Measurement of magnetic fields using the voltage generated

by a vibrating wire
Jason Gilberta and Cameron Baribeau
arXiv:2008.09898v1 [physics.acc-ph] 22 Aug 2020

Canadian Light Source, University of Saskatchewan, Saskatoon, Canada

August 25, 2020

Abstract
A vibrating wire may be used as an instrument with a variety of applications, one of
which is the measurement of magnetic fields. Often, the magnetic fields are determined by
measuring the amplitude of the wire vibration under the action of a Lorentz force. Though
generally adequate, this approach may be inconvenient in certain circumstances. One of
these occurs when it is necessary to measure the amplitude of high-frequency vibration, as
the amplitude is expected to decrease linearly with frequency, and thus becomes harder to
measure. Another example may be found in situations where the sensor must operate over a
wide range of vibration frequencies. In this case the sensor will be unresponsive to specific
frequencies of wire vibration, which are determined by the placement of the sensor. This
means that for the instrument to be robust, the sensor must be precisely mobile, or multiple
sensors must be used.
Here a technique which may be used to supplement the displacement sensor is described.
This technique makes use of the voltage generated by the motion of the wire in the mag-
netic field under measurement. It is predicted that the technique may be more suitable for
measurements requiring high frequency vibration, and is sensitive to all frequencies of vibra-
tion. Measurements of a magnetic field obtained using this technique are compared to those
found using only a displacement sensor, and the benefits and drawbacks of the technique are
discussed.

a
Corresponding author. Electronic mail: [email protected]

1
1 Introduction
In the field of accelerator physics the vibrating wire technique [1] is utilized to measure the magnetic
field of a variety of devices. These include quadrupoles [2, 3, 4, 5], undulators [6], and wigglers
[7, 8, 9, 10, 11].
This technique has recently been used at the Canadian Light Source [11] in an attempt to mea-
sure the magnetic field along the centerline of an insertion device, a hybrid in-vacuum wiggler with
a period of 80 mm. While taking these measurements, it was noted that the electrical impedance
of the wire appeared to be changing. It was proposed that this change may be explained by an
electromotive force (emf) generated by the motion of the wire. To test the validity of this expla-
nation, the hypothetical emf generated by the wire was calculated, and in doing so, it was found
that it may be possible to obtain useful information about the magnetic field from measurements
of the emf. For an example of this principle applied to the measurement of fluid viscosity, see Ref.
[12].
This contribution is organized as follows. In Section 2 the theory of the vibrating wire tech-
nique is briefly explained; after which, a description of the emf generated by the vibration of a
wire, in terms of this theory, is developed. In Section 3 the application of this emf for magnetic
measurement is discussed. In Section 4 methods used to test the practicality of this approach,
and the results of these tests, are discussed. Finally, the conclusions drawn from this work are
summarized in Section 5.

2 Theory
2.1 A summary of the vibrating wire technique
To provide context for the measurement method being proposed, the conventional method must
first be described. This description will be minimal, and those who are interested in more detail
may refer to Ref. [1].
The principle of the vibrating wire technique is that by measuring the displacement of a wire
due to a Lorentz force acting along its length, the magnetic field which causes the Lorentz force may
be inferred. It is well known that the transverse displacement of a taut wire under the influence
of some driving force is described by

T uzz − γut + µutt = F . (2.1)

Here u = u(t, z) is the transverse displacement of the wire, and F = F (t, z) is the force per unit
length acting on the wire; t is the time at which the wire is observed, and z is the direction which
the wire extends. Subscripts are used to denote differentiation with respect to the subscripted
symbol. The coefficients of the differential terms are T the tension in the wire, µ the mass per
unit length of the wire, and γ the damping coefficient.
In the context of magnetic field measurement, this differential equation relates a Lorentz force
F (t, z) acting on the wire to the transverse displacement u(t, z) of the wire; if the current in
the wire is known, the magnetic field acting on the wire may be determined by measuring the
displacement of the wire. A diagram which defines the coordinate system, depicting an insertion
device, may be found in Figure 2.1.

2
 𝒚

Girders

Wire 𝒙

Figure 2.1: A cross section of the measurement apparatus. The wire extends in the z direction,
perpendicular to the page. The magnetic field produced by the girders is in the y−direction. The
motion of the wire is ideally confined to the plane y = 0.

