Design of Magnetic Shield for Hall-Effect Current Sensor Using Integrated Magnetic
Concentrator
HwangHongsik1
ChoJunghyun1
LeeCheewoo†
-
(Ph.D. course Department of Electrical and Computer Engineering, Pusan National University)
Copyright © The Korean Institute of Illuminating and Electrical Engineers(KIIEE)
Key words
IMC Hall Sensor, Current Sensor, Magnetic Shield, Shielding Error Rate, Skin Effect Error Rate
1. Introduction
Recently, due to limited oil resources and global environmental problems, interest
in eco-friendly vehicles such as HEVs and EVs is growing globally. Especially, HEV
among eco-friendly vehicles has the highest expectation (1). At the same time, as the HEV market expands, the market for current sensors is also
growing together. The current sensor is a key component of an eco-friendly vehicles.
It is used to measure the current flow during charging and discharging of the charger
and to monitor the battery management system (2).
The presence of the IMC hall sensor based on the IMC technology among the current
sensors indicates that the sensor placement changes, and the sensor performance is
improved (3-4). In addition, a suitable shield shape is designed and presented so as not to cause
an error due to interference of external magnetic field generated from a 3-phase or
polyphase current sensor (5-6). In this paper, the optimal shape of current sensor shield is studied for miniaturization
of current sensor for HEV in vehicle components business which has space limitation.
In order to achieve miniaturization of IMC current sensors, accuracy and linearity
are important. Accuracy and linearity are critical to achieving miniaturization of
current sensors. In designing, external magnetic field shielding, internal magnetic
flux density, magnetic saturation, and skin effect should be considered.
The U-shaped shield structure, which is generally used, was designed as a reference
model. In order to miniaturize the IMC current sensor, it is necessary to reduce the
width and height of the shield. But if both are reduced, it is difficult to miniaturize
because the internal magnetic flux density and the external magnetic shielding error
rate performance are not satisfied. In this paper, we design a shield that satisfies
all four conditions to be considered and can be miniaturized by modifying the shape
of the base model.
2. U-Shaped Shield for IMC Hall Current Sensor
An IMC hall current sensor used in this paper has a thin soft magnetic material called
a concentrator inside a hall integrated circuit (IC) chip as shown in Fig. 1.
This soft magnetic material serves to concentrate the magnetic flux around the hall
sensor like the core of the conventional core type hall current sensor. Also, a hall
element is placed under two concentrators to change the magnetic flux entering parallel
to the hall element vertically. As a result, the hall sensor can be placed parallel
to the magnetic flux direction as shown in Fig. 2. Compared with conventional hall current sensors in the low current range, the IMC
hall current sensors have advantages in miniaturization of the current sensor because
they do not require a ferromagnetic core for magnetic flux concentration and can be
placed close to the measurement current source (5).
Fig. 1. IMC hall current sensor
Fig. 2. Comparison of current sensor position according to current sensor type, (a) conventional hall current sensor, (b) IMC hall current sensor
2.1 Shield Design Considerations
In order to eliminate measurement error caused by other power sources around the sensor
or external magnetic field in the high current system, the IMC hall current sensor
uses a shield. The shield has an important role in determining the performance and
size of an IMC hall current sensor. The shield of an IMC hall current sensor should
be optimized to ensure the effects of four considerations. First, the external magnetic
field shielding rate should be checked so as not to affect the magnetic field generated
outside the power source to be measured. The external magnetic field shielding error
rate is as follows.
where $B_{s}$ is the magnetic flux density of the shield center air gap measured under
single-phase conditions, and $B_{t}$ is the magnetic flux density measured when three
current sensors are side by side under three-phase conditions.
Second, the magnetic flux density inside the shield should not exceed 25mT to prevent
magnetic saturation of the concentrator made of soft magnetic material. Third, the
internal magnetic flux density of the shield should not be saturated. Finally, the
error rate due to the skin effect on the measuring busbar should be considered. The
skin effect error rate is expressed by equation (2).
where $B_{ts}$ represents the magnetic flux density measured when the skin effect
occurs on the bus bar under three-phase conditions.
In this study, with the size of a standard external magnetic field shielding rate
and skin effect error rate to about 1% was in progress the design.
2.2 Characteristic Analysis of U-Shaped Shield
Fig. 3 shows that the currently mass-produced U-shaped shield is applied to the HEV inverter
3-phase bus bar. Table 1 shows the current condition of 3-phase bus bar and specification of U-shaped shield.
