Home Search Collections Journals About Contact us My IOPscience

A Rogowski–Chattock coil for local magnetic field measurements: sources of error

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2010 Meas. Sci. Technol. 21 107003

(http://iopscience.iop.org/0957-0233/21/10/107003)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 193.255.248.150
This content was downloaded on 02/02/2015 at 07:58

Please note that terms and conditions apply.

IOP PUBLISHING MEASUREMENT SCIENCE AND TECHNOLOGY
Meas. Sci. Technol. 21 (2010) 107003 (5pp) doi:10.1088/0957-0233/21/10/107003

TECHNICAL DESIGN NOTE

A Rogowski–Chattock coil for local

magnetic field measurements:
sources of error
Ahmed Abou-Elyazied Abdallh and Luc Dupré
Department of Electrical Energy, Systems & Automation, Ghent University, Sint-Pietersnieuwstraat 41,
B-9000 Ghent, Belgium
E-mail: Ahmed.Abdallh@UGent.be

Received 1 June 2010, in final form 20 July 2010

Published 2 September 2010
Online at stacks.iop.org/MST/21/107003

Abstract
A magnetic field can be measured on a space scale of a few millimeters by means of magnetic
sensors, such as a Rogowski–Chattock coil. The main advantage of the Rogowski coil is its
capability to measure the field strength directly at the sample surface because both ends of the
coil can be installed very close to the specimen surface. However, the measurements are
affected by numerous errors, which are comprehensively discussed in this note.

Keywords: Rogowski–Chattock coil, measurements errors, local magnetic measurements

(Some figures in this article are in colour only in the electronic version)

1. Introduction 2. A Rogowski coil: theory of operation

Local magnetic field measurements have gained a lot of interest
during the last decade, since they are becoming important
tools in different research topics. The local magnetic field is
measured by means of magnetic sensors, such as a Rogowski–
Chattock coil. The main advantage of the Rogowski coil is its
capability to measure the field strength directly at the sample
surface because both ends of the coil can be installed very close
to the specimen surface [1]. Although the Rogowski–Chattock
coil was initially described in the 18th century [2], it is still used
for one-dimensional [3] and two-dimensional [4] magnetic
field measurements. However, the inaccuracies in Rogowski–
Chattock coil measurements have not been investigated fully,
up to now.
In this note, we present a theoretical analysis of the
sources of error in Rogowski–Chattock coil measurements,
which is very useful to enhance the accuracy of magnetic
measurements.
A Rogowski coil, or Chattock coil (also known as magnetic
potentiometer [2, 5]), is a special kind of helical coil sensor
uniformly wound onto a relatively long non-magnetic, flexible
strip bent in such a way that its end faces are placed in close
contact with the sample, as shown in figure 1 [6, 7].
The principle of the Rogowski coil is based on the property
that for any closed path not containing a current source
(L = C1 C2, see figure 1) the integral of the magnetic
field H is zero, L H · dl = 0. Therefore, a scalar magnetic
potential difference exists between two points A and B, which
can be determined by
 
FmA − FmB = H · dl = H · dl. (1)
C1 C2
If the magnetic field is uniform between points A and B,
the line integral of the magnetic field H along the path C1 can

0957-0233/10/107003+05$30.00 1 © 2010 IOP Publishing Ltd Printed in the UK & the USA
Meas. Sci. Technol. 21 (2010) 107003 Technical Design Note

n n
d
Non magnetic strip, VRogowski = μ0 S Hi × li (8)
Lc average = 0.04 m Lc average i=1
dt
C2

