4.2.2.

Large changes in resistance; potentiometers

Under this heading we shall deal exclusively with potentiometer-type
transducers. Other types with large changes in resistance, such as
semiconductor strain gauges, or thermistors, have characteristics
entirely different from metal-wire potentiometers and will be treated
separately. Our main interest will be the behaviour and design of wire
wound precision potentiometers as applied to instrument transducers,
which have a wide field of application, in spite of many shortcomings,
mainly because of their high output. Continuous-track potentiometers,
using conductive plastics, have only recently come into use [98].
4.2.2.1. Resolution and noise
The linked problems of resolution and noise are peculiar to poten
tiometer transducers and are here discussed before general performance
characteristics because of the infuence they have on sensitivity, linearity,
and associated circuit requirements. Consider a linear wire-wound poten
tiometer [3]. A cross'section of 10 turns is shown in Fig. 4.2.4(a) together
with awiper gliding over the bare wires. If the wiper could be arranged to
touch only one wire at a time the voltage resolution would be
AV= V]n, (4.2.8)
where V is the total voltage across the potentiometer, say 10 V, and in
the number of turns, say 10, the apparent resolution of such a poten
tiometer would thus be A V=1V, or 10 per cent, of the total voltage. In
106
Electrical output characteristics (4.2
ipr àcross
Wiper on n ire

Potentiometer wires

(a)

Major resolution
pulse
Minor resolution
pulse

Apparent resolution
pulse

Travel or timne

(b)
FIG. 4.2.4. (a) Linear potentiometer (schematic diagram) and (b) voltage diagram.
reality, however, the wiper is bound to short-circuit adjacent wires at
some time of its travel across the winding. This causes (n-2) minor
resolution pulses, since no change in output voltage occurs if the two first
and the two last turns of the winding are short-circuited. The total
number of resolution pulses is therefore equal to (2n-2). Fig. 4.2.4(6)
shows thbe distribution of major and minor resolution pulses along the
wiper travel- across the potentiometer of Fig. 4.2.4(a). The amplitude
AVini, of a minor pulse as we short turn x to turn (x+1), is
(4.2.9)

and of a major pulse AVma is the difference between the apparen!

resolution AV and the minor resolution,
AVma= AV-AVat=
The ratio of the duration times of the major and the
-vt (4.2.10)

minor resolunou
pulses depends upon the geometry of wire
in the first instance upon the ratio of the winding and wiper profiles,
wiper radius to the wire radiu
These considerations will be disturbed in
wire and wiper when the potentiometer haspractice by fats developing
been in use for some time.
4.2] Electrical output characteristics 107

small wiper diameter, especially, when used with relatively soft alloys,
favours the early development of flats, and may.even tear the wire track.
On the other hand, an excessively large radius only needs a small amount
of wear to short;Circuit three, or even more, turns of the potentiometer.
This results in reduced precision of the potentiometer, an effect which is
aggravated if the track is 'bumpy'. Infine-wire potentiometers of standard
design a wiper-wire radius ratio of 10 will normally give satisfactory
results. Details of wire. winding, coils and formers, and some wiper
patterns, willbe discussed in section 4.2.2.3, dealing with the construction
of potentiometer-type pick-offs. The normally bad resolution of low
resistance potentiometers can be avoided biy shunting a low-value fixed
resistor with a high-resistance potentiometer of good resolution which
will give satisfactory performance if little current is drawn from the
potentiometer-resi_tor combination.
Electric noise differs from resolution irregularities by its random
character and its peculiar frequency spectrum. Any fixed resistor is a
noise generator 4J. The electron current has a noise current superim
posed upon it, caused by the random motion of free electrons, which are
in equilibrium with the thermal motion of the molecules..This type of
noise has a uniform frequency spectrum, and its magnitude depends on
resistance and temperature only. The wider the frequency bandwidth of
the measuring equipment following the resistor, the greater the amount of
noise- transmitted. This so-called Johnson noise is- normally very smal,
except in extreme cases of high resistance, high temperature, and wide
frequency bandwidth. In carbon-composition fixed resistorS, and in some
continuous-track compound potentiometers, there is an additional type of
noise which is generated when a current is passed through the resistor. It
material and
is thought to be caused by random changes in the resistance
its characteristics. Its magnitude depends in acomplicated way on fre
on the
quency, on material and shape of the resistor, and, of course,
transmitted bandwidth. Typical noise values for fixed cracked-carbon
resistors are
Vfor 1000 resistor,
0-05 Vper
0-5 u V per V for 1 M resistor.
Wire-wound resistors are free from this type of noise.
are gener
Variable resistors suffer from further types of noise which
potentiometer track. In transducer
ated by the wiper travelling along the
noise, generated in the
Work they are of great importance. The contact
Wiper-wire contáct area, is caused by variation ofthe contact resistance.
fluctuation, espe
Contributors are contact area variation and pressure
on the track. This noise is
Clally in the presence of toreign particles potentiometer noise. It tends
Probably the most important single factor in
108 Electrical output characteristics
to increase with the lifetime of a variable resistance owing to wear
[4.2
contamination and oxidization of track and wiper. Major criticisms and ofto
potentiometer transducers are based on this progressive increase of noise
generation. Another noise source is the generation of small voltages by
the rubbing action between dissimilar metals of wiper and track. How-
ever, by suitable choice of materials this may be reduced to 100-300 V.
Thermoelectric efects which may come under this heading should also be
TABLE 4.2.2
potentiometere
Suitable combinations of materidls for wiper and wire of
Winding-wire material
Wiper material

