IET ELECTRICAL MEASUREMENT SERIES 12

Microwave Measurements
3rd Edition

Other volumes in this series:
Volume 4 Volume 5 Volume 7 Volume 8 Volume 9 Volume 11 The current comparator W.J.M. Moore and P.N. Miljanic Principles of microwave measurements G.H. Bryant Radio frequency and microwave power measurement A.E. Fantom A handbook for EMC testing and measurement D. Morgan Microwave circuit theory and foundations of microwave metrology G. Engen Digital and analogue instrumentation: testing and measurement N. Kularatna

Microwave Measurements
3rd Edition
Edited by R.J. Collier and A.D. Skinner

The Institution of Engineering and Technology

Published by The Institution of Engineering and Technology, London, United Kingdom © 1985, 1989 Peter Peregrinus Ltd © 2007 The Institution of Engineering and Technology First published 1985 (0 86341 048 0) Second edition 1989 (0 86341 184 3) Third edition 2007 (978 0 86341 735 1) This publication is copyright under the Berne Convention and the Universal Copyright Convention. All rights reserved. Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted, in any form or by any means, only with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms of licences issued by the Copyright Licensing Agency. Inquiries concerning reproduction outside those terms should be sent to the publishers at the undermentioned address: The Institution of Engineering and Technology Michael Faraday House Six Hills Way, Stevenage Herts, SG1 2AY, United Kingdom www.theiet.org While the authors and the publishers believe that the information and guidance given in this work are correct, all parties must rely upon their own skill and judgement when making use of them. Neither the authors nor the publishers assume any liability to anyone for any loss or damage caused by any error or omission in the work, whether such error or omission is the result of negligence or any other cause. Any and all such liability is disclaimed. The moral rights of the author to be identified as author of this work have been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.

British Library Cataloguing in Publication Data Microwave measurements. – 3rd ed. 1. Microwave measurements I. Collier, Richard II. Skinner, Douglas III. Institution of Engineering and Technology 621.3’813 ISBN 978-0-86341-735-1

Typeset in India by Newgen Imaging Systems (P) Ltd, Chennai Printed in the UK by Athenaeum Press Ltd, Gateshead, Tyne & Wear

Contents

List of contributors Preface 1 Transmission lines – basic principles R. J. Collier 1.1 1.2 Introduction Lossless two-conductor transmission lines – equivalent circuit and velocity of propagation 1.2.1 Characteristic impedance 1.2.2 Reflection coefficient 1.2.3 Phase velocity and phase constant for sinusoidal waves 1.2.4 Power flow for sinusoidal waves 1.2.5 Standing waves resulting from sinusoidal waves Two-conductor transmission lines with losses – equivalent circuit and low-loss approximation 1.3.1 Pulses on transmission lines with losses 1.3.2 Sinusoidal waves on transmission lines with losses Lossless waveguides 1.4.1 Plane (or transverse) electromagnetic waves 1.4.2 Rectangular metallic waveguides 1.4.3 The cut-off condition 1.4.4 The phase velocity 1.4.5 The wave impedance 1.4.6 The group velocity 1.4.7 General solution Further reading

xvii xix 1 1 1 4 5 5 6 7 8 9 10 10 10 12 14 15 15 16 16 17 19 19 19

1.3

1.4

2

Scattering parameters and circuit analysis P. R. Young 2.1 2.2 Introduction One-port devices

vi

Contents 2.3 2.4 2.5 2.6 2.7 2.8 Generalised scattering parameters Impedance and admittance parameters 2.4.1 Examples of S-parameter matrices Cascade parameters Renormalisation of S-parameters De-embedding of S-parameters Characteristic impedance 2.8.1 Characteristic impedance in real transmission lines 2.8.2 Characteristic impedance in non-TEM waveguides 2.8.3 Measurement of Z0 Signal flow graphs Appendices 2.A Reciprocity 2.B Losslessness 2.C Two-port transforms References Further reading 22 24 27 27 28 29 30 30 33 35 36 37 37 39 40 41 41 43 43 52 52 54 54 54

2.9

3

Uncertainty and confidence in measurements John Hurll 3.1 3.2 Introduction Sources of uncertainty in RF and microwave measurements 3.2.1 RF mismatch errors and uncertainty 3.2.2 Directivity 3.2.3 Test port match 3.2.4 RF connector repeatability 3.2.5 Example – calibration of a coaxial power sensor at a frequency of 18 GHz References

54 56 59 59 60 61 61 61 62 62 62 63 64 64 65

4

Using coaxial connectors in measurement Doug Skinner 4.1 4.2 Introduction 4.1.1 Coaxial line sizes Connector repeatability 4.2.1 Handling of airlines 4.2.2 Assessment of connector repeatability Coaxial connector specifications Interface dimensions and gauging 4.4.1 Gauging connectors Connector cleaning 4.5.1 Cleaning procedure 4.5.2 Cleaning connectors on static sensitive devices Connector life

4.3 4.4 4.5

4.6

Contents 4.7 4.8 4.9 4.A 4.B 4.C 4.D 4.E Adaptors Connector recession Conclusions Appendix A Appendix B Appendix C Appendix D Appendix E Further reading

vii 65 65 66 66 66 85 86 87 88 91 91 91 93 94 97 98 104 105 107 108 110 110 112 112 114 115 115 115 115 116 116 117 119 120 121 121 122 122 124

5

Attenuation measurement Alan Coster 5.1 5.2 5.3 Introduction Basic principles Measurement systems 5.3.1 Power ratio method 5.3.2 Voltage ratio method 5.3.3 The inductive voltage divider 5.3.4 AF substitution method 5.3.5 IF substitution method 5.3.6 RF substitution method 5.3.7 The automatic network analyser Important considerations when making attenuation measurements 5.4.1 Mismatch uncertainty 5.4.2 RF leakage 5.4.3 Detector linearity 5.4.4 Detector linearity measurement uncertainty budget 5.4.5 System resolution 5.4.6 System noise 5.4.7 Stability and drift 5.4.8 Repeatability 5.4.9 Calibration standard A worked example of a 30 dB attenuation measurement 5.5.1 Contributions to measurement uncertainty References Further reading

5.4

5.5

6

RF voltage measurement Paul C. A. Roberts 6.1 6.2 Introduction RF voltage measuring instruments 6.2.1 Wideband AC voltmeters 6.2.2 Fast sampling and digitising DMMs

viii

Contents 6.2.3 RF millivoltmeters 6.2.4 Sampling RF voltmeters 6.2.5 Oscilloscopes 6.2.6 Switched input impedance oscilloscopes 6.2.7 Instrument input impedance effects 6.2.8 Source loading and bandwidth AC and RF/microwave traceability 6.3.1 Thermal converters and micropotentiometers Impedance matching and mismatch errors 6.4.1 Uncertainty analysis considerations 6.4.2 Example: Oscilloscope bandwidth test 6.4.3 Harmonic content errors 6.4.4 Example: Oscilloscope calibrator calibration 6.4.5 RF millivoltmeter calibration Further reading 125 126 127 129 130 132 133 133 135 136 137 137 138 140 143 147 147 148 150 150 151 152 153 154 154 155 156 157 157 158 158 159 159 160 162 162 163 163 163

6.3 6.4

7

Structures and properties of transmission lines R. J. Collier 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 Introduction Coaxial lines Rectangular waveguides Ridged waveguide Microstrip Slot guide Coplanar waveguide Finline Dielectric waveguide References Further reading

8

Noise measurements David Adamson 8.1 8.2 Introduction Types of noise 8.2.1 Thermal noise 8.2.2 Shot noise 8.2.3 Flicker noise Definitions Types of noise source 8.4.1 Thermal noise sources 8.4.2 The temperature-limited diode 8.4.3 Gas discharge tubes 8.4.4 Avalanche diode noise sources

8.3 8.4

Contents 8.5 Measuring noise 8.5.1 The total power radiometer 8.5.2 Radiometer sensitivity Measurement accuracy 8.6.1 Cascaded receivers 8.6.2 Noise from passive two-ports Mismatch effects 8.7.1 Measurement of receivers and amplifiers Automated noise measurements 8.8.1 Noise figure meters or analysers 8.8.2 On-wafer measurements Conclusion Acknowledgements References

ix 164 164 166 166 169 169 171 172 174 175 175 176 176 176 179 179 180 180 181 181 182 182 185 187 188 189 190 192 193 194 198 200 201 203 207 207 208 208 214 216

8.6

8.7 8.8

8.9

9

Connectors, air lines and RF impedance N. M. Ridler 9.1 9.2 Introduction Historical perspective 9.2.1 Coaxial connectors 9.2.2 Coaxial air lines 9.2.3 RF impedance Connectors 9.3.1 Types of coaxial connector 9.3.2 Mechanical characteristics 9.3.3 Electrical characteristics Air lines 9.4.1 Types of precision air line 9.4.2 Air line standards 9.4.3 Conductor imperfections RF impedance 9.5.1 Air lines 9.5.2 Terminations Future developments Appendix: 7/16 connectors References

9.3

9.4

9.5

9.6

10

Microwave network analysers Roger D. Pollard 10.1 10.2 10.3 Introduction Reference plane 10.2.1 Elements of a microwave network analyser Network analyser block diagram Further reading

x 11

Contents RFIC and MMIC measurement techniques Stepan Lucyszyn 11.1 11.2 Introduction Test fixture measurements 11.2.1 Two-tier calibration 11.2.2 One-tier calibration 11.2.3 Test fixture design considerations Probe station measurements 11.3.1 Passive microwave probe design 11.3.2 Probe calibration 11.3.3 Measurement errors 11.3.4 DC biasing 11.3.5 MMIC layout considerations 11.3.6 Low-cost multiple DC biasing technique 11.3.7 Upper-millimetre-wave measurements Thermal and cryogenic measurements 11.4.1 Thermal measurements 11.4.2 Cryogenic measurements Experimental field probing techniques 11.5.1 Electromagnetic-field probing 11.5.2 Magnetic-field probing 11.5.3 Electric-field probing Summary References 217 217 218 220 229 230 230 231 236 240 240 241 243 243 246 246 247 249 249 250 251 254 255 263 263 263 263 266 267 267 267 269 269 270 270 270 271 272 273 273

11.3

11.4

11.5

11.6

12

Calibration of automatic network analysers Ian Instone 12.1 12.2 12.3 12.4 12.5 Introduction Definition of calibration Scalar network analysers Vector network analyser Calibration of a scalar network analyser 12.5.1 Transmission measurements 12.5.2 Reflection measurements Problems associated with scalar network analyser measurements Calibration of a vector network analyser Accuracy enhancement 12.8.1 What causes measurement errors? 12.8.2 Directivity 12.8.3 Source match 12.8.4 Load match 12.8.5 Isolation (crosstalk) 12.8.6 Frequency response (tracking)

12.6 12.7 12.8

Contents 12.9 12.10 12.11 12.12 Characterising microwave systematic errors 12.9.1 One-port error model One-port device measurement Two-port error model TRL calibration 12.12.1 TRL terminology 12.12.2 True TRL/LRL 12.12.3 The TRL calibration procedure Data-based calibrations References

xi 273 273 276 279 284 284 286 287 289 289 291 291 291 292 292 292 293 293 293 296 299 299 299 300 301 301 302 304 305 305 306 309 310 310 312 318 319 320

12.13

13

Verification of automatic network analysers Ian Instone 13.1 13.2 13.3 Introduction Definition of verification Types of verification 13.3.1 Verification of error terms 13.3.2 Verification of measurements Calibration scheme Error term verification 13.5.1 Effective directivity 13.5.2 Effective source match 13.5.3 Effective load match 13.5.4 Effective isolation 13.5.5 Transmission and reflection tracking 13.5.6 Effective linearity Verification of measurements 13.6.1 Customised verification example 13.6.2 Manufacturer supplied verification example References

13.4 13.5

13.6

14

Balanced device characterisation Bernd A. Schincke 14.1 14.2 Introduction 14.1.1 Physical background of differential structures Characterisation of balanced structures 14.2.1 Balanced device characterisation using network analysis 14.2.2 Characterisation using physical transformers 14.2.3 Modal decomposition method 14.2.4 Mixed-mode-S-parameter-matrix 14.2.5 Characterisation of single-ended to balanced devices 14.2.6 Typical measurements

xii

Contents 14.3 Measurement examples 14.3.1 Example 1: Differential through connection 14.3.2 Example 2: SAW-filter measurement (De)Embedding for balanced device characterisation Further reading 321 321 326 326 328 329 329 329 329 332 333 333 333 333 334 336 336 337 337 337 338 338 339 340 340 341 343 343 344 346 346 347 349 349 349 350 350 351 351

14.4

15

RF power measurement James Miall 15.1 15.2 Introduction Theory 15.2.1 Basic theory 15.2.2 Mismatch uncertainty 15.3 Power sensors 15.3.1 Thermocouples and other thermoelectric sensors 15.3.2 Diode sensors 15.3.3 Thermistors and other bolometers 15.3.4 Calorimeters 15.3.5 Force and field based sensors 15.3.6 Acoustic meter 15.4 Power measurements and calibration 15.4.1 Direct power measurement 15.4.2 Uncertainty budgets 15.5 Calibration and transfer standards 15.5.1 Ratio measurements 15.6 Power splitters 15.6.1 Typical power splitter properties 15.6.2 Measurement of splitter output match 15.6.3 The direct method of measuring splitter output 15.7 Couplers and reflectometers 15.7.1 Reflectometers 15.8 Pulsed power 15.9 Conclusion 15.10 Acknowledgements References

16

Spectrum analyser measurements and applications Doug Skinner 16.1 Part 1: Introduction 16.1.1 Signal analysis using a spectrum analyser 16.1.2 Measurement domains 16.1.3 The oscilloscope display 16.1.4 The spectrum analyser display 16.1.5 Analysing an amplitude-modulated signal

Contents 16.2 Part 2: How the spectrum analyser works 16.2.1 Basic spectrum analyser block diagram 16.2.2 Microwave spectrum analyser with harmonic mixer 16.2.3 The problem of multiple responses 16.2.4 Microwave spectrum analyser with a tracking preselector 16.2.5 Effect of the preselector 16.2.6 Microwave spectrum analyser block diagram 16.2.7 Spectrum analyser with tracking generator Part 3: Spectrum analyser important specification points 16.3.1 The input attenuator and IF gain controls 16.3.2 Sweep speed control 16.3.3 Resolution bandwidth 16.3.4 Shape factor of the resolution filter 16.3.5 Video bandwidth controls 16.3.6 Measuring low-level signals – noise 16.3.7 Dynamic range 16.3.8 Amplitude accuracy 16.3.9 Effect of input VSWR 16.3.10 Sideband noise characteristics 16.3.11 Residual responses 16.3.12 Residual FM 16.3.13 Uncertainty contributions 16.3.14 Display detection mode Spectrum analyser applications 16.4.1 Measurement of harmonic distortion 16.4.2 Example of a tracking generator measurement 16.4.3 Zero span 16.4.4 The use of zero span 16.4.5 Meter Mode 16.4.6 Intermodulation measurement 16.4.7 Intermodulation analysis 16.4.8 Intermodulation intercept point 16.4.9 Nomograph to determine intermodulation products using intercept point method 16.4.10 Amplitude modulation 16.4.11 AM spectrum with modulation distortion 16.4.12 Frequency modulation 16.4.13 FM measurement using the Bessel zero method 16.4.14 FM demodulation 16.4.15 FM demodulation display 16.4.16 Modulation asymmetry – combined AM and FM 16.4.17 Spectrum of a square wave 16.4.18 Pulse modulation 16.4.19 Varying the pulse modulation conditions

xiii 354 354 354 355 356 356 356 358 359 360 360 361 362 365 366 366 372 372 373 373 374 375 376 376 378 378 378 379 379 380 381 382 383 383 384 384 385 386 387 387 388 389 389

16.3

16.4

xiv

Contents 16.4.20 ‘Line’ and ‘Pulse’ modes 16.4.21 Extending the range of microwave spectrum analysers 16.4.22 EMC measurements 16.4.23 Overloading a spectrum analyser Conclusion Further reading 391 392 392 392 393 394

16.5

17

Measurement of frequency stability and phase noise David Owen 17.1 17.2 17.3 17.4 17.5 17.6 17.7 17.8 17.9 Measuring phase noise Spectrum analysers Use of preselecting filter with spectrum analysers Delay line discriminator Quadrature technique FM discriminator method Measurement uncertainty issues Future method of measurements Summary

395 396 397 399 400 401 404 405 406 406

18

Measurement of the dielectric properties of materials at RF and microwave frequencies Bob Clarke 18.1 18.2 18.3 Introduction Dielectrics – basic parameters Basic dielectric measurement theory 18.3.1 Lumped-impedance methods 18.3.2 Wave methods 18.3.3 Resonators, cavities and standing-wave methods 18.3.4 The frequency coverage of measurement techniques Loss processes: conduction, dielectric relaxation, resonances International standard measurement methods for dielectrics Preliminary considerations for practical dielectric measurements 18.6.1 Do we need to measure our dielectric materials at all? 18.6.2 Matching the measurement method to the dielectric material Some common themes in dielectric measurement 18.7.1 Electronic instrumentation: sources and detectors 18.7.2 Measurement cells 18.7.3 Q-factor and its measurement Good practices in RF and MW dielectric measurements

409 409 410 413 414 414 416 417 418 422 422 422 423 425 425 426 427 429

18.4 18.5 18.6

18.7

18.8

Contents A survey of measurement methods 18.9.1 Admittance methods in general and two- and three-terminal admittance cells 18.9.2 Resonant admittance cells and their derivatives 18.9.3 TE01 -mode cavities 18.9.4 Split-post dielectric resonators 18.9.5 Substrate methods, including ring resonators 18.9.6 Coaxial and waveguide transmission lines 18.9.7 Coaxial probes, waveguide and other dielectric probes 18.9.8 Dielectric resonators 18.9.9 Free-field methods 18.9.10 The resonator perturbation technique 18.9.11 Open-resonators 18.9.12 Time domain techniques 18.10 How should one choose the best measurement technique? 18.11 Further information References 19 Calibration of ELF to UHF wire antennas, primarily for EMC testing M. J. Alexander 19.1 19.2 Introduction Traceability of E-field strength 19.2.1 High feed impedance half wave dipole 19.2.2 Three-antenna method 19.2.3 Calculability of coupling between two resonant dipole antennas 19.2.4 Calculable field in a transverse electromagnetic (TEM) cell 19.2.5 Uncertainty budget for EMC-radiated E-field emission Antenna factors 19.3.1 Measurement of free-space AFs 19.3.2 The calculable dipole antenna 19.3.3 Calibration of biconical antennas in the frequency range 20–300 MHz 19.3.4 Calibration of LPDA antennas in the frequency range 200 MHz to 5 GHz 19.3.5 Calibration of hybrid antennas 19.3.6 Calibration of rod antennas 19.3.7 Calibration of loop antennas 19.3.8 Other antenna characteristics 18.9

xv 430 430 432 434 436 437 437 439 442 444 446 446 448 449 449 450

459 459 460 460 461 462 462 462 464 466 466 466 467 467 467 468 468

19.3

xvi

Contents 19.4 Electro-optic sensors and traceability of fields in TEM cells Acknowledgements References

469 470 470 473

Index

Contributors

David Adamson National Physical Laboratory Teddington, Middlesex, TW11 0LW david.adamson@npl.co.uk Martin J. Alexander National Physical Laboratory Teddington, Middlesex, TW11 0LW martin.alexander@npl.co.uk Bob Clarke National Physical Laboratory Teddington, Middlesex, TW11 0LW bob.clarke@npl.co.uk Richard J. Collier Corpus Christi College University of Cambridge Trumpington Street Cambridge, CB2 1RH rjc48@cam.ac.uk Alan Coster Consultant 5 Fieldfare Avenue, Yateley Hampshire, GU46 6PD ajcoster@theiet.org

John Hurll United Kingdom Accreditation Service 21–47 High Street, Feltham Middlesex, TW13 4UN john.hurll@ukas.com Ian Instone Agilent Technologies UK Ltd Scotstoun Avenue, South Queensferry West Lothian, EH30 9TG ian_instone@agilent.com Stepan Lucyszyn Dept Electrical and Electronic Engineering Imperial College London Exhibition Road, London, SW7 2AZ s.lucyszyn@imperial.ac.uk James Miall National Physical Laboratory Teddington, Middlesex, TW11 0LW james.miall@npl.co.uk David Owen Business Development Manager Pickering Interfaces 183A Poynters Road, Dunstable Bedfordshire, LU5 4SH david.owen@pickeringtest.com

xviii Contributors Roger D. Pollard School of Electronic and Electrical Engineering University of Leeds Leeds, LS2 9JT r.d.pollard@leeds.ac.uk Nick M. Ridler National Physical Laboratory Teddington, Middlesex, TW11 0LW nick.ridler@npl.co.uk Paul C.A. Roberts Fluke Precision Measurement Ltd Hurricane Way Norwich, NR6 6JB paul.roberts@fluke.com Bernd Schincke Rohde and Schwarz Training Centre Germany bernd.schincke@rohde-schwarz.com Doug Skinner Metrology Consultant doug.skinner@theiet.org Paul R. Young Department of Electronics University of Kent Canterbury, Kent, CT2 7NT P.R.Young@kent.ac.uk

