CHAPTER 34

Transducers and Sensors

The terms transducers, detector, and sensors are debatably interchanged depending on the country,
company, and technological application area.
A transducer, by its very name, is an electronic device that measures a physical quantity, therein
converting energy from one form to another. But in power electronics the overall transformation can be
from an electrical current to electrical current as with a Hall effect based current transformer. That is, to
classify a transducer as converting from one energy form to another is misleading, given the dc current
transformer converts current to flux, then flux back to current. The transducers to be considered have in
common that the output is an electrical quantity, suitable for control and monitoring purposes. While the
energy source type to be transformed include electrical, mechanical, electromagnetic (including light),
chemical, acoustic or thermal energy.
A terminology distinction is to use sensor for the actual sensing element and transducer for the sensing
element plus any associated circuitry. All transducers would thus contain a sensor and most sensors
would also be transducers.
Electromechanical output devices are generically called actuators.
It is important to know what the various applicable transducers do, what kind of signal they provide, and
how to make them compatible.

The terms sensor, detector, and transducer are not differentiated in this chapter.

Broadly the transducers are classified into two main types: active transducers and passive transducers.

An active transducer is a device, which does not require external electrical power (a battery) to
converter and generate an electrical circuit compatible output. It is a self generating transducer.
For example
 Solar cell:- when it is exposed to a light source, it converts light energy into a proportional dc
voltage.
 Piezo-electric crystal:- when it is subjected to changing pressure it produces a proportional ac
voltage.

A passive transducer is a device which requires external electrical power like battery to create an
electrical circuit compatible output. It cannot generate its own voltage or current. It only changes its
resistance or capacitance, etc. during conversion.
For example
 Light Dependent Resistor:- when exposed to light, its resistance decreases many orders (from
less than 10Ω) proportionally and when it is dark its resistance is high (several MΩ ).
 Thermistor:- when thermistor is exposed to heat its resistance decreases and when it is cooled
its resistance increases.

BWW
 Chapter 34 Transducers and Sensors 1548

34.1 General sensor/transducer properties

Transducers include image, current, magnetic, proximity, vibration, temperature, light, accelerometers,
Hall Effect sensors, piezoelectric, level sensors, pressure sensors, barometers, thermistors,
microphones, and speakers.
Physical and chemical transduction
principles
Table 34.1

Transducers can be classified according to the operating principle and stimulus energy form namely
 Mechanical:- for example, strain gauge, LVDT, etc.
 Thermal:- for example, thermistor, thermocouple, etc.
 Magnetic:- for example, search coil, Rogowski Coil, etc.
 Radiation (electromagnetic) transducers:- for example, solar cell, photo diode, Hall effect.
 Photoelectric induction - frequency and pressure measurement,
 Electromechanical:- for example, strain gauge, load cell, etc.
1549 Power Electronics

 Electrochemical:- for example, fuel cell, etc.

 Electroacoustic:- for example, microphone, load speaker, piezoelectric crystal, etc.
 Thermoelectric:- for example, thermocouple, Peltier cooler, etc.
 Ultrasonic:- detector/sensor for vehicle parking

General sensor properties are

 Range:- predictable region of predictable performance and reliability
 Zero drift:- associated with changes in temperature, aging, and conditioning circuitry
 Sensitivity:- change in output for a per unit change in the parameter being measured.
 Resolution:- the smallest detectable change of input parameter
 Response:- time taken by a sensor to approach its true output
 Linearity:- output that is directly proportional to input over its entire range
 Hysteresis:- inability to faithfully reproduce data in both directions of operation
 Calibration:- accurate output measurement based on knowledge of the input
 Span:- dynamic range of measureable input parameter
 Full scale output:- algebraic difference between the electrical output signals measured with
maximum input stimulus and the lowest input stimulus applied.
 Accuracy:- actually inaccuracy - ratio of the highest deviation of a value represented by the
sensor to the ideal value.
 Threshold:- input variation limit that does not cause the sensor output to change
 Cross-Talk:- the gain of the sensor with respect to an unintended stimulus

