ffi----***^-*- :m

ffi:::::-;-
ilr:-*::-.*--iru
ffi*-::.-*--tr
m-:::*:*-_ffi

Application
of ts&KEqiriprnent
to
 Applicationof B & K Equipment

to

STRAINMEASUREMENTS

by

John Vaughan

October 1975

tsBN a7 87355 08 6
 CONTENTS

Foreword

L i st of S y m bols Use d .........

Ghapter 1 lntroduction

Chapter 2 Stressand Strain 11

Chapter 3 T he S tra i nG a u g e 17
Wire and foil strain gauges 18
S em ic o n d u c tosrtra i ng a u g e s(s e ea l so C hapter10) ... 20
Specialgaugetypes 22
Stressgauges 24
Gaugesfor other applications 25
T he s tra i ng a u g ea s a tra n s d u c e r 26
B ac k in gm a te ri a l s 28

Chapter 4 S t r ainG a u g eAd h e s i v e sa n d S e a l i n gMethods 30

Adhesives 30
A " M o d e l " me th o dfo r c e me n ti n gth e gauge 31
Surveyof adhesivesused for mounting
s t r aing a u g e s . ............, 32
Sealingmethods 34

Chapter 5 S elec ti o no f th e c o rre c ts tra i n g a u g e ........ 36

S t r aing a u g es u p p l i e rs 38

Chapter 6 T he M e a s u ri n gC i rc u i t 39
T he qu a rte rb ri d g e ...... .. 41
T he ha l f b ri d g e 45
T he f u l l b ri d g e ........ 48

Chapter 7 The PracticalMeasuringSystem 52

Ch a pt er 8 S our c e so f E rro ri n Stra i n Ga u g eM e asurements....... 61

Electricalsourcesof error 61
S uppl yl e a d s .......... 62
B r idgen o n -l i n e a ri ty 66
Us e of th e b a l a n c ec i rc u i t .. 67
Error due to input impedanceof the
measuringinstrument 69
 T em p e ra tu rev a ri a ti o n s 71
G aug em o u n ti n ga n d c o n n e c ti o n............ 72

Chapter 9 Measurementof Strain with Other

B & K In s tru m e n ts........... 74
Dev ia ti o nBri d g eT y p e 1 5 2 1 ......... .. 74
Voltmeter Tvpe 2425 78
Voltmeter Type 2425 and Conditioning Amplifier
Tvpe 2626 80
V ibr a ti o nM e te r T y p e 2 5 1 1 ......... 82
M eas u ri n gAmp l i fi e rsa n d An a | y 2 e rs........................ 85

Chapt er 10 S emi c o n d u c toStra r i n Ga u g e s ...... ......... 88

Accuratedeterminationof strain with
semiconductorgauges 92
Accurate measurementwith the
S t r ai n In d i c a to r1 5 2 6 94
1 . Determinationof the resistanceof the
m o u n te dg a u g e ........ 95
2. M e a s u re me not f th e i n d i c a te ds trai n ........ 97
3. Determinationof the actualstrain level 98
4. Determinationof apparentstrain due to
t em p e ra tu rev a ri a ti o n ..... 1O1

Bibliogr aphy ... 1O4

Appendix 1 Derivationof expressionsfor the positionof the

principalaxes referredto a set of arbitrarily
positionedaxes.(lncludesexpressionsvalid for
delt a , re c ta n g u l a r,a n d T -d e l taro s ettes) ........ 105
T he d e l ta ro s e tte ................108
T he r e c ta n g u l a ro
r s e tte ...... 112
T heT-d e l ta ro s e tte ............. 113

Appendix 2 Practicalexamplesshowing the use of the equations

developedin Appendix 1, and the constructionand
ex pla n a ti o no f Mo h r' s s tra i nc i rc l e ......... 1 1 5
Calc u l a ti o nme th o d ......,..... 115
G r ap h i c a me
l th o du s i n g Mo h r' s s trai n ci rcl e .............119
 F OR E WOR D

The object of this publicationis to give the prospectivestrain gauge

user a simple explanationof the basic theory and practiceof measure-
me n t us ing s t r ain g a u g e s .l t w i l l s h o w h o w v ari ous factors,such as
typ e of gauge,gau g e ma te ri a l ,me a s u ri n gc o n d i t i onsand type of i nstru-
me n t at ionc an inf lu e n c eth e re s u l tso b ta i n e d

Elementarystress and strain theory will be outlined and the working

p ri n c ipleof t he s t r a i ng a u g ed e s c ri b e dVa . ri o u sty pesof gaugeand mea-
su r ing ar r angem e n w t i l l b e d i s c u s s e da, n d s o m e practi calti ps gi ven for
g a u ge handlings o th a t th e u s e r c a n g e t th e b e s t from a measuri ngar-
rangement.Adhesivesfor strain gauge mountingwill be discussed.Al-
ternative applicationsfor strain gauges, as stress, force, pressure,or
torque transducerswill also be mentioned.

The B r iiel and K j e r p ro g ra mo f s tra i n me a s u rementi nstrumentati on

wi l l be int r oduc e d ,a n d g u i d a n c eg i v e n to h e l p i n the choi ceof gauge
a n d s y s t emf or pa rti c u l a ra p p l i c a ti o n sA s a n a ddi ti onalai d to the pro-
spective user, information is given on various sources of gauges and
adhesives
 The following letters and symbols have been used throughout this
book:

a resistanceratio
A Amperes
C degreesCentigrade
d differentialsign
E excitationvoltageand YoungsModulus
F degreesFahrenheit
g accelerationdue to gravity
G Modulus of Rigidity
I current
k gaugefactor
I length
N error factor due to input resistanceof measuringinstrument
P error factor due to lead resistance
r relativechange in resistance
R resistance
t temperature
T error factor due to bridgenon-linearity
V voltage
x'l
I
y I perpendicularaxes
.J
a angle
B angle
d small increment
dR change in resistance
6l change in length
dk change in gauge factor
strain
0r temperatureexpansioncoefficient
u PoissonsRatio
pe m ic r os t r ain
T circle constant
o stress
r shear stress
O ohm s
 CHA P TE R 1
INTRODUCTION

In the early days, machineryand structureswere developedand built

on a trial and error basis, making extensiveuse of the "rule of thumb"
approachto the problems of design. This "rule" worked very success-
fully in areas of slow technologicaldevelopment,where principlesand
methods were unsophisticated.A new part would be designed using
very conservativesafety factors, constructed, possibly tested before be-
ing put into service,and then modifiedby "beefing-up" or lighteningas
experience indicated. With the advent of mass production and the de-
mand for high speed machineryand really efficient construction,these
older design methods were no longer technically or economicallyvi-
able. This was a particularproblem in the newly developingaircraft in-
dustry, to take just one example, where massive safety factors could
not be used, as every unit of weight had to be deployedto give the max-
imum advantage.Further, innovationin the new industrycame so rap-
idly that technical developmentssoon outpacedthe ability of the older
empirical methods to resolve the problems that arose. lt soon became
impossibleto perform all the calculationsnecessaryto design aircraft
p a r t s hav ing m ax i mu m s tre n g tha n d mi n i mu m w ei ght w i thout empl oy-
ing teams of mathematicians,incurring costly time penalties,or risking
technical obsolescence before the development was completed. What
was needed was a thorough knowledge of the actual load and stress
distribution, which in turn called for accurate determination of the
strains (c stresses)present, so that the calculation and design process
could be acceleratedwithout becomingunreliable.

This climate led to the evolutionof new methodsof calculationto be

used in stress analysis,and to the developmentof several new instru-
ments for determiningstrain experimentally,under actual service con-
ditions.Earliermethodsof strainmeasurementemployeddirect mechani-
cal measurementon the part, often supplementedby opticalamplifica-
tion to overcome the difficulty of detecting minuie changes in length
with the unaided eye. Later, electrical instrumentationwas developed
that used transducersemployingcapacitance,inductanceor piezoelec-
 =
Strain
f

L@d
<__

Fig.l -l . The definition of strain

tric effects These electricalmethods sufferedfrom many of the disad-

vantagesassociatedwith the mechanicalextensometers,such as com-
parativelylarge size and mass, which could influence the behaviourof
the item under test and create clamping difficulties.They also suffered
from their own problems, an over-sensitivityto vibration or tempera-
ture. complex detectioncircuitry, a lack of transducerrobustness,or a
price sufficientto precludetheir use in the large numbers necessaryfor
really effective stress determination on complex constructions.Nev-
ertheless,all these measuringmethodsare still in use, and where con-
ditions are suitable, they yield satisfactoryresults. More recently, the
electrical resistance"strain gauge" has come into widespreaduse re-
p l a c ingt he ear lierm e th o d si n m o s t a p p l i c a ti o n s .

The strain gauge was developedin the late thirties by two research-
e rs i n t he US A . W o rk i n g i n d e p e n d e n tloy f e a c h o t her,S i mmonsat C A L-
TEC and Ruge at MIT developeda strain gauge consistingof a length of
wire glued to the test object so that changes in length (strainslon the
surface are transferredto the wire. These length changescause altera-

Gauge length

Conductor Brcking material Cement Test pirce ,tu321

Fig 1.2. The electrical resistance strain gauge

tions in the resistanceof the wire which can be measuredby compara-
ti vely s im pleelec tri c a lc i rc u i try .T h e m o d e rn s trai n gauge,exampl esof
whic h ar e s hown i n F i g .1 .2 a n d F i g .1 3 , w o r ks i n exactl ythe same
way, with strain being detectedby measuringthe resistancevariations
caus ed by c hang e s i n th e g a u g e l e n g th o f th e w i re. The strai n gauge
can be v er y s m al l a n d c o mp a c th a v i n g n e g l i g i bl emass to exert a mi ni -
mum of influence on the measuringobject, and be easily mounted on
th e t es t s pec im en ,u s u a l l yb y c e me n ti n g .

The electricaldetection circuits required to measure the very small

changes in the gauge resistanceare comparativelyuncomplicated,be-
ing variations of the familiar Wheatstonebridge. When suitable com-
pensationcircuits are employedor self-compensating gaugesused, tem-
p er at ur es ens it iv i tyi s v i rtu a l l ye l i mi n a te d .T h e resi stancestrai n gauge
allows a very economicallypriced measuringsystem to be made,where
the actual cost per gauge is often so low as to be of virtually no conse-
q uenc e. G auge c o s ts a re n o l o n g e r a h i n d e ra nceto the use of strai n
g auges ,c em ent e di n th e i r h u n d re d so n a s tru c t ure,to sol veany parti cu-
lar stressanalysistask by actual multiple measurements,insteadof the
laborious calculation procedures,based on extrapolationfrom a few
measurements,that had been used previously.

E-

Fig-l .3. Examplesof typical strain gauges

Typicalwell known applicationsfor strain gauges include experimen-

tal strain and stress measurementon aircraft, boats, cars and other
forms of transportation.Strain gauges are also used for the measure-
ment of stress in larger structures, for example apartment buildings
and office blocks,pressurizedcontainers,bridges,dams, etc. The strain
gauge is an important laboratoryimplement used for pure research,as
a design tool in the developmentstages of many machinesand struc-
tu res , and as a t ea c h i n ga i d to d e m o n s tra teb a s i cengi neeri ngconcepts
in educationalestablishments.

Less well known applicationsof strain gauges include their use in

transducers,where some physicalproperty,about which informationis
required, will be arranged to deflect a strain gauged member, the
amount of deflectionbeing relatedto the propertyto be measured.Typi-
cal examples of transducer applicationsfor strain gauges are dynam-
ometer rings for force or load measurement,pressuretransducersthat
use a strain gaugeddiaphragm,and displacementmeasurementwith a
thin strain-gauged"feeler".

10
 CHA P TE R2
S TRE S SA ND S TRA I N

The m ax im um b e n e fi t fro m s tra i n g a u g e me asurementscan onl y be

o bt ainedwhen a co rre c tl ya s s e mb l e dme a s u ri ngsystemi s al l i edw i th a
thorough knowledgeof the factors governingthe strength and elasticity
. his k n o w l e d g ea l l o w s th e s tra i n gaugesto be depl oyedi n
o f m at er ials T
th e m os t ef f ec t iv ema n n e r. s o th a t re l i a b l em e asurementscan be ob-
ta i nedt hat lend t h e m s e l v e sto c l e a r u n a mb i g u o usi nterpretati on.

During the design and constructionof machines and structures,the

strengt hoft hem ate ri a lto b e u s e d p l a y sa v e ry i mportantpart i n the cal -
cu lat ions T he s t r e n g tho f th e m a te ri a li s u s e d t o fi nd w hetherthe parts
can carry the loads demandedof them without excessivedeformation
o r f ailur e T hes e l o a d c a rry i n g a b i l i ti e sa re n o rmal l ycharacteri sedi n
te r m s of S T RE S S,th a t i s th e a m o u n t o f l o a d carri edby a gi ven area.
Thereforestress is quoted in pressure units, force per area (for example
Pa = N/ m 2, k p/ m m 2 , l b t/i n 2 l . S i mi l a rl y ,o th e r i mportant cri teri afor a
mat er ial,s uc h as th e l i m i t o f p ro p o rti o n a l i tyth, e ul ti matestrength,and
th e br eak ings t r e n g th ,a re a l s o u s u a l l y g i v e n i n terms of stress.H ow -
ever stress itself cannot be measured,it must be deducedfrom the me-
ch anic aldim ens i o n sa n d th e a p p l i e dl o a d . F o r exampl e,a know n ten-
si on load c an be a p p l i e dto a te s t s p e c i m e n ,a nd the stresscal cul ated
fro m t he applied l o a d a n d th e c ro s s .s e c ti o n The speci menw i l l not
change colour as with elevatedtemperatures,or affect a magneticnee-
d l e as wit h t he pa s s a g eo f a n e l e c tri cc u rre n t The onl y physi calsi gn of
l o adingis t he m e c h a n i c adl e fo rma ti o nd u e to th e l oad,and i t i s thi s de-
fo rm at iont hat c an b e me a s u re db y th e s tra i n g auge.H ookeshow edthe
re lat ions hipbet we e n l o a d a n d e x te n s i o nto b e a l i nearfuncti on,there-
fo re if t he def lec t i o ni s m e a s u re da n d th i s re l a ti onshi pi s know n, the ap-
p l i ed loadc an be c a l c u l a te d .

In t he s am e wa y th a t l o a d sa re c h a ra c te ri sed i n terms of stress,ex-

te ns ion is c har ac te ri s e di n te rm s o f S T R A IN w , here strai ni s usual l yde-
fi n ed as t he c han g e i n l e n g thp e r u n i t l e n g th(m m,/mm,i n,/i n),a di men-
si onles s r at io A d i re c t re l a ti o n s h i pk n o w n a s Y oung' s Modul us (E ),
si m ilar t o Hook e' sL a w . e x i s tsb e tw e e n s tre s sa nd strai n and i s show n

11
 Ultimate and
Br@king strain

74437

Fig 2 | . Stress/Strain curve for a typical Metal

i n Fi g. 2. 1 A s s t r ai n i s d i m e n s i o n l e s sE, h a s s tre ss di mensi ons,force

p e r ar ea.

Ta k inga gener alp u rp o s ec o n s tru c ti o ns te e l a s an exampl e,the yi el d

point and limit of proportionalitylie at a stress level of approximately
2 6 kp / m m 2 ( 37O OOl b f/i n 2 l , a n d th e u l ti m a t e stress i s 45 to
50 kp,/mm2(65 000 to 72 00O lbt/in2l. Using a value for Young'sModu-
fu s o f 21 x 103k p/ m m2(3 0 x 1 0 6 l b f/i n 2 l ,th e y i el d poi nt strai n can be
found from.

s tra i n : =
2 1 x' 61 C
: 1240ut

It i s wor t h r em em b e ri n gth i s v a l u e a s i t i s ty p i calfor a constructi on

type steel.

From t he f or egoin g ,i t c a n b e s e e n th a t i f th e Y o ung' sModul usof the

ma ter ial is k nown, a n d th e s tra i n i s m e a s u re d ,i t shoul d be a si mpl e
ma tter t o c alc ulat eth e s tre s s .T h i s i s tru e w h e n the di recti onof the
stress is known, which is the case when a simple strut is exposedto
te n sion. Howev er , i n a l l o th e r c a s e s th e re i s a further compl i cati on,
w h i c h is ex am inedb e l o w

Fig.2.2showsthe exaggerated deformationof a strut under simple ten-

si o n loadingwhic h c a u s e s a n i n c re a s ei n l e n g th and a correspondi ng
decreasein cross section This is known as the PoissonEffect, and it
me a ns t hat if s t r ain i s me a s u re di n e i th e r o f th e pl anesperpendi cul ar

12
 I
I

Fi9.2.2. Deformation of a strut

to th e a pp lied loa d, a negat iv e s t r ain wit h les s er m a g n i t u d e w i l l b e d e -

tect ed . Th e ma gn itude of t his les s er s t r ain depend s u p o n P o i s s o n 's R a -
t io(l), a ma teria l's co ns t ant t hat v ar ies f r om m at er i a l t o m a t e r i a l . l t i s u s -
ually ab ou t 0,3 , so t hat t he s t r ain m eas ur ed per p e n d i c u l a r t o l o a d i n g
will be ap pro xima tely - O , 3 t im es t he s t r ain m ea s u r e d p a r a l l e l t o t h e
loading. These strains have no associated stress as there is no load ap-
plied in th eir dire ctio n.

S imp ly state d, the r elat ions hip

_ Sfress
Strain

is only valid in th e d ir ec t ion of t he applied load, an d n o t i n a n y o t h e r d i -

rect ron.

Fi g. 2 3 f ur t her ill u s tra te sth e re a s o n b y p l o tting the magni tudesof

stre s s and s t r ain, t o g e th e rw i th th e i r d i re c ti o n sfo r a si ngl epoi nt i n the
si mp lyloadeds t r ain fi e l d

Fig.2 3. Stress and strain about a point in a simply loaded system

13
 Str es s lies in t he s a m e p l a n e a s th e l o a d i n g ,and so does the maxi -
mu m s t r ain, but t he mi n i mu m s tra i n s(e " ,e rl l i e i n pl anesperpendi cu-
l a r to t he loadingp l a n e .T h e s ep l a n e sp a ra l l e la nd perpendi cul arto the
direction of loading are the PRINCIPALPLANES,and the stressesthat
a ct upon t hem ar e k n o w n a s th e P R IN C IP AL S T RE S S E Sl .t shoul d al so
be noted that there are no shearingstresseson principalplanes.

It will now be appreciatedthat it is not possiblejust to multiply the

me a s ur eds t r ain by th e M o d u l u sE to o b ta i nth e m axi mumstress.A ddi -
ti o nal dat a on t he d i re c ti o no f l o a d i n g(l o c a ti o nof the pri nci palpl anes)
wi l l be nec es s ar ye, s p e c i a l l yw h e re l o a d sa re a p pl i edi n more than one
p l a ne

Th e pr oblemc an b e s i m p l i fi e db y d i v i d i n gth e i n formati oni nto vari ous

componentsthat can be resolvedinto the three perpendicularaxes.The
strain in each axis can be consideredto be composedof the primary
strain due to the loading,and the strain producedby the PoissonEffect.
Fo r t he ex am ples h o w n i n F i g 2 .2 , th e c o mp o nentscan be w ri tten as
follows:

6x ox 6y Fox dz uo
--_ ax:- : _' 1: e r : _ Q.1l
xEyE E

S imilar e qu atio ns c an be wr it t en f or loads a p p l i e d i n t h e y a n d z

axes. lf no w the lo ads ar e applied s im ult aneous ly i n t h e t h r e e a x e s , t h e
stra in in an y d irect ion c an be f ound by adding t he c o m p o n e n t s a l g e b r a i -
cally:

":+-+-+
'
-
lY
-l1o' -Uo' 12:21
EEE
oz
,- : - llo | - !o '
EEE

C onvertin g th ese equat ions int o a f or m th a t yi el ds stress, i l l ustrates

the p oin t tha t the st r es s in any plane is a function of the stress in the
other planes:

o, : Eq1' trtoyl lro,

oy : Eer*' po r-f lto, (2:3)
or : Et 7* 1t oy 1 lt o,

14
so that even when the locationof the principalplanes is known, meas-
u re ment sm us t be ma d e i n e a c h o f th e a x e s .

The pr ac t ic allim i ta ti o n so f s tra i n g a u g e m e a surementprovi de the

me ans of s im plif y i n gth e e x p re s s i o n si n e q u a t i ons 2:3, as normal l y
strain gaugesare cementedto the surfaceof the part being tested (one
exception is when gauges are used for measurementson concrete).
Stress at a surface cannot act perpendicularto the surfaceplane so ef-
fe ct iv elyt he gauge i s me a s u ri n ga tw o d i me n si onalstrai n system.A l l
the terms for one of the axes (for examplez) can be eliminatedso that
the equationsfor principalstresscan be rewritten as follows:
F
o , : zu 2 ( tr * p tr )
t-F
(2:41
E
oy : -, ( Ey+ lt x)
t -tt -2

Often the directions of the principal stresses are not known in the
practicalmeasurementsituation, so it is advantaoeousto have expres-
si o ns f or pr inc ipa l s tra i n s th a t re fe r to a rb i trari l yposi ti onedaxes as
sh o wn in equat ion2 :5

o: txcos2a * t, sin2a* Brysina cosa (2:5)

Th e equat ionde fi n e s th e s tra i n e a a l a p o i n t, w here a i s the angl e

th a t t he s t r ain m a k e sw i th th e X a x i s i n a rb i tra r yX -Y axes. The strai n
d c an be m eas u re db y a s tra i n g a u g e , a n d w h en measurementsare
made at three different angles to obtain three values lor eo and a,
th ree equat ionsc a n b e s o l v e ds i m u l ta n e o u s l yto gi ve x (the strai n i n
d i re c t ionX ) , y ( t h e s tra i n i n d i re c ti o nY ), a n d Fry (the sheari ngstrai n
p re sentat t he poin t).F u rth e r,w h e n th e v a l u e so btai nedare substi tuted
i n to equat ion2: 6,

tan2ao: l3'v (2:6)

Ex_ ty

t he an gle ao tha t the pr inc ipal planes m ak e wit h t h e a r b i t r a r y a x e s c a n

be d ete rmine d. Henc e, when ao is s ubs t it ut ed b a c k i n t o e q u a t i o n 2 : 5
t he prin cip al strain s and s hear ing s t r ain at t he m e a s u r i n g p o i n t c a n b e
f ound an d sub stitu t ed int o equat ion 2: 4 t o giv e t h e s t r e s s e s a c t i n g a t
t he p oin t.

The derivation of equations 2:5 and 2:6 is shown in Appendix 1 . Ap-

pendix 2 contains worked examples to illustrate how the principal

15
stresses and strains can be calculated from actual gauge measure-
ments, and also demonstrates how Mohr's Strain Circle can also be
use d to find the se v alues .

The ca lcu latio n pr oc es s es det ailed in Append i x 2 a r e s u f f i c i e n t l y t e d i -

ous, e sp ecially wher e m any m eas ur ing point s h a v e t o b e e v a l u a t e d , f o r
various researchers to have devised ways to try and simplify them One
commo n meth od us es c om put er s f or high s pee d d a t a r e d u c t i o n ( s o m e -
tim e s in rea l-tim e) , t his is par t ic ular ly ef f ec t iv e o n m u l t i - c h a n n e l m e a -
su ring syste ms. W her e f ewer m eas ur em ent s h a v e t o b e e v a l u a t e d , o r
w he n time is o f les s er im por t anc e, s pec ial nom o g r a p h s h a v e b e e n e m -
plo ye d with co nsider able s uc c es s .

Nomographs have been devised for several groupings of strain

gau ge s, an d for v ar ious t es t s am ple m at er ials . C o m m o n t y p e s o f n o m o -
gr a ph a re availa ble f or us e wit h s pec ial gr ou p s o f s t r a i n g a u g e s t h a t
have be en arra nged at pr edet er m ined angles to e a c h o t h e r , a l l o n t h e
same backing. The strain gauge groups are known as "Rosette
Ga ug es", an d the y will be m or e f ully des c r ibed i n C h a p t e r 3 . T h e n o m o -
gra ph s can yield t he pos it ion of t he pr inc ipal a x e s , m a g n i t u d e s o f t h e
m aximum a nd m inim um nor m al s t r ains , m agn i t u d e o f t h e s h e a r s t r a i n ,
and with so me t y pes , t he s t r es s es c an als o b e d e t e r m i n e d f r o m t h e
nomog rap h. Man y t y pes of nom ogr aph ar e base d u p o n t h e g e o m e t r y o f
t he Mo hr Stra in Cir c le, but as t he s t eps t owar d s a s o l u t i o n d i f f e r f r o m
nomog rap h to nom ogr aph, t hey will not be d e s c r i b e d f u r t h e r i n t h i s
book.

