This is a repository copy of Analysis and design of a differential sampled-line six-port reflectometer.

White Rose Research Online URL for this paper:http://eprints.whiterose.ac.uk/82385/

Version: Accepted Version

Article:

Julrat, S, Chongcheawchamnan, M, Khaoraphapong, T et al. (1 more author) (2012) Analysis and design of a differential sampled-line six-port reflectometer. IEEE Transactionson Microwave Theory and Techniques, 61 (1). 244 - 255. ISSN 0018-9480

https://doi.org/10.1109/TMTT.2012.2227788

eprints@whiterose.ac.ukhttps://eprints.whiterose.ac.uk/

Reuse

Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version - refer to the White Rose Research Online record for this item. Where records identify the publisher as the copyright holder, users can verify any specific terms of use on the publisher’s website.

Takedown

If you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.

AbreprinfoanpeevimdeMwivestrm

im

of

de

un

Si

de

sy

ap

m

br

sy

en

el

no

a

re

m

vi

ci

an

De

So

ma

Un

bstract�The eflectometer (resented in this

n Coplanar Strior interfacing wnd four balunserformance ofvaluated with mpedance meaesigned and fa

Measurement reith the measurector networkructure gives

mode interferen

Index Termmpedance measu

The six-port

f microwave

etermines the

nder test (DU

ince phase d

etectors can b

ystem can

pplications tha

measurement [4

reast cancer d

ystems [11]-[1

Previously, m

nded designs

lectromagnetic

oise, etc. This

differential sy

ejection can be

measurement te

iewpoint, a di

ircuit in terms

nd offers high

Manuscript rece

S. Julrat, M. C

epartment of Co

ongkla University

ail: mitchai@coe

I. D. Robertso

niversity of Leed

An

analysis and(SPR) based s paper. The pip (CPS) transwith the RF sos connected tof the propos

a common-masurement. A bricated at 1 Gesults for varired results obtak analyzer. I

a significant nce signals.

ms�Reflectom

urement, transm

I. INT

reflectometer

measuremen

e complex re

UT) from four

does not nee

be used and

be realized.

at can use SP

4]-[5], moistu

detection [9],

13], to name b

most SPRs h

s which ar

c interferenc

s will be mani

ystem. It is w

e considerably

echnique [14]

ifferential circ

s of a capabili

her dynamic

eived xxxxxxx

Chongcheawcham

omputer Enginee

y, Songkhla, 90

.psu.ac.th).

on is with the I

s, Leeds LS2 9JT

alysis

Sakol Julrat,

d design of a on a sampl

practical differsmission line foource and dev the RF logared differentia

mode rejection prototype dGHz in order tous load impeained from a ct is shown tadvantage in

metry, microw

mission line me

TRODUCTION

r (SPR) is we

nts and sens

eflection coef

r signal ampli

d to be mea

a low cost m

There are

PR, such as d

ure content de

imaging sen

but a few.

ave been imp

re fundamen

ce, ground b

ifested as com

ell known tha

y improved by

]. From a cir

cuit outperfor

ity to reduce c

range [15].

mnan and T. Kh

ering, Faculty of

0112, Thailand. (

Institute of Mic

T, UK

and D

Line

Mitchai Chon

differential sled-line strucrential SPR is rorm, using twovice under testithmic detectoal SPR strucn ratio analydifferential SPto verify the a

edances are comcommercial authat the diffn rejecting co

wave measur

asurements

ll known in th

ing. The tec

fficient of a

itude measure

asured direct

measurement

e several lo

dielectric perm

etermination

nsors [10] and

plemented as

ntally sensiti

bouncing, su

mmon-mode n

at the common

y using a diffe

rcuit impleme

rms the single

common-mod

Thus, to effe

aorapapong are

f Engineering, P

(corresponding a

crowaves and P

Design

Six-P

ngcheawcham

Ian D.

six-port ture is realized o baluns t (DUT) ors. The cture is sis and

PR was analysis. mpared tomatic

ferential ommon-

rements,

he field

chnique

device

ements.

tly, RF

[1]-[3]

ow-cost

mittivity

[6]-[8],

d radar

single-

ive to

ubstrate

noise in

n-mode

ferential

entation

e-ended

de noise

ectively

with the

Prince of

author e-

hotonics,

red

an

the

stru

alg

stu

est

SP

rej

to

rea

suf

dif

par

coe

com

stru

and

giv

and

Ex

con

Se

is

com

mo

con

res

sig

fro

DC

n of a

Port R

mnan Senior M

. Robertson Fe

duce these unw

SPR system,

e design. The

ucture and d

gorithm based

In this paper

udy the conc

tablish a meth

PR performan

ection ratio (C

demonstrate

adily be app

fficient accura

fferential SPR

rameters and

efficient and

mmon-mode

ucture is pres

d simulation

ven in Section

d fabricated

xperimental re

ncept in Secti

ction VI.

II

The proposed

illustrated in

mposed of a

ode; ports 1 a

nnected to ba

spectively. Fo

gnal power me

om these four

C outputs are u

Fig. 1 P

Differ

eflecto

Member, IEEE

ellow, IEEE

wanted signal

, a differentia

e key to achi

develop a refl

d on the differe

r, mixed-mode

cept of the

hod to determi

nce is conside

CMRR). The

the feasibili

plied to a d

acy for many

R structure is

d the metho

calibration

rejection ra

sented and dis

results of th

n IV. The diffe

to operate

esults of this p

ion V. Finally

I. DIFFERENT

d SPR system

n Fig. 1. Th

a six-port ne

and 2 of the

aluns that con

our more balu

easurement po

baluns are co

used to determ

Proposed different

rentia

omete

E, Thanate Kha

ls and enhanc

al technique sh

ieve this is to

lection coeffi

ential-mode n

e S-parameter

differential S

ine the reflect

ered in terms

structure prop

ty of the co

differential S

y applications

s analyzed w

ods for dete

are presente

atio for the

cussed in Sec

he differential

erential SPR s

at 1 GHz

prototype are

y, the paper w

TIAL SPR CON

m based on a d

he differential

etwork operat

differential si

nnect to the R

uns are empl

ort, numbers

onnected to R

mine the reflec

tial six-port netw

l Sam

er

aoraphapong,

ce dynamic ra

hould be app

o use a diffe

icient determi

etwork.

r analysis is u

SPR structur

tion coefficien

s of common

posed in [16] i

oncept since

PR structure

s. In Section

with mixed-mo

ermining refl

d. An analy

e differential

ction III. The

SPR structu

structure is de

on FR4 sub

used to valida

will be conclu

NCEPT

differential str

l SPR struct

ting in diffe

ix-port netwo

RF source and

oyed, one for

3 to 6. The o

RF detectors a

ction coefficie

work structure

mpled-

and

1

ange in

lied to

erential

ination

used to

re and

nt. The

n-mode

is used

it can

e with

II, the

ode S-

flection

ysis of

SPR

design

ure are

signed

bstrate.

ate the

uded in

ructure

ture is

erential

ork are

DUT,

r each

outputs

and the

ent.