Preface

This book contains most of the lecture notes used during the 14th IET Training Course on Microwave Measurements in May 2005 and is intended for use at the next course in 2007. These courses began in 1970 at the University of Kent with the title ‘RF Electrical Measurements’ and were held subsequently at the University of Surrey (1973) and the University of Lancaster (1976 and 1979). In 1983 the course returned to the University of Kent with the new title ‘Microwave Measurements’. In recent years it has been held appropriately at the National Physical Laboratory in Teddington, where the National Microwave Measurement Facilities and Standards are housed. In 1985 and again in 1989, the late A.E. Bailey was the editor of a publication of these notes in book form. This third edition has been jointly edited by myself and A.D. Skinner and includes a large number of new authors since the last edition. Although the book is primarily intended for the course, it has proved popular over the years to anyone starting to measure microwaves. The book begins with some revision chapters on transmission lines and scattering parameters, and these topics are used in the later chapters. Although these topics were included in most Physics and Electronics degree programmes in 1970, this is not the case now. As a result the reader who is totally unfamiliar with electromagnetic waves is advised to consult one of the many first-rate introductory books beforehand. The uncertainty of measurement is introduced next and this is followed by the techniques for the measurement of attenuation, voltage and noise. Many of these measurements are made using a microwave network analyser, and this remarkable instrument is discussed in the next few chapters. After this, the measurement of power, the use of spectrum analysers and aspects of digital modulation and phase noise measurements are covered in separate chapters. Finally, with the measurement of material properties, and both antenna and free field measurements, the wide range of topics is completed. A feature of the training course is that in addition to the lectures there are workshops and demonstrations using the excellent facilities available in the National Physical Laboratory. The course is organised by a small committee of the IET, many of whom are involved in both the lectures and the workshops. Without their hard work, the course would not have survived the endless changes that have occurred since 1970. Although the book is a good reference source for those starting in the field of microwave measurements, the course is strongly recommended, as not only

xx

Preface

will the lectures and workshops enliven the subject, but meeting others on the course will also help in forming useful links for the future. Finally, today microwaves are being used more extensively than ever before and yet there is a serious shortage of microwave metrologists. This book, and the course linked with it, are intended to help redress this imbalance. R.J. Collier Chairman – Organising Committee of the IET Training Course on Microwave Measurements

Chapter 1

Transmission lines – basic principles
R. J. Collier

1.1

Introduction

The aim of this chapter is to revise the basic principles of transmission lines in preparation for many of the chapters that follow in this book. Obviously one chapter cannot cover such a wide topic in any depth so at the end some textbooks are listed that may prove useful for those wishing to go further into the subject. Microwave measurements involve transmission lines because many of the circuits used are larger than the wavelength of the signals being measured. In such circuits, the propagation time for the signals is not negligible as it is at lower frequencies. Therefore, some knowledge of transmission lines is essential before sensible measurements can be made at microwave frequencies. For many of the transmission lines, such as coaxial cable and twisted pair lines, there are two separate conductors separated by an insulating dielectric. These lines can be described by using voltages and currents in an equivalent circuit. However, another group of transmission lines, often called waveguides, such as metallic waveguide and optical fibre have no equivalent circuit and these are described in terms of their electric and magnetic fields. This chapter will describe the two-conductor transmission lines first, followed by a description of waveguides and will end with some general comments about attenuation, dispersion and power. A subsequent chapter will describe the properties of some transmission lines that are in common use today.

1.2

Lossless two-conductor transmission lines – equivalent circuit and velocity of propagation

All two-conductor transmission lines can be described by using a distributed equivalent circuit. To simplify the treatment, the lines with no losses will be considered

2

Microwave measurements
I − C∆x I ∂V ∂t

L ∆x V

V − L∆x

∂I ∂t

∆x

Figure 1.1

The equivalent circuit of a short length of transmission line with no losses

first. The lines have an inductance per metre, L, because the current going along one conductor and returning along the other produces a magnetic flux linking the conductors. Normally at high frequencies, the skin effect reduces the self-inductance of the conductors to zero so that only this ‘loop’ inductance remains. The wires will also have a capacitance per metre, C, because any charges on one conductor will induce equal and opposite charges on the other. This capacitance between the conductors is the dominant term and is much larger than any self-capacitance. The equivalent circuit of a short length of transmission line is shown in Figure 1.1. If a voltage, V , is applied to the left-hand side of the equivalent circuit, the voltage at the right-hand side will be reduced by the voltage drop across the inductance. In mathematical terms V becomes V − L x Therefore, the change V = −L x Hence V ∂I = −L x ∂t and Lim V ∂V ∂I = = −L x ∂x ∂t x→0 ∂I ∂t ∂I in a distance ∂t x

V in that distance is given by

Transmission lines – basic principles

3

In a similar manner the current, I , entering the circuit on the left-hand side is reduced by the small current going through the capacitor. Again, in mathematical terms I becomes I − C x Therefore, the change I = −C x Hence ∂V I = −C x ∂t and Lim ∂I ∂V I = = −C x ∂x ∂t x→0 ∂V ∂t ∂V in a distance ∂t x

I in that distance is given by

The following equations are called Telegraphists’ equations. ∂I ∂V = −L ∂x ∂t ∂I ∂V = −C ∂x ∂t Differentiating these equations with respect to both x and t gives ∂ 2I ∂ 2V = −L ; ∂x∂t ∂x2 ∂ 2V ∂ 2I = −L 2 ; ∂t∂x ∂t ∂ 2V ∂ 2I = −C ∂x∂t ∂x2 2I ∂ ∂ 2V = −C 2 ∂t∂x ∂t

(1.1)

Given that x and t are independent variables, the order of the differentiation is not important; thus, the equations can be reformed into wave equations. ∂ 2V ∂ 2I ∂ 2I ∂ 2V = LC 2 ; = LC 2 2 2 ∂x ∂t ∂x ∂t (1.2)

The equations have general solutions of the form of any function of the variable (t ∓ x/v). So if any signal, which is a function of time, is introduced at one end of a lossless transmission line then at a distance x down the line, this function will be delayed by x/v. If the signal is travelling in the opposite direction then the delay will be the same except that x will be negative and the positive sign in the variable will be necessary.

4

Microwave measurements

If the function f (t − x/v) is substituted into either of the wave equations this gives 1 f v2 t− x = LCf v t− x v

This shows that for all the types of signal – pulse, triangular and sinusoidal – there is a unique velocity on lossless lines, v, given by 1 v=√ LC (1.3)

This is the velocity for both the current and voltage waveforms, as the same wave equation governs both parameters.

1.2.1 Characteristic impedance
The relationship between the voltage waveform and the current waveform is derived from the Telegraphists’ equations. If the voltage waveform is V = V0 f t − then V0 ∂V =− f ∂x v V0 ∂I = f ∂t Lv t− x v x v

Using the first Telegraphists’ equation t− x v

Integrating with respect to time gives I= x V V0 f t− = Lv v Lv L C

Therefore V = Lv = I V = Z0 I (1.4)

This ratio is called the characteristic impedance, Z0 and for lossless lines (1.5)

This is for waves travelling in a positive x direction. If the wave was travelling in a negative x direction, i.e. a reverse or backward wave, then the ratio of V to I would be equal to −Z0 .

Transmission lines – basic principles

5

1.2.2 Reflection coefficient
A transmission line may have at its end an impedance, ZL , which is not equal to the characteristic impedance of the line Z0 . Thus, a wave on the line faces the dilemma of obeying two different Ohm’s laws. To achieve this, a reflected wave is formed. Giving positive suffices to the incident waves and negative suffices to the reflected waves, the Ohm’s law relationships become V+ V− = Z0 ; = −Z0 I− I+ VL V+ + V− = = ZL IL I+ + I − where VL and IL are the voltage and current in the terminating impedance ZL . A reflection coefficient, ρ or , is defined as the ratio of the reflected wave to the incident wave. Thus = So ZL = and = ZL − Z0 ZL + Z0 (1.7) Z0 (1 + ) V+ (1 + ) = I+ (1 − ) (1 − ) V− I− =− V+ I+ (1.6)

As Z0 for lossless lines is real and ZL may be complex, , in general, will also be complex. One of the main parts of microwave impedance measurement is to measure the value of and hence ZL .

1.2.3 Phase velocity and phase constant for sinusoidal waves
So far, the treatment has been perfectly general for any shape of wave. In this section, just the sine waves will be considered. In Figure 1.2, a sine wave is shown at one instant in time. As the waves move down the line with a velocity, v, the phase of the waves further down the line will be delayed compared with the phase of the oscillator on the left-hand side of Figure 1.2. The phase delay for a whole wavelength is equal to 2π . The phase delay per metre is called β and is given by β= 2π λ (1.8)

6

Microwave measurements
2π radians of phase delay per wavelength

Figure 1.2

Sine waves on transmission lines

Multiplying numerator and denominator of the right-hand side by frequency gives β= √ ω 2πf = = ω LC λf v (1.9)

where v is now the phase velocity, i.e. the velocity of a point of constant phase and is the same velocity as that given in Section 1.2.

1.2.4 Power flow for sinusoidal waves
If a transmission line is terminated in impedance equal to Z0 then all the power in the wave will be dissipated in the matching terminating impedance. For lossless lines, a sinusoidal wave with amplitude V1 the power in the termination would be
2 V1 2Z0

This is also the power in the wave arriving at the matched termination. If a transmission line is not matched then part of the incident power is reflected (see Section 1.2.2) and if the amplitude of the reflected wave is V2 then the reflected power is
2 V2 2Z0

Since | |= then | |2 = Power reflected Incident power (1.10) V2 V1

Transmission lines – basic principles

7

Clearly, for a good match the value of | | should be near to zero. The return loss is often used to express the match Return loss = 10 log10 1 | |2 (1.11)

In microwave circuits a return loss of greater than 20 dB means that less than 1% of the incident power is reflected. Finally, the power transmitted into the load is equal to the incident power minus the reflected power. A transmission coefficient, τ , is used as follows: Transmitted power = |τ |2 = 1 − | |2 Incident power (1.12)

1.2.5 Standing waves resulting from sinusoidal waves
When a sinusoidal wave is reflected by terminating impedance, which is not equal to Z0 , the incident and reflected waves form together a standing wave. If the incident wave is V+ = V1 sin(ωt − βx) and the reflected wave is V− = V2 sin(ωt + βx) where x =0 at the termination, then at some points on the line the two waves will be in phase and the voltage will be VMAX = (V1 + V2 ) where VMAX is the maximum of the standing wave pattern. At other points on the line the two waves will be out of phase and the voltage will be VMIN = (V1 − V2 ) whereVMIN is the minimum of the standing wave pattern. The Voltage Standing Wave Ratio (VSWR) or S is defined as S= Now S= V1 (1 + | |) V1 + V2 = V1 − V 2 V1 (1 − | |) VMAX VMIN (1.13)

8 So

Microwave measurements

S= or

1+| | 1−| | S −1 S +1

(1.14)

| |=

(1.15)

Measuring S is relatively easy and, therefore, a value for | | can be obtained. From the position of the maxima and minima the argument or phase of can be found. For instance, if a minimum of the standing wave pattern occurs at a distance D from a termination then the phase difference between the incident and the reflected waves at that point must be nπ(n = 1, 3, 5, . . .). Now the phase delay as the incident wave goes from that point to the termination is βD. The phase change on reflection is the argument of . Finally, the further phase delay as the reflected wave travels back to D is also βD. Hence nπ = 2βD + arg( ) Therefore by a measurement of D and from knowledge of β, the phase of be measured. (1.16) can also

1.3

Two-conductor transmission lines with losses – equivalent circuit and low-loss approximation

In many two-conductor transmission lines there are two sources of loss, which cause the waves to be attenuated as they travel along the line. One source of loss is the ohmic resistance of the conductors. This can be added to the equivalent circuit by using a distributed resistance, R, whose units are Ohms per metre. Another source of loss is the ohmic resistance of the dielectric between the lines. Since, this is in parallel with the capacitance it is usually added to the equivalent circuit using a distributed conductance, G, whose units are Siemens per metre. The full equivalent circuit is shown in Figure 1.3.

L

R

C

G

Figure 1.3

The equivalent circuit of a line with losses

Transmission lines – basic principles The Telegraphists’ equations become ∂I ∂V = −L − RI ∂x ∂t ∂I ∂V = −C − GV ∂x ∂t

9

(1.17)

Again, wave equations for lines with losses can be found by differentiating with respect to both x and t. ∂ 2V ∂V ∂ 2V = LC 2 + (LG + RC) + RGV 2 ∂t ∂x ∂t ∂ 2I ∂ 2I ∂I = LC 2 + (LG + RC) + RGI 2 ∂t ∂x ∂t

(1.18)

These equations are not easy to solve in the general case. However, for sinusoidal waves on lines with small losses, i.e. ωL R; ωC G, there is a solution of the form: V = V0 exp (−αx) sin(ωt − βx) where 1 v = √ m s−1 (same as for lossless lines) LC R GZ0 α= + nepers m−1 2Z0 2 or α = 8.686 GZ0 R + dB m−1 2Z0 2 (1.21) (1.22) (1.23) (1.19) (1.20)

√ β = ω LC radians m−1 (same as for lossless lines) Z0 = L (same as for lossless lines) C

1.3.1 Pulses on transmission lines with losses
As well as attenuation, a pulse on a transmission line with losses will also change its shape. This is mainly caused by the fact that a full solution of the wave equations for lines with losses gives a pulse shape which is time dependent. In addition, the components of the transmission lines L, C, G and R are often different functions of frequency. Therefore if the sinusoidal components of the pulse are considered separately, they all travel at different velocities and with different attenuation. This frequency dependence is called dispersion. For a limited range of frequencies, it is sometimes possible to describe a group velocity, which is the velocity of the pulse rather than the velocity of the individual sine waves that make up the pulse. One effect

10

Microwave measurements

of dispersion on pulses is that the rise time is reduced and often the pulse width is increased. It is beyond the scope of these notes to include a more detailed treatment of this topic.

1.3.2 Sinusoidal waves on transmission lines with losses
For sinusoidal waves, there is a general solution of the wave equation for lines with losses and it is α + jβ = (R + jωL) (G + jωC) and Z0 = R + jωL G + jωC (1.24)

In general, α, β and Z0 are all functions of frequency. In particular, Z0 at low frequencies can be complex and deviate considerably from its high-frequency value. As R, G, L and C also vary with frequency, a careful measurement of these properties at each frequency is required to characterise completely the frequency variation of Z0 .

1.4

Lossless waveguides

These transmission lines cannot be easily described in terms of voltage and current as they sometimes have only one conductor (e.g. metallic waveguide) or no conductor (e.g. optical fibre). The only way to describe their electrical properties is in terms of the electromagnetic fields that exist in and, in some cases, around their structure. This section will begin with a revision of the properties of a plane or transverse electromagnetic (TEM) wave. The characteristics of metallic waveguides will then be described using these waves. The properties of other wave guiding structures will be given in Chapter 7.

1.4.1 Plane (or transverse) electromagnetic waves
A plane or TEM wave has two fields that are perpendicular or transverse to the direction of propagation. One of the fields is the electric field and the direction of this field is usually called the direction of polarisation (e.g. vertical or horizontal). The other field which is at right angles to both the electric field and the direction of propagation is the magnetic field. These two fields together form the electromagnetic wave. The electromagnetic wave equations for waves propagating in the z direction are as follows: ∂ 2E ∂ 2E = µε 2 ∂z 2 ∂t (1.25) ∂ 2H ∂ 2H = µε 2 ∂z 2 ∂t where µ is the permeability of the medium. If µR is the relative permeability then µ = µR µ0 and µ0 is the free space permeability and has a value of 4π ×10−7 H m−1 .

Transmission lines – basic principles

11

Similarly, ε is the permittivity of the medium and if εR is the relative permittivity then ε = εR ε0 and ε0 is the free space permittivity and has a value of 8.854 × 10−12 F m−1 . These wave equations are analogous to those in Section 1.2. The variables V , I , L and C are replaced with the new variables E, H , µ and ε and the same results follow. For a plane wave the velocity of the wave, v, in the z direction is given by 1 1 see Section 2 where v = √ v= √ µε LC (1.26)

If µ = µ0 and ε = ε0 , then v0 = 2.99792458 × 108 m s−1 . The ratio of the amplitude of the electric field to the magnetic field is called the intrinsic impedance and has the symbol η. η= µ E = ε H see Section 1.2 where Z0 = L C (1.27)

If µ = µ0 and ε = ε0 , then η0 = 376.61 or 120 π . As in Section 1.2, fields propagating in the negative z direction are related by using −η. The only difference is the orthogonality of the two fields in space, which comes from Maxwell’s equations. For an electric field polarised in the x direction ∂Hy ∂Ex = −µ ∂z ∂t z If Ex is a function of (t − ) as before then v ∂Hy z 1 ∂Ex = − E0 f t − = −µ ∂z v v ∂t z 1 1 = Ex Hy = E0 f t − v µv µv Therefore Ex = µv = Hy µ =η ε (1.28)

If an Ey field was chosen, the magnetic field would be in the negative x direction. For sinusoidal waves, the phase constant is called the wave number and a symbol k is assigned. √ √ (1.29) k = ω µε see Section 1.2 where β = ω LC Inside all two-conductor transmission lines are various shapes of plane waves and it is possible to describe them completely in terms of fields rather than voltages and currents. The electromagnetic wave description is more fundamental but the equivalent circuit description is often easier to use. At high frequencies, two-conductor transmission lines also have higher order modes and the equivalent circuit model for

12

Microwave measurements

these becomes more awkward to use whereas the electromagnetic wave model is able to accommodate all such modes.

1.4.2 Rectangular metallic waveguides
Figure 1.4 shows a rectangular metallic waveguide. If a plane wave enters the waveguide such that its electric field is in the y (or vertical) direction and its direction of propagation is not in the z direction then it will be reflected back and forth by the metal walls in the y direction. Each time the wave is reflected it will have its phase reversed so that the sum of the electric fields on the surfaces of the walls in the y direction is zero. This is consistent with the walls being metallic and therefore good conductors capable of short circuiting any electric fields. The walls in the x direction are also good conductors but are able to sustain these electric fields perpendicular to their surfaces. Now, if the wave after two reflections has its peaks and troughs in the same positions as the original wave then the waves will add together and form a mode. If there is a slight difference in the phase, then the vector addition after many reflections will be zero and so no mode is formed. The condition for forming a mode is thus a phase condition and it can be found as follows. Figure 1.5 shows a plane wave in a rectangular metallic waveguide with its electric field in the y direction and the direction of propagation at an angle θ to the z direction. The rate of change of phase in the direction of propagation is k0 , the free space wave number. As this wave is incident on the right wall of the waveguide in Figure 1.5, it will be reflected according to the usual laws of reflection as shown in Figure 1.6. On further reflection, this wave must ‘rejoin’ the original wave to form a mode. Therefore, Figure 1.6 also shows the sum of all the reflections forming two waves – one incident on the right wall and the other on the left wall. The waves form a mode if they are linked together in phase. Consider the line AB. This is a line of constant phase for the wave moving towards the right. Part of that wave at B reflects and moves along BA to A where it reflects again and rejoins the wave with the same phase. At the first reflection there is a phase shift of π . Then along BA there is a phase delay followed by another phase shift of π at the second reflection. The phase condition is 2π + phase delay along BA for wave moving to the left = 2mπ where m = 0, 1, 2, 3, . . ..

a b a = 2b

Figure 1.4

A rectangular metallic waveguide

Transmission lines – basic principles
0° Direction of propagation or wave vector 180° Metal waveguide wall θ 0° Metal waveguide wall

13

180° z 0° y⊕ x Lines of constant phase on waveform

a

Figure 1.5

A plane wave in a rectangular metallic waveguide
0° 0°

180°

180°

θ A 90° − 2θ 0°

θ 0°

Fields cancel at metal walls

180°

180°

B

0°

0°

Fields add at the centre TE10 mode

Figure 1.6

Two plane waves in a rectangular metallic waveguide. The phases 0o and 180o refer to the lines below each figure

14

Microwave measurements The phase delay can be found by resolving the wave number along BA and is k0 sin 2θ radians m−1

If the walls in the y direction are separated by a distance a, then AB = a cos θ

Therefore the phase delay is k0 sin 2θ(a/cos θ) or 2k0 sin θ . Hence, the phase condition is k0 a sin θ = mπ where m = 0, 1, 2, 3, … (1.30)

The solutions to this phase condition give the various waveguide modes for waves with fields only in the y direction, i.e. the TEmo modes, ‘m’ is the number of half sine variations in the x direction.

1.4.3 The cut-off condition
Using the above phase condition, if k0 = ω/v0 then ω sin θ = v0 mπ a

The terms on the right-hand side are constant. For very high frequencies the value of θ tends to zero and the two waves just propagate in the z direction. However, if ω reduces in value the largest value of sin θ is 1 and at this point the mode is cut-off and can no longer propagate. The cut-off frequency is ωc and is given by ωc = or fc = v0 m 2a (1.31) v0 mπ a

where fc is the cut-off frequency. If v0 = λc fc then λc = 2a m (1.32)

where λc is the cut-off wavelength.