Rhodium, rhodium plating

40% copper-palladium 80% nickel, 20% chromium
Gold-silver (Nichrome)
Osmium-iridium
10% ruthenium-platinum
Gold
2-5% graphite in silver (Nichrome
Constantan
10% graphite in copper
40% nickel, 60% silver Manganin
Gold
55% copper, 45% nickel
Gold-silver (Constantan)
Platinum-iridium
[Silver-palladium
lPlatinum-iridium
10% gold, 13% copper, 30% Silver-palladium
silver, 47% palladium

watched especially in d.c. applications at high environmental tempera

tures. Combinations of wiper and wire materials which have been found
satisfactory are given in Table 4.2.2 [5]. Finally, there is vibrational noise,
or high-velocity noise, which is caused by a jumping or bouncing move
ment of the wiper. Its magnitude, when it occurs, is comparable with the
full-scale voltage variation, since it is caused by the temporary disengage
ment of wiper and track. It must be avoided at all costs. For a given
wiper-track combination there is a maximum wiper speed which must not
be exceeded. The remedy is often an improved wiper design which should
include arevision of the contact pressure and oscillatory characteristics of
the wiper structure.
4.2.2.2. Sensitivity andlinearity
The sensitivity of an unloaded potentiometer is given in volts o
full-scale mechanical travel of the wiper. This voltage is ideally equa
the input voltage across the total potentiometer winding, though there
may be slight irregularities at the extreme ends of the track. The input
4.2] Electrical output characteristics 109
voltage is set by the dissipated wattage which causes the temperature of
the winding to rise to, but not beyond, aspecified level. Its magnitude
thus depends on the cooling conditions under which the potentiometer
has to work which, in turn, depend on the thermal characteristics of the
materials used in the pick-of, and on the transducer design, including its
shape and size. In general, the current density in the winding wire should
not exceed 10 Amm". This rough guide is applicable to small fine-wire
pick-offs wound on anodized aluminium formers. Under worse conditions
of heat conduction, e.g. for coils on formers of solid insulating materials,
the current density should probably not be greater than 5 A mm2.
In order to meet a specified linearity the minimum resolution is fixed. If
the apparent resolution, as discussed above, is n per cent of full scale, the

FIG. 4.2.5. Potentiometer circuit.