m

di

sc

th

re

po

an

co

po

di

ar

po

di

eq

w鯨怠di

re

en

an

tra

an

(1

Si欠態eq

w

po

Fig. 2

A. Mixed-m

To analyze t

mixed-mode S

ifferential and

chematic diag

he differential

eflected and i

ort 券 [17]. 欠鳥nd reflected w

ommon-mode

ort 券 [14].

ifferential wav

re differential

ort 券, respectiv

To obtain t

ifferential SP

quations can b

where 倦 = 3,4,怠態, 鯨態怠, and

ifferential SPR

epresenting th

nded port 1 an

nd 鯨頂賃態 are

ansmission co

nd 2 to the com

Rewriting (1

1d), we obtain

決 決

ince the refle態【決態, (2a) an

quation form a

where 】決鳥賃】態 a

ower waves

2 Conceptual sche

mode S-param

the proposed

S-parameters

d common-mo

ram defining

SPR. From

incident wav鳥津 and 決鳥津 a

waves at differ

incident and

The convers

ve variable ca

and common

vely.

the power w

PR structure

be defined as f決怠 噺 鯨決態 噺 鯨決鳥賃 噺決頂賃 噺5 and 6 deno鯨態態 are sin

R. 鯨鳥賃怠 and 鯨鳥he transmissio

nd 2 to the dif

mixed-mode

oefficients fro

mmon-mode p

b) in terms o

n, 決鳥賃 噺 岾鯨鳥賃態 伐決頂賃 噺 岾鯨頂賃態 伐ection coeffici

nd (2b) can b

as follows, 】決鳥賃】態 噺 】決頂賃】態 噺and 】決頂賃】態 ar

at port k, re

ematic of the diff

meter analysis

differential SP

are adopted

ode signals. F

the various w

Fig. 2, 欠津 a

ves, respective

are the differe

rential port 券.

reflected wav

sion between

an be found i

n mode power

wave equatio

in Fig. 2,

follows, 鯨怠怠欠怠 髪 鯨怠態欠態鯨態怠欠怠 髪 鯨態態欠態鯨鳥賃怠欠怠 髪 鯨鳥賃鯨頂賃怠欠怠 髪 鯨頂賃態otes the detect

ngle-ended S鳥賃態 are mixed

on coefficien

fferential port

e S-parameter

om the single-

port 倦, respect

f 欠怠 and apply

伐 聴匂入迭聴鉄鉄聴鉄迭 峇 欠態 髪伐 聴迩入迭聴鉄鉄聴鉄迭 峇 欠態 髪ient of the D

be rewritten

】畦鳥賃康 髪 稽鳥賃噺 】畦頂賃康 髪 稽頂賃re differential

spectively. 畦鳥

ferential SPR

s

PR shown in

d to describ

Fig. 2 illustra

wave paramet

and 決津 repres

ely, at single

ential-mode in

欠頂津 and 決頂津ves at common

n single-ende

in [18]. 鶏鳥津 a

r wave at diffe

n of the pr

the wave v

態 態 賃態欠態 態欠態

tor port numb

S-parameters

d-mode S-para

nts from the 倦, respectivel

rs representin

-ended signal

tively.

ying this to (

髪 岾聴匂入迭聴鉄迭 峇 決態

髪 岾聴迩入迭聴鉄迭 峇 決態

DUT (康) is e

in the power

賃】態】決態】態 賃】態】決態】態

and common鳥賃 and 稽鳥賃 a

Fig. 1,

be the

ates the

ters for

sent the

e-ended

ncident

are the

n-mode

ed and

and 鶏頂津

ferential

roposed

variable

(1a),

(1b),

(1c),

(1d),

ber. 鯨怠怠,

of the

ameters

single-

ly. 鯨頂賃怠

ng the

l port 1

1c) and

(2a),

(2b).

qual to

r wave

(3a),

(3b),

n-mode

are the

cha

Th

res

To

Fo

and

pra

com

com

eff

SP

the

sys

SP

To

(3a

per

equ

wh鶏鳥賃thr

can

exa

Co

fro

fol

wh鶏鳥the

res

ma

cal

dif

III

dif

tra

no

beh

dif

Th

qu

aracteristic pa

he differential

sidual commo

o determine 康or a perfect b

d thus only

actice, since

mmon-mode

mmon-mode s

fects to the per

B. Calibrat

In order to m

PR structure, a

e characterist

stematic error

PR hardware c

o calibrate the

a) is applied.

rform mathem

uation system

頒鶏鳥戴鶏鳥替鶏鳥泰鶏鳥滞番here 察 is the 賃 噺 】決穴倦】に is

rough the balu

n be determ

ample five sta

onsequently, th

om (4) by pe

llowing equati

here 捲沈珍 denot態袋珍 denotes t

e balun; for 倹spectively. Th

anner to the t

libration algo

fferential SPR

I. ANALYSIS

In the differ

fferential-mod

ansmission lin

ise and unba

havior relatin

fferential tran

his will degr

antify the pe

arameters of t

l signal is ob

on-mode signa

, either (3a) o

balun, the com

(3a) will be

imbalance in

noise cannot

signal adds to

rformance.

tion method

measure 康 by

a calibration p

tic parameters

rs caused by th

can also be rem

differential S

By substituti

matical manip

m presented in

番 噺 】決態】態 頒系怠怠系態怠系戴怠系替怠characteristic

the differenti

un at port k, w

mined with se

andard loads

he real and im

erforming ma

ions:-

Re岶康岼 噺 デデIm岶康岼 噺 デデ

tes the eleme

the differentia倹 = 1-4, this

he solution fo

technique pro

orithms [20]-[

R.

OF COMMON-

rential SPR

de operation

nes and balu

alanced effec

ng to the po

smission lines

rade the diff

erformance of

the differentia

btained throug

al appears fro

or (3b) is theo

mmon-mode s

applied to th

n the balun i

be complete

o the power re

y using the pr

procedure is n

s of the diff

he non-ideal c

moved in the

PR, the powe

ing k=3, 4, 5

pulation, we o

matrix form a系怠態 系怠戴 系系態態 系態戴 系系戴態系替態 系戴戴 系系替戴 系 matrix for th

ial-mode pow

where k=3, 4,

everal calibra

can be used f

maginary parts

atrix inversio

デ 捲ぬ倹ね倹噺な ぉ 牒匂 鉄甜乳デ 捲な倹ね倹噺な ぉ 牒匂 鉄甜乳

デ 捲ね倹ね倹噺な ぉ 牒匂 鉄甜乳デ 捲な倹ね倹噺な ぉ 牒匂 鉄甜乳

ent at row i a

al-mode powe

leads to 鶏鳥戴 ,or 康 in (5) is

oposed by Ho

22] can also

-MODE REJECT

structure sho

is obtained

uns. In pract

cts can occur

ower source,

s and baluns

ferential SPR

f the propose

al six-port ne

gh a balun w

om the same

oretically suff

signal is elim

he computati

is unavoidabl

ely eliminated

eading which i

roposed diffe

needed to dete

fferential SPR

characteristics

calibration pr

er equation sho

5 and 6 in (3

obtain a calib

as follows:- 系怠替系態替系戴替系替替番 頒 な】康】態Re岶康岼Im岶康岼番

he differential

wer that is ob

5, 6. In this c

ation scheme

for calibration

s of ī can be

n, resulting

and column j

er obtained th

鶏鳥替 , 鶏鳥泰 an

found in a s

oer [19] and s

be applied f

TION RATIO (C

own in Fig.

