IMTC 2004 ~ Inslrummtatitionand Measurement

Technology Conference
CO~O, MY, 18-20 M ~ 2004
Y

Design of a Magneto-optic Residual Current Device using

Aqueous Ferrofluid as the Sensing Material

Luis Martinez', Ryszard Rakowski', Franjo Cecelja'

'Department of Design and Systems Engineering, Brnnel University, Uxbridge, UEi8 2PH, England
'Department of Engineering, Universidad Iberoamericana, Mexico City, 012 IO, Mexico
Phone: +44 1895 274000, Fax: + 44 1895 812556, Email: richard.rakowski@brel.ac.uk

Abstract - The paper presents a design for U Magneto-Optic Table I Performance improvement in matends selection
Residuul Current Device (MORCD) using ferroflurd as the sensing
material. As the selection of suitable magneto-optic material was Year Material Length Cost MOPI'
seen as the critical phase in the design, a novel Magneto-Optical
Pefortnunce Indicator is prolimed und verified rising known oprical 1995 DEDF lucm f100 3
materials. I f was experimentally esrablished that the designed
MORCD is siificiently Just rcnd capable of derecting residrcal 1998 Cd,Mn,.,Te I cm fi.1100 8
curreIZt (Islow as 1.51 m4.
x103 Ferrolluid 0.1 C m fl >II

Keyword.$: magneto-optics, residual current device, safefery, (*) Magnetooptic Performance Indicator defined in Section 2
current meusuremnt. of this paper

The research work presented in this paper identifies the

1. INTRODUCTION key variables, other than optical absorbtion and Verdet
constant, that need optimisation if a material is to be selected
For over forty years in domestic and industrial electrical for RCD applications. Consequently, a new Magneto-Optical
safety systems, residual current fault detection has been Performance Indicator (MOPI) for materials development, for
achieved using ferromagnetic (transformer) technology. use in high-sensitivity applications, was established. More
Inherent problems with reliability, false tripping, size and importantly, we found that an aqueous ferrofluid is a suitable
cost of these devices have resulted in consumer resistance to material for RCD applications. It has been shown that the
their use in the domestic environment. The search for MOP1 for the aqueous ferrofluid is 11.30 compared to a
improvements to the design of Residual Current Devices MOPI of 3 for a rod of DEDF glass. It was experimentally
(RCD), using alternative technologies, came a step closer, established that the sensor can detect currents in the
when in 1995 some authors [I] showed the feasibility of milliampere (better than 5 mA) range with an accuracy better
using the linear magneto-optical effect or Faraday effect for than 5%. With a materials cost o f f I and a length of 0.1 cm
the detection of currents in the milliampere range. They used the ferrofluid-based RCD can provide a viable alternative to
differential fields generated by two coils wound around a ferromagnetic alternatives.
Double Extra Dense Flint glass rod (DEDF 785258).
However, the length of I O cm of the glass rod was too long 11. PROPOSED APPROACH
and the coil arrangement too bulky for practical applications.
Another arrangement using a CdMnTe diluted magnetic A . Design Requirements
semiconductor crystal that reduced the size of the system in a
10:l proportion was developed [2] but, unfortunately, A Magneto-optic Residual Current Device (MORCD) can
increased the costs in the same proportion, hence producing a be represented by a three-block model as shown in figure I .
device that would not fulfil the initial component cost The electromagnetic circuit that senses the residual current is
requirements for a commercial device. The comparison of used to generate a residual magnetic field, which in t u n
some of magneto-optic materials used for RCD applications induces an observable Faraday effect within the magneto-
is summarized in Table I . optical material.

0-7803-8248-X/04/$17.00 02004 IEEE 804

 ;i_?
FuT17FH1 ,
L^YI-. f e*,”.
..~
Lli“.,

,.............~ ~ ~ . . . . . . .................................
oL”#UI-I a-,*. *.“.IO
......
,__-

olI.IBix* i 8w#,Pl~jsuI*l
~............ ~~ ....... ~~ ~ ...,.. ~~ ........~~~ ~~ ~