n d
VRogowski = μ0 S Hdl. (9)
Lc average dt C2
Specimen Substituting (1) and (2) into (9) yields
d D n d dHAB
S = π x (0.4e -3)2 m2 VRogowski = μ0 S (HAB lAB ) = KR , (10)
Lc average dt dt
A C1 B where
lAB = 0.005 m μ0 SnlAB
KR = (11)
Lc average
is the Rogowski coil coefficient ‘coil sensitivity’, which can
Figure 1. Schematic overview of the Rogowski–Chattock coil used, be calculated from this equation, or determined by calibration
with 1305 turns. using a Helmholtz coil. Therefore, the magnetic field strength
be calculated as H on the specimen surface can be obtained by the time integral
 of the potential difference VRogowski .
H · dl = HAB lAB , (2) A higher coil sensitivity can be achieved by a larger
C1
Rogowski coil coefficient KR , which is proportional to the
where HAB and lAB are the magnetic field component
number of coil turns. Assuming that the wire used in the coil
tangential to the path C1 and the distance between points
with a diameter d is uniformly wound on the non-magnetic
A and B in the magnetic material, respectively.
coil strip, the total number of coil turns is n = (Lc average /d).
When the magnetic field varies with time, the induced
Hence, the coil sensitivity (11) can be reformulated as
terminal output voltage of the Rogowski coil is the sum of
voltages induced in each turn (all n turns are connected in μ0 SlAB
KR = . (12)
series): d
n The Rogowski coil is fabricated ‘manually’ in such a
VRogowski = Vone turn . (3) way that the coil turns are uniformly distributed on the
i=1 complete circumference of the non-magnetic coil strip. The
Assuming that the cross-sectional area Si of the turn i manufacturer has to pay attention to ensure that the end turns of
of the Rogowski coil is so small that the field value may the Rogowski coil are at the end of the non-magnetic strip and
be supposed to be constant over any given cross section Si , in parallel with the specimen surface, so that a good contact
although its direction and value may vary along C2 for each with the specimen can be achieved. Moreover, the diameter
turn, we have of the non-magnetic strip has to be greatly smaller than the
length between the two end faces, i.e. D  lAB , in order to
dφ n
d
VRogowski = = Bi Si , (4) reduce the error of taking the average coil length Lc average .
dt i=1
dt On the other hand, the Rogowski coil can be calibrated
using Helmholtz coils. The Helmholtz coil system produces
where φ is the total flux linkage, Si is the coil cross-sectional
a uniform magnetic field (0.1% error) in a central cylinder
area and n is the number of coil turns. Equation (4) can be
parallel to the Helmholtz coil axis. The volume of the
rewritten as
uniform cylinder has a radius equal to 15% of the Helmholtz
 n
d
VRogowski = μ0 Si Hi . (5) coil radius and a length equal to 40% of the distance
i=1
dt between the two coils (these percentage values depend on
the manufacturer). Specifically, in the Helmholtz coil used,
Next, let us assume that the cross-sectional area Si is
the uniform cylinder has a diameter of 50 mm and a length of
constant for every coil turn, i.e. Si = S, ∀i:
50 mm [8]. The Rogowski coil sensor is placed in the middle of
n
d the uniform magnetic field cylinder region. The water-cooled
VRogowski = μ0 S Hi . (6)
dt Helmholtz coils system, used in our laboratory ‘EELAB’, is
i=1
capable of creating a magnetic induction in the air gap up to
By multiplying equation (6) with l i
li
, where li is the 0.2 T, which corresponds to a magnetic field of 16 kA m−1 .
magnetic path length between two neighboring coil turns, we The rated current in the Helmholtz coils is 35 A at a power
obtain frequency of 50 Hz. The Rogowski coil used has 1305 turns,
n
d li with a calibrated and calculated sensitivity of 0.412 and
VRogowski = μ0 S Hi × . (7) 0.103 04 nV (A m−1 )−1 , respectively. The dimension of the
i=1
dt li
Rogowski coil is shown in figure 1. The deviation between
Again, assuming that the magnetic path length li is constant the calibrated and calculated sensitivity could be explained
L
for every coil turn, i.e. li = l, ∀i, therefore, l = c average
n
, due to the assumptions that have been made in the theoretical
for a single layer sensor analysis, i.e. equations (6) and (8), which are not realized

2
Meas. Sci. Technol. 21 (2010) 107003 Technical Design Note

in practice. Therefore, it is clear that the Rogowski coil Air Magnetic field lines
calibration using the Helmholtz coil is necessary.
The main advantage of the Rogowski coil is its capability Coil turn
to measure the field strength directly at the sample surface
Hi
because both ends A and B of the coil can be installed very
H H Hi
close to the specimen surface. A correctly designed and θi
manufactured Rogowski coil should give the same output Hi H
signal independent of the shape of the coil between points
A and B.