d with diffe

ice, common

r due to non

grounding sy

used in the d

R performanc

ed differential

2

etwork.

while a

balun.

ficient.

minated

on. In

le, the

d. This

in turn

erential

ermine

R. The

s of the

rocess.

own in

a) and

bration

(4)

l SPR.

btained

case, 察

es; for

n [19].

solved

in the

(5a),

(5b),

of 察-1.

hrough

d 鶏鳥滞 , similar

several

for the

CMRR)

1, the

erential

n-mode

n-ideal

ystem,

design.

ce. To

l SPR,

3

common-mode rejection ratio (CMRR) analysis is applied.

Normally, CMRR is defined by the differential-mode

transmission versus the common-mode signal transmission

from the input port (port A) to output port (port B) as follows

[14];-

系警迎迎 噺 にど log怠待 嵳聴匂匂遁豚聴迩迩遁豚 嵳 (6),

where CMRR is in decibels (dB). 鯨鳥鳥喋凋 and 鯨頂頂喋凋 denote the

differential transmission coefficients and common mode

transmission coefficients from input port A to output port B,

respectively. Similarly, the CMRRs of the 2-port differential

transmission line that is used in the proposed differential SPR

can be defined as follows,

系警迎迎態怠 噺 にど log怠待 嵳聴匂匂鉄迭聴迩迩鉄迭 嵳 (7a), 系警迎迎怠態 噺 にど log怠待 嵳聴匂匂迭鉄聴迩迩迭鉄 嵳 (7b),

where 系警迎迎態怠 and 系警迎迎怠態 refer to the CMRR from

differential port 1 to 2 and port 2 to 1, respectively. 鯨鳥鳥態怠 and 鯨鳥鳥怠態 are differential-mode transmission parameters from port

1 to 2 and from port 2 to 1, respectively. 鯨頂頂態怠 and 鯨頂頂怠態 are

common-mode transmission parameters from port 1 to 2 and

from port 2 to 1, respectively.

For the 3-port balun circuit that is used in the proposed

differential SPR, the CMRR is defined in a different way

since the common-mode port is unavailable at the single-

ended port of the balun. This can be defined as follows [14];-

系警迎迎鳥鎚 噺 にど log怠待 嵳聴匂濡聴迩濡 嵳 (8a), 系警迎迎鎚鳥 噺 にど log怠待 嵳聴濡匂聴濡迩嵳 (8b),

where 系警迎迎鳥鎚 and 系警迎迎鎚鳥 refer to the single-ended to

differential ports and from differential to single-ended ports,

respectively. 鯨鳥鎚 and 鯨頂鎚 denote the transmission coefficients

from single-ended port to differential and to common modes,

respectively. 鯨鎚鳥 and 鯨鎚頂 are the transmission coefficients

from differential and common-mode to single-ended modes,

respectively.

By applying the CMRR definitions of differential line and

balun circuit to the proposed differential SPR shown in Fig. 1,

we define its CMRR as the ratio between the differential and

common-mode signal transmissions at power reading port k (k

being 3, 4, 5 or 6) referred to the single-ended RF input port 1

as follows,

系警迎迎賃怠 噺 にど log怠待 嵳聴匂入迭銚迭聴迩入迭銚迭嵳 (9)

Eliminating 欠怠, by using (3b) and rewriting in term of 康, we

obtain,

系警迎迎賃怠 噺 にど log怠待 嵳岫聴匂入鉄聴鉄迭貸聴匂入迭聴鉄鉄岻辻袋聴匂入迭岫聴迩入鉄聴鉄迭貸聴迩入迭聴鉄鉄岻辻袋聴迩入迭 嵳 (10)

From (10), it is shown that the CMRR of the proposed

differential SPR is a function of the load, with reflection

coefficient 康, at port 2. This means that to evaluate the CMRR

of the differential SPR, various DUT measurements are

required - in particular, five standard loads 康 of 0, -1, 1,+j and

-j. However, 鯨頂賃怠 and 鯨頂賃態 in (10) will not be obtained by

direct measurement since the common-mode signal is

unavailable at single-ended port of the 3-port balun. In this

paper, 鯨鳥賃沈, and 鯨頂賃沈 where i is either 1 or 2, can be obtained

by the steps given below:-

Step 1: Measure the 3-port single-ended S-parameters of the

balun connected at power reading port k. Convert the single-

ended S-parameters of the balun to mixed-mode S-parameters.

Step 2: Measure the 3-port single-ended S-parameter of the

differential SPR structure. These three single-ended ports are

port i, k+ and k-. Noted that port k+ and k- are differential ports

of differential line at power reading port k that is connected to

the balun, and can be measured by removing the balun at

power reading port k.

Step 3: Convert the 3-port single-ended S-parameters

obtained from Step 2 to mixed-mode S-parameters.

Step 4: Connect the mixed-mode S-parameters obtained

from Step 1 and Step 3 in the flow graph shown in Fig.3.

The above measurement procedure is based on a two-port

network analyzer measurement. During the measurement

procedure, only two ports are measured and the other ports are

terminated by its characteristics impedance. This is similar to

the 3-port balun S-parameter extraction [23].

In Fig. 3, the mixed-mode S-parameters obtained from Step

3 are indicated with the bar symbol. The symbol i can be

either 1 or 2, for port 1 or port 2 of the differential SPR and

k={3, 4, 5, 6} for the power reading ports. The dark, gray and

dotted lines represent the differential-mode, common-mode

and conversion-mode paths, respectively. From Fig. 3, we

apply Mason�s rule [24] to solve for 鯨鳥賃沈 and 鯨頂賃沈, giving the

expressions (11a) and (11b).

Fig. 3 Signal flow graph for calculating Sdki and Scki

TA

(s

lin

lin

ill

de

di

in

CP

re

us

en

po

de

po

an

IV

is

ca

to

gi

in

of

ch

w

tra

m

qu

at

w

di

ex

lin

m

sin

m

tra

5(

di

cr

鯨鳥賃沈 噺鯨頂賃沈 噺

ABLE I. DIMEN

substrate height=

Chara

impeda

5

1

2

IV.

To demonstr

ne SPR (SL-S

ne without

lustrated in F

evelopment

ifferential SPR

n the coplanar

PS line are c

espectively. T

sing coupling

nded, microst

ower is subs

etector.