Table 34.2: Transducers and sensors

Quantity being Input Device Output Device

Measured (Sensor) (Actuator)
Light Dependant Resistor (LDR)
Lights & Lamps
Photodiode
Light Level LED's & Displays
Photo-transistor
Fibre Optics
Solar Cell
Thermocouple
Thermistor Heater
Temperature
Thermostat Fan
Resistive temperature detectors (RTD)
Strain Gauge Lifts & Jacks
Force/Pressure Pressure Switch Electromagnet
Load Cells Vibration
Potentiometer
Motor
Encoders
Position Solenoid
Reflective/Slotted Opto-switch
Panel Meters
LVDT
Tacho-generator AC and DC Motors
Speed Reflective/Slotted Opto-coupler Stepper Motor
Doppler Effect Sensors Brake
Bell
Carbon Microphone
Sound Buzzer
Piezo-electric Crystal
Loudspeaker

The generic stimulus are: acoustic, electrical, magnetic, optical, thermal, chemical, and mechanical.
There are innumerable applications for transducers in power electronics and everyday life, in cars,
machines, aerospace, medicine, manufacturing and robotics.

34.2 Current measurement

34.2.1 Current measurement: closed loop ferrite transformer

Figure 34.1 shows a ferrite current measurement transformer where a compensation winding maintains
the air gap flux at zero, enabling dc (as well as ac) currents to be measured. (Transducers without the
feedback compensation winding are termed open loop.) Measurement bandwidth is typically dc to
200kHz. The current to be measured, primary current Ip, produces an mmf in the ferrite toroidal core. A
Hall effect transducer，with a Hall constant K, detects the flux in the core air gap and an op amp
 Chapter 34 Transducers and Sensors 1550

compensation circuit drives current through the high turns winding in an attempt to zero the core flux.
The current in the compensation winding is therefore proportional to the current being measured,
according to
NpI p  Ns I s (34.1)
A Hall effect generator temperature dependence, offset voltage VOT, can be compensated electronically.
The same transducer can be used to measure voltage by adding an external series resistor in the
primary, which produces a current that is measured, which is proportional to the voltage. The number of
primary turns is usually large so as to minimize the resistor current. The resistance, in conjunction with
the primary self inductance (and leakage), limit the measurement bandwidth, to the time constant L/R.

VH VH  K I B Φ=o
B
d
Hall effect
Is
transducer Ip
Hall
+
d -
1:ηT
i(t) i(t) Is
RB Vo=IsRB
=ηTIpRB

Figure [Link] measurement transducer using a flux compensated toroidal ferrite core.

34.2.2 Current measurement: Rogowski Coil

Rogowski coils, shown in figure 34.2, are used for passive detection and versatile non-invasive
measurement of alternating current (ac) or high speed current pulses (non-dc). It is typically wound on
an air-core so in theory there are no hysteresis, saturation, or non-linearity effects.
The main part is the measuring head which is basically a coil uniformly wound around a flexible cylinder
of insulating material. At one end, the winding is connected to a conductor located in the centre of the
cylinder of insulated material. At the other end, the winding and centre conductor provide the signal
output. This construction makes both electrical connections available at a single point.
The operating principle is that if a closely uniformly wound air-cored toroidal coil of N turns/m is placed
axially around a straight conductor carrying current i in a closed path, the alternating magnetic field
produced by the current in the conductor induces a coil output voltage E that is proportional to the rate of
change of the cross section area A sq m which encircles any flux linked component produced by the
current i, given by the expression:
di
E  M
dt
where M is the mutual inductance (or sensitivity, Vs/A) between the Rogowski coil and the conductor
and di/dt is the rate of change of current in the conductor. If the coil outputs are connected to an
integrator, the output signal reproduces the current waveform.
Rather than measuring the short circuit current through the coil directly, the measurement is the integral
of the open circuit voltage.
There are two advantages to the Rogowski coil.
One advantage of a Rogowski coil over other types of current transformers is that it can be made open-
ended and flexible, allowing it to be wrapped around a live conductor without disturbing the conductor.
A second advantage is that the Rogowski coil does not use a magnetically permeable core like a
standard current transformer, making it of low inductance. The insertion impedance can be less than
1nH. Since it has no permeable core to saturate, it can respond linearly to extremely large currents.
Being of low inductance it can also respond to very fast frequency pulses. A standard current
transformer can have its core saturated at very high currents, and the inductance limits its frequency
response. The closer in form to a perfectly symmetric toroidal uniform coil of wire, with equally spaced
windings, the Rogowski coil is less susceptible to external electromagnetic interference.