3. Sources of errors in Rogowski coil measurements

Specimen
The accuracy of the magnetic field measurements using the
Rogowski coil can be affected by numerous errors, such as Figure 2. Schematic overview of the Rogowski–Chattock coil used
faulty integration by the Rogowski coil and non-uniformity subjected to a uniform magnetic field pattern.
of the magnetic field in the specimen over the region of
integration [9]. The local magnetic field at each coil turn Hi is the
(i) Faulty integration by the Rogowski coil. In the above magnetic field component perpendicular to the ith
theory, it has been assumed that Si and li are constant coil turn:
along the path C2 of the Rogowski coil. Unless these Hi = |Hi | cos θi , (20)
conditions are carefully observed in the winding of the where θi is the angle between the local magnetic field
coil, the integration is incorrect. vector Hi and its perpendicular component on the ith
(a) Inaccurate Si : the error could occur, for example, coil turn Hi (see figure 2).
when the coil cross-sectional area S is not constant The Rogowski coil can be used to measure the
for every coil turn, i.e. Si = S, ∀i. Recalling equation local magnetic field strength in magnetic circuits with
(5), uniform or non-uniform field patterns. Below, the
n
d sources of errors for both cases are analyzed.
VRogowski = μ0 Si Hi . (13)
i=1
dt Uniform magnetic field pattern. In the case of a
uniform magnetic field pattern (see figure 2), the
Let us assume S̃ is the average cross-sectional area
value of the magnetic field H is constant for each
of the coil, so
coil turn, so S̃ and L̃ can be reformulated as
n
d  n
VRogowski = μ0 S̃ Hi , (14) d
H ni=1 Si cos θi i=1 Si cos θi
i=1
dt S̃ = dtd n =  n , (21)
dt
H i=1 cos θi i=1 cos θi
where S̃ is
n d d n
i=1 Si Hi H cos θi li
S̃ = n ddt . (15) L̃ = dt
d n i=11
i=1 dt Hi dt
H i=1 li cos θi li
n
(b) Inaccurate li : moreover, the error could also occur, cos θi li
for example, when the magnetic path length is not = n i=11 . (22)
i=1 li cos θi li
constant for every coil turn, i.e. li = l, ∀i.
Recalling equation (7), It is clear from equations (21) and (22) that the
n
d li inaccuracy in Si and li may give different results
VRogowski = μ0 S Hi × , (16) compared to equation (9) as S̃, L̃ and also KR depend
i=1
dt li
on some geometrical coil parameters, i.e. θi and li .
n
1 d Once these values are calibrated using a Helmholtz
VRogowski = μ0 S Hi × li . (17) coil system, the Rogowski coil can be used accurately
l i dt
i=1 for measuring magnetic fields in magnetic circuits
Let us assume L̃ is the average magnetic path length with uniform magnetic field patterns.
of the coil turn, so
Non-uniform magnetic field pattern. However, in
μ0 S  d
n
VRogowski = Hi × li , (18) the case of a non-uniform magnetic field pattern,
L̃ i=1 dt the value of the magnetic field H is not constant for
each coil turn. This means that S̃ and L̃ are not
where L̃ is
n d only dependent on the geometrical coil parameters
Hi li
L̃ = n i=1 1dt d . (19) θi and li , as in the case of uniform field patterns
i=1 li dt Hi li (see equations (21) and (22)) but may also strongly

