Current Sensors Design Guide Application Note Melexis
Current Sensors Design Guide Application Note Melexis
CURRENT SENSORS
REFERENCE DESIGN GUIDE
Application Note
TABLE OF CONTENTS
INTRODUCTION ................................................................................................................................ 3
HALL-EFFECT CURRENT SENSORS ..................................................................................................... 3
HALL EFFECT ..................................................................................................................... 3
HALL EFFECT CURRENT SENSORS ........................................................................................... 4
TYPES OF HALL SENSORS USED WITH FERROMAGNETIC MATERIALS .............................................. 5
Conventional Hall sensors.............................................................................................. 5
Planar IMC-Hall® sensors .............................................................................................. 5
CONVENTIONAL HALL SENSORS........................................................................................................ 6
INTRODUCTION ................................................................................................................. 6
QUICK SELECTION GUIDE .................................................................................................... 7
IMC-HALL® SENSORS ...................................................................................................................... 10
IMC VERSIONS ................................................................................................................ 11
MAGNETIC DESIGN........................................................................................................... 12
Magnetic field estimation............................................................................................ 12
Magnetic field estimation with shield ......................................................................................................... 13
Magnetic field estimation without shield ................................................................................................... 14
Conductor with neck-down.......................................................................................... 14
Stray field immunity .................................................................................................... 15
Stray field reduction ..................................................................................................................................... 15
Stray field rejection ...................................................................................................................................... 16
Use of laminated shields.............................................................................................. 17
Mechanical tolerances and vibrations ......................................................................... 17
Cancelling stray field without shield ............................................................................ 18
Avoiding cross-talk without shield ............................................................................... 18
PCB DESIGN ................................................................................................................... 19
REFERENCE DESIGNS ........................................................................................................ 20
PCB application, 2-10A range, multi-layer/multi-turn solution.................................... 21
PCB application, 2-10A range, ferromagnetic shield solution...................................... 22
PCB application, 10-50A range .................................................................................... 22
Bus bar application, 50-250A range ............................................................................ 23
Bus bar application, 300-700A range .......................................................................... 24
Bus bar application, dual range 5A/200A .................................................................... 25
Cable application, 10-100A range ............................................................................... 25
QUICK SELECTION GUIDE .................................................................................................. 26
FERROMAGNETIC MATERIALS......................................................................................................... 27
FERROMAGNETIC SUPPLIERS .............................................................................................. 27
FERROMAGNETIC SHIELDS ................................................................................................. 28
U-Shield ....................................................................................................................... 28
Mechanical assembly .................................................................................................. 28
FERROMAGNETIC CORES ................................................................................................... 29
FERROMAGNETIC MATERIALS COMPARISON .......................................................................... 31
END-OF-LINE CALIBRATION............................................................................................................. 31
ADDITIONAL INFORMATION ........................................................................................................... 32
CONTACT US ................................................................................................................................... 32
DISCLAIMER .................................................................................................................................... 33
Introduction
Current sensing has become more and more important in the modern world. Many applications, such
as traction inverters, electric motors, ACDC and DCDC converters, on-board chargers, battery
management systems, and many others, are relying on current sensing to improve efficiency and
safety. Melexis offers a wide portfolio of current sensors. In this document, we are going to present
the high current sensing solutions of melexis: IMC-Hall® and Conventional Hall sensors, together with
examples of assembly structures, and suggestions on how to design and build a current sensing
module. This guide is divided in 5 main sections:
1. Hall-effect current sensors: an introduction to hall effect sensors, with some basic theory.
2. Conventional Hall sensors: a description of Conventional Hall sensors, with some insight on
how to obtain an optimal design.
3. IMC-Hall® sensors: a description of IMC-Hall® sensors, an advanced technology enabling
simple assembly. This section will also focus on examples and tips to reach the optimal design.
4. Ferromagnetic materials: a description of ferromagnetic materials used in a current sensing
module, with also a focus on some assembly solutions.
5. End-of-line calibration: a description on how to reduce at minimum errors, to obtain the best
design for your application.
Where 𝑟 is the distance from the conductor and 𝜇0 is the vacuum permeability. This field can be
substituted into Equation 1, obtaining the link between the Hall voltage and the current inside the
wire (Equation 3).
