RP150

SENSITIVITY OF A GALVANOMETER AS A FUNCTION

OF

RESISTANCE

ITS

By H.

B. Brooks

ABSTRACT
Maxwell derived a theorem which states that for the maximum deflection of a
galvanometer connected to a given external circuit containing a given electromotive force the ratio of galvanometer resistance to external resistance should
be equal to the ratio of the diameter of the wire in the galvanometer coil (before
Ayrton and Perry later
insulation) to the diameter including the insulation.
stated that the galvanometer resistance should be equal to the external resistance.
This widely quoted statement

differs

from Maxwell's because Maxwell assumed

and Perry a thickness varying

of insulation, Ayrton
directly as the diameter of the bare wire.

an unchanging thickness

The present paper demonstrates that while both statements are correct for
their respective assumptions, the sensitivity function in each case is very flat in
the vicinity of the maximum; that aside from considerations of damping the
user has a wide range of choice of galvanometer resistance with relatively small
loss of sensitivity, and can in fact sometimes obtain much better sensitivity with
moving-coil galvanometers (when critical damping is considered) by departing
from the theoretical optimum value of resistance. A general curve of the sensitivity function for the Ayrton-Perry assumption is given with a similar curve for
a particular case for the Maxwell assumption. The performance of four lines of
commercial galvanometers is shown by plotted points.
The constructional limitations and those related to damping which must be
considered in applying the conclusions of the paper are outlined, and the practical
advantages of the facts brought out, to both maker and user, are briefly summarized.

CONTENTS
Page
I.

II.

III.

IV.
V.
VI.
VII.

297
299
302
303
309
311
311

Introduction
Sensitivity of the moving-magnet galvanometer
Sensitivity of the moving-coil galvanometer
Sensitivity for conditions widely different from the optimum
Damping limitations of the moving-coil galvanometer
Constructional limitations
Practical considerations
I.

The question

of the

INTRODUCTION

most suitable

size of wire for the coil of a gal-

vanometer, in order to secure a maximum sensitivity for certain

definite conditions of use, arose during the development of the electric
telegraph.

Maxwell

investigated the case of a moving-magnet

galvanometer used to measure the current in an external circuit of

resistance R, and concluded that for the maximum deflection per volt
in the external circuit the galvanometer should be wound with wire
of such size that its resistance will be G, where G is defined by a relation which in the symbols of the present paper is

G/R = D/(D + 2t)

which the right-hand member

(1)

the ratio of the diameter

the bare wire to the diameter over the insulation of thickness t.

in

Maxwell, Electricity and Magnetism,

is

2, sec. 716; 1873.

297

of
It

"

298

Bureau

of Standards Journal of Research

was assumed that the wire was

[V01.4

of uniform diameter throughout the

coil.

Maxwell 2 also investigated the best form of cross section for the
circular coil of a moving-magnet galvanometer, and the manner
in which the size of wire should be varied as the winding proceeded,
in order to obtain the most efficient utilization of the winding space.
Ayrton and Perry 3 republished Maxwell's results and determined
the condition for maximum sensitivity of a galvanometer having this
most efficient form of winding. They found that the maximum
sensitivity was obtained when the resistance of the galvanometer
was equal to that of the external circuit. Heaviside 4 examined
Ayrton and Perry's analysis and demonstrated that the simplicity of
their result did not contradict Maxwell's sensitivity theorem, nor
did it depend on the particular form of cross section of the coil, nor
on the law governing the variation of the diameter of the wire from
layer to layer, but solely upon a different assumption which Ayrton
and Perry had made, namely, that the thickness of the insulation
was in every case to be a constant fraction of the diameter of the
wire, or in present-day terms, that the space factor was to be constant.
An examination of Maxwell's analysis shows that the peculiar
result he obtained (as given in equation (1)) depends upon his tacit
assumption that t } the absolute thickness of the insulation, is constant
for all sizes of wire.
If in his preliminary equations t be taken as a
constant fraction of D, the result obtained by Ayrton and Perry will
be obtained, namely, that for maximum sensitivity the galvanometer
should be wound to have a resistance equal to that of the external
circuit.

This theorem has often been quoted. For example, in the case
wheats tone bridge having a resistance of 100 ohms in each of the
four arms the battery diagonal may be neglected for the condition of
approximate balance, and, hence, the "internal resistance" of the
bridge, which is the external resistance as far as the galvanometer
is concerned, is 100 ohms.
It is customary to say that in this case
for maximum sensitivity one should choose a galvanometer having a
100-ohm coil if the choice lies between galvanometers which are
identical in every respect except in size of wire, and, consequently,
in number of turns and resistance of coil.
While this advice can be
followed with moving-magnet galvanometers, it will, in general,
result in serious loss of time in taking readings if the galvanometers
in question are of the moving-coil type without damping frames,
because as a rule such galvanometers as now made are much overdamped when connected to an external circuit having a resistance
equal to the coil resistance. It will, however, be convenient, in
the analysis now to be presented, to assume initially that the galvanometers are either of the moving-magnet type, or that for other
reasons considerations of damping may be neglected. The manner
in which the conclusions reached may need to be modified for movingcoil galvanometers will be indicated in a later section of this article
entitled "Damping Limitations of the Moving-Coil Galvanometer.
It is often the case, in the mathematical treatment of problems of
maxima and minima, that no investigation is made concerning the
manner of variation of the dependent variable on both sides of the
of a

2
3

Maxwell, Electricity and Magnetism, 2, sees. 717-720; 1873.