Fig. I Thrcc blmk modcl Tor MORCD

A set of characteristics for the primary element are

generally based on the requirements of the residual current
Fig. 2 Cause-effst diagram for the MORCD
devices, available standards and the idealised design
characteristics. In particular, the suitability of a material for
C. Figure of Merit
the MORCD application is driven by four main factors:
Size;
The figure of merit M ( = v l a ) a s defined in ref. [4] has
Magneto-optical characteristics (Verdet constant,
transmissivity and birefringence); proven to he a useful tool for evaluating the performance of
cost; magneto-opticai materials. Here V is the Verdet constant
Insulation capabilities ( non-conductive material). and a is the absorption coefficient at specific wavelength.
We found that the size and magneto-optical characteristics This figure of merit is now in common use amongst scientists
are correlated and result in the sensitivity of the device. This and engineers, and it is particularly useful in determining
correlation can be expressed by the following equation optimal operating wavelength for a particular Faraday
material. A main disadvantage of this Comparison, however,
K,VL >172.77.10-3 (1) is that most optical materials of engineering interest are
where K , is the transmisivity of the material, V is the designed in such a way that their transmittance is matched to
Verdet constant and L is the length of the primary element. the visible spectrum. Therefore, it can be demonstrated that
However, if the maximum length, which satisfies the such a figure of merit depends mainly on the derivative of the
manufacturers requirements (IO mm) is used. the material optical dispersion through the range. To overcome this
selection can initially rely on the Verdet constant and disadvantage and to consider other important factors, we have
transmisivity. established a multi-objective performance metric, namely the
Magneto-optical Performance Indicator (MOPI), whose
B. Optimisation ofthe MORCD metrics are summarised in Table 2.

The refinement of any design occurs as a result of Table 2 Objectives and indexes far magnetc-optical detrclar design
analysis, testing and redesign of the three main areas for Objeelive tnder
possible improvement: measuring methods, materials and Sensitivity lo magnetic fields Max. Verdct constant (V)
systems. The options for the optimisation of a MORCD can Dispersion of light throughout the Min vd
be visualised with a cause-effect diagram [3], as shown in spectrum Max. d n l d l
Fig. 2. We have found that potentially successful areas for Natural or induced optical activity Min I,
optimisation for the MORCD design are: Transmission loss Max T/L
a) materials with improved magneto-optical properties, i.e. Size Min L
large Verdet constant;
b) photodetectors with increased sensitivity; In addition to the indexes given in Table 2, we have also
c) electronic circuits with reduced noise for increasing the added a geometry factor. To this end, we have defined a
resolving power of the device; sensitivity factor S, which represents the size of the device
d) detection algorithms. [m] required for measuring a magnetic flux density of 125 FT
and considering the resolution of the detection system as I‘,
This paper focuses on the optimisation of materials selection and compared to a standard length of Imm:

805
 where S, [m] is the sensitivity factor for a magneto- SF57 117.32 10.7
SF58 28.41 Y .Y
optical device, 8, [rad] is the threshold of detection system,

it. The index S, is aimed to clearly differentiate those difference results from the inclusion of o;her indexes rather
materials that are highly mabmeto-optically efficient. Hence, than just those relying on the absorbance. If the MOP1 is used
the MOPl for MORCD designs can be calculated as, as a selection tool for the MORCD application, it is expected
that a suitable material will exhibit a MOPI larger than IO,
llv AnlM
V
MOP/= log(-)+
M
log(-)+-+- +- S, (3) which is the value exhibited by most of the suitable glasses.
vo -
MO l l v , . A n - I M Imm
where, MOPI is the magneto-optical performance indicator D. Muteriul selection
for the candidate material at wavelength 1[m] , V is the
Verdet constant [rad T-' m-'1, M the figure of merit [rad T-'1, In the MORCD design we have initially focused on optical
vd is the Abbe number, A d A A is the derivative of the optical glasses, which tend to be readily available at a relatively low
dispersion [m-'1, S, the sensitivity factor [m] defined by cost (El - f1.5 per millimeter of material) [I]. However,
MOPI for these materials is IO or lover, making them
equation (2). Also, V, M O ,vdo and Ano/lM are Verdet unsuitable for MORCD applications from the sensitivity
constant, figure of merit, Abbe number and derivative of the point of view. Other readily available materials with much
optical dispersion for the reference material, respectively. We higher sensitivity and consequently higher value for MOPI,
have quartz as the reference material since it is a material that such as polycrystalline materials in a form of Cadmium
has been extensively studied. The physical properties of Manganese Telluride (Cdn.ssMnn.asTe) or ferromabmetic
quartz are shown in Table 3. materials in a form of YIG tend to be far too expensive and
hence also unsuitable for MORCD applications.
Table 3 Quartz properties at 633 nm Consequently, our focus shifted over alternative materials in
Physical property Value the form of ferrofluids. Along with the commercially
Vcrdet constant (!I) 4.82 rad T I m-' available kerosene based ferrofluid EMGYOS, we have
Index of refraction at the d 1.532 423 developed a novel ferrofluid FFA [6], the characteristics of
line (w) which are summarized in Table 5 .
Abbe number ( vd) 67.96
'Transmittance 0.997 Table 5 MOPl f z t m for fmofluids
Length of the sample 0.01 m
Absorption coefficient 0.3 m? Fcrrafluid Carrier MOP1 M
[rad mi']
I/") 0.0147
FFA Water II II 31 1.68
AdAl 16 049.247 m?
lla 3.328 m
M 16.0425 rad T-'
s 0.00397
Table 5 indicates that the sensitivity of ferrofluids
represented in a form of MOPl for both commercially
available and in house designed materials are suitable for
In order to verify the validity of the selection method
MORCD applications as they are >IO. Also, ferrofluids tend
proposed in this paper, a set of materials with known
to be inexpensive materials with a typical cost of fO.15 per
properties is used. Optical glasses are selected since these
millitre [6]. Therefore, we used the in house designed
materials have been previously used as primary elements and
ferrofluid FFA.
their magneto-optical properties have been extensively
studied. A sample of three optical glasses were chosen as a
111. RESULTS
test group to validate the MOPI; BK7, SFS7 and SFS8. The
concomitant figure of merit M and MOPI for these glasses
are summarized in Table 4. The suitability of the selected ferrofluid as a current
sensing element is verified experimentally by the
implementation of a MORCD, as shown in Fig. 3. The
Table 4 Figure of mcnt and MOPl for opticvl glasses ferrofluid sensing element was prepared as described in ref.
[6] and positioned inside a solenoid (30 turns, 3.5 cm
Material M (rad T') MOPI diameter, 1 cm length). The assembly is mounted between
BK7 33.36 4.8 two linear polarisers whose transmission axis have an offset