Rogowski coil operating principle

A Rogowski coil works by sensing the magnetic field in the space around the conductor that carries the
current. The relationship is given by the Ampere’s Law. Accordingly, the line integral of the magnetic
field around a closed loop is equal to the net current encircled by it, no matter what path the loop takes.
 H cos  d  i t  (34.2)
1551 Power Electronics

current i(t)
er
e
B(t)
Flexible central insulating cylinder of magnetic field produced
area A around which is by conductor current
N turns per metre
±
emf induced by changes in
magnetic field produced by
conductor current

Rogowski currenti(t)
coil

LP filter
i( t )

R C
ucto r
c ond

Vcoil Vout
integrator

Figure 34.2. Rogowski coil and active integrator basic operation.

The mathematical expression that shows this effect where dℓ is a small element of length along the loop,
H is the magnetic field in dℓ and is the angle between the direction of the field and the direction of the
element.

The magnetic field due to a long straight conductor carrying current i, in air, is
o i
B  o H 
2 R
where μo=4π×10-7 and R is the perpendicular radial distance from the conductor to the point at which the
magnetic field is calculated (the major radius of the toroid). The direction of the magnetic field being
tangentially perpendicular to the current and to the radius r, and determined by use of the right hand
rule.

Each turn of the Rogowski coil N turns produces a voltage proportional to the rate of change of the
magnetic flux B through the turn. Assuming a uniform magnetic field density throughout the turn of area
A, by Faraday’s equation, the rate of change magnetic flux is equal to the rate of change of magnetic
field density times the cross-sectional area of the turn π×r2 (toroid cross section radius, r).
d dB
Vturn    A
dt dt
The output voltage from the coil with N turns, effectively series connected, is
dB t 
V coil t   NA
dt
Substitution of B gives
dB  NA di  NA di di
Vcoil  NA  o  o  M (34.3)
dt 2 R dt S dt dt
where S is the mean circumference of the toroid and M is the mutual inductance (sensitivity）between
the coil and the conductor and is independent of the frequency.

The self-inductance L of a coil uniformly wound with a toroidal shape toroid, which affects the output
voltage frequency response, is
 Chapter 34 Transducers and Sensors 1552

L  o N 2 R  R 2  r 2  (34.4)
 
If a rectangular cross section ring is used then the emf produced is given by
o NH c di
Vcoil   n
2 b dt
where H is the rectangular core height and b and c are the inner and outer diameter of the coil.

In order to get a voltage proportional to current an integrator - either active or passive - must be used.
An active integrator, as shown in figure 34.2, using an operational amplifier is a common solution. The
op-amp needs to have sufficient frequency response (both upper and lower cut off half-power points)
and current sourcing and sinking capability to drive the capacitor at the expected frequency.

The integrator needs a resistor placed across the capacitor in order to be made into a leaky integrator.
The resistance placed across the capacitor should be just small enough to leak off the capacitor and
keep it zeroed but not so small that it interferes with the integration performance in the frequencies of
interest.

Ignoring any leaky resistance added to the integrator of figure 34.2 the output is
1
Vout   Vcoil dt
RC
After integrating the signal of equation (34.3), the total output voltage is
1 di M
Vout 
RC  M dt  i
dt 

The transducer/amplifier sensitivity, or transfer function, is:

Vout M

i 
where Vout is the output voltage of the integrator, τ=RC is its time constant, and i is the conductor
current. Changing τ, the operation range can be modified and it is possible to operate from mA to MA.
dB

phase
M/τ
V/A

-3dB
V out/I

20Hz 1MHz
Gain

fL FH
frequency (log scale) f Hz

Figure 34.3. Frequency response of a Rogowski coil and integrator.

It is important to take into account linearity and bandwidth of the integrator, and design it according to
the type of current to be measured. For high frequencies it is appropriate to use a passive integrator
composed only of R and C, such that the mid-band gain is M/CR (V/A)

The relationship Vout proportional to i is constant across the transducer bandwidth. The bandwidth is
defined as the range of frequencies from fL to fH for which sinusoidal currents can be measured to within
3dB of the specified sensitivity M/CR, as shown in figure 34.3.