Transmission lines – basic principles

15

A simple rule for TEmo modes is that at cut-off, the wave just fits in ‘sideways’. Indeed, since θ = 90◦ at cut-off, the two plane waves are propagating from side to side with a perfect standing wave between the walls.

1.4.4 The phase velocity
All waveguide modes can be considered in terms of plane waves. As the simpler modes just considered consist of only two plane waves they form a standing wave pattern in the x direction and form a travelling wave in the z direction. As the phase velocity in the z direction is related to the rate of change of phase, i.e. the wave number, then Velocity in the z direction = ω Wave number in the z direction

Using Figure 1.5 or 1.6 the wave number in the z direction is k0 cos θ Now from the phase condition sin θ = mπ/k0 a. So k0 cos θ = k0 1 − Also k0 = mπ k0 a
2 1/ 2

2π 2π f ω = = λ0 λ0 f v0

√ where v0 is the free space velocity = 1/ µ0 ε0 . Hence the velocity in the z direction vz = vz = 1− ω k0 cos θ

v0 mλ0 2a

2 1/ 2

= 1−

v0 λ0 λc

2 1/ 2

(1.33)

As can be seen from this condition, when λ0 is equal to the cut-off wavelength (see Section 1.4.3) then vz is infinite. As λ0 gets smaller than λc then the velocity approaches v0 . Thus the phase velocity is always greater than v0 . Waveguides are not normally operated near cut-off as the high rate of change of velocity means both impossible design criteria and high dispersion.

1.4.5 The wave impedance
The ratio of the electric field to the magnetic field for a plane wave has already been discussed in Section 1.4.1. Although the waveguide has two plane waves in it the

16

Microwave measurements

wave impedance is defined as the ratio of the transverse electric and magnetic fields. For TE modes this is represented by using the symbol ZTE . ZTEmo = Ey η E0 = = cos θ Hx H0 cos θ

where E0 and H0 refer to the plane waves. The electric fields of the plane waves are in the y direction but the magnetic fields are at an angle θ to the x direction. Therefore ZTEmo = 1− Thus, for the λ0 might be 500 . η mλ0 2a
2 1/ 2

= 1−

η λ0 λc
2 1/ 2

(1.34)

λc the value of ZTEmo is always greater than η0 . A typical value

1.4.6 The group velocity
Since a plane wave in air has no frequency-dependent parameters similar to those of the two-conductor transmission lines, i.e. µ0 and ε0 are constant, then there is no dispersion and so the phase velocity is equal to the group velocity. A pulse in a waveguide therefore would travel at v0 at an angle of θ to the z-axis. The group velocity, vg , along the z-axis is given by vg = v0 cos θ = v0 1 − mλ0 2a
2 1/2

(1.35)

The group velocity is always less than v0 and is a function of frequency. For rectangular metallic waveguides
2 Phase velocity × Group velocity = v0

(1.36)

1.4.7 General solution
To obtain all the possible modes in a rectangular metallic waveguide the plane wave must also have an angle ψ to the z-axis in the y–z plane. This will involve the wave reflecting from all four walls. If the two walls in the x direction are separated by a distance b then the following are valid for all modes. If A= 1− mλ0 2a
2

−

nλ0 2b

2 1/2

(1.37)

m = 0, 1, 2, . . . and n = 0, 1, 2, . . .

Transmission lines – basic principles
TE32 TM32 TE50 fcTE 5 0 a = 2b Region of monomode propagation TM11 TE11 TM21 TE21 TM31 TE31 TE22 TM22 TM41 TE41 1 2 8 3 13 4 20 5 =

17

TE10

TE01 TE20

TE30

TE02 TE40

mm 10

fcTE

m 2 + 4n 2

Figure 1.7

Relative cut-off frequencies for rectangular metallic waveguides

then the velocity v0 v= A η ZTE = A vg = Av0

(1.38)

The modes with the magnetic field in the y direction – the dual of TE modes – are called transverse magnetic modes or TM modes. They have a constraint that neither m nor n can be zero as the electric field for these modes has to be zero at all four walls. ZTM = ηA (1.39) The relative cut-off frequencies are shown in Figure 1.7, which also shows that monomode propagation using the TE10 mode is possible at up to twice the cut-off frequency. However, the full octave bandwidth is not used because propagation near cut-off is difficult and just below the next mode it can be hampered by energy coupling into that mode as well.

Further reading
1 Ramo, S., Whinnery, J. R., and VanDuzer, T.: Fields and Waves in Communication Electronics, 3rd edn (Wiley, New York, 1994) 2 Marcowitz, N.: Waveguide Handbook (Peter Peregrinus, London, 1986) 3 Cheng, D. K.: Field and Wave Electromagnetics, 2nd edn (Addison-Wesley, New York, 1989)

18

Microwave measurements

4 Magid, L. M.: Electromagnetic Fields and Waves (Wiley, New York, 1972) 5 Kraus, J. D., and Fleisch, D. A.: Electromagnetics with Applications (McGrawHill, Singapore, 1999) 6 Jordan, E. C., and Balmain, K. G.: Electromagnetic Waves and Radiating Systems, 2nd edn (Prentice-Hall, New Jersey, 1968) 7 Chipman, R. A.: Transmission Lines, Schaum’s Outline Series (McGraw-Hill, New York, 1968)

Chapter 2

Scattering parameters and circuit analysis
P. R. Young

2.1

Introduction

Scattering parameters or scattering coefficients are fundamental to the design, analysis and measurement of all microwave and millimetre-wave circuits and systems. Scattering parameters define the forward and reverse wave amplitudes at the inputs and outputs of a network. Microwave networks take on various forms and can be as simple as a shunt capacitor or as complicated as a complete system. Common microwave networks are one-, two-, three- or four-port devices. The definition of a scattering parameter is intrinsically linked to the form of transmission medium used at the ports of the network. Transmission lines and waveguides come in four distinct classes: (1) transverse electromagnetic (TEM), which includes coaxial lines and parallel pairs; (2) quasi-TEM lines, such as microstrip and coplanar waveguide (CPW); (3) transverse electric (TE) and transverse magnetic (TM) waveguides, such as rectangular waveguide; and (4) hybrid waveguides, which include dielectric guides and most lossy transmission lines and waveguides. For simplicity, the one-port scattering parameter or reflection coefficient will be defined first for TEM lines before the more complicated multi-port and waveguide networks are analysed.

2.2

One-port devices

Consider a simple two-conductor transmission line, such as a coaxial cable or parallel pair. These types of transmission lines support TEM waves allowing the wave transmission to be expressed purely in terms of voltage between the conductors and the current flowing through the conductors. If the line is terminated by a load Z (Figure 2.1), which is not perfectly matched with the transmission line, then some

20

Microwave measurements

V+ Z V−

Figure 2.1

Transmission line terminated in a mismatched load

of the incident waves will be reflected back from the load. In terms of voltage and current along the line we have, at a point z, V (z) = V+ e−jβz + V− e+jβz and I (z) = I+ e−jβz − I− e+jβz where Z0 = V+ V− = I+ I− or I (z) = 1 V+ e−jβz − V− e+jβz Z0 (2.2) (2.1)

β is the phase constant (rad m−1 ), V+ and I+ are, respectively, amplitude of the voltage and current of the forward propagating wave; V− and I− are that of the reverse. The e−jβz terms denote forward propagation (towards the load), whereas the e+jβz terms denote reverse propagation (away from the load). Z0 is the characteristic impedance of the transmission line and is dependent on the geometry and material of the structure. For simplicity, the time dependence has been omitted from Equations 2.1 and 2.2. The actual voltage and current are given by Re{V(z)ejωt } and Re{I(z)ejωt }, respectively. Suppose we choose a reference plane at the termination where we set z = 0. We define the reflection coefficient at this point by = V− V+ (2.3)

Since Ohm’s law must apply at the termination Z= V I (2.4)

Note that we are free to set the reference plane anywhere along the line; hence, this might be at the connector interface, at the load element or some distance along the line.

Scattering parameters and circuit analysis

21

Substituting (2.1) and (2.2) into (2.4) gives the well-known relationship between the reflection coefficient of the termination and its impedance at the reference plane Z = Z0 V+ + V − 1+ = Z0 1− V+ − V − (2.5)

Similarly, the relationship between the admittance of the termination and its reflection coefficient is given by Y = Z −1 = Y0 V+ − V − 1− = Y0 V+ + V − 1+ (2.6)

−1 where Y0 is the admittance of the transmission line, Y0 = Z0 . Another useful expression is given by solving for in (2.5)

=

Y0 − Y Z − Z0 = Z + Z0 Y0 + Y

(2.7)

Clearly, is dependent on the impedance of the termination but we note that it is also dependent on the characteristic impedance of the line. A knowledge of Z0 is therefore required to define . Relationships for the power flow can also be defined. It is well known that, using phasor notation, the root mean square power is given by P= 1 Re{VI ∗ } 2

where ‘*’ denotes the complex conjugate. Substituting for V and I from (2.1) and (2.2) yields, for z = 0, P= 1 Re (V+ + V− ) (V+ − V− )∗ 2Z0

which gives P= 1 |V+ |2 − |V− |2 2Z0

where we have made use of the fact that V+ ∗ V− − V+ V− ∗ is purely imaginary and V+ V+ ∗ = |V+ |2 . We also assume that Z0 is purely real. Note that the power is dependent on the characteristic impedance. We can, however, define the network in terms of another set of amplitude constants such that the impedance is not required in power calculations. Let V = Z0 ae−jβz + be+jβz (2.8)

22 and

Microwave measurements

1 ae−jβz − be+jβz I=√ Z0

(2.9)

where a and b are defined as the wave amplitudes of the forward and backward √ propagating waves. With reference to (2.1) it is easy to show that a = V+ / Z0 and √ b = V− / Z0 . We find now, if Z0 is purely real, that the power is simply given by P= 1 |a|2 − |b|2 2 (2.10)

|a|2 /2 is the power in the forward propagating wave and |b|2 /2 is the power in the backward propagating wave. Equation (2.10) is a very satisfying result since it allows propagation to be defined in terms of wave amplitudes that are directly related to the power in the wave. This is particularly useful for measurement purposes since power is more easy to measure than voltage or current. In fact we shall see that for many microwave networks, voltage and current cannot be measured or even defined. The analysis so far has dealt with one-port devices. These are completely specified by their impedance Z or reflection coefficient (with respect to Z0 ). The more important case of the two-port, or multi-port, device requires a more complicated model.

2.3

Generalised scattering parameters

Consider the two-port network shown in Figure 2.2. There will, in general, be waves propagating into and out of each of the ports. If the device is linear, the output waves can be defined in terms of the input waves. Thus, b1 = S11 a1 + S12 a2 b2 = S21 a1 + S22 a2 (2.11) (2.12)

where b1 and b2 are the wave amplitudes of the waves flowing out of ports 1 and 2, respectively. Similarly, a1 and a2 are the wave amplitudes of the waves flowing into ports 1 and 2, respectively. S11 , S21 , S12 and S22 are the scattering coefficients or

a1 S b1

a2 b2

Figure 2.2

Two-port device represented by S-parameter matrix

Scattering parameters and circuit analysis

23

scattering parameters. Using the definition of wave amplitude the voltages at port 1 and port 2 are given by V1 = Z01 (a1 + b1 ) and V2 = Z02 (a2 + b2 )

respectively, where it is assumed that the characteristic impedance is different at each port: Z01 at port 1 and Z02 at port 2. Similarly, the currents entering port 1 and port 2 are I1 = √ 1 (a1 − b1 ) Z01 and 1 I2 = √ (a2 − b2 ) Z02

respectively. Equations 2.11 and 2.12 can be more neatly written in matrix notation S b1 = 11 b2 S21 or b = Sa, where a= b a1 ,b = 1 a2 b2 and S= S11 S21 S12 S22 S12 S22 a1 a2 (2.13)

where S is the scattering matrix or S-parameter matrix of the two-port network. If port 2 is terminated by a perfect match of impedance Z02 , that is, all of the incident power is absorbed in the termination, then we have the following properties: S11 = b1 a1 and S21 = b2 a1

a2 =0

a2 =0

Similarly, if port 1 is terminated by a perfect match of impedance Z01 then S22 = b2 a2 and S12 = b1 a2

a1 =0

a1 =0

By using the above definitions we can obtain some insight into the meaning of the individual S-parameters. S11 is the reflection coefficient at port 1 with port 2 terminated in a matched load. It, therefore, gives a measure of the mismatch due to the network and not any other devices that may be connected to port 2. S21 is the transmission coefficient from port 1 to port 2 with port 2 terminated in a perfect match. It gives a measure of the amount of signal that is transmitted from port 1 to port 2. S22 and S12 are similarly defined with S22 giving the reflection from port 2 and S12 the transmission from port 2 to port 1.

24

Microwave measurements

The S-parameter representation equally applies to multi-port devices. For an n-port device, the S-parameter matrix is given by      S11 S12 · · · S1n a1 b1 b2  S21 S22 · · · S2n  a2       .  = . . .  .  .. . .  .  .  . . . . . . . an bn Sn1 Sn2 · · · Snn where bk is the amplitude of the wave travelling away from the junction at port k. Similarly, ak is the wave amplitude travelling into the junction at port k. The S-parameters are defined as Sij = bi aj

ak =0 for k =j

2.4

Impedance and admittance parameters

Expressions similar to (2.5) and (2.7) can be obtained for n-port devices. If we have an n-port device then the voltage and current at the reference plane of port k are given by Vk = and Ik = √ 1 (ak − bk ) Z0k (2.15) Z0k (ak + bk ) (2.14)

where Z0k is the characteristic impedance of the transmission line connected to port k. Equations 2.14 and 2.15 can be written in matrix notation as V = Z0 (a + b) and I = Z0
−1/2 1/2

(2.16)

(a − b)

(2.17)

respectively. Where a and b are column vectors containing the wave amplitudes and V and I are column vectors containing the port voltages and currents         b1 V1 I1 a1 b2  V2  I2  a2          a = .  , b = .  , V = .  , and I = .  .  .  .  .  . . . . an bn Vn In

Scattering parameters and circuit analysis Z0 is a diagonal matrix with Z0k as its diagonal elements   0 ··· 0 Z01  0 Z02 · · · 0    Z0 =  . . .  .. . .   . . . . . 0 0 · · · Z0n

25

(2.18)

√ 1/2 Z0 denotes a diagonal matrix with Z0k as its diagonal elements. Often, the characteristic impedance of each of the ports is identical, in which case each of the diagonal elements is equal. From (2.8) and (2.9), with z = 0, we have a= and b= 1 −1/2 (V − Z0 I) Z 2 0 (2.20) 1 −1/2 (V + Z0 I) Z 2 0 (2.19)

Let V = ZI, where Z is the impedance matrix, extensively used in electrical circuit theory   Z11 Z12 · · · Z1n Z21 Z22 · · · Z2n    Z= . . .  .. . .   . . . . . Zn1 Zn2 · · · Znn Also let I = YV, where Y is the admittance matrix   Y11 Y12 · · · Y1n Y21 Y22 · · · Y2n    Y= . . . . .. . .   . . . . . Yn1 Yn2 · · · Ynn We note that Z = Y−1 . The individual elements of the impedance and admittance matrices are defined as follows: Zij = Vi Ij and Yij = Ii Vj

Ik =0 for k =j

Vk =0 for k =j

That is, Zij is the ratio of voltage at port i to the current at port j with all other port currents set to zero, that is, short circuit. Yij is defined as the ratio of current at port i to the voltage at port j with all other port voltages set to zero, that is, open circuit. Substituting V = ZI and b = Sa into (2.19) and (2.20) yields Z = Z0 (U − S)−1 (U + S) Z0
1/2 1/2

= Y−1

(2.21)

26

Microwave measurements Network parameters for common microwave networks
Network parameters

Table 2.1

Circuit
Lossless transmission line of length L, phase constant and characteristic impedance Z0

S=

e−jβL

0

e−jβL 0

Shunt admittance Y

S= Y Z=

− Y 2Y0 1 Y + 2Y0 2Y0 −Y Z Z Z Z

where Z =Y −1
Series impedanceZ

S= Z

1 2Z0 + Z

Z 2Z0 2Z0 Z

Y=

Y −Y −Y Y

where Y = Z −1 π network

YA

YC YB

Y=

YA + YC −YC −YC YB + YC

T network

ZA ZC

ZB

Z=

ZA + ZC ZC ZC ZB + ZC

Scattering parameters and circuit analysis or solving for S S = Z0
−1/2

27

(Z − Z0 ) (Z + Z0 )−1 Z0

1/2

(2.22)

where U is the unit matrix:   1 0 ··· 0 0 1 · · · 0   U = . . . .. . . . . . . . 0 0 ··· 1 Examples of (2.21) and (2.22) for two-port networks are given in Appendix 2.C. Z and Y parameters can be very useful in the analysis of microwave networks since they can be related directly to simple π or T networks (refer to Table 2.1). These circuits are fundamental in lumped element circuits, such as attenuators, and are also important in equivalent circuits for waveguide junctions and discontinuities.

2.4.1 Examples of S-parameter matrices
Table 2.1 shows some common examples of microwave networks and their network parameters. Parameters are only shown for the simplest form. The associated S, Z or Y parameters can be determined using (2.21) and (2.22). In each case it is assumed −1 that the characteristic impedance is identical at each port and equal to Z0 = Y0 . We notice from the table that the S-parameter matrices are symmetrical, that is, Smn = Snm . This is a demonstration of reciprocity in microwave networks and applies to most networks (see Appendix 2.A). A property of lossless scattering matrices is also seen for the line section. Here, S T S ∗ = U, which applies to all lossless networks; refer to Appendix 2.B.

2.5

Cascade parameters

Another useful transformation of the S-parameter matrix is the cascade matrix. The two-port cascade matrix is given by T a1 = 11 T21 b1 T12 T22 b2 a2 (2.23)

where we notice that the wave amplitudes on port 1 are given in terms of the wave amplitudes on port 2. Note that some textbooks interchange a1 with b1 and b2 with a2 . Comparing (2.23) with (2.13) gives the following relationships between the cascade matrix elements and the scattering coefficients: T = T11 T21 1 T12 1 = T22 S11 S21 −S22 S12 S21 − S11 S22 (2.24)

28

Microwave measurements a1 b1 b2 a2 a1′ b1′ a2′ b2′

S T

S′ T′

Figure 2.3

Two two-port networks cascaded together

Similarly, the reverse transform is given by S= 1 T21 T11 1 T22 T11 − T12 T21 −T12 (2.25)

Suppose we have two two-port devices cascaded together (refer to Figure 2.3). The first network is given by T a1 = 11 b1 T21 T12 T22 b2 a2

and the second network T a1 = 11 T21 b1 T12 T22 b2 a2

where, by inspecting Figure 2.3 we see that a b2 = 1 a2 b1 Therefore, T a1 = 11 b1 T21 T12 T22 T11 T21 T12 T22 b2 a2

We see that in order to calculate the input wave amplitudes in terms of the output amplitudes we simply multiply the cascade matrices together. Often the cascaded two port is converted back to an S-parameter matrix using (2.25). Any number of cascaded two-port networks can then be replaced by a single equivalent two-port network.

2.6

Renormalisation of S-parameters

We have already seen that S-parameters are defined with respect to a reference characteristic impedance at each of the network’s ports. Often we require that the S-parameters are renormalised to another set of port characteristic impedances. This is important as measured S-parameters are usually with respect to the transmission line Z0 or matched load impedance of the calibration items used in the measurement system. These often differ from the nominal 50 . To convert an S-parameter

Scattering parameters and circuit analysis matrix S, that is, with respect to the port impedance matrix   Z01 0 · · · 0  0 Z02 · · · 0    Z0 =  . . .  .. . .   . . . . . 0 0 · · · Z0n we, first, transform S to an impedance matrix using (2.21) Z = Z0 (U − S)−1 (U + S) Z0 S = Z0
−1/2 1/2 1/2

29

Next the impedance matrix is transformed into the S-parameter matrix S Z − Z0 Z + Z0
−1

Z0

1/2

where now a reference impedance matrix of Z0 is used   0 ··· 0 Z01  0 Z 0  02 · · ·   Z0 =  . . .  .. . .   . . . . . 0 0 · · · Z0n S is then with respect to Z 0 . Often the renormalised S-parameters are with respect to 50 in which case all the diagonal elements of Z 0 are equal to 50 .

2.7

De-embedding of S-parameters

Another very important operation on an S-parameter matrix is the de-embedding of a length of transmission line from each of the ports. This is extremely important in measurement since often the device under test is connected to the measurement instrument by a length of transmission line and, therefore, the actual measured value includes the phase and attenuation of the line. It can be shown that the measured n-port S-parameters S are related to the network’s actual S-parameters S by S = where  S 0 ··· ··· .. . 0 0 . . .     

e−γ1 L1  0  = .  . . 0

e−γ2 L2 . . . 0

· · · e−γN LN

It is assumed that all of the lines are matched to their respective ports. If we know the length of line Lk at each port and the complex propagation constant γk = αk + jβk then we can de-embed the effect of the lines. Thus, S=
−1

S

−1

30

Microwave measurements

L1 γ1

L2

S

γ2

Figure 2.4

Two-port network with feeding transmission lines at each port

Due to the diagonal nature of , the inverse operation −1 simply changes the −γk Lk terms to +γk Lk . Figure 2.4 shows a typical two-port network with lines connected to both ports. In this case the actual network parameters S are related to the S-parameters S by S= S11 e+2γ1 L1 S21 e+γ1 L1 +γ2 L2 S12 e+γ1 L1 +γ2 L2 S22 e+2γ2 L2

In the lossless case γk would degenerate to jβk and only a phase shift would be introduced by the lines.