To validate t

ort network is

nd Ansoft H

V(A). The CM

then calculate

A. Symme

The dimens

alculated for a

ool [27] and th

iven in Table I

nsertion loss o

f 50 and 100

haracteristic im

width and 4.30

ansformers co

measurement w

uarter-wavelen

t 1 GHz. The

was verified

ifferential an

xtracted by ter

ne with char

measuring sing

ngle-ended S-

mixed-mode S-

Fig. 5(a) s

ansmission re

(a)-5(b), it is

ifferential-mod

ross-mode con

噺 怠貸岫聴違匂匂入入聴匂噺 怠貸岫聴違匂匂入入聴匂匂NSIONS OF SYM

=1.6 mm, copper t

acteristic

ance (ȍ) W

50

100

200

DIFFERENTIA

rate the diffe

SPR) based on

reference gr

Fig. 4. This S

of the sing

R uses a diffe

strip (CPS) m

connected to

The RF signal

resistors, as p

trip-based, S

sequently con

the system ope

s designed an

HFSSTM softw

MRR of the C

ed based on th

trical CPS des

sions of the

a FR-4 substr

he parameters

I (The lower i

of differential a

ȍ are difficul

mpedance of

0 mm gap. C

ould be emplo

with a stand

ngth of the se

performance

by simulatin

nd common-m

rminating all f

racteristic im

gle-ended S-pa

-parameters o

-parameters, w

shows the

esponses of th

clear that the

de signals bet

nversion is ob

岫聴違匂入日聴匂袋聴違迩匂入入聴匂迩袋聴違岫聴違迩入日聴濡匂袋聴違迩匂入入聴匂迩袋聴違迩

MMETRICAL CP

thickness=0.05mm

Width (mm) S

3

2.54

2.54

AL SAMPLED-L

erential SPR c

n a purely diff

round plane

SL-SPR is pr

gle-ended st

erential transm

medium [25].

the baluns at

ls are sample

previously rep

SL-SPR [26].

nverted to D

eration, the C

nd simulated w

ware, as desc

CPS sampled-l

he simulated r

sign

e symmetrica

rate using a r

s for 50, 100

impedance of

and common-

lt to implemen

200 ȍ was u

Commercially-

oyed to the CP

ard 50 ȍ ne

elected CPS d

of a CPS line

ng with An

modes of th

four ports of t

mpedance 100

arameters. Th

of the CPS lin

which are plott

differential

he symmetric

e designed C

tter than the c

btained as sho

聴濡匂岻盤怠貸岫聴違匂迩入入聴聴違迩迩入入聴迩迩袋聴違匂迩入入濡迩岻盤怠貸岫聴違匂匂入入聴匂迩迩入入聴迩迩袋聴違匂迩入入聴PS ON FR4 SUBS

m, 綱追=4.55, tan 絞Separation (mm)

0.14

0.48

4.30

LINE SPR

concept, a sa

ferential transm

is consider

roposed as a

tructure [12]

mission line r

The two ends

t port 1 and

d from ports

ported for the

. The sampl

DC voltages

PS sampled-li

with Agilent A

cribed in sub

line six-port n

results.

al CPS line

recognized sy

and 200 ȍ li

CPS line allo

-modes). Impe

nt practically,

used, with 2.

-available 1:4

PS design and

etwork analy

design is 45 m

e with 70 mm

nsoft HFSSTM

he CPS line

the symmetric

0 ȍ to groun

he resulting fo

ne were conve

ted in Fig. 5.

and common

al CPS line.

CPS line can s

common mod

own in Fig. 5(

聴迩匂袋聴違迩迩入入聴迩迩岻匪入聴迩匂岻袋聴違匂匂入入聴違迩迩匂匂袋聴違匂迩入入聴迩匂岻匪聴迩匂岻袋聴違匂匂入入聴違迩迩STRATE 絞 =0.02)

ampled-

mission

red, as

simple

]. The

realized

s of the

port 2,

3 to 6

single-

led RF

by RF

ine six-

ADSTM

bsection

network

e were

ynthesis

nes are

ows low

edances

, so the

54 mm

4 balun

d allow

yzer. A

mm long

m length M. The

e were

cal CPS

nd and

our-port

erted to

n-mode

In Fig.

support

de. Low

c). The

com

tha

N

net

mm

gap

res

per

wit

lum

CP

to

im

dif

por

sin

par

tra

dif

the

Ar

coe

mi

1.2

net

bet

and

to

dif

are

bec

sho

of

dep

F

匪袋聴違匂入日聴濡迩岫聴匂匂聴違迩入入岫聴匂匂聴迩迩貸聴匂袋聴違迩入日聴濡匂岫聴匂迩聴違迩入入岫聴匂匂聴迩迩貸聴匂迩mputed CMR

an 3 dB CMRR

B. Differen

Next, the c

twork, as show

m long and ea

p. To couple

sistor couplin

rformance of

thin the work

mped resistor

PS line, with e

100 ȍ imped

mpedance of

fferential port

rt network w

ngle-ended S-

rameters were

Fig. 6(a) s

ansmission co

fferential trans

e differential

round 1 GHz,

efficients are

inimum comm

26 GHz. Th

twork achiev

tween port 1

d common-mo

the power re

fferential and

e plotted. Low

cause a desig

ould be noted

the SL six-p

pend on the co

Fig. 4 Differentia

聴違迩匂入入袋聴迩匂聴違迩迩入入匂迩聴迩匂岻袋聴違匂迩入入聴違迩匂匂入入袋聴迩迩聴違匂迩入入迩聴迩匂岻袋聴違匂迩入入聴違迩匂

RR of the CPS

R at 1 GHz is

ntial sampled-

complete dif

wn in Fig. 4, w

ach differentia

e the RF sign

ng is required

f CPS line an

ing range of c

is applied fo

each single-en

ance. The 100

the 200 ȍt of balun. Th

was simulated

parameters ar

e calculated [2

hows the d

efficients from

smission coef

reflection co

the common-

less than -4

mon-mode tran

is means tha

ves low com

and 2. Fig. 6(

ode transmiss

eading port (p

eight commo

w differential

gn with large

d that the diff

port network

oupling resisto

al sampled-line si

入岻迩匂入入岫聴匂迩聴迩匂貸聴匂入岻匂入入岫聴匂迩聴迩匂貸聴匂S line is shown

s obtained.

lined six-port

fferential sam

was modeled.

al port is sepa

nal from the

d to minimize

nd keep the

commercial R

or coupling R

nded power m

0 ȍ is single-

differential

he differentia

d with Ansof

re obtained. T

28] and are plo

differential a

m port 1 to 2

fficient is bette

oefficient is

mode transmi

4 and -3 dB

nsmission coe

at the differ

mmon-mode s

(b) and 6(c) s

ion coefficien

port 3 to 6),

n-mode transm

l signal transm

coupling res

fferential trans

and its freque

ors.

ix-port network fo

匂匂聴迩迩岻

匂匂聴迩迩岻

n in Fig. 5(d).

