Investigation and Study of Crosstalk

Department of Science and Technology Institutionen för teknik och naturvetenskap

Linköping University Linköpings universitet

g n i p ö k r r o N 4 7 1 0 6 n e d e w S , g n i p ö k r r o N 4 7 1 0 6 -E S

LiU-ITN-TEK-A-15/008--SE

Investigation and Study of

Crosstalk

Krishna Prasad Rao Pasupuleti

2015-02-02

LiU-ITN-TEK-A-15/008--SE

Investigation and Study of

Crosstalk

Examensarbete utfört i Elektroteknik

vid Tekniska högskolan vid

Linköpings universitet

Krishna Prasad Rao Pasupuleti

Handledare Magnus Karlsson

Examinator Shaofang Gong

Upphovsrätt

Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare –

under en längre tid från publiceringsdatum under förutsättning att inga

extra-ordinära omständigheter uppstår.

Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner,

skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för

ickekommersiell forskning och för undervisning. Överföring av upphovsrätten

vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av

dokumentet kräver upphovsmannens medgivande. För att garantera äktheten,

säkerheten och tillgängligheten finns det lösningar av teknisk och administrativ

art.

Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i

den omfattning som god sed kräver vid användning av dokumentet på ovan

beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan

form eller i sådant sammanhang som är kränkande för upphovsmannens litterära

eller konstnärliga anseende eller egenart.

För ytterligare information om Linköping University Electronic Press se

förlagets hemsida

http://www.ep.liu.se/

Copyright

The publishers will keep this document online on the Internet - or its possible

replacement - for a considerable time from the date of publication barring

exceptional circumstances.

The online availability of the document implies a permanent permission for

anyone to read, to download, to print out single copies for your own use and to

use it unchanged for any non-commercial research and educational purpose.

Subsequent transfers of copyright cannot revoke this permission. All other uses

of the document are conditional on the consent of the copyright owner. The

publisher has taken technical and administrative measures to assure authenticity,

security and accessibility.

According to intellectual property law the author has the right to be

mentioned when his/her work is accessed as described above and to be protected

against infringement.

For additional information about the Linköping University Electronic Press

and its procedures for publication and for assurance of document integrity,

please refer to its WWW home page:

http://www.ep.liu.se/

LITH Institution för Teknik och Naturvetenskap

Investigation and Study of Crosstalk

Krishna Pasupuleti

January 2015

Examiner: Shaofang Gong, Professor, ITN.

Supervisor: Magnus Karlsson, Universitetslektor, ITN.

Master of Science Thesis

LITH ITN Department

Wireless Networks and Electronics

Campus Norrkoping 601 74.

i

Thesis WNE 2015

Investigation and Study of Crosstalk

Krishna Pasupuleti Approved Examiner Shaofang Gong Supervisor Magnus Karlsson

Abstract

Crosstalk is defined as an unwanted coupling between the conductors. By this it is meant that signals from one of the signal conductors (a generator in this case) are coupled to another signal conductor (receptor), or conductors (receptors), depending on the number of conductors in the vicinity of the generator. Crosstalk in this way affects the signal level on the receptor and thereby affects the total system performance within the system. This can happen in several ways, one of which is through edge coupling. Edge coupling is a process where two signal conductors are placed beside each other in the same layer while the ground conductor could have been placed either under these conductors, in a separate layer like Mclin (Microstrip coupled lines) and Sclin (Coupled striplines), or beside the signal conductors as in Cpwcpl2 (Coplanar wave guide coupled lines). This then means that edge coupling occurs through the sides where the generator and the receptor are facing each other. Broadside coupling is another way, where it occurs when the signal conductors are broadside faced to each other in different layers with reference planes above and below these signal conductors.

Coupling of the signals from the generator to the receptor can occur through capacitive coupling or inductive coupling. Capacitive coupling, also known as electrical coupling, occurs due to the difference in the characteristic impedance of the generator (usually 50 or 100 Ω) and its heavy load (1 kΩ or more) which results in high voltage difference between the generator and the reference conductor (ground). This leads to the creation of a charge across the generator and the receptor-facing sides and finally results in the electric field coupling between them. On the other hand, inductive coupling, also known as magnetic coupling, occurs when the load is less than the characteristic impedance of the generator, and this thereby results in a heavy current flow through the generator which in return results in a strong magnetic field around itself and so leads to magnetic coupling to the receptor. The aim in this thesis is to measure both the capacitive and inductive coupling load’s impacts on both the edge coupling and the broadside coupling models through crosstalk on the receptor. This thesis starts with the background and corresponding theory and equations to the crosstalk coupling. Later on it tests both the edge- and broadside coupling models with different physical properties exploitation. Inductive and capacitive loads are used to measure the resulting crosstalk coupling. Particularly to see the effect of capacitive and inductive coupling in reality in multi layered PCB, a Sbclin (Broadside coupled striplines) model has been used with different angular placement of the generator. Finally mclin physical models are compared with the simulated models and corresponding differences are discussed. It can be concluded that crosstalk effect increases or decreases with physical properties exploitation. Crosstalk also increases with the wrong termination of the load.

ii

Sammanfattning

Överhörning definieras som en ofrivillig koppling mellan signalledare på ett kretskort. Det vill säga att signaler från den ena signalledaren (generatorn i detta fall) är kopplad till den andra signalledaren (receptorn i detta fall), eller andra ledare (receptorer), beroende på antalet signalledare i närheten av generatorn. Överhörning påverkar på så vis signalstyrkan på receptorn och detta påverkar den totala system funktionen. Detta kan ske på olika sätt, ett sätt är genom sidokoppling. Sidokoppling sker där två signalledare är placerad sida vid sida av varandra i ett och samma lager. Vidare är jordplanen placerad antingen under lederna, i ett särskilt lager som Mclin (Microstrip coupled lines) och Sclin (Coupled striplines), eller bredvid signallederna så som i Cpwcpl2 (Coplanar wave guide coupled lines). Således sker sidokoppling genom sidorna där generatorn och receptorn är vända mot varandra. Bredsidokoppling är ett annat sätt där överhörning sker när två signalledare är placerad bredsidvända mot varandra och har ideal referensledare både på topp och botten av kretsen. Koppling av signaler från generatorn till receptorn i sidokoppling och bredsidokoppling sker antingen genom kapacitivkopplingen eller induktivkopplingen. Kapacitivkoppling (även kallas för elektriskt-koppling), sker på grund av skillnaden mellan karakteristisk impedans av generatorn (brukar vara 50 eller 100 Ω) och sin belastning (1 kΩ eller mer) vilket leder till hög spänning mellan generatorn och sin referensledare (jordplan). Detta leder till skapandet av en laddning över generatorn och receptorvända sidor och slutligen resulterar det i en elektrisk fältkoppling mellan dem. Dock sker induktivkoppling (även kallas för magnetiskt-koppling) när belastningen är mindre än själva karakteristisk impedans av generatorn, och detta resulterar i en hög strömledning igenom generatorn. Slutligen leder detta till skapandet av ett starkt magnetiskt-fält runt om generatorn vilket leder till magnetiskt-fältkoppling till receptorn.