2.8

Characteristic impedance

We have seen that a microwave network can be characterised in terms of its S-parameters and that the S-parameters are defined with respect to the characteristic impedance at the ports of the network. A fundamental understanding of the nature of Z0 is therefore essential in microwave circuit analysis and measurement. Unfortunately, the true nature of characteristic impedance is often overlooked by microwave engineers and Z0 is usually considered to be a real-valued constant, such as 50 . In many cases this is a very good assumption. However, the careful metrologist does not make assumptions and the true nature of the characteristic impedance is imperative in precision microwave measurements. In fact without the knowledge of characteristic impedance, S-parameter measurements have little meaning and this lack of knowledge is so often the cause of poor measurements. This is particularly important in measured S-parameters from network analysers. S-parameters measured on a network analyser are with respect to either the Z0 of the calibration items or impedance of the matched element used to calibrate the analyser. If this value is ill-defined then so are the measured S-parameters.

2.8.1 Characteristic impedance in real transmission lines
If a TEM or quasi-TEM line contains dielectric and conductive losses then (2.1) and (2.2) become V (z) = V+ e−γ z + V− e+γ z (2.26)

Scattering parameters and circuit analysis and I (z) = 1 V+ e−γ z − V− e+γ z Z0

31

(2.27)

where the complex propagation constant is defined as γ = α + jβ. Equations 2.26 and 2.27 are very similar to (2.1) and (2.2); however, the attenuation constant α adds an exponential decay to the wave’s amplitude as it propagates along the line. It is easy to show that in terms of the transmission line’s per length series impedance Z and shunt admittance Y the propagation constant γ and characteristic impedance Z0 are given by [1] γ = √ ZY and Z0 = Z Y

In the lossless case Z = jωL and Y = jωC, representing the series inductance of the conductors and the shunt capacitance between them. The complex propagation constant then reduces to the familiar phase constant jβ and Z0 degenerates to a realvalued constant dependent only on L and C √ γ = jβ = jω LC and Z0 = L C

In lossy lines there is a series resistive component due to conduction losses and a shunt conductance due to dielectric losses. Hence, Z = R + jωL and Y = G + jωC and therefore, γ = and Z0 = R + jωL R + jωL = G + jωC γ (2.29) (R + jωL) (G + jωC) = Z0 (G + jωC) (2.28)

where R, L, G and C are often functions of ω. Two important facts about Z0 are immediately evident from (2.29): Z0 is complex and a function of frequency. Therefore, the assumption that Z0 is a real-valued constant which is independent of frequency is only an approximation. Fortunately, for many transmission lines the loss is small. In this case R ωL and G ωC and an approximate expression for the propagation constant is obtained by using a first-order binomial expansion. Thus, α≈ 1√ LC 2 R G + , L C √ β ≈ ω LC and Z0 ≈ L C

We see that first-order Z0 is identical to the lossless expression and, hence, the assumption that Z0 is a real-valued constant is often used. However, there are many cases when this approximation is far from valid. For example, transmission lines and waveguides operating at millimetre-wave frequencies often have very large losses

32

Microwave measurements

due to the increase in conduction and dielectric loss with frequency. In these cases, precision measurements must consider the complex nature of the transmission line. Furthermore at low frequencies where ω is small, we find that R ωL and G ωC. The complex nature of both γ and Z0 then plays a very important role. By way of example, Figures 2.5 and 2.6 show how the real and imaginary parts of Z0 vary with frequency for a CPW. The parameters of the line are typical for a microwave monolithic integrated circuit (MMIC) with a 400 µm thick gallium arsenide (GaAs) substrate and gold conductors of 1.2 µm thickness. We see that above a few GHz the real component of Z0 approaches the nominal 50 of the

120 110 100 Z0 90 80 70 60 50 0 2 4 f: GHz 6

Theory Measurement

8

10

Figure 2.5

Real part of characteristic impedance of CPW transmission line on GaAs

0 − 10 − 20 − 30 − 40 Z0 − 50 − 60 − 70 − 80 − 90 −100 0 2 4 f: GHz 6

Theory Measurement

8

10

Figure 2.6

Imaginary part of characteristic impedance of CPW transmission line on GaAs

Scattering parameters and circuit analysis

33

design but with a small imaginary part of a few ohms. At low frequencies the picture is very different. We see that as the frequency decreases there is a rapid increase in the magnitude of both the real and imaginary component of Z0 . Although the results are shown for CPW, similar results would be seen for microstrip, stripline and even coaxial cable. If Z0 has an appreciable imaginary part associated with it then a more complicated network analysis is required. The normal definitions of a and b are a = V+ Z0 and b = V− W1/2 . Z0

which have units of If the mode travelling on the transmission line carries power p0 then a and b can be written as √ √ a = C+ p0 and b = C− p0 where C+ and C− are constants with = C− /C+ . If Z0 is complex then it follows that the power p0 will also be complex having a real, travelling component and an imaginary, stored component. As a and b are travelling wave amplitudes they must be defined in terms of the real component of the power. Hence, √ √ a = C+ Re p0 and b = C− Re p0 Equation 2.10 then becomes P= Im(Z0 ) 1 |a|2 − |b|2 + Im(ab∗ ) Re(Z0 ) 2

That is, the power is not simply the difference in the forward and reverse waves unless Im(Z0 ) = 0. This results in a more complicated network theory where the definitions of impedance and admittance matrices have to be modified, resulting in different expressions for the conversion and renormalisation equations. The interested reader should consult Reference 2 for a thorough study of network theory with complex characteristic impedance.

2.8.2 Characteristic impedance in non-TEM waveguides
The usual definition of characteristic impedance is the ratio of the forward voltage to forward current. These are easily determined for simple TEM transmission lines, such as coaxial cable, where the voltage between the two conductors and the current flowing through them is uniquely defined. However, it can be more difficult to define voltages and currents in quasi-TEM transmission lines, such as microstrip and CPW, due to their hybrid nature. In fact, many waveguides used in microwave systems may only have a single conductor, such as rectangular waveguide, or no conductors at all as in a dielectric waveguide. In these cases, it becomes impossible to define a unique voltage or current and guides of this type are better explained in terms of their electric and magnetic fields E(x, y, z) = C+ et (x, y)e−γ z + C− et (x, y)e+γ z H(x, y, z) = C+ ht (x, y)e
−γ z

(2.30) (2.31)

− C− ht (x, y)e

+γ z

34

Microwave measurements

where et and ht are the electric and magnetic fields in the transverse plane, respectively. C− and C+ are complex-valued constants. In general, all transmission lines are described by (2.30) and (2.31) and not by (2.1) and (2.2). Equations 2.30 and 2.31 can be expressed as [2] E= and H= I (z) ht i0 (2.33) V (z) et v0 (2.32)

where the equivalent waveguide voltage and current are given by V (z) = v0 C+ e−γ z + C− e+γ z and I (z) = i0 C+ e−γ z − C− e+γ z respectively. v0 and i0 are normalisation constants such that Z0 = v0 i0 (2.36) (2.35) (2.34)

Both V (z) and v0 have units of voltage and I (z) and i0 have units of current. In order to extend the concept of voltage and current to the general waveguide structure, (2.34) and (2.35) must satisfy the same power relationships as (2.8) and (2.9). It can be shown that the power flow in a waveguide across a transverse surface S is given by [1]     1 1 V (z)I (z)∗ (2.37) P = Re E × H∗ · dS = Re p0  2 2  v0 i0 ∗
S

with modal power p0 =
S

et × ht ∗ · dS

(2.38)

Therefore, in order to retain the analogy with (2.8) and (2.9) we require P= and thus v0 i0 ∗ = p0 (2.40) 1 Re{V (z)I (z)∗ } 2 (2.39)

Scattering parameters and circuit analysis

35

We see that the magnitude of Z0 is not uniquely defined since we are free to choose any value of v0 and i0 as long as (2.40) is satisfied. For example, |Z0 | is often set to the wave impedance of the propagating mode. Another popular choice, used in network analysers, is |Z0 | = 1. Note, however, that the phase of Z0 is set by (2.40) and is an inherent characteristic of the propagating mode. To ensure that the characteristic impedance satisfies causality, Z0 should be equal to, within a constant multiplier, the TE or TM wave impedance of the waveguide [3]. This is essential in time-domain analysis and synthesis where responses cannot precede inputs. For hybrid structures, such as dielectric waveguide, causality is ensured if the following condition is satisfied: |Z0 (ω)| = λe−H (arg[p0 (ω)]) where λ is an arbitrary constant and H is the Hilbert transform [4]. Since we cannot define a unique value of Z0 , we cannot define S-parameter measurements with respect to a nominal characteristic impedance. This is not a problem for standard rectangular waveguide and coaxial cable which have set dimensions, since we can specify measurements with respect to WG-22 or APC7, etc. However, if we are using non-standard waveguides, such as image or dielectric waveguide then all we can say is our S-parameters are with respect to the propagating mode on the structure. Another important difference in waveguide networks is that in general a waveguide will support more than one mode. Multimode structures can be analysed using multimodal S-parameters [5]. Fortunately, under usual operating conditions, only the fundamental mode propagates. However, at a discontinuity, evanescent modes will always be present. These exponentially decaying fields will exist in the vicinity of the discontinuity and are required to completely explain the waveguide fields and network parameters. It is important to remember that (2.34) and (2.35) are only equivalent waveguide voltages and currents, which do not have all the properties of (2.1) and (2.2). For example, Z0 is dependent on normalisation and therefore we could define two different values of Z0 for the same waveguide. Furthermore, even if we do use the same normalisation scheme it is possible for two different waveguides to have the same Z0 . Clearly, a transition from one of these guides to the other will not result in a reflectionless transition, as conventional transmission theory would suggest. We also have to be very careful when converting to Z-parameters using (2.21), since Z is related to Z0 . Since Z0 is not uniquely defined the absolute value of Z-parameters cannot be determined. This is not surprising since impedance is intrinsically linked to current and voltage. However, even though absolute values cannot be defined, they can be useful in the development of equivalent circuit models for waveguide devices and junctions.

2.8.3 Measurement of Z0
For precision air-filled coaxial lines, the characteristic impedance can be approximately obtained from measurements of the geometry of the line. For quasi-TEM

36

Microwave measurements

lines several techniques can be used to measure Z0 . These include the constant capacitance technique [6,7] and the calibration comparison technique [8]. In the constant capacitance technique the characteristic impedance is derived from measured values of the propagation constant using the thru-reflect-line (TRL) technique [9] and a DC resistance measurement of the line. In the calibration comparison technique a two-tier calibration is performed, first in a known transmission line and then in the unknown line. This allows an ‘error-box’ to be determined, which acts as an impedance transformer from the known transmission line impedance to that of the lines under test.

2.9

Signal flow graphs

The analysis so far has relied on matrix algebra. However, another important technique can also be used to analyse microwave circuits, or indeed their low-frequency counterparts. This technique is known as the signal flow graph. Signal flow graphs express the network pictorially (Figure 2.7). The wave amplitudes are denoted by nodes, with the S-parameters being the gain achieved by the paths between nodes. To analyse signal flow graphs the following rules can be applied [10]. Rule 1. Two series branches, joined by a common node, can be replaced by one branch with gain equal to the product of the individual branches. Rule 2. Two parallel branches joining two common nodes can be replaced with a single branch with gain equal to the sum of the two individual branches. Rule 3. A branch that begins and ends on a single node can be eliminated by dividing the gains of all branches entering the node by one minus the gain of the loop. Rule 4. Any node can be duplicated as long as all paths are retained. These four rules are illustrated in Figure 2.8. Figure 2.9 gives an example of applying the above rules to analyse a microwave circuit. The network is a simple two-port network terminated by an impedance with a1 S21 b2

S11

S22

b1

S12

a2

Figure 2.7

Signal flow graph for two-port network

Scattering parameters and circuit analysis
Rule 1 Sa Sb SaSb

37

Sa

Rule 2 Sa + Sb

Sb Sa

Rule 3

Sa 1−Sb

Sb Sa

Rule 4 Sa

S

a

Figure 2.8

Kuhn’s rules for signal flow graph analysis

reflection coefficient L . We wish to calculate the reflection coefficient at the input terminal, i.e. b1 /a1 . First, we apply Rule 4 to duplicate the a2 node. Then, using Rule 1, we eliminate both the a2 nodes. The closed loop, S22 L is eliminated using Rule 3. Next, Rule 1 is applied to eliminate node b2 . Finally, applying Rule 2, we obtain a value for b1 /a1 . Clearly, for larger networks, the signal flow graph technique can be very difficult to apply. However, it can often be useful for analysing simple networks – giving a more intuitive approach to the problem.

2.10 Appendix
The following relationships are true only for purely real port characteristic impedances. For complex characteristic impedance the reader is referred to Reference [2].

2.A Reciprocity
Using the Lorentz reciprocity relation [11] it can be shown that, in general, Zmn = Znm . This is only true for networks that do not contain anisotropic media, such as ferrites.

38

Microwave measurements a1 S11 S21 b2 S22 a2 b2 ΓL S22 b1 a1 Rule 1 S11 S21 S12 b2 a2 a2 S22 ΓL ΓL ΓL

b1

a1 Rule 4 S11

S12 S21

ΓLS12 S21 b1 a1 1−S22 ΓL b2

Rule 3

S11

ΓLS12 b1 a1

Rule 1

S11

ΓLS12S21 1−S22 + ΓL b1 a1 S11+ ΓLS12S21 1−S22 + ΓL

Rule 2

b1

Figure 2.9

Example of the use of signal flow graphs to analyse a microwave network

Therefore, in matrix notation we have Z = ZT where ZT is the transpose of Z. From (2.21) it becomes apparent that if the characteristic impedance is identical at every port then (U − S)−1 (U + S) = U + ST U − ST
−1

Therefore, S = ST and provided the impedance matrices are symmetrical, Smn = Snm .

Scattering parameters and circuit analysis

39

2.B Losslessness
An n-port network can be described by an n × n S-parameter matrix:      b1 S11 S12 · · · S1n a1 b2  S21 S22 · · · S2n  a2       .= . . .  .  .. . .  .  .  . . . . . . . bn b = Sa If the network is lossless, then the power entering the network must be equal to the power flowing out of the network. Therefore, from (2.10) we have n i=1

(2B.1)

Sn1

Sn2

· · · Snn

an

|bi |2 =

n i=1

|ai |2

(2B.2)

But |bi |2 = bi bi ∗ , therefore, n i=1

 b1 ∗ b2 ∗    |bi |2 = (b1 , b2 , . . . , bn )  .   .  .  bn ∗

We see that the column matrix in the above equation is given by the conjugate of the right-hand side of (2B.1), that is, S∗ a∗ . Similarly, the row matrix is given by the transpose of the right-hand side of (2B.1), that is, (Sa)T . Therefore, we have n i=1

|bi |2 = (Sa)T S∗ a∗ = aT ST S∗ a∗

(2B.3)

where we have used the fact that the transpose of the product of two matrices is equal to the product of the transposes in reverse. Substituting (2B.3) into (2B.2) yields a T ST S ∗ a ∗ = a T a ∗ where we have used the fact that n i=1

(2B.4)

|ai |2 = aT a∗

Equation (2B.4) can be written as a T S T S∗ − U a ∗ = 0 where U is the unit matrix. Therefore, S T S∗ = U

40

Microwave measurements

In other words, the product of the transpose of the S-parameter matrix S with its complex conjugate is equal to the unit matrix U. In the case of a two port we have
∗ ∗ S11 S11 + S21 S21 = 1 ∗ ∗ S12 S12 + S22 S22 = 1 ∗ ∗ S11 S12 + S21 S22 = 0

and
∗ ∗ S12 S11 + S22 S21 = 0

2.C Two-port transforms
Using (2.21) for a two-port network gives √ 1 2 Z01 Z02 Z12 √ Z1 S= Z2 Z 2 Z01 Z02 Z21 where Z = (Z11 + Z01 )(Z22 + Z02 ) − Z12 Z21 Z1 = (Z11 − Z01 )(Z22 + Z02 ) − Z12 Z21 and Z2 = (Z11 + Z01 )(Z22 − Z02 ) − Z12 Z21 Similarly, using (2.22) Z= where S = (1 − S11 )(1 − S22 ) − S12 S21 S1 = (1 + S11 )(1 − S22 ) + S12 S21 and S2 = (1 − S11 )(1 + S22 ) + S12 S21 1 Z01 S1 S 2Z02 S12 2Z01 S21 Z02 S2

References
1 Ramo, S., Whinnery, J. R., and Van Duzer, T.: Fields and Waves in Communication Electronics, 2nd edn (John Wiley & Sons Inc., 1984) 2 Marks, R. B., and Williams, D. F.: ‘A general waveguide circuit theory’, Journal of Research of the National Institute of Standards and Technology, 1992;97: 533–62

Scattering parameters and circuit analysis

41

3 Williams, D. F., and Alpert, B. K.: ‘Characteristic impedance, power and causality’, IEEE Microwave Guided Wave Letters, 1999;9 (5):181–3 4 Williams, D. F., and Alpert, B. K.: ‘Causality and waveguide circuit theory’, IEEE Transactions on Microwave Theory and Techniques, 2001;49 (4):615–23 5 Shibata, T., and Itoh, T.: ‘Generalized-scattering-matrix modelling of waveguide circuits using FDTD field simulations’ IEEE Transactions on Microwave Theory and Techniques, 1998;46(11):1742–51 6 Marks, R. B., and Williams, D. F.: ‘Characteristic impedance determination using propagation constant measurement’, IEEE Microwave Guided Wave Letters, 1991;1 (6):141–3 7 Williams, D. F., and Marks, R. B.: ‘Transmission line capacitance measurement’, IEEE Microwave Guided Wave Letters, 1991;1 (9):243–5 8 Williams, D. F., Arz, U., and Grabinski, H.: ‘Characteristic-impedance measurement error on lossy substrates’, IEEE Microwave Wireless Components Letters, 2001;1 (7):299–301 9 Engen, G. F., and Hoer, C. A.: ‘Thru-reflect-line: an improved technique for calibrating the dual six-port automatic network analyser’, IEEE Transactions on Microwave Theory and Techniques, 1979;MTT-27:987–93 10 Kuhn, N.: ‘Simplified signal flow graph analysis’, Microwave Journal, 1963;6:59–66 11 Collin, R. E.: Foundations for Microwave Engineering, 2nd edn (McGraw-Hill, New York, 1966)

Further reading
1 Bryant, G. H.: Principles of Microwave Measurements, IEE Electrical measurement series 5 (Peter Peregrinus, London, 1997) 2 Engen, G. F.: Microwave Circuit Theory and Foundations of Microwave Metrology, IEE Electrical Measurement Series 9 (Peter Peregrinus, London, 1992) 3 Kerns, D. M., and Beatty, R. W.: Basic Theory of Waveguide Junctions and Introductory Microwave Network Analysis, International series of monographs in electromagnetic waves volume 13 (Pergamon Press, Oxford, 1969) 4 Somlo, P. I., and Hunter, J. D.: Microwave Impedance Measurement, IEE Electrical Measurement Series 2 (Peter Peregrinus, London, 1985)

Chapter 3

Uncertainty and confidence in measurements
John Hurll

3.1

Introduction

The objective of a measurement is to determine the value of the measurand, that is, the specific quantity subject to measurement. A measurement begins with an appropriate specification of the measurand, the generic method of measurement and the specific detailed measurement procedure. Knowledge of the influence quantities involved for a given procedure is important so that the sources of uncertainty can be identified. Each of these sources of uncertainty will contribute to the uncertainty associated with the value assigned to the measurand. The guidance in this chapter is based on information in the Guide to the Expression of Uncertainty in Measurement [1], hereinafter referred to as the GUM. The reader is also referred to the UKAS document M3003 Edition 2 [2], which uses terminology and methodology that are compatible with the GUM. M3003 is available as a free download at www.ukas.com. A quantity (Q) is a property of a phenomenon, body or substance to which a magnitude can be assigned. The purpose of a measurement is to assign a magnitude to the measurand: the quantity intended to be measured. The assigned magnitude is considered to be the best estimate of the value of the measurand. The uncertainty evaluation process will encompass a number of influence quantities that affect the result obtained for the measurand. These influence, or input, quantities are referred to as X and the output quantity, that is, the measurand, is referred to as Y . As there will usually be several influence quantities, they are differentiated from each other by the subscript i, so there will be several input quantities called Xi , where i represents integer values from 1 to N (N being the number of such quantities). In other words, there will be input quantities of X1 , X2 , . . . , XN . One of the first steps is to establish the mathematical relationship between the values of the input quantities,