Ayrton and Perry, J. Soc. Teleg. Engrs., 7, pp. 297-300; 1878.
Heaviside,

J. Soc.

Teleg. Engrs., 9, pp. 202-206; 1880.

Galvanometer Sensitivity

Brook*]

299

maximum or minimum. Such incomplete treatment of the

problem is apt to be very misleading, and one should not blindly
adopt the optimum condition at the expense of other qualities without first investigating to see what happens for a considerable range
of values of the independent variable on both sides of that which

point of

gives this optimum condition. 5

The object of this paper is to show that while the theorems of
Maxwell and of Ayrton and Perry concerning what may be called
the working sensitivity of a galvanometer are theoretically correct,
there is nevertheless a great deal of latitude in the choice of the
resistance of the galvanometer, with only a relatively small resulting
The manner
loss in sensitivity, and with some material advantages.
in which the user may take advantage of this fact will be indicated. 8
II.

THE MOVING-MAGNET
GALVANOMETER

SENSITIVITY OF

There are various ways of expressing the sensitivity of a galvanomwhich ultimately relate to its current sensitivity; that is,
its response (deflection) for unit current in its coil.
There are cases,
such as in the measurement of insulation resistance by the directdeflection method, where the external resistance is so high that the
current sensitivity is the important property, but the scope of this
paper is limited to cases where the resistance of the external circuit
does not differ from that of the galvanometer by more than s&j a
few orders of magnitude. For such cases the user is concerned with
what might be called the " working microvolt sensitivity," but which
will be called for brevity the " working sensitivity" in the following
discussion.
The working sensitivity applies only to the user's particular problem at the moment, and may be defined as the response
of the galvanometer for unit electromotive force in a circuit which
includes the galvanometer and a particular external circuit. As
illustrations may be cited
(a) The measurement, by direct deflection,
of a small thermal emf. set up in a thermocouple of given resistance;
eter, all of

(b) the detection (and reduction to zero) of the small electromotive

force across the galvanometer diagonal of an unbalanced wheatstone
bridge.
The moving-magnet galvanometers in use in Maxwell's time had
coils of circular form in which all portions of the length of any given
turn of wire were equally effective in setting up a magnetic field at
the center of the coil. For simplicity this form of galvanometer will
be assumed initially in this treatment of the problem, and the case
of the permanent-magnet moving-coil (d'Arsonval) galvanometer will

be considered later.
1 An
example, although unrelated to the present case, will serve to demonstrate this point. Suppose
that it is desired to make a metal "weight" in cylindrical form. It may easily be shown that to have
the necessary volume of metal with the desirable minimum of exposed surface the diameter of the weight
should equal its height. However, other considerations may call for a different form with the height
either much smaller or much greater than the diameter. It may be shown that when the diameter is
twice (or one-half) the height, the volume-surface ratio is only 5 per cent below its desirable maximum
value, and for the diameter equal to four times (or one-fourth) the height the ratio is still within 13 per
cent of the maximum. It is obvious that one does not have to sacrifice much in other requirements to
obtain virtually the full benefit of the optimum form of weight.
That galvanometer makers and users do not realize how much latitude is possible is shown by statements in catalogues and text books. For example, a recent (1928) textbook on electrical measurements
states that "In the choice of a galvanometer for a definite measurement one should be careful to have the
resistance of the galvanometer as nearly as possible of the same order of magnitude as the resistance of
the measurement circuit." An equally recent catalogue states: "It should be remembered that for a
given constant electromotive force in circuit, the resistance of the coil should, if possible, equal all the
resistance of the external circuit, including the suspension."


300

Bureau

of Standards Journal of Research

Referring to the circular coil shown in cross section in Figure

Z = mean

length of turn; that

[V01.4

1, let

is, total length of wire divided

of turns.
diameter of bare wire.
p = volume resistivity of material of wire, assumed constant.
t = thickness of insulation.
number of turns of wire in the coil.
= resistance of the coil.
=4 resistance of external circuit.
electromotive force acting in the circuit.

by

number

N=
G

E=

Then

the resistance of the coil

is

G^lNLp/irD*
and the current

(2)

is

I=E/(R + G)
= E/(R + 4NL p /tD 2 )

(3)

For any given shape and given cross section of the

coil, and any number of turns in it, the magnetic
field strength at its center (or any point within
or near the coil) will be proportional to the ampere
Since the deflecting torque exerted upon
turns. 7
a given magnetic needle within or near any coil
directly proportional to the strength of
the magnetic field set up by the ampere
turns in the coil, we can consider the
deflection of the galvanometer as proportional to the ampere turns.
In a discussion of relative sensitivity, when two or
more coils which differ only in number of
turns are concerned, the deflecting torques
is

H^^
^

Figure

1.