Howe ve r, with t he elem ent ar y t heor y of s t r e s s a n d s t r a i n d e s c r i b e d

in this ch ap ter, and wit h t he us e of M ohr ' s S t r a i n C i r c l e w h i c h i s d e -
ta jled in App en di x 2, t he engineer is able t o ge t a l l t h e n e c e s s a r y s t r e s s
information from the measurements made by the most commonly used
stra in g au ge a r r angem ent s The c alc ulat ion s o r c o n s t r u c t i o n s a r e
stra igh t forwa rd, and t he only addit ional inf or m a t i o n r e q u i r e d i s Y o u n g s
mod ulu s for th e tes t m at er ial, and it s Pois s ons r a t i o

16
 CHA P TE R3
rHE S TRA IN GA UGE

A s me ntio ne d in t he int r oduc t ion, t he elec tr i c a l r e s i s t a n c e s t r a r n

gauge consists of a conductor cemented to the test object. The conduc-
tor has a very small cross section so that the adhesive cementing it
to the spe cime n is s t r ong enough t o hold it s ec u r e l y . T h i s a l l o w s t h e
strains to be transmitted from the test object directly to the conductor
without relative slip between test object and conductor, or buckling of
the co nd uctor u nd er c om pr es s ion. The s m all c h a n g e s i n t h e g a u g e
length o f th e co nd uc t or t hat ar e c aus ed by a load a p p l i e d t o t h e t e s t o b -
ject in du ce sma ll changes in t he r es is t anc e of t h e c o n d u c t o r ( a n e f f e c t
first described by Lord Kelvin), and these changes in gauge resistance
are d ete cte d b y th e m eas ur ing ins t r um ent at ion. Th e c h a n g e i n g a u g e r e -
sista nce is re late d to t he c hange in gauge lengt h ( s t r a i n ) b y t h e G a u g e
F acto r k:

, 6 R ,6 1 6R
(3:11
R l tR
where

R = g a u g ere s i s ta n c e
dR = c h a n g ei n g a u g ere s i s tance
| = g a u g el e n g th
dl = c h a n g ei n g a u g el e n g th
= s tra i n

O ne of the majo r f ac t or s t hat af f ec t t he per f or m a n c e a n d u s e f u l n e s s

of any strain ga ug e is t he m at er ial f r om whic h t h e c o n d u c t o r i s m a d e
ldeally the con du ctor s hould hav e a high G auge F a c t o r , s o t h a t s m a l l
strain s give a s larg e c hanges as pos s ible t o t he r e s i s t a n c e T h e s p e c i f i c
resistan ce o f th e mat er ial s hould als o be high, t o g i v e t h e b i g g e s t p o s s i -
ble c ha ng es in resis t anc e when s t r ained i. e. be t t e r r e s o l u t i o n T h e s e
qualitie s make th e s t r ain gauge s ens it iv e t o s m all s t r a i n s .

F urth er, multip les of a giv en load ( s t r ain) m us t g i v e t h e s a m e m u l t i p l e

of th e resista nce ch ange. i- e. t he G auge Fac t or m u s t b e l i n e a r , a s i m -

17
p l i e d by equat ion3 :1 , a n d n o t v a ry w i th th e d e greeof l oadi ng.S i mi -
larly, the Gauge Factor should not vary with time, then repeatedappli-
cationsof a given load will always give the same resistancechange'

The conductor material must also be as insensitiveto temperature

changes as possibleto prevent variationsin temperaturefrom causing
apparentstrains that could be as large as the actual strain due to me-
chanical loading. lt should be noted that with modern gauge materials.
it is often possibleto use the temperaturesensitivityof the conductor
material to compensatefor the temperaturesensitivid of the material
in the test object. Temperature compensation methods will be de-
scribedlater.

Normally, the choice of a strain gauge conductor material is taken

out of the hands of the engineer who is interested in making strain
measurements.Gauges are items to be purchasedready-made,and it
is unusual (but not unknown) for them to be fabricatedon the iob. The
gauge manufacturerhas done the work of materialsselectionand treat-
ment, and suppliesa range of strain gaugeswith guaranteedcharacter-
istics to suit most common measuring situations. The test engineer
merely has to select the gauge characteristicsthat most accurately
match the particularmeasurementrequirement.

Wire and Foil Strain Gauges

Earliertypesof straingaugewere madeof thin copper-nickel or chrome-

nickef alloy wire approximatelyO,O25mm (O,OO1in) in diameter. To
achieve the longest practicalgauge length, which gives bigger resist-
a n ce c hanges ,and a t th e s a me ti me th e m i n i m um occupi edarea so
that the measurementscan approximatepoint strain determination,the
conductoris normally folded into a grid pattern similar to that shown in
Fi g .3. 1. lt c an be s e e n th a t th e g ri d l a y o u t s ti l l has a l ong gauge
length, but the transversesensitivity has not been increasedby any
great amount, however, it still exists and should be taken into account.

Fig.3.l. Typical wire strain gauge

18
 Fi9.3.2. Typicalfoil strain gauge

Considerablesavings in the weight of backing material and adhesive

a re a ls o m adewit h th i s a rra n g e me n t.

More recently,the many advantagesof the foil gauge have led to its
wide-spreaduse, but it has not completelyreplacedthe wire gauge. In
the foil gauge the conductoris made by etching a grid pattern in a thin
metal foil only a few micrometresin thickness,made of a similar alloy
to the wire, or by cutting the grid from foil using accurate dies. These
processesallow accurateand cheap productionof'almost any conceiv-
able grid pattern, includingcomplexshapes;or very small gauges.A ty-
p i ca l f oil gaugegr id i s i l l u s tra te di n F i g .3 .2 .T h e l argetabs at each turn
of the conductor path make the foil gauge very insensitiveto strains
acrossthe grid due to the comparativelylow resistance in the tabs.
Another important feature of the tabs, is that their large surface en-
sures that linear conditionsexist over the completeactive length of the
grid. End effects are minimised, and the creep problem is also greatly
reduced in this way too. End effects are illustratedin the greatly exag-
g e ra t eddr awing s ho w n i n F i 9 .3 .3 , n o te th a t th e actualgauge l ength of
the grid is unaffected.

Foil gauges also have a greater ratio of surface area to cross sec-
tional area than wire gauges,which gives them greatly enhancedheat

Fi9.3.3. Diagram illustrating the end effects

19
d i ss ipat ionqualit i e s th a t p e rm i t h i g h e r v o l ta g e l evel s to be used for
gauge excitation.These have a greater effect on the measuringcircuit,
and hence much improved resolving power and accuracy. The in-
creased surface area also gives a larger contact area for cementing
o n to t he t es t obje c t,th u s m i n i m i s i n gth e p ro b l e mof creep.

Ty pic alv alues fo r s tra i n g a u g e s i n s e ri e s p roducti onthat are avai l -

a b le " of f t he s helf " fro m s u p p l i e rsa re s i m i l a rfo r w i re and foi l gauges:

Gauge Factor: approximately2. This is usually quoted individually

to two decimal placeswith the gauge or pack of gaugesand often a
toferance,for example! 1o/o,will also be given.
Res is t anc e: s ta n d a rd i z e dv a l u e s , 1 2 0 4 , 350O, 600O, and
1OOOO. A resistance tolerance is often quoted, for example
! O.25o/o.
Linearity: measurementsare accuratewithin O,lVo up to 4OOOpe,
and wit hin 1% u p to l O OO0 /e .
Breakingstrain: 2OOO0to 25OOOpe
Fatigue life: up to 107 strain reversals.
Temperature compensation: normally gauges are available with
automatic compensationthat matches the temperatureexpansion
coefficient9, of one of the three most commonly used construction
metals:
6
G ener alpur po s es te e l sw i th 0 t:1 1 x 1 0 -6p er" C (6,1 x 10 per" F)
Stain less stee ls wit h 0 t:1 7 x 1 0 -6p er " C (9,5x 10-6per' F)
Alumin ium with 0 r2 3 x 1 0 6 p er" C (12,8x 70-6per" F)

Some manufacturers also supply gauges compensatedfor use on tita-

nium, magnesium alloys, or plastic materials. When these are cem-
ented to the materialfor which the resistance/temperature characteris-
tic has been matched with the expansion coefficient, the apparent
strain due to temperature variations on normal gauges can be held
d o wn t o f es s t han !1 ,5 p e /o C (!O ,8 p e /o F l o v er a temperaturerange
from -2Oo to + l5OoC (-50 to + 3OOoF).

Semiconductor Strain Gauges

In the last decade,semiconductorstrain gauges have come into use

with Gauge Factors5O or 60 times greaterthan for wire or foil gauges.
Semiconductorgauges consist of a strip conductor made from a single
cry s t al of s ilic on o r g e rm a n i u m th a t c o n ta i n san accuratel yadj usted
amount of impurityto give the characteristicdesired.

20
 semiconductorgauges are more sensitive to temperaturevariations,
a n d g e ner allynot s o ru g g e da s w i re o r fo i l g a u g e s.They are moresui t-
able for dynamic measurementsbut can be used for short term mea-
surements of static strain levels. They require special handling to get
the best out of them.

Typicalvaluesfor productionsemiconductorstrain gaugesare:

Gauge Factor: 10O (individuallycalibrated).

Resistance:120 O.
Linearity:measurementsare accuratewithin 1% up to lOOO/e.
Breakingstrain: approximately5OOOge.
Fatigue life: 1Oo strain reversals.

716Q

F ig. 3. 4. Typical semiconductor gauge

At the present time, there is only a very limited selectionof semicon-

d u cto r s t r ain gauges a v a i l a b l e T h i s to g e th e r w i th thei r hi gher pri ce
serve to restrict them to more specialized applicationswhere the high
Gauge Factor is an advantage,for examplein the measurementof very
low strain levels The least strain that can be measuredwith semicon-
d u cto r gaugesis of t h e o rd e ro f 0 ,O 0 1 /e , w h i l e metal gaugescan mea-
su re only O , 1pe S ome o f th e s p e c i a lte c h n i q u e srequi redfor use w i th
se mi c onduc t or gauge sw i l l b e d i s c u s s e di n C h a p te r10.

However, it is not really possibleto think in terms of one type of gen-

eral purpose gauge, because many different factors influencethe type
selectedfor any particular measurementapplication.Outsidethe above
me n ti oned" nor m al" ra n g e s ,o th e r fa c to rsl i k e b a c k i ngmateri al s,or the
a d h e siv esem ploy ed l i m i t th e u s e o f th e g a u g e , parti cul arl yw i th ex-
tended temperature ranges Therefore special gauge types have been
developedwith conductorcharacteristics.backing materials,and adhe-
sives tailor-madefor more specializedapplications

21
Special Gauge Types

So f ar in t his c h a p te r,th e o n l y s tra i n g a u g e st hat have been consi d_

e re d hav e been s in g l ee l e me n tg a u g e sfo r m e a s uri ngl i nearstrai n.H ow -
e ve r , when dif f er e n tg ri d c o n fi g u ra ti o n so r m u l ti -gri darrangementsare
used, improved measurementsor additional information can be ob_
tained with a minimum of extra effort

In chapter 2 it was shown that when the directiongof the principal

axes are not known, three gaugescan be used to give the necessaryin-
formation. However, three gauges give almost three times as much la-
b o u r in pr epar at io na n d m o u n ti n g ,a s s u m i n gth a t there i s enough room
fo r them , and a de a l o f e x tra c a l c u l a ti o n ,e s p e ci al l yw hen the gauges
have not been mountedat an "easy" angle from each other.

To get around these complicationsas mentioned earlier. strain ,,ro_

settes" are often used where two, three, or four measuringgrids are
mounted on the same backingso that they can all be cementedat the
same time on to the test specimen in one easy operation.Strain ro_
settes have standard, accuratelydetermined 45o, 60", or 9Oo angles
between the different grids to help sirnplifythe calculation.charts or
calculatorsare often availablefor use with standard rosette types, so
that strain and the direction of the principalplanes can either be read

Fig.3.5a Delta rosette Fig.3.5b Stacked grid 9Oo rosette

Fig.3.5c. Two element gOo grid Fig. 3. 5d. Herringbone grid

22
d i re ct ly , or f ound w i th a v e ry m i n i m u m o f s i mpl e ari thmeti c.Fi g.3
shows some of the more common strain rosettearrangements.

The Delta rosette in Fig.3 5a is a very common type used to deter-

mi n e t hedir ec t ionsof th e p ri n c i p a la x e s ,a n d th e p ri nci palstrai nsacti ng
o n th e m A r os et t ef o r s i mi l a ra p p l i c a ti o n si s s h o w n i n Fi g.3.5b,but i n
this case the gauge grids have been stacked in a sandwich arrange-
ment to save space, and to give a closer approximationto a point strain
measurement when this type of rosette is used with higher excitation
voltages,heat dissipationcan be a problembecauseof the close proxim-
ity of the grids to each other.

Th e t wo elem ent g a u g e s h o w n i n F i g .3 .5 c h a s i ts gri ds arrangedat

9 Oo to eac h ot her . T h i s a rra n g e m e n ct a n b e u s e d t o augmentthe mea-
su re d c hange in r es i s ta n c e i,. e . g i v e a l a rg e rG a uge Factor,w hen the
two grids are connected into adjacent arms of a measuringbridge in
th e w ay s hown in F i 9 .6 .4 . Ga u g e F a c to rsa p p ro x i matel y1,3 ti mes the
n o rmalv aluec an be o b ta i n e dw i th th i s a rra n g e me nt.

Fi g .3. 5ds hows a t w o g a u g e ro s e ttew i th th e g ri d s arrangedi n a her-

ri n g b o nepat t er n.T hi s ty p e o f g a u g e ,a n d s i mi l a ra rrangementsi ncorpo-
rating four grids at 90o to each other are frequentlyemployedfor mea-
su ri n g t or s ionals t r ai n so n a x l e sa n d s h a fts .

Another type of sandwich construction is that used in the "Flexa-

g a g e " m anuf ac t ur edb y th e Bu d d C o mp a n y .T h i s g auge i s empl oyedto
measure bending strains on plates and panels where one surface is in-
accessible.The two measuringgrids are mounted, one on each surface
of a thick backingpiece with an accuratelyknown thicknessof material
between them. When the assembly is cementedto the test panel one
gauge is in contact with the panel surface and measuresthe surface
stra i n. F r om t he m ea s u re me n ot b ta i n e db y th e o ther gri d, i t i s possi bl e

Msred Strain 1
MeasuredStain 2
Lin@r St.ain - BendingStrain

)
Proicted Strain on Lin@r Strain
InnerSurface
7@11

Fi9.3.6. Principle of the Flexagage

23
to deducethe position of the neutral axis, the amount of bending,and
th e linears t r ainpre s e n t.F i g .3 .6 i l l u s tra te sth e ori nci pl e.

A simple geometrygives the positionof the neutral axis, while the dif-
ference between the strains projected onto the two surfaces of the
panel is a measure of the linear strain present. The algebraicaverage
of the projectedsurface strains gives the bending strain at the panel
surface.

The gaugesillus tra te di n F i g .3 .7 a re u s e d to measui -etangenti al ,ra-

d i al. or c om bineds tra i n so n th i n m e m b ra n e sa n d di aphragms.Gauges
o f t hes et y pesar e a v a i l a b l ew i th d i a me te rsl e s st han 5 mm (O,2i n).

Fig.3.7a. Gauge for measuring Fig.3 7b. Gauge for measuring

tangential strain radial strain

7fl$l

Fig 3.7c Gauge for measuring radial and/or tangential strain

Fig 3 7. Diaphragm strain gauges

Stress Gauges

The stress gauges shown in Fig.3 8 can be calibratedto give read-

ings directly in stress. They make use of the relationshipdemonstrated
in Equation2:4 that the stress in the measureddirection is dependent
u p o n t he s t r ain in t h a t d i re c ti o n ,a n d a l s o u p o n th e transversestrai n.

24
 3. 8 a

3.8b
Fig.3.8. Typical stressgauge configurationc

T he gauges how n i n F i g .3 .8 ah a s tw o g ri d s w i th conductorl engthsi n

the ratio of 1 '.p (Poissonsratio).The longer grid must be mountedwith
filaments parallel to the direction of the required stress measurement'
When foaded,the two grids in series give the er + Uev term in the
equation, so that the change in gauge resistance is proportionalto
stress. The other gauge in the figure uses a slightly different arrange-
ment to achievethe same effect. When the central axis of the gauge is
priented in the direction of the desiredmeasurement,the proportionof
the gauge conductor in the direction of measurement and at right
a n glest o it ar e in t h e ra ti o 1 :g , s o a g a i nth e e x + U ey term i s gi ven.

Gauges for Other Applications

Parametersother than stress or strain can be measuredusing strain

gauge measuring instrumentationand suitable gauges. The examples
sh own in F ig. 3. 9 a re fa i rl y ty p i c a lo f n o n -s tra i nmeasuri nggauges.The

7ffi

3. 9a. Crack propagation 3. 9b. Crackpropagation 3.9c. Temperature

gauge gauge sensor

Fig.3.9. Gauges for non-strain measurement applications

25
crack propagationgauges change in resistanceas the growing crack
progressivelybreaks the gauge conductors.Alternativelywhen a suit-
able gauge material is used that has a high sensitivityto temperature
and very low sensitivityto strain, the resultant gauge makes an excel-
lent temperaturesensor Typical sensors of the type illustratedhave a
useful temperaturerangefrom -8Oo to + 3OOoC(-1O0o to + 6OOoF).

The Strain Gauge as a Transducer

Resistancestrain gauges are also used as the active element in

several transducerapplications,and the following short section briefly
sketches some of the more typical examples.The significanceof the
bridge connectiondiagramswill be explainedmore fully in subsequent
ch a pt er son " T he M e a s u ri n gC i rc u i t" a n d " T h e P r acti calMeasuri ngS ys-
te m " .
Load

l.

--\r7-

Fig.3.lO. Arrangement and connection of gauges in a load transducer

The Load Transducershown here makes use of four strain gauges

cementedon the inside of the cylinder (for protection).The gaugesare
connectedtogether to make use of Poissonsratio to increasethe effec-
tive gauge factor, so that the bridge sensitivitywill be approximately
2,6 times the sensitivityof an individualgauge.(seealso Fig.6.7!.

Under the action of a tensile load, the curvature of the Ring in

Fi g . 3. 11 is " f lat t e n e d " s o th a t th e i n n e r g a u ges experi encetensi on
while the outer gauges are in compression.The bridge sensitivitywith
this arrangementwill be four times that of a single gauge, (see also
Fi s.6. 8) .

26
 Fig.3.l l. Ring Dynamometer for measuring tensile loads

Th e P r es s ur eT r a n s d u c e rs h o w n i n F i 9 .3 .1 2 has four strai n gauges

mo unt ed on a t hin c i rc u l a rd i a p h ra g mth a t b e n ds under the acti on of
pressure.The gauges at the edge of the diaphragm in a comparatively
unstressedarea give automatictemperaturecompensationas described
in Chapter6. while the gaugesat the centre of the diaphragmare bent
to give an output that is directlyproportionalto pressure.

Fig.3.l2. Diaphragm transducer for measuring pressure

The output from the strain gauges on the feeler arm shown in
Fig.3.13 will be directly proportionalto bending' which is proportional
to displacement.The two gaugesgive automaticcompensationtor tem-
perature changes at the gauges.They will have a combinedsensitivity
that is two times that of a single gauge.

27
 1 7fiA3

Fig.3.l3. Displacement Transducer for measuring liquid level

Backing Materials

Str ain gaugebac k i n gm a te ri a l sh a v e tw o m a i n f uncti ons,they protect

th e gauge gr id f r om d a ma g ed u ri n g h a n d l i n ga n d mounti ng,and they
tra ns m itt he s t r ainfro m th e te s t o b j e c tto th e g a u g econductor.

The bac k ingm us t h a v e c o m p a ra ti v e l yl o w s ti ff nessso that i t can fol -

l o w s t r ain c hange s i n th e te s t s p e c i me nw i th o u t i tsel f affecti ngthem
(b y dam pingf or ex a mp l e ).O n th e o th e r h a n d , i t m ust be sti ff enoughto
supportthe gauge conductorand not deform to give a distortedor irreg-
ular change in length. Simply stated, the backing must be stiff com-
p a redwit h t he c on d u c to r,b u t fl e x i b l ec o m p a re dw i th the test materi al .
As will be s een, t h i s i s n o t to o d i ffi c u l t a re q ui rementto meet, and
severaldifferenttypes of backingmaterialare in common use

Pa per bac k ingsa re o fte n e mp l o y e dw i th w i re strai n gauges.P aper

h a s t he t win adv an ta g e so f b e i n g re a d i l yo b ta i n a bl eand easi l yw orked,
wh i le it s s t if f nes si s a d e q u a teto s u p p o rt th e g auge duri ng handl i ng.
Wh e n pr oper lyapp l i e dw i th a th i c k l a y e r o f n i tro-cel l ul ose cementthat
co m plet ely penet r a te sa l l th e p o re s i n th e p a p er, the backi ngi s qui te
strong enough to transmit strains from the measuring object to the
g a uge c onduc t or .Th i s c o mb i n a ti o no f c e m e n t a nd paper typi cal l yw i l l
only suffer a breakdownof adhesionat room temperaturewith strains
g re a t ert han lO O O O Op e ,w h i c h i s w e l l a b o v eth e l i mi t of most conduc-
to r m at er ials .T his b a c k i n gh a s a u s e fu l te mp e raturerange up to TOoC
(1 sO oF )f or s t at icstra i n m e a s u re me n tsH. i g h e rtemperaturescan be tol -
erated for short periods- up to a few hours, or for measuringdynamic
stra rn

28
 Ep ox yplas t ic sar e a l s o u s e d fre q u e n tl ya s b a c ki ngmateri al sfor many
typ e s of s t r ain gau g e , p a rti c u l a rl yfo r fo i l g a u g e s.E poxypl asti ccan op-
e ra t e s uc c es s f ullya t s l i g h tl y h i g h e r te m p e ra tu resthan paper, and i t
has a strain limit in excessof most conductormaterials.

Ap plic at ionsf or h i g h e r te mp e ra tu re s ,u p to a bout 2OOoC(4OOoFl ,

ca n be s at is f iedby u s i n g p h e n o l i cre s i n (b a k e l i te ),
or temperaturecured
e p ox y plas t icas a b a c k i n gma te ri a l Ba k e l i teb a ckedstrai n gaugesare
a l so par t ic ular lyr e c o m m e n d e dfo r th e i r l o n g te rm stabi l i tyand resi st-
a n c e t o c r eep T h e h i g h e r te m p e ra tu rec a p a b il i ti esmade possi bl eby
these backingmaterialsmean that a carefulchoiceof cement will be re-
q u i r ed. T he pr oper ti e so f v a ri o u sc e m e n tsa n d a dhesi vesare descri bed
in the next chapter

Stainlesssteel is also used as a carrierfor gaugesintendedfor opera-

tion at very high temperatures,up to 4OO"C (7sO"F). The conductor
g ri d is bondedt o t h e me ta l b a c k i n gd u ri n g m a n u facture,so that the di f-
ficuft cfampingand heat curing processesare carriedout under accu-
ra tely c ont r olledc o n d i ti o n s L a te r th e w h o l e g a uge assembl yi s si mpl y
mo unt edon t he t est s p e c i me nb y s p o t w e l d i n g .

For applic at ionsa b o v e 4 O 0 o C , h i g h te m p e ra tureunbackedor stri p-

pable gaugesare employed.These can be used typicallyup to tempera-
tu res of 9O O oC( 16 5 0 o F ) T h e u n b a c k e dc o n d u ctori s cementedto the
te st s pec im enus in g a h i g h te mp e ra tu rec e ra m ic cement. l t shoul d be
n o tedt hat s pec iali z e d
i n s ta l l a ti o na n d c u ri n gte c h ni quesare requi red.

29
 CHA P TE R4
STRA IN GA UGE A DHE S IV E SA ND
S E A LING ME THODS

Adhesives

A s with th e b ack ing m at er ial, t he m ajor r equ i r e m e n t f o r a s t r a i n

gauge a dh esive is s uf f ic ient elas t ic it y t o f ollo w t h e d i m e n s i o n a l
changes in the te st objec t wit hout s ac r if ic ing ac c ur a c y i n t h e t r a n s f e r o f
st rain to th e ga ug e c onduc t or . W hen s elec t ing a s u i t a b l e a d h e s i v e f o r
cementin g the strain gauge t o t he t es t objec t , s ev e r a l i m p o r t a n t f a c t o r s
have to be considered, as not all adhesives are usable for all gauges or
for all measurement conditions. some adhesives give a chemical bond,
others make a mechanical connection, and still others a combination of
the two.