𝜇0
𝑉𝐻 = 𝑆𝐻 𝐼 = 𝑆𝐻 ∙ 𝐹𝐹 ∙ 𝐼 Equation 3
2𝜋𝑟
𝜇0
The parameter 𝐹𝐹 = 2𝜋𝑟 is the field factor. For geometries different than a wire, the field factor is
computed with different equations, and in general it is equal to the ratio between the field and the
generating current (𝐹𝐹 = 𝐵/𝐼). Normally, the farther is the sensor from the conductor, the smaller
the signal will be.
Signal processing plays a fundamental role in the quality of the sensor. Thanks to the integration of
the Hall element on a CMOS processed silicon IC, we can take advantage of integrated circuits for
signal variable amplification (to obtain the desired sensitivity), offset and sensitivity correction and
temperature compensation.
To concentrate the magnetic field on the sensor, and also to reduce the influence of external fields, it
is possible to use ferromagnetic elements (cores and shields).
Conventional Hall current sensors (Figure 2) are sensitive to the magnetic field perpendicular to the
chip surface. They are meant to be used in combination with a ferromagnetic core. In a typical
application, the core is wrapped around the current conductor and concentrates the magnetic flux on
a small air gap (typically 2-5mm) where the sensor is inserted.
Pros
▪ Strong magnetic gain from the core
▪ Very robust against cross-talk
▪ Suitable for medium to very high currents
Cons
▪ Performance limited by the core (geometry
and material): saturation, hysteresis,
frequency response and thermal drift
▪ Bigger footprint (size, weight) than solutions
based on IMC-Hall® sensors
The IMC-Hall® sensor (Figure 3) includes on top of the silicon IC a ferromagnetic layer (the Integrated
Magnetic Concentrator, IMC) whose purpose is to concentrate the magnetic field on the sensing
elements. Thanks to this technology, IMC-Hall® sensors are sensitive to magnetic fields parallel to the
chip surface. Thus, the sensors can directly measure the current flowing in a bus bar or a PCB trace
below the package, without the need for a core.
Pros
▪ Sensitive to magnetic field parallel to the chip
surface, enabling an easy SMD assembly,
vertical stacking and minimum footprint.
▪ IMC is made of magnetic material featuring
very high permeability and very low
hysteresis
▪ Magnetic gain from IMC
Cons
▪ Requires magnetic shielding or specific design
to avoid cross-talk and/or noise from external
fields
Figure 3: IMC-Hall® sensor
In the following document, we will discuss how to design at best a system with these 2 types of
sensors.
I
d
𝐼 [𝐴]
𝐵 [𝑚𝑇] = 1.25 × Equation 4
𝑑 [𝑚𝑚]
The naming convention for ferromagnetic cores is C followed by the air gap dimension (for instance,
C5 for a 5mm air gap shield). A development kit (link to application note on website) is available for
designs testing.
Option code: xxA Option code: xxR Option code: xxS Option code: xxT Option code: xxZ
(a) (b) (c) (d) (e)
Figure 5: Different leads bending options: (a) straight leads, (b, c, d) bent leads, (e) and planar leads.
IMC-Hall® Sensors
IMC-Hall® sensors are based on an Integrated Magnetic Concentrator (IMC, a ferromagnetic layer
placed on top of the silicon IC) that locally converts the horizontal magnetic field (𝐵𝑥 ) into a vertical
component (𝐵𝑧 ) that can be measured by high performing silicon Hall elements. To concentrate the
field on the sensor, and to protect it from external fields, ferromagnetic shield are normally used (see
Figure 7).
Figure 7: IMC-Hall® sensor. The yellow regions show the approximate position of the IMCs.
Figure 8 shows how a horizontal field is bent to be converted into a vertical field at the hall plates
position (2 sensing points are present). This enables a differential sensing, rejecting any other field
that is not parallel to the sensor plane. IMC are available in different types, as explained in section 4.1.
Figure 8: IMC and field bending. Any field in the Z direction is rejected by the differential sensing.
IMC versions
IMC-Hall® sensors are available in 4 different versions covering a broad range of sensitivities and
magnetic field ranges: Low Field (LF), High Field (HF), Very High Field (VHF), Extra High Field (XHF).