The maximum phase current of the HEV inverter used in this study is 600A. To confirm
an influence of the external magnetic field on the V phase, the strongest current
condition is -300A for the U phase and V phase bus bars and 600A for the W phase bus
bar. Under the current conditions, finite element analysis is performed to verity
the performance of the U-shaped shield and to analyze the effect of the width and
height changes of the shield on external magnetic shielding rate and skin effect error
rate.
Fig. 3. U-shaped shield for three phase bus bar in HEV inverter drive.
Table 1. Specification of U-shaped shield
|
Design factor and parameters
|
Symbol
|
Values
|
|
Shield width
|
W
|
30.0mm
|
|
Shield height
|
H
|
22.0mm
|
|
Shield thickness
|
T
|
1.0mm
|
|
Bus bar width
|
$B_{w}$
|
28.0mm
|
|
Bus bar thickness
|
$B_{t}$
|
2.5mm
|
|
Bus bar interval
|
$B_{i}$
|
9.4mm
|
|
U-phase current
|
$I_{u}$
|
-300A
|
|
V-phase current
|
$I_{v}$
|
-300A
|
|
W-phase current
|
$I_{w}$
|
600A
|
Fig. 4. U-shaped shield performance with respect to variation of shield width, (a) magnetic field shielding error of V-phase, (b) flux density in inner space of W-phase shield, (c) skin effect error of W-phase
Fig. 4 shows comparison of external magnetic shielding rate, the magnetic flux density of
the shield internal space, and skin effect of error rate as the shield width becomes
narrower at a given current condition. Fig. 4 (a) shows external magnetic shielding rate of the V-phase bus bar, and it can be confirmed
that the external magnetic shielding effect is improved as the shield width is narrowed.
Fig. 4 (b) and Fig. 4 (c) shows the magnetic flux density and skin effect of the shield internal space of the
W phase on which the largest current flows. As the shield width is decreased, the
magnetic flux density in the inner space is increased to 25mT, causing the concentrator’s
magnetic saturation problem, but the skin effect error rate is improved. As a result,
the reduction of the shield width has a positive effect of reducing the external magnetic
shielding rate and skin effect error rate, but it causes the problem of magnetic saturation
of the concentrator. Therefore, to miniaturize IMC hall current sensor, magnetic saturation
problem of concentrator should be improved when shield width is reduced.
Fig. 5. U-shaped shield performance with respect to variation of shield height, (a) magnetic field shielding error of V-phase, (b) flux density in inner space of W-phase shield
Fig. 5 (a) and 5 (b) show the change of the external magnetic shielding rate of the V-phase shield and
the internal magnetic flux density of the W-phase shield when the shield height is
changed from 12mm to 30mm at the shield width of 30mm. As the shield height decreases,
the external magnetic field shielding error rate increases and the internal magnetic
flux density decreases slightly. From the results of figures 4 and 5, the miniaturization
of the shield should be designed in consideration of the problem of the internal magnetic
flux density due to the reduction of the shield width and the increase of the shielding
error rate due to the reduction of the shield height.
3. Compact Design of Shield for IMC Hall Current Sensor
Fig. 6 shows the shield shape proposed in this study for miniaturization of an IMC hall
current sensor for HEVs. Unlike U-shaped shield structure, it has a top shield that
covers the upper part of bus bar and has a structure with four air gaps between the
tope shield. The role of top shield is to allow most of the main flux created in the
bus bar to flow through the shield. Therefore, the magnetic flux density of the internal
space can be reduced, and the magnetic field can sufficiently push the external magnetic
field to prevent the external magnetic field from penetrating the inside. Four air
gaps reduce the amount of magnetic flux produced by the bus bar, preventing the shield
from saturation.
The width and height of the shield proposed in this study are fixed at 15mm and the
performance of the shield is verified under the current conditions of Table 1. In addition, four models are selected as shown in Fig. 7 to confirm the performance according to the position of top shield. Fig. 8 compares the analysis results of the internal magnetic flux density and the external
magnetic shielding rate of the four models.
Fig. 6. Proposed IMC hall current sensor, (a) shield structure, (b) magnetic flux line
Fig. 7. Classification of shield design model by top shield position, (a) model 1 (b) model 2, (c) model 3, (d) model 4.
Fig. 8 (a) shows the magnetic flux density of the inner space of the shield of the W-phase bus
bar with a maximum current of 600A. Although the shield width is reduced from 30mm
to 15mm compared to the U-shaped shield, the internal flux density satisfies the condition
of less than 25mT regardless of the model. Fig. 8 (b) shows the external magnetic field shielding error rate of the V-phase bus bar, and
even though the shield height is reduced to 15mm, the other models expect the model
3 satisfy the error rate within 1%. In this study, the model 1 which is easy to manufacture
is selected as the final model and optimized.