t network

mpled-line si

. The CPS line

arated with a 1

CPS line, a

e the insertio

RF coupled

RF detectors. A

RF signals fro

measurement p

ended charact

line CPS an

al sampled-lin

ft HFSSTM. T

The mixed-mo

otted in Fig. 6

and common

2. In Fig. 6(a

er than -4 dB,

less than -1

ission and refl

, respectively

fficient is -14

rential SL si

signal transm

how the diffe

nts from port 1

respectively.

mission coeff

mission is ob

sistors is cho

smission coef

ency characte

for HFSS� simul

4

(11a)

(11b)

. More

ix-port

e is 70

10 mm

a large

on loss

signal

A 2 kȍ

om the

port set

teristic

nd the

ne six-

Twelve

ode S-

6.

n-mode

a), the

, while

7 dB.

flection

y. The

4 dB at

ix-port

mission

erential

1(or 2)

Eight

ficients

btained

sen. It

fficient

eristics

lation

Tr

Cr

H

co

sim

ba

Fig. 5 Compu

ransmission for d

ross-mode conver

C. Differen

Here, the diffe

ombining the

mulation as d

aluns. The per

uted mixed-mod

differential mode

rsion, and (d) Com

ntial SL six-po

erential SPR

e differential

described in

rformances of

(a)

(b)

(c)

(d)

de S-parameters

, (b) Reflection

mmon-mode reje

ort network w

shown in Fi

sampled-lin

Section IV(C

f the SPR in te

of the CPS

for differential m

ection ratio.

with ideal balu

g. 1 is mode

e six-port n

C) with six id

erms of CMRR

line, (a)

mode, (c)

uns

eled by

network

deal 1:2

Rs,

Figsix-

coe

from

coe

dif

fro

sin

obt

cal

of

Fig

pow

low

of

eli

ref

des

loa

sho

CM

g. 6 Simulated m

-port network:

efficients from po

m port 1 and 2

efficients from po

fferential and

om (10)-(11b)

ngle-ended, d

tained from (

lculated differ

the differenti

g. 7(c), the co

wer sampling

w. Such neglig

using ideal b

minate comm

Figs. 7(e)-7(g

ferred to por

signed with i

ads. The resul

own in Fig. 7

MRRs vary wi

mixed-mode S-par

(a) Differential-

ort 1 to 2, (b) Di

2 to ports 3 to

ort 1 and 2 to port

common-mod

). Figs. 7(a),

differential an

(12a) and (12b

rential-mode

ial SL six-por

ommon-mode

g ports, either

gible common

baluns, which

mon-mode sign

g) show the ca

rt 1. The di

ideal baluns i

lts are obtaine

7(a)-7(c) to e

ith the termina

(a)

(b)

(c) rameters of the

mode and comm

ifferential-mode t

6, and (c) Com

ts 3 to 6

de transmissio

(b) and 7(c) s

nd common-m

b), respective

S-parameters

rt network sho

e transmission

r referred to p

n-mode transm

h have an exc

nals.

alculated CM

ifferential SL

is terminated

ed by applyin

eqns. (10)-(11

ating load at p

differential samp

mon-mode trans

transmission coe

mmon-mode trans

on can be calc

show the calc

mode S-param

ely. In Fig. 7(

are similar t

own in Fig. 6

n coefficients

port1 or 2, ar

missions are a

cellent capabi

MRRs of ports

L six-port ne

with five di

ng the S-param

1b). As show

port 2. The ob

5

pled-line

smission

fficients

smission

culated

culated

meters

a), the

to that

6(a). In

at the

re very

a result

ility to

3 to 6

etwork

fferent

meters

wn, the

tained

6

(a) (b)

(c) (d)

(e) (f)

(g)

Fig. 7 Simulated performance of the differential sampled-line six-port network with ideal baluns: (a) Differential mode transmission, (b) Single-ended mode

transmission, (c) Common-mode transmission, (d) CMRR at port 3, (e) CMRR at port 4, (f) CMRR at port 5, and (g) CMRR at port 6.

CMRRs at all power sampling ports are better than 60 dB,

which is mainly due to the ideal balun performance. Clearly

the actual balun performance is very important for eliminating

common-mode noise in the proposed differential SPR � more

so than the CPS differential line performance. This is because

the CMRR of the balun is much greater than the CPS

differential line.

D. Performance for impedance measurement

In this section, the proposed differential SPR is simulated to

predict its behavior with several load impedances. The

simulation is performed for two different design cases � with

ideal and imperfect baluns. In practice, baluns will always

exhibit imbalance and we model this by inserting a phase

shifter and attenuator to artificially produce imbalance effects

as illustrated in Fig. 8. The phase shifter and attenuator are

varied from ±20 degree and ±2 dB, respectively, to producing

imbalance effects. For the ideal balun case, the phase shifter

and attenuator are set to zero.

Prior to impedance measurement, a calibration process is

performed at frequencies around 1 GHz. The 1 GHz simulated

re

im

re

ca

qu

di

W

fr

th

ca

on

lin

to

50

LD

Fig. 8 Simula

Fig. 9 Simul

(a) A

esults of the

mbalance of

espectively. T

alculated and

uantify the i

ifferential SPR

We performed

equencies and

hat imbalance

alibration proc

V

The different

n FR4 substra

ne with 2.54 m

o achieve 200

0-to-200 ȍ

DB21942M20

ation setup to test

ation results of im

Amplitude imbalan

SPR with

baluns are

The magnitud

the actual lo

impedance m

R. As shown,

d the impe

d still obtaine

effect in the

cess.

V. DESIGN A

tial SPR struc

ate for operati

mm width and

ȍ differential

baluns [2

0C-001) as the

imbalance effect

(a)

(b)

mbalance effect fo

nce and (b) Phase

amplitude im

illustrated in

des of the

oad impedanc

measurement p

, we obtain v

edance meas

ed accurate re

e baluns can

AND MEASURE

cture was des

on at 1 GHz.

d 4.30 mm sep

l characteristic

29] were s

e differential p

t for differential S

for differential SP

e imbalance

mbalance and

n Fig. 9(a) an

error betwe

ces are compu

performance

very accurate

surement at

esults. This su

be removed

EMENT

igned and fab

A symmetric

paration was s

c impedance.

selected (par

port is matche

SPR

PR:

d phase

nd (b),

en the

uted to

of the

results.

other

uggests

by the

bricated

cal CPS

selected

Murata

rt no.

ed to

the

and

net

fro

06

res

cho

[30

the

sin

(0

fro

dig

con

dat

usi

SP

con

thr

the

inv

per

(C

con

the

kn

eva

Un

LD

S-p

2-p

con

bal

of

mo

sho

the

LD

and

be

Fig. 10 P

e 200 ȍ char

d the single-en

twork analyz

om the CPS lin

03). The RF

sistors is abou

osen power d

0] are used fo

e baluns. Its

ngle-ended po

to -60 dBm)

om the differ

gital converte

nvert the DC

ta is then tran

ing a program

Fig. 10 show

PR. Port �P1�

nnected to th

rough baluns t

e performanc

vestigated and

rformance of

MRR and

nsidered in de

A. Balun

As mentioned

e design for re

owledge of t

aluate the pe

nfortunately,

DB21942M20

parameters of

port single-e

nnected to a

lun ports so t

the balun are

After the 3-p

ode S-parame

ows the comp

e measured

DB21942M20

d cross-mode

seen that the

hotograph of the

racteristic imp

nded port can

zer. The RF

ne by using 2

signal coupli

ut -17 dB, whi

detector. LT®5

or determining

input impeda

rt of the balun

) that can cov

rential CPS l

ers are conn

signals from

nsferred to a P

m developed in

ws a photogra

is connected

he DUT and p

to the power-s

ce of the f

d discussed in

f the baluns

impedance

etail.