806
of 45", as shown in Fig. 4, and illuminated with linear
polarised light generated by a 0.5 m W HeNe laser. The
intensity of the emerging light was sensed with a
photodetector OPTIO I,
OPTlOl
r----------.

--.
Fig. 6 Results from the current detection experiment (Oscilloscope set
Mains at SV / div, l 0 m idiv); frequency 59.5 Hz, coil current of 0.25 A

Fig. 3 Current ddrction expcrimenl circuit

Additionally, experiments with an AC signal from the mains
(1 17 VAC @ 60 Hz) applied to the sensor through a current
The coil in the current detection arrangement is connected
limiting resistor with a magnitude of 1.0 A indicated that the
to a 240W 117V AC unloaded motor (Fig. 3). system has a sensitivity of 4.81 VIA. As the total noise at the
detection system was measured to he 5.124 mV, the
concomitant resolution of the system is 1.5 1 mA.

IV. CONCLUSION

(0 As the material selection is the most critical phase in

Fig. 4 C w e n l detection armngement
MORCD design, a new Magneto-optic Performance
lndicator (MOPI), that was verified using known materials,
The experimental results with the coil current of 1.95 A at has proven to be a valuable design tool. MOPI has also
a frequency of 59.5 Hz are shown in Fig. 5 and for the coil shown that ferrofluids are very attractive and inexpensive
current of 0.25 A at a frequency of 59.5 Hz in Fig. 6. The materials for MORCD applications. The designed MORCD
observed time response confirms that that the system is was experimentally verified showing suitable response and
sensitivity to residual cwents as low as 1.51 mA.

REFERENCES
-7.- i
j
-I:
[I] L.M. Maniner, P.Adl, R.T. Rakowski and A. Cheshmehdaast, "Optical
current transducer for residual current sensing in the milliampere
range", Sensors and Actuators, A: Physical, vol. 67, num. 1-3 pt I , pp.
in2-10~,1998.
[Z] P. Kaervtrakulpong, “Magneto-optical residual current dwiccs", MSc.
Industrial Measuremen1 Systems Dissertation, Dept. o f Manufacturing
and Enginecnng Systems, London: Bruncl University, 1998.
[?] K. Ishikawa, Guide ID quality control, 2nd rev.,Tokyu Arbor MI: Asian
Producti vity Organization, 1986.
[4] C.C. Robinson, "The Faraday Rotation of Diamagnetic Glasses from
0.334 U to 0.19 U", Applicd Optics, vul. 3, num. 10, pp. 1163-1166,
1964.
151 G.A. Massey, D.C. Erickson, R.A. Kadlec, "Elrctromgnctic field
components: their measurement using linear elrctrooptic and
magnetooptic cffects(TE)". Applied Optics vol. 14 mum. I I , pp. 2712.
1975.
Fig. 5 Results from thc current detection expcriment (Oscilloscope set [ 6 ] L.M.Maniner, R.T. Rakowski, F. Crcclia, A Novel Magneto-uptic
Ferrotluid Material for Sensor Applications to be published a1
at SV I div, l0ms idiv); frequency 59.5 Hz. coil current of 1.95 A
lnstmmentation and Measurcmcnt 'Technology Conference, Italy 2004

807