Fig. 8. Performance comparison of four proposed shield design models, (a) flux density in inner space of W-phase shield, (b) magnetic field shielding error of V-phase
Fig. 9. Design parameters of model 1
The shield design parameters for the miniaturization of model 1 are shown in Fig. 9, and the parameters are selected as three types: shield width, height, and top shield
width. The design range of the shield width and height is from 8mm to 14mm considering
the size of the bus bar, and the design range of the top shield is selected from 2mm
to 6mm. When each design variable is changed by 1mm interval, 245 design combinations
are obtained. In order to reduce the number of design combinations, the main effect
of the internal magnetic flux density and the external magnetic shielding error rate
is performed. In the case of the error rate due to the skin effect, the smaller the
shield width, the more advantageous. Therefore, the skin effect error rate is performed
only for the final design of the model 1.
Fig. 10. Main effect analysis of shield width, height, and top shield width, (a) internal magnetic flux density, (b) external magnetic shielding error rate.
Fig. 10 shows the main effect analysis results for three design variables, shield width and
height, and top shield width. The results show that the internal magnetic flux density
is sensitive to the shield width and the variable value of the top shield, and the
external magnetic shielding error rate is sensitive to all three variables. By combining
the two results, the most sensitive areas of the three variables are adjusted to the
final design range, reducing the number of cases with 27 design combinations. Table 2 compares the analysis results of the internal magnetic flux density and the external
magnetic field shielding error ratio for 27 combinations of design variables. From
the Table 2, the width and height of the shield are 10mm and 8mm, respectively, and the top shield
width 5mm is selected as the final model. Fig. 11 shows the results of the analysis of the skin effect error rate of the final selected
model. The maximum bandwidth of 600 ampere class IMC current sensor has 50kHz to 100kHz
and the skin effect error rate is analyzed in this area. The greater the distance
from the bus bar, the less susceptible to the skin effect and the minimum error rate
at 3.65mm from the bus bar. Regardless of the distance to the bus bar and the frequency,
the skin effect error rate is within 1% and the error rate within 0.4% at the point
where the actual hall censor is mounted.
Table 2. Specification of U-shaped shield
|
Design variables
|
Constraints
|
|
|
W
(mm)
|
H
(mm)
|
$L_{ms}$
(mm)
|
Flux density
(mT)
|
Error Rate
(%)
|
|
1
|
10
|
8
|
4
|
29.1
|
0.08
|
|
2
|
10
|
8
|
5
|
24.1
|
0.27
|
|
3
|
10
|
8
|
6
|
21.1
|
0.97
|
|
4
|
10
|
9
|
4
|
28.2
|
0.21
|
|
5
|
10
|
9
|
5
|
23.2
|
0.27
|
|
6
|
10
|
9
|
6
|
20.0
|
0.78
|
|
7
|
10
|
10
|
4
|
26.4
|
0.31
|
|
8
|
10
|
10
|
5
|
21.4
|
0.31
|
|
9
|
10
|
10
|
6
|
18.2
|
0.90
|
|
10
|
11
|
8
|
4
|
28.0
|
0.01
|
|
11
|
11
|
8
|
5
|
23.0
|
0.05
|
|
12
|
11
|
8
|
6
|
19.8
|
0.98
|
|
13
|
11
|
9
|
4
|
27.2
|
0.04
|
|
14
|
11
|
9
|
5
|
22.1
|
0.24
|
|
15
|
11
|
9
|
6
|
18.8
|
0.91
|
|
16
|
11
|
10
|
4
|
25.5
|
0.25
|
|
17
|
11
|
10
|
5
|
20.4
|
0.47
|
|
18
|
11
|
10
|
6
|
17.2
|
0.83
|
|
19
|
12
|
8
|
4
|
27.1
|
0.05
|
|
20
|
12
|
8
|
5
|
22.0
|
0.53
|
|
21
|
12
|
8
|
6
|
18.8
|
0.98
|
|
22
|
12
|
9
|
4
|
16.5
|
0.14
|
|
23
|
12
|
9
|
5
|
21.3
|
0.17
|
|
24
|
12
|
9
|
6
|
17.9
|
0.79
|
|
25
|
12
|
10
|
4
|
25.0
|
0.11
|
|
26
|
12
|
10
|
5
|
19.8
|
0.13
|
|
27
|
12
|
10
|
6
|
16.4
|
0.74
|
Fig. 12 shows the IMC hall current sensor with U-shaped shield designed for the current conditions
in Table 1 and the IMC hall current sensor with shield proposed in this study. When the proposed
shield is applied to the IMC hall current sensor, it is confirmed that it satisfies
all four current sensor design considerations: external magnetic shielding rate, internal
magnetic flux density, shield saturation, and skin effect error rate and it is possible
to reduce the size by 83.7% compared with the conventional U-shaped shield. A detailed
performance comparisons improved from the initial U-shaped model are shown in Table 3. From the comparison, in addition to the volume reduction, the external magnetic
shielding error rate is improved by 0.33% and the sikin effect error rate was improved
by 2.85%. The proposed shield structure is fabricated by inserting shield pieces into
plastic injection molding.