P ro ba bly the most im por t ant f ac t or t o be c ons ide r e d w i l l b e c o m p a t i -

bilit y with th e ga ug e bac k ing m at er ial and wit h t he t e s t o b j e c t m a t e r i a l
T he ce men t must no t dam age eit her m at er ial, bu t m a k e a s o u n d a n d
last ing bo nd with th em . lt s hould hav e good long te r m s t a b i l i t y a n d n o t
decom p ose o r fa ll a par t af t er a s hor t t im e. Fur t he r , t h e c e m e n t s h o u l d
not be subject to appreciable creep during the lifetime of the test Re-
sist ance to de co mpo s it ion and c r eep m us t be m a i n t a i n e d t h r o u g h o u t
the a n ticip ate d te mper at ur e r ange of t he t es t , and b e m a i n t a i n e d u p t o
the maximum strain level expected Some tests demand stability up to
very hig h temp era tur es and at high s t r ain lev els , a n d t h e s e r e q u i r e -
ments must be satisfied bY the cement

T he ce men t sho ul d be a good elec t r ic al ins ulat o r e l i m i n a t i n g t h e p o s -

sibility of cu rren t le ak age t hr ough t he t es t s pec im e n t h a t c o u l d g i v e r i s e
t o er ron eo us ba lan c e c ondit ions . This is par t ic ul a r l y i m p o r t a n t w h e r e
st rippe d strain ga ug es ar e em ploy ed, es pec ially if it i s b o r n e i n m i n d t h a t
the resistance differences being measured can be of the order of thou-
sandths of an ohm After cementing the gauge in place, it is important
t hat the insula tion is c hec k ed us ing a good qualit y o h m m e t e r

S om e ad he sives need s pec ial c ar e and t r eat m e n t d u r i n g s t o r a g e ,

30
strain ga ug e mo un ting, and c ur ing. Som e c em ents h a v e t o b e s t o r e d a t
low temperatures to yield the best storage life and avoid wastage.
Others need special curing processes, high temperatures for periods of
several hours which the test object may not be able to tolerate. Ce-
ments b ased o n re f r ac t or y ox ides r equir e s pec ial e q u i p m e n t t o h a n d l e
the molten spray Yet other cements present something of a health ha-
zard as the y rele ase t ox ic f um es dur ing c ur e, or r e q u i r e s p e c i a l p r e c a u -
tions during handling to prevent harmful contact with the skin that
causes in flama tion .

R esistan ce to h um idit y is anot her v er y im por t a n t q u a l i t y f o r a s t r a i n

gauge ceme nt to po s s es s Som e adhes iv es and c a r r i e r m a t e r i a l s a r e f a r
more re sp on sive to c hanges in at m os pher ic h u m i d i t y , a n d e x h i b i t
chang es in insula tion r es is t anc e t hat c an giv e r is e t o z e r o d r i f t a n d d i f f i -
culties in ba lan cin g. Longer t er m , or out door s t r a i n g a u g e i n s t a l l a t i o n s
will req uire some fo r m of m ois t ur e pr ot ec t ion.

T he ch art sh ows t he m ajor t y pes of adhes iv e u s e d f o r m o u n t i n g

strain gauges, and lists some representative exdmples of manufactur-
ers, tog eth er with the m or e im por t ant c har ac t er is t i c s t h a t i n f l u e n c e s e -
lection of the most suitable cement tor any particular measurement as-
srgnme nt.

A "Model" method for cementing the gauge

Whe n ce men ting t he gauges in plac e, c leanline s s i s o f t h e u t m o s t t m -

port an ce Th e surfa c e of t he t es t s pec im en m us t b e s m o o t h e n o u g h f o r
t he ce men t typ e selec t ed t o adher e pr oper ly , and s h o u l d a l s o b e f r e e
f rom rust, scale , p aint , et c , t hat c ould int er f er e w i t h t h e q u a t i t y o f t h e
bond . Whe re n eces s ar y t he s ur f ac e c an be m ac h i n e d t o g i v e a s u i t a b l e
f inish , o r g lass pa per and em er y c lot h c an be us e d . S o m e t i m e s i t m a y
even b e n ecessa ry t o r oughen a highly polis hed su r f a c e t o e n s u r e a s a -
tisfactory joint.

The prepared surface must be washed down thoroughly to remove

metal or dirt particles and grease. Trichlorethylene, methyl-ethyl-ke_
tone, or carbon tetrachloride are suitable cleaning fluids for this pur-
pose Ve ntila tion p r ec aut ions s hould be obs er v ed . S e v e r a l w a s h e s w i l l
be necessa ry to achiev e t he nec es s ar y degr ee of c l e a n l i n e s s , a n d i t r s a
good p olicy to ke ep was hing wit h c lean paper t ow e l s o r l i n t f r e e c l o t h s
unt il th e wipin g cloth no longer pic k s up dir t

Q1
 ? E
E E!
t
E
E
ad
e i:
;i !
E
;E:E
iu 3
z
aEe
: 15
!:i
Eie

I g
;5s
E
6
@
o
E
9F
E6

*F 2
I
R
E E! 3i
Ei: iee
;; e 8E- :i:g
5
i3
s I sEi
3ii
t ;:
3-
iEB
E 6 l-
.EE
EiEi
: -ri I6t 88
1T
I
8
8
cEc :q T
g!
o!

1E! ir"i
-T
3b

E :S EEE

5iE Es;83
a
hb
iT ? + g
Eb 8
3
<p
-1 E:
o

o
c
A Eb
i!
IE
i
33
6A
I
E
.t 3 FEg'
Eb !
o
I t EA E r3
c
6 =al t ::i g
13g
E 3
o
i tI 4i I F
;- 6 : E
E
i-

i'i:
eR EA
E
g . E
8 3i 3
E
E
Eg 8 !3
o uj
l:

*i-r' E
9s EE: 3
E:6
!
;; 6
iR
!g J:
I
s;i
<!!
i !E
:c;
t

o o ol

E
o E" EE,:
ao 9i E"l ,
;Eigi
J bB
o
o
g:5za AA 6

I
|o
o
o
o
t '-a
i:

o iii:
o t &
E 5" 99
q6 i6

ss:
o
o E
F9
c r 35
!!
3e E
E:

32
 Cementingshould take place immediatelyafter the test surface has
been cleaned,to prevent the collectionof dust particles,or the forma-
tion of an oxide layer. The backing of the gauge should also be de-
greased immediately prior to cementing. Mark the position that the
gauge is to occupy so that it can be accurately located while cement-
ing, or lay the gauge in place and tape the supply leads so that it can
be raised up for cementingwithout losing its location.lt is often a good
idea to tape a small piece of polythenefoil so that it can cover and pro-
tect the gauge,but can also be hinged up.

When an adhesive is being used that requires mixing, it should be

ready mixed and pre-cu,ring- following the makers instructions- so
that it can be applied immediatelythe gauge and surfaceare ready.Ap-
ply a thin layer of cement to the back of the gauge, or to the test speci-
me n, and lay t he g a u g e i n p l a c e e n s u ri n gth a t it l i es i n the correctposi -
tion and orientation.Press it down gently with a thumb to ensure that
there are no air bubblestrapped between the specimenand the back-
i n g . A pply pr es s u rea n d te mp e ra tu red u ri n g the curi ng peri od as re-
quired for the particular adhesive selected.lf terininal tabs are to be
cem ent edin plac e ,th i s c a n a l s o b e d o n e b e fo recuri ng.

After the cement has cured, the strain gauge leads can be soldered
to the terminals. Whenever possible the gauge resistanceshould be
checkedfor changes caused by damage during mounting' lf it is found
to be markedlydifferent from the rated value, a new gauge will have to
be cemented in its place. The gauge insulation from the test piece
should also be checkedusing a megohmmeterthat employsa low exci-
ta t ion v olt age.Us e o f a h i g h v o l ta g e c a n d a magethe gauge w i re and
possibly cause overheatingand breakdownof the adhesiveor backing'
50 to 100 megohms of insulation is consideredto be adequateto en-
sure stableoperationof the gauge.

The widely used method of running a finger tip over the mounted
gauge to check whether it is firmly bondedis a useful test, providedthe
results are treated with caution. Any movement of the gauge, caused
by faulty bonding should show up as a violent change in the indicated
strain level on the measuringinstrument.

lf these tests of the gauge give a satisfactoryresult, the whole gauge

arrangementcan be given a water-proofingtreatment if this is neces-
sary for the particulartest environment,or test programduration.

33
Se alingM et hods

In t he s hor t t er m , i t i s p e rmi s s i b l eto u s e u n protectedgauges under

co ndit ionsof c lean l i n e s sa n d v e ry l o w h u m i d i ty .Indoorappl i cati onsi n
the test or researchlaboratorycan employ unprotectedgaugesfor perr-
o d s up t o a f ew d a y s w i th o u t e x p e ri e n c i n gi n s ul ati onchangesor zero
drift due to the penetrationof moisture.However,any strain gauge in-
stallation that is intended for use over extendedperiodsof time must
have some form of protectionto excludemoisture.Outdoor applications
will require particularlystringent measuresto counterdbtthe effects of
th e weat her . V ar i o u s m e th o d s h a v e b e e n s u c cessful l yused, varyrng
fro m c om plet eenc a p s u l a ti o ni n e p o x yo r p h e n o l i ccompounds,to mul ti -
l a ye r r ubber and me ta l s a n d w i c h w i th a l i b e ral use of seal i ng com-
p o u nds

Fig. 4. 1, s hows o n e ty p i c a l me th o d re c o m m e ndedby several strai n

g a u ge s upplier swh e re b y th e w h o l e g a u g e , a n d a short secti onof the
supply leads are completely covered by a mastic compound. Most
g a u ge m anuf ac t u re rsa re a b l e to s u p p l ya s e a l i n gki t for use w i th thei r
gaugesto protectthem in this way

Hard protrctive cove.ing

(Thiokol or nitrile rubber) Tight sal around @ble

(Synthetic varnish)
Mastic $aling compound
(Bitumastic or mineral waxl

Fig.4.l. Typicalstrain gauge protection for longer term applications

The ar r angem en ts h o w n i n F i g .4 .2 i s re c o m mendedfor real l y l ong

te rm s t r ain gauge i n s ta l l a ti o n as n d g i v e se x c e l l e ntprotecti onagai nstse-
vere environments.The hard, metal or plastic cover gives protection
a g ains t m ec hanica ld a ma g e a n d a l s o a c ts a s a fi rst stage barri er
against the ingress of moisture lf a metal cover is used it should be
made from as thin a sheet as possible to minimise its effect on the
strain measured by the gauge. A small amount of silica gel ensures
that any air that is trappedwill be as dry as possible.

34
 Ha.d cover
{mtal or polv@rbonate} Sealed in @ble tub

Mastic s6ling compound 'ws

Fi g. 4, 2. s t r ainga u g e p ro te c ti o n fo rl o n g te rmi n sta| l ati onsi nsevereen-

vironments

35
 CHA P TE R5
SELE CTIONOF THE COR RE CT
S TRA !N GA UGE

After discussingthe merits and otherwise of strain g"rg"r, backing

materials and adhesivesin the precedingchapters, the following sec-
tion is a quick, common sense guide to the selection of the correct
g a ugef or any par ti c u l a ra p p l i c a ti o n .

As the individual measuringproblem will have a great influence on

the actual gauge type and configuration chosen, the following points
should be considered:

1 . Straight forward linear strain distributionswhere the directions of

the principal planes are known can be measured by the classic
singleelem entg a u g e .

2. Where some augmentationof the Gauge Factor is desirable,a dou-

b le gr id gauge,s u c h a s th a t s h o w n i n F i g .3 .5c may be empl oyed,or
two separategauges.

3. lf the strain distributionis more complex,one of the various types of

strain rosette may be more useful or more convenient.

4. The size of the gauge must be suitable for the test object. lf high
strain gradientsare expecteduse the smallestpossiblegauge, When
it is necessaryto employ higher voltages for gauge excitation, a
larger gauge grid may be required to dissipate the extra heat.
5OOmW per square centimetre(5W per in2) is a fair approximation
for the maximum amount of,power that a typical foil gauge can han-
dle when mounted on a metal test piece. This is the equivalentof
1V on a 12O O g a u g e th a t o c c u p i e sO ,1 7 cm2 (O,O25i n2|of area,
but it should be remembered that a wire gauge may be able to take
only a quarter of this heatingvalue.

5. The gauge, backing material,and cement chosen must be able to op-

erate at the temperaturesand strain levelsanticipated.Where these
levels are not known suitable safety factors should be applied.

36
6. The gauge,backing material,and cement should have a long enough
fatigue life to survive the test program without premature break-
downs.

7. The backing material and cement must be chemically compatible

with the test object material. and not produce unwanted side-
effects.

8. The temperatureexpansioncoefficientof the gauge should suit the

material to which it is going to be attached. This item is rather less
important when dummy gauges or full bridge connectionsare used
for temperaturecompensationas describedin Chapter6.

9. Finally as this is a common sense guide, use the least expensive

gaugethat satisfiesthe conditionslistedabove.

Strain gaugesand strain gauge transducerscan be obtainedfrom the

sourceslisted in the followingTable.

37
Strain Gauge Suppliers

Ge n e r a l 5emr- S pec|al
Trans-
Supplier Purpose conductor Purpose
ducers
Ga u g e s Ga u q e s Gauges
Concrete,
Ail t e c h .U S A H i qh tem p
Akers Electronics, Norwav
H i gh ex tens i on
Ba ld w i n L i ma Ha mi l ro n, USA x X x
H r qh l em p
W i l l i a mT B e a n I n c , U SA High extension

B e l l & H o w e l l .U S A x

B u d d C o m p a n yU
, SA H i gh tem p

Dentronics Inc . USA x H r gh tem p

E I h E T L td , UK x x
F r isch e n E l e ctro n i c, x
G e rma n F e d Re p
F ri t zH e l l i g e& C o , x High iemp
G e r m a nF e d B e D
G au g e T e ch n i q u e , UK x x
H o tta n g e r B a l d wi n Me Btechnik, H i gh ex tens i on
x x x
German Fed Reo H i qh tem o
Hit e cC o r p , U S A X H i gh tem p
K u l i te S e mi co n d u cto r Pr oducts x x
USA
Kyowa. Japan x
Herbert Lembcke, Sweden X

M a g n a fl u x, U S A x H i gh ex tens i on

M i cro Ga u g e l n c , U S A x
H i gh ex tens r on
M i cro Me a su re me n ts, USA x x
H i gh tem p
Pe e ke l . H o l l a n d High extension
Ph i l i p s, Ho l l a n d x x
Proud. Czechoslovakia x

Sh o w a , Ja p a n x

C onc r ete
S t r a i n s t a l lU
,K H i oh tem p
S t a t h a mI n s l r u m e n t sIn c . USA x
Technograph, UK x x
Transducers{CEL)Ltd.. UK x x
Vi b ro Me te r Co ro . S witzer land x x H i gh temp
H i gh ex tens i on
Vi sh a y In te rte ch n o l o g y, USA x x
H i qh tem p
High extension
W e l w vn E l e ctri c. U K x x
H i qh tem o,
Whittake(,USA x x

38
 CHA P TE R6
T HE ME A S URING CIRCUI T

As me ntio ne d in Chapt er 1, t he ex t r em ely s m a l l c h a n g e s o f t h e o r d e r

of th ou sa nd ths of an ohm , t hat oc c ur in t he ga u g e r e s i t a n c e ( t y p i c a l l y
120 O to 2O0 OO) due t o v ar iat ions in t he appl i e d s t r a i n c a n b e m e a -
sure d by th e Wh ea t s t one br idge. The br idge is s h o w n i n F i g . 6 1 , w h e r e
t he re sistan ce sh own in eac h of t he f our br idge a r m s c o u l d r e p r e s e n t a
strain gauge. The measuring indicator can be a galvanometer. or the
signal can be led to the input terminals of a measuring amplifier - as
in the B & K Strain Indicator Type 1526.

Fig.6.l The Wheatstone Bridge

The n umb erin g c onv ent ion us ed her e is t he s a m e a s t h a t a d o p t e d f o r

the Typ e 1 52 6 Inst r uc t ion Book

The symbols used in this section are as follows:

B = re sisranc e ( Q )
dR = cha ng e in r es is t anc e ( O )
i = curre nt ( am p)
V = voltage (volt)
E = excitation voltage (volt)

F irst co nside r the ideal c as e wher e t he r es is t an c e i n e a c h a r m i s i d e n t i -

cal, i. e. the bridge is "balanced" so there is no potential difference

39
across the indicator.Hence the voltagedrop across R ro must be equal
to the voltagedrop acrossR23,and a definiterelationshipmust exist be-
tween the resistancesand their positionsin the bridgearms.

As Vzq: izq.,Rzq: Vzs: izi; Rzs

and Vtt: izar R'rq: V1.;': izs.,Rrs
then iro, Rro: izy Rzg (6:1)
and i"o, Rro: izs, R.ts ( 6:2)
dividing 6:1 by 6 :2
iro.,Rro _ iril Rn
i"o, Rro ir", R.,,
Rro - Rrt
(6:3)
R' o Rt s

T his is a ve ry imp or t ant r elat ions hip whic h indic a t e s t h a t a n y c h a n g e

in t he re sistan ce s o n one s ide of t he br idge c an be b a l a n c e d b y a d j u s t -
ment s to th e resista nc e v alues on t he ot her s ide o f t h e b r i d g e A s w i l l
be shown late r, th is is t he bas is f or one of t he m e t h o d s u s e d f o r t e m -
perature comp en sa tion Fur t her , it indic at es t he po s s i b i l i t y o f i n c r e a s e -
ing t he e ffective g au ge f ac t or when s dv er al gaug e s a r e u s e d i n t h e
bridge arms

S olv ing for o ne b ridge r es is t or ( f or ex am ple R2a ) ,

^r^:+?
E quation 6:4 in dicat es t hat when t he v alue of R 1 a i s k n o w n a c c u -
(6:4)

rately, an d th e ra tio of t he t wo r es is t anc es on t h e o t h e r s i d e o f t h e

bridge is a lso kn own ac c ur at ely , t he v alue of t he u n k n o w n R 2 4 c a n b e
det ermin ed with th e s am e degr ee of ac c ur ac y

T aking the examp le a s t age f ur t her , if t he unk n o w n r e s i s t a n c e R 2 4

was a stra in g au ge on a t es t piec e, t he br idge c oul d b e b a l a n c e d a c c u -
rately before loading, the loading could be applied to the specimen, and
t hen the brid ge co ul d be balanc ed again by adjus t i n g t h e r a t i o o f t h e
t wo re sistors Rn /R1 3 lf t he c hange in t his r at io i s k n o w n , t h e n t h e
difference in strain gauge resistance between the two loading condi-
t ions can be d ete rm ined. This dif f er enc e will be d u e t o t h e s t r a i n
caused b y th e ap plie d load, and t he s t r ain c an be c a l c u l a t e d f r o m ,

strain :
#:+"1
(3:1)

40
 6. 2a 6 .2 b 6.2c

Fig 6 2 Methods of balancing the Wheatstone Bridge

T h e ratio o f th e two r es is t or s c an be c hanged b y a n a c c u r a t e l y d e t e r -

mined amo un t whe n one or m or e of t he br idge a r m s c o n t a i n s a c a l i -
brat ed po ten tiome ter Fig. 6 2 s hows s om e of t he t y p i c a l a r r a n g e m e n t s
t hat a re u se d in co m m er c ial ins t r um ent s A v ar i a t i o n o f t h e a r r a n g e -
ment sho wn in Fig 6 2c is us ed in t he B & K Ty pe 1 5 2 6 t o g i v e a c o n v e -
nient meth od of b al anc ing t he br idge.

A s sho wn in eq ua t ion 3: 1 abov e, t he r es is t anc e c h a n g e i s d i r e c t l y p r o -

port ion al to th e strain, s o t he v ar iable r es is t or c on t r . o lc a n e a s i l y b e c a l i -
brate d to re ad stra in, t her eby elim inat ing s om e te d i o u s a r i t h m a t i c A n
even more convenient arrangement is achieved when the indicator it-
self is calib rate d d ir ec t lv in s t r ain v alues . lf t he b r i d g e i s b a l a n c e d a t
the sta rt o f e ach tes t , t he s t r ain lev el c an be r ead f r o m t h e m e t e r w i t h -
out the ne ed fo r re- balanc ing t he br idge f or eac h l o a d c o n d i t i o n T h i s a r -
rangeme nt works v er y well wit h s m all s t r ain lev e l s , b u t a t h i g h e r l e v e l s
as the bridge goes further out of balance the galvanometer begins to
load th e brid ge in troduc ing dis t or t ion

Wh en the in dicat ing ins t r um ent is a m eas ur ing a m p l i f i e r , t h e l o a d i n g

effe ct ca n b e keo t t o a m inim um bec aus e m od e r n a m p l i f i e r s c a n b e
built to work with very low input levels. lf a digital voltrneter is used as
t he d isp lay it will be jus t as eas y and ec onom ic a l t o o b t a i n f i n e r e s o l u -
tion in volta ge me as ur em ent as it would be in r e s i s t a n c e F u r t h e r m o r e ,
the gain can be adjusted to suit the different gauge factors encountered
w hen ga ug es o f var ious m at er ials ar e us ed, t hus p r e s e r v i n g t h e d i r e c t
readin g fe atu re

The O.uarter Bridge

Wh at e ffect will a r es is t anc e c hange due t o s t r a i n h a v e o n t h e v o l -

t age me asure d by t he indic at or ? Ref er r ing onc e a g a i n t o t h e b r i d g e
shown in Fig 6 1, let a c hange in r es is t anc e be i n t r o d u c e d i n t o o n e o f

41
the a rms to u nb ala nc e t he br idge, s o t hat R24 b e c o m e s R + d R , w h i l e
the o the r re sistan c es r em ain unc hanged ( = R) . Th i s i s e q u i v a l e n t t o t h e
situation wh ere o nly one ( ac t iv e) gauge is bein g s u b j e c t e d t o s t r a i n ,
w hile the oth er res is t anc es ( k nown as dum m ies ) c a n b e e i t h e r i n a c t i v e
gaug es, o r ju st re si s t or s , t his ar r angem ent is t he " O u a r t e r B r i d g e " .

ir^,:a^?n'r:ET#T"
E
2R+6R
V2 a : i2 a 1R2 a
tr
: (R+ 6R)
ZdTn
ER+ E6R
- (6:5)
2R+ 6R

whife irrr:-J--:i
F23+ R13 2R

V4: i2s 1R2s : #

E
( 6:6)
2

S o th e vo ltag e in t he m eas ur ing link ,

v : v 4 - v 2 a : z+ - + lzjR4+ 6R
- 2 ER + E6 R -2 ER -2 E6 R
4R + 26R
E6R
: +
4FJ?6F-

Fo r v er y s m all v a l u e so f d R ,

4R + 2 6 R+ 4 ft ,.
F AF],
S o th at V: + - y l' t (6:7)
4B
S ub stitu ting th e v alues f r om equat ion 3: 1 , and r e a r r a n g i n g ,

,: y (strain) (6:8)
Ek

42
 T his is a ve ry im por t ant r elat ions hip s howing t h a t t h e r e i s a d i r e c t
connection be twee n t he applied s t r ain and t he v o l t a g e i n t h e m e a s u r i n g
link. lf th e e xcitatio n v olt age and t he gauge f ac t o r a r e k n o w n , i t i s o n l y
necessary to measure the out-of-balance voltage to obtain the strain
level.

T he strain given by equat ion 6: 8 c ould hav e b e e n p r o d u c e d e q u a l l y

easily by mechanical loading, or by the effects of a temperature change
if the strain gauge had not been matched to the expansion coefficient
of the test specimen material lt could also have been caused by a com-
binatio n o f lo ad an d t em per at ur e c hanges , but th e r e i s n o s i m p l e w a y
of t elling wh ich For t unat ely t he W heat s t one br id g e p r i n c i p l e a l l o w s f o r
the insertion of a compensation resistor into one of the bridge arms.
T his resisto r h as a t em per at ur e c har ac t er is t ic c al c u l a t e d t o c o m p e n s a t e
f or the re sistan ce change in t he ac t iv e gauge due t o a t e m p e r a t u r e v a r -
ralron .

The simplest and most commonly employed method to obtain a comp-

ens atio n resisto r w it h t he des ir ed c har ac t er is ti c ; i s t o u s e a s t r a i n
gaug e with id en tic al s pec if ic at ion t o t he ac t iv e g a u g e T h i s c o m p e n s a -
t ion ga ug e is mo unt ed on an uns t r es s ed par t of t h e t e s t s p e c i m e n , o r
on a separate piece of the same material placed close to the test piece
so that it experiences the same temperature changes without the strain
chan ge s. Th e com pens at ion gauge m us t be c on n e c t e d i n t o o n e o f t h e
bridg e arms ad jacent t o t he ac t iv e gauge ar m , ( p o s i t i o n s R 1 4 o r R 2 3 i n
F ig.6.1 ) Man y me as ur ing ar r angem ent s us e c al i b r a t e d r e s i s t o r s i n p o -
sit ion s R23 an d R i3, s o t he c om pens at ion r es i s t o r i s m o s t c o n v e n i -
ently po sitio ne d at R14

Re ferrin g ag ain t o Fig 61, let R24 and R14 b o t h i n c r e a s e b y a n

amou nt dR1 d ue to a t em per at ur e c hange.