linearity error cannot be .smaller than tn per cent of full scale. Since
there are additional errors affecting linearitysuch as mechanical inac
curacies in the wiper movement, irregularities in winding pitch, variations
in wire and former dimensions, etc.in a practical design the computed
number of turns required to achieve a specified linearity should be
multiplied by 2 for coil formers of circular cross-section, and by about 3
for other than circular cross-sections, such as for phenolic.card formers.
For potentiometers of large travel, especially for multi-turn helical types,
asafe factor on the required turn number is probably 4-5. In general it
can also be said that the finer the wire the greater the factor to achieve a
specified linearity.
Apart from the above considerations which limit the 'inherent' linearity
of a potentiometer transducer in the presence of finite resolution and
noise, with loaded potentiometers there is what we might call its 'circuit
non-linearity'.This, in the first instance, is determined by the ratio of
total potentiometer resistance R to the load resistance RL. In Fig. 4.2.5,
V is the input voltage and Vo the output voltage. R, is the portion of the
total potentiometer resistance R which occurs between wiper and 'bot
tom' end of R at the potentiometer setting x= R/R. In uniformally
WOund potentiometers x is identical with the relative travel t, which is
zero for R, =0 and unity for R,=R. R' is a variable resistan ce in series
with the load resistance RL, which will concern us later on. For a simple
 110 Electrical output characteristics [4.2
80

70
voltage
m=l0
60
ideal

of 50
%
in 40
error
30
Voltage

0-5 -m=0

O-2 O"4 06 0-8

Resistance ratio X=R,/R
Relative travel t
various resistance
a loaded linear potentiometer at travel t.
FIG. 4.2.6. Percentage voltage error of resistance ratiox =RJR, or relative
ratios m= R/RL, plotted against
voltage ratio
circuit, where R'= 0, it can be shown that the
(4.2.11)
V 1+mx(1-x)
to the load
where m is the ratio of total potentiometer resistance R
resistance RL, i.e. m= RJRL. In general, if Va/V=x the voltage ratio of
the perfectly linear potentiometer, then the percentage error [ of V/V
with reference to Va/V, is

10(VV-(1--10. (4.2.12)

A family of efror curves for some resistance ratios m have been plotted
in Fig. 4.2.6 against the resistance ratio x. which is identical with tne
relative wiper travel t if the potentiòmeter is uniformly wound. t
obvious from these.curves that the load resistance R, must be at leas
say, 10-20 times higher than the total potentiometer resistance K
order to keep the non-linearity within 1-2 per-cent- of ful scale
however, this condition cannot be met, linearity can be improved in e
4.2] Electrical output characteristics 111

curye

Vea
cru r v/
ey
nea
c r
V e

Equalising

Non-linear

CLe 42.7. Non-linear potentiometer characteristic with

equalization, schematic diagram.
different ways:
(a) by a non-linear potentiometer characteristic of suitable shape;
(b) by the introduction of a non-linear variable resistance R of suit
able characteristic in series with the load R. (Fig. 4.2.5).
Fig. 4.2.7 shows schematically the first method applied to a poten
tiometer circuit of given non-linearity. Linearity is obtained [6] by means
of a potentiometer winding which follows a law given by the inverse
function of eqn (4.2.11), i.e.
R_mt-1
R 2mt {( (4.2.13)
The two curves are in mirror symmetry to the ideal straight line at 45°.
The non-linear resistanue function of egn (4.2.13) can be realized by a
card-wound potentiometer of a variable winding height h, as shown in
Fig. 4.2.8. Now it can be shown that the function h= holt) must follow
the differential curve of R/R for full equalization of the original non
linearity. Thus
h=.d(R,/R) mt-1
(4.2.14)
dt
2mt1(mt-ly+4mt
The second method employs a variable series resistance (Fig. 4.2.5) of
maximum value R'. The wiper is coupled with the wiper of the poten
tiometer, and it is convenient to make its travel of the same length as the

ho

Fio. 4.2.8. Scheme of winding card for the purpose of equalizing potentiometer non
linearities.
4.2.2.3. Construction