Fig. 11. Skin effect error rate of the final shield model
Fig. 12. Size comparison of two shield type IMC hall current sensors, (a) conventional U-shaped shield, (b) proposed shield
Table 3. Performance comparison between the initial model and proposed model
|
Design factor
|
Initial Model
|
Proposed Model
|
|
Shield width, W
|
mm
|
30
|
10
|
|
Shield height, H
|
mm
|
22
|
8
|
|
Shield thickness
|
mm
|
1
|
1
|
|
Bus bar width
|
mm
|
2.5
|
2.5
|
|
Bus bar interval
|
mm
|
9.4
|
9.4
|
|
Maximum current
|
A
|
600
|
600
|
|
Air gap length, g
|
mm
|
-
|
0.5
|
|
Top middle shield length, Lms
|
mm
|
-
|
5.0
|
|
Shield area
|
mm2
|
736.0
(100%)
|
120.0
(16.3%)
|
|
Internal magnetic flux density
|
mT
|
25.0
|
24.1
|
|
External magnetic shielding error rate
|
%
|
0.6
|
0.27
|
|
Skin effect error rate
|
%
|
3.2
|
0.35
|
4. Conclusion
This paper proposed shape design for miniaturization of IMC hall current sensor shield
for HEV. An effect of the shield parameter change on the existing U-shaped shield
structure was analyzed and it was confirmed that the miniaturization of the U-shaped
shield required improvement of the internal magnetic flux density and the external
magnetic shielding error rate. In this paper, a new shield structure is proposed to
improve the problems in miniaturizing the shield. The proposed shield structure is
optimized in consideration of the four shield performance indicators; external magnetic
shielding rate, internal magnetic flux density, shield saturation, and skin effect
error rate. Through the orthogonal array table and the main effect analysis method,
which are one of the methods of the experimental design method, the miniaturization
design work was carried out by simplifying the number of experiments and it was confirmed
that the proposed model as the final miniaturization model satisfied the fo ur design
considerations. The proposed model confirmed that the area of the shield is reduced
by 83.7% than the initial model, and it is verified that the shape conforms to the
miniaturization.
Acknowledgements
This research was supported by Basic Research Laboratory through the National Research
Foundations of Korea funded by the Ministry of Science, ICT and Future Planning (NRF-2015R1A4A1041584).
References
Global Automotive Industry Outlook 2009: Impact of Economic Slowdown on the Future
of Auto Sales and Production, Frost & Sulivan, 2009, p31, online edition accessed
June 2010
Kim Ho-Gi, Kang Gu-Bae, Nam Dong-Jin, 2009, Coreless hall current sensor for automotive
inverters decoupling cross-coupled field, Journal of Power Electronics 9.1, pp. 68-73
Palumbo V., 2013, Hall current sensor IC with integrated Co-based alloy thin film
magnetic concentrator, EPJ Web of Conferences, EDP Sciences, Vol. 40
Popovic R. S., 2002, Hall ASICs with integrated magnetic concentrators, Proceedings
of the Sensors Expo & Conference, Boston, USA.
Kim Ho-Gi, Kang Gu-Bae, Nam Dong-Jin, 2009, Coreless hall current sensor for automotive
inverters decoupling cross-coupled field, Journal of Power Electronics 9.1, pp. 68-73
Imamura , Masaaki , 1998, Analysis of magnetic fields due to three-phase bus bar currents
for the design of an optical current transformer, IEEE transactions on magnetics 34.4,
pp. 2274-2279
Biography
He received M.S degree in electrical engineering from Pusan national university, Busan,
Korea, in 2017.
He is currently working toward Ph.D. degree.
His research interests are design and control of electric motor.
He received M.S degree in electrical engineering from Pusan national university, Busan,
Korea in 2017.
He is currently working toward Ph.D. degree.
His research interests are design and control of electric motor.
He received the B.S. and M.S. degrees in electrical engineering from Pusan national
university, Busan, Korea, in 1996 and 1998, respectively.
He received Ph.D. in electrical and computer engineering at university, Blacksburg,
USA.
He is an professor of electrical engineering at Pusan national university.
His research interests include the design of electric machines and their optimal operation.