With its strong magnetic gain, the LF IMC is ideally suited for applications with low currents, requiring
high magnetic sensitivities (up to 700mV/mT). At the other end of the scale, the XHF IMC can linearly
sense strong magnetic fields up to ±90mT, for current sensing applications with very high-power
densities. In the following pages, Figure 9 shows the different sensitivity ranges that IMCs can reach.
Figure 10 shows the linearity ranges for each one of the IMCs, and Figure 11 shows the current ranges
in which IMCs can be used.
Figure 11: Optimal IMC configuration for different current ranges. Saturation of shields should also be
considered. Melexis offers an online simulator tool, to choose the most optimized solution for a
current sensing application, the current sensor simulator.
The design of the magnetic system around the sensor is important to be able to take advantage of the
best performances of Melexis sensors.
Magnetic design
The next sections provide a guide for an optimal magnetic design of an IMC-Hall® sensing system. First,
simple rules for magnetic field estimation will be described (section 4.2.1). Then, solutions will be
described in order to improve signal to noise ratio and stray field immunity (sections 4.2.2 and 4.2.3),
to increase frequency response and linearity (section 4.2.4), to remove the effect of mechanical
tolerances (section 4.2.5), and finally to implement the IMC-Hall® sensor without ferromagnetic
shields (sections 4.2.6 and 4.2.7). For a focus on the ferromagnetic materials, see section 5.
To compute the needed sensitivity, the magnetic field at the sensor location can be estimated,
considering the configuration in which the IMC-Hall® sensor is used: with or without shield.
In a typical application, a U-shaped ferromagnetic shield is wrapped around the current conductor to
protect the sensor from external fields and improve the overall robustness of the sensing solution (see
Figure 12).
𝐼 [𝐴]
𝐵 [𝑚𝑇] = 1.25 × Equation 5
𝑊 [𝑚𝑚]
Depending on application environment and requirements, the ferromagnetic shield is not necessarily
required. However, this configuration would lead to a drastic reduction of stray field immunity.
Without shield, the magnetic field B measured by the sensor for a current I, a trace width W and a
vertical position H can be approximated as:
𝐼[𝐴] 𝑊[𝑚𝑚]
𝐵 [𝑚𝑇] = 0.4 × ∗ atan ( ) Equation 6
𝑊[𝑚𝑚] 2𝐻[𝑚𝑚]
In order to limit the cost, size and weight of the shield on systems with wide bus bars, we recommend
reducing its cross-section locally, as illustrated in Figure 15. Such a neck-down has minimal impact on
the electrical resistance and allows for a much more compact current sensing solution.
IMC Hall® sensors are protected from external or stray fields (i.e. fields that are not generated by the
current under measurement) by shields. Stray fields can be generated by external sources or, in case
of an inverter application, by the other phases, that should be measured by other sensors (cross talk).
Stray fields can affect both offset (in case of a constant field, not related to the measured current) and
sensitivity (when the external field changes with the measured current).
Stray field from an external source can be reduced first of all by increasing its distance with the sensor.
Figure 16 shows how cross talk decreases with the increase of the distance. These simulated values
are calculated for distances between 20mm and 170mm, with 2 cases: one with the sensor perfectly
centred in the middle of the shield, and one with the sensor displaced in the horizontal direction by
1mm. It is possible to notice that a centred sensor represents the best configuration to minimize the
stray field.
Figure 16: Cross talk error and stray field dependence with the increase of bus bar spacing.
When the distance cannot be increased enough to reduce the error from a stray field source, it is
necessary to use the ferromagnetic shield properties. We saw that by decreasing the shield width, the
field intensity on the sensor increases (Equation 5). Moreover, this also reduces the quantity of
external field that reaches the sensor, or, in other words, increases the ratio between the external
field 𝐵𝑒𝑥𝑡 and its component inside the shield 𝐵𝑖𝑛𝑡 , called the Shielding factor 𝑆𝐹 = 𝐵𝑒𝑥𝑡 /𝐵𝑖𝑛𝑡 (Figure
17).
Figure 18 shows how increasing the shield width reduces the shielding factor, leading to a larger
influence from stray field sources. In case the shield width is limited by the bus bar dimensions, it is
possible to adopt the necking down solution (see Figure 15), to locally reduce the bus bar thickness to
be able to use small shields.