44

Microwave measurements

xi , and that of the measurand, y. Details about the derivation of the mathematical model can be found in Appendix D of M3003 [2]. Each of these input quantities will have a corresponding value. For example, one quantity might be the temperature of the environment – this will have a value, say 23 ◦ C. A lower-case ‘x’ represents the values of the quantities. Hence the value of X1 will be x1 , that of X2 will be x2 , and so on. The purpose of the measurement is to determine the value of the measurand, Y . As with the input uncertainties, the value of the measurand is represented by the lowercase letter, that is, y. The uncertainty associated with y will comprise a combination of the input, or xi , uncertainties. The values xi of the input quantities Xi will all have an associated uncertainty. This is referred to as u(xi ), that is, ‘the uncertainty of xi ’. These values of u(xi ) are, in fact, something known as the standard uncertainty. Some uncertainties, particularly those associated with evaluation of repeatability, have to be evaluated by statistical methods. Others have been evaluated by examining other information, such as data in calibration certificates, evaluation of long-term drift, consideration of the effects of environment, etc. The GUM [1] differentiates between statistical evaluations and those using other methods. It categorises them into two types – Type A and Type B. Type A evaluation of uncertainty is carried out using statistical analysis of a series of observations. Type B evaluation of uncertainty is carried out using methods other than statistical analysis of a series of observations. In paragraph 3.3.4 of the GUM [1] it is stated that the purpose of the Type A and Type B classification is to indicate the two different ways of evaluating uncertainty components, and is for convenience in discussion only. Whether components of uncertainty are classified as ‘random’ or ‘systematic’ in relation to a specific measurement process, or described as Type A or Type B depending on the method of evaluation, all components regardless of classification are modelled by probability distributions quantified by variances or standard deviations. Therefore any convention as to how they are classified does not affect the estimation of the total uncertainty. But it should always be remembered that when the terms ‘random’ and ‘systematic’ are used they refer to the effects of uncertainty on a specific measurement process. It is the usual case that random components require Type A evaluations and systematic components require Type B evaluations, but there are exceptions. For example, a random effect can produce a fluctuation in an instrument’s indication, which is both noise-like in character and significant in terms of uncertainty. It may then only be possible to estimate limits to the range of indicated values. This is not a common situation but when it occurs, a Type B evaluation of the uncertainty component will be required. This is done by assigning limit values and an associated probability distribution, as in the case of other Type B uncertainties. The input uncertainties, associated with the values of the influence quantities Xi , arise in a number of forms. Some may be characterised as limit values within which

Uncertainty and confidence in measurements a a

45

Probability p

xi − a

xi

xi + a

Figure 3.1

The expectation value xi lies in the centre of a distribution of possible values with a half-width, or semi-range, of a

little is known about the most likely place within the limits where the ‘true’ value may lie. It therefore has to be assumed that the ‘true’ value is equally likely to lie anywhere within the assigned limits. This concept is illustrated in Figure 3.1, from which it can be seen that there is equal probability of the value of xi being anywhere within the range xi − a to xi + a and zero probability of it being outside these limits. Thus, a contribution of uncertainty from the influence quantity can be characterised as a probability distribution, that is, a range of possible values with information about the most likely value of the input quantity xi . In this example, it is not possible to say that any particular position of xi within the range is more or less likely than any other. This is because there is no information available upon which to make such a judgement. The probability distributions associated with the input uncertainties are therefore a reflection of the available knowledge about that particular quantity. In many cases, there will be insufficient information available to make a reasonable judgement and therefore a uniform, or rectangular, probability distribution has to be assumed. Figure 3.1 is an example of such a distribution. If more information is available, it may be possible to assign a different probability distribution to the value of a particular input quantity. For example, a measurement may be taken as the difference in readings on a digital scale – typically, the zero reading will be subtracted from a reading taken further up the scale. If the scale is linear, both of these readings will have an associated rectangular distribution of identical size. If two identical rectangular distributions, each of magnitude ±a, are combined then the resulting distribution will be triangular with a semi-range of ±2a (Figure 3.2). There are other possible distributions that may be encountered. For example, when making measurements of radio-frequency power, an uncertainty arises due to imperfect matching between the source and the termination. The imperfect match usually involves an unknown phase angle. This means that a cosine function characterises the probability distribution for the uncertainty. Harris and Warner [3] have

46

Microwave measurements

Probability p

xi − 2a

xi

xi + 2a

Figure 3.2

Combination of two identical rectangular distributions, each with semirange limits of ±a, yields a triangular distribution with a semi-range of ±2a

shown that a symmetrical U-shaped probability distribution arises from this effect. In this example, the distribution has been evaluated from a theoretical analysis of the principles involved (Figure 3.3). An evaluation of the effects of non-repeatability, performed by statistical methods, will usually yield a Gaussian or normal distribution. When a number of distributions of whatever form are combined it can be shown that, apart from in one exceptional case, the resulting probability distribution tends to the normal form in accordance with the Central Limit Theorem. The importance of this is that it makes it possible to assign a level of confidence in terms of probability to the combined uncertainty. The exceptional case arises when one contribution to the total uncertainty dominates; in this circumstance, the resulting distribution departs little from that of the dominant contribution.
Note: If the dominant contribution is itself normal in form, then clearly the resulting distribution will also be normal.

When the input uncertainties are combined, a normal distribution will usually be obtained. The normal distribution is described in terms of a standard deviation. It will therefore be necessary to express the input uncertainties in terms that, when combined, will cause the resulting normal distribution to be expressed at the one standard deviation level, like the example in Figure 3.4. As some of the input uncertainties are expressed as limit values (e.g. the rectangular distribution), some processing is needed to convert them into this form, which is known as a standard uncertainty and is referred to as u(xi ).

Uncertainty and confidence in measurements

47

xi − a

xi

xi + a

Figure 3.3

U-shaped distribution, associated with RF mismatch uncertainty. For this situation, xi is likely to be close to one or other of the edges of the distribution

Figure 3.4

The normal, or Gaussian, probability distribution. This is obtained when a number of distributions, of any form, are combined and the conditions of the Central Limit Theorem are met. In practice, if three or more distributions of similar magnitude are present, they will combine to form a reasonable approximation to the normal distribution. The size of the distribution is described in terms of a standard deviation. The shaded area represents ±1 standard deviation from the centre of the distribution. This corresponds to approximately 68 per cent of the area under the curve

48

Microwave measurements

When it is possible to assess only the upper and lower bounds of an error, a rectangular probability distribution should be assumed for the uncertainty associated with this error. Then, if ai is the semi-range limit, the standard uncertainty √ is given by u(xi ) = ai / 3. Table 3.1 gives the expressions for this and for other situations. The quantities Xi that affect the measurand Y may not have a direct, one to one, relationship with it. Indeed, they may be entirely different units altogether. For example, a dimensional laboratory may use steel gauge blocks for calibration of measuring tools. A significant influence quantity is temperature. Because the gauge blocks have a significant temperature coefficient of expansion, there is an uncertainty that arises in their length due to an uncertainty in temperature units. In order to translate the temperature uncertainty into an uncertainty in length units, it is necessary to know how sensitive the length of the gauge block is to temperature. In other words, a sensitivity coefficient is required. The sensitivity coefficient simply describes how sensitive the result is to a particular influence quantity. In this example, the steel used in the manufacture of gauge blocks has a temperature coefficient of expansion of approximately +11.5 × 10−6 per ◦ C. So, in this case, this figure can be used as the sensitivity coefficient. The sensitivity coefficient associated with each input quantity Xi is referred to as ci . It is the partial derivative ∂f /∂xi , where f is the functional relationship between the input quantities and the measurand. In other words, it describes how the output estimate y varies with a corresponding small change in an input estimate xi . The calculations required to obtain sensitivity coefficients by partial differentiation can be a lengthy process, particularly when there are many input contributions and uncertainty estimates are needed for a range of values. If the functional relationship is not known for a particular measurement system the sensitivity coefficients can sometimes be obtained by the practical approach of changing one of the input variables by a known amount, while keeping all other inputs constant, with no change in the output estimate. This approach can also be used if f is known, but if f is not a straightforward function the determination of partial derivatives required is likely to be error-prone. A more straightforward approach is to replace the partial derivative ∂f /∂xi by the quotient f / xi , where f is the change in f resulting from a change xi in xi . It is important to choose the magnitude of the change xi carefully. It should be balanced between being sufficiently large to obtain adequate numerical accuracy in f and sufficiently small to provide a mathematically sound approximation to the partial derivative. The example in Figure 3.5 illustrates this, and why it is necessary to know the functional relationship between the influence quantities and the measurand. If the uncertainty in d is, say, ±0.1 m then the estimate of h could be anywhere between (7.0 − 0.1) tan (37) and (7.0 + 0.1) tan (37), that is, between 5.200 and 5.350 m. A change of ±0.1 m in the input quantity xi has resulted in a change of ±0.075 m in the output estimate y. The sensitivity coefficient is therefore (0.075/0.1) = 0.75.

Uncertainty and confidence in measurements Table 3.1
Assumed probability distribution Rectangular

49

Probability distributions and standard uncertainties
Expression used to obtain the standard uncertainty ai u(xi ) = √ 3 Comments or examples

U-shaped

ai u(xi ) = √ 2

A digital thermometer has a least significant digit of 0.1 ◦ C. The numeric rounding caused by finite resolution will have semi-range limits of 0.05 ◦ C. Thus the corresponding standard uncertainty will be 0.05 ai = 0.029 ◦ C. u(xi ) = √ = 1.732 3 A mismatch uncertainty associated with the calibration of an RF power sensor has been evaluated as having semi-range limits of 1.3%. Thus the corresponding standard uncertainty will be 1.3 ai u(xi ) = √ = = 0.92%. 1.414 2 A tensile testing machine is used in a testing laboratory where the air temperature can vary randomly but does not depart from the nominal value by more than 3 ◦ C. The machine has a large thermal mass and is therefore most likely to be at the mean air temperature, with no probability of being outside the 3 ◦ C limits. It is reasonable to assume a triangular distribution, therefore the standard uncertainty for its 3 ai temperature is u(xi ) = √ = = 1.2 ◦ C. 2.449 6 A statistical evaluation of repeatability gives the result in terms of one standard deviation; therefore no further processing is required. A calibration certificate normally quotes an expanded uncertainty U at a specified, high level of confidence. A coverage factor, k, will have been used to obtain this expanded uncertainty from the combination of standard uncertainties. It is therefore necessary to divide the expanded uncertainty by the same coverage factor to obtain the standard uncertainty. Some manufacturers’ specifications are quoted at a given confidence level, for example, 95% or 99%. In such cases, a normal distribution can be assumed and the tolerance limit is divided by the coverage factor k for the stated confidence level. For a confidence level of 95%, k = 2 and for a confidence level of 99%, k = 2.58. If a confidence level is not stated then a rectangular distribution should be assumed.

Triangular

ai u(xi ) = √ 6

Normal (from repeatability evaluation) Normal (from a calibration certificate)

u(xi ) = s(¯ ) q U k

u(xi ) =

Normal (from a manufacturer’s specification)

u(xi ) =

Tolerance limit k

50

Microwave measurements

Φ d

Figure 3.5

The height h of a flagpole is determined by measuring the angle obtained when observing the top of the pole at a specified distance d. Thus h = d tan . Both h and d are in units of length but the related by tan . In other words, h = f (d) = d tan . If the measured distance is 7.0 m and the measured angle is 37◦ , the estimated height is 7.0 × tan(37) = 5.275 m.

Similar reasoning can be applied to the uncertainty in the angle . If the uncertainty in is ±0.5◦ , then the estimate of h could be anywhere between 7.0 tan (36.5) and 7.0 tan (37.5), that is, between 5.179 and 5.371 m. A change of ±0.5◦ in the input quantity xi has resulted in a change of ±0.096 m in the output estimate y. The sensitivity coefficient is therefore (0.096/0.5) = 0.192 m per degree. Once the standard uncertainties xi and the sensitivity coefficients ci have been evaluated, the uncertainties have to be combined in order to give a single value of uncertainty to be associated with the estimate y of the measurand Y . This is known as the combined standard uncertainty and is given the symbol uc (y). The combined standard uncertainty is calculated as follows:
N

uc (y) = i=1 ci2 u2 (xi )

N

≡ i=1 2 ui (y).

(3.1)

In other words, the individual standard uncertainties, expressed in terms of the measurand, are squared; these squared values are added and the square root is taken. An example of this process is presented in Table 3.2, using the data from the measurement of the flagpole height described previously. For the purposes of the example, it is assumed that the repeatability of the process has been evaluated by making repeat measurements of the flagpole height, giving an estimated standard deviation of the mean of 0.05 m. In accordance with the Central Limit Theorem, this combined standard uncertainty takes the form of a normal distribution. As the input uncertainties had been expressed in terms of a standard uncertainty, the resulting normal distribution is expressed as one standard deviation, as illustrated in Figure 3.6.

Uncertainty and confidence in measurements Table 3.2

51

Calculation of combined standard uncertainty ui (y) from standard uncertainties yi (y)
Value Probability distribution Rectangular Divisor Sensitivity coefficient √ 3 √ 3 0.75 0.192 m per degree 1 Standard uncertainty ui (y) 0.1 √ × 0.75 = 0.0433 m 3 0.5 √ × 0.192 = 0.0554 m 3 0.05 × 1 = 0.05 m 1

Source of uncertainty Distance from flagpole Angle measurement Repeatability

0.1 m

0.5 m per Rectangular degree 0.05 m Normal

1

Combined standard uncertainty uc (y) =

0.04332 + 0.05542 + 0.052 = 0.0863 m

y 5.275 m y – 0.0863 m y + 0.0863 m

Figure 3.6

The measured value y is at the centre of a normal distribution with a standard deviation equal to uc (y). The figures shown relate to the example discussed in the text

For a normal distribution, the 1 standard deviation limits encompass 68.27 per cent of the area under the curve. This means that there is about 68 per cent confidence that the measured value y lies within the stated limits. The GUM [1] recognises the need for providing a high level of confidence associated with an uncertainty and uses the term expanded uncertainty, U , which is obtained by multiplying the combined standard uncertainty by a coverage factor. The coverage

52

Microwave measurements

factor is given the symbol k, thus the expanded uncertainty is given by U = kuc (y) In accordance with generally accepted international practice, it is recommended that a coverage factor of k = 2 is used to calculate the expanded uncertainty. This value of k will give a level of confidence, or coverage probability, of approximately 95 per cent, assuming a normal distribution.
Note: A coverage factor of k = 2 actually provides a coverage probability of 95.45 per cent for a normal distribution. For convenience this is approximated to 95 per cent which relates to a coverage factor of k = 1.96. However, the difference is not generally significant since, in practice, the level of confidence is based on conservative assumptions and approximations to the true probability distributions. Example: The measurement of the height of the flagpole had a combined standard uncertainty uc (y) of 0.0863 m. Hence the expanded uncertainty U = kuc (y) = 2 × 0.0863 = 0.173 m.

There may be situations where a normal distribution cannot be assumed and a different coverage factor may be needed in order to obtain a confidence level of approximately 95 per cent. This is done by obtaining a new coverage factor based on the effective degrees of freedom of uc (y). Details of this process can be found in Appendix B of M3003 [2]. There may also be situations where a normal distribution can be assumed, but a different level of confidence is required. For example, in safety-critical situations a higher coverage probability may be more appropriate. Table 3.3 gives the coverage factor necessary to obtain various levels of confidence for a normal distribution.

3.2

Sources of uncertainty in RF and microwave measurements

3.2.1 RF mismatch errors and uncertainty
At RF and microwave frequencies the mismatch of components to the characteristic impedance of the measurement system transmission line can be one of the Table 3.3 Coverage factor necessary to obtain various levels of confidence for a normal distribution
Coverage factor k 1.65 1.96 2.00 2.58 3.00

Coverage probability p (%) 90 95 95.45 99 99.73

Uncertainty and confidence in measurements

53

most important sources of error and of the systematic component of uncertainty in power and attenuation measurements. This is because the phases of voltage reflection coefficients (VRCs) are not usually known and hence corrections cannot be applied. In a power measurement system, the power, P0 , that would be absorbed in a load equal to the characteristic impedance of the transmission line has been shown [3] to be related to the actual power, PL , absorbed in a wattmeter terminating the line by the equation P0 = PL (1 − 2| 1 − | L |2
G || L | cos ϕ

+|

2 2 G| | L| ) G

(3.2) and
L.

where ϕ is the relative phase of the generator and load VRCs and L are small, this becomes P0 = PL (1 − 2| 1 − | L |2
G || L | cos ϕ)

When

G

(3.3)

When ϕ is unknown, this expression for absorbed power can have limits. P0 (limits) = PL (1 ± 2| 1 − | L |2
G || L |)

(3.4)

The calculable mismatch error is 1−| L |2 and is accounted for in the calibration factor, while the limits of mismatch uncertainty are ±2| L || G |. Because a cosine function characterises the probability distribution for the uncertainty, Harris and Warner [3] showed that the distribution is U-shaped with a standard deviation given by u(mismatch) = 2|
G || √ 2 L|

= 1.414

G L

(3.5)

When a measurement is made of the attenuation of a two-port component inserted between a generator and load that are not perfectly matched to the transmission line, Harris and Warner [3] have shown that the standard deviation of mismatch, M , expressed in dB is approximated by 8.686 M = √ [| 2 +|
2 G| G

|2 (|s11a |2 + |s11b |2 ) + |
2 4 L | (|s21a |

2 2 L | (|s22a |

+ |s22b |2 ) (3.6)

·|

+ |s21b |4 )]0.5

where G and L are the source and load VRCs, respectively, and s11 , s22 , s21 are the scattering coefficients of the two-port component with the suffix a referring to the starting value of the attenuator and b referring to the finishing value of the attenuator. Harris and Warner [3] concluded that the distribution for M would approximate to that of a normal distribution due to the combination of its component distributions. The values of G and L used in (3.4) and (3.5) and the scattering coefficients used in (3.6) will themselves be subject to uncertainty because they are derived from measurements. This uncertainty has to be considered when calculating the mismatch uncertainty and it is recommended that this is done by adding it in quadrature with the

54

Microwave measurements

measured or derived value of the reflection coefficient; for example, if the measured value of L is 0.03 ± 0.02 then the value of L that should be used to calculate the √ mismatch uncertainty is 0.032 + 0.022 , that is, 0.036.

3.2.2 Directivity
When making VRC measurements at RF and microwave frequencies, the finite directivity of the bridge or reflectometer gives rise to an uncertainty in the measured value of the VRC, if only the magnitude and not the phase of the directivity component is known. The uncertainty will be equal to the directivity, expressed in linear terms; for example, a directivity of 30 dB is equivalent to an uncertainty of ±0.0316 VRC. As above, it is recommended that the uncertainty in the measurement of directivity is taken into account by adding the measured value in quadrature with the uncertainty, in linear quantities; for example, if the measured directivity of a bridge is 36 dB (0.016) and √ has an uncertainty of +8 dB −4 dB (±0.01) then the directivity to be used is 0.0162 + 0.012 = 0.019 (34.4 dB).

3.2.3 Test port match
The test port match of a bridge or reflectometer used for reflection coefficient measurements will give rise to an error in the measured VRC due to re-reflection. The 2 uncertainty, u(TP), is calculated from u(TP) = TP. X where TP is the test port match, expressed as a VRC, and X is the measured reflection coefficient. When a directional coupler is used to monitor incident power in the calibration of a power meter, it is the effective source match of the coupler that defines the value of G . The measured value of test port match will have an uncertainty that should be taken into account by using quadrature summation.

3.2.4 RF connector repeatability
The lack of repeatability of coaxial pair insertion loss and, to a lesser extent, VRC is a problem when calibrating devices in a coaxial line measurement system and subsequently using them in some other system. Although connecting and disconnecting the device can evaluate the repeatability of particular connector pairs in use, these connector pairs are only samples from a whole population. To obtain representative data for all the various types of connector in use is beyond the resources of most measurement laboratories; however, some useful guidance can be obtained from the ANAMET Connector Guide [4].

3.2.5 Example – calibration of a coaxial power sensor at a frequency of 18 GHz
The measurement involves the calibration of an unknown power sensor against a standard power sensor by substitution on a stable, monitored source of known source

Uncertainty and confidence in measurements impedance. The measurement is made in terms of Calibration factor, defined as Incident power at reference frequency Incident power at calibration frequency for the same power sensor response and is determined from the following: Calibration factor, KX = (Ks + Ds ) × δDC × δM × δREF

55

(3.7)

where KS is the calibration factor of the standard sensor, DS is the drift in standard sensor since the previous calibration, δDC is the ratio of DC voltage outputs, δM is the ratio of mismatch losses and δREF is the ratio of reference power source (short-term stability of 50 MHz reference). Four separate measurements were made which involved disconnection and reconnection of both the unknown sensor and the standard sensor on a power transfer system. All measurements were made in terms of voltage ratios that are proportional to calibration factor. There will be mismatch uncertainties associated with the source/standard sensor combination and with the source/unknown sensor combination. These will be 200 G S per cent and 200 G X per cent, respectively, where
G S X

= 0.02 at 50 MHz and 0.07 at 18 GHz = 0.02 at 50 MHz and 0.10 at 18 GHz = 0.02 at 50 MHz and 0.12 at 18 GHz.