Galvanometer

coil

connected to an external circuit containing an electromotive force

themselves need not be known, but may

be represented by the corresponding values of ampere turns, to which the torques

are directly proportional.

and divide by y
sides of equation (3) by
the result will be an expression for the ampere turns per volt, which
is directly proportional to the working sensitivity of the galvanometer
for the case of the particular value of resistance
in the external
circuit.
Denoting this quantity, which is a function of the size of
the wire, by r we have
If

we multiply both

F= NI/E
= N/(R + G)
= N/(R + 4NL P/irD 2 )

(4)

7 This is true for a coil of any form having a cross section which may be the same everywhere or may
change in form and dimensions from point to point in any manner, provided the current distribution over
a plane intersecting the cross section anywhere is not changed when the number of turns is changed. For
a constant value of ampere turns such a plane will obviously cut the same total current regardless of the

number

of turns.

301

Galvanometer Sensitivity

Brooks]

and D are not only interdependent, but depend

on t, the thickness of the insulation. We may eliminate N and introduce t by using the fact that the volume V occupied by the winding is
constant for all sizes of wire and thicknesses of insulation and is
In this expression

V=NL(D + 2t)

(5)

from which

N=V/[L(D + 2t)*\

(6)

in equation (4), and simplifying,

Substituting this value of
a general expression for the ampere turns per volt, namely,

F= LR(D + 2t) /V+4 L/irD

P
2

we

get

(7)

which the only variables are the diameter of the bare wire D, and
the thickness of insulation t. Such a general relation would be
represented by a surface. Since in wire-manufacturing practice there
are fairly definite relations already established between the diameter
and thickness of insulation t, we may make
a function of a single
variable by making t a constant or by introducing a suitable relation
between
and t. For example, in the case of silk-covered wire the
thickness of the insulation is usually constant regardless of the diameter of the bare wire.
Differentiating equation (7) on this basis, we
in

have

2LR(D + 2t)/V-S PL/wD3

dF

dD~

(8)

[LR{D + 2t) 2 /V+4:pL/irD 212

condition giving a maximum value of F is obtained by equating

the numerator of the right-hand member of equation (8) to zero, from

The

whick
4:

that

P /<irD*

= R(D + 2t)/V

is

4 P V/[<irD*(D + 2t)]

which on substituting the value of

G from equation (2) becomes

=R

V from

equation

(5)

and that

of

G(D + 2t)/D = R
or

G/R = D/(D + 2t)

(9)

is identical with the expression found by Maxwell for the case

constant for all sizes of wire.
As a second relation, which may be taken to hold with sufficient
approximation for the sizes of enamel-insulated wire used in the construction of moving-coil galvanometers, the thickness of the insulation may be assumed to be directly proportional to the diameter of
the bare wire that is,

which
of

When

this value of

is

F==

= aD

substituted, equation (7)

LRD

(10)

becomes

(l+2a) 2 /V+4:pL/TrD 2

(11)

302

Bureau of Standards Journal

of Research

won

from which

+ 2a) 2LRD/ F- SpL/irB*

lLRD (l + 2ay/V+*pL/irD 2

dF

2(1

dD~

(12)

Equating the numerator

value of the function

to zero we have the condition for a

this case, namely,

F for

Ap/ttB3

maximum

- (l + 2a) RD/ V
2

from which
4pV/[irD\l
Substituting for Fits value

NLD

+ 2a) = R
2

(l+2a) 2 there
,

results

4NL p/tD = R
2

that

is

O-R

(13)

which is the result found by Ayrton and Perry after introducing the
condition that the thickness of the insulation bears a constant ratio
to the diameter of the bare wire.
For the sake of having a simple form of coil in mind the preceding
analysis has been referred to the simple circular coil of rectangular
However, the conclusions are applicross section shown in Figure 1
cable to a galvanometer having a coil of any form with a cross section
which may be either of the same form everywhere or may change from
point to point. This follows from the fact already stated, preceding
the writing of equation (4), namely, that the magnetic field at any
point within or near a coil of any form whatever will depend solely
upon the ampere- turns regardless of the number of turns in the coil.
.

III.