.EE
R24+ R14 2( R + 6 R )
V2a: i2o', R2a

E
:Z@it;nT (R+ dF1)
E
( 6:9)
2
and asV23 is u nchanged V: V4- V2a

EE
-z-T :0 (6:10)

43
 So the bridge is still balanced,and the temperatureeffect is compen-
sated. lf the active strain gauge had been subjectedto a strain and a
temperaturechange,while the compensationgauge was only subjected
to a temperaturechange,the temperatureeffectswould have been can-
celled out. Only the strain componentwould have appearedas a volt-
age in the measuringlink, which is illustratedas follows.

L et R24 inc r eas eb y a n a mo u n t d R t d u e to a temperaturechange

a n d an am ount dR d u e to a n a p p l i e dl o a d ,w h i l e R 14 i ncreasesby dR r
o n l y, due t o t he t em p e ra tu rec h a n g e
EE :
,241 :-=- -
n2a*R1a F+dFr+dR+R+dFl
E
2R+26R1+6R
V2a: l2 a 1 R2 a

:zF j zl aF =m(R +6 P " +6 R)

ER+E6R1+EdF
=-2R (6:11)
+ 6R
+ 28fr,
F
From6:6 Vu:7

And V: Vzg-Vzq

_ E _ER+E6fu+EdF
2 2R+2dF1+dF
_2ER + 2E6Rt+ E6R-2ER -2E6fu -2E6R
4R+46R,+26R
Ed,B
4F+46Ft+26R

For very small values of dR 1 and of dR

4R+46R,+26R+4R

so that n: t'::

The temperature components have cancelled each other out, and

only the strain component due to loading remains, i. e. this is
equa tion 6:7 ag ain .

44
The Half Bridge

U p unt il t his poin t, o n l y a b ri d g ew i th o n e a c ti vegaugehas been con-

si d er ed. Howev er e q u a ti o n 6 :3 i n d i c a te dth e p o ssi bi l i tythat i f more
g a u gesar e pos it io n e do n th e te s t s p e c i me ni n s u ch a w ay that thei r re-
sistanceschange under the action of an appliedload, and these gauges
are then connectedinto the measuringbridge, several advantagescan
be achieved.Small strain levels will be detected more easily because
the additional active gauges will give a greater out-of-balancesignal
when the bridge is loaded,temperatureinfluencescan be cancelledout
as shown in the precedingexample,and in many arrangementsgreater
me a s ur ingac c ur ac yw i l l a l s o b e o b ta i n e d T h e fol l ow i ngexampl esi l l us-
trate some of the possibilities.

Fi9.6.3. Test specimen in simple bending with Half Bridge connection

Fig. 6. 3 s hows a te s t s p e c i m e nw i th a p p l i e dl o adi ngthat puts one sur-

fa ce int o c om pr es s i o nw h i l e th e o th e r s u rfa c ei s i n a numeri cal l yequal
te n sion.W hen t he c o n n e c ti o n si n th e b ri d g ea re as show n i n the di a-
g ra m ( k nown as " H a l f Bri d g e " , R 2 a w i l l i n c re a seto R + dR under the
a cti on of t he bend i n gw h i l e R 1 4 w i l l d e c re a s eto R - dR i f i denti cal
stra ingaugesar e u s e d
E
1241 -
R2 4 + R1 4
E E
F+dR+F-dR 2R
V2 a : i2 a 1 R2 a : fi n*ant
-nER+E6R

45
 E
Uncha ng ed from 6:6 Vzg :
2
E EF+EdR
so v: 2 2R
E R -E R -E d F
2R
E6R
- +
-2_R (6:12)

When t his r es ult i s c o m p a re dw i th e q u a ti o n6 :7 i t w i l l be seen that

th e br idges ens it iv i tyh a s b e e n d o u b l e d .

Providedthe test specimenhas its neutral plane at the centre of the

test strip, the bending stresses can be calculated from the surface
stra insf ound by t hi s m e a s u ri n gm e th o d .

Fig.6.4.Stress distribution on a teststrip in ,,^o,Jt"ro,rn

lf in addition to the pure bending, the specimenhad also been sub-

jected to a tensile force, both gauges would also have been stretched
b y the s am e am ou n t, i n a d d i ti o nto th e c h a n g e scausedby the bendi ng.
Th e s t r es sdis t r ibu ti o nfo r th i s s i tu a ti o ni s s h o w n i n Fi 9.6.5.

As the two active strain gauges are connectedin adiacent arms of

th e m eas ur ingbr id g e ,th e e ffe c t o f th e te n s i o n l oad w i l l be cancel l ed
o u t, leav ingjus t t he b e n d i n gc o mp o n e n tsto b e measured.

7tu375

Fig.6 5 Stress distribution on a test strip in combined bending and ten-

slon

46
 Fur t her m or e,if s tra i n g a u g e sw i th o u t te m p e r aturecompensati on,
or
gauges with poorly matchedcompensationhad been used, an apparent
strain will be detectedby both gauges it the temperaturevaries.Again,
as the gauges have identicaltemperaturecharacteristics,and are con-
nected in adjacentarms of the bridge, the temperatureeffectswill can-
cel each other out.

The arrangementis automaticallymade independentof temperature

variations,the only conditionbeing that the two gaugesreally must be
experiencingthe same temperature,Note that with large specimens.or
w h e r e t em per at u re sv a ry g re a tl y ,th i s c o n d i ti o nmi ght not be met, and
self-compensating gaugesmust be used.

Fig. 6. 6 s hows a n o th e r te s t s p e c i m e nw i th a tensi l e l oad appl i ed.

When the gauge connectionsare as indicated,R24increasesto R + dR,
w h i le R1a dec r ea s e sto R - 0 ,3 d R , b e c a u s ethe test materi ali n thi s
ca s e has a P ois s o n sR a ti oo f O .3 .

.4
,.jor,.4

-o) $s
Fi9.6.6. Specimen in pure tension with Half Bridge connection
7@6

,E
lzat: p^1
4
E
R+dR+R-0,3dR
E
2R+ 0,76R
V2a: i2a1 R2a

F
=:= - : (R+6R)
2R+0,76R'," '
ER+EdR
-
2R+0,76R
F
A gain V"":7
tr ER+E6R
A nd V= i- -
2 2R+0.76R

47
 2ER+O,7E6R
-2ER -2E6R
4 R + 1 ,4 6 R
1 ,3 E 6 8
- 4 R + 1 ,4 6 R

A nd f or v er y s m a l l v a l u e so f 6 R

4R+ 1,46R
+ 4R
.. . 1.3E6R (6:13)
so that
4R
A gain, c om par i n gth e re s u l t w i th e q u a ti o n6 :7, the bri dgesensi ti vi ty
will be s een t o b e 1 ,3 ti m e s th e s e n s i ti v i tyfor a si ngl e acti ve gauge.
The m ain us e f o r th i s ty p e o f c o n n e c ti o n(o t her than the mi nor i n-
crease in sensitivity it affords) is to obtain automatic temperature
compensation.lt will be readily understoodthat if temperaturechanges
occur during the test, both gauges will experiencethe sarne alteration
of resistanceso that the temperatureeffectsare cancelledout as in the
previous example. Note that if the compensationgauge is mounted on
a strainedpart of the test sample,the bridgesensitivitywill be affected.

The Full Bridge

The foregoing examples have illustrated that it is possible to improve
the sensitivityobtainablefrom a single gauge. while maintainingtem-
perature compensation. by suitable connection of an extra loaded
gauge into the bridge.An even greaterincreasein sensitivitycan be ob-
tained when all four bridgearms containactivestrain gauges.In this ar-
rangement,called the "Full Bridge", balancemust be obtainedby some
additionalresistor in the bridge that can be adjusted,for example like
the arrangementsshown in Fig.6.2b or Fig.6.2c. When all of the active
gauges are subject to the same temperaturevariations,the apparent
strainsdue to temperaturecancel out as before.

I n t he f ollowin g e x a mp l e ,a te n s i l e te s t s p e ci menhas four i denti cal

active strain gauges mounted upon it with their connectionsas shown
i n F i9. 6. 7.lf t he P o i s s o n sR a ti oo f th e te s t ma teri ali s O,3 agai n,the fol -
lowing changes in gauge resistancewill occur under the action of the
a p pliedload,
R24 increasesto R + 6R
R13 inc r eas esto R + d R
R14 dec r eas e sto R - O,3 6 R
R23 dec r eas e sto R - O,3 6 R

48
 Fi9.6.7. Specimen in tension with Fuil Bridge connection

E
t241 -
R2 4 + R1 4

F+ dF+R -0,36R
E
:2R+0,?6R

V2a: i2a1 R2a

tr
:-=:---:- (R+6R)
2R+0,76R',
ER+E6R
-
2R+0,?6R

When t he f ull b ri d g e i s u s e d V 2 3 i s s ti l l th e product of i 231 and

R23, but t he r es u l t i s n o l o n g e rE/2 a s th i s s i de of the bri dgenow con-
ta ins ac t iv egauge s .
,E
trtr:
Rr"+Ru
E
R - 0,36R+F+ dF
E
2R+0,?6R
V123: i2y R2s

F
ZFnm 8-o'3bR)
ER - O,3E6R
-
2 R + 0 ,7 6 R
And V: V2s-V2a

49
 ER -O,3E6R ER + E 6R
2R+0,76R 2R+0,76R
ER-O,3E6R - ER- E6R
2R+0,76R
1,3bR
- -
2R+0,76R

Forverysmallvaluesof dR,

2R+0,76R
+ 2R
sothat y-a1'3QR
-2R (6;14)

lf th is re su lt is c om par ed wit h equat ion 6: 7 ag a i n , i t w i l l b e s e e n t h a t

the full bridge connection on a tensile (or compressive) specimen has
had the effect of making the bridge sensitivity 2,6 times the sensitivity
of a sing le a ctive g auge.

Fig.6 .8 sh ows a not her s pec im en in bending, w i t h a f u l l b r i d g e s t r a i n

gau ge a rran ge men t when t he gauges ar e c onne c t e d a s i l l u s t r a t e d , t h e
follo wing ch an ge s in r es is t anc e oc c ur under t he a c t i o n o f t h e l o a d i n g ,

R2 4 in cre ases to R + dR
R1 3 in cre ases to R + dR
R1 4 de cre ases t o R - dR
R23 decreases to R - dR

7@

Fi9.6.8. Specimen in Bending with Full Bridge connection

.E
I"or :
Rr o+ R_
E E
F+dF+R-dR 2R

50
An d v24-- i241 R24

tr
:;= (F+dR)
zn
ER+ E6R
-n
.EE
Si milar ly . ZJ I
R- dR+ R+ dR 2R

An d V2s: i1s2 R2s

E6R
- nER -

T h ere fore th e voltage in t he m eas ur ing link ,

V:V 2 s -V 2 a
E R -E 6 R ER+EdF
-n-n
, E6R (6:15)
R

I n th e be nd ing ca s e wit h a f ull ac t iv e br idge, it i s p o s s i b l e t o i n c r e a s e

the effective bridge sensitivity to four times the sensitivity of a guarter
bridge with o nly o ne ac t iv e gauge ( 6: 7) . Not e t ha t t h i s i s o n l y p o s s i b l e
with a bending condition where two gauges can be placed to measure
positive strain, and the other two placed to measure an equal negative
strarn.

The foregoing examples are not intended to be an exhaustive list of

all possib le meth od s of c onnec t ing m eas ur ing br i d g e s , o r o f b a l a n c i n g
them. However it is hoped that the basic principles established here
w ill e na ble th e follo wing c hapt er s on pr ac t ic al m e a s u r i n g s y s t e m s , w i t h
their limita tion s an d pos s ible s our c es of er r or t o b e u n d e r s t o o d . F u r -
t her, th at o the r pra c t ic al ar r angem ent s will s ugg e s t t h e m s e l v e s t o t h e
test engineer faced with the problem of strain measurement.

51
 CHAPTER7
THE PRA CTICA LME A S URING S Y S T E M

The B & K S t r ai n In d i c a to rT y p e1 5 2 6 (s h o w n in Fi g.7.1)i sa di rect i n-

dicating instrument with a digital displayof the measOredstrain level.
Severalfeatures help to make this instrumentparticularlysimple to set-
u p and us e. T he 1 5 2 6 h a s fa c i l i ti e sfo r m e a s uri ngw i th ful l , hal f, and
q u ar t er br idgec on n e c ti o n sa, n d i ts s e n s i ti v i tyrange i s from + 199,9/e
to t ' l 99909 with o n e a c ti v es tra i n g a u g e .l t c an empl oygaugesw i th
resistancesbetween 50Q and 2O0OO, and there is also provisionfor
varying the gauge factor to suit differentgauges.The Strain Indicatoris
able to measure both static and dynamic strain at frequenciesup to
3OOHz, and there is a hold function to capture transients. A 3 kHz
squarewavecarrier system is employed,with bridge voltagesadjustable
to 3V , 1V , and O ,3 V . U p to fi v e b ri d g e sc a n b e connectedto the i nstru-
ment at the same time. and then be selectedone at a time for measure-
ment .

,*:*,
r llla
a at
J*."L
j,
:i'-1}'
I
r- T " il',
Lil?e
r
Fig.7.l. The B & K Strain lndicatorType | 526

The relative significanceof the various features will now be dis-

cussedin more detail.

Digit al dis play sa re s i m p l e to u s e , c o mp l e te l yel i mi nati ngthe needl e

p a r allaxt hat c an g i v e b a l a n c i n go r re a d i n gi n a c curaci es.
The di spl ayon

52
th i s ins t r um entis p a rti c u l a rl yg o o d h a v i n g O ,O5%resol uti onand hi gh
a ccur ac y ,wit h all t h e i l l u mi n a te de l e m e n tsi n the same pl ane to gi ve
e a sy r eadingat wide v i e w i n ga n g l e s .T h e d i g i ta lsi gnal i s al so avai l abl e
in BCD form from a TTL compatibleoutput if it should be neededfor fur-
th e r d igit alpr oc es s i n g .

This instrument is also intendedto be used with B & K LevelRecord-

ers Type 2305 and Type 2307, therefore known DC levels can be
sh o wn on t he dis pl a yw h i l e a ma tc h i n gc a l i b ra tedD C output i s avai l -
able for settingup the Recorder.

Bu ilt - in c ir c uit r ya l l o w s th e u s e r to s e l e c tu s e w i th ful l or hal f bri dge

me as ur em entand , a re m i n d e ro n th e fro n t p a n el i ndi catesthe gauge
connectionsto be used.

Fi g. 7. 2s hows t he b ri d g ec o n n e c ti o n su s e d fo r f ul l and hal f bri dgear-

. will b e s e e n th a t w h e n h a l f b ri d gei s sel ected,a pai r of
ra n g em ent s lt
precision 1 kO resistors make up the dummy half of the bridge, and
th a t t he t wo ac t iv e g a u g e sa re c o n n e c te di n p o s i tl onsR 23 and R 13.l t
wi l l als o be s een t h a t th e b a l a n c ere s i s to ri s p l a c edi n paral l elw i th the
a cti v e gauges .T he o th e r re s i s to rsi n th e b a l a n cel i ne enabl e the bal -
ance resolutionto be varied. The values shown on the Mode Selector
swi tc h r ef er t o t he b a l a n c ere s o l u ti o no b ta i n a b l e.W i th 120O gauges
a n d a gaugef ac t or o f 2 ,0 O , p o s i ti o n" 2 o /o "a l l o w s 120O/6,O4kf) = 2Y o
to be balanced,equivalentto t lOOOO/c which gives a resolution of
2OOOueper turn of the potentiometer.A better resolutionand finer bal-
ance can be achieved(at the cost of a reduction in the total balancing
range available)when the Selector is switched to a lower percentage
va l u e.

W hen quar t er br i d g e c o n n e c ti o ni s re q u i re d ,t he S trai n Indi catori s

switchedto half bridge mode and a specialOuarter BridgeAdaptor con-

u",[l''JJ-oPZ"n

74@46

Fig.7.2. Bridge and balance connectionsused in the Type | 526

CJ
 Fig.7 3 The Auarter Bridge Adaptor ZR OOI4

tainin g a p recisio n r es is t or as a dum m y gauge is p l u g g e d i n t o t h e i n s t r u -

men t. Fig 7.3 is a phot ogr aph of t he Adapt or T h e p i n s o n t h e r e a r e n -
gage with th e te rm inals on t he 1526 while t he c o n n e c t i o n s a r e s h o w n
on the top surface. The photograph shows a strap joining two of the
t ermin als, bu t this int r oduc es a r es is t anc e er r o r i n t o t h e b r i d g e a n d i s
not re co mmen de d f or r eallv ac c ur at e m eas ur em e n t s

In the recomm ended ar r angem ent f or c onn e c t i n g t h e o n e a c t i v e

strain ga ug e to th e O uar t er Br idge Adapt or ( Fig . 7 . 4 ) t h e c o n d u c t o r s R L
shou ld be as sim ilar t o eac h ot her as pos s ib l e . T h i s c a n e a s i l y b e
achie ve d by using t he s am e t y pe of c onduc t or l e a d , a n d k e e p i n g t h e
lead le ng ths th e sam e. The t wo leads ar e c onn e c t e d a s s h o w n , a n d a
third lea d (R) co nn ec t s one of t he t er m inal t abs o n t h e s t r a i n g a u g e , v i a
the Ad ap tor, d irect ly t o Ter m inal 3 on t he 152 6 s o t h a t e a c h a r m i n
t he ha lf b ridg e cont ains 120O plus t he lead r e s i s t a n c e R 1 . T h i s a r -

Active Gauge
120 sl

L- .J
ZR 0014

Fig 7 4. Connections recommended for the Auarter Bridge Adaptor

54
ra n gem entens ur e sth a t a n y re s i s ta n c ec h a n g e in the l eadsdue to tem-
p e rat ur e v ar iat io n s w i l l b e e q u a l l y d i v i d e d b etw een the tw o bri dge
a rms , t hus m aint a i n i n gth e b a l a n c e .T h e e ffe c to f the addi ti onall ead i n
th e m eas ur inglin k i s n e g l i g i b l eb e c a u s ere s i s tanceR i s very smal l com-
p ar ed wit h t he in p u t i mp e d a n c eo f th e m e a s uri ngampl i fi er,but there
will be a very small reductionin the effectivek as describedin Chapter
8

The connectionR may be replacedby the strap - shown with a bro-

ken line, but t he u s e o f th i s s tra p i s re c o m m e n dedonl y w here connec-
ti o n leads ar e s h o rt, a n d w h e re th e re a re n o temperaturevari ati ons.
This is bec aus ew h e n th e s tra p i s u s e d th e b ri dgearm contai ni ngthe
a ct iv e gauge als o c o n ta i n stw o l e a d re s i s ta n c es,w hi l e the other arm
w i l l only c ont ainth e 1 2 0 Q d u m m y T h e l a tte r si tuati oncan cause er-
ro neousm eas ur e m e n tsb e c a u s e ,a l th o u g hth e bri dgecan be bal anced,
a l l t he t em per at u rei n d u c e dre s i s ta n c ec h a n g esi n the l eads w i l l now
o cc ur in only one a rm o f th e b ri d g e .T h i s p robl emof temperatureef-
fe c t s on t he leadsi s d i s c u s s e dfu rth e r i n C h a p ter8.

T he O uar t er B r i d g eAd a p to rZ R 0 0 1 8 i s a v a il abl efor use w i th 35OO

str ain gauges T h e c o n s tru c ti o na n d u s e o f the ZR 0018 i s i n al l re-
spec t sident ic alt o th e Z R OO1 4 , e x c e p tfo r th e resi stanceval ue of the
d um m y gauge.

Not et hat when a q u a rte rb ri d g ea rra n g e m e nits bei ngused,a temper-

a tur e c om pens a te d s tra i n g a u g e mu s t b e e mp l o yedw hi ch has a temper-
a tur e c har ac t er i s ti cm a tc h e dto th e te s t m a te ri alas w as descri bedi n
Chapt er3
A s was dem o n s tra te di n C h a p te r 6 , th e b ri dge sensi ti vi tymust be
cons ider edf or e a c h i n d i v i d u a lme a s u ri n ga rr angementbecause both
th e gaugef ac t or o f th e g a u g e su s e d ,a n d th e b r i dgeconnecti onsw i l l af-
fec t it . T he ex am p l e sg i v e n s h o w th a t th e b r i dge sensi ti vi tycan take
ver y dif f er ent v a l u e s , a n d i t s h o u l d a l s o b e born i n mi nd that the
e ngineerm ak ing th e me a s u re m e n tsma y b e m ore i nterestedi n obtai n-
i n g a s t r es s v alu e , o r th e l e v e l o f th e l o a d i ngthat causesthe strai n'
Th es e par am et e rsw i l l re q u i re a n e x tra mu l ti pl i cati onfactor to be ap-
to have an i nstrument
p l ied t o t he br idg e s e n s i ti v i tyl.t i s a d v a n ta g eous
where the gauge factor can be adjustedover a full decadeso that the
l e v elsdis play edc a n h a v e th e c o rre c tn u me ri c alval ue for the parameter
re quir ed.

T he gaugef ac to r, b ri d g es e n s i ti v i tyo, r a n y other mul ti pl i cati onfactor

can be adjus t edb y m e a n bo f th e " Ga u g e F a ctor"controlon the 1 526

55
 Fig 7 5 The CalibrationBridge ZR OOI3

The range of adjustment is between approximately0,90 and 10,70

w i th an ac c ur ac yo l !O ,O1 o /oAn
. a c c u ra c yo f 1 O ,OO57ocan be obtai ned
wh e n t he Calibr ati o nB ri d g eZ R O O1 3 i s u s e d .T he gaugefactor can be
shown on the displayfor adjustmentwhile a measuringarrangementis
b e i n g s et - up.

The Calibr at ionB ri d g ec o n ta i n sa fi n e l y b a l a n cedset of resi storsthat

h a ve an ohm ic di ffe re n c ec o rre s p o n d i n gto a 2 OOOgedi spl ay val ue
w i th a gaugef ac t oro f 2 ,0 0 .

In some circumstancesDC excitationcan cause a thermocoupleef-

fect becauseof temperaturedifferencesbetween solderedjoints, or at
other placesin the arrangementwhere dissimilarconductorsare in con-
tact. To avoid these effects,the Type 1 526 employsan AC carrier wave
system. AC levels are easier to amplify, and an additionaladvantageof
the carrier system is that it permits low bridge excitation voltages,
wh i c h in t ur n m inimi s eh e a t g e n e ra ti o nO. n th e 1526the mi ni mum ex-
ci tat ionv olt ageis O ,3 V, a n d w h e n th i s i s u s e d w i th 120Q gauges,i t
gives a total of only 75 mW per gauge to be dissipatedas heat, which
permits the instrument to be used for measurementson materialswith
very low thermal conductivities.When a higher excitationvoltage is
used, the signal to noise ratio in the amplifiers is improved, so that
much lower strain levelscan be resolved.With the highestexcitationle-
vel availableon this instrument (3,0V) it is possibleto measure up to
p l u s or m inus 2O O l e fu l l s c a l e ,w i th a re s o l u ti onof O,1pe.

A major disadvantageof AC excitationis the sensitivityof the measur-

ing system to reactive unbalance,usually in the form of capacitiveun-

56
balance caused by long bridge leads. Various methods have been de-
vised to counteractcapacitiveimbalance,the most common being to in-
clude variable capacitors in the bridge arms and then to achieve a
capacitiveand a resistive balance. The Type 1526 makes the extra
stage - capacitivebalancing- unnecessaryby using a 3 kHz square
wave as the excitationvoltage,and by gating only a portionof the mea-
su ri ng s ignal f r om th e b ri d g e fo r me a s u re m e n tas show n i n Fi g.7.6.
This gating effectivelyeliminatesthe signal peakscausedby the capaci-
tive unbalancein the bridge. With the demodulatorprinciple used, the
residual unbalanceis sensed as resistiveunbalance.which can be bal-
anced out with the resistance balance potentiometer. Typically a
0 ,0 5 f / F im balanc eo n a 1 2 0 O b ri d g ew i l l c a u s ean error i n readi ngof
t 1 digit in t he 1 V : 2 O OOg era n g e .

Signal from bridge

with capacitive unbalance

Bridge Supply Signal

Fig.7 6. Operation of the gating signal

When m eas ur in gs ta ti c s tra i n , th e m e a s u re dval ue can be read di -

rectly from the digital displqy, together with an indicationof whether
th e s t r ain is t ensi l e o r c o mp re s s i v es, h o w n b y a + or a - si gn. The
sign always follows the strain being experiencedby the gaugesin posi-
ti o n s R13 and R2a .