It has been shown [1] that if the wire is held fixed at points which are outside the influence of
the magnetic field, a solution to the above differential equation may be obtained in the form of a
sine-series. This means that the magnetic field may be expressed as
∞
X  nπ 
B(z) = Bn sin z , (2.2)
n=0
L

where the series coefficients Bn are related to the displacement of the wire, and L is the distance
between the points where the wire is held fixed. If the wire is underdamped, and the current in
the wire is of the form

I(t) = I0 eiωt , (2.3)

which is meant to represent a sinusoidal signal with amplitude I0 and angular frequency ω, written
in the form of a complex exponential for convenient mathematical manipulation, then the series
coefficients are related to the vibration of the wire by
 nπ  I0 cos (ωt − φ)
un (t, z) = Bn sin z . (2.4)
L µ [(γ 0 ω)2 − (ω 2 − ωn2 )2 ]1/2

Here un is the displacement of the wire when driven at a frequency near the nth harmonic of the
fundamental frequency of the wire; ωn is the frequency of the nth harmonic, which corresponds to
the nth series coefficient, and is given by
s
nπ T
ωn = ; (2.5)
L µ

3
γ 0 = γµ−1 , and φ is the phase shift between the driving current and the wire vibration, defined as
γ 0ω
tan φ = . (2.6)
ωn2 − ω 2
Examining equation (2.4), it can be seen that the measured amplitude of the wire vibration is
given by the time-independent component of the equation,
 nπ  I  −1/2
0
An = Bn sin zs (γ 0 ω)2 − (ω 2 − ωn2 )2 , (2.7)
L µ
where zs is the position of the displacement sensor. From this the series coefficients may be
obtained by fitting the following equation to measurements of vibration amplitude,
−1/2
An = an cn ω 2 − (ω 2 − bn )2

, (2.8)

assuming that T , µ, and γ are effectively constant for the curve. From which the series coefficients
may be found to be
µ  nπ −1
|Bn | = an sin zs . (2.9)
I0 L
One should note that equation (2.7), from which the series coefficients are calculated, is strictly
of one sign, as determined by the sensor position zs . This means that it can not by itself be
used to calculate the corresponding magnetic field via equation (2.2), as it must be known which
of the coefficients are positive and which are negative. This information may be obtained from
measurements of the phase shift between the wire vibration and the driving current, as was done
in [1].
To briefly elaborate on the need for assuming the wire is underdamped in the derivation of
equation (2.4), it is required so that the curves of equation (2.4) are sufficiently localized, so that
each un is near zero in the vicinity of un±1 . If this is not the case, equations (2.7)-(2.9) will be
better represented as a sum over n, and a more complicated approach to fitting may be required.

2.2 Electromotive force as a signal

In the previous section it was explained how an electrical signal may be used to induce motion in
a wire by means of a Lorentz force, and that this motion acts as a signal which carries information
about the strength of the magnetic field along the length of the wire. Here an alternative approach
will be explored. In this case, the motion of the wire will be considered to be a way of generating
an electrical signal, and this electrical signal will be shown to carry information about the magnetic
field.
The fact that an electrical signal is generated by the vibration of the wire may be explained
by Faraday’s Law. The circuit containing the wire is a loop of conducting material, meaning
the vibration of the wire causes the area of, and therefore the magnetic flux through, the circuit
to vary over time, giving rise to a voltage via ε = −dΦ/dt. This means that a voltage which
opposes the driving current will be generated across the wire, so that the wire appears to resist
the driving current. In terms of an electric motor, the signal we mean to discuss may be referred
to as ”counter-electromotive force” or ”back emf”. Here it will simply be referred to as the emf.
At this point a mathematical description of ε, the emf generated by a vibrating wire, will be
derived for a wire located at the center of an insertion device. In the case of a wire positioned

4
at x = y = 0, and a magnetic field directed along the vertical direction y, such as is depicted in
Figure 2.1, the magnetic field acting on the wire may be written By (x, z) = B(x, 0, z), where the
subscript indicates the component of the magnetic field. If the motion of the wire is confined to
the plane y = 0, then the emf can be expressed in terms of the flux enclosed by the circuit as
d L u
Z Z
dΦ
ε = − = − By (x, z) dxdz , (2.10)
dt dt 0 −∞
where u is the displacement of the wire. Since u = u(t), by the fundamental theorem of calculus
Z L
∂u
ε = − By (u, z) dz . (2.11)
0 ∂t