nnsducer work we are concerned almost

exclusively with small
straight or curved potentiometers of round or rectangular
Suaer aspects of potentiometer design and cross-section.
construction have been
dealt within Dummer's book [5] on variable resistors, which also
a comprehensive bibliography. The contains
of a
components of complete
potentiometer itself is only one of
several
actuated bythe moving mass of an
instrument transducer. It may be
accelerometer, by a pressure-sensitive
Dourdon tube, by a rotating shaft, or in many other ways. In instrument
transducers it must be oOI a preciSion type, and its design and construc
ion must be viewed in this light. In previous sub-sections of section 4.2
have discussed the conditions under which noise and
non-linearity are
educed to a minimum and resolution is an optimum. These, with the
oddition of stability in adverse environments, are the primary require
ments in transducer performance.
It is convenient to discuss the construction of potentiometers under
three headings:
winding wire and wire winding,
winding former,
wiper.
Briefly, the following are the more important points.
Winding wire for precision potentiometers is precision-drawn and
annealed in a reducing atmosphere to avoid surface oxidization. Its
resistivity may vary between 0-4 ulmand 1·3 pnm and its temperature
coefficient of resistance between 0-002per cent per °C and 0-01 per cent
per °C. (See also Table 4.2.3, p. 120, on wire resistance strain gauges.)
114 Electrical output characteristics [4.2
copper should be as small as possible,
Fine wire may have to be
The thermuelectric e.m.f. against
protected from surface corrosion by enamelling
and ductile
or oxidization. In small
it must be strong
gauges corners. The uniformity enough
of wire to
diame-
small-radius
stand winding round controlled'by BS 115 and 117, calling, for instance,
ter and resistance is resistance of S. W.G.50 (0-025
5 per cent on accordinomm
for a tolerance of finer tolerances down to ±1 per cent, w
diameter) wire, or that these tolerances are rather wide
notice wire :
grade. The reader will requirements. Higher-precision
linearity
respect to resolution and The resistance stability with time depends un
obtainable at higher cost. of the wire material
to withstand
ability ease, but i
annealing and upon the be soldered or welded with
wire should silver-soldered
sion. Resistance may be' copper-plated and
the terminal ends
difficult cases
should have a high melting-point to avoid creep at high
The wire material
working temperatures. alloys are in common use.
resistance
The following families of
alloys (Constantan, Ferry, Advance, Eureka) have
(a) Copper-nickel of resistance and a
medium-high
the lowest thermal coefficient adequate, but their ther
resistivity. Their mechanical strength is
moelectrical potential with copper is high.
Nickel-chromium alloys (Nichrome V, Karma) have high resistivity
(b) of rsistance. Their resistivity
and fairly low thermal coefficient can be obtained with a thin
depends on the state of annealing.Theywound potentiometers. They
film of oxide for insulation in closely
have a high mechanical strength and high ultimate working temper
somewhat higher resistivity and
atures. Karma is a material of than Nichrome V.
lower thermal coeficient of resistance
cheaper than nickel
(c) Nickel-chromium-iron alloys (several) are temperature and
chromium alloys but are ferromagnetic at room
have higher thermal coefficients of resistance.
resistance to corrosion and thus
(d) Silver-palladium alloys have high similar to the alloys
lower contaçt resistance, but are otherwise
under (b) above.
the wiper matena
The choice'of the wire material will also be affected by
and Table 4.2.2 should be consulted. or
oxidized
The windings of bare wire must be spaced, but enamelled must is
the wipermethod
wire can be wound touching. In this case the track for
cleaned either mechanically or chemically. The mechanical enanie
normally preferred since it is almost impossible to dissolve the under-
and
the track without softening the remaining insulation between
neath the windings.
 Electrical output characteristics 115
4.2]

(a) Wiper support

Wiper (ontat
Direction of
wiper movement

Potentiometer trak

(b)
-Wiper

Direction of
wiper movement
Potentiometer track

(c)

-Wiper
Direction of
wipcr movement 4
Potentiometer track

(a)