d previously, t

educing unwa

the S-parame

erformance of

the 3-po

C-001 Murata

f the balun we

ended S-para

matched load

that all the ele

obtained.

port S-parame

eters of the

puted mixed-m

d single-en

C-001 balun.

transmission

differential re

fabricated differe

pedance of di

n be measured

differential s

2 kȍ lumped S

ng obtained w

ich suits the dy

5534 logarithm

g the RF diffe

ance (50 ȍ)

n, and it has w

ver the range

line. On-boar

nected to the

the detectors

PC and the cal

n LabVIEW�

aph of the fab

d to the RF s

ports �P3� to

sampling dete

fabricated di

n the followin

s and the d

measuremen

the balun is th

anted common

eters of the b

f the propose

ort S-param

a balun were

re obtained by

ameters with

d. This proces

ements in the

eters were ob

balun were

mode S-param

nded S-para

Fig. 11(a) s

n coefficients

eflection coeff

ential SL-SPR

ifferential line

d by a standard

signals are co

SMT resistors

with 2 kȍ co

ynamic range

mic power det

erential powe

is matched

wide dynamic

of power re

rd 12-bit ana

detectors [3

to binary dat

culations perf

�.

bricated diffe

source, port �

�P6� are con

ectors. Validat

ifferential SP

ng subsection

differential SL

nt accuracy)

he key compon

n-mode noise.

balun is need

ed differential

meters of

unavailable. S

y measuring a

h the third

ss is repeated

3-port S-para

btained, the m

calculated. F

meters, derived

ameters of

hows the refl

of the balun.

ficient is well

7

e CPS

d 50 ȍ

oupled

s (style

oupling

e of the

tectors

r from

to the

c range

adings

alog to

31] to

ta. The

formed

erential

�P2� is

nnected

tion of

PR is

ns. The

L-SPR

) are

nent in

Thus,

ded to

l SPR.

the

So, the

a set of

d port

d to all

ameter

mixed-

ig. 11

d from

the

flection

It can

below

F

-2

di

co

G

co

tra

co

Th

fo

fr

be

us

SP

di

sim

Fig.11 Measured

transmission coe

25 dB, the c

ifferential and

ommon-mode

GHz, respectiv

oefficients and

ansmission c

ommon-mode

his confirms t

or both differe

equency. As

etter than 20

sed to reduce t

B. CMRR

To evaluate

PR, the me

ifferential SL

mulated S-pa

d S-parameters of

efficients, (b) Tran

mode r

cross-mode t

d common-mo

reflection co

vely. Fig. 1

d it can be se

coefficient is

transmission

that the perfo

ential and com

shown in Fig

dB at 1 GHz

the common-m

of the differen

the CMRR o

easured sing

six-port netw

arameters of

(a)

(b)

(c)

f the balun (a) Ref

nsmission coeffic

rejection ratio.

transmission

ode is less tha

efficient is be

11 (b) show

een at 1 GHz

better than

coefficient is

ormance of th

mmon-mode p

g. 11(c), the

z, meaning th

mode noise ef

ntial SL-SPR

of the fabrica

gle-ended S-

work and balu

the differenti

flection and cross

cient, and (c) Com

coefficient b

an -15 dB wh

etter than -5 d

ws the transm

z that the diffe

-3 dB wh

s well below -

he balun is su

ports at the op

computed CM

hat the balun

ffectively.

ated differenti

parameters o

uns are require

ial six-port n

s-mode

mmon-

between

hile the

dB at 1

mission

ferential

ile the

-17 dB.

ufficient

perating

MRR is

can be

ial SL-

of the

ed. The

network

obt

me

dif

end

wit

illu

coe

tra

ref

bet

12

2 t

bel

por

coe

sho

res

bal

key

go

int

tained from

easured resu

fferential SL-

ded measurem

th matched

ustrated in Fig

Fig. 12(a)

efficients betw

ansmission co

flection coeff

tween simula

(b)-12(e) show

to the power

low -15 dB ar

Fig. 13 shows

rts of the di

efficients. Co

own in Fig.

sponse of the

lun performan

y to achieving

od capability

terference.

Ansoft HF

ults. The S-

SPR were m

ments made

loads. A co

g. 12.

shows the

ween port 1

oefficient is

ficient is less

ated and mea

w the transmi

r sampling po

re achieved ov

s the calculate

ifferential SL

omparing with

11 (c), it can

differential SL

nce. Thus, a c

g a high-perfo

y to suppre

(a)

(b)

(c)

SSTM are co

-parameters

measured by u

with the othe

omparison o

transmissio

and 2. It ca

better than

than -15 dB

asured results

ission coeffici

orts (ports 3

ver the range 0

ed CMRRs at

L-SPR for fiv

h the CMRRs

n be confirm

L-SPR is main

careful selectio

ormance differ

ess common

ompared wit

of the fabr

using 2-port s

er ports term

of these resu

on and refl

an be seen th

-6 dB whil

B. Good agre

s is obtained.

ients from po

to 6). Value

0.8-1.25 GHz.

the power sam

ve DUT refl

of the balun

med that the C

nly governed

on for the balu

rential SPR th

n-mode noise

8

th the

ricated

single-

minated

ults is

flection

hat the

le the

eement

. Figs.

rt 1 or

s well

.

mpling

flection

itself,

CMRR

by the

un is a

hat has

e and

FiTr

co

an

Tr

ig. 12 Measure

ransmission and

oefficients of port

nd 2 to port 4, (d)

ransmission coeff

(d)

(e)

ed results of th

reflection coeffic

ts 1 and 2 to port

Transmission co

ficients of ports 1

(

(

)

)

he fabricated d

cients of ports 1

3, (c) Transmiss

oefficients of port

and 2 to port 6.