At low frequencies, the integrator gain increases and theoretically becomes infinite as the frequency
approaches zero. This would result in unacceptable dc drift and low frequency noise; hence the
integrator gain is limited at low frequencies. This limitation is controlled by a low pass filter in parallel
with the integrating capacitor C. The low pass filter sets the low frequency bandwidth fL, typically less
than 1Hz.
1553 Power Electronics

Furthermore, due to the distributed inductance, equation (34.4), and inter-turn capacitance of the
Rogowski coil, there is a high frequency bandwidth fH, (generally >1MHz) above which the measurement
is attenuated and significant phase delay occurs.

Construction
There are a number of Rogowski coil current transformer types.
1 - Flexible Rogowski coil
The insulated winding is placed over a long and flexible plastic former typically between 3.5 and 15mm
in diameter. The coil is fitted by wrapping it round the conductor to be measured and bringing the ends
together. External insulation can be composed of one or several insulation layers (to increase the
sensitivity), thermal shrinkable protection, electrostatic screen, which affects the flexibility of the coil.
Electrostatic screen can be added to improve insulation of external influences.
Although less sensitive and less accurate than the rigid form, a flexible coil is better for high frequency
measurements.
It is useful with large size or awkward shaped conductors or in places with limited access or where a
lightweight transducer is needed which can be suspended on the conductor. As an open coil, it is not
necessary to disconnect the conductor that carries the current to be measured and the user has only to
connect the ends after the coil is placed around the conductor. Its form is compact and versatile.

Typical electrical features are:

Mutual inductance M: between 30 and 300nH.
Maximum frequency: between 100kHz and 1MHz depending on M.
Minimum frequency: between 1 and 10Hz, depending on the integrator.
Current range: from 1A to >1MA.
Accuracy: 1%.

2 - Rigid Rogowski coil

The rigid coils are wound on a solid plastic former, normally in a toroidal shape, and tend to be bulkier
than flexible coils but have better stability. External insulation can be composed of one or several
insulation layers, or varnished or encapsulated and potted. An electrostatic screen can be added to
improve insulation to external influences.
The output voltage is stable and the accuracy is good. A rigid coil lower measurement frequency range
is lower than with a flexible coil, hence is more applicable for low current and low frequency
measurements. Disadvantageously, being a continuous ring, the current-carrying conductor must be
disconnected and placed through the core centre hole before the measurement. It can be used for high
precision measurements or for permanent installation.

Typical electrical features are:

Mutual inductance M: between 3 and 5μH.
Maximum frequency: between 10kHz and 30kHz depending on M.
Minimum frequency: down to 0.1Hz, depending on the integrator
Current range: from 100mA to >100A.
Accuracy: 0.1%

3 - Planar Rogowski coil

The sensor can be manufactured using a planar coil rather than a toroidal coil. In order to reject the
influence of conductors outside the sensors measurement region, planar Rogowski current sensors use
a concentric coil geometry instead of a toroidal geometry to limit the response to external fields. The
main advantage of the planar Rogowski current sensor is that the coil winding precision that is a
requirement for accuracy can be achieved using low cost printed circuit board manufacturing.

Features and applications

A Rogowski coil used as current sensor has numerous advantages:

 The air coil has no hysteresis or remanence, it does not saturate, and is linear. The mutual
inductance is independent of the current, while its inductance is low, a few nH.
 Non-intrusive and flexible. Not closed loop during attachment. Can be retro-fitted.
 Good response to current transients, so they are appropriate for current pulse measurements
or for protection systems. A simple RC output filter suffices for a basic protection scheme.
 High bandwidth. The high-frequency limit is determined by the self-resonance of the coil and
depends on the coil design. Although not applicable to DC measurement, with an accurate
integrator design, it is possible to measure frequencies lower than 1Hz.
 Can measure ac signals superimposed on large dc currents. Huge over current capability
 Chapter 34 Transducers and Sensors 1554

 The same coil can measure a wide range of currents, from mA to MA, with a typical sensitivity
of 1.0mV/A for ± 6000A peak, and a di/dt typically from 1 to 25kA/μs, over a coil temperature
range from -20Cº to 100ºC.
 Easy calibration. Because of its linearity, coils may be calibrated at any current level.
 It is lightweight, compact and easy to install and to transport. Importantly, it is easy to use.
 Output variation with temperature is low.
 Low power consumption, and can be totally passive. Low cost.
 High frequency bandwidth (3dB) decreases with coil length, for example, 100mm 12MHz,
200mm 8MHz.
 Rogowski Coil cross sectional diameter specifies electrical isolation, for example, 3.5mm for
2kV isolation and 4.5mm for 5kV isolation. The open circuit output voltage is low, viz., safe.