If the displacement of the wire takes the form of a standing wave this becomes
∂Λ L
Z  nπ 
ε = − sin z By (u, z) dz , (2.12)
∂t 0 L
where Λ = Λ(t) is the amplitude of the wire displacement at time t. Noting that u = u(z), meaning
By (u, z) = By (z), and recalling the definition of the sine-series,
∞
2 L
X  nπ  Z  nπ 
B(z) = Bn sin z and Bn = B(z) sin z dz , (2.13)
n=0
L L 0 L

then it becomes clear that

L ∂Λ
ε = −Bn . (2.14)
2 ∂t
At this point, one should recall that these results are obtained for a driving force in the form of
F = F (z), meaning that it is assumed that the magnetic field does not vary appreciably in either
of the transverse directions within the span of the wire vibration.
If the motion of the wire is described by equation (2.4), then un = Λ sin nπ


L
z , and
I0 L h
2 i−1/2
ε(t, ω) = Bn2 sin (ωt − φ) ω (γ 0 ω)2 − ω 2 − ωn2 ; (2.15)
2µ
comparing the time-dependent factors of equations (2.15) and (2.4), it can be seen that the
phase difference between the emf and the driving current will be φε = φ + π/2, so that
ωn2 − ω 2
tan φε = . (2.16)
γ 0ω
The series coefficient from equation (2.2) may then be obtained using a curve fitting approach
similar to that described in the previous section, by measuring the amplitude of the emf as opposed
to the amplitude of the wire vibration. Applied to equation (2.15), this gives
−1/2
εn (ω) = an ω cn ω 2 − (ω 2 − bn )2

, (2.17)

and
r
2µ
|Bn | = an . (2.18)
I0 L

5
These results suggest that measurements of the voltage across the wire may be used to obtain
information about the magnetic field acting on the wire, in a manner similar that described in the
previous section.
There is, however, one key difference between the two which one must take note. This is
the fact that measurement of the voltage can only be used to determine the magnitude of the
series coefficients. In mathematical terms, this is evidenced by the relationship εn ∝ Bn2 , which is
independent of the sign of Bn . In physical terms, one may consider the fact the voltage generated
by the wire depends only on the speed of the wire, meaning it is independent of the direction of
the wire moves with respect to the driving current.

3 Application
In this section the application of the theory described in Section 2.2 will be discussed.
To begin, the shortcomings of using emf as a signal will be considered. The first is that it may be
unwieldy as a measurand for non-planar magnetic fields. Though the derivation of equation (2.18)
explicitly treats only the case of planar motion in a field of one component, equation (2.11) should
be valid for any wire displacement in any magnetic field if By (x, z) is replaced with B(x, y, z).
It must be remarked that in the case that there is a significant field in both of the directions
transverse to the motion of the wire, each component of the field will contribute to the voltage
which is produced. This would mean that the emf would carry information regarding both field
components, and special care would need to be taken to interpret the meaning of the signal. If
optical displacement sensors are used, the contribution of each field component may be easier to
determine, as the motion in each direction may be measured independently.
Next, it should be reiterated that measurement of the amplitude of the emf is not by itself
sufficient to calculate the magnetic field distribution. Unlike those results, in this case the signal
being measured is proportional to the square of the series coefficient. This means that the sign of
the coefficient may not be determined from measurement of the emf alone, making it unsuitable for
determining the field distribution over space via equation (2.2). Despite this, applications which do
not require the sign of the coefficients still exits. One such is the determination of the magnitude
of coefficients related to specific frequencies of vibration, which may be used to characterize error
fields such as in Ref. [10].
Turning now to the advantages of measuring the emf, the first is that it may be used to
complement some of the shortcomings of a displacement sensor. For example, the sensor may
be positioned near the node of one of the harmonics, meaning that it would be unable to detect
the vibration of the wire at that frequency. To elaborate on the problematic nature of the sensor
response, consider equation (2.9) in the form
µ  nπ −1
Bn = an sin zs ∝ sin (x)−1 , (3.1)
I0 L
and apply simple calculus-based error analysis to obtain
∂Bn
∝ Bn cot (x) . (3.2)
∂x
This shows that for (x mod π)  1, or nzs /L near an integer, the uncertainty in the series coeffi-
cient value becomes very large. This uncertainty becomes significant when many series coefficients
are needed, and the distribution of these coefficients is difficult to predict, such as in the detailed