3
Meas. Sci. Technol. 21 (2010) 107003
or weakly depend on the time dependence of the
magnetic field to be measured.
When calibrating the Rogowski coil in the
Helmholtz coil system, the Rogowski coil is
subjected to a uniform sinusoidal magnetic field at the
power frequency. This calibration procedure, which
gives a specific coil sensitivity value based on the
characteristics of the specific magnetic field used in
the Helmholtz coil system, may result in inaccurate
measurements due to the different characteristics of
the non-uniform magnetic field patterns to be tested.
This might be the reason for the errors observed for
Rogowski coil measurements in magnetic circuits
with a non-uniform magnetic field pattern. The
higher the magnetic field non-uniformity, the higher
the expected measurement errors. This issue is
discussed in detail in section 4.
Further inaccuracies occur if the Rogowski coil is not
in contact with the specimen. That could occur, for
example, when the end turns of the Rogowski coil are
not parallel with the specimen surface [9]. The magnetic
potential measured is then not the one on the surface
of the specimen, and equation (2) does not hold. Also,
especially at high frequency, the Rogowski coil suffers
from the disadvantage of being unstable to measure a
field without there being a relative motion of the sample
and the coil, due to sample vibration [10].
(ii) Non-uniformity of the magnetic field in the specimen over
the region of integration. In order to maintain equation (2),
the field in the specimen over the region of the Rogowski
coil must be approximately constant. In order to fulfill
this condition,
• the size of the Rogowski coil should be small,
especially the distance between the coil end faces
points A and B in figure 1;
• the Rogowski coil should be placed where the field is
most uniform.
4. Measurements and tests
In order to show the effect of the degree of the magnetic
field non-uniformity on the Rogowski coil measurements,
local magnetic field measurements are carried out in magnetic
circuits with two different magnetic field pattern cases:
uniform and non-uniform magnetic field patterns. In both
cases, the local magnetic field measurements using the
Rogowski coil are coupled with the local magnetic induction
measurements at the same position, measured using a search
coil wound over the complete thickness of the sample in order
to recover the normal magnetizing B–H curve of the material
under test. The normal magnetizing B–H curve is composed
of two magnetic quantities. The first magnetic quantity is
the magnetic induction, which is obtained in the two cases
using the search coil technique, which is considered as a
golden standard technique [1]. The second magnetic quantity
is the magnetic field strength, which is measured using the
Rogowski coil sensor. We assume here that the expected error
4
Technical Design Note
Ferromagnetic yoke
96
Excitation
winding
Rogowski coil
. x x .
50
x . . x
Search coil Magnetic
sample
32 128
Ferromagnetic yoke
Figure 3. Schematic diagram of the single-sheet tester used as a
uniform magnetic field pattern. Dimensions are in mm.
in the obtained normal magnetizing B–H curve is due to the
error in the magnetic field measurements using the Rogowski
coil sensor rather than the error in the magnetic induction
measurements.
The material under test is a non-oriented electrical steel
(M 700/50 A). It is assumed that the material is isotropic, and
hence the B–H curve is the same everywhere. The obtained B–
H curve for each magnetic circuit condition is compared with
the original normal magnetizing B–H curve measured using
the IEEE standard 393-19911 , in which a magnetic ring core
is fully and uniformly wound with two windings, an excitation
winding and a measurement winding. The magnetic field
strength and magnetic induction are obtained using Ampere’s
law and Faraday’s law, respectively [11].
The quasi-static magnetic measurements are performed
at 1 Hz for a sinusoidal current excitation, in order to have
a negligible presence of eddy current effects in the magnetic
core. The object under test is demagnetized between two
successive measurements. The measurements were carried out
20 times in order to ensure better measurement reliability. The
objects, which are used for the two different circuit conditions,
are made from the same material and constructed in such a
way that the influence of the cutting is the same for all objects.
In the results presented below, the calibrated Rogowski coil
sensitivity is used.
4.1. Uniform magnetic field patterns
A double-yoke single-sheet tester is used to generate a
magnetic field, with a high degree of field uniformity, in the
magnetic strip between the upper and lower yokes, as shown
in figure 3. The measurements were carried out in the middle
of the magnetic strip.
4.2. Non-uniform magnetic field patterns
A partially excited magnetic ring core (figure 4), can be seen as
a magnetic circuit with a non-uniform magnetic field pattern.
1 IEEE standard 393-1991 IEEE standard for test procedures for magnetic
cores.
Meas. Sci. Technol. 21 (2010) 107003 Technical Design Note