Another useful feature of a Rogowski sensor is immunity to far-field interference. EMF components
induced by the same far field source will cancel each other. Two components of EMF induced by the
current passing through the wire inside the coil will be added to each other.

One disadvantage is that the Rogowski coil produces an EMF proportional to di/dt. Therefore, at
connecting or disconnecting instants, the EMF goes ‘infinite’. Transient voltage suppressors or other
voltage protection is needed to prevent overloading the interfacing electronics. Also, accuracy is slightly
dependant on the position of the current carrying conductor in loop.

34.2.3 Flux-gate Transformer

The term fluxgate refers to the operational principle of specific isolated current measurement
transducers. The magnetic field generated by the current to be measured is detected by sensor. Current
transducers based on the Hall-effect, usually called standard fluxgate transducers use a gapped toroidal
magnetic circuit with the field measuring element in the gap, plus the toroid has a secondary winding as
considered in section 34.2.1. Other fluxgate transducers avoid gapping the core and use magnetic
saturation and magnetic balance as part of the measurement mechanism.

The flux-gate current measurement transducer detects the saturation state of a magnetic circuit. The
magnetic core is a high permeability material, which surrounds the magnetic field source to be
measured. Usually, magnetic core fluxgate sensors have a high magnetic permeability, very low
coercitivity and many turns in a coil on the core to increase sensitivity. Ferrite is the standard core
material, because of its good frequency response.

The magnetic material is excited by a square wave signal (typically 200Hz to 600Hz) that bidirectionally
and symmetrically saturates the core, in the absence of an external magnetic field. This saturation
symmetry is lost in the presence of an external magnetic field produced by the current to be measured.
Current injected in an auxiliary winding creates a compensating magnetic field that restores the
symmetry of the hysteresis cycle. The injected compensating current is proportional to the current being
measured. Thus inductance changes are use to detect the external field generated by primary current,
and therefore the current flowing through the enclosed conductor.
This system is suitable for the measurement of dc and ac currents, with high accuracy, high frequency
(in excess of 100kHz), and a high current range.

Figure 34.4. Main types of fluxgate transducers.

Main types of fluxgate transducers

The main types of fluxgate transducers are shown in Figure 34.4, namely:
i. Standard fluxgate, as with the Hall effect current measurement transducer in section 34.2.1.
ii. Two magnetic cores fluxgate, where performance is improved by: (a) making the field sensing head
one (shaded) of the two magnetic cores, and (b) ensuring high frequency performance by using a
separate core for the transformer effect, both without an airgap.
iii. Three magnetic cores fluxgate, where the performance is further improved by (a) duplicating the
field sensing head, using two magnetic cores (dark color) with an excitation coil wound around each of
1555 Power Electronics

them; and (b) improving the way the high frequency current transformer is made and processed
electronically.
iv. Low frequency fluxgate, using only the low frequency part of the two magnetic cores fluxgate
transducer, not including the current transformer.

ΦP+Φs=0

Figure 34.5. Structure and principal of a Flux-gate transducer without an air gap in the magnetic
path.

Operating Principle
The gapless transducer uses its toroidal core as field detector, without an gap in the magnetic path. An
auxiliary winding is added to the core forming a core-wounded set, which is used as a saturable inductor
detector.

u(t Equal areas

) Average equals zero

i(t
)

Figure 34.6. Voltage waveforms of excitation and current by the auxiliary winding under zero flux
(current) conditions.

In order to detect a null field in the magnetic circuit, the secondary winding is excited with a
compensating current. Hence the transducer operates with a zero field condition, as shown in Fig. 34.5.
This null condition means the current IS in the secondary winding is directly proportional to the primary
current to be measured, IP. The relationship between the primary and secondary current is equation
(34.1), where NS the number of the secondary winding turns.
I p  Ns I s (34.1)
The detection of the zero flux condition in the magnetic path of the transducer is based on the change of
the inductance of the saturable inductor formed by the ferromagnetic core and the auxiliary winding. In
absence of current to be measured IP, the net flux through the saturable inductor core is zero. Under
these conditions, with a squared voltage waveform u(t) applied to the auxiliary winding NA, the
magnetising current i(t) waveform in auxiliary winding, bidirectionally saturating the core in a zero
average flux condition, is as shown Fig. 34.6.
The effect of the current to be measured IP on the auxiliary winding current IA, when a low frequency
square voltage waveform is applied, leads to a non-zero average current, as shows Fig. 34.7. The
average value and its sign depend on the value and direction of the current IP.
 Chapter 34 Transducers and Sensors 1556

Unequal areas

Ip<0

Figure 34.7. Effect of square wave voltage source excitation current with external field produced by a
primary current.