SENSITIVITY OF

THE MOVING-COIL GALVANOMETER

In the preceding discussion every portion of the length of the wire

composing the galvanometer coil was effective in setting up a magnetic
field at the center of the coil upon which the deflection depends.
In
(d'Arsonval) galvanometer,
the permanent-magnet moving-coil
which is the favorite instrument at the present time for all but very
special work where the sensitivity must be pushed to extreme values,
the galvanometer resistance is made up of three parts; (a) that of the
elements of the coil which may be considered as active, that is, those
parts of the vertical sides which cut the flux of the permanent magnet
and provide the torque; (b) that of the top and bottom parts of the
coil which do not cut the flux and which may be considered as serving
merely to connect the active wires in series; (c) the resistance of the
suspensions. 8 Since the suspensions are assumed to remain unchanged
as the size of wire in the coil is changed, their resistance may be taken
as constant.
Obviously, this resistance should be counted in as part
of the external circuit.
This procedure is particularly important for
galvanometers having coils with a resistance of only a few ohms, in
which case the resistance of the suspensions may be several times the

Many

galvanometers of the moving-coil type have the moving coil supported by pivot-and-jewel
and have spiral springs which serve the dual purpose of carrying current to and from the coil
supplying the counter force. For example, many pyrometer galvanometers used with thermocouples are of this description. It will be understood that whenever the word "suspension" is used in this
article the use of spiral springs as an electrical and mechanical equivalent is understood

bearings,

and

of

Galvanometer Sensitivity

Brooks]

303

It might seem at first glance that the resistance

resistance of the coil.
of the "inert" horizontal wires at the top and bottom of the coil
ought also be counted as part of the external resistance, but it will
be shown that this is not the case.
In order that the preceding reasoning for the moving-magnet gal-

vanometer may be applied to the moving-coil galvanometer it is

only necessary to show that in the latter the torque varies directly
as the ampere turns in the coil, just as in the former the magnetic
field at the center of the coil, and, consequently, the torque on the
moving magnet varies as the ampere turns. In the moving-coil
galvanometer the torque T is equal to the product of four factors;
that is

T=H-I-(2Nh).b/2

(14)

in which
is the magnetic field intensity in the air gap, / the current,
2 Nil the total length of active wire, made up of two vertical sides of
each of the
turns of effective height h, and 6/2 is one-half the
effective breadth of the coil; that is, it is the average distance from the
individual vertical wires to the axis of rotation of the coil. Equation (14) may be written

T=HKbxNI

(15)

= Constant X ampere

turns

In this expression the coefficient of NI is a constant for the given

galvanometer regardless of the size of the wire of which the coil is
wound. Hence, all the relations previously derived for the movingmagnet galvanometer apply to the moving-coil galvanometer also,
if the resistance of the suspensions be included as part of the external
resistance.

IV.

SENSITIVITY FOR CONDITIONS WIDELY DIFFERENT

FROM THE OPTIMUM

It is not possible to obtain stock galvanometers in closely graded

steps of coil resistance and it would be necessary to draw special
sizes of wire to produce them, as may be seen from the fact that in a
series of coils of constant space factor wound from consecutive A. W.
G. sizes of wire the resistances of any two adjacent coils in the series
would be in the ratio of 1.6 to 1. It will be of interest, therefore, to
see what happens when the resistance of the galvanometer varies
from a value very much lower than the optimum value to one very
much higher, and having thus discovered how much latitude is possible without serious loss of sensitivity, to consider in what way this
latitude may be turned to practical advantage.
In order to do this
conveniently, a relation will be developed from the preceding equations, as follows:
The general expression for the function, ampere turns per volt,
which is proportional to the working sensitivity, is

F=N/(R + G)

(4)

This relation holds without restriction as to the ratio of the diameter

of the wire to the thickness of insulation.
From this point on one

must

select

some

8850030

definite relation

between these quantities.

304

Bureau

of Standards Journal of Research

[vol 4

Taking the case of thickness of insulation directly proportional to

wire diameter; that is, a constant space factor, from equations (4)
and (13) we have that

Fmzx = N /2B

(16)

where Ni denotes the particular number of turns in the coil of resistance Gi = R, which gives maximum sensitivity when used with an
external circuit of resistence R.

Dividing equation

F
N
Fm&x R + G
2R
R+G

(4)

by equation

2R
'

,_
{i7)

N^
'

Nt

Since the resistances of two coils wound in identical winding

spaces, using wires of the same space factor, are to each other as the
squares of their numbers of turns,

miN^GIGr
and since by equation

(13)

G,
it

=R

follows that

N/N, = JGJR
Introducing this value into equation (17),

F _

it

becomes

f\^x ~- tJG/R + jR/6

(18)

which can also be written in a form more convenient for computation

namely,

F~m~~-y/G/R

+ I/JG/R

(19)

first term of the denominator has been computed its reciprocal gives the second term.
This function has the value unity
when G = R. It is plotted as curve
in Figure 2, in which the ordinates are values of F/Fmax plotted in per cent of maximum sensitivity,,
and the abscissas are values of the ratio of galvanometer resistance
to external resistance.
The curve has such a flat maximum that it.
is necessary to plot the abscissas to a logarithmic scale in order to
show the behavior of the function over an adequate range of relative
values of galvanometer resistance to external resistance. 9
As an example of the use of curve (fig. 2) let the external resistance
be 100 ohms, so that a 100-ohm galvanometer is indicated as the most
sensitive.
It may be seen that one can use a galvanometer of any
resistance between 20 ohms and 500 ohms and still obtain a sensitivity
not less than 74 per cent of the maximum.