(a)

(b)

differential SL-S

and 2, (b) Tran

ion coefficients o

ts 1 and 2 to port

SPR, (a)

nsmission

of ports 1

5 and (e)

Fig. 13 Calc

(c

(d

culated CMRRs a

c)

d)

at port, (a) 3, (b) 4

(a)

(b)

4, (c) 5 and (d) 6

9

10

(c)

Fig.14 Measured reflection coefficients of the differential SPR structure (x), compares to automatic network analyzer (o) at (a) 0.950 GHz, (b) 0.975 GHz

and (c) 1 GHz

C. Reflection coefficient measurement

The reflection coefficient measurement of the proposed

differential SPR structure is demonstrated and discussed in

this subsection at frequencies around 1 GHz. Prior to

reflection coefficient measurement, calibration must be

performed to eliminate the systematic noise and some

uncertainties in the measurement system from various

components such as the logarithmic detector and analog-to-

digital converter. To counteract these uncertainties, a

calibration scheme based on a statistical method that uses the

weighted squared error technique [22] was chosen. This was

achieved by applying (7a) and (7b). After calibration was

performed, measurements on ten different DUTs were carried

out. These measurements are compared to those obtained from

a commercial automatic vector network analyzer (HP8510C)

in Fig. 14. Figs. 14(a)-14(c) show the reflection coefficient

measurements made with the differential SPR at three

frequencies - 0.95 GHz, 0.975 GHz and 1 GHz. It is found

that the measured results of the proposed differential SPR are

close to the measured results from the 8510C network

analyzer. The mean absolute errors compared to the results

obtained from the network analyzer are 0.032, 0.029 and

0.033 at 0.95 GHz, 0.975 GHz and 1 GHz, respectively.

D. Interference rejection in the differential SPR

In this subsection, the performance of the differential SPR

in a hostile electromagnetic environment is investigated and

compared to the single-ended SPR. A 2.5 GHz rectangular

microstrip patch antenna on Duroid 5880 substrate (substrate

height=0.787 mm, copper thickness=0.017 mm, 綱追=2.20, tan 絞 =0.0009) is chosen to generate the electromagnetic wave

interference with the differential and single-ended SPRs, as

illustrated in Fig. 15. This antenna has patch size 12x16 mm

and a feed line 2x10 mm that is connected to the wide side at

distance 2 mm away from the edge of the patch). The patch

antenna is placed above the center of the differential and

single-ended SPRs at a distance of 150 mm. The input signal

to the patch antenna was varied from -20 to +5 dBm. This

emulates common-mode noise in the SPR systems. To

validate the common-mode rejection of the differential SPR,

the reflection coefficient of various electrical loads that cover

the Smith Chart area were measured by the differential and

single-ended SPR with and without the 2.5 GHz inference

present, respectively.

(a) (b)

Fig. 15 Experimental setup to investigate the effect of electromagnetic wave

interference on: (a) differential SPR and (b) single-ended SPR

(a)

(b)

Fig.16 Effect of emulated interference on (a) differential SPR and (b) single-

ended SPR measurements.

Fig. 16 (a) and (b) show the magnitude error of reflection

coefficients at 1 GHz when using the differential and single-

ended SPRs, respectively, in the presence of interference.

From -20 to -10 dBm interference signal power, the reflection

coefficient error for the single-ended SPR is comparable to the

differential SPR. However, as the antenna feed power is

increased beyond -10dBm, it can be seen that the single-ended

SPR is no longer useable at all, whilst the differential SPR still

maintains its performance thanks to its common-mode

rejection. It should be noted for completeness that the

11

reflection coefficient error of different electrical loads is not

equal as the common-mode rejection ratio is a function of load

reflection coefficient.

VI. CONCLUSIONS

The analysis and design of a differential six-port

reflectometer has been presented. The system has been

demonstrated experimentally at 1 GHz and produced DUT

measurements that match the results from a commercial vector

network analyzer. The proposed differential SPR structure

aims to reduce the inherent common-mode noise in the

system. The differential SPR structure is composed of a

sampled-line differential six-port network with baluns. The

balanced ports of the baluns are connected to the differential

six-port network while the single-ended ports serve as the

interface with the signal source, DUT and RF detectors.

Determination of the DUT reflection coefficient with the

differential SPR structure has been analyzed by using mixed-

mode S-parameters as well as studying the common-mode

rejection ratio. The effect of imbalance of the balun can be

removed by the calibration procedure prior to the

measurement. From the CMRR results, it is shown that the

technique is capable of reducing the common-mode noise. It is

noted that the common-mode noise rejection ratio depends

mainly on the performance of balun rather than CPS

differential line, which is clearly one of the key components

determining the rejection of common-mode noise and

interference. The reflection coefficient measurements obtained

from the proposed differential SPR are in good agreement

with the commercial network analyzer. It is believed that the

proposed differential SPR structure will find various

applications for reflection coefficient measurement where

common-mode noise or interference is of concern.

ACKNOWLEDGMENT

This research is supported by The Thailand Research Fund

(TRF) for The Royal Golden Jubilee (RGJ) Ph.D Program

Scholarship (PHD/0193/2551) and Prince of Songkla

University.

REFERENCES

[1] G. F. Engen, �The six-port reflectometer: An alternative network

analyzer,� IEEE Transactions on Microwave Theory and Techniques,

vol. 25, no. 12, pp. 1075- 1080, Dec. 1977.

[2] G. F. Engen, �An improved circuit for implementing the six-sort

technique of microwave measurements,� IEEE Transactions on

Microwave Theory and Techniques, vol. 25, no. 12, pp. 1080- 1083,

Dec. 1977.

[3] C. A. Hoer, �A network analyzer incorporating two six-port

reflectometers,� IEEE Transactions on Microwave Theory and

Techniques, vol. 25, no. 12, pp. 1070- 1074, Dec. 1977.

[4] F. M. Ghannouchi and R. G. Bosisio, �Measurement of microwave

permittivity using a six-portreflectometer with an open-ended coaxial

line,� IEEE Transactions on Instrumentation and Measurement, vol. 38,

no. 2, pp. 505-508, Apr. 1989.

[5] J. Munoz, M. Rojo, and J. Margineda, �A method for measuring the

permittivity without ambiguity using six-port reflectometer,� IEEE

Transactions on Instrumentation and Measurement, vol. 42, no. 2, pp.

222-226, Apr. 1993.

[6] J. A. R. Ball, B. Horsfield, J. R. Holdem, R. B. Keam, W. S. Holmes,

and A. Green, �Cheese curd moisture measurement using a six port

reflectometer,� presented at the Proc. Asia Pacific Microwave Conf.,

New Delhi, India, vol. 2, pp. 479�482.

[7] K. Y. Lee, �Development of a microwave instrumentation system for

the determination of moisture content in oil palm fruits,� 2008.

[Online]. Available: http://psasir.upm.edu.my/5092/. [Accessed: 06-

Oct-2011].

[8] S. Julrat, M. Chongcheawchamnan, T. Kaoraopaphong, O.

Pattrapiboonchai, M. Krairiksh and I. D. Robertson, �Single frequency

based dry rubber content determination technique for in-field

measurement application�, IEEE Sensors Journal, vol. 12(10), pp. 3019

� 3030, Oct. 2012. [9] N. Seman and M. E. Bialkowski, �Design of a UWB 6-port

reflectometer formed by microstrip-slot couplers for use in a microwave

breast cancer detection system,� in 2007 IEEE Antennas and

Propagation Society International Symposium, 2007, pp. 245-248.

[10] H. C. Lu and T. H. Chu, �Microwave diversity imaging using six-port

reflectometer,� IEEE Transactions on Microwave Theory and

Techniques, vol. 47, no. 1, pp. 84-87, Jan. 1999.