Syftet med studien är att mäta både den kapacitiva- och induktiva kopplingens påverkan på båda sidokoppling och bredsidokoppling genom överhörningen på receptorn. Studien inleds med bakgrund och motsvarande teori till överhörningen. Därefter blir det lite bredare testning på sidokopplings- och bredsidokopplings modeller genom att utnyttja fysiska egenskaper. Induktiv och kapacitiv belastning har använts på modellerna för att mäta motsvarande överhörning. Framförallt för att kolla effekten av kapacitiv och induktivkoppling i fallet med ett vanligt multi lager kretskort i verkligheten, Sbclin (Broadside coupled striplines) model har använts med olika vinkelplaceringar av generatorn. Slutligen har fysiska mclin modeller jämförts med motsvarande simulerade modeller och därefter har skillnaderna mellan modellerna diskuterats. Slutsatsen i denna studie är att överhörning kan öka eller minska beroende på utnyttjandet av de fysiska egenskaperna. Överhörning ökar också när det är fel belastning på generatorn.

iii

FOREWORD

This report is a result of Master of Science final year thesis project in Wireless Networks and Electronics with specialization in electromagnetic compatibility at Linkoping’s technological university.

I would like to thank Allan Huynh for the all help he contributed during the initial phase of this work. I would like to direct many thanks to my supervisor Magnus Karlsson at ITN, who gave me ideas, advice and constructive criticism during the work time. I would like to also thank Gustav Knutsson for his help during the fabrication of PCB models in the lab. Moreover I would like to thank Marjorie Carleberg from Språkverkstad for her patience in scrutinizing all my report work for linguistic accuracy.

Last but not least, I thank all the people, especially my parents and friends, who gave me support and commitment over the time.

Linköping, Sweden, January 2015 Krishna Pasupuleti

iv Contents FOREWORD ... iii NOMENCLATURE ... x Chapter 1 Introduction ... 1 1.1 Background ... 1

1.2 Problem definition and the purpose ... 2

1.3 Method ... 3

1.4 Theory ... 3

The Capacitive coupling (Zo < Load) ... 6

The Inductive coupling (Zo > Load) ... 6

1.5 Delimitations ... 6

Chapter 2 Edge coupling: Coplanar waveguide coupled lines ... 7

2.1 Introduction ... 7

2.2 Substrate model ... 7

2.3 Signal conductor’s model ... 7

2.3.1 Simulation results ... 8

2.4 Variation of the distance ‘G’ between the signal conductor and the reference conductor ... 9

2.4.1 ‘G’ as 0.2 mm ... 9

2.4.2 ‘G’ as 6.0 mm ... 9

2.4.3 Comparison of results of the signal conductor with varying ‘G’ ... 10

2.5 The basic model of the coplanar waveguide coupler ... 10

2.5.1 Simulation results ... 11

2.6 Variation of distance ‘S’ between the generator and the receptor ... 12

2.6.1 ‘S’ as 10.0mm ... 12 2.6.2 ‘S’ as 1.0 mm ... 13 2.7 Capacitive load ... 14 2.7.1 Simulation results ... 15 2.8 Inductive load ... 16 2.8.1 Simulation results ... 16 2.9 Tables ... 17

2.10 Summary of the second chapter ... 17

Chapter 3 Edge coupling: Microstrip coupled Lines ... 18

3.1 Motivation ... 18

3.2 Substrate model ... 18

3.3 The model of the signal conductor ... 18

3.3.1 Simulation results ... 19

3.4 Microstrip coupled lines model... 20

3.4.1 Simulation results ... 20

3.5 Variation of distance ‘S’ between the generator and the receptor ... 21

3.5.1 ‘S’ as 10.0 mm ... 21 3.5.2 ‘S’ as 1.0 mm ... 22 3.6 Capacitive load ... 23 3.6.1 Simulation results ... 23 3.7 Inductive load ... 24 3.7.1 Simulation results ... 24

3.8 The basic model with ideal ground plane ... 25

3.8.1 Simulation results ... 25

3.9 Table ... 26

3.10 Conclusion ... 26

v

4.1 Motivation ... 27

4.2 Substrate model ... 27

4.3 Signal conductor’s model ... 27

4.3.1 Simulation results ... 28

4.4 Edge-coupled lines in stripline model ... 28

4.4.1 Simulation results ... 28

4.5 Variation of distance ‘S’ between the generator and the receptor ... 29

4.5.1 ‘S’ as 10.0 mm ... 29 4.5.2 ‘S’ as 1.0 mm ... 30 4.6 Capacitive load ... 31 4.6.1 Simulation results ... 31 4.7 Inductive load ... 32 4.7.1 Simulation results ... 32 4.8 Table ... 34 4.9 Conclusion ... 34

Chapter 5 Broadside coupling: Broadside-coupled lines in stripline ... 35

5.1 Motivation ... 35

5.2 Substrate model ... 35

5.3 Signal conductor’s model ... 35

5.4 The basic model of broadside coupled lines in stripline model ... 36

5.4.1 Simulation results ... 36

5.5 Capacitive load ... 37

5.5.1 Simulation results ... 37

5.6 Inductive load ... 38

5.6.1 Simulation results ... 38

5.7 The angular placement of the generator ... 39

5.7.1 Orthogonal placement of the generator ... 39

5.7.2 Another angular placement of the generator ... 41

5.7.3 Mirror reflection of the previous experiment with the ports interchanged ... 42

5.7.4 Capacitive load termination to the orthogonal placement of the generator ... 44

5.7.5 Inductive load termination to the orthogonal placement of the generator ... 45

5.8 Table ... 46

5.9 Conclusion ... 46

Chapter 6 The physical models of microstrip coupled lines and the comparison of results with simulated models ... 47

6.1 Introduction ... 47

6.2 The cable loss ... 47

6.2.1 Test results ... 47

6.3 The physical model of the signal conductor ... 48

6.3.1 Comparison of results ... 48

6.4 The physical model of the microstrip coupled line model ... 49

6.4.1 Comparison of results ... 50

6.5 The physical model with distance S = 1.0 mm between the generator and the receptor... 51

6.5.1 Comparison of the results ... 52

6.6 The physical model with distance S = 10.0 mm between the generator and the receptor ... 53

6.6.1 Comparison of results ... 54

Chapter 7 Discussions and conclusions ... 56

Chapter 8 Suggestions for future studies... 57 References 58

vi

Figure 1. The model of crosstalk with three conductors generator, receptor and reference conductor. ... 4

Figure 2. The substrate model of cpwcpl2 with air as the medium above and below. ... 7

Figure 3. The signal conductor and its reference conductor according to the calculated model. ... 8

Figure 4. Signal transmission from port 1 to port2 and port reflections at port 1. ... 8

Figure 5. Signal transmission from port 1 to port 2 and port reflections at port 1 for G = 0.2 mm. ... 9

Figure 6. Signal transmission from port 1 to port 2 and port reflections at port 1 for G = 6.0 mm. ... 10

Figure 7. Variation of G and corresponding effect on the signal transmission and port matching. ... 10

Figure 8. Basic model with all the parameters described. ... 11

Figure 9. Signal transmission capability and port matching at port 1 of the basic model. ... 11

Figure 10. Crosstalk at both the port 3 and port 4 with respect to port 1. ... 12

Figure 11. Comparison of the signal transmission and port reflections of the generator (S=10.0 mm) with the basic model... 13

Figure 12. Comparison of the crosstalk at both the near end and the far end (S=10.0 mm) with the basic model... 13

Figure 13. Comparison of the signal transmission and port reflections of the generator (S=1.0 mm) with the basic model... 14

