Richard Dickinson and Shaun Milano 7/11/2002

Isolated Open Loop Current Sensing Using Hall Effect

Technology in an Optimized Magnetic Circuit

Richard Dickinson & Shaun Milano

Allegro MicroSystems, Inc.

162 Pembroke Rd.
Concord NH, 03301 USA
Tel: 001 (603) 228-5533
Fax: 001 (603) 224-2466

Abstract:

With the expected arrival of a 42V parallel bus power supply aboard automobiles and
new energy efficiency standards being imposed on most household appliances, there is a growing
need for current sensing as a means of monitoring and controlling power consumption.
There are three rival technologies that are typically used for measuring current: sense
resistors, Hall effect sensors and current transformers. Each have attributes that differentiate
them on a cost versus performance scale. Galvanic isolation, ease of implementation, robustness
and low cost are a few attributes that users demand when choosing a current sensor.
Galvanic isolation is needed to protect the sensing device from potentially damaging high
power signals (over-current spikes) and to minimize the power dissipated (and heat generated)
by sense resistors. Because Hall effect sensors measure the magnetic field strength in close
proximity to the current conductor, they can be separated by a few millimeters from the current
signal, providing several kilovolts of isolation. Other open loop Hall effect current sensor
designs internalize the current carrying conductor, allowing the user to optimize the current
sensor package’s size and thermal characteristics.
While closed loop sensing can provide excellent accuracy, size and cost prohibit the use of these
types of transducers in many applications. Open loop sensors usually have a more limited range
of linearity and cannot compensate for offset and residual field errors. However, magnetic
circuits and the linear Hall effect IC can be designed to reduce these types of errors without the
external control circuitry that is typically required by closed loop systems.

This paper will explore magnetic circuit design, IC circuit designs, and integrated packaging
techniques that attempt to minimize the errors associated with open circuit designs in order to
achieve a sensor with accuracy equivalent to that of sense resistors, but without the isolation
tradeoff.

Page 1 of 12
Shaun Milano Page 2 7/11/2002

Introduction
Applications requiring current sensing are becoming ubiquitous, from battery
management systems aboard automobiles to dc motor controllers within household
appliances. While sense resistors may have previously represented the state of the art in
current sensing, the need for safe, isolated detection of electrical current has spurred the
development of non-intrusive current sensing methods and devices. Among a handful of
alternative current-sensing methods, Hall-Effect sensors may be implemented in high
volume with relatively low cost for many of these applications.

Current Sensing Technology Overview

There are three technologies that are typically used for measuring current: sense
resistors, current transformers and Hall effect sensors.
Sense resistors are simply a resistor placed in series with the load. By ohms law,
the voltage drop across the device is proportional to the current. For low currents, these
provide very accurate measurement given the resistance value has a tight tolerance.
Although sense resistors with high performance thermal packages have been developed
for larger currents, they still result in insertion loss. In addition, they do not provide a
measurement isolated from transient voltage potentials on the load. Sense resistors also
require other circuitry such as instrumentation amplifiers to generate a distinguishable
signal.
Current transformers are relatively simple to implement and are passive devices
that do not require driving circuitry to operate. The primary current (AC) will generate a
magnetic field that is coupled into a secondary coil by Faraday’s Law. The magnitude of
the secondary current is proportional to the number of turns in the coil, which is typically
as high as >1000. The secondary current is then sensed through a sense resistor to
convert the output into a voltage.
There are two techniques for sensing current using Hall effect devices. According
to the Hall effect, a magnetic field passing through a semiconductor resistor will generate
a differential voltage proportional to the field (figure 1).
Figure 1: Representation of the Hall effect and its electrical parameters

Concentric magnetic field lines are generated around a current carrying conductor.
Approximating the primary current conductor as infinitely long, the magnetic field
strength may be defined B = µo I/ 2pr, where µo is the permeability of free space, I is the
current and r is the distance from the center of the current conductor. In order to induce a
larger signal out of the Hall element; the current conductor may be wrapped around a

Page 2 of 12
Shaun Milano Page 3 7/11/2002

slotted ferrous toroid N number of times, such that B = µo NI/ 2pr. In an open loop
topology, the Hall element output is simply amplified and the output is read as a voltage
that represents the measured current through a scaling factor as depicted in figure 2.
Figure 2: Basic Topology of Open Loop Hall Effect Current Sensor

In a closed loop topology, the output of the Hall element drives a secondary coil that will
generate a magnetic field to cancel the primary current field. The secondary current,
scaled proportional to the primary current by the secondary coil ratio, can then be
measured as voltage across a sense resistor.

