IEEE TRANSACTIONS ON MAGNETICS, VOL. 51, NO.

1, JANUARY 2015 4000104

Setup for Generating an AC Magnetic Field From 3 to 100 kHz

Michal Ulvr
Faculty of Electrical Engineering, Czech Technical University in Prague, Prague 166 36, Czech Republic
Department of Electromagnetic Quantities, Czech Metrology Institute, Prague 150 72, Czech Republic

Generating a stable, homogeneous reference ac magnetic field is important for traceable calibrations of ac magnetic field analyzers
with a triaxial coil probe (e.g., EFA 300, C.A. 42). These probes are widely used by health and safety professionals, in manufacturing,
and in service industries. This paper presents a single-layer Helmholtz-type solenoid for generating the reference ac magnetic field
in the frequency range from 3 to 100 kHz. Special attention is given to a full characterization of this setup, including a comparison
between the analytical and finite element method design of the solenoid, on the one hand, and experimental results, on the other.
The magnetic flux density inside the solenoid can be generated with expanded uncertainty of (0.4%–0.8 %) for k = 2 using a special
search coil.
Index Terms— AC magnetic flux density (MFD), Helmholtz coils, search coil, solenoid.

I. I NTRODUCTION It is necessary to subtract one solenoid from another to

calculate the MFD—the internal solenoid with number of
H ELMHOLTZ coils as a magnetic flux density (MFD)
standard in various sizes and designs are widely used for
generating a reference ac magnetic field in a wide frequency
turns N2 and length l2 , which represents the empty turns,
has to be subtracted from the external solenoid with num-
range [1], [2] or for a single frequency value [3], [4]. The ber of turns N1 and length l1 of the winding including
design of the MFD standard should meet the requirements the empty turns between the two Helmholtz-type solenoid
for the amplitude value and the frequency range of the windings. The equation for MFD inside the Helmholtz-type
generated magnetic field for occupational and general public solenoid is then
exposure levels [5], and also for homogeneity in the volume Bz = Bz1 − Bz2 . (2)
of the probe size. Each coil is wound on its own frame in a
classical Helmholtz coils design, and these frames, therefore, After substituting into (2), we obtain the equation for calcu-
need to be mounted in a very stable manner. We have a lating the MFD inside the Helmholtz-type solenoid as
multilayer Helmholtz-type solenoid in the Czech Metrology  
N1 I 0.5l1 + z 0.5l1 − z
Institute for generating an ac magnetic field up to 3 kHz. B z = μ0  +
Multilayer coils cannot be applied for a higher frequency 2l1 R 2 + (0.5l1 + z)2 R 2 + (0.5l1 − z)2
 
range, due to the parasitic capacities of the winding. For these N2 I 0.5l2 + z 0.5l2 − z
−μ0  + .
reasons, a single-layer Helmholtz-type solenoid was designed 2l2 R 2 + (0.5l2 + z)2 R 2 + (0.5l2 − z)2
and produced for the frequency range from 3 to 100 kHz. (3)