After the

9 After this paper was written the

author discovered that an equation identical in substance with equation (18) had been given by Schuster in a paper entitled "Electrical Notes," in Phil. Mag. 39, p. 175, 1895.
Although Schuster does not so state, his reasoning applies only to galvanometers wound with wires of a
constant space factor. In view of the fact that the useful information given by Schuster seems to have

escaped general attention, possibly because of the rather indefinite title of the paper, it appears desirable
to point out again the relationships involved in greater detail and to extend the treatment to the case of a
constant thickness of insulation as exemplified by silk-covered wire.


305

Galvanometer Sensitivity

Brooks]

'

IxJ

o
<

to

^3

CI

sa
p o
*
feS

S.

as

a
>r-

cc

^A

>

s&

9.

-J
<" o ^
-

SOB
> o

2^3

cc

J*

ess

LiJ

1-

Q,

PQ

a>

LU
CD

'

a5
S5

LlJ

/
/
c

h-

5gg
*-

co

\ \\

CO-

LLI
* o'O
ca-c

CC

DO

LuJ

1-

**0

"c3

...

C^

\s

<Pr\

Sal

2
o

sal

<"

ot fe

S3

-rt

*2S

O
O
<
CC

t\

fI

AXIAlllSNi?

iI

5i

I\

s\

a S3

V
1

ss

SoS

>

$I

WOWiXVW dO XN303d

8:

306

Bureau of Standards Journal

of Research

[Vol. 4

The plotted points of curve

(fig. 2) represent computed values of
relative sensitivity of actual galvanometers for which representative
average values of coil resistance, current sensitivity, etc., were kindly
supplied by the makers. Four groups of galvanometers are included,
nominally differing only in the resistance of the coil. For each
group one galvanometer, of an intermediate value of resistance, was
taken as a basis of reference, and its deflection for a given voltage in
an external circuit having a resistance equal to its own resistance was
Point 1 is thus common
plotted as point No. 1, namely, 100 per cent.
The deflection which each of the other galvanometo all the groups.
ters of the group would give when connected to the same external
circuit with the same electromotive force acting was then computed
as a percentage of the deflection given by the reference galvanometer
and plotted as a point for which the abscissa is the ratio of galvanometer resistance to external resistance.
The following table gives particulars concerning the galvanometers
Table
Maker

1.

Group
designation

Coil resistance

Plotted
points

Catalogue

Ohms

No.
Leeds

&

Northrup Co

2420
2260
2310
41221

4.5

16
20
15
10

312
85
250
100

1,100
1, 340
1,100
1,000

2,1,3
4,1,5
6, 1, 7,

9, 1,

10

The galvanometers of the first three groups are of the permanentmagnet moving-coil type. The Cambridge galvanometers are of the
Broca moving-magnet type. The Leeds & Northrup galvanometers

wound with

enamel-insulated wire for which the thickness of the

very nearly a constant fraction of the diameter of the bare
wire.
The coils of the Cambridge Instrument Co. galvanometers are
made in England, and it has been assumed in this paper that the wire
used in them has a constant ratio of thickness of enamel to diameter
of bare wire.
In the manufacture of galvanometers it is not found necessary to
are

enamel

is

adjust the strength of individual magnets and the torsion of individual

suspensions to standard values. It is sufficient to control these quantities so that at least a certain stated minimum value of sensitivity is
attained.
The failure of some of the points of Figure 2 to He on or
near curve A is the result of these manufacturing variations. From
the maker's data on the number of turns in the coils and the current
sensitivity it has been possible to compute some values of relative
sensitivity corrected to the basis of a standard strength of magnet and
a standard torsion of the suspension. The corresponding points are
denoted on Figure 2 by primed numerals; thus 4' denotes the relative
sensitivity of catalogue No. 2260 galvanometer of 20 ohms resistance
reduced to the basis of the same strength of magnet and torsion of
suspension as the 85-ohm galvanometer of this type, which is the reference galvanometer of this group; and similarly for the other points

and 10'.
Although the greater space factor of the enamel-insulated wire is an
important advantage in the case of moving-coil galvanometers, particularly for the finer sizes of wire, there are instances where single-silk
5', 9',

307

Galvanometer Sensitivity

Brooks]

For moving-magnet galvanometers of relatively

insulation is used.
low resistance (say 10 ohms or less) the space factor of the single-silk
and the enamel-insulated wires may be so nearly equal that other conThe relative working sensitivsiderations may determine the choice.
ity of galvanometers wound to have various resistances, using wire
with a constant thickness of insulation, will now be discussed. It is
assumed that the galvanometers are identical in all respects other than
the size of wire in the coil.
Returning to equation (9),. and denoting by
and G the particular
and of G which give maximum sensitivity, we have
values of

G /R = D
l

The length
tion being

(/D

+ 2i)

(20)

of wire in the coil, the diameter of the wire over the insula-

D + 2t,
x

is

N L=V/(D
1

+ 2tY

(21)

is the particular value of

corresponding to
The
x
resistance of the coil is obtained by multiplying this value of the
length by p and dividing the product by the cross-sectional area of
the conductor; that is,

where Ni

^ = 47p/[xA
Substituting this value of

(A + 20

in equation (20)

(22)

and simplifying,

B + 2tD = 4:Vp/7rR
z

which

is

(23)

not a general equation, but holds only for the diameter of

wire giving maximum sensitivity.