Figure 3: Basic Topology of Closed Loop Hall Effect Current Sensor

By keeping the resultant field at zero, the errors associated with offset drift,
sensitivity drift and saturation of the magnetic core will also be effectively canceled.
Closed-loop Hall effect current sensors also provide the fastest response times. However,
with a secondary coil that may be needed to drive up to several milli-amps of current,
power consumption is much higher in closed loop Hall effect devices than open loop
topologies. The closed loop configuration also limits the magnitude of the current that

Page 3 of 12
 Shaun Milano Page 4 7/11/2002

can be sensed since the device may only drive a finite amount of compensation current.
Figure 4 provides a simple comparison of different current sensing techniques.

Figure 4: Summary of current sensing techniques and attribute comparisoni

Current-sensing Accuracy Galvanic Power Relative Typical current
1 2
Technique Isolation dissipation Cost ranges
Sense resistor >95% None High Low <20A, DC-100kHz
Transformer ~95 Yes Moderate Med. Up to 1000A, AC
Open loop Hall 90-95% Yes Low Med. Up to 1000A, DC-
effect sensor 20kHz
Closed loop Hall >95% Yes Moderate- High <500A, DC-150kHz
effect sensor High

Design Considerations for Open Loop Hall Effect Current Sensors

The measurable current range, the output linearity and measurement accuracy are
attributes that influence the design of a current sensor package. The linearity and
accuracy are determined through the magnetic circuit that couples the primary current to
the magnetic field sensed by the Hall element. Offsets and inaccuracy introduced in the
magnetic circuit can often be minimized in the linear Hall effect IC design.
The measurable current range often dictates the size of an open loop current
sensor. For current levels less than100A, a reduced package size can be achieved by
internalizing the primary conductor. Combining the Hall effect sensor, flux concentrator
and primary conductor into a single assembly for PCB installation opens up applications
that previously relied on sense resistors. For current levels of100A and higher, any
discontinuity in the bus-bar could present reliability risks. Therefore, rather than
soldering or welding on a current sensor, end users desire a package that can be
clamped/installed around the conductor. If the primary conductor is internalized into the
package the design objective is to minimize the insertion losses. Any power dissipated
through the primary conductor will generate heat in close proximity to the Hall effect IC
and will be seen as a temperature coefficient. Figure 5 provides an example of a
proposed 100A full-scale current sensor package integrating the primary conductor.

1
Including power loss through primary current carrying conductor & power consumed by device
2
Estimated, including cost of auxillary circuitry needed to implement device in operating environment

Page 4 of 12
Shaun Milano Page 5 7/11/2002

Figure 5: Open loop Hall effect current sensor integrating primary conductor
Open loop Hall effect current sensor package under development at Allegro Microsystems, Inc.

In addition to the tradeoff between insertion losses and size for a given current
range, there is also a relationship between the flux concentrator and the measurable
current range. The magnetic field sensed by the Hall effect transducer is proportional to
the product of the primary current and a gain factor related to a flux concentrator. In
order to keep a magnetic circuit in its linear region for a given current, the dimensions of
the flux concentrator and its effective magnetic path length, must be appropriately scaled.
A typical design rule is to use a toroid with a cross sectional area that is twice as large as
the area seen in the gap where the sensing element sits. Choosing a magnetic flux
concentrator material with the appropriate effective permeability can be determined as a
function of the primary current range,

µe = Bl e/0.4pNI

where N is the number of turns (typically 1), el is the effective path length and I is the full
scale current. The effective path length is

lg = le(1/µe - 1/µI)(0.397)

As the primary current range increases, it is often necessary to increase the air gap size to
ensure a linear output. A material commonly used as a flux concentrator is powdered
iron, having magnetic permeability 2000 to 5000 times greater than air. Figure 6
provides a diagram of an open loop Hall effect sensor package, with a ferrous core
concentrating flux onto a Hall effect IC.