The difference between the waveforms shown corresponds to the current direction applied. Then a
magnetic field is generated to cancel this effect, to produce a zero flux condition. Knowing the current
applied to cancel this effect by returning the system to the zero flux condition, allows accurate
calculation of the current passing through the primary conductor.

A disadvantage of this structure is noise injection on the primary current IP measure. This noise is
induced from the auxiliary current IA, because of coupling through the magnetic core of the transducer.
The solution is to use of second core with its own auxiliary winding, for compensation. These two cores
with their auxiliary windings must be identical. Now the secondary winding NS in which the compensation
current of flux in the transducer is applied is common to both auxiliary cores. If the second auxiliary core
NA2 is excited with an equal current but in reverse direction to the current used for exciting the first
auxiliary core NA1, the currents induced on the primary current conductor (IP) will be equal and opposite,
cancelling the coupled effects.

Better performance, at increased cost, is obtained with a common compensation winding placed around
both cores. Identical ampere-turns cancellation is ensured to cancel the external field.

NS

NA1 NA2

(a) (b)

Figure 34.8. Two saturable inductors and compensating windings:

(a) two separate windings and (b) a common winding.
1557 Power Electronics

As the winding NA1 of the flux detector is coupled with the compensation winding NS, the applied square
wave voltage is re-injected into the compensation winding and creates a parasitic current in the
measurement resistor. However, the square wave voltage induced in the NS winding by this flux is
practically cancelled when a second NA2 winding is mounted on a second detector core (identical to NA1)
within the compensation winding NS. The residual flux (the sum of the opposed fluxes in NA1 and NA2)
create small voltage peaks that cause the remaining signal to correlate with the fluxgate excitation.
In each NS winding arrangement case, the system is completed with an electronic circuit to impose the
zero flux condition monitoring the compensation current.

A fourth winding parallel to NS is wound under the compensation winding NS on the main core to extend
the frequency range of the transformer effect to lower frequencies. It is connected to a circuit that adds
voltage via the power amplifier to compensate the small induced voltage in the frequency range too high
for the fluxgate detector.

Figure 34.9. Block diagram of the flux-gate transducer circuit.

The block diagram of the flux-gate transducer measurement system is shown in Fig. 34.9.

PWM is used for zero flux condition control, where feedback forces the currents associated with the
saturable cores to a net average value of zero.

i. Signal generator for excitation of the auxiliary windings, NA

The generator is based on a comparator circuit with hysteresis (or Schmitt trigger). The circuit changes
the output voltage when the circulating excitation current IA in the main winding exceeds a threshold.
The magnetic components form part of the oscillator circuit, and their electrical characteristics influence
the frequency of the squared signal generator circuit, which is around 300Hz.

ii. Symmetry detector of the auxiliary current IA

In the absence of primary current IP the average value of the current of excitation IA is zero. The effect
produced by primary current is the appearance of an average value different from zero and sign
dependent of the sense of this current. A PI controller is to automatic adjustment the value of the
secondary current winding IS, to ensure that the primary current excitation winding has zero mean value.
This controller cannot guarantee the proper measurement system functionality at start-up if primary
current IP pre-exists, because under these conditions the two inductors will be saturated. A primary
current through the measurement system in non zero flux conditions, produce a high frequency current
(tens of kHz) in the main excitation winding IA and a non zero average value, with independent sign of
the primary current circulating sense IP. To alleviate this drawback, an additional controller ensures that
the zero flux condition is reached regardless the initial pre-existing primary current IP. The operation of
this controller is based on the property where the frequency of the current in the main excitation winding
IA is high frequency when the system is not balanced and low frequency when the system is operating in
the vicinity of the point of zero flux. As shown in figure 34.9, this controller includes a triangular low-
frequency oscillator, a frequency detector for the excitation current IA and an analogue switch controlled
by the frequency detector.
 Chapter 34 Transducers and Sensors 1558