These values are assumed to include the uncertainty in the measurement of . The standard power sensor was calibrated by an accredited laboratory 6 months before use; the expanded uncertainty of ±1.1 per cent was quoted for a coverage factor k = 2. The long-term stability of the standard sensor was estimated from the results of five annual calibrations to have rectangular limits not greater than ±0.4 per cent per year. A value of ±0.2 per cent is assumed as the previous calibration was within 6 months. The instrumentation linearity uncertainty was estimated from measurements against a reference attenuation standard. The expanded uncertainty for k = 2 of ±0.1 per cent applies to ratios up to 2:1. Type A evaluation The four measurements resulted in the following values of Calibration factor: 93.45, 92.20, 93.95 and 93.02 per cent. ¯ The mean value KX = 93.16 per cent. The standard deviation of the mean, per cent. √ ¯ u(KR ) = s(KX ) = 0.7415/ 4 = 0.3707

56

Microwave measurements Uncertainty budget
Value ±% Probability distribution 1.1 0.2 0.1 0.2 Normal Rectangular Normal Rectangular Divisor ci 2.0 √ 3 1.0 1.0 1.0 1.0 ui (Kx )% vi or veff 0.55 0.116 0.05 0.116 ∞ ∞ ∞ ∞

Table 3.4

Symbol Source of uncertainty KS DS δDC δM Calibration factor of standard Drift since last calibration Instrumentation linearity Stability of 50 MHz reference Mismatch: Standard sensor at 50 MHz Unknown sensor at 50 MHz Standard sensor at 18 GHz Unknown sensor at 18 GHz Repeatability of indication Combined standard uncertainty Expanded uncertainty

2.0 √ 3

M1 M2 M3 M4 KR u(KX ) U

0.08 0.08 1.40 1.68 0.37

U-shaped U-shaped U-shaped U-shaped Normal Normal Normal (k = 2)

√ √ √ √

2 2 2 2

1.0 1.0 1.0 1.0 1.0

0.06 0.06 0.99 1.19 0.37 1.69 3.39

∞ ∞ ∞ ∞ 3 >500 >500

1.0

3.2.5.1 Reported result The measured calibration factor at 18 GHz is 93.2% ± 3.4%. The reported expanded uncertainty (Table 3.4) is based on a standard uncertainty multiplied by a coverage factor k = 2, providing a coverage probability of approximately 95 per cent. 3.2.5.2 Notes (1) For the measurement of calibration factor, the uncertainty in the absolute value of the 50 MHz reference source need not be included if the standard and unknown sensors are calibrated using the same source, within the timescale allowed for its short-term stability.

Uncertainty and confidence in measurements (2) This example illustrates the significance of mismatch uncertainty in measurements at relatively high frequencies. (3) In a subsequent use of a sensor further random components of uncertainty may arise due to the use of different connector pairs.

57

References
1 BIPM, IEC, IFCC, ISO, IUPAC, IUPAP, OIML. Guide to the Expression of Uncertainty in Measurement. International Organisation for Standardization, Geneva, Switzerland. ISBN 92-67-10188-9, First Edition 1993. BSI Equivalent: BSI PD 6461: 1995, Vocabulary of Metrology, Part 3. Guide to the Expression of Uncertainty in Measurement. BSI ISBN 0 580 23482 7 2 United Kingdom Accreditation Service, The Expression of Uncertainty and Confidence in Measurement, M3003, 2nd edn, January 2007 3 Harris, I. A., and Warner, F. L.: ‘Re-examination of mismatch uncertainty when measuring microwave power and attenuation’, IEE Proc. H, Microw. Opt. Antennas, 1981;128 (1):35–41 4 National Physical Laboratory, ANAMET Connector Guide, 3rd edn, 2007

Chapter 4

Using coaxial connectors in measurement
Doug Skinner

Disclaimer
Every effort has been made to ensure that this chapter contains accurate information obtained from many sources that are acknowledged where possible. However, the NPL, ANAMET and the Compiler cannot accept liability for any errors, omissions or misleading statements in the information. The compiler would also like to thank all those members of ANAMET who have supplied information, commented on and given advice on the preparation of this chapter.

4.1

Introduction

The importance of the correct use of coaxial connectors not only applies at radio and microwave frequencies but also at DC and low frequency. The requirement for ‘Traceability to National Standards’ for measurements throughout the industry may depend on several different calibration systems ‘seeing’ the same values for the parameters presented by a device at its coaxial terminals. It is not possible to include all the many different types of connector in this guide and the selection has been made on those connectors used on measuring instruments and for metrology use. It is of vital importance to note that mechanical damage can be inflicted on a connector when a connection or disconnection is made at any time during its use. The common types of precision and general-purpose coaxial connector that are in volume use worldwide are the Type N, GPC 3.5 mm, Type K, 7/16, TNC, BNC and SMA connectors. These connectors are employed for interconnection of components and cables in military, space, industrial and domestic applications. The simple concept of a coaxial connector comprises an outer conductor contact, an inner conductor contact, and means for mechanical coupling to a cable and/or

60

Microwave measurements

to another connector. Most connectors, in particular the general-purpose types, are composed of a pin and socket construction. There are basically two types of coaxial connector in use and they are known as laboratory precision connectors (LPC), using only air dielectric, and general precision connectors (GPC), having a self contained, low reflection dielectric support. There are also a number of general-purpose connectors in use within the GPC group but they are not recommended for metrology use. Some connectors are hermaphroditic (non-sexed), particularly some of the laboratory precision types, and any two connectors may be joined together. They have planar butt contacts and are principally employed for use on measurement standards and on equipment and calibration systems where the best possible uncertainty of measurement is essential. Most of the non-sexed connectors have a reference plane that is common to both the outer and inner conductors. The mechanical and electrical reference planes coincide and, in the case of the precision connectors, a physically realised reference plane is clearly defined. Hermaphroditic connectors are used where the electrical length and the characteristic impedance are required at the highest accuracy. The 14 mm and 7 mm connectors are examples of this type of construction and they are expensive. GPCs of the plug and socket construction look similar but the materials used in the construction can vary. The best quality connectors are more robust, which use stainless steel, and the mechanical tolerances are more precise. It is important to be clear on the quality of the connector being used. The GPC types in common use are the Type N 7 mm, GPC 3.5 mm, Type K 2.92 mm, GPC 2.4 mm and Type V 1.85 mm. The choice of connectors, from the range of established designs, must be appropriate to the proposed function and specification of the device or measurement system. Often a user requirement is for a long-life quality connector with minimum effect on the performance of the device it is used on and the repeatability of the connection is generally one of the most important parameters.

4.1.1 Coaxial line sizes
Some coaxial line sizes for establishing a characteristic impedance of 50 are shown in Table 4.1. They are chosen to achieve the desired performance over their operating frequency range up to 110 GHz. Table 4.1 Coaxial line sizes for 50 characteristic impedance
14.29 8.5 9.5 7.00 18.0 19.4 3.50 33.0 38.8 2.92 40.0 46.5 2.40 50.0 56.5 1.85 65.0 73.3 1.00 110.0 135.7

Inside diameter of the outer conductor in mm (nominal) Rated minimum upper frequency limit in GHz Theoretical limit in GHz for the onset of the TE11 (H11 ) mode

Using coaxial connectors in measurement

61

4.2

Connector repeatability

Connectors in use on test apparatus and measuring instruments at all levels need to be maintained in pristine condition in order to retain the performance of the test apparatus. The connector repeatability is a key contribution to the performance of a measurement system. Connector repeatability can be greatly impaired because of careless assembly, misalignment, over-tightening, inappropriate handling, poor storage and unclean working conditions. In extreme cases, permanent damage can be caused to the connectors concerned and possibly to other originally sound connectors to which they are coupled. Connectors should never be rotated relative to one another when being connected and disconnected. Special care should be taken to avoid rotating the mating plane surfaces against one another.

4.2.1 Handling of airlines
When handling or using airlines and similar devices used in automatic network analyser, calibration and verification kits, it is extremely important to avoid contamination of the component parts due to moisture and finger marks on the lines. Protective lintfree cotton gloves should always be worn. The failure to follow this advice may significantly reduce performance and useful life of the airlines.

4.2.2 Assessment of connector repeatability
In a particular calibration or measurement system, repeatability of the coaxial interconnections can be assessed from measurements made after repeatedly disconnecting and reconnecting the device. It is clearly necessary to ensure that all the other conditions likely to influence the alignment are maintained as constant as possible. In some measurement situations, it is important that the number of repeat connections made uses the same positional alignment of the connectors. In other situations, it is best to rotate one connector relative to the other between connections and reconnections. For example, when calibrating or using devices fitted with Type N connectors (e.g. power sensors or attenuator pads) three rotations of 120◦ or five rotations of 72◦ are made. However, it is important to remember to make the rotation before making the contact. The repeatability determination will normally be carried out when trying to achieve the best measurement capability on a particular device, or when initially calibrating a measuring system. The number of reconnections and rotations can then be recommended in the measurement procedure. Repeatability of the insertion loss of coaxial connectors introduces a major contribution to the Type A component of uncertainty in a measurement process. If a measurement involving connectors is repeated several times, the Type A uncertainty contribution deduced from the results will include that arising from the connector repeatability provided that the connection concerned is broken and remade at each repetition.

62

Microwave measurements

It should be remembered that a connection has to be made at least once when connecting an item under test to the test equipment and this gives rise to a contribution to the Type A uncertainty contribution associated with the connector repeatability. Experience has shown that there is little difference in performance between precision and ordinary connectors (when new) so far as the repeatability of connection is concerned, but with many connections and disconnections the ordinary connector performance will become progressively inferior when compared with the precision connector.

4.3

Coaxial connector specifications

The following specifications are some of those that provide information and define the parameters of established designs of coaxial connectors and they should be consulted for full information on electrical performance, mechanical dimensions and mechanical tolerances. IEEE Standard 287-2007 IEC Publication 457 MIL-STD-348A incorporates MIL-STD-39012C IEC Publication 169 CECC 22000 British Standard 9210 DIN Standards Users of coaxial connectors should also take into account any manufacturer’s performance specifications relating to a particular connector type.

4.4

Interface dimensions and gauging

It is of utmost importance that connectors do not damage the test equipment interfaces to which they are offered for calibration. Poor performance of many coaxial devices and cable assemblies can often be traced to poor construction and non-compliance with the mechanical specifications. The mechanical gauging of connectors is essential to ensure correct fit and to achieve the best performance. This means that all coaxial connectors fitted on all equipments, cables and terminations should be gauged on a routine basis in order to detect any out of tolerance conditions that may impair the electrical performance.

4.4.1 Gauging connectors
A connector should be gauged before it is used for the very first time or if someone else has used the device on which it is fitted.

Using coaxial connectors in measurement

63

If the connector is to be used on another item of equipment, the connector on the equipment to be tested should also be gauged. Connectors should never be forced together when making a connection since forcing often indicates incorrectness and incompatibility. Many connector screw coupling mechanisms, for instance, rarely need to be more than finger-tight for electrical calibration purposes; most coaxial connectors usually function satisfactorily, giving adequately repeatable results, unless damaged. There are some dimensions that are critical for the mechanical integrity, non-destructive mating and electrical performance of the connector. Connector gauge kits are available for many connector types but it is also easy to manufacture simple low-cost test pieces for use with a micrometer depth gauge or other device to ensure that the important dimensions can be measured or verified. The mechanical gauging of coaxial connectors will detect and prevent the following problems. Inner conductor protrusion. This may result in buckling of the socket contacts or damage the internal structure of a device due to the axial forces generated. Inner conductor recession. This will result in poor voltage reflection coefficient, possibly unreliable contact and could even cause breakdown under peak power conditions. Appendix 4.A shows a list of the most common types of coaxial connector in use. Appendix 4.B gives information on the various connector types including the critical mechanical dimensions that need to be measured for the selected connector types.

4.5

Connector cleaning

To ensure a long and reliable connector life, careful and regular inspection of connectors is necessary and cleaning of connectors is essential to maintaining good performance. Connectors should be inspected initially for dents, raised edges, and scratches on the mating surfaces. Connectors that have dents on the mating surfaces will usually also have raised edges around them and will make less than perfect contact; further to this, raised edges on mating interfaces will make dents in other connectors to which they are mated. Connectors should be replaced unless the damage is very slight. Awareness of the advantage of ensuring good connector repeatability and its effect on the overall uncertainty of a measurement procedure should encourage careful inspection, interface gauging and handling of coaxial connectors. Prior to use, a visual examination should be made of a connector or adaptor, particularly for concentricity of the centre contacts and for dirt on the dielectric. It is essential that the axial position of the centre contact of all items offered for calibration should be gauged because the butting surfaces of mated centre contacts must not touch. If the centre contacts do touch, there could be damage to the connector or possibly to

64

Microwave measurements

other parts of the device to which the connector is fitted. For precision hermaphroditic connectors the two centre conductor petals do butt up and the dimensions are critical for safe connections. Small particles, usually of metal, are often found on the inside connector mating planes, threads, and on the dielectric. They should be removed to prevent damage to the connector surfaces. The items required for cleaning connectors and the procedure to be followed is described in the next section.

4.5.1 Cleaning procedure
Items required: (1) (2) (3) (4) (5) Low-pressure compressed air (solvent free); cotton swabs (special swabs can be obtained for this purpose); lint-free cleaning cloth; isopropanol and illuminated magnifier or a jeweller’s eye glass.

Note: Isopropanol that contains additives should not be used for cleaning connectors as it may cause damage to plastic dielectric support beads in coaxial and microwave connectors. It is important to take any necessary safety precautions when using chemicals or solvents.

4.5.1.1 First step Remove loose particles on the mating surfaces and threads using low-pressure compressed air. A wooden cocktail stick can be used to carefully remove any small particles that the compressed air does not remove. 4.5.1.2 Second step Clean surfaces using isopropanol on cotton swabs or lint-free cloth. Use only sufficient solvent to clean the surface. When using swabs or lint-free cloth, use the least possible pressure to avoid damaging connector surfaces. Do not spray solvents directly on to connector surfaces or use contaminated solvents. 4.5.1.3 Third step Use the low-pressure compressed air once again to remove any remaining small particles and to dry the surfaces of the connector to complete the cleaning process before using the connector.

4.5.2 Cleaning connectors on static sensitive devices
Special care is required when cleaning connectors on test equipment containing static sensitive devices. When cleaning such connectors always wear a grounded wrist strap and observe correct procedures. The cleaning should be carried out in a special handling area. These precautions will prevent electrostatic discharge (ESD) and possible damage to circuits.

Using coaxial connectors in measurement

65

4.6

Connector life

The number of times that a connector can be used is very difficult to predict and it is quite clear that the number of connections and disconnections that can be achieved is dependent on the use, environmental conditions and the care taken when making a connection. Some connector bodies such as those used on the Type N connector are made using stainless steel and are generally more rugged, have a superior mechanical performance and a longer useable life. The inner connections are often gold plated to give improved electrical performance. For many connector types the manufacturer’s specification will quote the number of connections and disconnections that can be made. The figure quoted may be as high or greater than 5000 times but this figure assumes that the connectors are maintained in pristine condition and correctly used. For example, the Type SMA connector was developed for making interconnections within equipment and its connector life is therefore relatively short in repetitive use situations. However, by following the guidance given in this document it should be possible to maximise the lifetime of a connector used in the laboratory.

4.7

Adaptors

Buffer adaptors or ‘connector savers’ can be used in order to reduce possible damage to output connectors on signal sources and other similar devices. It should be remembered that the use of buffer adaptors and connector savers may have an adverse effect on the performance of a measurement system and may result in significant contributions to uncertainty budgets. Adaptors are often used for the following reasons: (1) To reduce wear on expensive or difficult to replace connectors on measuring instruments where the reduction in performance can be tolerated. (2) When measuring a coaxial device that is fitted with an SMA connector.

4.8

Connector recession

The ideal connector pair would be constructed in such a way in order to eliminate any discontinuities in the transmission line system into which the connector pair is connected. In practice, due to mechanical tolerances there will almost always be a small gap between the mated plug and socket connectors. This small gap is often referred to as ‘recession’. It may be that both the connectors will have some recession because of the mechanical tolerances and the combined effect of the recession is to produce a very small section of transmission line that will have different characteristic impedance than the remainder of the line causing a discontinuity. The effect of the recession could be calculated but there are a number of other effects present in the mechanical construction of connectors that could make the

66

Microwave measurements

result unreliable. Some practical experimental work has been carried out at Agilent Technologies on the effect of recession. For more information a reference is given in Further reading to an ANAMET paper where the results of some practical measurements have been published. The connector specifications give limit values for the recession of the plug and socket connectors when joined (e.g. see Appendix 4.B for the Type N connector). The effect on electrical performance caused by recession in connectors is a subject of special interest to users of network analysers and more experimental work needs to be carried out.

4.9

Conclusions

The importance of the interconnections in measurement work should never be underestimated and the replacement of a connector may enable the Type A uncertainty contribution in a measurement process to be reduced significantly. Careful consideration must be given, when choosing a connector, to select the correct connector for the measurement task. In modern measuring instruments, such as power meters, spectrum analysers and signal generators, the coaxial connector socket on the front panel is often an integral part of a complex sub-assembly and any damage to this connector may result in a very expensive repair. It is particularly important when using coaxial cables, with connectors that are locally fitted or repaired that they are tested before use to ensure that the connector complies with the relevant mechanical specification limits. All cables, even those obtained from specialist manufacturers, should be tested before use. Any connector that does not pass the relevant mechanical tests should be rejected and replaced. Further information on coaxial connectors can be obtained direct from manufacturers. Many connector manufacturers have a website and there are other manufacturers, documents and specifications that can be found by using World Wide Web and searching by connector type.

4.A Appendix A
4.A.1 Frequency range of some common coaxial connectors Table 4A.1 lists common types of coaxial connector used on measurement systems showing the frequency range over which they are often used and the approximate upper frequency limit for the various line sizes.

4.B Appendix B
4.B.1 The 14 mm precision connector The 14 mm precision connector was developed in the early 1960s by the General Radio Company and is known as the GR900 connector. It has limited usage and is mainly used in primary standards laboratories and in military metrology. It is probably the best coaxial connector ever built in terms of its performance and it has

Using coaxial connectors in measurement Table 4A.1
Title

67

Common types of coaxial connectors used on measurement systems
Line size Impedance Upper frequency range (for normal use) Upper frequency limit (approximate value)

Precision non-sexed connectors GPC 14 GPC 14 GPC 7 Type N GPC 3.5 Type K Type Q Type V Type W 1.0 mm Type N Type N 7/16 SMA 14.2875 mm 14.2875 mm 7.0 mm 7.0 mm 3.5 mm 2.92 mm 2.4 mm 1.85 mm 1.0 mm 1.0 mm 7.0 mm 7.0 mm 16.0 mm 3.5 mm 50 75 50 50 50 50 50 50 50 50 50 75 50 50 8.5 GHz 3.0 GHz* 18 GHz 18 GHz 26.5 GHz 40 GHz 50 GHz 65 GHz 110 GHz 110 GHz 18 GHz 3 GHz* 7.5 GHz 26.5 GHz 9 GHz 8.5 GHz 18 GHz 22 GHz 34 GHz 46 GHz 60 GHz 75 GHz 110 GHz 110 GHz 22 GHz 22 GHz 9 GHz 34 GHz

Precision sexed connectors

Generel purpose conncetors

Note: ∗ Measurements made in 75 3 GHz.

impedance are normally restricted to an upper frequency limit of

low insertion loss, low reflection and extremely good repeatability. However, it is bulky and expensive. The interface dimensions for the GPC 14 mm connector are given in IEEE Standard 287 and IEC Publication 457. Before use, a visual examination, particularly of the centre contacts, should be made. Contact in the centre is made through sprung inserts and these should be examined carefully. A flat smooth disc pressed against the interface can be used to verify correct functioning of the centre contact. The disc must fit inside the castellated coupling ring that protects the end surface of the outer connector and ensures correct alignment of the two connectors when mated. The inner connector should be gauged with the collet removed. There is a special tool kit available for use with GR900 connectors. There are 50 and 75 versions of the GR900 connector available. The GR900 14 mm connector is made in two types. LPC GPC Laboratory precision connector General precision connector Air dielectric Dielectric support

68

Microwave measurements

The LPC version is usually fitted to devices such as precision airlines for use in calibration and verification kits for automatic network analysers and reflectometers (Figure 4B.1). There is also a lower performance version of the GR900 connector designated the GR890. The GR890 connector can be identified by the marking on the locking ring and it has a much reduced frequency range of operation, for example, approx. 3 GHz. 4.B.2 The 7 mm precision connector This connector series was developed to meet the need for precision connectors for use in laboratory measurements over the frequency range DC to 18 GHz. This connector is often known as the Type GPC7 connector and it is designed as a hermaphroditic connector with an elaborate coupling mechanism. The connector interface features a butt coplanar contact for the inner and outer contacts, with both the mechanical and electrical interfaces at the same location. A feature of the GPC7 connector is its ruggedness and good repeatability over multiple connections in a laboratory environment. The connector is made in two types. GPC LPC General precision connector Laboratory precision connector Dielectric support Air dielectric

The LPC version is usually fitted to devices such as precision airlines for use in calibration and verification kits for automatic network analysers and reflectometers. The interface dimensions for the GPC7 connector are given in IEEE Standard 287 and IEC 457. The most common connector of this type in the UK has a centre contact comprising a slotted resilient insert within a fixed centre conductor. The solid part of the centre conductor must not protrude beyond the planar connector reference plane, although the resilient inserts must protrude beyond the reference plane. However, the inserts must be capable of taking up coplanar position under pressure. A flat, smooth plate or disc, pressed against the interfaces can verify correct functioning of the centre contact. There are two versions of the collet for this connector: one has four slots and the other has six slots. For best performance it is good practice to replace the four-slot version with the six-slot type (see Figure 4B.2). Figure 4B.2a shows the construction of the GPC7 connector and the location of the outer conductor mating plane. The use of GPC7 connector is normally restricted to making precision measurements in calibration laboratories. 4.B.2.1 Connection and disconnection of GPC7 connectors It is important to use the correct procedure when connecting or disconnecting GPC7 connectors to prevent damage and to ensure a long working life and consistent electrical performance. The following procedure is recommended for use with GPC7 connectors.