Page 5 of 12
Shaun Milano Page 6 7/11/2002

Figure 6: Open loop Hall effect current sensor magnetic circuit topology
Computer simulation screen capture of open loop Hall effect current sensor package

Another property of the magnetic flux concentrator that affects the linearity of the
open loop output is hysteresis. In addition to having high permeability, ferromagnetic
materials tend to retain some magnetic induction after an applied field is removed. This
leads to inaccuracy at low current levels. In order to minimize this error, it is best to use
materials with low magnetic remenance, Br and low magnetic coercivity, Hc. Soft ferrites
and round loop NiFe alloys provide the closest balance between high permeability and
low hysteresis. Another way to minimize hysteresis-offset error is to scale the toroid
such that the entire core will not saturate until current levels well beyond the intended
full- scale current.
The voltage generated on the Hall element is filtered and amplified into an output
voltage proportional to the sensed primary current. Techniques such as chopper
stabilization, (alternating the Vcc and sense terminals on the Hall element) help to
minimize offset voltages and zero current error. Factory trimming of resistors within the
IC adjusts the gain so that the linear region of the amplifier matches the linear region of
the magnetic circuit. The block diagram in Figure 7 summarizes the functionality
provided by a typical linear Hall effect sensor.

Page 6 of 12
Shaun Milano Page 7 7/11/2002

Figure 7: Functionality of linear Hall effect sensor used in Allegro ACS750KCA-

100 Current sensor

Unfortunately, the linear Hall effect sensor cannot compensate and remove all
sources of error in an open loop current sensor. At zero amps, magnetic hysteresis and
temperature drift of the quiescent voltage will act as a source of error. At full scale
current, the non-linearity of the flux concentrator, the resolution of the sensitivity trim
and temperature coefficient from the packaged device will also create inaccuracy.
Transfer curves of Iprimary versus Vout are presented in figure 8.

Page 7 of 12
Shaun Milano Page 8 7/11/2002

Figure 8a: Typical Iprimary versus Linear Hall-Effect Transducer Output

Across full-scale and current operating conditions

Accuracy
Full Scale
25°C Only
Vout (Volts)

Accuracy
Average
Full Scale
Output
Over Temp

Accuracy
0Amp
25°C Only

I (Amps)
Accuracy +/- Full Scale
0Amp
Over Temp

Accuracy
Full Scale
25°C Only

Accuracy
Full Scale
Over Temp

Page 8 of 12
Shaun Milano Page 9 7/11/2002

Figure 8b: Typical Iprimary versus Vout, Allegro ACS750LCA-075

Iprimary vs V OUT
Allegro ACS750LCA-075 Current Sensor
5
4.5
4
3.5
3
Vout

2.5
2 -40C

1.5 20C
85C
1 150C

0.5
0
-100 -80 -60 -40 -20 0 20 40 60 80 100
Amps

The dynamic accuracy of a current sensor is another important attribute in

applications in which the primary current is a transient signal. Although open loop Hall
effect sensors are frequency limited compared to sense resistors and closed loop
transducer topologies, there are design techniques that can improve their performance.
As with many other performance traits, the composition of the flux concentrator is a
relevant factor. According to Lenz’ Law, dB/dt = µdI/dt, any change in magnetic field
will induce eddy currents which produce magnetic fields opposite to the externally
applied field. Any change in magnetic field will induce in the material, “a voltage that is
in the opposite direction to the voltage producing the magnetizing current and the
alternating magnetic field. The induced voltage will set up circular currents in the
material which produce magnetic fields opposite to the original magnetic field.”ii The
power losses from eddy currents in a given material is expressed by

Pe = Bm2 f2 d2 /?

Where Bm is the maximum magnetic induction, f is the frequency, d is the smallest

dimension transverse to flux and ? is the material resistivity.
Since ferrites typically have high resistance, power losses are not a concern through the
frequency range that open loop Hall effect current sensors operate. However, other core
materials, such as steel alloys, are low resistance mediums may suffer significantly from
eddy current losses. Fortunately, without significantly altering the magnetic reluctance,
breaking a core into several thin layers (such that the value of d in the aforementioned
equation approaches zero) can increase the transverse electrical resistance by an order of
magnitude. For this reason, laminated cores are often used in current sensor packages
requiring some degree of dynamic accuracy.