Direction of Wiper
wiper movement
Potentiometer track

constructions.
FIG. 4.2.11. Some wiper

discussed here in any detail. Poten

Wire-winding techniques cannot be
a specialized field and must be learned like a craft.
tiometer winding is helical
principles of some winding machines for linear, toroidal, and refer
The Dummer's book [5), and further
potentiometers are explained in
ences are given.
requirements for winding formers are form stability and high
The main are probably the most suit
surface-insulation resistance. Ceramic bodies varnish if highly porous;
with silicone
able, but may have to be treated wanted in large quantities, Some
ceramic
are rather expensive unless
they stage before firing (Steatite, Lava),
bodies can be shaped in their 'green' be produced by means of ablind
wire can
and winding groves for bare
Winding of slightly larger-diameter wire.
116
Electrical output characteristics [4.2
Some materials can be machined even in their fired state [99).
stable winding formers can also be made from aluminium which has Very
been,
anodized by a special process. This latter type of former also permits
higher current densities because of its good heat conductivity.
formers have also been made of polymethyl methacrylate Sat(Perspex),
isfactory
phenolicresins (Bakelite), and moulded epoxies (Araldite). These materi.
fniel
is are easily machinable and fairly stable. A smooth and accurate
of all coil formers is of great import an ce.
The wiper construction is frequently responsible for the noise level of
potentiometer. Suitable combinations of wiper and wire materials ar
listed in Table 4.2.2. A great variety of diverse shapes have been tried but
com.
the results obtained from tests are anything but conclusive. Some
of the winer
mon constructions are shown in Fig. 4.2.11. The flexibility
the head mass
arm has to provide the contact pressure, and, together With
high-velocity
of the wiper, constitutes a vibratory system which may cause
noise when unsuitably dimensioned. The contact pressure required de.
pends on the materials employed in the design, the ratio of wire and
wiper diameter, and the expected life of the potentiometer. Practical
values vary between 5 mN and 50mN.
4.2.3. Small changes -in resistance
This section deals with snall resistance changes due to variable strain or
variable temperature. One of the major dificulties in application is the
mutual interference between the two effects, i.e. strain gauges are sensi
tive to temperature and certain resistance thermometers are sensitive to
strain. Suppression of the one or the other effect is essential when
accurate results are expected, but since the magnitudes of both influences
are of similar order this is not an easy task.
4.2.3.1. Wire and foil strain gauges
The application of fine wire to the nmeasurement of strain was estab
lished in the U.S.A. during the Second World War but the basicprinciple
has been known for a long time [8] and results of experimental investiga
tions had been published at earlier dates 9]. In recent years the strain
gauge literature has grown beyond limits. The subject has also been
treated in monographs [10H15] in greater detail than is possible or
desirable for the purpose of this book. In the present section our aim is to
cover the elements of wire and foil strain gauges (and of semiconductol
strain gauges in the section to follow) thoroughly enough to enable the
reader to understand and assess strain-gauge performance, and to destg
4.2.3.4).
strain-gauge transducers for particular applications (see section
It will, however, be impossible to discuss all the finer points of stra
gauge design, manufacture, application, and analysis.
4.2) Electrical output characteristics 117
The modern wire resistance strain gauge usually consists of a fine
resistance wire of, say, 0-025 mm diameter which is arranged in the
form
of a grid in order to obtain higher resistances. The grid is bonded to the
bmcture under test with an insulation layer between the two (Fig.
A2.12(a)). The bonding layer transmits the surface strain in the structure
to the wire, thus producing a change of resistance in the wire, which is
proportional to strain. Since the bonded-resistance strain gauge is not an
extensometer its length does not matter in analysis and need nÍt be

Length Fine wire

grid
Terminal
(a) wires Width

Insulating layer and

bonding cement
Fine wire grid

(b)
Structure under Neutral axis
bending
FIG. 4.2.12. Wire resistance strain gauge (flat-grid type).

known accurately. In very thin structures, such as sheet metal, under

bending stress (Fig. 4.2.12(6))a correction factor may have to be applied
which is
D
e=E D+H (4.2.17)
where
E =actual surface strain,
E,=apparent strain measured,
D= distance between surface and neutral axis of structure,
H=height of gauge wire above structure surface.
AS0, since the strain-gauge grid has a finite width, the gauge has a
Sensitivity totransverse strains [16]which may amount to 0-5-2 per cent of
the longitudinal sensitivity [14]. The gauge factor quoted by the manufac
turer is normally obtained on a test rod under tension and is thus related
tO uniaxial stress involving both longitudinal and transverse strains.
118 Electrical output characteristics (4.2
4.2.3.1.1. Sensitivity and linearity. The resistance of a single long wire is
Rplla ()
where (4.2.18)
p=resistivity (m),
l=length of wire (m),
of wire (m).
a= crosS-sectional area
If a uniform stress oo(Nm) is applied to this wire along its total length
the variation in resistance is
da, dp
dR_dp!a)_p olpl2'ag
d¡ d¡
a dg' (4.2.19a)
resistance R,
or, if referred to the initial
1 dR_1 Yl 1 a, 1 ôp (4.2.196)
R' d o o a 'd¡
thus
dR_•l ôa dp (4.2.19c)
RI a p
but finite stress variation the total
Eqns (4.2.19) show that for a small
resistance change is due to:
(4.2.20a)
(a) fractional change in length, A/l;
or in wire diameter d, -Aala
(b) fraction change in cross-section a, (4.2.20b)
-2Ad/d;
(4.2.20c)
(c) fractional change in resistivity, Aplp.
With Poisson's ratjo v the lateral contraction is
(4.2.21)
Ad/d= -vAV,
and we have with egn (4.2.19c)
(4.2.22)