Using coaxial connectors in measurement

69

Teflon insulator Cup-shaped self-aligning spider spring

A

A .001 to .003 Gap

Retaining stud

View A-A

Figure 4B.1

The GR 900 14 mm connector [photograph NPL]

70

Microwave measurements
(a) Dielectric support bead Outer Connector conductor nut Sleeve (fully extended)

Centre conductor Collet

Outer conductor mating plane (b)

Collet protrusion

Figure 4B.2

(a) The GPC 7 connector and (b) GPC 7 connector with the 6 slot collet [photograph A.D. Skinner]

Connection (1) On one connector, retract the coupling sleeve by turning the coupling nut until the sleeve and the nut become disengaged. The coupling nut can then be spun freely with no motion of the coupling sleeve. (2) On the other connector, the coupling sleeve should be fully extended by turning the coupling nut in the appropriate direction. Once again the coupling nut can be spun freely with no motion of the coupling sleeve. (3) Put the connectors together carefully but firmly, and thread the coupling nut of the connector with the retracted sleeve over the extended sleeve. Finally tighten using a torque spanner set to the correct torque (see Appendix D).

Using coaxial connectors in measurement Disconnection (1) Loosen the fixed coupling nut of the connector showing the wide gold band behind the coupling nut. This is the one that had the coupling sleeve fully retracted when connected. (2) Part the connectors carefully to prevent damage to the inner conductor collet.

71

It is a common but bad practice with hermaphroditic connectors, to screw the second coupling ring against the first in the belief that there should be no loose parts in the coupled pair. This reduces the pressure between the two outer contacts of the connectors, leading to higher contact resistance and less reliable contact. When connecting terminations or mismatches do not allow the body of the termination to rotate. To avoid damage, connectors with retractable sleeves (e.g. GPC7) should not be placed face down on their reference plane on work surfaces. When not in use withdraw the threaded sleeve from under the coupling nut and fit the plastic protective caps. 4.B.3 The Type N 7 mm connector The Type N connector (Figure 4B.3a) is a rugged connector that is often used on portable equipment and military systems because of its large size and robust nature. The design of the connector makes it relatively immune to accidental damage due to misalignment during mating (subject to it being made and aligned correctly). The Type N connector is made in both 50 and 75 versions and both types are in common use (Figure 4B.3). Two different types of inner socket are at present available for Type N socket connectors. They are referred to as ‘slotted’ or ‘slotless’ sockets. The slotted Type N (Figure 4B.3c) normally has either four or six slots cut along the inner conductor axis to form the socket. This means that the diameter and, therefore, the characteristic impedance are determined by the diameter of the mating pin and they are easy to damage or distort. The development of the slotless socket (Figure 4B.3b) by Agilent has resulted in a solid inner conductor with internal contacts and is independent of the
(a) (b) (c)

Figure 4B.3

(a) Type N plug, (b) Type N socket slotless and (c) Type N socket slotted [photographs NPL]

72

Microwave measurements

mating pin providing improved and more consistent performance. The slotted inner is normally only fitted to general-purpose versions of the Type N connector. The reference plane for the Type N connector is the junction surface of the outer conductors. Unlike some other pin and socket connectors the junction surface of the inner connector is offset from the reference plane by 5.258 mm (0.207 inches). The offset is designed this way in order to reduce the possibility of mechanical damage due to misalignment during the connection process. The construction and mechanical gauging requirements for the Type N connector are shown in Figure 4B.4. The offset specifications can vary and the different values shown in the table for A and B in the diagram show various values depending on the specification used. The electrical performance of a beadless airline is particularly dependent on the size of the gap at the inner connector junction due to manufacturing tolerances in both the airline and the test port connectors. Therefore for the best performance it is important to minimise the gap to significantly improve the electrical performance. It is important to note the manufacturer’s specification for any particular Type N connector being used. To meet the present MIL-STD-348A requirement the minimum recession on the pin centre contact is 5.283 mm (0.208 inches). For some applications Type N connectors need only be connected finger-tight but torque settings are given in Appendix D that should be used in metrology applications.
Type N connector Mating plane

Socket

A Gap Connector pair

Plug

B

Figure 4B.4

The Type N connector

Using coaxial connectors in measurement

73

The gauging limits are listed in Table 4B.1 and apply to both 50 and 75 connector types. For the convenience of users the dimensions are given in Imperial and Metric Units. The metric values are shown in brackets (Table 4B.1). The Type N connector is designed to operate up to 18 GHz but special versions are available that can operate up to 22 GHz and also to 26.5 GHz. (Traceability of measurement is not at present available for devices fitted with 7 mm connectors above 18 GHz.) 4.B.3.1 Gauging a plug Type N connector When gauging a plug Type N connector a clockwise deflection of the gauge pointer (a ‘plus’) indicates that the shoulder of the plug contact pin is recessed less than the minimum recession of 0.207 inches behind the outer conductor mating plane. This will cause damage to other connectors to which it is mated. 4.B.3.2 Gauging a socket Type N connector When gauging a socket Type N connector a clockwise deflection of the gauge pointer (a ‘plus’) indicates that the tip of the socket mating fingers are protruding more than the maximum of 0.207 inches in front of the outer conductor mating plane. This will cause damage to other connectors to which it is mated.

Warning
75 Type N connectors. On the 75 connector the centre contact of the socket can be physically destroyed by a 50 pin centre contact so that cross coupling of 50 and 75 connectors is not admissible. Special adaptors can be purchased, which are commonly known as ‘short transitions’, to enable the connection to be made if necessary, but these transitions should be used with caution. If possible it is best to use a minimum loss attenuation pad to change the impedance to another value.

4.B.4 The 7/16 connector This connector was developed in Germany during the 1960s for high-performance military systems and was later developed for commercial applications in analogue cellular systems and GSM base station installations. This connector is now being widely used in the telecommunications industry and it has a frequency range covering from DC to 7.5 GHz. The ‘7/16’ represents a nominal value of 16 mm at the interface for the internal diameter of the external conductor, and a nominal value of 7 mm for the external diameter of the internal conductor to achieve 50 . High-quality 7/16 connectors are available to be used as standards for the calibration of automatic network analysers, reflection analysers and other similar devices. A range of push on adaptors is available to eliminate the time-consuming need for tightening, and disconnecting using a torque spanner.

74

Table 4B.1
Dimensions in inches (mm) Plug (B) +0.003 (+0.0762) 0.207 (5.2578) 0.000 (0.0000) +0.003 (+0.0762) 0.208 (5.2832) 0.000 (0.0000) 0.208 min ( 5.2832) +0.005 (+0.1270) 0.223 (5.6642) −0.005 (−0.1270) +0.010 (+0.2540) 0.223 (5.6642) −0.010 (−0.2540) – 0.016 (0.4064) 0.001 (0.0254) 0.000 (0.000) 0.000 (0.000) Min. Nom. Max. Gap between mated centre contacts

Type N connector

Type N specification

Socket (A)

Microwave measurements
0.006 (0.1524)

MMC precision also HP precision

0.000 (0.000) 0.207 (5.2578) −0.003 (−0.0762)

MIL-STD-348A standard test

0.000 (0.0000) 0.207 (5.2578) −0.003 (−0.0762)

0.001 (0.0254)

0.007 (0.1778)

MIL-STD-348A Class 2 present Type N

0.207 max (5.2578)

0.001 (0.0254) 0.026 (0.6604)

– 0.036 (0.9144)

MMC Type N equivalent to MIL-C-71B

+0.005 (+0.1270) 0.197 (5.0038) −0.005 (−0.1270)

MIL-C-71B old Type N

+0.010 (+0.254) 0.197 (5.0038) −0.010 (−0.2540)

0.006 (0.1524)

0.026 (0.6604)

0.046 (1.1684)

Using coaxial connectors in measurement

75

It is a repeatable long-life connector with a low return loss. It also has a good specification for inter-modulation performance and a high-power handling capability. Terminations, mismatches, open and short circuits are also made and back-toback adaptors, such as plug-to-plug, socket-to-plug and socket-to-socket are also available. They are designed to a DIN specification number 47223 (Figure 4B.5). For further information on the 7/16 connector the article by Paynter and Smith is recommended. This article describes the 7/16 connector and discusses whether to use Type N connector technology or to replace it with the 7/16 DIN interface for use in mobile radio GSM base stations.

B

A

(a)

(b)

Figure 4B.5

(a) The 7/16 socket connector and (b) the 7/16 plug connector [photographs A.D. Skinner]

76

Microwave measurements 7/16 connector
Dimensions A Plug (Inches) Plug (mm) min. 1.47 1.73 max. 1.77 1.75 B Socket (Inches) min. 0.697 0.0705 max. 0.815 0.0713 Socket (mm) min. 1.77 1.79 max. 2.07 1.81

Table 4B.2
7-16

Specification General purpose Reference/test

min. 0.0579 0.0681

max. 0.0697 0.0689

Figure 4B.6

The SMA connector [photograph NPL]

The mechanical gauging requirements for the 7/16 connector are shown in Table 4B.2. 4.B.5 The SMA connector The interface dimensions for SMA connectors are listed in MIL-STD 348A. BS 9210 N0006 Part 2 published primarily for manufacturers and inspectorates, also gives details for the SMA interface but some of the requirements and specification details differ. For example, wall thickness may be a little thinner and hence a little weaker. However, MIL-STD-348A does not preclude thin walls in connectors meeting this specification although the physical requirements and arrangements will probably ensure that thicker walls are used for both specifications (Figure 4B.6). The SMA connector is a semi-precision connector and should be carefully gauged and inspected before use as the tolerances and quality can vary between manufacturers. The user should be aware of the SMA connector’s limitations and look for possible problems with the solid plastic dielectric and any damage to the plug pin. In a good quality SMA connector the tolerances are fairly tight. However the SMA connector is not designed for repeated connections and they can wear out quickly, be out of specification, and potentially destructive to other connectors. The SMA connector is widely used in many applications as it is a very cost-effective connector and suitable

Using coaxial connectors in measurement

77

for many purposes; however, precision metrology is not normally possible using SMA connectors. Connector users are advised that manufacturers’ specifications vary in the value of coupling torque needed to make a good connection. Unsatisfactory performance with hand tightening can indicate damage or dirty connector interfaces. It is common, but bad practice to use ordinary spanners to tighten SMA connectors. However, excessive tightening (>15 lb in) can easily cause collapse of the tubular portion of the pin connector. Destructive interference may result if the contacts protrude beyond the outer conductor mating planes; this may cause buckling of the socket contact fingers or damage to associated equipment during mating. The dielectric interface is also critical since protrusion beyond the outer conductor mating plane may prevent proper electrical contact, whereas an excessively recessed condition can introduce unwanted reflections in a mated pair. The critical axial interface of SMA type connectors is shown in Figures 4B.7a and b and Table 4B.3 where the dimensions are given in inches, with the equivalent in millimetres shown in parentheses. The specification allows dielectric to protrude past the outer conductor mating plane to 0.002 inches (0.0508 mm) maximum. However, there is some doubt if the SMA standards permit the dielectric to protrude beyond the reference plane. There is a high voltage version that does allow the dielectric to protrude beyond the reference plane, but it does not claim to be compatible with the SMA standard. 4.B.6 The 3.5 mm connector This connector is physically compatible with the SMA connector and is often known as the GPC 3.5 mm connector. It has an air dielectric interface and closely controlled centre conductor support bead providing mechanical interface tolerances similar to hermaphroditic connectors. However, although in some ways planar, it is not an IEEE 287 precision connector. There is a discontinuity capacitance when coupled with SMA connectors.
(a) Outer conductor Plastic dielectric Center conductor (b) Outer conductor Plastic dielectric Center conductor

Outer conductor mating plane

Outer conductor mating plane

Figure 4B.7

(a) SMA plug connector and (b) SMA socket connector

78

Table 4B.3
Pin depth in inches (mm) Socket dielectric −0.002 max 0.000 (0.000) min Plug pin Plug dielectric −0.002 max

SMA connector

SMA specification

Microwave measurements

Socket pin

MIL-STD-348A Class 2

+0.030 (0.7620) 0.000 (0.000) 0.000 (0.000) +0.002 (0.0508) 0.000 (0.000) −0.002 (−0.0508) +0.002 (−0.0508) 0.000 (0.000) 0.000 (0.000) +0.002 (0.0508) 0.000 (0.000) 0.000 (0.000) +0.005 (0.1270) 0.000 (0.000) 0.000 (0.000) +0.005 (0.1270) 0.000 (0.000) 0.000 (0.000) +0.003 (0.0762) 0.000 (0.000) 0.000 (0.000)

MMC Standard

+0.005 (0.1270) 0.000 (0.000) 0.000 (0.000)

+0.002 (0.0508) 0.000 (0.000) −0.002 (−0.0508) +0.002 (0.0508) 0.000 (0.000) 0.000 (0.000) +0.002 (0.0508) 0.000 (0.000) 0.000 (0.000)

MMC precision

+0.005 (0.1270) 0.000 (0.000) 0.000 (0.000)

MIL-STD-348A standard test

+0.003 (0.0762) 0.000 (0.000) 0.000 (0.000)

Using coaxial connectors in measurement
(a)

79

(b) Outer conductor

Outer conductor

Centre conductor Outer conductor mating plane

Centre conductor

Outer conductor mating plane

Figure 4B.8

(a) GPC 3.5 mm plug connector and (b) GPC 3.5 mm socket connector [photograph NPL] The 3.5 mm connector
Pin depth in inches (mm) Socket Plug 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

Table 4B.4

3.5 mm Specification

LPC GPC

0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

Note: A plus (+) tolerance indicates a recessed condition below the outer mating plane.

A special version of the GPC 3.5 mm connector has been designed. The design incorporates a shortened plug pin and allows the centre conductors to be pre-aligned before contact thus considerably reducing the likelihood of damage when connecting or disconnecting the 3.5 mm connector. Figure 4B.8 shows the plug and socket types of GPC 3.5 mm connector and Table 4B.4 shows the gauging dimensions. 4.B.7 The 2.92 mm connector The 2.92 mm connector is a reliable connector that operates up to 46 GHz and it is used in measurement systems and on high-performance components, calibration and

80

Microwave measurements
(a)

(b) 0.000 0.000 mm 0.005 0.127 mm

(c)

(

)

0.000 0.000 mm 0.005 0.127 mm

(

)

Figure 4B.9

(a) Type K plug and socket connector [photograph NPL], (b) Type K socket connector and (c) Type K plug connector

verification standards. It is also known as the Type K™ connector. The K connector interfaces mechanically with 3.5 mm and SMA connectors. However, when mated with the 3.5 mm or SMA connector the junction creates a discontinuity that must be accounted for in use. Compared with the 3.5 mm and the SMA connector the 2.92 mm connector has a shorter pin that allows the outer conductor alignment before the pin encounters the socket contact when mating a connector pair. The type K connector is therefore less prone to damage in industrial use. Figure 4B.9 shows the diagram of the Type K connector and Table 4B.5 gives the important gauging dimensions. 4.B.8 The 2.4 mm connector The 2.4 mm connector was designed by the Hewlett Packard Company (now Agilent Technologies) and the connector assures mode free operation up to 60 GHz. It is also known as the Type Q connector. The 2.4 mm connector is a pin and socket type

Using coaxial connectors in measurement Table 4B.5 Type K connector
Pin depth in inches (mm) Socket LPC GPC 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508) Plug 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

81

2.92 mm Specification

Table 4B.6

Type 2.4 mm connector
Pin depth in inches (mm) Socket Plug 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

2.4 mm Specification

LPC GPC

0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

connector that utilises an air dielectric filled interface. The 2.4 mm interface is also mechanically compatible with the 1.85 mm connector.
Note: The manufacturers of small coaxial connectors have agreed to the mechanical dimensions so that they can be mated non-destructively. This has led to the use of the term ‘mechanically compatible’ and because both lines are nominally 50 it has been assumed that ‘mechanically compatible’ equates to electrical compatibility. The effect of the electrical compatibility of mechanically mateable coaxial lines is discussed in ANAlyse Note No. 3 January 1994 included in the list of further reading at the end of this chapter.

As for other connectors of this type, the coupling engagement of the outer conductors is designed to ensure that the outer conductors are coupled together before the inner conductors can engage to prevent damage to the inner conductor. Figure 4B.10 shows the diagram of the 2.4 mm connector and Table 4B.6 gives the important gauging dimensions. 4.B.9 The 1.85 mm connector The 1.85 mm connector was designed by Anritsu and the connector assures mode-free operation up to 75 GHz. It is also known as the Type V™ connector. The 1.85 mm connector is a pin and socket type connector that uses an air dielectric filled interface. The coupling engagement of the outer conductors is designed to ensure that the outer conductors are coupled before the inner conductors can engage to ensure a damage-free fit. Figure 4B.11 shows the diagram of the 1.85 mm connector and Table 4B.7 shows the important gauging dimensions.

82

Microwave measurements

0.000 (0.000) to + 0.002 (+ 0.0508)

Figure 4B.10 Table 4B.7

The 2.4 mm socket and plug connector [photograph NPL] Type 1.85 mm connector
Dimensions in inches (mm) Socket Plug 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

1.85 mm Specification

LPC GPC

0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

Note: A plus (+) tolerance indicates a recessed condition below the outer mating plane.

4.B.10 The 1.0 mm connector The 1.0 mm connector was designed by Hewlett Packard (now Agilent Technologies) and is described in IEEE 287 Standard. No patent applications were filed to protect the design of the 1.0 mm connector as it is intended by Agilent to allow free use of

Using coaxial connectors in measurement
(a)

83

(b)

(c)

0.000 (0.0000) 0.004 (–0.1016)

Figure 4B.11

(a) The 1.85 mm socket and plug connector [photograph NPL], (b) Type V socket connector and (c) Type V plug connector

the interface by everyone. Any manufacturer of connectors is free to manufacture its own version of the 1.0 mm connector. The 1.0 mm connector is also known as the Type W Connector. It is a pin and socket type connector that utilises an air dielectric filled interface and assures mode-free operation up to 110 GHz. The coupling diameter and thread size are chosen to maximise strength and increase durability. The coupling engagement of the outer conductors is designed to ensure that they are coupled together before the inner conductors can engage, toe ensure a damage-free fit. Figures 4B.12 and 4B.13 show two versions of the 1.0 mm connector available from Agilent Technologies and the Anritsu Company. They are based on the dimensions shown in IEEE 287. Figure 4B.14 is a diagram of Anritsu’s W connector; a 1.0 mm connector based on the dimensions in the IEEE 287 Standard. Table 4B.8 shows the important gauging dimensions for a 1.0 mm connector.

84

Microwave measurements

Figure 4B.12

The Agilent 1.0 mm socket and plug connector [photograph NPL]

Figure 4B.13

The Agilent 1.0 mm socket and plug connector [photograph Anritsu]
Male Female

UT 47 coax

Female contact

Cable sleeve

Lock nut

Figure 4B.14

Diagram of the plug and socket arrangement

Using coaxial connectors in measurement Table 4B.8 Maximum pin depth for a 1.0 mm connector
Pin depth in inches (mm) Socket LPC GPC 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508) Plug 0 to +0.0005 (+0.0127) 0 to +0.002 (+0.0508)

85

1.0 mm Specification

Note: A plus (+) tolerance indicates a recessed condition below the outer conductor mating plate.