Figure 14. Comparison of the crosstalk at both the near end and the far end (S=1.0 mm) with the basic model... 14

Figure 15. Comparison of the signal transmission and port reflection of the Capacitive load model with the basic model... 15

Figure 16. Comparison of the crosstalk at both the near end and the far end of capacitive load model with the basic model... 15

Figure 17. Comparison of the signal transmission and port reflection of the Inductive load model with the basic model... 16

Figure 18. Comparison of the crosstalk at both the near end and the far end of the Inductive load model with the basic model. ... 16

Figure 19. The substrate model of mclin with air as the medium above and below the pcb board... 18

Figure 20. The signal conductor and its reference conductor according to the calculated model. ... 19

Figure 21. Signal transmission from port 1 to port2 and port reflections at port 1. ... 19

Figure 22. Mclin basic model with the parameters described. ... 20

Figure 23. Simulated results of the basic model. ... 20

Figure 24. Comparison of the signal transmission and port reflection of the generator (S=10.0 mm) with the basic model... 21

Figure 25. Comparison of the crosstalk at both the near end and the far end (S=10.0 mm) with the basic model... 21

Figure 26. Comparison of the signal transmission and port reflection of the generator (S=1.0 mm) with the basic model... 22

Figure 27. Comparison of the crosstalk at both the near end and the far end (S=1.0 mm) with the basic model... 22

Figure 28. Comparison of the signal transmission and port reflections of the Capacitive load model with the basic model... 23

Figure 29. Comparison of the crosstalk at both the near end and the far end of the Capacitive load model with the basic model. ... 23

Figure 30. Comparison of the signal transmission and port reflection of inductive load model with the basic model. ... 24

Figure 31. Comparison of the crosstalk at both the near end and the far end of the inductive load model with the basic model. ... 24

Figure 32. Comparison of the signal transmission and port reflection of the ideal ground model with the basic model. ... 25

Figure 33. Comparison of the crosstalk at both the near end and the far end of the ideal ground model with the basic model ... 25

Figure 35. The substrate model of sclin with dielectric as the medium above and below the signal conductors. ... 27

Figure 36. Signal conductor according to the calculated model. ... 27

Figure 37. Signal transmission from port 1 to port 2 and port reflections at port 1. ... 28

Figure 38. Basic model with the generator and the receptor in the middle layer. ... 28

Figure 39. Signal transmission and port reflections at port 1 of the basic model. ... 29

vii

Figure 40. Crosstalk coupling at both the near end and the far end of the basic model. ... 29

Figure 41. Signal transmission and port reflections at port 1 for ‘S’ as 10.0 mm ... 30

Figure 42. Comparison of the crosstalk at the near end and the far end (S = 10.0 mm) to the basic model .... 30

Figure 43. Signal transmission and port reflections at port 1 for ‘S’ as 1.0 mm ... 31

Figure 44. Comparison of the crosstalk at the near end and the far end (S = 1.0 mm) to the basic model. ... 31

Figure 45. Comparison of the signal transmission and port reflections at port 1 of the Capacitive load model to the basic model... 32

Figure 46. Comparison of the crosstalk at both the near end and the far end of the Capacitive load model with the basic model. ... 32

Figure 47. Comparison of the signal transmission and port reflection of the Inductive load model with the basic model... 33

Figure 48. Comparison of the crosstalk at both the near end and the far end of the Inductive load model with the basic model. ... 33

Figure 49. The substrate model of the sbclin with dielectric as the medium above and below the signal conductors. ... 35

Figure 50. Basic model with the receptor and the generator in the second and the third layers respectively. ... 36

Figure 51. Port reflections at port 1 and signal transmission from port 1 to 2 of the basic model. ... 37

Figure 52. Crosstalk coupling at the far end and the near end of the basic model. ... 37

Figure 53. Comparison of the signal transmission and port reflections at port 1 of the capacitive load model to the basic model... 38

Figure 54. Comparison of crosstalk at both the near end and the far end of the Capacitive load model with the basic model. ... 38

Figure 55. Comparison of the signal transmission and port reflection of the inductive load model with the basic model. ... 39

Figure 56. Comparison of the crosstalk at both the near end and the far end of the inductive load model with the basic model. ... 39

Figure 57. Orthogonal placement of the generator in the SBCLIN model ... 40

Figure 58. Signal transmission and port reflections at port 1 of the orthogonal model. ... 40

Figure 59. Crosstalk coupling at both the near end and the far end of the orthogonal model. ... 41

Figure 60. The layout model of the experiment. ... 41

Figure 61. Signal transmission and port reflections in comparison with the previous experiment. ... 42

Figure 62. Crosstalk coupling at the near end and the far end in comparison with the previous experiment. ... 42

Figure 63. Lay out drawing of the experiment. ... 43

Figure 64. Signal transmission and port reflection in comparison with the previous experiment. ... 43

Figure 65. Crosstalk at the near end and the far end in comparison with the previous experiment. ... 44

Figure 66. Signal transmission and port reflections at port 1 in comparison with the orthogonal model. ... 44

Figure 67. Crosstalk at the near end and the far end in comparison with the orthogonal model. ... 45

Figure 68. Signal transmission and port reflections at port 1 in comparison with the orthogonal model. ... 45

Figure 69. Crosstalk comparison at both the near end and far end in relation to orthogonal mode. ... 45

Figure 70. Cable connection loss over the frequency range ... 48

Figure 71. Front side of the fabricated signal conductor with the ground layer on the other side. ... 48

Figure 72. Comparison of result of the physical model with the simulation model of the signal conductor. .. 49

Figure 73. Layout view of the basic model before fabrication. ... 49

Figure 74. Fabricated basic model. ... 50

Figure 75. Comparison of the signal transmission of both the physical model and the simulated model... 50

Figure 76. Comparison of the near end coupling of the physical model with the simulated model. ... 51

Figure 77. Comparison of the far end coupling of the physical model with the simulated model. ... 51

Figure 78. Fabricated model for S = 1.0mm ... 52

Figure 79. Comparison of the signal transmission of both the S = 1.0 mm physical model and simulation model... 52

Figure 80. Comparison of the near end coupling of both the S = 1.0 mm physical model and simulation model... 53

Figure 81. Comparison of the far end coupling of both the S = 1.0 mm physical model and simulation model... 53

Figure 82. Fabricated model for S = 10.0mm ... 54

Figure 83. Comparison of the signal transmission of both the S = 10.0 mm physical model and simulation model. ... 54

viii

Figure 84. Comparison of the near end coupling of both the S = 10.0 mm physical model and simulation model... 55 Figure 85. Comparison of the far end coupling of both the S = 10.0 mm physical model and simulation

ix

Table Of Contents

Table 1. Signal conductor model ... 17

Table 2. Coplanar waveguide coupler model ... 17

Table 3. Microstrip coupled lines model ... 26

Table 4. Edge coupled lines in stripline ... 34

x

NOMENCLATURE

Name Sign Units

Characteristic Impedance Z_o Ω

Permittivity ϵ F/m

Permeability µ H/m

Natural frequency ω radian/sec

Wavelength λ m

Electric field E V/m

Magnetic field H A/m

Capacitance C farad (F)

Inductance L henry (H)

Electric potential V volt (V)

Electric current I ampere (A)

Resistance R ohm (Ω)

Frequency f Hertz (1/sec)

Notations

Symbol

Description

NE

R

Resistance at the near end.