While a solution of equation (23) by analytical methods is possible,
the procedure is laborious, and for practical purposes a graphical
method is much simpler and quicker. The curve (fig. 3) has as
abscissas values of optimum diameter of bare wire
ly and as ordinates values of the function
+ 2tDi3 using for 2t its usual value
i
for single-silk covering, namely, 0.002 inch (0.05 mm.).
For a given
external resistance R, and a series of galvanometers which differ only
in the resistance of the coil, that is, only in the diameter
of the
bare wire, the value of the quantity 4 Vp/irR, the right-hand member
of equation (23) can be computed.
The value of the abscissa x of
the curve (fig. 3) corresponding to this value of 4Vp/irR as ordinate
can then be found directly.
Returning to the general expression for F, the ampere turns per
volt, to which the working sensitivity is proportional, namely,

4:

F=N/(R + G)

(4)

using the symbols Fmax Ni, and

the prex in the same sense as in
ceding discussion, and expressing the coil resistance G in terms of
,

and x by equation (9), it may be shown by an analysis similar to

the preceding that the equation for the relative sensitivity for the
case of a constant thickness of insulation takes the form

F
Fmax

(A +

2R

R+

4,

VpHirD 2 (D + 2t) 2

(A + 2f)

(D + 2tf

(24)

of D for any given case by using the abovementioned curve, the right-hand member of equation (24) contains
only one variable, D. Values of F/Fmax may then be computed for a

Having found the value


308

Bureau of Standards Journal

number

of Research

[Vol.4

of values of
above and below the particular value
x for
sensitivity, and a curve may be plotted giving

maximum working
100

so

/
/

/
//

4Vp

OJ

0.01

/
/

1
I
OlOOJ

10

12

DIAMETER OF BARE WIRE

14

IN

18

20

22

24

26

28

THOUSANDTHS OF AN INCH

Figure

3.
Curve for the graphical solution of the biquadratic equation (23)
for the optimum diameter of wire for the case of a constant thickness of
insulation, namely, 0.001 inch

For a given winding volume

(cubic inches), resistivity p (numerically equal to the resistance between opposite faces of a 1-inch cube) and external resistance li ohms, the curve is
entered with the product 4Vp/irE. The corresponding ordinate has as abscissa the diameter of
the wire.

values of relative sensitivity

4Vp/[wD 2 (D + 2t) 2 ], values

F/Fma x
of

as a function of the coil resistance

for the various chosen values

which

Galvanometer Sensitivity

Brooks]

309

being obtained during the process of computing the values of

F/Fmax from equation (24).
Curve B (fig. 2) has been computed for Cambridge Instrument Co.
Broca galvanometers on the assumptions that the coils are wound
with single-silk insulated copper wire and that the external resistance
Comparing the two curves of Figure 2 it is seen that
is 100 ohms.
the maximum working sensitivity, using silk-covered wire, would be
about 90 per cent of the maximum for the 100-ohm Broca galvanomThe maxieter as it is actually wound with enamel-insulated wire.
mum working sensitivity, using silk-covered wire, occurs when the
galvanometer resistance is a little over 80 ohms as required by
Maxwell's theorem, equation (1). The two curves intersect when
the ratio of galvanometer resistance to external resistance is 0.062;
that is, when the resistance of the galvanometer is 6.2 ohms for either
kind of insulation. This is the case where the silk and the enamel
have the same thickness. For galvanometers of lower resistance
curve B lies a little above curve A, which means that for the coarse
wires used in these coils the silk is thinner than the enamel. 10
Going to the right of this point of intersection the ratio of the working sensitivity of the galvanometer wound with silk-covered wire to
that of the galvanometer of the same resistance having enamelinsulated wire decreases continuously, reaching the value 0.54 when
the galvanometer resistance is 10,000 ohms.
of

V.