In a non-zero flux initial condition, the input of the current compensation driver IS is connected to the
triangular low-frequency signal generator. This winding compensation current IS ensures the necessary
zero flux condition occurs. When detected, the frequency of the current of the main excitation winding IA
decreases; then the PI controller is connected to the input of the current compensation driver.

iii. Valid measure indicator

The output of the low-frequency detector circuit is connected to a valid measurement indicator. This
indicator is activated having detected a low frequency primary winding excitation current, that is when
the system operates in zero flux conditions. A LED and relay are the output elements to indicate the
zero flux valid measurement condition.

iv. Driver to generate the compensation current

This D class amplifier circuit is used to generate the current that flows through the secondary winding
compensation NS. Although the amplifier is highly efficiency compared to a linear amplifier, it adds
harmonics at the same switching frequency and higher. It is based on a pulse width modulator (PWM),
which generates a squared voltage waveform with a duty cycle proportional to the PI controller output
signal.
The modulator output PWM square waveform is applied to the compensation winding NS through a
current driver implemented with a half or full bridge inverter.
The self inductance of NS filters the current. The system output voltage is measured across a shunt
resistance connected in series with this winding, and is proportional to the primary current being
measured.

v. Measure of high frequency currents

The basic flux-gate principle is only suitable for the measurement of dc or low frequencies ac currents.
The maximum frequency for the ac measurement is fixed by the operating frequency of the zero flux
detection system. For the measure of high frequency ac currents and to obtain a suitable dynamic
behaviour in case of fast transient currents, a third core is included. This third core is coupled only by the
compensation winding NS and it operates as a conventional current transformer, as shown in figure
34.8(b).

vi. Power supply

Two stable output voltages, ±12V, are derived from a flyback converter with a dc input voltage between
10V and 30V.
Typical main characteristics of the system are:
 maximum peak current: 1000A
 primary rated current, IN: 700A
 small signal bandwidth, 5% of IN: dc to 100kHz
 conversion ratio: 1:1000
 supply voltage: 10 to 30Vdc
 di/dt: > 100A/μs
 offset current: < 5μA
 linear error: < 5 ppm
 burden resistor > 20Ω, ½W

Main advantages are:

 High Accuracy over high bandwidth
 Very low output noise and offset drift (high frequency PWM ripple in output and references)
 Negligible insertion losses
 High immunity to interference
 Overload capability
 Excellent linearity

Applications include:
 Precise and high stability inverters
 Energy measurements
 High precision power supplies
 Feed back element in high performance gradient amplifiers for MRI
 Medical equipment

[Link]
1559 Power Electronics

34.2.4 Resistive Sensor

The resistive shunt, although a simple concept, v=iR, can be a complicated device due to three features,
at any current, viz.,
i. shunt power losses, necessitating possible transducer cooling,
ii. internal shunt inductance which distorts voltage with increasing frequency transients, and
iii. galvanic isolation between the high voltage circuit and the control circuit, especially at high nominal
and pulse voltages.

In the low current range film technology SMD-resistor shunts (2 and 4 terminal - Kelvin) are used, which
provide good heat sinking and low inductivity. Typical characteristics are 0.3mΩ to 1Ω, ±1% tolerance,
0.1to 5W dissipation. The next higher power class uses thick film technology in a TO-220 housing for
through-hole mounting; these resistors have good thermal characteristics because of the external
heatsink and offer low inductivity.

For high accuracy measurement of high frequencies coaxial shunts are used, where influence of the
internal inductivity is compensated what allows to measure current in the range up to several MHz.
Coaxial shunts can register peak current in excess of 100kA in the 200MHz frequency range.
Resistance ranges from 1mΩ to 0.1Ω, typically, and has a positive temperature coefficient, typically 10-
12ppm/°C. The concentric manganin and copper cylinder structure, is ideal for forced air cooling and
water cooling. As resistance decreases, bandwidth decreases and rise time increases, viz., 25μΩ
[250kHz, 2μs, 840kJ, 50W] to 50mΩ, [1.2GHz, 0.3ns 75J, 20W].