[11] C. G. Miguelez, B. Huyart, E. Bergeault, and L. P. Jallet, �A new

automobile radar based on the six-port phase/frequencydiscriminator,�

IEEE Transactions on Vehicular Technology, vol. 49, no. 4, pp. 1416-

1423, Jul. 2000.

[12] A. Khy and B. Huyart, �A 94 GHz radar using a six-port reflectometer

as a phase/frequency discriminator,� in Radar Conference, 2005.

EURAD 2005. European, 2005, pp. 213-215.

[13] B. Boukari, E. Moldovan, S. Affes, K. Wu, R. G. Bosisio, and S. O.

Tatu, �A heterodyne six-port FMCW radar sensor architecture based on

beat signal phase slope techniques,� Progress In Electromagnetics

Research, vol. 93, pp. 307-322, 2009.

[14] W. R. Eisenstadt, B. Stengel, and B. M. Thompson, Microwave

differential circuit design using mixed mode S-parameters. Artech

House Publishers, 2006.

[15] N. Yang, C. Caloz and K. Wu, "Greater than the sum of its part,"

Microwave Magazine, IEEE, vol 11, no. 4, pp. 69-82 June 2010.

[16] E. J. Griffin, �Six-port reflectometer circuit comprising three directional

couplers,� Electronics Letters, vol. 18, no. 12, pp. 491-493, Jun. 1982.

[17] K. Kurokawa, �Power waves and the scattering matrix,� IEEE

Transactions on Microwave Theory and Techniques, vol. 13, no. 2, pp.

194- 202, Mar. 1965.

[18] D. E. Bockelman and W. R. Eisenstadt, �Combined differential and

common-mode scattering parameters: theoryand simulation,� IEEE

Transactions on Microwave Theory and Techniques, vol. 43, no. 7, pp.

1530-1539, Jul. 1995.

[19] C. A. Hoer, �Using six-port and eight-port junctions to measure active

and passive circuit parameters,� NBS Tech, Note 673, Sep. 1975.

[20] F. M. Ghannouchi and R. G. Bosisio, �An alternative explicit six-port

matrix calibration formalism usingfive standards,� IEEE Transactions

on Microwave Theory and Techniques, vol. 36, no. 3, pp. 494-498, Mar.

1988.

[21] P. I. Somlo and J. D. Hunter, �A six-port reflectometer and its complete

characterization by convenient calibration procedures,� IEEE

Transactions on Microwave Theory and Techniques, vol. 30, no. 2, pp.

186-192, Feb. 1982.

[22] S. P. Jachim and W. D. Gutscher, �A statistical method for calibrating

the six-port reflectometerusing nonideal standards,� IEEE Transactions

on Microwave Theory and Techniques, vol. 37, no. 11, pp. 1825-1828,

Nov. 1989.

[23] K. Jung et al.,"Marchand balun embedded probe," IEEE

Transaction on Microwave Theory and Techniques, IEEE

Transactions, vol.56, no.5, pp.1207-1214, May 2008 [24] N. S. Nise, Control Systems Engineering, 5th ed. Wiley, 2007.

[25] R. N. Simons, Coplanar waveguide circuits, components, and systems,

John Wiley & Sons, Inc, 2001.

[26] B. M. Altrabsheh and I. D. Robertson, �A multi-probe microwave

reflectometer using an improved calibration algorithm,� in Microwave

Conference, 2006. 36th European, 2006, pp. 979-982.

[27] J.W.P.Ng, �Ascps software version 1.1,� 1999.

[28] L. B. Lok, �Balanced microwave circuit and adaptive single-sideband

mixers�, Ph.D. Thesis, University of Leeds, Leeds, London. May 2007.

[29] Murata Manufacturing Co., �Datasheet chip multilayer hybrid baluns

LDB21942M20C-001�.

[30] Linear technology cooperation, �Data sheet LT5534 50MHz to 3GHz

RF power detector with 60 dB dynamic range�, 2004.

[31] ARM Ltd., �Cortex-M3 technical reference manual�, 2005.

20

So

inc

ap

So

Ch

in

gra

cu

wa

joi

In

Ki

008, he joined th

ongkhla, Thailand

clude microwave

pplications. Dr. C

ociety.

hair for the 2011

microwave and

ants as PI and

urrently leads an

ave LTCC comp

ining Leeds he

stitute at the Un

ing's College Lon

Sakol JMechatro

degree in

Prince of

disk driv

Bachelor

in Comp

Universit

applicatio

measurem

Mitchai Bangkok,

in teleco

Institute o

1992, the

signal pro

U.K., in

engineerin

Guildford

He join

Technolo

he Faculty of En

d, as an Associat

e circuit design a

Chongcheawcham

Ian RobeCentenary

Wave Cir

Electronic

published

and millim

editing th

Design an

the IEEE i

and millim

He was G

European Micro

millimetre-wave

led 5 major E

IeMRC-funded p

onents with Lou

was a founder

niversity of Surre

ndon.

Julrat obtained

onics Engineerin

n Electrical Eng

f Songkla Unive

ves industry for

rs degree. He is

puter Engineerin

ty. His research

on, dielectric m

ment systems.

Chongcheawch, Thailand. He re

ommunication

of Technology L

e M.Sc. degree

ocessing from Im

1995, and the P

ng from the

d, U.K., in 2001.

ned the Maha

gy, Bangkok, in

ngineering, Princ

te Professor. His

and microwave te

mnan is a member

ertson holds th

y Chair in Mic

rcuits and is H

c & Electrical

over 400 papers

metre-wave engi

he well-known

nd Technology. H

in 2012 for contr

metre-wave system

General Technica

owave Week � the

e engineering. H

EPSRC-funded co

project in 3D m

ghborough and I

member of the

ey and led the M

d Bachelors de

ng in 2005 and

gineering in 20

rsity. He worked

two years follow

currently a Ph.D

ng, Prince of

h interests are

measurement, sen

hamnan was

eceived the B.En

from King M

Ladkrabang, Ban

in communica

mperial College,

Ph.D. degree in

University of

anakorn Univer

n 1992, as a Lec

ce of Songkla Un

current research

echniques for ag

r of the IEEE M

he University o

crowave and Mi

Head of the Sc

Engineering.

in the area of m

ineering as wel

book RFIC and

He was elected F

ributions to MMI

m-in-package tec

al Programme Co

e premier Europe

He has held over

ollaborative proj

microwave and mi

Imperial College.

e Advanced Tec

MMIC Research

egree in

d Master

009 from

d in hard

wing the

D. student

Songkla

six-port

nsors and

born in

ng.degree

Mongkut�s

ngkok, in

tion and

London,

electrical

Surrey,

rsity of

cturer. In

niversity,

interests

ricultural

icrowave

of Leeds

illimetre-

chool of

He has

microwave

ll as co-

d MMIC

Fellow of

IC design

chnology.

ommittee

ean event

r £3M in

jects. He

illimetre-

. Prior to

chnology

Team at

12