Eqn (4.2.22) yields the strain sensitivity ('gauge factor')

(4.2.23)
K=RIK= 1+2v+ Ape
and with Poisson's ratio v=0:3 for most metals,
egn (4.2.23) becomes
K=1-6+Aplp (4.2.24)
 4.2] Electrical output characteristics 119
It is known, from a great number of-experiments, that most materials
under elastic strain have gauge factors different from 1-6, which implies
that achange in resistivity must be involved. In the purely plastic region
of deformation of any wire material no volume change is possible
(AV=0), as the wire cannot store energy under these conditions. For the
same reason nochange in resistivity can be expected (Aplp =0). There
fore. since Poisson's ratio in the plastic constant-volume case is v=0-5,

Strain Strain Strain Strain Strain

(b) (c) (a) (e)
Iron,
Hard Drawn
408 Silver
Palladium
'FAnnealed
erry Nickel 'Minalpha
Copper, Copper
Silver,
Platinum,
10% Iridium -
Platinum.
I08 Rhodium
PMatinium

FIG.4.2.13. Typical strain-sensitivity curves.

the gauge factor of eqn (4.2.23) becomes approximately

K=1+2x0-5= 2. (4.2.25)
If some materials exhibit a gauge factor of approximately 2 also in the
elastic region of deformation, this means that the volume and the resistiv
ity changes are 'accidentally' cancelled [16]. Such a material willhave an
almost constant gauge factor over its elastic and plasticstrain range, thus
providing a wide linear stresS-strain relationship. Some annealed nickel
copper alloys are of this type.
Fig, 4.2.13 shows schematically the strain sensitivity for some'knees'wire
materials in an annealed as well as in a hard drawn state [17]. The
in these curves indicate the change-over points between essentially elastic
the simple
and essentially plastic strain. However, not all materials followunexpected
patterns of Fig. 4.2.13(a), (b), and (c). Nickel (d) has quite .an
cnange-over from negative to positive values of strain sensitivity, and
Minalpha (e) exhibits a very gradual transition period. For practical uses
ype (c) of Fig. 4.2,13 has the most attractive characteristic, covering a
120 Electrical output characteristics
(MN
m)
Ultimate
strength 8002000
1250
460 1000 400200

°C)
coefficient
of
per
resistance 0-0175
Thermal 0-002 0-010-024
+0-002 0-680-4
cent
gauges
(per

strain
Resistivity
(ulm) 0-065
resistance 0.480-45
0-451-251-031-00-62 0-1

metal Gauge
factor2-12-2 2-12-13-62-5 4-7-124-8
(
4.2.3in
TABLE
use etc.
for molybdenum,
materials

Nominal
composition
some etc..5%
chromium,
tungsten
chromium,
chromium
of copper
Characteristics
55% 20% 20%
8%
8% platinum,
nickel,
nickel,nickel,
nickel,