6
measurement of insertion device fields [11]. To compensate for this, the sensor would need to be
mobile so that it could be moved away from the nodes at these problematic frequencies, see Ref.
[13] for example, or multiple sensors would need to be used to reduce the number of cases where
there is no detectable motion. In these cases, the voltage produced by the motion of the wire may
still be detectable, despite the insensitivity of the displacement sensor.
Another benefit to measuring the emf is that it may be better suited to determine the magnitude
of the series coefficients corresponding to higher harmonics of vibration, and therefore to higher
driving frequency. Examining the expression for wire displacement, equation (2.4), it can be seen
that un ∝ ω −1 when ω ≈ ωn , meaning that the amplitude of the wire vibration will become
smaller as the driving frequency increases and, consequently, harder to measure. In comparison,
the expression for emf, equation (2.15), predicts the amplitude of the emf to be independent of
the driving frequency. This implies that the emf may be better suited for measurements at high
driving frequency.
Finally, it may be possible to use this technique for quadrupole fiducialization. By moving the
current-carrying wire relative to the quadrupole, the magnetic center may be found by measuring
the position where no emf is produced by the wire. A similar approach is described in in Ref.
[14], in which the motion of the wire is mechanically driven. Another attempt at using a vibrating
wire for this purpose may be found in Ref. [2], in which the displacement of the wire is measured
instead of the voltage.
In the interest of predicting the sensitivity of the emf to magnetic field strength, one may
consider that the derivative of the signal dε/dBn ∝ Bn , meaning that the sensitivity should vary
linearly with the magnitude of the coefficient being measured. A more meaningful approach may
be to compare the emf signal to that of the vibration amplitude, the sensitivity of which has
already been compared to other magnetic measurement techniques (Ref. [15]). To predict the
relative sensitivity of each technique to the strength of the magnetic field, one may consider the
ratio of one measurand to the other. Comparing equations (2.7) and (2.15), it can be found that

Bn ωL  nπ −1
max |εn | = sin zs An , (3.3)
2 L
which, for driving frequencies near a resonant frequency, can be written as
s ! s !
T  nπ  −1 ω Bn T  nπ −1 B
n
max |εn | = nπ sin zs An ≈ nπ sin zs An (3.4)
µ L ωn 2 µ L 2

upon substitution of equation (2.5). From this it can be seen that when the same magnet is
measured using both methods, and the same measurement setup, the emf will be scaled with
respect to the vibration amplitude by two factors. The first is a factor proportional to the driving
frequency, as has already been noted, and the second is the magnitude of the coefficient being
measured, which is the projection of the field onto that particular mode of vibration. This implies
that measurement of the emf is less effective for small-magnitude coefficients, unless the frequency
of vibration is sufficiently high. In general, equation (3.4) may be used as a crude way of estimating
the response of one approach compared to the other.

7
4 Measurement and Analysis
4.1 Measurement
This section contains a description of the approach used to measure the emf generated by the wire,
as well as a discussion of the measurements taken. Due to a shortage of time with the equipment, it
was only possible to measure 9 harmonics, and only one measurement per harmonic was possible.
In order to test the hypothesis that the emf generated by the vibrating wire may be used to
determine the magnetic field acting on the wire, the voltage drop across the wire was measured
as a function of driving frequency. This voltage is related to the emf generated by the wire by
Kirchoff’s Law,

V = ε + IR , (4.1)

where V is the measured voltage across the wire, I is the drive current, and R is the resistance of
the wire. A plot of the predicted form of the measurable quantities, voltage and phase (equations
(4.1) and (2.16), respectively), can be found in Figure 4.1.
The amplitude of the wire voltage, as well as the phase offset with respect to the driving voltage,
was measured using a Signal Recovery model 7265 lock-in amplifier which took the voltage across
the wire as a direct input. For these measurements, the magnetic field of an in-vaccum wiggler
with a period of 80 mm and a length of about 1.41 m was used. For the measurement a beryllium-
copper wire with a diameter of 0.1 mm, and length of 3.516 m, was used, and tension was applied
using a hanging mass. Parameters of the setup can be found in Table 1. The data collected may
be found in Figure 8.2.
It should be noted that the measurements presented were taken using two different current
amplitudes to drive the wire. One set of data, for modes 61 through 70, were collected using
a 20 mA drive current, while a 1 mA current was used for modes 85 through 90. The choice of
current amplitude was made based on the observed amplitude of the vibration near the resonant
frequency being considered. The driving force had to be sufficient to produce a clear signal but
at the same time not be so large as to excite non-linear phenomena, meaning that frequencies at
which the wire was particulary sensitive to driving had to be measured using smaller currents.
Those interested in details concerning non-linear phenomena in the motion of vibrating wires may
refer to [16, 17], and especially [18].