4.C Appendix C
4.C.1 Repeatability of connector pair insertion loss The values shown in Table 4C.1 show some insertion loss repeatability (dB) figures provided that the connector-pairs are in good mechanical condition and clean; further, that in use, they are not subjected to stress and strain due to misalignment or transverse loads. For any particular measurement process the connector repeatability in the uncertainty budget is calculated in the same units of the final measurements. For example, when measuring the calibration factor of a power sensor the repeatability is measured in per cent. These guidance figures will serve two purposes: (1) They show limits for connector repeatability for normal use in uncertainty estimates where unknown connectors may be involved. (2) Provide a measure against which a ‘real’ repeatability assessment can be judged. The figures in Table 4C.1 are based on practical measurement experience at NPL, SESC, and in UKAS Calibration Laboratories. In practice, connector repeatability is an important contribution to measurement uncertainties and should be carefully determined when verifying measurement systems or calculated for each set of measurements made. In some cases values better than those shown in Table 4C.1 can be obtained. Table 4C.1
Connector GR900 – 14 mm GPC7 – 7 mm Type N – 7 mm GPC3.5 – 3.5 mm SMA 3.5 mm

Typical connector insertion loss repeatability
Connector Insertion loss repeatability dB 0.001 (DC to 0.5 GHz) 0.001 (DC to 2 GHz) 0.001 (DC to 1 GHz) 0.002 (DC to 1 GHz) 0.002 (DC to 1 GHz) 0.002 (0.5–8.5 GHz) 0.004 (2–8 GHz) 0.004 (1–12 GHz) 0.006 (1–12 GHz) 0.006 (1–12 GHz)

0.006 (8–18 GHz) 0.008 (12–18 GHz)

86

Microwave measurements

4.D Appendix D
4.D.1 Torque wrench setting values for coaxial connectors Table 4D.1 gives a list of recommended connector tightening torque values to be used for metrology purposes for each connector type. This list is based on the best available information from various sources and should be used with care. Some manufacturers recommend slightly different values for the torque settings in their published performance data. Where this is the case the manufacturers’ data should be used. With all torque spanners, it is possible to get substantially the wrong torque by twisting the handle axially and by a variety of other incorrect methods of using the torque spanner. There are also some differences on the torque settings used when making a permanent connection (within an instrument) rather than for metrology purposes. Many manufacturers quote a maximum coupling torque which if exceeded will result in permanent mechanical damage to the connector. For combinations of GPC3.5/SMA connectors the torque should be set to the lower value, for example, 5 in-lb. Torque spanners used should be regularly calibrated, and set to the correct torque settings for the connector in use and clearly marked. On some torque spanners, the handles are colour coded to represent the torque value set for ease of identification, for example, 12 in-lb (1.36 N-m) blue and 8 in-lb (0.90 N-m) red. However, for safety, always check the torque setting before use especially if it is a spanner not owned by or normally used every day in the laboratory (Table 4D.1).

Table 4D.1

Torque spanner setting values
Torque in-lb 12 12 12 20 8 5 5–8 8 8 4 3 N-m 1.36 1.36 1.36 2.26 0.90 0.56 0.56–0.90 0.90 0.90 0.45 0.34

Connector Type GR900 GPC 7 N 7/16 GPC 3.5 SMA K Q V W1 W Size (mm) 14 7 7 16.5 3.5 3.5 2.92 2.4 1.85 1.1 1.0

Using coaxial connectors in measurement

87

4.E Appendix E
4.E.1 Calibrating dial gauges and test pieces There are a number of different types of dial gauge and gauge calibration block used for gauging connectors. They require regular calibration to ensure that they are performing correctly. There is a British Standard BS 907: ‘Specification for dial gauges for linear measurement’ dated 1965 that covers the procedure for the calibration of dial gauges and this should be used. However, the calibration of the gauge calibration blocks is not covered by a British Standard, but they can be measured in a mechanical metrology laboratory. It is important to use the correct gauge for each connector type to avoid damage to the connector under test. Some gauges have very strong gauge plunger springs that, if used on the wrong connector, can push the centre block through the connector resulting in damage. Also if gauges are used incorrectly they can compress the centre conductor collet in precision GPC 7 mm connectors, during a measurement, resulting in inaccurate readings when measuring the collet protrusion. 4.E.2 Types of dial gauge Dial gauges used for the testing of connectors for correct mechanical compliance are basically of two types: 4.E.2.1 Push on type The push on type is used for measuring the general-purpose type of connector. For plug and socket connectors two gauges are normally used (one plug and one socket) or a single gauge with plug and socket adaptor bushings. 4.E.2.2 Screw on type The screw on type is mainly used (except GR 900) in calibration kits for network analysers and reflectometers. They are used for the GPC7 and sexed connectors and for the latter they are made in both plug and socket versions. The screw on type is made in the form of a connector of the opposite sex to the one being measured. When a gauge block is used to initially calibrate the dial gauge, a torque spanner should be used to tighten up the connection to the correct torque. 4.E.3 Connector gauge measurement resolution Because of connector gauge measurement resolution uncertainties (one small division on the dial) and variations in measurement technique from user to user connector dimensions may be difficult to measure. Dirt and contamination can cause differences of 0.0001 inch (0.00254 mm) and in addition the way that the gauge is used can result

88

Microwave measurements

in larger variations. When using a gauge system for mechanical compliance testing of connectors carry out the following procedures each time: (1) (2) (3) (4) (5) (6) carefully inspect the connector to be tested and clean if necessary; clean and inspect the dial gauge, and the gauge calibration block; carefully zero the dial gauge with the gauge calibration block in place; remove the gauge calibration block; measure the connector using the dial gauge and note the reading and repeat the process at least once or more times as necessary.

4.E.4 Gauge calibration blocks Every connector gauge requires a gauge calibration block that is used to zero the gauge to a pre-set value before use. The diagram in Figure 4E.1 shows a set of dial gauges and gauge calibration blocks for a Type N connector screw on type gauge. The diagram in Figure 4E.2 shows an SMA dial gauge of the push on type with its gauge calibration block. There are a number of different types and manufacturers of connector gauge kits in general use and the manufacturer’s specification and calibration instructions should be used.

90 0 10 80 20 70 30 Marconi Instruments 60 40 50 40 30 20
0.001 mm 0.2 0 0.2 0.4

90 0 10 80 20 70 30 Marconi Instruments 60 40 50 40 30 20
0.001 mm 0.2 0 0.2 0.4

50

50

0.4

10 0 90

60 70 80

0.4

10 0 90

60 70 80

60823

60823

Dial Gauge unit

Calibration standard

Figure 4E.1

Type N screw on dial gauge and calibration block

60823

60823

Using coaxial connectors in measurement

89

5 10

5 10 10

5

5 10

15 15 MAURY MICROWAVE 20 25 20

15 15 MAURY MICROWAVE 20 25 20

SMA FP

SMA FD

M

M

Figure 4E.2

Type SMA push – on type dial gauge for socket pin depth (FP) and dielectric FD with calibration block

Further reading Uncertainties of measurement
UKAS: The expression of uncertainty and confidence in measurement, M3003, 2nd edn (HMSO, London)

ANAMET Reports1
Ridler, N. M., and Medley, J. C.: ANAMET-962: dial gauge comparison exercise, ANAMET Report, no. 001, Jul 1996 Ridler, N. M., and Medley, J. C.: ANAMET-963 live dial gauge comparison exercise: ANAMET Report, no. 007, May 1997 Ridler, N. M., and Graham, C.: An investigation into the variation of torque values obtained using coaxial connector torque spanners, ANAMET Report, no. 018, Sep 1998. French, G. J.: ANAMET-982: live torque comparison exercise, ANAMET Report, no. 022, Feb 1999 Ridler, N. M., and Morgan, A. G.: ANAMET-032: ‘live’ dial gauge measurement investigation using Type-N connectors, ANAMET Report, no. 041, Nov 2003
1 Other

reports which are fore-runners of the 2nd edition of the ANAMET Connector Guide Skinner, A. D.: Guidance on using coaxial connectors in measurement – draft for comment, ANAMET Report, no. 015, Feb 1998

F

F

90

Microwave measurements

Skinner, A. D.: ANAMET connector guide, ANAMET Report, no. 032, Jan 2001, Revised Mar 2006

ANAlyse notes
Ide, J. P. L.: A study of the electrical compatibility of mechanically mateable coaxial lines, ANAlyse Note, no. 3, Jan 1994 Ridler, N. M.: How much variation should we expect from coaxial connector dial gauge measurements?, ANAlyse Note, no. 14, Feb 1996

ANA tips notes
Smith, A. J. A., and Ridler, N. M.: Gauge compatibility for the smaller coaxial line sizes, ANA-tips Note, no. 1, Oct 1999 Woolliams, P. D. and Ridler, N. M.: Tips on using coaxial connector torque spanners, ANA-tips Note, no. 2, Jan 2000 ANAMET news articles [The first three items in this list are short, amusing, articles (albeit containing important information).] Ide, J. P.: ‘Are two collets better than one?’, ANAMET News, Issue 2, Spring 1994, p. 3 Ide, J. P.: ‘Masters of the microverse’, ANAMET News, Issue 2, Spring 1994, p. 2 Ide, J. P.: ‘More from the gotcha! files: out of my depth’, ANAMET News, Issue 9, Autumn 1997, p. 9 Instone, I.: ‘The effects of port recession on ANA accuracy’, ANAMET News, Issue 11, Autumn 1998, pp. 4–6

For further publications on connectors
Ridler, N. M.: ‘Connectors, air lines and RF impedance’, notes to accompany the IEE training course on Microwave Measurements, Milton Keynes, UK, 13–17 May 2002

Connector repeatability
Bergfield, D., and Fischer, H.: ‘Insertion loss repeatability versus life of some coaxial connectors’. IEEE Transactions on Instrumentation and Measurement Nov 1970, vol. Im-19, no. 4, pp. 349–53 Type 7/16 coaxial connector Paynter, J. D., and Smith, R.: ‘Coaxial connectors: 7/16 DIN and Type N’, Mobile Radio Technology, April 1995 (Intertec Publishing Corp.) Ridler, N. M.: ‘Traceability to National Standards for S-parameter measurements of devices fitted with precision 1.85 mm coaxial connectors’, presented at 68th ARFTG Conference, Broomfield, Colorado, Dec 2006

Chapter 5

Attenuation measurement
Alan Coster

5.1

Introduction

Accurate attenuation measurement is an important part of characterising radio frequency (RF) or microwave circuits and devices. For example, attenuation measurement of the component parts of a radar system will enable a designer to calculate the power delivered to the antenna from the transmitter, the noise figure of the receiver and hence the fidelity or bit error rate of the system. A precision power measurement system, such as the calorimeter described by Oldfield [1], requires the transmission line preceding the measurement element to be characterised to determine the effective efficiency of the system. The thermal electrical noise standard described by Sinclair [2] requires accurate attenuation measurement of the transition or thermal block between the hot termination and the ambient temperature output connector to determine its excess noise ratio.

5.2

Basic principles

With reference to Figure 5.1, when a generator with a reflection coefficient G is connected directly to a load of reflection coefficient L , let the power dissipated in the load be denoted by P1 . Now if a two-port network is connected between the same generator and load, let the power dissipated in the load be reduced to P2 . Insertion loss in decibels of this two-port network is defined as follows: L(dB) = 10 log10 P1 P2 (5.1)
G

Attenuation is defined as the insertion loss where the reflection coefficients L = 0.

and

92

Microwave measurements

Generator

ΓG

ΓL

Load

P1

Generator

ΓG

Two-port network

ΓL

Load

P2

Figure 5.1
E

Insertion loss
S21

ΓG

S11

S22

ΓL

S12

Figure 5.2

Signal flow

Note that insertion loss depends on the value of G and L , whereas attenuation depends only on the two-port network. If the source and load are not perfectly matched in an attenuation measurement, there will be an error associated with the result. This error is called the ‘mismatch error’ and it is defined as the difference between the insertion loss and attenuation. Hence Mismatch error (M ) = L − A (5.2)

Figure 5.2 shows a signal flow diagram of a two-port network between a generator and load where S11 is the voltage reflection coefficient looking into the input port when the output port is perfectly matched. S22 is the voltage reflection coefficient looking into the output port when the input port is perfectly matched. S21 is the ratio of the complex wave amplitude emerging from the output port to that incident upon the input port when the output port is perfectly matched. S12 is the ratio of the complex wave amplitude emerging from the input port to that incident upon the output port when the input port is perfectly matched.

Attenuation measurement

93

From the above definitions, the equation from Warner [3] for insertion loss is given as follows: L = 20 log10 |(1 −
G S11 ) (1 − L S22 ) − L |S21 | · |1 − G L | G S12 S21 |

(5.3)

It can now be seen that the insertion loss is dependent upon G and L as well as the S-parameters of the two-port network. When G and L are matched ( G and L = 0), then (5.3) is simplified to A = 20 log10 1 |S21 | (5.4)

where A represents attenuation (dB). From (5.2), the mismatch error M is the difference between (5.3) and (5.4). M = 20 log10 |(1 −
G S11 ) (1 − L S22 ) − G L| G L S12 S21 |

|1 −

(5.5)

Note that all the independent variables of (5.5) are complex. In practice, it may be difficult to measure the phase relationships where the magnitudes are small. Where this is the case, and only magnitudes are known, the mismatch uncertainty is given as a maximum and minimum limit. Thus 1 ± (| G S11 | + | L S22 | + | G L S11 S22 | + | G L S12 S21 |) M (limit) = 20 log10 1 ∓ | G L| (5.6) If the two-port network is a variable attenuator, then the mismatch limit is expanded from (5.6) to give M (limit) = 20 log10 × 1 ± (| 1 ∓ (|
G S11e | + | L S22e | + | G L S11e S22e | + | G L S12e S21e |) G S11b | + | L S22b | + | G L S11b S22b | + | G L S12b S21b |)

(5.7)

where suffix b denotes the attenuator at zero or datum position (residual attenuation) and suffix e denotes the attenuator incremented to another setting (incremental attenuation).

5.3

Measurement systems

Many different and ingenious ways of measuring attenuation have been developed over the years, and most methods in use today embody the following principles: (1) (2) (3) (4) (5) Power ratio Voltage ratio AF substitution IF substitution RF substitution

94

Microwave measurements
Power meter Signal generator

Insertion point

Matching attenuator

Device under test

Power sensor

Figure 5.3

Power ratio

5.3.1 Power ratio method
The power ratio method of measuring attenuation is perhaps one of the easiest to configure. Figure 5.3 represents a simple power ratio configuration. First, the power sensor is connected directly to the matching attenuator and the power meter indication noted P1 . Next, the device under test is inserted between the matching pad and power sensor and the power meter indication again noted P2 . Insertion loss is then calculated using L(dB) = 10 log10 P1 P2 (5.8)

Note that unless the reflection coefficient of the generator and load at the insertion point is known to be zero, or that the mismatch factor has been calculated and taken into consideration, measured insertion loss and not attenuation is quoted. This simple method has some limitations: (1) (2) (3) (4) Amplitude stability and drift of the signal generator Power linearity of the power sensor Zero carry over Range switching and resolution

Amplitude drift of the signal generator. Measurement accuracy is directly proportional to the signal generator output amplitude drift. Power linearity of the power sensor. The modern semiconductor thermocouple power sensor embodies a tantalum nitride film resistor, shaped so that it is thin in the centre and thick at the outside, such that when RF power is absorbed, there is a temperature gradient giving rise to a thermoelectric emf. The RF match due to the deposited resistor is extremely good and the sensor will operate over a 50 dB range (+20 dBm to –30 dBm) but there is considerable departure from linearity (10 per cent at 100 mW), and it is necessary to compensate for this and the temperature dependence of the sensitivity by electronics.

Attenuation measurement

95

Diode power sensors, described by Cherry et al. [4] may be modelled by the following equation Pin = kVdc exp(yVdc ) (5.9)

where Pin is the incident power, Vdc is the rectified dc voltage, and k and y are constants which are functions of parameters such as temperature, ideality factor and video impedance. At levels below 1 µW (−30 dBm), the exponent tends to zero, leading to a linear relationship between diode output voltage and input power (dc output voltage proportional to the square of the rms RF input voltage). For power levels above 10 µW (−20 dBm), correction for linearity must be made. Modern (smart) diode power sensors embody an RF attenuator preceding the sensing element. The attenuator is electronically switched in or out to maintain best sensor linearity over a wide input level range. These sensors have a claimed dynamic range of 90 dB (+20 dBm to −70 dBm) and may be corrected for power linearity, frequency response and temperature coefficient by using the manufacturer’s calibration data stored within the sensor e2prom. Experiments by Orford and Abbot [5] show that power meters based on the thermistor mount and self compensating bridge are extremely linear. Here the thermistor forms one arm of a Wheatstone bridge which is powered by dc current, heating the thermistor until its resistance is such that the bridge balances. The RF power changes the thermistor resistance but the bridge is automatically rebalanced by reducing the applied dc current. This reduction in dc power to the bridge is called retracted power and is directly proportional to the RF power absorbed in the thermistor. The advantage of this system is that the thermistor impedance is maintained constant as the RF power changes. Thermistors, however, have a slower response time than thermocouple and diode power sensors and have a useful dynamic range of only 30 dB. Zero carry over. Due to the electronic circuits, small errors may occur when a power meter is zeroed on one range and then used on another. Range switching and resolution. Most power meters operate over several ranges, each of approximately 5 dB. With a near full scale reading, the resolution and noise of the power meter indication will be good. However, the resolution and noise contribution at a low scale reading may be an order worse (this should be borne in mind when making measurements and all effort made to ensure that the power meter is at near full scale for at least one of the two power measurements P1 , P2 ). Figure 5.4 shows a dual channel power ratio attenuation measurement system, which uses a two-resistor power splitter to improve the source match and monitor the source output level. First, power sensor A is connected directly to the two-resistor splitter and the power meter indication is noted as PA1 , PB1 . Next the device under test is inserted between power sensor A and the two-resistor splitter and again the power meter indication is noted as PA2 , PB2 . Insertion loss may

96

Microwave measurements
Dual channel power meter Signal generator
B A

Two resistor Insertion point splitter

Power sensor B

Device under test

Power sensor A

Figure 5.4

Dual channel power ratio

now be calculated using L(dB) = 10 log10 PA1 PB2 · PB1 PA2 (5.10)

The dual channel power ratio method has two advantages over the simple system of Figure 5.3. (1) The signal level is constantly monitored by power sensor B, reducing the error due to signal generator RF output level drift. (2) Using a two-resistor power splitter or high directivity coupler will improve the source match [6,7]. The fixed frequency performance of this system may be improved by using tuners or isolators to reduce the generator and detector mismatch. A system has been described by Stelzried and co-workers [8,9] that is capable of measuring attenuation in waveguide, WG22 (26.5–40 GHz), with an uncertainty of ±0.005 dB/10 dB to 30 dB. The author has assembled a measurement system similar to that shown in Figure 5.4, with the equipment computer controlled through the general-purpose interface bus (GPIB). The system operates from 10 MHz to 18 GHz and has a measurement uncertainty of ±0.03 dB up to 30 dB, and ±0.06 to ±0.3 dB from 30 dB up to 70 dB. Return loss of the device under test is measured using an RF bridge at the insertion point. Insertion loss is then measured as described above and the mismatch error is calculated so that the attenuation may be quoted. Figure 5.5 shows a scalar network analyser attenuation and voltage standing wave ratio (VSWR) measurement system. The analyser comprises a levelled swept frequency generator, three detector channels and a display. The DC output from the detectors is digitised and a microprocessor is used to make temperature, frequency response and linearity corrections for the detectors. Mathematical functions such

Attenuation measurement

97

Scatar network analyser

Detector B

A

B

C

RF

Device under test

Open short

Detector A

Two resistor splitter

C VSWR autotester

Figure 5.5

Scalar network analyser

as addition, subtraction, and averaging may be performed, making this set up very fast and versatile. The system has a claimed dynamic range of 70 dB (+20 dBm to −50 dBm), although this will be reduced if the source output is reduced by using a two-resistor splitter or padding attenuator. Attenuation measurement uncertainty varies according to the frequency and applied power level to the detectors but is generally ±0.1 to ±1.5 dB from 10 to 50 dB. Although this may not be considered as being highly accurate, the system is fast and can identify resonances which may be hidden when using a stepped frequency measurement technique. It is also more useful when adjusting a device under test. An important consideration for all of the above power ratio (homodyne) systems is that they make wide band measurements. When calibrating a narrow band device, such as a coupler or filter, it is important that the signal source be free of harmonic or spurious signals as they may pass through the device un-attenuated and be measured at the detector.

5.3.2 Voltage ratio method
Figure 5.6 represents a simple voltage attenuation measurement system, where a DVM (digital voltmeter) is used to measure the potential difference across a feedthrough termination, first when it is connected directly to a matching attenuator, V1 , and then when the device under test has been inserted, V2 . Insertion loss may be calculated from: L(dB) = 20 log10 V1 V2 (5.11)

(Note that the generator and load impedance are matched.)

98

Microwave measurements
Signal generator Digital voltmeter

Insertion point

Matching attenuator

Device under test

Feed–through termination

Figure 5.6

Voltage ratio

This simple system is limited by the frequency response and resolution of the DVM as well as variations in the output of the signal generator. The voltage coefficient of the device under test and resolution of the DVM will determine the range, typically 40–50 dB from dc to 100 kHz. A major contribution to the measurement uncertainty is the linearity of the DVM used, which may be typically 0.01 dB/10 dB for a good quality eight digit DVM. This may be measured using an inductive voltage divider, and corrections made.

5.3.3 The inductive voltage divider
Figure 5.7 is a simplified circuit diagram of an eight-decade IVD (inductive voltage divider), described by Hill and Miller [10]. This instrument is an extremely accurate variable attenuation standard operating over a nominal frequency range of 20 Hz to 10 kHz. (Special instruments have been constructed to operate at 50 kHz and 1 MHz.) It consists of a number of auto-transformers, the toroidal cores of which are constructed of wound insulated Supermalloy tape. (Supermalloy is an alloy that is 70 per cent Ni, 15 per cent Fe, 5 per cent Mo and 1 per cent other elements, and has a permeability >100,000 and hysteresis