Laminated shields are made of a stack of thin sheets separated by insulators. This allows to reduce the
effect of eddy currents (that are generated in the shield by time oscillating fields, and cause gain
reduction and phase shift of the sensor signal), boosting the sensing performances. Moreover, for very
high currents (typically above 800A), it is often more efficient and cost-effective to use a laminated
shield so that the in-plane thickness can be increased, decreasing the depth and overall footprint for
the same performance.
The mechanical assembly of the sensor and shield has tolerances, that may affect the measurements
of the system. Any static error can be compensated by end of line calibration (see section 6). However,
vibrations will introduce a time variable error of the sensitivity, since movements of the sensor with
respect to bus bar and shield affect directly the field factor. The influence of displacements in the
three Cartesian directions is shown in Figure 19, while Figure 20 shows how the reference system is
chosen on the bus bar and shield assembly.
Figure 20: definition of reference system on the bus bar and shield assembly
From Figure 19 we can observe that displacements in the Z direction affect the most the sensitivity of
the sensor. Displacements in X and Y directions influence less the error.
We can conclude that the IMC-Hall technology is very robust against mechanical tolerances in the X
and Y directions. For an optimal design it is recommended to strengthen the fixation along the Z
direction by placing the fixation screws close to the sensor and shield assembly.
In AC applications, external stray fields can be cancelled out by the microcontroller. Computing the
difference between max and min sensor output values provides a signal independent of any parasitic
DC field (see Figure 21).
Even without ferromagnetic shield, cross-talk between adjacent current tracks can be avoided by
design. Figure 22 illustrates a concept of current trace layout with slots to force the current to flow
perpendicular to the main track axis. The sensors are rotated by 90° with their sensitive axis (blue
arrow) parallel to the current trace. With such a configuration, there is virtually no cross-talk between
phases.
Figure 22: Current trace layout with slots and rotated sensors to avoid cross-talk between phases.
PCB design is also important to have an optimal current sensing solution. Section 4.3 shows some
suggestions on how to optimally integrate the sensor and to design a PCB.
PCB design
The PCB design and layout play an important role in the final performances of the current sensing
module. More specifically, two different aspects are to be considered when designing the ground layer
on the PCB.
When an application implies high voltage switching (for instance in motor control applications), an
expanded ground layer, as depicted in Figure 23, will help reducing the parasitic coupling capacitance
generated by voltage time transients (dV/dt).
Figure 23: Current sensing structure. Expanded Ground Layer on PCB in order to reduce the parasitic
coupling capacitance (coming from voltage transients).
At the same time, the ground layer can have a big impact on the response time.
If the ground layer covers all the surface of the PCB and surrounds the two legs of the shield, Eddy
currents will start to flow circularly around them, generating a counter-magnetic field which slows
down the response time of the sensor.
In order to avoid increased response time, the ground layer should be divided such that it interrupts
the circulations of Eddy currents around the shield (see Figure 24).
Figure 24: Current Sensing Structure – Designing the Ground Layer on the PCB such that Eddy Currents
circulations is interrupted
Reference Designs
Table 4 shows an overview of reference designs based on IMC-Hall® current sensors for different
target applications.
Non-intrusive current
Simple PCB with clamp-on shield
sensing from cable
wrapped around the cable
(section 4.4.7).
10-100A
Table 4: Overview of the reference designs based on planar IMC-Hall® current sensors.
In the following sections, each solution is illustrated and described in details.
This solution is based on PCB with multiple layers and trace windings (current loops) for very high
sensitivity. It can be used with or without ferromagnetic shield, depending on sensitivity and accuracy
requirements. Figure 25 and Table 5 show the different designs and sensitivities achievable.
Figure 25: PCB layout example for very high sensitivity with 6 windings on 3 layers.
This solution is based on a PCB with one layer and a single current trace (no windings). A closed
ferromagnetic shield for high magnetic gain is used to boost the current sensitivity. Figure 26 shows
the full system and some assembly details.
(a) (b)
(c) (d)
Figure 26: Top (a) and side (b) view of the full system. (c) Shield in one piece inserted through slots on
the PCB edge. and (d) shield in two parts inserted in PCB slits and assembled together.
This solution is based on a PCB with one layer and a single current trace. It can be used with or without
ferromagnetic shield (U-shaped). Two different sensitivity ranges are available: up to 60mV/A (without
shield) or up to 170mV/A (with shield). Figure 27 shows the PCB with (a) and without shield. Figure
27.c shows how to assemble the shield on the PCB.