Parameter Symbol Value

Density µ 71.74 mg · m−1
Length L 3.516 m
Tension T 9.022 N
Fundamental Frequency f0 51.6 Hz

Table 1: Parameters of the experimental setup.

8
 1 100

V / max|V|
0.9
50

[deg]
Normalized Wire Voltage 0.8

Phase
0.7

-50
0.6

0.5 -100
0.9 0.95 1 1.05 1.1
Driving Frequency / n

Figure 4.1: Voltage drop across the wire, predicted using equation (4.1) with IR = 1.1 V, and the
phase shift between the wire and drive voltages, predicted using equation (2.16).

4.2 Analysis
The measurements of the wire voltage were then compared to the theory outlined in Section 2.2
by fitting the following altered version of equation (2.17). In order to account for variations in the
driving voltage of the wire, such as those caused by a change in ambient temperature, a DC offset
parameter was introduced to the curve-fitting equation so that
−1/2
εn (ω) = an ω cn ω 2 − (ω 2 − bn )2

+ dn . (4.2)

The result of this curve fitting process can be found in Figure 8.3. A measure of the quality of the
fit for each data set can be found in Table 2. The quantity R found there is the sum of square
differences between the data set and the fit, normalized by the sum of squared measurements. This
normalization is applied to ensure the measure is independent of the scale of the data. Expressed
mathematically,
!−1
X X
R = yk 2 [yk − f (ωk )]2 , (4.3)
k k

where yk is the k th element of the data set and f (ωk ) is the fit-function evaluated at the k th
frequency measured.
Examining the data illustrated in Figure 8.2, it can be seen that, in general, the measured
curves are of the same shape as the predicted curves found in Figure 4.1. However, there are
some features which appear unique to the driving current used. In particular, the measurements
taken using a drive current of 20 mA appear to have an irregular phase shift, in that the phase is
bound between ±10◦ while the measurements taken using a drive current of 1 mA have a phase
bound between ±60◦ . Referring to Figure 4.1, it appears the latter case agrees more closely with

9
theory. This implies either an unaccounted for dependence on the drive current, or an experimental
error. It seems likely that this difference is due to measurement error related to the difference in
signal strength. Typically, the curves measured at a smaller drive current have heights between 0.5
and 1 volts, while the other curves are an order of magnitude smaller, and the latter are notably
narrower than the former. If measurement error is responsible for this discrepancy, it could possibly
be rectified by sampling the curve at a higher frequency resolution.
To check the accuracy of the series coefficients calculated using measurements of the emf Bnε ,
the values of which may be found in Table 2, comparison was made to those calculated from
measurements of wire displacement BnA . In theory, the two approaches to measurement should
give the same series coefficients. It should be noted that this comparison makes use of data sets
collected several weeks apart. In this time the apparatus was used intermittently, and the same
wire remained strung. Based on observations made throughout the use of the vibrating wire, it
does not seem likely that the time interval between measurements will affect the values of the
series coefficients.
A comparison of the two sets of coefficients may be found in Table 2, and is visualized in
Figure 8.1, which shows the relative difference between the two sets of coefficients, taking BnA
as the reference. These figures show close agreement of coefficients 87, 89, and 91, as well as the
apparent relationship between driving current and measurement discrepancy. To elaborate on the
latter claim, comparison of the two sets of coefficients showed that Bnε measured at 20 mA differed
from BnA . It is noteworthy that these Bnε correspond to data sets with irregular phase shifts,
seen in Figure 8.2. In addition, the Bnε for n = 85 is exceptional in that it was measured using
a different drive current than the BnA it is being compared to. If there is an unaccounted for
dependence on current, this comparison may be invalid.
Examining Figure 8.2, it is interesting to note that in the cases of n = 89 and n = 91 some
signs of non-linear oscillation, resulting from excessive vibration amplitude, may be seen. The main
feature which implies non-linear oscillation is the asymmetry of the amplitude curves, namely that
there appears to be a sharp drop on the high-frequency side of the curve associated with the collapse
of transverse whirling motion [17]. Even though the motion of the wire appears to be outside the
regime described by equation (2.1), the series coefficients calculated using either method are in
close agreement. This would indicate that non-linear wire motion may have either a similar, or
minor, impact on the values of the series coefficients. In the least, this would mean that the
measurement of emf is no more sensitive to this problematic oscillation than the measurement of
displacement.