FE

R

Resistance at the far end. L

R Resistance at load.

S

R Source resistance.

NE

V

Voltage at the near end.

FE

V

Voltage at the far end.

)

,

( t

z

I

_G

Generator current at place z and at time t.

)

,

( t

z

I

_R Receptor current at place z and at timet.

)

,

( t

z

V

_G Voltage between generator and reference conductor at place z and at time t.

)

,

( t

z

V

_R Voltage between receptor and reference conductor at place z and at time t.

G Distance between generator and reference conductor in mm. L Length of the signal conductors.

xi

W Width of the signal conductor in mm.

t Thickness (mm). m l Mutual inductance. m c Mutual conductance.

Abbreviations

CAD Computer Aided Design

ADS Advanced Design System

PCB Printed Circuit Board

CRO Cathode Ray Oscilloscope

CPWCPL2 Coplanar waveguide coupled lines.

CWG Coplanar waveguide.

MLIN Microstrip line.

MCLIN Microstrip Coupled lines.

SLIN Stripline.

SCLIN Edge coupled lines in stripline.

SBCLIN Broadside coupled line in stripline.

1

This chapter mainly deals with the background to crosstalk, its purpose, methods, theory and delimitations.

1.1 Background

Crosstalk between conductors in a cable, or on a PCB board, has been a very serious problem in modern communications. This problem particularly arises when it concerns electromagnetic compatibility with other products and/or interfering with itself, particularly within high frequency applications with close proximity placement of PCB lands thereby making it easier for coupling of information from one land to another. There are also cases where crosstalk can affect the radiated and/or conducted emissions of a product ([1], pg559). In the article ( [2], pg1551), it was clearly presented how the crosstalk from other bands in the bandwidth of MODIS(Moderate Resolution Imaging Spectroradiometer) affects one of its operating bands. It is an unwanted coupling of signals between the conductors within a system ( [3], pg6). Accumulation of crosstalk noise on a large scale constrains the network scalability, causes severe performance degradation and diminishes the signal to noise ratio (SNR) of a system ( [4], pg437). There are two types of crosstalk generated within the system: one is interfering with itself and the other one is the alien crosstalk or crosstalk from other systems interfering with the original system [5].

In addition, crosstalk can be classified into three types depending on where it occurs within a system i.e., linear-, nonlinear- and nonlinear & linear crosstalk. Power amplifiers are one of the major sources of non-linear crosstalk in a system and thus coupling before and after the power amplifiers results in nonlinear crosstalk and linear crosstalk respectively [6]. Crosstalk generally occurs when there are two or more conductors spaced within close vicinity to each other apart from the reference or ground conductor. By this it is meant that there is a generator, reference conductor, and a receptor or several receptors depending the field of application. In particular, crosstalk occurs either through magnetic coupling (i.e., through magnetic field also known as inductive coupling), or through electrical coupling (i.e., through electrical field also known as capacitive coupling).

Moreover crosstalk coupling is very complicated, because it is a problem of near field coupling (i.e., the radiated waves have not travelled the sufficient distance to take the form of stationary waves) and it happens in one and the same system and thereby concerns the intra system interference performance of the product ([1], pg559). Depending on the type of load applied at the generator rear end, the type of coupling is defined i.e., a heavy load results in capacitive coupling and the lighter load results in inductive coupling. Whether a load is considered to be heavier or lighter is defined with respect to the characteristic impedance of the generator.

During the inductive coupling a noise current is induced in the receptor because of the magnetic field produced by the generator. On the other hand, during the capacitive coupling, the charge accumulates on the receptor surface facing the generator because of the electric field produced by the high voltage difference between the generator and its reference conductor. In this way, crosstalk destroys the signal conducting capability of the receptor

2

and/or efficiently radiates from the receptor to other products thereby decreasing the product compatibility with other products in the vicinity by interfering with their performance.

Crosstalk is mainly dependent on the environment around the cable, or the PCB land i.e., whether it is homogeneous or inhomogeneous. A homogeneous environment around the cable implies that there is the same permittivity all around the cable, while an inhomogeneous environment has varying permittivity. Dielectric insulations around the PCB lands, or cables, can complicate the determination of per unit length capacitance but do not affect the per unit length inductance because all dielectrics have the same permeability as that of the air i.e., one ([1], pg568). By this it is meant that inhomogeneous environments are hard to determine when it comes to per unit length capacitance and therefore complicate the process of minimizing the crosstalk effect.

In “Introduction to Electromagnetic Compatibility” by Paul R. Clayton, there are several ways proposed to minimize the crosstalk depending on the source of origin. For example if the crosstalk between the two wires in a cable arises due to capacitive effect, then drawing a shield around the generator and grounding it at one of the ends would minimize it i.e., it results in a discharge of the charge accumulated on the generator side through the ground and thereby cancels the coupling between generator and receptor. However, if the source is inductive, then shielding the receptor and grounding the shield at both ends of the receptor would minimize the effect, i.e., this would result in grounding the reverse current produced due to the magnetic field and thereby minimize the coupling effect. Moreover, reducing the return current loop area by placing the PCB lands (source and ground) as close to each other as possible would drastically minimize the inductive coupling. Twisted wire pair is another useful method to reduce inductive coupling. Generally it is either inductive or capacitive coupling that is dominant when it comes to crosstalk, so concentrating on the dominant one would drastically minimize the total cross talk.

There are of course other available ways to minimize the crosstalk noise digitally by encoding the data with different techniques to encounter the crosstalk noise in the digital world [7] [8] [4] [5] .The article “Crosstalk Correlations for Coplanar Waveguide Scattering-Parameter Calibration ” by Dylan F Williams et al has successfully presented the ways to minimize the crosstalk noise added in coplanar waveguide vector network analyzer calibrations. Data recovery methods after coupling of crosstalk noise to the received signal are presented in [5]. In semiconductor physics, abnormal increase of crosstalk in sub-threshold logic circuits is absorbed for the first time i.e., when the sub-threshold voltages of nMOS and pMOS are varied and the turn on resistance of the Generator (Aggressor circuit) is much lower than the Receptor (Victim Circuit), a large crosstalk noise is observed. A method to reduce this is also presented in [9]. Moreover crosstalk noise reduction between interconnects in VLSC circuits is presented in [10].

1.2 Problem definition and the purpose

The main problem that this thesis investigates is

How the level of the crosstalk coupling between the PCB lands on a PCB board varies with a variation of the physical properties between the lands or the load.

The purpose is to investigate the crosstalk coupling by the exploitation of the physical properties between the lands such as distance variation. For example the distance between the generator and its reference conductor is varied and the resulting effect on the crosstalk coupling to the receptor is measured. Another way of exploiting the distance is by varying

3

the distance between the generator and the receptor conductors. Moreover, by varying the distance between the generator and the receptor and, at the same time, by varying the distance between the generator and the reference conductor, is another method of exploiting the distance and measuring the overall effect on the crosstalk coupling. This thesis also concentrates on investigating the effect of capacitive and inductive loads on the overall coupling. When it comes to a real life scenario, several other factors are taken into consideration such as by placing the generator and the receptor conductors in different positions on each other and seeing the overall effect.