II. D ESIGN AND C HARACTERIZATION OF THE S OLENOID Equation (3) will be simplified for calculating the MFD in the
center of the solenoid (z = 0) as
A. Analytical Method
(N1 − N2 )I
Two approaches can be used in the analytical design of B z = μ0
a Helmholtz-type solenoid. The first approach is based on  l1 −l2  
the Biot–Savart law, from which an equation can be obtained 0.5l1 R 2 + (0.5l2 )2 − 0.5l2 R 2 + (0.5l1 )2
×  . (4)
for determining the MFD of a long thin solenoid in the (R 2 + (0.5l1 )2 )(R 2 + (0.5l2 )2 )
longitudinal z-axis of the solenoid. Thus
  The second option is the calculation published in [6], which
NI 0.5l + z 0.5l − z also offers the possibility of calculating the MFD in the
B z = μ0  + (1)
2l R 2 + (0.5l + z)2 R 2 + (0.5l − z)2 ρ-axis perpendicular to the z-axis of the solenoid. The vector
of magnetic field B inside the single-layer solenoid can be
where N is a count of the winding turns, I is the current described by
through the solenoid (A), R is the radius of the winding (m),  
l is the total winding length (m), and z is the distance from u2 u4 u6
Bz = μ0 I αkσ 1 + k2 2 + k4 4 + k4 6 · · · (5)
the center of the solenoid in the axis of the solenoid (m). R R R
This equation is valid, if the solenoid winding is wound  
v2 v4
homogeneously along the whole length and is not interrupted. Bρ = μ0 I αkσ −k2 2 − k4 4 − · · · (6)
Ri R
Manuscript received June 3, 2014; revised September 5, 2014; accepted where I is the current through the solenoid (A), α is the
September 15, 2014. Date of current version January 26, 2015. Corresponding
author: M. Ulvr (e-mail: mulvr@cmi.cz). number of turns to 1 m, R is the radius of the winding
Digital Object Identifier 10.1109/TMAG.2014.2359296 (mm), kσ , k2 , k4 , and k6 are constants dependent on the
0018-9464 © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

Authorized licensed use limited to: University of Zagreb: Faculty of Electrical Engineering and Computing. Downloaded on December 07,2021 at 13:37:55 UTC from IEEE Xplore. Restrictions apply.
 4000104 IEEE TRANSACTIONS ON MAGNETICS, VOL. 51, NO. 1, JANUARY 2015

geometry of the solenoid (they can be found in [6]), and u 2 , u 4 ,

v2 , and v4 are functions of the coordinates derived from the
Legendre polynomials. In addition, all odd constants k1 , k3 , . . .
are equal to zero and the solenoid was designed with k2 = 0.
After substituting for u 4 , u 6 , and v4 into (6) and (7), we obtain
⎛ ⎞
8z 4 − 2z 2 ρ 2 + 3xρ 4
⎜ 1 + k4 8R 4
+ ⎟
Bz = μ0 I αkσ ⎜ ⎝ 16z 2 − 120z 4 ρ 2 + 90z 2 ρ 4 − 5ρ 6
⎟
⎠
k6 + · · ·
16R 6
(7)
 
