR (4.1) We see that NF requirements are stringent for 80 MHz than 40 MHz . Also we have stringent requirements on NF for low frequency of operation of 2 MHz . So in terms of NF, 10-40 MHz band is good. We should have choose low order modulation schemes which have low requirement on SNR. This would end up with low requirements on NF and hence low power consumption and simpler design.

0 5 10 15 20 25 −20 −15 −10 −5 0 5 10 15 20 25 SNR [db] NF [db]

Noise figure requirement vs. SNR

NF−2MHz NF−10MHz NF−40MHz NF−80MHz

Figure 4.4. Noise figure versus signal-to-noise ratio at different frequencies.

4.5 In–band interferences due to body antenna

effect and electrode

Figure 4.5 shows the interference strength in the band of interest. This is a setback for working at higher frequencies which have better performance. The cordless phone band and FM radio interference levels are taken from the previous work [1] which are due to body antenna effect and RFID signal strength is what is measured

(53)

4.6 Conclusion 39 in the lab during the measurements. So frequency band of 5-10 MHz and 20-40 MHz is best from interference point of view.

Cordless phone Frequency range : 46-50Mhz Strength : -48 dBm FM radio Frequency range : 80-110Mhz Strength : -30 dBm Frequency 0 Mhz 120 Mhz RFID Frequency r: 13.56 Mhz Strength : -69 dBm

4.6 Conclusion

The specifications for AFE are analyzed based on the simulation model and initial measurements results. The in-band interferences due to body antenna effect and electrode in the cordless phone band, RFID and FM radio band etc seem to be a setback for selecting higher frequencies to work with. We conclude that the best frequency band of interest is 5-10 MHz and 20-40 MHz giving a requirement on NF of around 10 dB for a low order modulation scheme which would require a SNR of 10 dB and DR of 50 dB with sensitivity of -75 dBm. These requirements suggest pulse based communication with a low order modulation scheme would be a suitable for this communication. So pulse based communication is considered for further AFE implementation.

(54)

(55)

Chapter 5

AFE modeling and

verification for proof of

concept

The derived AFE requirements suggest that pulse based communication with a low order modulation scheme would be a suitable for this communication. So a suitable AFE for this kind of communication is modeled and verified for functionality along with a baseband modulation scheme. The AFE model is built based on the off-shelf components. It uses various parameters of the submodules based on the datasheets of respective components. The transient response of the model and general take cares for the digital baseband is described.

5.1 Baseband data Encoding

Manchester data encoding for baseband data encoding is selected based on the conclusions drawn from the AFE requirement analysis, since it is a low order modulation scheme. Main advantages of using this encoding scheme are simpler hardware and clock recovery circuit. Some more advantages and implementation details are discussed in subproject1.

In Manchester encoding, data bit ’0’ and ’1’ is coded as "01" and "10" respec-tively as shown in the figure 5.1. This shows that Manchester code ensures fre-quent line voltage transitions, directly proportional to the clock rate which helps clock recovery. The DC component of the encoded signal is not dependent on the data and therefore carries no information, allowing the signal to be conveyed conveniently by human body which do not convey a DC component in capacitive communication because of the series capacitance in the loop.

(56)

42 AFE modeling and verification for proof of concept

Clock Data Manchester encoded signal

1 1 0 0 1

Figure 5.1. Manchester encoding.

5.2 AFE driver and discrete fourier transform

ver-sus transition time

In the previous section we saw that the Manchester code transitions are propor-tional to the clock rate. One extreme condition is when we have continuous 1’s or 0’s transmitted which correspond to the code voltage transitions corresponding to the clock rate used. In the other extreme we have pattern of alternating ’1’ and ’0’s which correspond to half the clock rate . So the encoded data would have frequencies in between half the clock rate and clock rate provided we try to restrict the encoded data to fundamental tone . An ideal square wave is mathematically equivalent to the sum of a sine wave at that same frequency, plus an infinite se-ries of odd-multiple frequency sine waves at diminishing amplitude by a factor of 1/(2n-1), where ’n’ is an integer. But this is valid if there are instantaneous transitions between the high and low levels which is practically impossible due to infinite bandwidth requirement. We have finite rise/fall times for the square wave due to limited bandwidth . This also means that if we have finite rise and fall times , we are restricting the odd harmonics of the square. Figure 5.2 shows how increasing the transition time at the driver would restrict the frequency content. We have a peak at the fundamental followed by diminishing odd order harmonics. The transmitted signal is a clock at 10 MHz. When we were selecting the driver required for the AFE we saw that the off-the-shelf drivers have huge transition times which led us to check what happens due to that and found that it is rather a boon for this encoding style as it restricts the encoded signal to fundamental tone as much as possible but this could mean a restriction to using higher frequencies. The transmitter driver is modeled by the nominal transition times in the datasheet. It consists of voltage source and series resistance with capacitive self load for bandwidth limitation as shown in figure 5.4. The driver used is CD74AC244 from Texas instruments.

(57)

5.3 AFE receiver using discrete components 43 0 50 100 0 0.2 0.4 0.6 0.8 1 Normalized power Rise time = 1ns 0 50 100 0 0.2 0.4 0.6 0.8 1 Frequency [Mhz] Rise time = 20ns 0 50 100 0 0.2 0.4 0.6 0.8 1 Rise time = 40ns

Figure 5.2. Discrete fourier transform versus transition time.