1.3 Method

A theoretical study has been done to acquire the relevant knowledge, guidance and information from scientific articles and especially from the course book “Introduction to Electromagnetic Compatibility” by Paul R. Clayton. This literature study will cover all the basic questions such as what is crosstalk, when does it occur, what are the conditions that result in increase in coupling and so on. Moreover all the models that are going to be exploited in this thesis, such as mclin, sclin, cpwcpl2, sbclin, mlin and slin, are simulated in the Agilent ADS2011.05 CAD software for measuring the edge coupling and the broadside coupling. The physical models for mclin are tested using the Agilent CRO and the resulting graphs from physical models and simulating models are compared using MATLAB software. All the conductor specifications are calculated using Agilent ADS Linecalc.

1.4 Theory

One of the major problems in high density circuits and high data rate circuits is the crosstalk between the neighboring circuits or transmission lines in these systems causing degradation of signal integrity and creating logical errors. Even though the solution to reducing crosstalk is one of the simplest methods i.e., to draw apart circuitry or lines from each other so that there is no coupling between them, it is hard to implement this for the given size of the machines as they are continuously decreasing with high density circuits spaced very close to each other [11].

Moreover Estimating crosstalk problem in a complex system is not an easy task since it requires methods to quickly predict where the crosstalk coupling occurs and find a corresponding solution. Currently there is only a little information in the literature for predicting the maximum crosstalk in electrically large circuits. However, in the article “Maximum crosstalk estimation in weakly coupled transmission lines” by Mathew S. Halligan and Daryl G. Beetner, there is presented a mathematically rigorous, worst case high frequency crosstalk including transmission lines losses [11].

In general, crosstalk occurs only when there are more than two conductors. However, the notions involved in two conductor transmission line theory carry over to a large extent to the case of multi conductor transmission lines( [1], pg560). So, for example, adding a third conductor to a two conductor transmission line would possibly generate interference between the circuits that are connected at the ends of these conductors.From the Figure 1, according to Clayton, 2006, one can see how these transmission line equations are calculated.

4

Generator conductor (Aggressor circuit)

Rs + IG(z, t)

VG(z, t)

Receptor conductor (Victim circuit)

Vs(t)

+ + IR(z, t) ₊ RL RNE VNE(t) V ( tz, ) R VFE(t) R FE - - - I_G(z, t) I_R(z, t) - Reference conductor

Figure 1. The model of crosstalk with three conductors generator, receptor and reference conductor.

The voltage source Vs(t) is connected to the generator circuit (i.e., between the generator conductor and the reference conductor) having the driving loadR_L. This load connected to the generator results in radiation of an electromagnetic field from the generator and couples to the receptor circuit either by inductive coupling or capacitive coupling. In this way coupling results in the generation of the near end and the far end voltages V_NE(t) and V_FE(t)

respectively, which results in the flow of current I_R(z, t)in the receptor circuit.

Moreover the induced current due to inductive coupling is generally stronger at the far end than at the near end, which is of course due to less resistance through the load resulting in high flow of current. On the other hand, the result is the opposite for capacitive coupling, which is quite a natural response because of the high load, so therefore capacitive coupling is more effective at the near end than at the far end.

Both the mutual inductance (

l

_m) and mutual capacitance (

c

_m) are dependent on the PCB land cross-sectional dimensions such as length, separation from the ground and the separation between the generator conductor and receptor conductor and so on. Drawing apart two lines reduces these mutual impedances, and doubling the line lengths doubles the crosstalk. Crosstalk is very dependent on the frequency of the operation of the generator conductor i.e., the higher the frequency of operation of the generator circuit source the better the crosstalk coupling ( [3], pg6). It is these dimensions that are being exploited in this project to create a better understanding of crosstalk. In this project, models such as the coplanar waveguide, microstrip lines, and striplines are used to exploit these characteristics.

So from the Figure 1 according to Clayton, 2006, we can write the near end and the far end voltages in the frequency domain is as follows.

+

-5 S L S L m FE NE FE NE S L S m FE NE NE NE V R R R C j R R R R V R R L j R R R V      



1



……….. (1) S L S L m FE NE FE NE S L S m FE NE FE FE V R R R C j R R R R V R R L j R R R V       



1



………. (2) Where

L

l

L

_m



_m



,

L

c

C

_m



_m



.

(Note: Here it is assumed that all the lands have same length i.e., L.)

So depending on the load applied, one of the parts of the equation (1 or 2) can be reduced to zero, which is because, as stated before, it is often either capacitive coupling or inductive coupling that is dominant at a particular load and so, therefore the other one has a very minute effect on total coupling.

Weak coupling is assumed in this project, and by this it is meant that the current and voltage developed on the receptor circuit due to generator circuit will not develop any voltage or current on the generator circuit in return. Generally this is assumed because these currents and voltages produced on the receptor circuit due to the generator circuit do not have sufficient power to produce second order effects on the generator i.e., this is a one way process.

Equation (1 & 2) can also be derived by using the time domain model as presented in Clayton, 2006. Using Kirchhoff’s law from the generator circuit we obtain the following equations.

0 ) , ( ) , (       t t z I l z t z V G G G 0 ) , ( ) ( ) , (        t t z V c c z t z I _G m G G

Now, as the receptor circuit is dependent on the generator circuit, we can use Kirchhoff’s law once again to determine the time domain equations as follows,

t t z I l t t z I l z t z V _G m R R R          ( , ) ( , ) ( , )_{……….. (3a)} t t z V c t t z V c c z t z I G m R m R R          ( , ) ( , ) ) ( ) , ( _{……….. (3b)}

The right hand part of the Equation (3a and 3b) are due to mutual inductance and capacitance between the generator circuit and the receptor circuit. So from the above equations one can clearly say that receptor circuit is completely driven by generator circuit’s electromagnetic fields, and it is these fields that negatively impact the original signals in the receptor circuitry.

6

The Capacitive coupling (Zo < Load)

Capacitive coupling occurs generally when the load impedance is greater than the characteristic impedance of the generator conductor. This then results in less flow of current in the generator circuit and thereby creates a high voltage difference between the generator conductor and reference conductor. This high voltage then creates an electric field and couples to the receptor circuit through mutual capacitance

c

_m.

Even though some current flow exists in the generator circuit that creates the magnetic field and couples to the receptor circuit through mutual inductance, it is much smaller compared to the capacitive coupling. Generally the load range for capacitive coupling is between 50.0Ω to 1.0kΩ or above. So when it comes the equations shown above, only capacitive coupling parts are included to calculate the effects i.e., neglecting the inductive parts in all the above mentioned equations.

The Inductive coupling (Zo > Load)

By inductive coupling it is meant that the load impedance is less than the characteristic impedance of the generator resulting in a high current flow in the generator. Due to this impedance mismatching, there are lot of signal waves that are reflected back from the load resulting in a building up of signal strength above the threshold level. And so the heavy current flows through the generator circuit producing a strong magnetic field around it and thereby radiating efficiently from the generator circuit to the receptor circuit through mutual inductance

l

_m .

Generally it is said that inductive coupling has most effect when it comes to low impedance loads, which is because even though the capacitive coupling exists, it is much weaker and almost negligible compared to strong inductive coupling. The load range for inductive coupling is in the range of 1.0 Ω or so, and less than 50.0 Ω, but generally it is much less than characteristic impedance which is generally 50.0 Ω. Similar to capacitive coupling, when it comes to calculating the effects due to inductive coupling, capacitive coupling parts are neglected in the above mentioned equations.