T he T y pe 1526 c a n a l s o b e u s e d to m e a s u redynami cstrai n,but now

the reading procedureis slightly different becausethe dynamic levels
cause the displayedvaluesto fluctuate unreadablyand the positiveand

57
negative signs to alternate. This situation is immediatelystopped by
p u s hing in t he " H o l d Ma x ." O N b u tto n to b ri n g the peak hol d functi on
into operation.This capturesthe maximumpositiveor negativestrain le-
vel, whicheverwas the greater,and a steadydisplayresults.

lf both static and dynamic strain levels are present, as in the exam-
ples shown, it is still possibleto use the features built in to the instru-
ment to separatethem. One of the low pass Filterswhich suppressfre-
quenciesabove 3Hz, 30Hz, or 3OOHz can be used to_selectively sup-
press the dynamic componentof the strain so that just the static level
is displayed.

Alternatively,the methoddetailedin the following examplecan be fol-

lowed. Push the Hold ON button as for a dynamic signal to yield a posi-
tive or negative maximum level as before. This level, which is shown
a s " A " in F i9. 7. 7 , s h o u l d b e n o te d . T h e F u n cti onS el ectorshoul d be
switched to BALANCEPOTENTIOMETER POSITIONand this displayed
value noted.SelectOPERATEmode again and adjust the BalancePoten-
tiometer carefully,depressingthe Hold RESETbutton from time to time
to keep a check on the display until it just changessign. Note the new
display level - this is value "B" in the figure. The static strain level
can be found by subtractingB from A, and the dynamic level by multi-
plying B by 2. The lesser maximum can be found by subtracting2 X B
from A. Examinationof the figure shows that the changedbalanceset-
ting has in effect moved the base line by an amount equivalentto the
static level, so that the dynamic levelsfluctuatesequallyon either side
of it.

Should it be necessaryto restore the BalancingPotentiometerto its

original position for any reason, this can be done quite easily without
unloadingthe specimen.BALANCEPOTENTIOMETER POSITIONis se-

Fig.7 7. Examples of combined static and dynamic strains

58
lected again and the Potentiometeradjusteduntil the displayedvalue is
the same as the value originallyrecordedfor the BalancePotentiometer
Position. The Potentiometer is now on exactly the same setting as it
was when value "A" was measured,and this is also the same as the
original zero balancefrom the commencementof the test.

Another method of determiningthe dynamic componentof the strain

is to use an AC voltmeter(for examplea Type 2425, or Type 24261,or
an oscilloscopeto measurethe AC componenton the DC level coming
from one of the DC outputs.

It is often required that measurementsfrom several different points

on the test specimenbe obtained,and it is very convenientwhen this
can be done without re-connectingand re-balancingmeasuring bridges
during the course of the test. Up to five bridges can be connectedto
th e 1 526 at t he s a m e ti m e , i n a n y c o m b i n a ti onof quarter-, hal f-, or
futl-bridges,and the bridgescan be selectedone after another for meas-
u rem ent .

When more than five points have to be measured,the Type 1544

Multipoint Selector and Control enables the 1 526 to be connectedin
turn to a seriesof measuringpoints automatically,so that the strain lev-
el can be measured. Up to ten measuringpoints can be connectedto
the 1544, and i{ even more points are required,Type 1 545 Multipoint
Selectorswith provisionfor ten more points in each unit can be con-
nected to the same Selector and Control 1 544 to give a maximum of
four hundred measuringpoints Selectionof each point is controlledby
the 1 544 which also has a digital displayto indicatethe identification
n u m berof t he m e a s u ri n gp o i n t c u rre n tl yb e i n gs ampl ed.

Each measuring point connectionon both 'l 544 and 1 545 has two
te rm inalsf or t he e x c i ta ti o ns i g n a lw h i c h i s d e ri vedfrom the 1526, and
two t er m inalst o re c e i v eth e m e a s u ri n gs i g n a l. E ach poi nt connecti on
has its own "Bridge Mode" switch and "Balance" potentiometer.

Automatic sweep rates from ten measuringpoints per secondto one

point every ten secondscan be selected Point selectioncan also be ac-
complished remotely by a TTL source, or by the oam switch of a Level
RecorderType 2305 or Type 2307. When necessary,individualmea-
suring points can be switched out so that the point sweep passesover
th e m in t he aut o ma ti c s e q u e n c e ,w h o l e d e c a des- i ndi vi dual 1 545
u n it s - c an als o b e b y -p a s s e di n th i s w a y . On e si ngl e measuri ngpoi nt
can be selected for measurementby a "Call" switch, and if several

59
 ..1:,::,o 3 rl 13 |) o |} {9 i9 rp

r r

+ *
:f:: e 0ftscnoss

co
| l

s nr r nr r r nr r

U r
Fig.7 8. The Multipoint Selector and Control Type | 544 and the Multi-
point Selector Type | 545

points are "called" in this manner, a restrictedsweep sequenceover

j u st thes epoint swil l o c c u r.

The built-in interface in the Selector and Control 1544 is a B & K

StandardInterface.lt is similar to the proposedIEC Standard,the excep-
tion being that it is not addressable.The data transmitted from the
1 544 consists of the measuring point identity number, and the strain
l e ve l.

The instrumentsthat have been describedin this chapter make up a

ve ry f lex iblem eas u ri n ga rra n g e me nth t a t i s s i mp l eto operate,and cap-
a b l e of per f or m inga w i d e ra n g eo f m e a s u ri n ga s s ignments

60
 CHA P TE R8
S OURCE SOF E RRO R
IN STRA IN GA UGE ME A S UR E ME NT S

Many different factors can cause measuring errors in strain gauge

systems.These sources fall into three broad categories,electricaler-
rors that arise in the measuring system, errors that have their source
in temperatureinduced effects, and errors due to faulty gauge moun-
ti ng or s elec t ion.

This chapter will analysesome of the more common sourcesof error

and suggestways to improvemeasuringaccuracyby properselectionof
system components, and by careful preparationof each item. lt will
also show how some measuringerrors can be compensatedfor auto-
matically,and how the effects of others can be minimisedby suitable
calibrationprocedures.

ElectricalSources of Error

Fig.8.l. Full-bridge showing some electric", iT"rr", of error

The figure shows the familiar bridge circuit, with the additionof lead
resistancesR1 and input resistanceto the measuringinstrument R;. A
generalexpressionfor the amplificationcan be written as follows,

v R". R
E (R2s+ R.3) (R24+R14)

61
lf the bridge is in balance to begin with.

E n :E :l : (8:2)
Rr o Rt . "
dR (8:3)
And if
R

e x p re s s i o ns i m p l i fi e sto ,
Th e n t he am plif ic a ti o n

Va (8:4)
(rro -r" rIrrs -tra ) fi -T) NP
Z: t;*

Wher e t he t er ms (1 - T ), N , a n d P a re e rro r factorsfrom vari ous

so u r c esas will be e x p l a i n e dl a te r.

Supply leads

The leadst hat s u p p l yth e m e a s u ri n gb ri d g ea r e by far the most seri -

o u s pot ent ials our c e o f e rro rs i n a n y s tra i n g auge arrangement.The
l e a d s s hould be s u i ta b l efo r th e p a rti c u l a rte s t appl i cati on'for exampl e
it is pointless to waste time and money on thoroughly waterproofing
th e gaugeins t alla ti o nw h e re p o o r q u a l i tyl e a d sa l l ow the penetrati onof
mo r s t ur e.

C hapt er7 m ent i o n e dm e th o d sfo r mi n i mi s i n gerrors causedby tem-

perature changes on long leads, but wherever possible, long leads
sh o uld be ex c lude dfro m th e a c tu a l b ri d g ea rm s . W here thi s i s not pos-
sible, the supply leads should still be as short as practicable,and made
up from the same length and type of cableto minimise resistiveand ca-
pacitiveeffects. The lead resistanceshould be kept as low as possible
to ex er t t he s m all e s ti n fl u e n c eo n th e o v e ra l l resi stanceof the bri dge
sys t em .

To get t he lead r e s i s ta n c ep ro b l e mi n to p ro p o r ti on,consi derthefol l ow -

i n g ex am pleas ill u s tra te di n F i g .8 .2 , w i th a s i n gl e acti ve 120Q gauge
h a v ingk = 2, O O .

The leadswill a c t a s a n i n te g ra lp a rt o f th e gaugeresi stanceso that

a n y v ar iat ionin t h e a mb i e n tte m p e ra tu rei n fl u encesthe i ndi catorread-
i n g dir ec t ly Heavyg a u g el e a d sw i th a c o mp a rati vel lyow resi stancew i l l
b e r equir ed,but e v e n s o th e .l e a d i n fl u e n c ec an be consi derabl ew i th

62
 RL

I 2OO Active Gauge

' RL 750316

Fig 8.2. Q.uartet bridge arrartgement with all lead resistance in the
same arm as the active stratn gauge

th i s ar r angem entl.f th e l e a d sa re 1 0 me tre s(3 3ft) i n l engthw i th a di -

a me t er of O , 8 m m (0 ,0 3 i n = 2 1 g a u g e ),a n d are made from copper
w i th a r es is t anc eo f 0 ,0 1 7 Q fo r a 1 mm2 c o n d u ctor1 metre i n l ength.

n ': o '0 r!^ :ro

:0 3 4 a
0,5
So t hat t he t ot al a d d i ti o n are
l s i s ta n c ei n th e b ridgearm i s 0,68Q.

Copper has a temperaturecoefficientof O,OO4/"C (O,OO22/oFl,so

th at a loC c hange i n te m p e ra tu rew i l l c h a n g ethe total l ead resi stance
by'
dRr : 0.68x 0.004

:0 ,0 0 2 7 { 2 /" c
This v alue c an b e s u b s ti tu te dfo r th e d R te rm in equati on3:1, to yi el d
an apparentstrain componentdue to temperatureeffects on the supply
l e ads .
dF
Apparentstrain
RK
0,0027
120x2,00
: 0,000011

: 11pd" C (6 tn/' F)

T o elim inat et hi s s o u rc e o f e rro r, th e s i n g l e a cti ve gauge shoul d be

connected into the measuring bridge with the three lead system that

63
 750372
L

Fi9.8.3. Recommended method of connecting single active gauge

with three lead system

i s s hown in F ig .8 .3 th e g u i d i n g p ri n c i p l e b ei ng to ensure that any

changesin resistancedue to temperatureeffectson the leadswill be di-
vided between two adjacent arms of the bridge cancellingeach other
out To further ensure that the leadsreally experiencethe same temper-
atures, it is quite commonto twist them together.

Hav ingnow ( ho p e fu l l ye) l i m i n a te dth i s s o u rc eof errordue to tempera-

ture effects on the supply leads, there still remains the basic modifica-
tion to the bridge resistancecausedby lead resistances.

Errordue to lead resistancecan be calculatedas follows.

7
Error Factor D- ( 8:5)
p
1L "cu
Rs

Where R", is the total resistancein the supply leadsand R, is the re-
sistanceof the bridge that must be overcomeby the supply (i. e. the re-

Fig 8.4. Full-bridge with lead resistances shown

64
sistance measuredbetween points D and E in Fig.8 4 when the bridge
is disconnectedfrom the measuringinstrument.
1 (8:6)
Relativeerror 1- P:
1+ RB
R",

As an ex am ple,c o n s i d e ra fu l l -b ri d g ew i th fo u r 120Q gauges con-

nected to the measuring instrument by leads that each have a resist-
a n ce RL = O , 23 Q .

1 -P :::;# d " :0 ,3 8 2 e '"

t
' 2 v g ,2 3

lEz@s

Fig.8.5. Half-bridge with lead resistances shown

For t he half - br id g es h o w n i n F i g .8 .5 ,c o n n e c tedto a Type 1526 hav-

ing precision 1 kQ resistors,and the same gauges and lead resistances
as before,
1x100 :0 ,1 9 2 %
1 -P:

A quarter bridge with the same resistancevalues, and connectedas

re com m endedin F i g .8 .3 w i l l h a v e th e s a me re l ati veerror as the hal f-
b ri dgeex am ple.

Pick-upin the supply leads can also be a problem,but it is greatlyre-

duced by twisting the leads around each other as for eliminatingtem-
perature error, or even better using a cable with a properlygrounded
screen.When the connectionshave been made, it is best to allow the
leads to be moved as little as possibleto avoid transients,especiallyaf-
ter the bridgehas been balanced.

65
Bridge non-linearity

When a direct reading instrument is being used, the output from the
measuring bridge will not always be an exact linear function of the
change in gauge resistance,becausethe bridge is out of balancewhen
th e m eas ur em en ti s ma d e .T h e (1 - T ) te rm i n equati ons8:1 and 8:4
is the error factor due to measuringstrain with the bridgeunbalanced.
(r -rrl (rz s -o rt/ (rr a-r.,) (r" o-arro)
1 + r2s+a(1 + r13) 1 + r24+a(1 + h4)
Th e r elat iv eer r or T : (8:7)
f2q - f2g* f1s - f14

When a = 1, which will be the case when the bridgeis made up from
a set of matched strain gauges,or matchedgauges and precisionresis-
tors, the expressioncan be simplifiedto,
r z4 2 - r 2 s2 *r p 2 - 1 1 4 2
T= (8:8)
2( r r o- r r "I r . "- t t q)

lf R2a and R1a are used as active gauges, the expressionfurther

si m plif iest o,
1
( r 24+ r t 4) (8:9)
2

When R2a and R 1 4 a re a c ti v eg a u g e sw i th th e opposi tesi gn and the

sa m e m agnit ude,l i k e th o s e s h o w n i n F i g .6 .3 , the rel ati veerror T w i l l
be O.

l f only one ac t i v e g a u g e i s i n u s e , a n d th e b al anceci rcui t i s al so i n

use, it is best to connectthe gauge into the oppositeside of the bridge
fro m t he balanc ec i rc u i tsw h e n e v e rp o s s i b l e .When the 1526 i s bei ng
used with a quarter bridge adaptor,the influenceof the balancecircuit
o n R13 s hould be ta k e n i n to c o n s i d e ra ti o nU. s ing equati on8:7 w i th a
= 1 and only R24 = O,

-r (8:10)
2
A rule of thumb which can be used for the case with one active
strain gauge says that the relative error factor T is the same size as the
measuredstrain (in strain)when gauge factor 2 is used,
as ek : r
t r*
th en t:-:- (8:11)
22

oo
 For example:with gauge lactor 2, the indicatedstrain for a system
w i th one ac t iv egau g ei s 1 7 1 3 9 e

Therefore T : strain: 0,001713

And 1713 x 0,001713 : 2,93ue

So t he c or r ec t edm e a s u re m e nst h o u l db e 1 7 1 6 pe.

Note that this can give large errors when semiconductorgaugeswith

high gaugefactors are being used.

Use of the balancecircuit

When a balancingcircuit is being used, the balancepotentiometeral-

ters the resistancevalue of one or more arms of the bridge.lf these re-
sistanceshappen to be strain gauges, their sensitivityto strain will be
reduced, thereby introducing an error. This effect can be shown by
means of an example.

6,.1tP
' Jwvll -r- z@4s
Fig.8.6. The components ol R1s

C ons idert he 15 2 6 u s e d w i th o n e a c ti v e 1 2OQ gauge i n posi ti on

R13 and the full 5OOQ range of the potentiometerused to balance it
when the "2o/o" resolution mode has been selected, (this is the worst
case with a 12OO gauge). In effect, resistanceR13 will be made up
from two resistorsas shown in Fig.8.6.
111
R rr: 1 n - 6 0 4 0
R .::1 1 7 ' 6 6 2 Q
So the error in measurementwill be,
:
,1 2 0 -1 1 7 ,6 6 2 ,.^
tvOo/o
t- t
' 120
: 1,948o/o
Similar calculationscan be made for the errors that occur with other
gauge resistancevalues,and with other balancingmethods

67
 ?N 0 200

,@/t

+1 8 0 0

Switch 1 0 6%

Fiq.8.7. Error due to use of balance circuit, with gauge resistancesas

shown, and a single active gauge used in position Rrr, (with
R ,, the curves will slope the other way)

68
Error due to input impedanceof the measuringinstrument

The error factor due to the input impedance,

(8:12)

where Ro is the resistanceof the bridge between points 4 and 3 with

the measuringinstrument disconnected,and R; is the input impedance
o f th e m eas ur ingins tru m e n t,(s e e F i g .8 .8 ).

error 7-N:
Therefative (8:13)
+
'*
^"

Fig.8. 8. lnput impedance of measuring'iff{tru^"rt

U s ing t he 1 526 w h i c h h a s a n i n p u t i m p e d a n ceof 67,5 kO as an ex-

a mp l e, and a f ull b ri d g ew i th fo u r 1 2 0 Q s tra i n g auges,the bri dgere-
sistancemeasuredbetweenpoints 4 and 3 will be,

1: 1 J- 1
Ro 120+120 120+120
So R o :1 2 0 Q
7 -N :
1+ 6 7 5 0 0
120
:1 ,7 8 x 1 0 -3 :0 ,1 7 8 o/"

Some measuring instruments are calibrated for use with strain

g a u ges hav ing one p a rti c u l a rre s i s ta n c ev a l u e . The 1526 i s cal i brated
in such a way that when it is used with a full bridge employingfour
1 2 0 Q s t r ain gaug e s th e e rro r j u s t c a l c u l a te da bove w i l l be compen-

69
 Fig.8. 9. Hatf-bridse with 35O n"ffi,

sated automatically.However, this compensatederror will have to be

deductedfrom the total error when a calibratedinstrumentis used with
a bridge arrangementhaving different resistancevalues.This can be il-
lustratedby means of an example.Fig.8.9 shows a half-bridgearrange-
ment using two l kQ resistorsand two 350O strain gauges.
The bridgeresistancebetween points 4 and 3,
1: 1 L' 7
Ro 1000+350 1000+350
R o :6 7 5 Q
with the1526 1-'N ::itd
'- 67s
1 0 -3 :0 ,9 9 0 %
:9 ,9 0 x
The measuring error with respect to the calibrationwith 12OQ full
b ridge, :0 ,9 9 0 _ 0 ,1 7 8
:0 .8 1 2 " /o
Error

0
Full Bridg with 4 equal Gauges

50 70 lm 120 200
200 300
300 500 700 1k 2K
Gauge Resistance,o ze2a4
Fig.8. I O. Measuring error on the | 526 due to input impedance
70
 A chart can be calculatedfor the errors arising from different combi-
nations of gauge resistancein full and half-bridgesfor any instrument.
Fig.8.1Ois a typicalgraph showing this type of measuringerror.

Temperature variations

Changes in the temperatureof the measuringsystem can cause er-

rors for several reasons.First, the resistance,and hence the k factor of
the gauge conductormaterialchange with temperatureto cause an ap-
parent strain. Another possiblecause of error is that the coefficientof
expansionwith temperaturefor the gauge may not be the same as that
of the test object's material.A third reason is that the resistanceof the
supply leads changeswith temperature,this was discussedin the previ-
ous section.

Various methods can be used to eliminate, or compensatetempera-

ture induced measurementerrors. Use of dummy gauges as described
in Chapter6, will compensatefor variationsin the k factor with temper-
ature, and for differencesin the coefficientof expansionbetween the
gauge conductor material and the test object material. The dummy is
cemented in the same way as the active gauge to a piece of similar
materialto the test object, and connectedin the measuringbridge in an
arm adjacent to the active gauge. Being exposedto exactly the same
temperaturevariationsas the active gauge,the resistancechangesdue
to temperaturewill be the same as in the active gauge, so their effects
wiff be cancelledout. Compensationwith this method is suitable lor ar-
rangementshaving one, two, or four activegauges.

When the dummy arm of a quarter bridge is made up from a cali-

brated resistor, or when it is impossibleto subject the dummy to pre-
cisely the same temperaturevariationsas the active 9auge, one of the
following alternativescan be applied.Self-compensatingstrain gauges
have been developedthat are immune to temperaturevariations.A cali-
brated length of another conductor material, or a thermocoupleis in-
cluded in the measuringgrid, and has the oppositetemperature/resist-
ance characteristicto the grid material so that errors due to variations
in the gauge resistancewith temperaturewill be cancelledout.

Furthermore, manufacturers have gone to great lengths to develop

conductor materials that exhibit only a very small resistancechange
over quite extensivetemperaturerangeswhen they are cementedon to
the correct test specimen material. Manufacturerscan supply gauges

71
with temperature expansion coefficients to suit the most commonly
used materials,and this type of gauge can often be employedwhen the
temperaturevariationsare not expectedto exceedperhaps 5O to 8OoC.
Even when the temperaturerange is greater,the error in measurement
i s u s ualf yf es st han 2 p e p e r o C .

Gauge mounting and connection

Incorrect mounting of the gauges on the test specimencan produce

major errors due to creep, moisture effects, poor electricalinsulation
fro m t he t es t piec e , e tc ., a s me n ti o n e de l s e w h e re.Gaugeal i gnmenti s
a l so v er y im por t an ti n e l i mi n a ti n ge rro rs ,a n d w hen possi bl e,a prel i mi -
n a ry t es t r un us ing a n a p p l i c a ti o no f b ri ttl el a c q u ercan gi ve val uabl ei n-
formation on the directionof strain "flow" to assist in the correct posi-
ti o n ingof gauges .
The effects of moisture changing the balance point by altering the
gauge insulationwere mentionedin the chapter on adhesives.The best
safeguardis a thorough water-proofingprogramto preventthe penetra-
ti o n of m ois t ur einto th e c e me n to r b a c k i n g .

Diffusion is a long term effect, where atoms actuallymigratefrom ce-

me n t t o t he bac k in go r to th e te s t m a te ri a la n d c a usechangesi n i nsul a-
tion, or even a breakdownof the cement bond. The use of inert materi-
a l s c an help gr eat l yto a l l e v i a teth i s e ffe c t.

After the first load cycle has been appliedto the test specimen,and
th e n r em ov edaga i n , i t i s s o me ti m e sfo u n d th at the bal anceposi ti on
has moved. This effect can have been caused by either of two condi-
tions. lf the load cycle has taken place in a comparativelyshort period
of time, the change in balance point is very often due to hysteresis,as
the other possiblecause, creep, may take a longer period to show it-
self. To minimisethe effectsof hysterisison the measurements,it is rec-
ommended that two or three load cycles are appliedto the test object
- if this is possible- before commencingthe test. The balancepoint
should be checked after each cycle, as after the secondcycle the bal-
ance point will probablybe stable within 0,1 or O,2o/o, so that measure-
ment can begin. After extendedperiodswith alternatingloading,an ef-
fect similar to hysterisiscan occur again due to gaugefatigue.

All m et alsex pe ri e n c eth e p h e n o m e n o nfa ti g u e w hen they have been

subjected to alternating stresses for longer periods of time. Strain
g a uge c onduc t or m a te ri a l s a re n o e x c e p ti o nto thi s rul e, and thi s

72
should be rememberedwhen measurementsare to be made on a sys-
tem with alternatingloads,particularlythose operatingat high frequen-
cies. The total permitted number of strain reversalscan be reached
very rapidly. For example,taking a typicalvalue for wire or foil gauges,
1 07 s t r ain r ev ers a l sa t t 1 0 0 0 l r c a n b e a c hi evedi n l ess than three
h o ur s at 10O O H z ,o r i n j u s t o v e r a n o rma lw o rki ngday at 3OOH z,i rre-
spectiveof whether or not measurementsare being made. Fatiguefail-
ure of a gauge usually takes the form of a fracture where the conduc-
tor leadsjoin the gauge grid, however, sometimesthe adhesivebreaks-
down first.
Creep is a much more complexproblemto diagnoseor cure, because
its effects,which are variable,rarely appearfrom the outset of the test.
When a strain gauge has been subjectedto a high strain level or has
b e en plac edin a h i g h te m p e ra tu ree n v i ro n me nt,or someother envi ron-
ment capableof breakingdown the bonding,the gauge may "creep" on
the surface to which it was attached, or the gauge conductor may
"creep" in the backing. Either of these effects, can cause substantial
ze r o dr if t , or s t at i c s tra i n re a d i n g sth a t fa l l s l o w l yw i th ti me as the el as-
ticity of the gauge conductor material tries to pull the grid back to its
u nloadedc onf igu ra ti o n G . e n e ra l l y ,a l te rn a ti n gl oads are l ess l i kel y to
give creep problemsthan are static loadingsystems.
Where only a few gaugesare showing signs of creep,the most satis-
factory remedy is to cement new gauges in place of the old, and simply
discardthe old gauge as attemptsto remount a faulty gaugewill almost
always be fruitless.
I n t he s it uat io nw h e re ma n y g a u g e si n a l a rge i nstal l ati onare show -
ing signs of balance point shift, or drifting strain levels,once the obvi-
o u s pos s ibilit yo f w ro n g l y a p p l i e d l o a d i n g h a s been el i mi nated and
creep definatelyidentified,a reappraisalof the gauging method is obvi-
ously called for. The adhesives and backing materials used must be
completelysuitable for the materials being tested. They must survive
the anticipatedtemperature range, and the maximum strain level ex-
pected. Furthermore,the utmost care must be taken during the actual
bonding of the gauge to the specimen,so that all mating surfacesare
thoroughlyprepared,and the adhesivecorrectlymixed.