This solution is based on the HF IMC sensor, that is mounted directly above the conductor on the PCB.
A simple, low-cost and compact U-shaped shield is mounted around the sensor to protect it from stray
fields and ensure good signal robustness against vibrations and displacements. With the dimensions
shown here, the linearity error is lower than ± 1.5A for currents up to ±250A. Figure 28 shows the
assembly detail, while Figure 29 shows the shield dimensions. Plot in Figure 30 shows an example of
output voltage, with the linearity error.
Figure 28: Demonstrator based on MLX91206 HF sensor and U12 shield with 0.8mm thickness.
20mV/A
Figure 30: Typical output and non-linearity of a sensor calibrated for ±100A.
This solution is based on the use of a VHF IMC-Hall® sensor. The measurement range can be extended
to 700A while keeping the same inner width than the one of the design described in 0, i.e. 12mm. To
obtain a linearity up to 650A, the shield thickness must be adapted from 0.8 to 1.5mm. Figure 31
shows the assembly and the shield dimensions. This compact solution allows to measure up to 700A
with a footprint of less than 2cm2. Figure 32 shows the linearity error lower than ±5A for currents up
to ±650A (i.e. no saturation of the signal).
12
1.5
Figure 31: Demonstrator based on MLX91208 VHF sensor and U12 shield with 1.5mm thickness.
This solution allows applications with a wide dynamic range. One sensor with C-shaped (closed) shield
is used for high accuracy at small currents (typical ±5A), while another sensor with U-shaped shield is
used for high saturation limit (typical ±200A). Other combinations of ranges are possible depending
on the application requirements. Figure 33 shows the assembly and shields geometry for this solution.
(a)
(b)
(c)
Figure 33: (a) Assembly. Images and sections for (b) closed shield and (c) U-shaped shield.
This solution is based on the clamp-on shield, that gathers the magnetic field around the cable and
concentrates it above the sensor package. Small air gap ensures high magnetic gain. The shield
geometry can be adapted to match various cable diameters and current ranges. Figure 34 shows the
designs, while Figure 35 shows the real implementation of this concept.
Figure 34: Cable clamp concepts (left: monolithic shield, right: two-part shield in plastic housing).
Figure 35: Implementation examples (monolithic and two-part shield in plastic housing).
Melexis has different IMC-Hall® sensors in its portfolio. The following section will introduce them,
describing their specifications.
Ferromagnetic materials
This section introduces Melexis’ main partners for ferromagnetic materials supply. Moreover, it
introduces standard dimensions for shield and core ordering.
Ferromagnetic suppliers
Melexis partnered with MagLab for ferromagnetic components supply (
www.maglab.ch
Figure 36).
www.maglab.ch
Ferromagnetic shields
U-Shield
Standard (U) and laminated (LU) shields can be ordered from the Maglab website using the following
order codes convention (valid for both types). Figure 37 shows shields image and cross sections. Table
8 shows an example of shield ordering, and Table 9 shows different shield thicknesses.
Mechanical assembly
Ferromagnetic shields can be assembled by crimping, screwing or bonding (glue or tape). They can
also be encapsulated in a pre-molded plastic case. The optimal solution depends on the application.
In any case, care should be taken to avoid mechanical stress on the part of the shield involved in the
magnetic measuring circuit. Figure 38 shows different solutions for shields assembly.
One of the most common solutions is to use a pre-molded plastic case, with slots to insert the shields,
as illustrated in Figure 39.
Ferromagnetic cores
Amorphous ferromagnetic cores can be ordered with the order codes convention shown in Figure 40.
Table 10 shows an example of core order number.
AMC1R5 – L
Figure 40: Amorphous ferromagnetic core sections, with ordering convention.
Order code example L [mm]
AMC1R5 – 5 5
Laminated cores can be ordered using the order codes convention shown in Figure 41. Table 11 shows
an example of core order number.
Ca – b – c
Figure 41: Ordering convention for laminated cores.
End-of-line Calibration
Each current sensor is individually tested and calibrated over temperature in the Melexis production
line. However, in order to achieve optimal accuracy, a final calibration is required at customer-side
after assembly to compensate for mechanical tolerances (sensor position deviations, shield
dimensions, etc.)