1.5 Delimitations

The frequency of application in this thesis is in between 100.0 MHz and 3.0 GHz. The motivation behind this is that there are not so many scientific articles in this area with high frequency applications such as 3.0 GHz. The substrate model used in this thesis is FR4 with a height of 0.8 mm height between the layers. In this thesis the main limits, apart from the functional limits, is the fabrication of multilayer PCB in the laboratory and the time factor for this. This is because laboratory equipment in the PCB lab is not sufficient for designing the multilayer PCB boards and time is not available for testing these physical models and comparing them with the simulation models.

7

This chapter mainly deals with the edge coupling in coplanar waveguide coupled lines.

2.1 Introduction

This PCB model is a one layer board that has the generator and the receptor lying in between the reference conductors in one and the same layer. The main motivation behind this model is to measure how the edge coupling dominates when both the generator and the receptor are placed close to each other. It is hard to say how strong the edge coupling can be based only on the distance between the generator and the receptor. However, it can often be said that the closer they are placed to each other, the better the chances are for stronger coupling. Edge coupling, as the name suggests, couples through the sides where the generator and the receptor are facing each other.

2.2 Substrate model

The substrate characteristics are calculated using Linecalc software with a characteristic impedance of 100.0 Ω at a center frequency of 3.0 GHz with 0.8 mm height.

Figure 2. The substrate model of cpwcpl2 with air as the medium above and below.

2.3 Signal conductor’s model

The signal conductor’s layout drawing is shown in Figure 3 with the described parameters. The motive behind this is to see the full performance of generator in the absence of the receptor conductor. This is a coplanar wave guide model, unlike the main model (cpwcpl2). This is because the characteristic impedance of the generator conductor changes when the receptor is removed.

Chapter 2

Edge coupling: Coplanar waveguide coupled

lines

8 L = 100.0 mm P5 P6 P1 P2 G = 3.0 mm P3 P4

Figure 3. The signal conductor and its reference conductor according to the calculated model.

2.3.1 Simulation results

Even though there are some standing wave reflections in the signal conductor in Figure 4, it is clear that the generator is perfectly matched with the load. The reason behind these reflections are due to the signal conductor length which is larger than the signal wavelength. This results in variation of the phase of the signal as it travels down the path.

Figure 4. Signal transmission from port 1 to port2 and port reflections at port 1.

Generator W=4.81 mm Reference conductor W = 40.0 mm Reference conductor S 2 1 M ag n it u d e in d B S 1 1 M ag n it u d e in d B Frequency (GHz) Frequency (GHz)

9

2.4 Variation of the distance ‘G’ between the signal conductor and

the reference conductor

Here all the calculated parameters of the signal conductor, according to Linecalc, are kept constant while varying the ‘G’ and seeing its effect on signal conduction through the signal conductor.

2.4.1 ‘G’ as 0.2 mm

The layout setup is almost same as in the Figure 3, except that the distance between generator and its reference conductor is minimized to 0.2 mm.

2.4.1.1 Simulation results

From Figure 5, it is clear that the signal transmission level has decreased to almost half the value compared to the basic model in Figure 4, while the signal reflections have become stronger at port 1. This is mainly due to the fact that the characteristic impedance of the generator has changed as the reference conductor is brought very close to it, resulting in mismatching at ports 1 and 2.

Figure 5. Signal transmission from port 1 to port 2 and port reflections at port 1 for G = 0.2 mm.

2.4.2 ‘G’ as 6.0 mm

This has the same experimental layout setup as in Figure 3, except that the ‘G’ now has been doubled to 6.0 mm instead. This is to study the impact of impedance mismatching between the generator and the reference conductor.

2.4.2.1 Simulation results

The signal transmission has worsened somewhat compared to both the basic model and with the ‘G’ as 0.2 mm. On the other hand, as expected, port mismatchning with port 1 has become stronger than in both the previous models. Despite the fact that there are only minute differences in the graphs in Figure 5 and 6, it can be said that having a large distance between the generator and its reference conductor is not a good choice.

S 2 1 M ag n it u d e in d B S1 1 M ag n it u d e in d B Frequency (GHz) Frequency (GHz)

10

Figure 6. Signal transmission from port 1 to port 2 and port reflections at port 1 for G = 6.0 mm.

2.4.3 Comparison of results of the signal conductor with varying ‘G’

Here, both the above stated models with different ‘G’ are compared with the basic calculated model. Signal conduction through the signal conductor is best with the calculated model compared with the both the other models where ‘G’ is varied. This also occurs with the port matching at port 1 i.e., low port reflections for the calculated model while high port reflections for both the other models, which is because of port mismatching, resulting from different characteristic impedances of the signal conductor as ‘G’ is varied. Figure 7 shows this very clearly.

Figure 7. Variation of G and corresponding effect on the signal transmission and port matching.

2.5 The basic model of the coplanar waveguide coupler

This, as the name suggests, is the basic model that is calculated using the Linecalc software. The Layout diagram is shown in Figure 8 with the parameters described. Note that the physical properties of the lands and the distance between them are different here compared with the signal conductor model except that the G is taken as 0.2 mm which showed stronger impedance matching of the generator with the reference conductor in the presence of an another conductor. The reason why the physical properties are different is that, since the

S2 1 M ag n itu d e i n d B S1 1 M ag n it u d e in d B Frequency (GHz) Frequency (GHz) S 2 1 M ag n it u d e in d B S1 1 M ag n itu d e i n d B G = 3.0 mm with S11 = -17.71 dB at fo = 500.3 MHz G = 0.2 mm with S11 = -10.60 dB at fo = 500.3 MHz G = 6.0 mm with S11 = -10.62 dB at fo = 500.3 MHz Frequency (GHz) Frequency (GHz)

11

model includes the receptor conductor, the characteristic impedance changes and hence the name, coplanar wave guide coupler.

L = 100.0 mm P7 P8 G = 0.2mm G = 0.2 mm P1 P2 P3 P4 P5 P6

Figure 8. Basic model with all the parameters described.

2.5.1 Simulation results

As expected, signal transmission through the signal conductor is very good and has very low port reflections at port 1, which can be seen in Figure 9. On the other hand, crosstalk at both the near end and the far end has gradually became constant around -20.0 dB, which can be observed in Figure 10.

Figure 9. Signal transmission capability and port matching at port 1 of the basic model.

Receptor W=3.40 mm W=40.0 mm Reference conductor W= 40.0 mm Reference conductor Generator W=3.40mm S = 4.71 mm Frequency (GHz) Frequency (GHz)

12

Figure 10. Crosstalk at both the port 3 and port 4 with respect to port 1.

2.6 Variation of distance ‘S’ between the generator and the receptor

Here, all the calculated parameters in the Figure 8 according to Linecalc are kept constant except that the distance ‘S’ between the generator and the receptor is varied and its corresponding effect on the crosstalk at both the near end and the far end is observed.

2.6.1 ‘S’ as 10.0mm

The layout setup looks almost like that in Figure 8, except that the distance ‘S’ is now increased to 10.0 mm. The motivation behind this is to see if the crosstalk decreases as the ‘S’ is now increased to almost double the calculated value.

2.6.1.1 Simulation results

Compared with the basic model, the port reflections at port 1 have increased by 10.0 dB when the distance ‘S’ is increased to 10.0 mm, whereas signal transmission has attained more ripples and almost has the same signal transmission quality. The probable reason behind this is that, when the model was initially designed, it was considered that it has another signal conductor in the vicinity and hence the corresponding characteristic impedance of 100.0 Ω is calculated keeping in mind that it existed. Now, as the distance is varied, the characteristic impedance of the signal generator is changed hence resulting in more ripples in signal transmission and stronger port reflections compared to the basic model.