zρ(4z − 3ρ )
2 2
Bρ = μ0 I αkσ −k4 4
− ··· (8)
2Rin
where z (mm) and ρ (mm) are cylindrical coordinates with
their origin in the center of the solenoid, and z is the longitu- Fig. 1. Model of the single-layer Helmholtz-type solenoid in FLUX 3D.
dinal axis of the solenoid. The following solenoid parameters
have been chosen: length of the frame 300 mm, I = 1 A,
N = 20 turns, R = 168.75 mm, turn pitch 1.8 mm, wire
diameter 0.75 mm, winding length 36 mm, and α = 555, 556.
When we substitute the parameters above for z = 0, ρ = 0 into
(7) and (8), the theoretical dc MFD value in the center of the
solenoid is 0.105882 mT. If we substitute the corresponding
parameters of N1 = 115 turns, l1 = 207 mm, N2 = 75 turns,
l2 = 135 mm into (4), we obtain a theoretical value of
dc MFD of 0.105724 mT. The difference of 0.15% between the Fig. 2. Substitute schema of the solenoid for ac simulation in FLUX 3D.
calculated values according to (4) and according to (7) and (8)
is caused by the rounded spreadsheet values and the calculated where K bdc is the solenoid constant with a dc current value of
dimensions that have been applied. Equations (3) and (7), 1 A (mT), f is the frequency of the generated ac MFD (Hz),
respectively, (8) can be used for calculating the theoretical and fr is the resonance frequency value (Hz). The change of
homogeneity values in the z-axis of the solenoid, respectively, K bac from K bdc is 1% for f / f r = 0.1, and 4% for f / fr = 0.2.
ρ at any distance from the center of the solenoid. Theoretical Equation (10) is then valid in its approach for small changes
homogeneity is calculated from the changes in the value of of K bac from K bdc .
the MFD for the given coordinates, relative to the value of the
MFD in the center of the solenoid. B. Finite Element Method
The design described previously needs to be analyzed for The solenoid was designed and simulated in the FLUX 3D
use in generating ac MFD. For ac analysis, it is impor- environment by the finite element method (FEM). Because, the
tant to know the value of the winding self-capacitance frame of the solenoid will be made of non-magnetic material,
C P , the self-inductance L S , the winding impedance, and it can be omitted in the model. This means that it was neces-
the resonance frequency value fr . The self-inductance value sary to create only a pair of windings, which should conform
of a single-layer solenoid can be calculated from the dimensionally as far as possible with the parameters given
optimized version of Wheeler’s formula [7]. The calcu- above. Finally, a model of the Helmholtz-type solenoid, Fig. 1,
lated value is then L S = 531 μH. The self-capacitance was created for magnetostatic simulation (magnetostatic 3D).
value C P = 137.1 pF was obtained from the optimized Subsequently, we calculated the MFD value of 0.105808 mT
Medhurst formula [8]. The resonance frequency value can be in the center of the solenoid model with a dc current value
calculated from of 1 A. The same model was also used for ac simulation
1 (steady-state ac magnetics 3D). Because, the coil conductor
fr =  (9)
2π L s C p in FLUX 3D behaves like a resistor (with a calculated value
the value for which is 589.8 kHz, after substituting for of 1.68 ), i.e., with no inductance or parasitic capacitance,
L S and C P . The required solenoid operating frequency range it was necessary to create a substitute coil schema and to
up to 100 kHz should be sufficiently below the resonance assign it to the model, to properly simulate the behavior of
frequency. In addition, to these parameters, it is also very the coil in the ac circuit (Fig. 2). The circuit is powered from
important to know the frequency dependence of the solenoid a voltage source V with a set frequency and voltage value.
constant K bac . This dependency is usually quadratic, and can The calculated theoretical values L S and C P were used in the
be determined approximately from [9] substitute schema. Resistor R N connected in series with the
  substitute schema with a value of 1  is used for calculating
f2 the ac current value. Then, an ac analysis of the MFD of the
K bac = K bdc 1 + 2 (10)
fr solenoid model can be made.