10
 n I0ε (mA) I0A (mA) Bnε (G) BnA (G) δBn (%) R(10−4 )
61 20 20 175.1 ± 9.3 241.2 27 12
63 20 20 130.3 ± 7.3 186.3 30 7.1
65 20 20 312 ± 21 -429.0 27 28
67 20 20 62.7 ± 3.8 89.42 30 0.63
70 20 20 49.5 ± 7.7 86.67 43 1.80
85 1 2 39590 ± 170 -42850 8 11
87 1 1 73810 ± 260 71640 3 5.5
89 1 1 73310 ± 290 -73280 0 7.3
91 1 1 37550 ± 180 38590 3 13

Table 2: Table of drive currents used for each measurement set, coefficients calculated for each
set, and corresponding percent difference between Bnε and BnA . BnA are reported to arbitrary
precision. R is a measure of the quality of the curve fit, normalized to be independent of the scale
of the data.

5 Conclusion
In Section 2.2 a theory of an alternative way of measuring magnetic fields using the vibrating wire
technique was developed. This alternative makes use of the electromotive force generated by the
motion of the wire in the field. The advantages, disadvantages, and potential applications of this
theory were discussed in Section 3. In Section 4 experimental tests of the theory are reported, the
results of which were found to agree with the conventional method for 3 of 9 test cases. The cases
which did not agree used a different drive current than the others, and produced considerably
smaller signals with narrower resonance curves. Based on this it is thought that the discrepancy
between the remaining 6 cases may be due to a much smaller measured signal, the amplitude of
the driving current used, or to the frequency resolution at which the curves were measured.
For comparison, the advantages and disadvantages of the technique with respect to the conven-
tional method will be summarized. Among the advantages is the lack of need for a displacement
sensor, which means simpler experimental design and setup. Another is, theoretically, a stronger
signal at high frequencies of vibration. The disadvantages include a lack of information on the
sign of the series coefficient, making the method unsuitable for determining field distributions, as
well as the fact that the orientation of the field under measurement is ambiguous. In addition, the
sensitivity of the technique is predicted to be more complicated than the conventional technique,
in that, comparatively, the sensitivity of the emf is worse for small-magnitude coefficients, but
becomes better as the magnitude increases.
To conclude, based on the results presented here, it would appear that the theory developed in
Section 2.2 is valid. This means that, in principle, it may be possible to determine magnetic field
through measurements of the voltage across the vibrating wire instead of the amplitude of the wire
vibration. A more thorough analysis will be needed to assess the practicality of this technique, as
well as identify unaccounted for factors, such as non-linear current dependence.

6 Acknowledgements
The authors are grateful to the following people. Jon Stampe for valuable discussion and aid in
experimental setup. Tor Pederson, Grant Henneberg, Garth Steel, Bruce Wu and Carl Jansen for

11
aid in developing the measurement system.
Research at the Canadian Light Source was funded by the Canada Foundation for Innovation,
the Natural Sciences and Engineering Research Council of Canada, the National Research Council
Canada, the Canadian Institutes of Health Research, the Government of Saskatchewan, Western
Economic Diversification Canada, and the University of Saskatchewan.

7 Author Contributions
JG planned and carried out measurements, developed the model, analysed the data, and wrote
the manuscript. CB contributed to the analysis of the data used as reference, and to the editing
of the final manuscript.

12
8 Figures

0.45
I0 = 20 mA
0.4
I0 = 1 mA I0A = 2 mA
0.35 I0 = 1 mA
1 - | Bn / BnA |

0.3

0.25

0.2

0.15

0.1

0.05

0
61 63 65 67 70 85 87 89 91
Harmonic Number

Figure 8.1: Plot of relative error in emf coefficients. Bars coloured to indicate data set. I0 indicates
driving current used for voltage measurements, while I0A indicates that used for displacement
measurements. For those labelled by I0 , I0 = I0A . Black bars represent the uncertainty in the
value. (If viewed in grayscale, Table 2 on p.11 may be used as reference).