As expected, the crosstalk did decrease as the ‘S’ increased. The coupling at the near end has decreased by approximately 5.0 dB while the coupling at the far end has decreased by approximately 3.25 dB.

13

Figure 11. Comparison of the signal transmission and port reflections of the generator (S=10.0 mm) with

the basic model.

Figure 12. Comparison of the crosstalk at both the near end and the far end (S=10.0 mm) with the basic

model.

2.6.2 ‘S’ as 1.0 mm

The experimental setup is the same as in the previous experiment, except that distance ‘S’ is now changed to 1.0 mm instead. The motivation behind this is to observe if the crosstalk increases in contrast to the previous experiment.

2.6.2.1 Simulation results

Port reflections at port 1 have increased by 1.37 dB compared with the basic model. This is because the characteristic impedance of the generator conductor has changed, but this change

M ag n it u d e o f S 1 1 i n d B Frequency (GHz) Basic model S11=-26.72 dB with S= 4.71 mm, fo =509.5 MHz S11 = -16.2 dB with S= 10.0 mm, fo= 507.3 MHz. Frequency (GHz) M ag n it u d e o f S 2 1 in d B M ag n it u d e o f S 4 1 in d B M ag n it u d e o f S 3 1 i n d B Frequency (GHz) Frequency (GHz) Basic model, S31=-20.09 dB & S41 =-36.60 dB with S= 4.71 mm, fo =509.5 MHz S31 = -25.03 dB & S41 = -39.85 dB with S= 10.0 mm, fo = 507.3 MHz.

14

is very minute compared to the previous experiment which was almost 10.0 dB greater than the basic model. On the other hand, crosstalk at the near end has increased by 6.21 dB and at the far end by 3.12 dB as expected.

Figure 13. Comparison of the signal transmission and port reflections of the generator (S=1.0 mm) with

the basic model.

Figure 14. Comparison of the crosstalk at both the near end and the far end (S=1.0 mm) with the basic

model.

2.7 Capacitive load

The experimental criteria and the layout setup is the same as in Figure 8, except that the port 2 of the generator is now terminated with a capacitive load i.e., 1.0 kΩ. The reason behind this is to measure and understand how the capacitive coupling works when there is a heavy load. M ag n it u d e o f S 1 1 i n dB M ag n it u d e o f S 2 1 i n dB Frequency (GHz) _{Frequency (GHz)} Basic model S11=-26.91 dB with S= 4.71 mm, fo =502.8 MHz S11 = -28.28 dB with S= 1.0 mm, fo= 502.8 MHz M ag n it u d e o f S 3 1 i n dB M ag n it u d e o f S 4 1 i n dB Frequency (GHz) Frequency (GHz) Basic model, S31=-20.10 dB & S41 =-36.78 dB with S= 4.71 mm, fo =502.8 MHz S31 = -14.31 dB & S41 = -33.66 dB with S= 1.0 mm, fo = 502.8 MHz

15

2.7.1 Simulation results

As expected, the capacitive termination has led to bad signal transmission as can be seen in Figure 15 and at the same time port reflections with port 1 have strengthened greatly compared to the basic model i.e., 24.42 dB rise in the S11 value, which is quite obvious due to heavy load mismatch. On the other hand, as expected, crosstalk coupling at both the near end and the far end has also increased. A rise of 1.2 dB at the near end and a tremendous rise of 15.44 dB at the far end has been observed compared with the basic model. The reason behind this high rise at the far end is that port 2 is poorly mismatched and hence is radiating efficiently from there.

Figure 15. Comparison of the signal transmission and port reflection of the Capacitive load model with

the basic model.

Figure 16. Comparison of the crosstalk at both the near end and the far end of capacitive load model with

the basic model.

M ag n it u d e o f S 1 1 i n dB M ag n it u d e o f S 2 1 i n dB Frequency (GHz) Frequency (GHz) Basic model with correct termination. S11=-26.91 dB

& S21 = -0.08 dB, at fo = 502.8 MHz

Capacitive termination at port2, S11=-2.49 dB & S21 = -3.92 dB at fo=502.8 MHz M ag n it u d e o f S 3 1 i n dB M ag n it u d e o f S 4 1 i n dB Frequency (GHz) Frequency (GHz) S31 = -18.90 dB & S41=-21.34 dB for capacitive termination at fo =502.8 MHz S31 = -20.10 dB & S41=-36.78 dB for basic model at fo = 502.8 MHz

16

2.8 Inductive load

The layout setup is same as in the previous experiment except that the load at port 2 is now changed to a lighter load of 5.0 Ω. This is done to measure the inductive coupling effect at both the near end and the far end.

2.8.1 Simulation results

In Figures 17 and 18, we see the inductive load simulation results. As can be seen, the port matching at port 2 is so poor that the port reflections at port 1 have the strongest value so far and there is worse signal transmission quality compared to both the capacitive load and the basic model. As compared to the basic model, a very significant rise of 25.96 dB for port reflections at port 1, and a fall of -7.71 dB for signal transmission from port 1 to port 2 can be observed. Hence it can be said that the generator in this experiment can be a good radiating antenna and can couple very easily.

On the other hand, crosstalk at the near end has decreased by 0.54 dB compared to the basic model, whereas coupling at the far end has increased by 15.67 dB, which also indicates that inductive coupling does have more effect at the far end than at the near end.

Figure 17. Comparison of the signal transmission and port reflection of the Inductive load model with the

basic model.

Figure 18. Comparison of the crosstalk at both the near end and the far end of the Inductive load model with

the basic model.

M ag n it u d e o f S 1 1 i n dB M ag n it u d e o f S 2 1 i n dB Frequency (GHz) Frequency (GHz) M ag n it u d e o f S 3 1 i n dB M ag n it u d e o f S 4 1 i n dB Frequency (GHz) Frequency (GHz) S31 = -20.54 dB & S41=-21.11 dB for inductive termination at fo =502.8 MHz S31 = -20.10 dB & S41=-36.78 dB for basic model at fo = 502.8 MHz Basic model with correct termination. S11=-26.91 dB

& S21 = -0.08 dB, at fo = 502.8 MHz

Inductive termination at port2, S11=-0.95 dB & S21 = -7.79 dB at fo=502.8 MHz

17

2.9 Tables

Table 1. Signal conductor model

At fo =500.3MHz. Basic model G = 0.2mm G = 6.0mm

S11

(port reflections at port 1)

-17.71 dB -10.60 dB -10.62 dB

S21

(Signal from port 1 to 2)

-0.09 dB -0.43 dB -0.41 dB

Table 2. Coplanar waveguide coupler model

Model Name S11

(port reflections at port 1)

Near end S31 Far end S41 Basic model At fo = 502.8MHz -26.91 dB -20.10 dB -36.78 dB ‘S’ as 10.0mm At fo =507.3MHz -16.21 dB -25.03 dB -39.85 dB ‘S’ as 1.0mm At fo = 502.8MHz -28.28 dB -14.31 dB -33.66 dB Capacitive load At fo = 502.8MHz -2.49 dB -18.90 dB -21.34 dB Inductive load At fo = 502.8MHz -0.95 dB -20.54 dB -21.11 dB

2.10Summary of the second chapter

Overall, the basic models have the ideal conditions in comparison with all the other values in the above Tables 1 and 2. Even though port reflections at port 1 are lowest for the model S = 1.0 mm in Table 2, it has strongest near end coupling in comparison with all the other models. Moreover, it can be concluded that crosstalk coupling increases or decreases depending on whether the distance between the generator and the receptor is decreased or increased. Load mismatching at the generator end always results in stronger crosstalk coupling to the receptor.