DAMPING LIMITATIONS OF THE MOVING-COIL

GALVANOMETER

Although the

effect of the resistance of the external circuit on the

of little or no consequence in the use of moving-magnet
galvanometers it is a very important matter when galvanometers of
the moving-coil type are concerned.
In these latter the following
cases arise in practice:

damping

is

The moving coil may be wound on a closed metal frame, or

be provided with a second closed damping winding which has
no connection to the external circuit. One form of this auxiliary
winding consists of a single closed turn of copper wire, and is sometimes
spoken of as a " damping rectangle." The use of this expedient in
any form makes it possible to construct (or alter) a galvanometer so
(a)

may

that the external critical damping resistance may have any desired
value (above a certain minimum), even infinity; that is, the motion
of the coil may be made aperiodic when its terminals are not joined
by any conducting path. A well-designed direct-current voltmeter
is a common example of this case.
In it the moving coil is wound
on an aluminum frame which provides nearly all the necessary

damping.
(b) The damping winding or rectangle of plan (a) has the disadvantage that, being of service only while the coil is in motion, it
adds otherwise needless weight and moment of inertia to the moving
system. Consequently, when high sensitivity and shortness of period
are desired, it is customary to avoid the use of damping coils, in which
case the kind of motion of the coil and, therefore, the time required to
obtain a reading depend very much on the value of the external
10

This

is on the assumption that the thickness of the enamel is actually a constant fraction of the diameter
While this is very nearly true for the sizes of wire used in the moving-coil galvanometers of
the practice of various makers of enamel-insulated wire is not uniform.

of the wire.

Table

1,

310

Bureau

of Standards Journal of Research

[von

Many users prefer a slight amount of underdamping,

such that the coil passes the point of final deflection by a small amount
and then returns to this point without further observable oscillation.
This slight overshoot gives the assurance that neither overdamping
nor friction ("sticking") is present.
One point of considerable importance to the galvanometer user
is that if the criterion be the time required to obtain a reading a
much wider tolerance is possible for the case of underdamping. This
may be illustrated by the results of an experiment made on a pivoted
pointer galvanometer having a frameless coil. The period of oscillation on open circuit was 2.6 seconds and the time to come to apparent
resistance.

rest

when

critically

damped

3.4 seconds.

reduced below the value for

critical

As the

total resistance

was

damping, the time to come

to
rest increased sharply, reaching a value of 8 seconds when the total
On increasing the
resistance was two-thirds of the critical value.
resistance beyond the critical value the time to come to rest became
5, 6, and 8 seconds for a total resistance twice, three times, and four

times the critical value. 11

The important conclusion to the galvanometer user from this
experiment is that if he must choose between underdamping and
overdamping the former will give him much greater latitude in the
total resistance for a given limit on the time to obtain a reading.
Of this total resistance the external resistance is the part with which
the user is usually more concerned. It sets a limit to the "internal
resistance" of the bridge, potentiometer, thermocouple, or other
device to which the galvanometer is connected. Assuming that
critical (or slight under) damping is required the user who has already
acquired a given galvanometer is usually concerned very little 12 about
the resistance of its coil, but very much concerned with the value of
external critical damping resistance. If the resistance of the apparatus to be connected to the galvanometer is less than the given value
of external critical damping resistance a rheostat in series with the
apparatus may be used to bring up the external resistance to its value
The working sensitivity will thus be constant
for critical damping.
for all values of apparatus resistance from zero to the value for
critical d amping.
If the resistance of the apparatus is greater than the
external critical damping resistance, the galvanometer must be shunted
to obtain critical damping, and it may be shown that if the external
circuit has a resistance n times the critical value and the galvanometer shunt has the proper resistance to produce critical damping, the
working sensitivity will be reduced to 1/n of its former value. 13
If, however, a galvanometer is to be selected from a number
which differ structurally only in size of wire and number of turns
in the coil, and, consequently, differ electrically in coil resistance
and external critical damping resistance, the analysis in the earlier
Such a result follows from the laws governing the motion of a galvanometer. The figures given in the
above example should not be construed to apply generally because one must decide on a definition of
"coming to rest." For example, if the motion of the coil is exactly aperiodic it reaches its theoretical rest
position only after an infinite time. However, it reaches this position, within 0.1 per cent of the deflection,
in about 1.5 T, where T is the undamped free period. For underdamped and overdamped motion also the
coil theoretically never reaches a condition of absolute rest.
1 3 An exception to this
statement must be made when one is concerned with the effect of ambient temperature variations on the indications of a galvanometer which is connected to an external device; for
example, a thermocouple. In this case there is a particular value of the temperature coefficient of resistance
of the entire circuit which will offset the temperature coefficient of torsional rigidity of the suspension. The
high temperature coefficient of resistance of the galvanometer coil makes the resistance of the coil a matter
of importance if independence of variations of ambient temperature is sought.
" This is shown graphically for a particular numerical example by fig. 1, p. 224, of a paper by Wenner on
General Design of Critically Damped Galvanometers; Bull. Bur. Stds., 13, 1916; Scientific Paper No. 273.