One importantpoint, that is always worth remembering,a point valid

for any type of strain gauge system, is that no matter how accuratethe
measuringinstrument, or how well reasonedthe choice of gaugesand
bridge arrangementsmay be, if the person who actually mounts the
gauges on the test specimen made a mistake, the measurementscan
b e m is leading,or u s e l e s s .

73
 CHAPTER 9
ME A S URE ME NTOF S TR A I N
WITH OTHE R B & K INS TRU ME NT S

The engineerwho has to measurestrain only at very infrequentinter-

vals, or who has to perform a preliminarysurvey to find whether strain
measurementis of value to a particulartest program,may be reluctant
to invest in a Type 1526 Strain Indicatorjust for these "once in a
while" measurements.This chapter will explain how useful strain
measurementscan be obtained from alternative arrangementsbased
on B & K instrumentsthat the user may alreadyhave in the test labora-
tory or inspectiondepartment.

Deviation Bridge Type 1521

A simple method of measuringstatic strains uses an instrumentthat

may be found in the productiontesting. or receivinginspectiondepart-
me n t . F ig. 9. 1 s how s th e a rra n g e me nfot r me a s u ri ngstrai nw i th a D evi -
a ti on B r idgeT y pe 1 5 2 1 . T h e 1 5 2 1 c o m p a re sth e val ue of a resi stance,
inductance,or capacitancewith a standardvalue, and gives a meter in-

O -t-

- ooo :
a aa- : a

9. I a. Measurement with one 9. I b. Measurement with two

active gauge acttve gauges
Fig.9.l. Measurement of static strain using a Deviation Bridge Type
1521

74
dication of the percentagedeviationfrom the standard.The instrument
contains a measuring bridge circuit where the Standard and the Un-
known form two arms of the bridge. When used in the Resistance
Mode, a dummy strain gauge can be employedas the Standardresis-
tor, and an active gauge as the Unknown. This arrangementis shown
i n Fig. 9. 1a.

Very briefly, setting up the instrumentfor measurementwith this ar-

rangementis as follows. The dummy gauge must have the same resist-
ance as the active gauge, it is cementedto similar materialto the test
piece and mounted near it to give automatic temperaturecompensa-
tion. The dummy is connectedto terminals 1 and 2 (Standard).The ac-
tive gauge is connectedto terminals 3 and 4 (Unknown).

a Fdb.

#
8r!d
erdg.

a
TYF

Ojxie'
1521

&

saa qqy"
Fiq.9.2. fhe Deviation Bridge Type | 521 -
-
75
 Select Bridge Mode "R" with 1 0O Hz excitation frequency, and set
the Function switch to RLC. Switch the RLC Deviationknob to "Zero"
and adjust the zero positionof the needle.Switch to "Ref." and adjust
the referencelevel of the meter to "1" on the auxilliaryscale.This sets
the excitationvoltageacrosseach gaugeto approximately0,36V.

U s ing t he m os t s e n s i ti v es c a l e (S A 0 1 1 4 , s h o w n i n Fi g.9.2)and the

most sensitivemeasuringrange (O,2o/o RLC Deviation),it should be pos-
sible to read the zero load out-of-balanceon the meter as a positiveor
negativepercentagedeviation.The tolerance on the gauges should be
good enough to permit this, but if the deviationis larger than the O,3%
full scale deflection on the most sensitivescale, the Standard resist-
ance can be adjustedto balancethe unknown. Note that it is not neces-
sary to balancethe bridge so that the needle rests at the "O" point, be-
cause as long as the needledeflectionfor the loadedcondition can be
read from the scale. the zero load deflectioncan be subtractedand the
changedue to loadingfound.

At no load. if the needlegoes off the scale at the positiveend (indicat-

i n g t hat t he Unk no w n i s th e l a rg e r),a s ma l l re s i stancei ncrement- for
example a section of resistancewire - can be put in series with the
Standardto increaseits value. lf the Standardis the larger, a large re-
si stanc ein par alle lw i th th e S ta n d a rdc a n re d u c ei ts val ue.

Note that the Unknown resistancecan also be modifiedto help bal-

ance the bridge, but this introducesan error to the value of the Gauge
Fa c t ors im ilar t o t h a t d i s c u s s e du n d e r " U s e o f t he B al anceC i rcui t" i n
C h a pt er8.

One final check can be made beforetaking measurements,when it is

suspectedthat there could be capacitiveunbalancein the system, (per-
haps due to the use of long measuringleads).This can be determined
by simply switching the Functionswitch to "Phase" and finding the de-
flection- if any. From a practical point of view, if the meter needle
falls anywhere on the scale when the Phase Deviationis at its most
sensitivesetting (0,002 tan6), the effects of capacitiveunbalancecan
be ignored.Similarly, if strain is being measuredon the 1% RLC Devia-
tion range, capacitiveunbalancecan be ignored if the needle stays on
the scale when O,01 tand Phase Deviationis selected.lf necessary,ca-
pacitivebalancingcan be performedin the same way as resistivebalan-
cing, by putting capacitorsin parallelwith the gauges.

to
 Meas ur em entwill b e e x p l a i n e db y m e a n s o f a n exampl e.A dummy
a n d an ac t iv e gau g e , b o th w i th n o mi n a l re s i s tance12OO are con-
n e cted in t he ar r a n g e me n ts h o w n i n F i g .9 .1 a . The gauges have a
Ga u ge F ac t ork = 2,O Oa n d i n th e u n l o a d e dc o n d iti onthe i nstrumenti n-
d i ca t es+ O , 08%dev i a ti o n .

When the specimen is loaded, the needle indicates -O,24o/o devia-

tro n .

Ref er r ingbac kt o E q u a ti o n3 :1 ,
, 6 R .d t
K:
n/ t
dR
T his can be re written as -k R (9:1)

W he re th e d R te r m c an be det er m ined f r om t h e instrument deflec-

t ion.

In th e examp le, the t ot al def lec t ion due t o loadin g

dR: - 0 ,2 4-(+ 0 ,0 8 )%
: - 0 ,3 2 Y" i . e . th e a cti vegaugei s i n
c o mpressl onl
-0 ,3 2 x 1 2 0
S ubstitu ting in to 9 :1 ,- _
2 ,0 0 x 1 2 0x 1 0 0
t:-0 ,0 0 1 6 t :-1 6 0 0 pt

Two point s ar e ob v i o u sfro m th i s re s u l t F i rs t,the actualval ue of the

resistanceis not too imoortant as the terms cancel out. Second, the
l o a d ing c aus ed a c h a n g e i n re a d i n g o f 3 2 g ra d uati onsof the scal e,
w h i c h is equiv alen tto 1 6 O 0 g e . T h a t i s , w i th Gauge Factor 2, one
sca l egr aduat ionr ep re s e n ts5 O g e.

The instrument can be used with Gauge Factorsthat do not give an

"easy" relationshipwith the meter indication, by adjustingthe Refer-
e n ce lev el.

For ex am ple,if t h e Ga u g eF a c to rfo r a p a rti c u largauge i s 2,2 set the

Re fe r enc elev el t o 2 /2 ,2 = O ,9 1 (w h i c h c a n b e e sti matedon the mai n
scale),to retain the simple relationship50ge,/graduation.

77
 When more sensitivity is required, two active gauges can be em-
ployed, one in the Standard arm and one in the Unknown arm, as
sh o w n in F ig. 9. 1b .M a k i n g u s e o f th e P o i s s o ne ffect w i l l gi ve approxi -
mately 40pe/graduation with Gauge Factor 2, and if the test piece is
in bending with one active gauge in tension and one in compression
25pe/graduation can be achieved.Note that only three leads are re-
quired with two activegaugesas terminals2 and 3 are connectedinter-
n a l l y.
D y nam ics t r ains ca n a l s o b e me a s u re du s i n g a 1521' as thi s i nstru-
ment features a DC output that can be fed to an oscilloscope,or to a
p e a k r eadingA C v o l tme te r.T h e o u tp u t g i v e s 1 ,1 V for ful l scal e defl ec-
ti o n , whic h is equ i v a l e n tto 3 6 ,6 mV ./g ra d u a ti onon the 1521. W hen
both static and dynamic strains are presentthe 1521 will indicate the
static strain, while the AC voltmeterwill measurethe amplitudeof the
ripple on the DC level (peak measurement)thereby showing the dy-
n a m ics t r ain.
Similar measurementscan be made with the DeviationBridge Type
1 5 1 9, but it s highe r e x c i ta ti o nfre q u e n c yc a n a c c entuateany capaci ti ve
u n balanc e.

Voftmeter fype 2425

The ElectronicVoltmeterType 2425 can be used on the DC output of

th e 1521 ( or 1519 ) to me a s u red y n a m i cs tra i n , but thi s vol tmeteral so
has severalfeatures that make it suitablefor measuringdynamic strain
a l o n e. F ig. 9. 3 s how s tw o m e a s u ri n ga rra n g e mentsthat can be used to
indicatestrain on a 2425.

One, two or four

active gaug6

7M59

9.3a. Measurement with 9.3b. Measurement with

Wheatstone bridge simplified bridge

Fi9.9.3. Measurement of dynamic strain using a Voltmeter Type 2425

78
 Excitationof either of the bridge circuits with a DC voltage(for exam-
ple a 1 ,5 V battery)will give no indicationon the AC meter when there
is no dynamic loading present on the specimen.As soon as a dynamic
load is applied, the voltmeter detects the superimposedAC that ap-
pears on the DC level, and the peak levelof this ripple is proportionalto
the dynamicstrain present.

The 2425 can measurepeak levels and it also has a Peak Hold func-
tion to capture transients. Other features of the 2425 lhat help to
make it suitable for strain measurementare its frequencyrange down
to O,5Hz, and its most sensitivevoltagerange of 1 mV full scale deflec-
tion, which gives 20gV/scale graduation.One feature that can be a
drawbackis the very wide frequencyrange up to O,5 MHz with no facili-
ties for filtering out high frequencynoise.The measuringbridgecircuits
operate quite well as antennae, so thorough screeningand grounding
is very important in reducing pick-up and noise. lf it should prove very
difficult to eliminate high frequency noise from the system, its effects
can be greatly reducedby making the measurementsin the RMS mode,
and then multiplying the measured levels by the Crest Factor (i.e. by
1 ,4 f or a s inus oidal o a d i n g ).

The dynamicstrain level is given by equation 6:8,

4V
- Ek
where
V - potentialmeasuredby the voltmeter
E = the excitationvoltage
k = Gauge Factor

Note that the Gauge Factor used here is actuallythe bridge sensitiv-
ity factor.

Considerthe examplewhere only one active gauge is used with k =

2, and the excitation voltage is 1,5V. Putting these values into the
equationand taking V = 1 mV (full scaledeflection)
1 x 1 0 -3
,,---4 x
1 .5 x 2
: 1,333 x 1 0-3e : 1333pe (f sd)

Which gives 26,66 ge/ scalegraduation

79
 For a Full-bridgewith bridge sensitivityfactor 8, and 3,OV excitation
voltage the equation gives 166/e full scale deflection and
3,33 pe/ scalegraduation.

Voftmeter 2425 and Conditioning Amplifier Type 2626

As previouslymentioned, noise pick-up can be a problem when the

2425 is used alone, and if the system noise is higher than the zero
load out-of-balancelevel, accuratecalibrationwill be impossible.An im-
provementin the signal-to-noiseratio can be obtainedwhen the signal
from the measuringbridge is passedthrough a preamplifierbefore be-
ing fed to the Voltmeter.A low noise preamplifiersuitablefor this pur-
pose is the ConditioningAmplifier Type 2626, which is normally used
as an accelerometerpreamplifier.At the price of some minor circuit
complication,severaladvantagesare gained from adding a 2626 to the
arrangementwith a 2425. Up to 60dB of amplificationis available
from the ConditioningAmplifier, and progressivehigh and low pass fil-
tering can be selected in steps to limit the width of the measuring
band. Furthermore, the three digit sensitivity adjustment allows the
gain to be set to give a direct readingof strain on the Voltmeterscale.

a 0iH.-05rH.

diib .U# ffi

@
Fi9.9.4. Type 2425 and Type 2626

The m inor c om p l i c a ti o ni s th e fa c t th a t th e 2626 i s a charge ampl i -

fier, so a 1 nF capacitorhas to be put in serieswith the bridgeto give a
charge source (in effect this constitutes a dummy accelerometer).
Fi g . 9.5 illus t r at esth e a rra n g e me n t

80
Fi9.9.5. Arrangement for measuring strain using a ConditioningAmpli-
fier Type 2626 and a Voltmeter Type2425

Setting-upand adjustmentis best explainedby means of the follow-

ing examples.Considera simple case first where the bridge contains
two active and two dummy gaugeswith k = 2 to give a bridge sensitiv-
ity for the arrangementof 4, and where the excitationvoltage is sup-
plied by a 1,5V battery.The limiting frequenciesare adjustedto cover
only the frequencyrange of interest. Setting the Volt/g Out to position
"1 " in t he r ange 1- 1 1 p C /g ,a n d k e e p i n g th e S ensi ti vi tysettoasi m-
ple figure of 1,OO a dynamic level of 22OmY peak is recordedby the
Voftmeter,now what does this 22OmY mean in terms of strain?

Becausethe output was set to 1Y/9, each volt recordedby the 2425
is the equivalentof 1 g measuredby an accelerometer.Further, when
the source impedancecontains 1 nF a Sensitivityof 1,OOpC/gwill give
the equivalent of 1 g for every millivolt coming from the source
(1 mV,zg),so that the Voltmetershows 1V for every 1 mV coming from
th e b r idge( i. e. 60 dB a m p l i fi c a ti o n ).

U s ingequat ion6: 8 a g a i n ,

t:H
3
_ 4 x 0 ,2 2 0 x 1 0
1 ,5 x 4
:0 .1 4 7 x 1 0 -3e

:1 4 7 u e

Another example showing the use of the Sensitivityadjustment,with

all parameters as before, except that Sensitivity is adjusted to
1 ,5 0 pC/ g ( i. e. nu m e ri c a l l yth e v o l ta g e ),th e V o l tmeter w i l l i ndi cate
1 4 7 m V ( num er ic a l l yth e s tra i n ).T h i s i s b e c a u sethe output i s 1V ,/g

81
fro m t he 2626, w h i l e i ts i n p u t i s 1 ,5 m V./g g i vi ng 1V i ndi cati onfor
e a c h 1, 5 m V f r om th e b ri d g e
4 x 0 ,14 7 x 1 0 -3
Hence Y 1, 5
1 ,5 x 4
:0 ,1 4 7 x 1 0 3 t

:1 4 7 u e

G augef ac t or so th e rth a n 2 c a n a l s ob e c o m p ensated for by the S ensi -

ti v it y adjus t m en t T a k i n gth e p re v i o u se x a mp l eagai n but usi ng gauges
wit h a G auge F a c to r2 ,2 (i . e b ri d g e s e n s i ti vi tyfactor w i th tw o acti ve
g auges wif l be 4 ,q , rt th e Se n s i ti v i tyi s a d j ustedto 1,5 x 2,2/2 =
1 , 65pC/ g, t he o u tp u t fro m th e 2 6 2 6 w i l l s ti l l be 1V ,/9, but the i nput
w ill be 1, 65m V , 2 g ,w h i c h g i v e s 1 V i n d i c a ti onfor each 1,65mV from
th e br idge,and t h e Vo l tm e te rw i l l s ti l l i n d i c a te147 mY

4 x 0 ,1 4 7x 7 0 3
x 1,65
1 ,5x 4 ,4
:0 ,1 4 7 x 1 0 -3t

:1 4 7 u t

Vibration Meter Type 2511

A not her ins t r u m e n tth a t c a n b e u s e d to me asuredynami cstrai n has

mos t of t he f ea tu re so f th e p re v i o u ss y s te m combi nedi n one i nstru-
m ent , t he G ene ra l P u rp o s eV i b ra ti o nM e te r Type 2511, and adds the
p o s s ibilit yof f r e q u e n c ya n a l y s i so f th e s tra i n usi ng an externalfi l ter.
W hen t he s im pli fi e db ri d g es h o w n i n F i g .9 .3 i s used,a separatebattery

'{'

Fi g.9.6a. Type251 |

82
 Fig.9.6b. Type | 621

w i l l not be r equir e da s th e 2 8 V D C s u p p l y fro m the 2511 can often

g i ve s uf f ic ientc urre n t to e x c i teth e b ri d g e .F i g .9.7 suggestsa sui tabl e
a rra ngem entt hat i n c l u d e sp ro v i s i o nfo r fre q u e ncyanal ysi sw i th a Tun-
a b l e B and P as sF ilte rT y p e 1 6 2 1 .

Fig 9 7 Arrangement for measurement and analysis of strain using a

General Purpose Vibration Meter Type 251| and a Tunable
Band Pass Filter Type 1 621

Whe n 1 20 O str ain gauges ar e us ed t oget her w i t h a 1 4 k Q v o l t a g e

dropp er, th e to tal res is t anc e in t he c ir c uit will be g i v e n b y ,

R r: 1 4 0 0 0 +1 2 0 + 1 2 0

:1 4 ,2 4 k A
W it h a nom inal s u p p l y v o l ta g eo f 2 8 V th e c urrent requi rementw i l l
be,
i :2 8
1 4 ,2 4 ^ A
-- 1,94mA

83
 T his will not ov e rl o a dth e 2 8 V o u tp u t w h i c h has a current l i mi t of
2 mA, so the voltageappliedto each gaugewill be,
E: 1.9 4x 1 2 0m V
:0 ,2 3 3 V
T he 2511 oper a te sa s a v o l tme te rw h e n a 1 nF capaci tori s con-
nected in series with the input voltage, so that when the instrument
has been calibratedfor lOOpC./g accelerometerseach "9" indicatedin
th e A c c eler at ionm o d e re p re s e n ts1 0 OmV (i .e. 1Og/V l . B eari ng i n
mi nd t hat t he 251 1 me a s u re sp e a k -to -p e a kl e vel s,a S cal e factor can
be derived from the excitation voltage, the g./V relationship,and the
b ri d geG augeF ac t o rb y u s i n g e q u a ti o n6 :8 a g a i n .

,: !!
Ek
Where

Y = peak me a s u re dv o l ta g e
E = bridgeexcitationvoltage
k = bridgesensitivityfactor

Now v :v P P
2
A nd Vo,: 1 0 g o -o

s o t hat Y :1 0 9 o o

59p p

N ow r ear r angin ge q u a ti o n 6 :8 to g e t th e re l a ti onshi pbetw een mea-

"9"
sured r,r",l:
"nO
a, g ?, o
EK
s_E k
e20

A nd put t ing in v a l u e s fro m th e a rra n g e m e n tshow n i n Fi g.9.7 w i th

two activegauges giving a bridge sensitivityfactor k = 4, and with exci-
ta t ionv of t age= 2x O,2 3 3 ,
g -2 x 0'233x 4
ScareFactor
e
: o,ogelz
gr,
84
 F or ex am pleif t h e me te r i n d i c a te s0 ,OO18 g p -p w i th thi s measuri ng
arransement'
e
:';;!i::,;io"'

W hen a T unabl e F i l te r 1 6 2 1 i s c o n n e c te da s the externalfi l ter for

th e V ibr at ionM et e r, th e s i g n a l fro m th e me a s u ri ngbri dgecan be fre-
q u enc y analy z ed.F u rth e r,th e n a rro w b a n d fi l teri ng avai l abl egi ves a
big improvementto the signal-to-noiseratio of the set-up.

Measuring Amplifiers and Analyzers

!,r

r^
-

Fi9.9.8. Type2l | 4. Type212O and Type2Ol O

85
 One final group of instrumentsthat can be used for measurementof
dynamic strain are the B & K MeasuringAmplifiers and FrequencyAna-
fyzers.The Measuring Amplifiers Type 26O6, Type 2607, Type 2608,
and Type 2609 can be used in exactlythe same way as the Voltmeter,
but as the 2608 and 2609 only record RMS levels the Crest Factorof
the appliedloading must be known when these two instrumentsare to
be used. The Type 2609 has a most sensitivevoltagerange of tOOIV
for full scale deflection,and can measure in the audio frequencyrange
2 OHz t o 2O k Hz . Th e o th e r M e a s u ri n gAmp l i fi erscan.measure l OgV
with full scale deflection and have a frequency range from 2Hz to
2 OOk Hz .

When one of the MeasuringAmplifiers(not 2609) is used with an ex-

ternal filter, or when one of the FrequencyAnalyzersType 2O1O,Type
2 1 13, T y pe 21 14 ,T y p e 2 1 2 0 , o rT y p e 2 1 2 1 i s used, the si gnal from
the measuringbridge can be filtered to improve the signal-to-noisera-
tio, or to make a frequencyanalysisof the dynamic strain levels.Again
th er e wilf be s ome re s tri c ti o n sa s th e 2 O 1 0 a nd 2121 can onl y mea-
su r e RM S lev els .

All analyzerscan make linear measurementsin the frequency range

2Hz to 2OOkHz,and have a maximum sensitivitythat gives 1O/V full
scaf e def lec t ion.T h e 2 1 1 3 a n d 2 1 2 1 c a n ma ke frequencyanal ysi sof
th e s t r ain s ignals i n th e ra n g e 2 0 H z to 2 O k H z, the 2113 w i th octave
or third-octave,and the 2121 with third-octave,lOYo,3o/o,or 'lo/ocon-
stant proportionalbandwidths.The 2120 can make narrow band (third-
o ct av e, 1U/ o, 3Y o ,o r 1 o /oal n a l y s i si n th e ra n g e 2H z to 20kH z, w hi l e
th e 2114 c ov er s th e fre q u e n c yra n g e 2 H z to 18OkH z i n octavesor
third-octaves.The 2O1O can make frequency analysis over the range
2Hz to 2OOkHz. with constant bandwidths selectable between
1 O00Hzand 3, 1 6 H z .

Ore, two
or four actiw irdi:
'o
o t-o.'
.oo
74N8

Fig.9.9 Dynamic strain measurement arrangement using a Frequency

Analyzer as the measuring instrument

86
 The method of zero calibration,and measurementwill be the same
as when the Voltmeter is used. For example,when one active gauge is
used with Gauge Factor 2, and the excitation voltage is 1.5V, the
strain for full scaledeflectionin the most sensitiverange(1O/V fsd),

,,rO:fi
_4x 1 0 x 1 0 -6
1 ,5 x 2
:1 3 ,3 p e (fs d )

Which gives O,266/e,/scale graduation(O to l OV scale).

87
 CHA P TE R 1O
SEM ICONDUCTORS TRA IN GA UG E S

Semiconductorstrain gauges are made from a thin strip conductor

that has been mechanically,or photo-chemicallycut from a single crys-
tal of silicon or germanium.These materialshave a high piezo-resistive
effect so that the electricalconductivityis very dependentupon the ap-
plied strain, therefore high gauge factors can be obtainedthat are 50
to 100 times greater than with metal gauges.The most common appli-
cations for semiconductorgauges are those that take advantageof the
high gauge factors available.Very low strain levels can be measured,
or high bridgeoutputscan be obtainedto drive relaysfor example.

The crystal material is prepared by "doping" silicon or germanium

with an accuratelypredeterminedamount of an impurityto give the re-
quired gauge characteristics.Semiconductorgaugescan be obtainedas
P-typeswith positivegauge factors, or as N-typeswith negativegauge
factors,

Fig. 10. 1 and F ig .1 O.2s h o w ty p i c a l re s i s ta n c echange,/strai n

curves
for a P-type and an N-type gauge respectively.lt will be observedthat
the slope of the curves (i. e. the gauge factor k) is nor a constant for
semiconductorgauges, but that it varies with the applied strain level.

Fig. I O. I Typical strain sensitivity curve for a P-type semiconductor

gauge

88
Fig.lO 2. Typical strain sensitivity curve for an N-tYpe semiconductor
gauge

SR
Ro
Typical P-type
0,10 semiconductor gau9e

0,08

0,06

0.04

o,o2

+200"
I emDeralureL
750324

Fig.l O 3. Typical temperature sensitivity of a semiconductor gauge

Fu r t her m or e,F ig .1 O3 i n d i c a te sth a t th e c h a n ge i n gaugeresi stancei s

also influencedfar more by temperaturefluctuationsthan is the case
with metal gauges, so that large apparent strains can be induced by
comparativelysmall changesin temperature.