13
 n = 61 I0 = 20 mA n = 63 I0 = 20 mA
1.145 1.5 1.125 1.8
Voltage Voltage
1.14 Phase Phase
1.6
1.135 1.12
1
1.4
[V]

[V]
1.13

[deg]

[deg]
1.115
Wire Voltage

Wire Voltage
1.125 1.2
0.5

Phase

Phase
1.12 1
1.11
1.115
0.8
0
1.11 1.105
0.6
1.105

1.1 -0.5 1.1 0.4

3130 3135 3140 3145 3150 3155 3160 3165 3235 3240 3245 3250 3255 3260 3265 3270
Frequency [Hz] Frequency [Hz]

n = 65 I0 = 20 mA n = 67 I0 = 20 mA
1.22 4 1.11 1.8
Voltage Voltage
Phase Phase 1.7
1.2 3 1.109
1.6
[V]

[V]
1.18 2 1.108 1.5
[deg]

[deg]
Wire Voltage

Wire Voltage
1.4
1.16 1 1.107
Phase

Phase
1.3

1.14 0 1.106 1.2

1.1
1.12 -1 1.105
1

1.1 -2 1.104 0.9

3335 3340 3345 3350 3355 3360 3365 3370 3440 3445 3450 3455 3460 3465 3470 3475
Frequency [Hz] Frequency [Hz]

n = 85 I0 = 1 mA n = 87 I0 = 1 mA
0.6 60 1.5 80
Voltage Voltage
Phase Phase 60
0.5 40
40
[V]

[V]

0.4 20 1
[deg]

[deg]
20
Wire Voltage

Wire Voltage

0.3 0 0
Phase

Phase
-20
0.2 -20 0.5
-40
0.1 -40
-60

0 -60 0 -80
4370 4375 4380 4385 4390 4395 4400 4405 4470 4475 4480 4485 4490 4495 4500 4505
Frequency [Hz] Frequency [Hz]

n = 89 I0 = 1 mA n = 91 I0 = 1 mA
1.5 80 0.5 60
Voltage Voltage
Phase 60 0.45 Phase
40
0.4
40
[V]

[V]

1 0.35 20
[deg]

[deg]

20
Wire Voltage

Wire Voltage

0.3
0 0
Phase

Phase

0.25
-20
0.5 0.2 -20
-40
0.15
-40
-60 0.1

0 -80 0.05 -60

4575 4580 4585 4590 4595 4600 4605 4610 4680 4685 4690 4695 4700 4705 4710 4715
Frequency [Hz] Frequency [Hz]

Figure 8.2: Measurements of wire voltage for several resonant frequencies. n is the harmonic
number, and I0 is the amplitude of the driving current.
14
 n = 61 I0 = 20 mA n = 63 I0 = 20 mA
45 25

40

35 20
[mV]

[mV]
30
15
Counter-EMF

Counter-EMF
25

20
10
15

10 5
5

0 0
3130 3135 3140 3145 3150 3155 3160 3165 3235 3240 3245 3250 3255 3260 3265 3270
Frequency [Hz] Frequency [Hz]

n = 65 I0 = 20 mA n = 67 I0 = 20 mA
120 10

100 9
[mV]

[mV]
80 8
Counter-EMF

Counter-EMF
60 7

40 6

20 5

0 4
3335 3340 3345 3350 3355 3360 3365 3370 3440 3445 3450 3455 3460 3465 3470 3475
Frequency [Hz] Frequency [Hz]

n = 85 I0 = 1 mA n = 87 I0 = 1 mA
0.5 1.5

0.4
[V]

[V]

0.3 1
Counter-EMF

Counter-EMF

0.2

0.1 0.5

-0.1 0
4370 4375 4380 4385 4390 4395 4400 4405 4470 4475 4480 4485 4490 4495 4500 4505
Frequency [Hz] Frequency [Hz]

n = 89 I0 = 1 mA n = 91 I0 = 1 mA
1.4 0.5

1.2 0.4
[V]

[V]

1
0.3
Counter-EMF

Counter-EMF

0.8
0.2
0.6
0.1
0.4

0.2 0

0 -0.1
4575 4580 4585 4590 4595 4600 4605 4610 4680 4685 4690 4695 4700 4705 4710 4715
Frequency [Hz] Frequency [Hz]

Figure 8.3: Measurements of emf with theoretical curve, equation (4.2), fit to data. n is the
harmonic number, and I0 is the amplitude of the driving current.
15
References
[1] Alexander Temnykh. Vibrating wire field-measuring technique. Nuclear Instruments and
Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated
Equipment, 399(2-3):185–194, 1997.