18

This chapter mainly deals with the edge coupling in Microstrip coupled lines.

3.1 Motivation

In this PCB model, both the generator and the receptor are lying on the same side of the PCB board while the ground or reference conductor is on the reverse side of the board. Unlike the previous model, this is a two layer PCB with signal conductors on one side and the ground layer on the other side. The main motive behind the usage of this model is to measure the edge coupling dominance when the reference conductor is on the other side. This model also replicates a real life scenario where all the signal circuitry is on one side and the ground layer is on the other side. Hence it is a good study case for measuring the edge coupling dominance.

3.2 Substrate model

The substrate characteristics are calculated using the Linecalc with the following parameters as input: FR4 substrate with height between the layers as 0.8 mm, with 50.0 Ωs characteristic impedance and with a center frequency of 3.0 GHz.

Figure 19. The substrate model of mclin with air as the medium above and below the pcb board.

3.3 The model of the signal conductor

Here the mlin model is used instead of the mclin model and the reason is quite obvious i.e., since there is only one signal conductor the characteristic impedance changes and hence the mlin model is preferred with same ideal conditions as in the mclin. The physical properties of the signal line conductor are as follows: W = 1.36 mm and L = 89.0 mm. The width of the reference conductor is W = 50.0 mm and has the same length as that of signal line conductor.

Chapter 3

Edge coupling: Microstrip coupled Lines

Generator & receptor as cond layer and reference conductor as cond2 lager.

19

Figure 20. The signal conductor and its reference conductor according to the calculated model.

3.3.1 Simulation results

The signal conduction capability is quite good and has quite low reflections at port 1 despite the fact that the ground is a non-ideal one. There are some ripples in the graphs in Figure 21, and the probable reason for this is that the signal conductor length, which is larger than the signal wavelength, results in a change of the phase of the signal over the distance.

Figure 21. Signal transmission from port 1 to port2 and port reflections at port 1.

20

3.4 Microstrip coupled lines model

This is the basic model having the parameters calculated using Linecalc, with a characteristic impedance of 50.0 Ω and a length of 83.0 mm. The width of the signal conductor is 1.29 mm. The reference layer has a width of 60.0 mm but has the same length as that of the signal conductors.

Figure 22. Mclin basic model with the parameters described.

3.4.1 Simulation results

This simulation result shows good signal transmission quality and has low port reflections at port 1 as expected from the basic model. Crosstalk coupling at both the near end and the far end is quite low and quite reasonable and is perfectly matched at all the ports as one can see in Figure 23.

Figure 23. Simulated results of the basic model.

21

3.5 Variation of distance ‘S’ between the generator and the receptor

The motivation behind this is to measure the impact of the distance ‘S’ between the generator and the receptor on crosstalk coupling to the receptor. All the other parameters that are calculated using the Linecalc are kept constant.

3.5.1 ‘S’ as 10.0 mm

This is done to see if the crosstalk coupling decreases as distance ‘S’ is now doubled. Moreover, it has also been done to see the effect of it on the signal transmission capability.

3.5.1.1 Simulation results

Port reflections at port 1 have almost the same values of port reflections values as in the basic model. Signal transmission capability is almost the same as in the basic model, and hence the variation of ‘S’ did not affect the signal transmission capability. On the other hand, crosstalk coupling at both the near end and the far end was almost the same as in the basic model at lower frequencies but as the frequency increased to 3.0 GHz both the near end and the far end coupling decreased almost by 10.0 dB.

Figure 24. Comparison of the signal transmission and port reflection of the generator (S=10.0 mm) with

the basic model.

Figure 25. Comparison of the crosstalk at both the near end and the far end (S=10.0 mm) with the basic

model. M ag n it u d e o f S 1 1 i n dB M ag n it u d e o f S 2 1 i n dB Frequency (GHz) Frequency (GHz) Frequency (GHz) Frequency (GHz) M ag n it u d e o f S 3 1 i n dB M ag n it u d e o f S 4 1 i n dB

22

3.5.2 ‘S’ as 1.0 mm

This experiment has the same motive as in the previous experiment i.e., ‘S’ is now decreased to 1.0mm and the corresponding effect on signal transmission quality and crosstalk is measured.

3.5.2.1 Simulation Result

Port reflections at port 1 have slightly decreased compared to the basic model but the signal transmission quality has also decreased slightly in comparison to the basic model. On the other hand, a rise of 8.01 dB at the near end and a rise of 9.23 dB at the far end has been observed in comparison with the basic model. The reason behind these strong couplings is that, since the distance between the generator and the receptor has decreased, there is a better chance for an increase in the coupling compared to both the basic model and the previous experiment.

Figure 26. Comparison of the signal transmission and port reflection of the generator (S=1.0 mm) with the

basic model.

Figure 27. Comparison of the crosstalk at both the near end and the far end (S=1.0 mm) with the basic

model. M ag n it u d e o f S 1 1 i n dB M ag n it u d e o f S 2 1 i n dB M ag n it u d e o f S 3 1 i n dB M ag n it u d e o f S 4 1 i n dB Frequency (GHz) Frequency (GHz) Frequency (GHz) Frequency (GHz) S31 = -21.33 dB & S41 = -25.64 dB with S= 1.0mm, fo = 500.3 MHz Basic model, S31=-29.34 dB & S41 =-34.87 dB with S= 4.71 mm, fo =500.3 MHz

23

3.6 Capacitive load

The experimental setup and the layout is the same as in Figure 22, except that port 2 of the generator is now terminated with a capacitive load i.e., 1.0 kΩ. The reason behind this is to measure and understand the impact of capacitive coupling in a mclin model when it is a heavy load. Moreover this setup will also check if the capacitive coupling has more effect on the near end than at the far end.

3.6.1 Simulation results

As expected the signal transmission capability of the signal line conductor has decreased to a low value of -6.2 dB at fo = 500.3 MHz and port reflections have increased to -1.0 dB. On the other hand, crosstalk at the near end has increased at higher frequencies. However, at the far end, coupling has the same effect as it does in the basic model for higher frequencies. So, it can be said that the capacitive coupling has more effect at the near end than at the far end.

Figure 28. Comparison of the signal transmission and port reflections of the Capacitive load model with

the basic model.

Figure 29. Comparison of the crosstalk at both the near end and the far end of the Capacitive load model

with the basic model.

Frequency (GHz) Frequency (GHz) M C L IN ( S 3 1 ) & 1 .0 k Ω (S 31 ) M C L IN ( S 4 1 ) & 1 .0 k Ω (S 41 ) Frequency (GHz) Frequency (GHz)

Basic model with correct termination. S11=-15.76 dB & S21 = -0.06 dB, at fo = 500.3 MHz

Capacitive termination at port2, S11=-1.0 dB & S21 = -6.2 dB at fo=500.3 MHz

S31 = -33.34 dB & S41=-25.81 dB for