A further complication is caused when the strain gauge crystal is

cementedonto the backing,the resistancechangesdue to shrinkagein-
duced strains as the adhesivedrys, and when the gauge is mountedon
the test specimen,the resistancechangesyet again. (lt might be noted
that this latter effect also occurs with metal gauges, but even though
the mountedgauge may be in a prestressedcondition,the gauge factor
for the metal gauge remains constant.Further, as the gauge factor for

89
the metal gauge is only one fiftieth of that for the semiconductor
g a u ge,t he c hange i n re s i s ta n c eo f th e m e ta l g a u gedue to shri nkageef-
fects will be fifty times smaller).

At the first sight of all these factors that influencethe resistanceof

the gauge - and hence its gauge factor - it might appear doubtful
whether accurate measurementscould ever be obtained, however the
si tu at ion is not qu i te a s b a d a s i t l o o k s . A s show n i n Fi g.1O.4and
Fi g . 1O . 5,f or a s in g l ea c ti v es e mi c o n d u c togr a u g eempl oyedw here tem-

Measuringerror %

Compression Tension ue

1000 2000 3000

Fig | 0.4. Typical accuracy of measurement when the nominal gauge

factor ko is used with a single active P-type gauge

Measuringerror %

8
6

Compressionte Tension !e

4000 3000 2000 1000 2000 3000

Fig I O 5 Typical accuracy of measutemenl when the nominal gauge

factor ko is used with a single active N-type gauge

90
peraturevariationsare less than 1 l OoC,use of the nominal gauge fac-
tor gives a typical measuring accuracybetter than !1oo/o at strain le-
ve l s u p t o + 2000ge , a n d d o w n to -4 0 0 0 / ,!5 o /o betw een+ 5O0ge
a n d -25009e.

For most measuringpurposeswhere very small levelsare to be mea-

sured, and where temperaturescan be kept stable, this order of accu-
racy will be sufficient. Further calibrationwill be unnecessary,and the
informationin the rest of this chaptercan be ignored.

However, where large temperature variations are encountered and

good accuracy required, several correctiontechniquescan be applied.
The best compensationis obtainedwhere a matchedpair of one P-type
and one N-type (or a double resistor P-N type) is used in a half bridge
arrangement,or two of each are connectedin full bridge.Completetem-
perature compensationis obtained,and the nominal gauge factor is ac-
curate and constantfor the whole operatingrange of the gauges.

Compensationgauges can be used in half or full bridgeconnections,

in the same way as for metal gauges,to eliminate temperatureeffects
completely.When the bridge is composedof all P-type gauges, there
will still be a constant overall measuring error of about + 3,5%. With
an all N-typebridge the error will be constant at approximately-3,5o/o.
These errors, due to shrinkage as the mounting cement drys, can be
correctedby a suitableadjustmentof the gauge factor.

Single self-compensatingN-type gauges can be obtained that have

specialtemperaturecharacteristicsdesignedto compensatefor the tem-
p e ra tur eex pans iono f th e te s t ma te ri a l .T y p i c a l l yth i s type of gaugew i l l
have a constant temperature deviation (0t) of + 5 or + EgeloC when
mounted on the specifiedmaterial. P-type gauges can not be tempera-
ture compensatedto the same degree,typically"0t" will be of the order
o f+9 to+ 1O pe/ oC.

A further method of temperaturecompensationfor single activesemi-

conductor gauges uses a constant current excitation, instead of con-
sta n t v olt age.Ref er ri n gto F i g .1 O.3i t w i l l b e s e e n that the changei n re-
si sta n c e( dR, / Ro)v ari e s fro m -O ,0 1 to + O ,1 0 o ver the temperature
ra n g e f r om about - 5 Oo to + 1 5 0 o C , w h i c h re presentsan apparent
stra i n v ar iat ionof - 1 0 0 g e to + 1 OO0 g e .Ov e r thi s range of apparent
strain the relationship6R/e can be assumedto be a constant,therefore
from the original gauge factor equationk = 6R/Re, k must be inversely
proportionalto R That is, when the resistanceR rises with tempera-

91
ture, the sensitivity(gaugefactor k) falls. A constant current appliedto
the gauge causes the excitiationvoltageto vary in direct proportionto
the gauge resistance.Thus when the temperatureincreases,the rise in
resistancecauses a rise in the excitationvoltage so that a proportion-
ally greater voltage appearsin the measuringlink and compensatesfor
the reducedsensitivityof the gauge. In this way the effectivegauge fac-
tor (- sensitivity)of the gauge can be kept virtually constant through
comparativelywide temperaturevariations.

Accurate determination of strain with semiconductor gauges

lf greater measuring accuracy is required from semiconductor

gauges, and P-N combinationscannot be used for some reason, a cor-
rection must be calculated.In addition to all the sourcesof error com-
mon to strain gauges in general that have been describedelsewhere in
this book, the nonlinear gauge factor of semiconductorgaugescontrib-
utes a major part of the measuringerror As describedin the previous
section, the great sensitivityof semiconductorgauges to temperature
can also give considerablemeasuring errors. However, when the na-
ture of these errors is understood,suitable calibrationtechniquescan
be appliedso that accuratemeasurementis possible.

The first point to bear in mind is that the gauge factor quoted by the
ma ker s hould be co n s i d e re dp u re l y a s a n o m i nal val ue for gui dance,
valid only for the temperatureat which the gauge was calibrated.This
is obviouslytrue where the characteristiccurve of resistancechange
against applied strain is of parabolicform, so that the effectivegauge
factor (the slope)is dependentupon the appliedstrain level.

When a semiconductorstrain gauge is received from the supplier,

th e f ollowinginf or ma ti o ni s u s u a l l yg i v e n fo r e a ch i ndi vi dualgauge,or
pack of gauges:

RoO i %, resistanceof the unstrainedgauge at some specifiedtem-

perature. Ro can only be measuredon an unmountedgauge con-
ductor.
k t %, nom inal g a u g e fa c to r, i .e ., th e g a u g e factor w hen the con-
ductor resistanceis Ro, which is the condition for an unstrained
conductorat the calibrationtemperature.(Herein lies the first ma-
jor sourceof error that must be compensatedfor.)
6k/ko/oper degree temperature,variation of gauge factor with tem-
perature.

92
 01pe per degreetemperature,variationof apparentstrain with tem-
perature.

From this information and a sensitivity curve like those shown in

Fi g .1O . 1and F ig. 1 O.2 ,th e a c tu a l s tra i n l e v e lc an be found. The sensi -
tivity curve should be availablefrom the gauge supplier or manufac-
tu rer .

Another typical gauge sensitivity curve is shown in Fig.1O.6 which

also indicateswhere measuringerrors can occur becauseof zero shift
caused by the cement shrinking as it drys during mounting.The sensi-
tivity curve obtained from the gauge manufacturerwill only show the
actual curve and the axes that pass through O, (marked6RlRo). lf the
resistanceof an installed gauge is measured and comparedwith the
quoted Ro value, it will be found that with a P-type gauge the resist-
ance of the mountedgauge R. is about 1O to 12o/olowerthan Ro, and
for an N-type gauge the mounted resistancewill be from 1O to 12%
higher. These changes have occured as the cement dried, and indicate
that shrinkage has induced a compressive stiain of the order of
1O00ge, (this is quite a typical value for shrinkagestrain),so that even
though no strain is being applied to the test specimen,the "zero" for
the mountedgauge has movedalong the curve to point M.

r 6R
R6
Gauge sensitivity
CUryE
Nominal k
(= slope of sensitivity curve
Shrinkage at point 0)
srratn Nominal k applied
at point M

,1 Actual k at
point M
Change in resistance
due to shrinkage Nomina! axes

Compressivestrain Tensile strain

-- -----;>
Practical ax6

7fiW

Fig.lO.6. Errors due to zero shift during mounting a P-type gauge

oe
 Th e nom inalgau g efa c to r k fo r th e g a u g e i s g i ven by the sl ope of the
sensitivitycurve where the gauge resistanceis equal to Ro and where
th e gaugec onduc t o ri s c o mp l e te l yu n s tra i n e d i,.e. the sl ope at poi nt O.
lf the nominal gauge factor is now used when the test piece is strained
th e dR, / Rt er m r eg i s te re db y th e m e a s u ri n gi n s trumentw i l l refer to the
ch a ngesin t he m o u n te dre s i s ta n c eR .. T h i s i s show n i n the fi gure by
th e nom inalk line p a s s i n gth ro u g h th e n e w z e ro strai n poi nt M, so that
measuringerrors will occur due to the differencebetween the sensitiv-
i ty c ur v e and t he n o m i n a l k l i n e p a s s i n gth ro u g h M. These measuri ng
errors are made up from two components,one is the difference be-
tw een a s t r aight " n o mi n a l s e n s i ti v i ty "l i n e a n d t he general l yparabol i c
true sensitivity"curve". This is the factor that causes the slope in the
e rro r c ur v es in F ig .1 0 .4 a n d F i g .1 O.5 .T h e o th er componenti s the di f-
ference between the slopesof the "true sensitivity"curve and the nomi-
nal sensitivitycurve at point "M". This is the differencethat causesthe
measuringertor at zero applied strain for single active gauges, or the
constant measuring error found when similar types of gauge are used
for temperaturecompensationin half or full bridgeconnections.

The only really sure method of avoidingerrors due to a parabolicsen-

sitivity curve, and errors due to zero shift, is to refer all measuredval-
ues back to the true gauge sensitivitycurve, and to the completelyun-
strainedc ondit iono f th e g a u g e c o n d u c to r.In thi s w ay the approxi ma-
ti o ns im plied in t h e a s s u m p ti o no f th e n o mi n a l gaugefactor val ue are
e l i m inat ed.

Accurate measurementwith the Strain Indicator 1526

Th e f ollowing ex a m p l ei s s u g g e s te da s a p ra c t i calmethodfor maki ng

a ccur at em eas ur e me n tsw i th a s i n g l e a c ti v e s e m i conductorgauge us-
i n g t he S t r ain I ndic a to r1 5 2 6 . S i mi l a rme th o d sc an be usedto get accu-
ra te s t r ain lev elsw h e n m u l ti -g a u g eb ri d g e sa re used w i th si mi l ar type
gauges.The method is dividedinto four stagesas follows: 1 , Determina-
tion of the resistanceof the mounted but unstrainedgauge to find the
a p p ar ent s t r ain du e to s h ri n k a g e , 2 , M e a s u rementof the i ndi cated
stra in due t o t he a p p l i e dme c h a n i c a l o a d i n go f th e speci menso that 3,
the actual appliedstrain level can be determined,and 4, Determination
o f the am ount of stra i n i n d u c e di n th e g a u g e w h en measurementsare
made at a differenttemperaturefrom the calibrationtemperature.

The manufacturersdata for the gauge used in the example is as fol-

lows:

94
 G auge:P - t y pes i l i c o n ,
N om inalgaugefa c to r.k = 1 1 2 !5 o /o
G auger es is t an c eR: o = 1 2 6 O !O,5 %
C alibr at iont em p e ra tu reto : = 2O"C
Temperaturesensitivity:6k/k = O,22Y"per oC
0 t = + 1O , 1P e P e r o C
S ens it iv it yc ur v e a t 2 O o Ca s s h o w n i n F i g .1O.8,and temperature
s ens it iv it yc ur v ea s s h o w n i n F i g .10 .3 .

1. Determination of the resistanceof the mounted gauge

As all the subsequentstrain levels measuredwill be based upon the

value obtained for the resistanceof the mounted gauge and will have
the same order of accuracy,the resistancemust be determinedwith as
much accuracy as possible,but there is no point in greater accuracy
th a n t he v alue giv e n fo r R o. W h e n a g o o dq u a l i tyohmmeteri s avai l abl e
that uses a low voltageexcitation,the resistancecan be measureddi-
mu l ti -me te ri s n ot sui tabl efor thi s pur-
re ctly .T he c om m o n ,i n e x p e n s i v e
p o s eas it s m eas ur i n ga c c u ra c yi s ty p i c a l l y1 5 % .

R6istance box in position D ldummy)

Active semiconductor gauge in position A (activel

7tu332

Fig.lO.7. Gauge and resistancebox connectionto the Type | 526 for

measurement of gauge resistance

The resistanceof the gauge on the unloadedspecimencan be deter-

mi n ed by us ing t he 1 5 2 6 a s a n o h mme te rw h e n an accuratedecadere-
sistancebox is available.A decadebox with an accuracyol O,5o/o or bet-
te r is quit e s uit ab l e .T h e a c ti v e s e mi c o n d u c to strai
r n gauge, and the
re s is t anc ebox s hou l db e c o n n e c te dto th e 1 5 2 6 as show n i n Fi g.1O.7.

1 . S et t he 1526 to " 0 ,3 V" e x c i ta ti o nv o l ta g e ,and " 2OO0ge" R ange.

2 . S et t he r es is t a n c eb o x to th e n o mi n a lg a u g eresi stanceval ue R o.
 3. Set the 1 526 gaugefactor to "about" 1O.

4. Switch the bridge mode to "Half-bridge,Off" to eliminate the bal-

ance potentiometerfrom the bridgecircuit.

5. Select "Operate". The displaywill blink when the resistanceof the

m ount edgau g e(R -) i s m o re th a n 2 % d i ffe rentfromR o. Fhe 1526
is only designedto balanceout 2o/oon its own potentiometerl.

6. Adjust the resistancein the decadebox in the direction indicated

by the display(- means decrease,+ means increasefuntil first the
dis plays t ops b l i n k i n g ,a n d th e n i n 1 O s te ps unti l the di spl aysi gn
c hanges .ln t h e e x a mp l eth i s o c c u re db e tw een 1l OQ and 1O9O
steps.

7 . Now s wit c hin gb a c kto th e p re v i o u so h m i c v al ue and then on agai n

so that the disptayalternatesbetween a small negativeand a small
positive value, adjust the "Gauge Factor" potentiometeron the
1 526 until there is a total of 1oOO/e differencebetween the two
values displayed.(The exact value of the gaugefactor is not import-
ant.) The 1 526 has now been adjusted so that it acts like an
ohm m et erwit h a re s o l u ti o no f 1 O O0d i g i ts per ohm at the gaugere-
sistance,and a measuringaccuracycomparableto the accuracyof
the resistancevalues in the decadebox.

8. The actual resistanceof the mountedgauge is given by the settings

on the decade box plus the displayedvalue. In the example these
valuesfor Rn. were:

1 0 9A + 4 1 0 p t: 1 0 9 ,4 1A
an d 1 1 0A -5 9 0 u e :1 0 9 .4 1 I

+ 7o accuracy of decade box, when the ambient temperature at the

g augewas 2O oC .

Strictly speaking,the out-of-balanceof the 1kO resistorsforming the

other side of the half bridgeshould also be checked,but as they are gar-
anteed to have less than 50 ppm differencebetween them. the effect
on the display will be very small. Nevertheless,unbalanceis checked
as follows:
9. Remove the semiconductor gauge and decade box connections
f r om t he t er m i n a l so f th e 1 5 2 6 , a n d re p l a cethem w i th the C al i bra-
t ion B r idgeZ R O 0 1 3 .

96
10. Without altering any of the other settings, check the displayed
v alue.T he posi ti o no f th e s w i tc h o n th e C a li brati onB ri dgei s uni m-
por t antas it is c o mp l e te l yb a l a n c e di n e i th e rsetti ng.

1 1 . A dd or s ubt ra c t th e d i s p l a y e dv a l u e (a c c o rdi ngto si gn) from the

gauge resistance.lt should be no morethan a few thousandthsof an
ohm . I n t he e x a mp l e th e u n b a l a n c ew a s + 1_pe, equi val entto
O,001 O which should be subtractedfrom the resistancevalue. In
this particularexample it has no significanteffect on the resistance
v aluef ound.
The differencein resistancebetween R. and Ro
6 R : 1 0 9 ,4 1- 1 2 6 : -1 6 ,5 9 A

so that
6 !:-1 9 :5 =e:-.o .r3 1 s
R^ 126
Fr om t he gaug e s e n s i ti v i tyc u rv e (F i g .1 O.8 )t he shri nkagestrai n i s
fo und t o be - 1 210 9 e = -O ,OO1 2 1 s tra i n

The slopeof the sensitivitycurve at M in terms of R,n is given by

k': l R (1O:1)
R -
dR dR R^
wh e r e (1O:2)
R .: R"" R.

k-: 2x 0,1315x126
so that
126x109,41x2500
k-: 96

This is the averagek, over a short portion of the curve having M at

rts centre.

A similar method of resistancemeasurementmay be employedusing

an accuratelycalibratedpotentiometer.

2. Measurementof the indicated strain

The knowledge of the amount of shrinkage strain there is on the

gauge, and of the actual gauge factor existing at the point where the
test piece is unstrainedopen up severalnew possibilities

97
 First, if the value k. just calculatedfor the gauge factor is set on the
1 5 2 6 ( ac t uallyk . , /1 O i s s e t).o r i f th e g a u g efa c tor on the 1526 i s set
to 1 0 and t he k m va l u e u s e d i n th e fo rm u l a

Displayedvalue x 10 : ( 10:31
microstrain
Km

the results obtained will have an accuracy(typically)better than t 5%

over the measuring range between t 1SOOpe, plus the accuracy to
which the resistance of the unstrained mounted gauge was deter-
mi n ed.T his is beca u s eth e e rro r c u rv e si n F i g .1 O.4and Fi g.1O' 5have
been movedto pass through zero while retainingtheir slopes.

For most practical measuring systems. the accuracyobtained with

this method will be sufficient. However, if a higher degreeof accuracy
is still required, one of the alternativemethods describedin section3
shouldbe followed.

3. Determination of the actual strain level

The method that has just been used to determinethe mounted resist-
ance of the gauge can be used again to find the resistanceof the strain
gauge when the test piece is subjectedto an appliedstatic mechanical
strain. A new 6R,/Rovalue can be found that yields a new strain value
from the sensitivitycurve which can be addedto the shrinkagestrain to
give the actual change in strain level causedby the appliedmechanical
stra in.
Returningto the worked example,when the strain was appliedto the
test specimen, the new resistance found by this method was
1 49. 22r ] .
Henc e 6 R :1 a 9 ,2 2 -1 2 6 : + 2 3 ,2 2 A

a nd yt :+?2-|2 : +0,184
t1 6 126

Which, accordingto the sensitivitycurve, is equivalentto a strain le-

vel of + 157Oue. Thereforethe change in strain level between the un-
strainedstate of the test piece and the strainedstate,

= 1 5 7 O te n s - l -1 2 1 O c o mp)

98
which gives a total of +2780pe change in level from the unstrained
state, i. e., the mechanical strain level applied to the test piece is
2780pe tensile, with an accuracyequivalentto the resistancedetermi-
n at ion.

The foregoing method for calculatingthe actual applied mechanical

strain can be used for measurementsup to the maximum permissible
level for the gauge, (4OOOto 5000/). However, as the method is
rather tedious, especiallywhere many strain measurementshave to be
determined, and as it is not suitable for measuring dynamic strain
either, the following method is generally recommendedfor measure-
ment of static or dynamic levels up to about 2OOOUe (with k = l OO).

Nominalk112, =slope
of snsitivitv curue at "O"
Gausesnsitivity
cu-"
+n.7 -'-
y L
!! where Ru = 10 O

+0,6 +0,4
/ :'rii,^n, =
1 I 2 applied
-,- +0,3
<. at point "B"

+o,4 True k at point "B" {=96)

+o,2

+0,3
+0,
r 550
+o,2

/,,
/l
Fig.l O.8. Construction of new axes for accurate strain measurement

99
 When the mounted resistanceof the semiconductorgauge has been
determined,the decade resistancebox or calibratedpotentiometercan
be removed and replacedin the half bridge by a resistor.

ldeally the resistanceof the dummy should be exactly the same as

the value found for R' so that it will not be necessaryto use the
built-in balance potentiometerin the 1526. When it is not possibleto
obtain or fabricate the exact resistance value required, the nearest ap-
proximation (less than 2o/oawayl can be used. The resistance of the
dummy gauge (R6) must also be known with a high degreeof accuracy
becausenow when the half bridge is balanced,the resistanceof the ac-
tive and the dummy arms will have the same value (Rs) which will be
the datum resistancefor all subsequentmeasurements.
R6 is given by,

ao: fuj!-t- a (1O:41

A new pair of axes can be drawn through point B where 6R,/Ro

crossesthe sensitivitycurve at a dR value equal to Ro - R5. fihis will
be point M when the dummy and active gauges have identical resisl-
ance = R. ). The scale of the new strain axis through B has exactlythe
same increments and size as the strain axis through O. lt has merely
been moved an amount equivalent to the strain level indicated by a re-
sistancechange of Ro - R6. The 6R/Rb axis will be scaledwith resist-
ance change incrementsreferred to R6, and these are given by a rela-
tionshipsimilar to that indicatedby equation 1O:2,

6R 6R R. (1O:5)
Thus -:- Y- -'
Ro Ro Ro

The resistor used as the dummy gauge in the worked example was
fo u n d t o hav e a r es i s ta n c eo f 1 1 0 ,5 9 Q !O ,5 o /o .
109,41+110,59
therefore Ro:
:1 1 0 ,0 0 A
dF 1 1 0 -1 2 6
Hence
Ro 126
: -0 ,1 2 7

Which is where the new strain axis crosses the sensitivitycurve at

point B. One unit of the dR,/Rbscalewill be given by.

100
 oR dF 126
nr: n" ^ tto
AFI
: 1 ,1 4 5 ;:

i . e . , t he inc r em e n tso n th e 6 R /R o s c a l e a re 1 , 145 ti mes greaterthan

the incrementson the dR,/R5scale.

Axes with these scales have been drawn through point B on the sen-
si t iv it yc ur v es how n i n F i g .1 O.8 .

The gauge factor on the 1 526 should now be reset to I O,OOand the
half bridgebalancedusing the built-in balancepotentiometer.The speci-
men is loaded with a static or dynamic strain, and an indicationob-
tained on the digital display,or output in the normal way. The quantity
displayedwill be dRlRb in units of pe/e x I O becausethe gauge fac-
tor set is 1o. The displayedvalue can readily be convertedto a resist-
ance value just by moving the decimal place.The.nthe actual strain ap-
plied to the test specimen can be read from the sensitivitycurve, re-
memberingto read from the axes passingthrough B.

Returning once again to the example,when the strain was applied,

th e dis playindic a te d+ 1 9 2 1 O ;re , w h i c h m e a n s that the strai n gauge
had undergonea resistancechange of + 1921OOpA/A (see above),so
th at dR, / Rbequal s+ 0 , 1 9 2 1 .

Applying this resistancechange to the sensitivity curve, it is seen

th a t it will be bro u g h t a b o u t b y a n a p p l i e d s trai n of + 155Oge. Thi s
measurementis accuratewithin ! 1 ,u/o, being the sum of the accura-
cies of the determinationof the gauge resistanceand of the dummv re-
si s t anc e.

Up until now, all the measurementsand correctionshave assumed

that the temperatureat the measuring point remains constant. Meth-
ods for calculationthe effect of temperaturevariationsare describedin
section4.

4. Determination of apparent strain due to temperaturevariation

The change in resistanceof a semiconductorstrain gauge that is de-

tected by the measuring instrument is made up from a component

101
caused by the applied mechanicalstrain, and from a componentdue to
temperatureinducedapparentstrain.

Total strain = Mechanicalstrain + Thermalstrain

When the strain gauge in the example had its temperature raised to
IOOoCwhile it was subjectedto a mechanicallyappliedstrain, the dis-
play indicated 17600/e. This value representsa change in resistance
of 176OOO/Q/O, so that

o R = o ,t 7 6
Rb

and R b = l l OA
th e r ef or e 6 R = l l o x o ,l 7 6 (10:6)
= 1 9 .3 5 f)

and the resastanceof the loaded strain gauge at l OOoC,

RU ro o )= l l O+ 1 9 ,3 5
= | 2 9 .3 5 f)

From the temperaturecurve Fig.1O.3,

Ro at I OOoC= Ro ot 2OoCx | 'O52

= l 2 6 x 1 ,O5 2
R{ ro o )= 1 3 2 ' 6 f)

And ..:- R _ _ t 2 9 .3 5 - 1 3 2 .6
Rotroo) | 32,6

= -O,O245

From the original gauge sensitivitycurve (with axes through O), the
6R/Ro value -O,0245 would have been producedby the applicationof
-2OO pe.

The total change in strain is the sum of the shrinkagestrain plus the
loaded strain,

= l 2 l O + (-2 O O)
= lOlOpe

102
 This value is made up from the apparent strain due to the difference
in expansion between the test object and the strain gauge (thermic
strain) and the mechanicalcomponentcausedby the appliedload.Ther-
mic strain can be determinedfrom,

,, !",!l*n* "
7;,
i;,:,:i
-_
Thereforethe strain due to mechanicalloading,

tot
'- =L1o'rou;
Again, the accuracyof this calculatedvalue will depend upon the ac-
curacywith which the resistanceswere determined.