5.3 AFE receiver using discrete components

The receiver model implementation as shown in figure 5.3 is similar to the one in Song, et al. [11]. The difference’s are that we have used off-the-shelf schmitt trigger and single supply operation. It consists of two preamplifiers to amplify the narrow small pulse signals with no DC offset,received from the body output. The Schmitt trigger triggers the output of the preamplifier to digital ’1’ and ’0’ which can be used by digital baseband. The free samples from Texas Instruments from their latest available components were chosen and simulated with their parameters to check if they can be used in the receiver before printed circuit board (PCB) design and how far can we go in terms of speed.

The simulation parameters are used based on the discrete off-the-shelf compo-nent’s datasheet. The opamp is modeled using the standard component "opamp" from ahdllib library of Cadence, as shown in appendix A.1. The Schmitt trig-ger is modeled using "hysteresis" component of the ahdllib library of Cadence as shown in appendix A.2. The table 5.1 shows the parameters used for the opamp, Schmitt trigger and biasing in the receiver. The resistor network R1 and R2 set the DC voltage at the amplifier to 1.3V , which is mid value of the shchmitt trigger triggering threshold voltages. So the voltage at the other end of the amplifier is maintained by Vref of 1.303V which cancels out the input offset voltage of the am-plifier as well. The single supply voltage of 3 V is used to simulate since batteries should be used as voltage source which can provide 3V with 2 batteries in series. The resistor pair network of (R3,R4) and (R5,R6) is used for providing a stable gain in the negative feedback configuration of the amplifier. The Schmitt trigger is modeled for positive and negative triggering threshold voltages and nominal output transition time. Both the amplifiers used “A1” and “A2”, are the same components. The amplifier used is THS4221 from Texas Instruments and Schmitt trigger is 74LV14 from NXP semiconductors.

(58)

44 AFE modeling and verification for proof of concept Ampliﬁer(A1) Ampliﬁer(A2) Schmitt trigger Reciever electrode R2 R3 R4 Vref C1 Vdc Vref C2 C3 R1 R5 R6 Vsupply Vsupply

Figure 5.3. AFE receiver for proof of concept.

5.4 Testbench for AFE model verification

Figure 6.2 shows the testbench to verify the transmitter and receiver model in the loop. As we have seen in section 5.2, if we are operating at a certain data rate say 15 Mbps, then in the encoded signal we have two extreme frequencies, i.e, 15 MHz and 7.5 MHz. The lower frequency has more attenuation according to the human body amplitude response. So we intend to set the amplifier gain for this frequency (7.5 MHz) and check if we are able to recover the signal from receiver. Higher frequency (15 MHz) is checked if we can recover the periodic clock ,in terms of its higher timing limitations. Once we see that both the frequencies can be recovered, it is believed that the other intermediate frequencies due to random data can be recovered. Later the receiver is given for test under noise and interference conditions in a software loop done in subproject2 where it measures the bit-error-rate (BER) with the digital baseband decoding of the signal. Manchester encoded data packet with random data is used to check the BER. The simulation results are presented in section 5.5 and 5.6.

(59)

5.4 Testbench for AFE model verification 45 bodyLinkModel_Distributed_RC Tx_HB_cc RX_SelfCc Rx_HB_cc TX_SelfCc Tx_Gnd_Cc _{Rx_Gnd_Cc} Gnd Vin Vout Gnd Tx_Rx_Cc

Tx_Vout _{Vin_body} Rx_Vin

Vout_body 15 Mhz Clock 7.5 Mhz Clock Manchester encoded Data packet @15 Mbps Clock Period measurement Decoded Packet data Comparison Reciever

Connecting the human body - Models, Connections and Competition

Institutionen för systemteknik

Department of Electrical Engineering

Examensarbete

Connecting the human body - Models, Connections

and Competition

Connecting the human body - Models, Connections

and Competition

Examensarbete utfört i Elektroniksystem

vid Tekniska högskolan i Linköping

av

Abstract

Acknowledgments

Contents

List of Figures

List of Tables

Chapter 1

Introduction

1.1

Background

1.2

System description

1.3

Purpose and goals of this thesis work

1.4

Outline

1.5

Goals that were met

Chapter 2

Literature study and review

2.1

Human body communication variants

2.1.1

Capacitive coupling

2.1.2

Galvanic coupling

2.2

Motivation to go for human body

communi-cation

2.3

Human body model for capacitive

communi-cation

2.4

Analog front end architectures

2.5

Conclusion

Chapter 3

Modelling and simulation of

human body for BCC

3.1

The distributed RC human body model

3.2

Simulation model including transceiver and

ground coupling

3.3

Frequency domain behaviour

3.3.1

Nominal frequency response : 120 cm arm-arm

10

10

10

10

10

−110

−100

−90

−80

−70

−60

−50

Frequency [hz]

S21 [dB]

S21 results for arm−arm (120cm) simulations

3.3.2

Impact of return path variations

3.3.3

Electrode to body coupling variations

S21 [dB]

S21 [dB]

3.3.4