36%80%
45% 75% 92%

479
No.
Alloy
VNichromne
Alloy Constantan Iso-elastic
Advance Platinum
Ferry Karma Nickel
4.2] Electrical output characteristics 121

wide range of apractically linear stress-strain relationship. The strain

sensitivity is nearly 2, as would be expected from eqn (4.2.25). A fair
number of wire materials have a higher gauge factor in the elastic range
(e.g. type (a)), but their application is normally limited to dynamic
measurements because of higher thermal coefficients of resistance. Nickel
is employed in surface thermometry because of its high ratio of thermal to
strain sensitivity.
Table 4.2.3 gives a survey of the materials most commonly used in wire
and foil strain gauges [15J. Ferry and Advance are copper-nickel alloys of
the Constantan type. Apart from the wide range of almost uniform gauge
factor they have a low thermal coefficient of resistance but should not be
used at temperatures above 400°C because of corrosion. Karma has a
higher resistivity and its temperature coefficient is low at temperatures
below 150 °C. With respect to corrosion it can be used up to 1000 °C. It is
also somewhat stronger than Constantan. Iso-elastic wire has a higher
gauge factor, but because of its high temperature coefficient it is normally
employed only in dynamic strain measurements, and so is Nichrome V
wire, at temperatures up to 1200°C. The alloy No. 479 is used in
unbonded strain-gauge transducers because of its high mechanical
strength and strain sensitivity. Nominal values for nickel and platinum
have been added for comparison.
4.2.3.1.2. Construction and bonding. In the construction of resistance
strain gauges we have to consider six particular aspects of gauges and
gauge application:
(a) gauge wire material;
(b) shape of gauges, and gauge manufacture;
(c) gauge backing;
(d) cements;
(e) connecting leads;
(f) protection of gauges.
(a) Gauge-wire materials have been discussed in the previous section.
From the manufacturing point of viewit is desirable to employ materials
which can be handled with ease as a fine wire. Materials for foil gauges
must respond to the appropriate etching processes. In both cases the
gauge material must weld to suitable lead material without danger of
electrolytic corrosion.
Ine most common shapes of metal resistance strain gauges are
shown in Fig, 4.2.14. The lat grid construction is normally preferred to
the wrap-around type because the grid is closer to the surface of the
122 Electrical output characteristics
(4.2
Felt
Gauge wire Cement
tFelt
Leads
Paper
Backing
(a)
Tube, flattened Cement

Leads

sPaper
(b) Wire helix

Gauge wire (weft)

Gauge wire,
silk insulated

(c) Textile (warp)

Etched foil

Tags
(a)
Backing.
FIG. 4.2.14. Some types of resistance strain gauges. (a) flat grid: (b) wrap-around: (c)
wOven; (d) etched foil.

Ahigher
 TABLE 4.2.4
Summary of application techniques for bonded strain gauges
Gauge backing Adhesive Gauge material Remarks

Paper Cellulose-acetone Copper-nickel Useful up to 60 °C; up to

about 100 °C with increased
drift.
Non-paper Acrylate (pressure Copper-nickel Useful up to 80 °C. Limited
sensitive) life
Paper or Polyester or cold-setting Copper-nickel Useful up to 80 °C. Limited
life.
epoxy type epoxy
Phenolic Or Heat-setting phenolic Copper-nickel Up to 180 °C: for short
epoxy type or epoxy, respectively periods up to 250 °C.
Requires bonding pressure.
Nickel-chromium 400 °C and above. Check
Glass weave or Ceramic cements insulation resistance.
none (transfer (phosphate, oxide, (various alloys)
Mainly for dynaMic strains.
gauge) glass)
 (e) Connecting leads çan aBso cause instability of the gauge. The
connection point between the thin gauge wire (or foil) and the thick leads
is the most likely location of fatigue fracture. This may be brought about
by stress concentrations due to change of cross-section or by material
changes in the wire due to heating (soldering or welding), or by both. A
'dual-lead' arrangement is said to improve upon the hysteresis, creep, and
124 Electrical output characteristics (4.2
fatigue characteristics of the gauge. The choice of a suitable lead materijal
rests with the gauge manufacturer and depends upon resistivity, upon
ability to produce reliable soldered or welded joints, and upon corrosion
resistance.
(f) The main problem in the protection of strain gauges at normal
temperatures is moisture. A large number of materials have been investi.
gated .and practical methods have been discussed elsewhere [14], [19)
Arranged in rough order of permissible ambient temperatures they are:
wax; bitumen; natural, synthetic, and silicone rubber; thiocol polyester:

FIG.4.2.15. General arrangement of strain-gauge brndge.

epoxy and ceramic cements. Moisture-absorbing materials such as silca

gel have occasionally been enclosed with the moisture-proving can.