[2] Alexander Temnykh. The magnetic center finding using vibrating wire technique. Preperint
CBN, pages 99–22, 1999.

[3] Z Wolf. Vibrating wire technique for quadrupole fiducialization. In 14th International Mag-
netic Measurement Workshop, pages 26–29, 2005.

[4] C. Wouters, M. Calvi, V. Vrankovic, S. Sidorov, and S. Sanfilippo. Vibrating wire technique
and phase lock loop for finding the magnetic axis of quadrupoles. IEEE Transactions on
Applied Superconductivity, 22(3):9001404–9001404, June 2012.

[5] V. Vrankovic, C. Wouters, S. Sidorov, P. Chevtsov, R. Deckardt, M. Emmenegger, M. Dach,

and S. Sanfilippo. A method for the submicrometer accuracy determination of quadrupole
magnetic axis. IEEE Transactions on Applied Superconductivity, 24(3):1–5, June 2014.

[6] Alexander Temnykh, Yurii Levashov, and Zachary Wolf. A study of undulator magnets
characterization using the vibrating wire technique. Nuclear Instruments and Methods in
Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equip-
ment, 622(3):650–656, 2010.

[7] Alexander B Temnykh. Some aspects of the use of vibrating wire technique for a wiggler
magnetic field measurement. Technical report, SIS-2002-426, 2001.

[8] Alexander B Temnykh and Kenneth D Finkelstein. The chess g-line wiggler tuning. In
PACS2001. Proceedings of the 2001 Particle Accelerator Conference (Cat. No. 01CH37268),
volume 4, pages 2456–2458. IEEE, 2001.

[9] A Temnykh. Vibrating wire and flipping coil magnetic measurement of a cesr-c 7-pole wiggler
magnet. In Proceedings of the 2003 Particle Accelerator Conference, volume 2, pages 1026–
1028. IEEE, 2003.

[10] Alexander B Temnykh. Vibrating wire apparatus for periodic magnetic structure measure-
ment. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spec-
trometers, Detectors and Associated Equipment, 515(3):387–393, 2003.

[11] Cameron Baribeau, Drew Bertwistle, Evan Ericson, Jason Gilbert, and Tor Pedersen. Mag-
netic measurements of insertion devices using the vibrating wire technique. In 10th Int. Par-
tile Accelerator Conf.(IPAC’19), Melbourne, Australia, 19-24 May 2019, pages 1683–1686.
JACOW Publishing, Geneva, Switzerland, 2019.

[12] AAH Pádua, JMNA Fareleira, JCG Calado, and WA Wakeham. Electromechanical model for
vibrating-wire instruments. Review of scientific instruments, 69(6):2392–2399, 1998.

[13] Renan Geraldes, Rodrigo Junqueira Leão, Geraldo Cernicchiaro, Regis Terenzi Neuenschwan-
der, James Francisco Citadini, and Antônio Ricardo Droher Rodrigues. New sensor and
non-contact geometrical survey for the vibrating wire technique. Nuclear Instruments and

16
 Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associ-
ated Equipment, 811:115–123, 2016.

[14] GE Fischer, VE Bressler, DR JKCobb, RE Ruland, HV Walz, and SH Williams. Precision

fiducialization of transport components. In Proceedings EPAC, page 138, 1992.

[15] Pasquale Arpaia, Domenico Caiazza, Carlo Petrone, and Stephan Russenschuck. Performance
of the stretched and vibrating wire techniques and correction of background fields in locating
quadrupole magnetic axes. In XXI IMEKO World Congress ’Measurement in Research and
Industry, 2015.

[16] R Narasimha. Non-linear vibration of an elastic string. Journal of Sound and Vibration,
8(1):134–146, 1968.

[17] Roger J Hanson, James M Anderson, and H Kent Macomber. Measurements of nonlin-
ear effects in a driven vibrating wire. The Journal of the Acoustical Society of America,
96(3):1549–1556, 1994.

[18] Henrik B Pedersen and Jeppe Langeland Knudsen. Direct determination of the non-linear
connection between tension and transverse amplitude for a vibrating string. European Journal
of Physics, 38(4):045003, 2017.

17