L̃ and also KR , which varies due to the time dependence of

Excitation
. winding
the non-uniform magnetic field to be measured (see equations
(21), (22)).
x
5. Conclusion
Do
In this note, the sources of error in Rogowski–Chattock coil
Di
measurements are investigated. We have shown that the
inaccurate coil fabrication may cause a considerable error in
the magnetic field to be measured. The discussions presented
Measurements
Position in this note clarify the necessity of coil calibration using the
Helmholtz coil. However, it is worth mentioning that the first
source of error (inaccurate integration) can be eliminated by a
proper calibration, which is not useful for the second source of
error (non-uniformity of the magnetic field in the specimen).
Magnetic measurements were carried out in order to support
the theoretical analysis.
Figure 4. Schematic diagram of the partially excited magnetic ring
core (non-uniform magnetic field pattern), Do = 145 mm, Di = 90
mm and  = 130◦ .
Acknowledgment
1.6
The authors gratefully acknowledge the financial support of
1.4
projects FWO-G.0082.06, GOA07/GOA/006 and IAP-P6/21.
1.2
Magnetic Induction B(T)

1
0.8
0.6
0.4
Uniform magnetic field pattern
0.2 Non−uniform magnetic field pattern
Reference characteristics
0 Electromagnetic modelling of short circuited coreplates
0 250 500 750 1000 1250 1500 1750
Magnetic Field Strength H (A/m)
Figure 5. The obtained B–H curve based on the local magnetic
measurements for magnetic circuits with uniform and non-uniform
magnetic field patterns, compared to the original characteristics
(IEEE standard 393–1991).
The degree of the magnetic field non-uniformity depends on
the value of the unwound excitation winding angle : the
higher the unwound angle value, the higher the degree of
the magnetic field non-uniformity. The measurements were
carried out in the middle of the unwound region.
4.3. Results and discussion
Figure 5 shows the obtained normal magnetizing B–H curve in
the two different magnetic circuits compared to the original one
(IEEE standard 393–1991). It is clear that the measurements
in the magnetic circuit with a uniform magnetic field pattern
could obtain almost the original B–H curve. However, the
measurements that were carried out in magnetic circuits with
a non-uniform magnetic field pattern result in an inaccurate
B–H curve. This could be explained due to the error in S̃,
References
[1] Abdallh A and Dupré L 2010 Local magnetic measurements in
magnetic circuits with highly non-uniform electromagnetic
fields Meas. Sci. Technol. 21 045109
[2] Chattock A 1887 On a magnetic potentiometer Proc. Phys.
Soc. Lond. 24 94–6
[3] Roger J, Vrignaud E, Henneron T and Benabou A 2009
2000 2250 2500 COMPEL 28 762–71
[4] Zhong J, Guo Y, Zhu J and Lin Z 2006 Characteristics of soft
magnetic composite material under rotating magnetic fluxes
J. Magn. Magn. Mater. 299 29–34
[5] Rogowski W and Steinhaus W 1912 Die Messung der
magnetischen Spannung (The measurements of magnetic
potential) Arch. Elektrotech. (Electrical Engineering) 1
141–50
[6] Tumanski S 2007 Induction coil sensors: a review Meas. Sci.
Technol. 18 R31–46
[7] Fiorillo F 2004 Measurement and Characterization of
Magnetic Materials (New York: Academic)
[8] Walker Inc. http://www.walkerscientific.com/Products/
Product Lines/Helmholtz/helmholtz.html
[9] Margerison T and Sucksmith W 1946 Some uses of the
magnetic potentiometer for the determination of
magnetization curves upon open-circuited specimens J. Sci.
Instrum. 94 182–4
[10] Cullity B and Graham C 2008 Introduction to Magnetic
Materials (New York: Wiley Interscience)
[11] Abdallh A, Sergeant P, Crevecoeur G, Vandenbossche L,
Dupré L and Sablik M J 2009 Magnetic material
identification in geometries with non uniform
electromagnetic fields using global and local
magnetic measurements IEEE Trans. Magn.
52 4157–60