Authorized licensed use limited to: University of Zagreb: Faculty of Electrical Engineering and Computing. Downloaded on December 07,2021 at 13:37:55 UTC from IEEE Xplore. Restrictions apply.
 ULVR: SETUP FOR GENERATING AN AC MAGNETIC FIELD FROM 3 TO 100 kHz
Fig. 3. Single-layer Helmholtz-type solenoid No. 1201.
TABLE I
C OMPARISON OF THE T HEORETICAL DC VALUES OF THE M AGNETIC
F LUX D ENSITY IN THE C ENTER OF THE S OLENOID W ITH THE
C ALIBRATED VALUE
III. S ETUP C HARACTERIZATION AND R ESULTS
A single-layer Helmholtz-type solenoid No. 1201 was
produced according to the frame and winding parameters
above (Fig. 3). The frame with inner diameter 285 mm was
made from glass epoxy with a refined groove and with a turn
pitch of 1.8 mm. The turn pitch was chosen large enough
to reduce the influence of the parasitic capacitance of the
winding. The winding was wound with enameled copper wire
0.75 mm in diameter. The winding is terminated by the BNC
connector. The solenoid was calibrated by direct comparison
with the national MFD standard. The calibrated solenoid dc
constant of 0.105497 mT/A was determined with expanded
uncertainty of 0.05% for k = 2. As is stated in Table I, the
relative difference δ (the difference between the theoretical
value in the center of the solenoid and the calibrated value)
is in the range from 0.21% to 0.36%, depending on the
method that is used. A real winding resistance value of
2.51  was measured. An inductance value of 600 μH was
measured by the LCR meter. The difference between the
calculated and measured resistance, respectively, inductance
value is given by the heterogeneous cross section of the wire
winding in the entire winding length and deviations in the turn
pitch from a value of 1.8 mm. This value could also affect
both the supply and the BNC connector. The actual winding
impedance will then be greater than the calculated values.
The frequency dependence of the calculated and measured
winding impedance values is shown in Fig. 4. Two methods
can be used for making it possible to determine the frequency
dependence of the MFD generated by the solenoid. The first
method, calculation according (10), requires knowledge of
the dc MFD value (determined by the calibration) and the
resonance frequency value. A resonance frequency value of
535 kHz was measured using a digital oscilloscope. The value
Authorized licensed use limited to: University of Zagreb: Faculty of Electrical Engineering and Computing. Downloaded on December 07,2021 at 13:37:55 UTC from IEEE Xplore. Restrictions apply.
4000104
Fig. 4. Frequency dependence of the impedance value.
Fig. 5. Frequency dependence of the solenoid constant.
of C P = 148 pF was calculated from the resonance frequency
value. Alternatively, it can be measured directly by a special
search coil with calibrated area turns. This method is also used
for precise adjustment of ac MFD in the solenoid. The special
search coil design and the measurement method is described
in detail in [10]. A special multilayer search coil No. K I with
calibrated constant K S = (1.3312 ± 0.0011) m2 was used for
measuring the adjustable magnetic field value inside a solenoid
in the frequency range up to 3 kHz. A single-layer search coil
No. EP 01/00 with calibrated constant K S = (0.045394 ±
0.000036) m2 was used for measuring the magnetic field value
inside the solenoid in the frequency range from 3 to 100 kHz.
The adjusted amplitude value of MFD Bm inside the solenoid
is then calculated from
UAVG
Bm = √ (11)
4K S f 2
where UAVG (V ) is the mean value of the output voltage
of the search coil measured using a Keithley 2001 digital
multimeter, K S (m2 ) is the value of the constant of the search
coil, and f (Hz) is the frequency measured by an HP 53131A
digital counter. The frequency dependence of the solenoid
constant calculated from (10), calculated by FEM in FLUX 3D
and measured by a special search coil (the search coil No.
K I was used up to 3 kHz and search coil No. EP 01/00 was
used from 3 to 100 kHz) is shown in Fig. 5. The relative
difference of the calculated values from the measured values
 4000104
TABLE II
R ELATIVE E XPANDED U NCERTAINTY (k = 2) OF THE AC MFD VALUE
M EASURED /S ET IN THE C ENTER OF THE S OLENOID
Fig. 6. Homogeneity inside the solenoid in the z-axis.
Fig. 7. Homogeneity inside the solenoid in the ρ-axis.
is <0.6% in the frequency range up to 500 Hz, and is in
the range from 0.25% to 0.7% in the frequency range from
3 to 100 kHz. Only the constant value at 100 kHz calculated