Galvanometer Sensitivity

Brooks]

311

part of this paper shows, first, that one can to a large extent ignore the
traditional statement concerning the condition for maximum working
sensitivity, and being free from this supposed limitation may select
a galvanometer having an external critical damping resistance most
nearly equal to the resistance of the apparatus with which it is to be
As moving-coil galvanometers are now made this would
used.
usually mean choosing a galvanometer having a net coil resistance of
say from one-third to one-tenth of the resistance of the connected
apparatus. Such a galvanometer will require a minimum amount
of reduction of sensitivity to effect the desired kind of damping and
may easily have several times the working sensitivity of a similar
galvanometer chosen according to the traditional rule, but requiring
an excessive amount of series resistance to effect the desired damping.
It may be pointed out that for some applications a heavily overdamped galvanometer is desired in order to smooth out transitory
fluctuations and give an average value.. The preceding analysis
shows that for this case also one may largely disregard the traditional
rule and select a galvanometer having a net coil resistance several
times the resistance of the external apparatus, and thus secure the
desired overdamping with relatively small loss in working sensitivity.
In so doing it should be kept in mind that the effect of changes of room

temperature will be relatively large. If independence of roomtemperature variations is necessary, the resistance of the galvanometer
u of the total resistance and the necessary
coil must be only a fraction
overdamping must be obtained by winding the coil on a damping
frame.

The use of a damping frame, coil, or rectangle (case (a)) is usually

confined to the original construction of the galvanometer and does
not lend itself readily to subsequent alterations. Properly applied
it makes it possible to produce galvanometers which have a very
large ratio of total critical damping resistance to coil resistance.
Such galvanometers are usually of high current sensitivity and are
suitable for work (such as insulation testing by the direct-deflection
method) where the external resistance will be so large that the ordinary coil damping is negligible.

VI.

CONSTRUCTIONAL LIMITATIONS

The theoretical treatment of the problem in this paper has neglected

for generality, certain practical limitations which operate to modify
the results obtained in actual galvanometers.
It is probable that
the maker usually limits himself to wires of certain standard diameThere
ters with corresponding standard thicknesses of insulation.
must be an integral number of turns per layer, and (what may be
more important) an integral number of layers in the coil. In consequence, when a particular size of wire is used the theoretical number
of layers for the assigned radial depth of winding might be, say, 4.6.
In this case 4 layers will have to be used if 5 would cut down the
clearances between the coil and the pole pieces and core below the
permissible minimum.
It will, therefore, be realized that in checking
the performance of a line of galvanometers which nominally differ
only in resistance some variation from the theoretical performance
must be expected, especially for coils of low resistance made of coarse
wire.
H. B. Brooks, Trans.

Am.

Inst. Elec. Engrs., 39, pp. 514-515; 1920.

312

Bureau

of Standards Journal of Research

[von

The maker sometimes intentionally weakens the magnet of one

more galvanometers of a given series, all nominally alike except as

or
to

resistance, in order to get for these exceptional ones particular values

Also, in some groups of
of external resistance for critical damping.
moving-coil galvanometers the volume of wire composing the coil
varies greatly within the group, the tendency being to use larger
volumes for the coils of finer sizes of wire, for which the microampere
In such cases one can not expect
sensitivity is the important feature.
the performance to follow a law which includes the assumption of
identical volumes of wire in the coils.
If galvanometers are nominally alike, except as to resistance and
period, the values of their current sensitivities should be reduced to
the basis of a common period before their working sensitivities are
compared by the method given in this paper. This reduction is made
by using the relation that when other things are equal the deflection
per microampere increases -as the square of the period.
VII.

PRACTICAL CONSIDERATIONS

of the practical advantages which may be derived from the

maxima of the relations shown by the curves of Figure 2
are as follows:
(a) The galvanometer maker can restrict the number of standard
windings to comparatively few and still meet nearly all ordinary

Some

very

flat

requirements.
(h) The user can purchase galvanometers of a correspondingly
small number of values of resistance, which values may be widely
different, with the knowledge that the galvanometers will serve for
all but extraordinary needs which require the construction of special
galvanometers.
(c) Since the user has so much latitude in the matter of galvanometer resistance he can select the one having a value of external
resistance for critical damping most nearly equal to the actual external
resistance.
This reduces to a minimum the time required to obtain
a reading, and conserves a certain amount of sensitivity that might
otherwise be lost by the use of shunt or series resistance to obtain
proper damping. This consideration of proper damping is a very
important one and will often govern the choice.
The function expressed by equation (18) and plotted as curve A
in Figure 2 has an interesting property which is evident from the form
of the right-hand member of the equation, namely, that the same value
of relative sensitivity F/Fm&x will result for any given ratio G/R or
for its reciprocal R/G.
That is, if a given percentage of the maximum
sensitivity is obtained when the resistance of the galvanometer coil
is
times the resistance of the external circuit, the same percentage
will be obtained when the galvanometer resistance is 1/m times the
external resistance.
Since galvanometers of lower resistance, being
wound with coarser wire, are less liable to damage by accidental
overload, and have some other advantages, such as higher microvolt
sensitivity when used with bridges of low resistance, one would
naturally choose galvanometers of lower resistance and equal working
sensitivity, except as the consideration of damping might prevent.

Washington, August

21, 1929.