Page 9 of 12
Shaun Milano Page 10 7/11/2002

Overview of Current Sensing Applications

Open-loop Hall effect current sensors are most appropriate for applications in
which galvanic isolation is absolutely necessary. Battery management circuits running
directly off of automobile batteries is one instance when isolating the primary current
from the supply powering the sensing device is needed.
Monitoring electrical current flow is an excellent method of gauging the
performance of motor-driven devices (such as fans, pumps, and chillers) controlled by a
building automation system (BAS). The torque supplied by a shunt motor is
proportional to the armature current, in which T = k? Ia. The speed of a dc motor can
also be monitored and controlled through the current driving it, in which S = [Va – (IR)]/
k? . Solid-state current switches can be used to provide a contact closure to show that a
motor is running, but many times it is helpful to be able to accurately measure current as
a diagnostic tool. The power consumption through an appliance can provide insight into
any component degradation or failure. For example, lower than normal current flow may
indicate problems such as a broken coupling on a pump, a broken or loose belt on a fan, a
dirty strainer on a pump application, or a dirty filter. A higher than normal current flow
may indicate bad or worn bearings, or belts that misaligned on motor loads. A monitored
current can also be used to determine the percent loading of a chiller.iii
Uninterruptible Power Supplies (UPS) are appliances that require current sensing
applications. Figure 9 provides a block diagram for one-type of UPS configuration, an
on-line or inverter-preferred topology that provides a regulated sinusoidal output under
any line condition.iv Regardless of the configuration, all UPS devices require fault
protection and current level measurement as inputs to their control circuitry.

Page 10 of 12
Shaun Milano Page 11 7/11/2002

Figure 9: Block diagram of UPS highlighting current sensing applications within

systemv

FUSE
AND DC BUS INVERTER OUTPUT
PFC
LINE CAPS POWER FILTER
AC FILTER STAGE AC
IN OUT

PWM1,
PWM2 PWM5,
PWM6
BATTERY
PWM4 BOOST
Is, Vs

SENSE GATE
BATTERY
AMPLIFIER DRIVE

Io, Vo

GATE BATTERY PWM3 GATE SENSE

DRIVE CHARGER DRIVE AMPLIFIER

Ib, VB V+, V-

SENSE AMPLIFIER

PWM1, PWM4 ADC11, ADC14, PWM3 PWM5,

PWM2 ADC7 ADC5 PWM6
ADC10,
ADC3,
ADC6
ADC12
DSP Controller FAULT
(TI TMS320C240) PDPINT DETECT

Page 11 of 12
Shaun Milano Page 12 7/11/2002

Conclusion
The number of applications requiring current sensing is continuing to expand.
Depending on the demands of an application such as the primary current range, the
desired steady state and dynamic accuracy and cost pressures, a number of different
techniques for sensing current are available. By providing galvanic isolation and a
relatively low cost solution, open loop Hall effect sensors are well positioned to meet the
needs of many of these applications.
Sources consulted

i
Paul Emerald, “’Non-Intrusive’ Hall-Effect Current-Sensing Techniques Provide Safe, Reliable Detection
and Protection for Power Electronics,” Allegro Microsystems, Inc. Technical Paper STP 98-1, page 2.
ii
Alex Goldman, Handbook of Modern Ferromagnetic materials , Boston: Kluwer Academic Publishers,
1999, 59.
iii
“Using Current Monitoring for Load Analysis,” Kele Technical Reference PM6, accessed 062802,
http://www.kele.com/Tech/Monitor/Power/TRefPM6.html.
iv
Shamin Choudhury, “Implementing Triple Conversion Single Phase On-line UPS using TMS320C240,”
Texas Instruments Application Report, 6-30-1999, page 2, accessed 062702,
http://wwws.ti.com/sc/psheets/spra589a/spra589a.pdf
v
Diagram adapted and modified from Figure 2 of Shamin Choudbury’s Texas Instruments Application
Report “Implementing Triple Conversion Single Phase On-line UPS using TMS320C240.”

Page 12 of 12