by FEM has a relative difference of ∼2% from the measured
value. The ac MFD amplitude value to 140 μT up to 40 kHz
and 40 μT at 100 kHz inside the solenoid can be measured/set
with expanded uncertainty of (0.4%–0.8%) for k = 2 in
the frequency range of (3–100) kHz, based on the detailed
uncertainty analysis in [10].
The uncertainty depends mainly on the uncertainty of the
search coil constant, on the uncertainty of the measured output
voltage from the search coil, and on the uncertainty of search
coil placement in the center of the solenoid. Homogene-
ity inside the solenoid must also be considered. Example
of the expanded uncertainty values of the ac MFD value
Authorized licensed use limited to: University of Zagreb: Faculty of Electrical Engineering and Computing. Downloaded on December 07,2021 at 13:37:55 UTC from IEEE Xplore. Restrictions apply.
IEEE TRANSACTIONS ON MAGNETICS, VOL. 51, NO. 1, JANUARY 2015
measured/set in the center of the solenoid in the frequency
range (3–100) kHz is presented in Table II. Finally, the real
homogeneity values were measured experimentally using a
special search coil No. EP 601 with a suppressed octupole
at a frequency of 50 Hz. The results for the theoretical values
calculated from (3), (7), and (8) and by FEM and the measured
homogeneity values are shown in Figs. 6 and 7 at a distance
of ±40 mm. The search coil was moved from the center
in the z- and ρ-axis, and the true homogeneity value was
calculated from the change in the measured output voltage.
The theoretical values calculated from (3) and (7) were the
same.
IV. C ONCLUSION
The design of the single-layer Helmholtz-type solenoid for
ac MFD generation in the frequency range of (3–100) kHz has
been proposed, and a full characterization has been presented.
The analytical design and the FEM design were compared with
a real solenoid. The results from the dc analysis show that the
relative difference of the MFD theoretical values in the center
of the solenoid is from 0.21% to 0.36%, depending on the
method that is used. The results from the ac analysis show that
the relative difference of the MFD frequency dependence in
the center of the solenoid is in the range of (0.25%–0.7%) for
the frequency range of (3–100) kHz—about 2% at 100 kHz—
compared with the values measured by a special search
coil. Good agreement between the theoretical (calculated) and
measured homogeneity values in the z- and ρ-axis was also
proven. Depending on the amplifier and the current shunt
that are used, an ac amplitude MFD value can be generated
by means of this solenoid with expanded uncertainty of
(0.4%–0.8%) for k = 2.
R EFERENCES
[1] M. Chiampi, G. Crotti, and D. Giordano, “Set up and characteri-
zation of a system for the generation of reference magnetic fields
from 1 to 100 kHz,” IEEE Trans. Instrum. Meas., vol. 56, no. 2,
pp. 300–304, Apr. 2007.
[2] D. A. Rollett and A. E. Drake, “Traceable magnetic field strength
measurements to 120 kHz,” IEE Proc. Sci., Meas. Technol., vol. 143,
no. 4, pp. 255–258, Jul. 1996.
[3] C. Fabricio de Melo, R. L. Araújo, L. M. Ardjomand, N. S. R. Quoirin,
M. Ikeda, and A. A. Costa, “Calibration of magnetic field meters
at 60 Hz using a Helmholtz coil: Constructive aspects and calcula-
tion of associated uncertainties,” in Proc. 18th IMEKO World Congr.,
Rio de Janeiro, Brazil, Sep. 2006, pp. 1942–1945.
[4] M. Ishii, R. Ketzler, M. Albrecht, S. Kurokawa, and Y. Shimada,
“Improvement of formula and uncertainty of the reference magnetic field
for ac magnetometer calibration,” IEEE Trans. Instrum. Meas., vol. 62,
no. 6, pp. 1443–1449, Jun. 2013.
[5] ICNIRP, “Guidelines for limiting exposure to time-varying electric and
magnetic fields (1 Hz to 100 kHz),” Health Phys., vol. 99, no. 6,
pp. 818–836, 2010.
[6] J. V. Afanasjev, N. V. Studencov, V. N. Khorev, E. N. Czeczurina, and
A. P. Shcolkin, Sredstva Izmerenij Parametrov Magnitnogo Polja. Saint
Petersburg, Russia: Ėnergia, 1979.
[7] R. Weaver. (2012). Numerical Methods for Inductance Calculation.
[Online]. Available: http://electronbunker.ca/CalcMethods.html
[8] D. W. Knight. (2013). The Self-Resonance and Self-Capacitance
of Solenoid Coils. [Online]. Available: http://www.g3ynh.info/zdocs/
magnetics/appendix/self_res/self-res.pdf
[9] V. E. Czernyshev, “Rasczot czastotnoj pogreshnosti mier magnitnoj
indukcii,” in Trudy Metrologičeskich Institutov SSSR, vol. 140. Moscow,
Russia: Izd. standartov, 1972.
[10] M. Ulvr, “Calibration of magnetic field meters up to 50 kHz at CMI,”
in Proc. 19th Symp. IMEKO TC, Barcelona, Spain, 2013, pp. 20–25.