This final calibration can be done in two ways: either by using the Melexis tools to directly program
the sensor EEPROM, or by adjusting the gain/offset at microcontroller level.
(a) (b) (c) (d)
Figure 42 resumes how the error are added. We can see that the mechanical assembly introduces an
error in the field factor, that affect the total error. A final calibration will lower the error down to the
case of the sensor after fabrication.Figure 1
Figure 42: Sum of error of sensor and mechanical assembly. We can see that the mechanical assemble
(a) introduces an error that will affect the total error (c). An end of line calibration (d) will lower again
the error to the value of the sensor after fabrication (b).
Additional information
Please refer to the following document for additional information on specific topics:
Contact us
To get in contact with our current sensors application team, please fill and submit the technical inquiry
form.
Disclaimer
The content of this document is believed to be correct and accurate. However, the content of this document is furnished "as
is" for informational use only and no representation, nor warranty is provided by Melexis about its accuracy, nor about the
results of its implementation. Melexis assumes no responsibility or liability for any errors or inaccuracies that may appear in
this document. Customer will follow the practices contained in this document under its sole responsibility. This documentation
is in fact provided without warranty, term, or condition of any kind, either implied or expressed, including but not limited to
warranties of merchantability, satisfactory quality, non-infringement, and fitness for purpose. Melexis, its employees and
agents and its affiliates' and their employees and agents will not be responsible for any loss, however arising, from the use
of, or reliance on this document. Notwithstanding the foregoing, contractual obligations expressly undertaken in writing by
Melexis prevail over this disclaimer.
This document is subject to change without notice, and should not be construed as a commitment by Melexis. Therefore,
before placing orders or prior to designing the product into a system, users or any third party should obtain the latest version
of the relevant information.
Users or any third party must determine the suitability of the product described in this document for its application, including
the level of reliability required and determine whether it is fit for a particular purpose.
This document as well as the product here described may be subject to export control regulations. Be aware that export might
require a prior authorization from competent authorities. The product is not designed, authorized or warranted to be suitable
in applications requiring extended temperature range and/or unusual environmental requirements. High reliability
applications, such as medical life-support or life-sustaining equipment or avionics application are specifically excluded by
Melexis. The product may not be used for the following applications subject to export control regulations: the development,
production, processing, operation, maintenance, storage, recognition or proliferation of:
1. chemical, biological or nuclear weapons, or for the development, production, maintenance or storage of missiles for such
weapons;
2. civil firearms, including spare parts or ammunition for such arms;
3. defense related products, or other material for military use or for law enforcement;
4. any applications that, alone or in combination with other goods, substances or organisms could cause serious harm to
persons or goods and that can be used as a means of violence in an armed conflict or any similar violent situation.
No license nor any other right or interest is granted to any of Melexis' or third party's intellectual property rights.
If this document is marked “restricted” or with similar words, or if in any case the content of this document is to be reasonably
understood as being confidential, the recipient of this document shall not communicate, nor disclose to any third party, any
part of the document without Melexis’ express written consent. The recipient shall take all necessary measures to apply and
preserve the confidential character of the document. In particular, the recipient shall (i) hold document in confidence with at
least the same degree of care by which it maintains the confidentiality of its own proprietary and confidential information,
but no less than reasonable care; (ii) restrict the disclosure of the document solely to its employees for the purpose for which
this document was received, on a strictly need to know basis and providing that such persons to whom the document is
disclosed are bound by confidentiality terms substantially similar to those in this disclaimer; (iii) use the document only in
connection with the purpose for which this document was received, and reproduce document only to the extent necessary for
such purposes; (iv) not use the document for commercial purposes or to the detriment of Melexis or its customers. The
confidentiality obligations set forth in this disclaimer will have indefinite duration and in any case they will be effective for no
less than 10 years from the receipt of this document.
This disclaimer will be governed by and construed in accordance with Belgian law and any disputes relating to this disclaimer
will be subject to the exclusive jurisdiction of the courts of Brussels, Belgium.
The invalidity or ineffectiveness of any of the provisions of this disclaimer does not affect the validity or effectiveness of the
other provisions.
The previous versions of this document are repealed.
Melexis © - No part of this document may be reproduced without the prior written consent of Melexis. (2022)