High strain levels can be measured, if necessary,when a resistance

box is used as the dummy. lf the bridge output goes outside the range
of the 1526 (flashingdisplay),the value of the dummy resistancecan
be altered until the display again shows a constant value. The change
in the value of the dummy will give a new value to the resistanceat
the balancecondition.The new balancevalue can be found from,

Rgtug" Rau^-v
R b"bnc" - !
2
a n d it s hould be ins e rte di n E q u a ti o n1 0 :6 i n p ta ceof the 11OO, w hi l e
the value from the displaygives a new dR,/R6.

103
Bibliography
The following books will be useful for engineers who wish to make
further study in strain gaugetechniques.

Strain Gauge Instrumentation,Aronson and Nelson,

InstrumentsPublishingCompany,Pittsburg.

Semiconductorand ConventionalStrain Gauges, Dean and Doug-

las ,
AcademicPress,New York.

Handbookof ExperimentalStressAnalysis,Helenyi,
J ohn W iley a n d So n s , N e w Yo rk .

Strain Gauges,Kinds and Uses, Neubert,

M ac m illan,L o n d o n .

The S t r ainG aug eP ri me r,Pe rrya n d L i s s n e r,

, e w Y o rk .
M c G r aw- H i l l N

Dehnungsmessuerfahren, Thamm, Ludvig,Huszar,und Szanto,

W ilhelm E rn s tu n d S o h n , B e rl i n .

ConciseGuide to StructuralAdhesives,Guttman,
Reinhold.N e w Y o rk .

AdhesiveMaterials,Their Propertiesand Usage,Katz,

F os t erP ubl i s h i n gC o mp a n y .C a l i fo rn i a .

AdhesivesHandbook,Shields,
Butterworths,London.

Die Technologieder Klebstoffe, Lilttgen,

v rl a g ,B e rl i n .
W ilhelm P a n s e g ra Ve

104
 A P P E NDIX I

Derivation of expressions for the position of the principal axes

referred to a set of arbitrarily positioned axes

Cons iderF ig. Al .1 , i t s h o w s a p a i r o f a x e s O-X and O-Y w i th a

gauge length O-A = / making an angle a with the X axis. lf a strain is
applied in the X direction so that the point A moves to A', for small
changesin a, the strain in the directiona will be:
6t
e":
T
t- x
where I --
cosd
and d/: dxcosa
dxcosacosa
so that E": ---
x-
dxcos2a (A 1:1)

hence Ea: txCos2a (A 1:2)

Simila rly, if a str ain is applied in t he y dir ec t io n t h e e x p r e s s i o n

eo: evsitfa can be derived (A 1:3)

Fig.Al .l . The effect of strain in X direction

105
 Fig.Al.2. The effect of shear strain

Finally,Fig.A1.2 shows the effect of applyinga shearingstrainpxv to

the gauge length /. The strain in the directiona will be:
_6_l_
.. _ 1 _6xcos2a
""
similar to equationA 1:1 above,

now 6r: ytanBrr* y|,v (for small anglesBay)

therefore ^ - -Y\,vcOS2'a
cd-
x
but Y-: tana.
x
and tanacosa: sina
therefore eo: Brrsinacosa (A 1:4)
lf all these strains are applied simultaneously,the effect on the
gauge length / can be found by adding the components algebraically,
th us

eo: erco*al ersin2a*B,,Sinacosa this is equation2:5 (A1:51

d can be measuredwith a strain gauge, so that if measurementsare

made at three different angles, to give three different values for eo and
d. the three unknowns can be calculated by solving three equations si-
multaneously.so that the strains at the measuringpoint are defined in
terms of the arbitrary axes X and Y.

Further, if the aboveequation(A1:5) is now rewritten in terms of the

doubfe angle 2a:

106
 u:txr@, ,*r, l#)*fury?o
,":++'j9cosz"+E !?S (A1:61

and then differentiated with respect to a, the angle of the principal

planescan be found:
dt,, _-2 (e,- e,r)sin2a^ *2fl,rcos2a, :0
da2- 2
sin2-ap:
tan2ao- F,v thi s i s equati on2:6 (A 1:7)
coszap Ex- y

hence the principal strains (."* , nl;n, dod B) acting on the principal
planes, can be calculatedby substitutionback into equationA1:5, and
the stressesfound by a further substitutioninto eguation2:4.

Appendix 2 contains worked examples using these equations, and

also gives a graphicalmethod devisedby Mohr that can be used to find
the magnitudesand directionsof the principalstrains at a point'

Fig.Al.3. Triangles that contain the angle 2a

The trigonometricalrelationshipsused in equationA1:6 can be found

from equation A1:7. A triangle can be drawn that contains the angle
2q, as well as si d e s w i th l e n g th so f 9 * v a n d e' -er. Tw o such tri -
angles are shown in Fig.A1.3. lt will be seen that the hypotenusemust
be,

107
 9,,
so thatsin2a: z t/G7'l4Fj

(er-e y)
and cos2a= -+

These values can be substitutedinto equation A1:6 to yield expres-

sions for the maximum and minimum principal strains. First using the
positivevalues to obtainthe maximum principalstrain ad follows,

__ erle, (er-ey) gry

, , Er-Ey ,Fr,
:
2 2 y'1e,-erf+92,y 2t/1t,-trf+92,y

_ E x*Ey

2 2\/(e,-erf + F2,y

(A 1:8)
2

a - + a .,
Similarly Emin: - (A 1:9)
- 2
Now standard equations can be developedfor the more commonly
used strain rosettes,to give the magnitudeand direction of the princi-
pal strains in terms of the strains measured by the individual gauge
g ri d s .

The Delta Rosette

Delta rosettes,examplesof which are shown in Fig.A1.4, have three

measuring grids arranged at 600 from each other. Taking the axis of
grid 1 as a datum, t is measured by grid 1, ot <r = Oo, e2 is mea-
sured by grid 2, ot d = 60", and 3 is measured by grid 3, at d =
120o. These measuredlevels can be substitutedinto equationA1:5 to
give x, e, and their direction referred to arbitary axes, and Fry in
terms of the measured levels so that,
in plane 1,e1= e ,c o * 0 + e rs i n 2 0 + B ry s i n0cos0 (A 1:101
in plane 2, e2:e ,c o * 6 0 + e rs i n 26 0 + B,rs i n60cos60 (A 1:11)
in ptane 3, es=6,sssz120+ersin2120+B,rsin | 20 cos 120 (A 1: 1 2)

108
 1

Delta Rosettes
Fig.Al.4. TYPical

T ak ingequat ionA 1 :1 0 a n d s o l v i n gfo re ' .

E 1= E rx1 2 + e ro + B rrx 0 x 1
s o t hat e, : e,

Tak ingequat ionA1 :1 1 a n d re s o l v i n gfo r y

t 1 \/3 Vf ^ V ls r
E2: ,x tr, F,r* ' V
V* V* i" V*
69*
:7-e1 , 3eu
4-- 4
4/ E1\ / 3 I
sothat i ('r-i* 29,,I
'r:

-s - F'z
:42-t'
s fr
Now r es olv inge q u a ti o nA1 :1 2 Io r e ,
1 1 * V { .^ ' V{ , 1
Es:E,x-2" -z* tri* i+ lJ,rr-z ^ -2

e 1 . 3e, *^
=z*i-29',
4/ Et \ / g_ I
sothat'r= V ("-Z* ZP,,
I
4 re e r, 9 ,y
=-T-3--V!
109
 Equatingthe two expressionsfor r.
4t, 1 Fr, 4ts et Frv
33V 333V 3
. 4eo E t 4eo t' = 2 F y
and i- s-t* v vs
4\62 4\6s
so thafg,r:3---

2(e2-e3)
^ r r:--T
P

Substitutingthis value into one of the expressionsfor er,

4ts E1 2(e2-e3)
-v 3 3 -- V3V3

_ 4es-e1*2e2-2e3,
3
-e1* 2e21 2es
,3

SO thzt e r= s t

- e1* 2e2* 2e3

,;J

andBxy--3!#

These three expressionscan be substituted in equationsA1:8 and

A1 :9 t o y ield t he m a x i m u ma n d m i n i m u m p ri n c i palstrai ns i n terms of
e t, 2. and e3.

Fi r s tus ing equat i o nA1 :8

E,*Ey V(;=;FF,
t _" , = T --- -

2(e 2-e 2)

110
 +.+){F (3.+.E)J|I
=z.(-+. Wr

't,-''.'i.%-t (o,,,3)
t#l'
'[,,-3+-"' (A,r:.l4
Ji [+]'

Which simplifiesto

'('--#)

1
1 Gz-es)
So that a^= -L,E (A 1:15)
P =.tsn-l
2 e1* e2l e3
-r--- 3
Whilep."r--emax-min (i eA l :13-A 1:14)

=t (A 1:16)

111
 These values can be substitutedinto equations2,4 to give maximum
and minimum stresslevels.so that,

",",:,[ ,t#.+ (A1:17)

"^,,=,[f,ff + (41:18
The basic relationshipbetween the shear stressand the shear strain

Shear Stress r = Shear Strainp x Modulus of RigidityG

is similar to the relationshipfor normal stressand strain (o = E).

a fs o E : 2G (1 + p " ) (A 1:19)

so rewriting the shear relationshipin terms of E,

Ep
r :Wl (A 1 : 2 o )
and insertingthe value for B.",

,^",: (A1:2r
)
h.

The Rectangular Rosette

Typical examples of rectangularrosettes are shown in Fig.A1.5. A

Fig.A 1.5. Typical Rectangular Rosettes 7M

112
similar analysis can be made to yield maximum, minimum, and shear
strains and stresses, together with the angle that the principal planes
ma k ewit h t he dat u m g a u g eg ri d (g ri d 1 i n F i g .A 1.5).

lA1:22li

(A1:23)

oo= -(t:t) (41:241

)-trn-'3-'

(A1:25)

'^"':#\6* ["'r"*"tJ' (A'l:281

The T-Delta Rosette

Fig.A | .6. Typical T-Delta Rosettes

113
 The T-delta rosette is basically the same as the delta rosette, but it
features a fourth strain gauge grid set at right angles to one of the
other grids, see Fig.A1.6. The equationsdevelopedfor the delta rosette
can be used, or the following expressionscan be used to take advan-
tage of all four gaugegrids.

(e;e)2+ (A1:291
!{rr-rr)'

(e;e)2+ (A1:3O)
!{rr-rl'

(A1:31)
",=f,trr-'ffi
9."r-' (e;e)2+ (A1:32)
!{rr-rrf

(e ,-et)2* (A1:33)
lfr-rd'J

(e ret)2* (A1:34)
$fr-rrl'J

'e;ea)2+ (A 1:35)
tGr-rl'

114
 A P P E NDIX 2

Practical examples showing the use of the equations developed in

Appendix 1, and the construction and explanation of
Mohr's strain circle

Fig.A2.1 shows a group of three strain gauges arrangedto measure

strain at three different angles about a point in a strain field. Gauge 1
measures a strain of + 85Ope (strain being a dimensionlessratio of
length to length, the units could equally well be pm/m or pin/inl.
Gauge 2 m eas ur e s-1 OOu e , a n d Ga u g e3 me asures+ 3509e.

Fig.A2.l. Strain gauge arrangement for the worked example

Calculation method

Taking the axis through Gauge 1 as the X axis, principalstrains will

be calculatedfrom equation 2:5,

Eo: Excos2a* ersin2a*,Bxrsinscosa

the positionof the principalplanesfrom equation2:6,

tan2cto- fl'v
' Ex.- Ey

115
andthe principalstresses
from equations2:4,
E
or: -7 (ex+ PE,/)
_U2
F
ov: pe,)
T=u2(erI

First, using the values obtained from Gauge 1 in Equation 2:5,

850: e,cos20 + e,sin20+ p, ysin1cos}

q: 850 pe

Now using the valuesfrom Gauge 2,

-100:850 (-0,0872)2 * q (0,99642+ p,v @,9962)(-0,0872)

-1 00 : 6,463+ 0,9926q - 0,8688B,y
-106,463+0,08688 B,v
'v 0.9926
ey : - 107, 2+ 0 ,0 8 7 5 2 p ,y

Using the valuesfrom Gauge 3,

350: 850 (- 0,2588)2t e, (- 0,965q2+ B,y (- 0,9659)(- 0,2588)

350 : 56,94+ 0,9328e, t 0,250B,y
293,06-0,250 pxY
',: o,g32g
ey : 314, 2- 0 ,2 6 8P,y

Equatingthe two values obtainedfor y

- 107,2+ 0,08752p,y : 314,2 - 0,268B,v
(0,08752+ 0,268)p, v : 314,2+ 107,2
4 2 1 ,4
^
P'v: agsss
p ,r:1 1 8 5 p t

(This is actually equivalent to twice the height to the intercept of the

Mohr strain circle and 1 axis in Fig.A2.3, as this axis is the datum for
the calculations.)
ey: - 107,2+ 0,08752x 1185
Therefore ev: - 107,2 + 103,8
e, -- - 3,4 pe
116
 Hencethe solutionfor er, er, and F' referredto the datum axis:
e, :8 5 0 p e t E y :-3 ,4 p e : 0 ,v:1185pe

Therefore tan2ao: (2:6)

w#Aql
1185
ta n'zaP:
^ gsgr-
tan2ao: 1'39
So 2 a r:5 4 ,3 " o r2 3 4 ,3 "

(Becausethe tangent is the ratio of two positive numbers, had they

been two negatives,234,3o would have been the correctsolution.l

s p = 2 7 ,1 5 0 fro m Ga u g e 1 d a tu m pl ane

Now putting ao back into the original eguations,

E : ercos2aoI ersi n2aot B, rsin arcosa,
-u,
t.", : 850cos227
,15" + (- 3,4)sin227,1 5" * 1185sin27,1
5" cos27,15"
:
Emax 850(0,89q2- 3,4(0,450)2 + 1185(0,450)(0,890)
e^u,: 673,3- 0,69+ 474,7
eaar:1147,3 pe

.;n occUrSat 9Oofrom .", so solvingfor e,n;n = (ao t 9Oo)

e-in: 850cos2 117,15" * 11 85sin117,15" cosl 17,15'

117,15" * (- 3,4)sin2
enin: 850(- 0,45q2 - 3,4(0,890)2+ 1185(0,890) (- 0,450)
t nin: 172, 1 - 2,7 -4 7 4 ,7
e6; n: - 305, 3Pe
An d d:max-Emin

f l: 1147, 3- (-s 0 5 ,3 )
0:1452,6 Pe

lf the three strain gauges in the example had been fastened on to an

aircraft structure made of aluminium with Young's Modulus E =
TOGPa (7OOokp./mm2= 1O7 lbf/in2l and Poissons Ratio tt = O,32,
these and the measuredvalues could be substitutedin equations2:4 to
yield the PrincipalStresses.

117
 o,",=
# (.** u',,,)
o^,,=
# (^,,.u,.",)
F ir s t in I S O Unit s :

,^",= (,,u,s-o,szx30fi Gpa

# )ro-6
7 0 x 1 0 4 9 ,6 x1 O6
_
0,8976

: 81,85MPa (8,185kp/mm2)

and (-tou,t. o,32xt,or,s)rnu e,,

'^,:-- 4F-
7 0 x 6 1 ,8 3 x 1 0 -6
_
0,8976

: 4,82 MPa (0,482 kp/mmz)

Sim ilar lyin B r it i s hU n i ts :

107 /.
- m a=x
ft - 11 sos,s tb/i n2
t -0 5 2 2 l ;1 4 7 ,3 -o ,3 zx )to-6
1 0 7x 1 0 4 9 ,6 x1 0 -6
:
0 ,8 9 7 6

= 11693lb/in2

107 / 6 tb/i n2
and n =-
- m ,n I- -3 0 5 ,3 +0 ,32x 11473
t -0 ,s 2 2 , )10
1 0 7x 6 1 ,8 3 x 1 0 -6
0 ,8 9 7 6

: 6 8 8 1 b /i n 2

118
 The shear stresscan be calculatedfrom
EB (A1:20)
7:
4ft oJ2)
7 0 x 1 4 5 2 ,6 x1 0 6
ISO Unit s : GPa
2(1+ 0,32)

: 38,5 MPa (3,85 kp/mm2)

1 0 7 x 1 4 5 2 ,6 x 1 06
Bri ti s hunit s :
2(1+0,32)

55021b/in2

Graphical method using Mohr's strain circle

There is a graphicalmethod, devisedby Mohr, that keepstrack of the

relationshipbetween strains and their directionsat a point in the strain
fi e l d . lt is eas ies tto d e m o n s tra teMo h r' s Stra i n C i rcl e by means of an
actual worked example. Fig.A2.2a shows the three strain gauges ar-
ranged to measure strain at three different angles about a point in a
strain f ield as in th e p re v i o u se x a mp l e .G a u g e 1 measures+ 850ge,
Ga u ge 2 m eas ur e s -1 O OU e , a n d G a u g e 3 m e asures+ 35ope. C on-
structionof the Strain Circle is realisedin the following seriesof steps'

+ 350 re - 10O pe

",rg".
M='*'i'

A2 2a. Strain gauge arrangement A2.2b. Rearrangement of axes

Fig.A2.2. Worked example for Mohr's strain circle

119
 1 . Rearrangement of geometry where necessary, to get the axes of
measurementof all three gaugeswithin a total includedangle less
than 18O", and to place the axis line of the gauge with intermedi-
ate strain magnitude between the axes of the other two gauges.
This is only a geometricalexercise,and it should be emphasised
that it does not involve actually moving the gauges. Fig.A2.2b
shows this rearrangementof the axis lines

2. Draw a vertical axis (shown as F/2 in Fig.A2.3),and with positive

values to the right, draw a series of parallel vertical axes at dist-
ances from the first axis that correspond to the strains measured
by each of the gauges.Theseaxes are shown as 1 , 2 and e3.

^\ \54'
= 2 x27"

By msurement
e.r, = 11rl0le

min = -3fl)/

fr" = 144Orc

Fig.A2.3. Mohr's Strain Circle for the worked example

120
 3. Draw the same angle that the axis having the smallest strain
makes with the axis having intermediate-strainin the rearranged
geometry diagram (2Oo) clockwise from an arbitary point A on the
axis correspondingto the intermediatestrain value (e3). Continue
the angle until it crossesthe axis for smallest-strain(e2) at point F.

4. Similarly draw the angle that the axis having the largest-strain
makes with the axis having intermediate-strainin the rearranged
geometrydiagram(75o) from the same point A on the other side of
the axis for the intermediatestrain value (es). Continue the angle
until the line crossesthe axis for largest-strain(e1) at point B.

5. Construct a circle passingthrough points A, B and F with its cen-

tre O where the bisectorsof A - B and A - F cross.

6. Draw a horizontal axis through the centre of the circle O.

7. Measure e-", which is given by the distanceJ - C.

8. Measure -6 which is given by the distanceG - J.

9. Measure B which is given by G - C the diameter of the circle

= max - min. (strictlyF = 2 x the radiusof the circle.)

1O. Find the angle of the principalplanes.Starting at one of the points

where the circle crossesthe smallest-strainaxis (e2), at point H or
F, move counterclockwise around the circumference of the circle
through 2 x the angle between the smallest-strainaxis and the in-
termediatestrain axis shown in the rearrangedgeometry diagram
until one of the points where the circle crosses the intermediate
axis (eg) at points A or E. ln the example shown here, the only
angfe that can satisfy the requirement is FOE = 2 x 2Oo = 4Oo.
Move on round the circumference from point E through 2 x the
angle that the largest-strain axis makes with the intermediate
strain axis in the rearranged geometry diagram. The new point
should also be where the circle crossesthe largest-strainaxis (tl;
it is at point B. The two angles FOE and EOB represent the angles
of the rearranged axes in Fig.A2.2b while the horizontal axis
through the centre of the circle represents the direction of the
principalplanes.One principalplane (the maximum O - C) lies be-
tween the intermediate and the largest-strain measured axes, mak-
ing an angle EOC with the line representingthe intermediate-strain
axis, and an angle COB with the line representing the largest-

121
 strain axis in the rearranged diagram. Therefore, in the rearranged
geometry diagram, the maximum principal plane lies at an angle
EOC/2 from the intermediateaxis, and at an angle BOC/2 trom
the largest-strainaxis. The other principalplane (the minimum) lies
at 9Oo to the first. Fig.A2.4a shows the rearranged geometry dia-
gram with the principal axes drawn in. and it is a comparatively
simple matter to extrapolate back to the orginal layout to find
where the principalplanes lie. This is shown in Fig.A2.4b.

l'*';r^^

+35ope zw

Fig.A2.4a. Principal planes on rear- Fig.A2.4b. Principal planes drawn

ranged axis diagram on the original layout

Fig.A2.4. Position of the principal planes

The final results of the worked example. as measured from the Mohr
Strain Circle are as follows:

t m ax : i1140pt
enin: - 300 Pt
0' v : t 1440Pe

and ap: 27" measuredfrom the Gauge 1 line.

These results can be compared with the results obtained by the calcu-
lations shown previously.

Now having obtained results, how does the method work? Reasoning
onwards from the construction employed, the convention used for the
axes is Normal Stress (e) horizontal with positive values to the right and
negativevalues to the left. Shear Stress (Fl graduatedin units of F/2,
is drawn vertically with positive values downwards. This suits F* in

't22
equation(A1:5 = 2:51,Bv is of equal magnitudebut oppositesign. The
sign convention and arrangement adopted permit one single point on
the diagram to represent the state of strain on some plane (X in
Fig.A2.5) in a strain field, so that all points on the circumferenceof the
circle representstrains on all the possibleplanes.

In equat ionsA 1 :6 a n d A 1 :7 (= 2 :6 ) th e a n gl esgi ven are tw o ti mes

the angle between the datum plane and the plane being calculated.Us-
ing the theorem that says that the angle subtendedat the centre of a
circle is twice the angle subtendedat the circumference,the strain cir-
cle is so constructedto yield these double angles at the centre. There-
fore the principal planes are at 1 8Oo from each other where the circle
crosses the Normal Strain axis, that is, where the Shearing Strain is
zefo.

Fig.A2.5. Geometry of Mohr's Strain Circle

Further examinationof the diagram in Fig.A2.5, shows that the right-

angle trianglethat containsthe angle 2a has a base length

_ .,_ ___T_
tr * t,

fx- tv
2
a nd height : F,v/2
' - tY
so that tan2a: BxY1
22
t an l a : t 0 " , w h i c h i s equati onA 1:7 (= 2:6)
( y- ty)

123
 The hypotenuseof the triangle

)"+(fuLF

so that sinZa:S Y
{e' UEv)z+{Eu,

and cos2a:ry 6t=:ty+t%f

From the method given for the worked example,the magnitudeof the
maximum Normal Strain is given by J - C. which is equivalentto the
distancefrom the verticalaxis to the centre of the circle plus the radius
(hypotenuse)of the circle.
e' tEY ev qlfzy
th a t is 6-
" 2: a 1t'= 1z

ev
y
1e'=tv1zq10Y 1e,= 1zq p2u-y
_ Er t Ey +
2 ev p
1e, - 1z lfu-y

: "i', +
pt=:ty+ftrf 1e'=
sv
1111fzy

e'* tv - eY
e":--V- -*tx cos2o+ FlY sinza
z z

w h i c h is equat ionA1 :6 .

ExaminingFig.A2.5 once again to find where the minimum Normal

Strain occurs, another right-angletriangle containing 2a is found. This
triangle is in the third quadrantof our sign conventionso that
- F'v (0v*: - F,v)
sin2a:
Ev
1t, - 1za 1ft-y

124
 fo;:'f *t$f
From equationA1:6 it can be deducedthat the radiusof the strain cir-
cl e
: t'l t,
cos2a+ 1iv sin2a
22

but here in the third quadrant

: -" -:', 0t, sin2o

cos2o-
22

Hence the minimum Normal Strain is given by the distancefrom the

vertical axis to the centre of the circle minus the circle radius as was
measured for the worked example (G - J). This still obeys equation
A1:6 as the terms are to be addedalgebraically.

Now it can be noted that the Mohr Strain Circle is a geometricalcon-

struction speciallydesignedto contain elements that satisfy the equa-
tions used for finding principal strains. lt can be employedto find the
positionof the principalplanes and hence the maximum and minimum
Normal Strains, and the Shearing Strain present at a point in a strain
field when the strain magnitudein three directionsis known. The accu-
racy of the final results dependsgreatly upon the accuracyand skill of
the draftsman constructingthe diagram. Thereforeit is suggestedthat
the circle be sketchedfreehand, and used purely as a reminder of the
geometrical relationships between the strain components, so that the
various strains can be calculatedon a basis of these geometricalrela-
tionships.

125
 s

Kjar
"ffiBriiel&
D K-? 8 5 0 N R U M, D E NMA RK T E LEPH: ( 02) 8005oo CABL: BFUKJA, C OPEN H AG EN T ELEX: 15316
E

15- O26