Shunt manganin is a copper (87%) – manganese (9.5%) plus nickel, resistance alloy used for precision
shunts in various high current applications. The alloy is long term stable (provided used below about
60°C) and has a low temperature coefficient of resistivity with peak (parabolic shaped) resistance at
about 50°C, ±15 ppm/°C. The thermal emf against copper is low, less than -1μV/°C, down to -0.2μV/°C
with a similar Cu/Mn based shunt alloy, zeranin.

It is possible to include a compensating element with the transfer function for approximate compensation
of reactance into the measurement scheme:
R1
H s 
L
s 1
R
Lω and Rω - shunt characteristics

However, frequency bandwidth of the overall device is limited by the bandwidth of the first order filter,
and accuracy of measured values.

(a)

(c) (d)
(b)

Figure 34.10. Current measurement resistors:

(a) thin film, four- terminal, (b) thick film, two terminal, (c) coaxial shunt and (d) its cross section.
 Chapter 34 Transducers and Sensors 1560

There are the following major features of the resistive shunts.

Advantages:
 Simple design and mechanically robust;
 Wide frequency range, from dc to over 2GHz, with coaxial shunts;
 High overload capability and possibility to measure current peak values;
 Need no auxiliary power supply for operation.
 High EMC rejection and not susceptible to noise with double coaxial type designs
 Circular symmetry minimizes skin effects that limit bandwidth
Disadvantages:
 Need to be directly connected to power bus, power losses during heating-up of the shunt;
 External galvanic isolation is required.

34.2.5 Magneto-optic Sensor

Current sensors based on magneto-optical (Faraday) effect are used for galvanically isolated
measurement of uni- or bidirectional DC currents high currents (up to ±500kA, ±600kA peak), with
±0.1% accuracy and linearity, in high voltage buses. Current sampling rate is 4kHz.
Because it does not need a magnetic yoke, an FOCS is smaller and lighter than Hall effect current
sensors, and suffers no reduction in accuracy due to saturation effects. Because magnetic field sensing
is distributed around circumference, it is unaffected by stray magnetic fields, and there is no need for
magnetic centring. It also does not need recalibration after installation or during its service life. Because
the optical fibre is inherently insulating, electrical isolation is easier to maintain.
A fibre loop is wound around a busbar, where the Faraday effect (a phase shift) occurs when polarized
light is experiences the magnetic field produced by the busbar current.
Right and left circularly polarized light waves through the coiled sensing fibre are reflected back at the
end of the fibre, (and their polarization direction is swapped) to the sensor electronics. A dc current,
accumulatively creates a phase difference which is proportional to the line integral of the magnetic field
along the sensing fibre. This difference is thus a direct and highly precise measure of the current. The
phase shift is
F  4VI
I   Hd s
where V is material strength constant, Verdot constant, wavelength and temperature dependent
H is magnetic field
s is length of sensing fibre

Figure 34.11. Magneto-optic current transducer.

Advantages:
 Possibility to measure high current values;
 Stability against the influence of the cross-magnetic field;
 Complete galvanic isolation;
 No power losses in the sensor.
 1561 Power Electronics

Disadvantages:
 High cost and large dimensions.
 Requires light source, etc.

34.2.6 Integrated ac/dc current sensors

Integrated Hall-effect current sensor and Hall-effect linear integrated circuits can provide highly accurate
(<1%, dc to 120kHz up to 1MHz, at up to 150°C), low noise output voltage signals that are proportional
to a conducted AC or DC current. Current sensing from 0 A to ±2.5A to ±50 A to ±200 A (uni and
bidirectional), with galvanic isolation up to 4.8 kV (for line-side or high-voltage applications) and low
resistance primary conductor (≤200µΩ down to 0.8 µΩ for low Joule losses), is in surface mount and
through-hole packages.

The block diagram in Figure 34.12 shows the internal block circuit diagram and construction of the
device.

The IC consists of a low-offset linear Hall-effect analogue output circuit with a copper conduction primary
path located near the Hall die. Applied current flowing through the copper conduction path generates a
magnetic field that is concentrated by a low magnetic hysteresis ferromagnetic core, which is converted
by the Hall IC into a proportional voltage. Device accuracy is optimized due the close proximity of the
magnetic signal to the Hall transducer. A precise, proportional output voltage is provided by the low-
offset, chopper-stabilized BiCMOS Hall IC, which is factory programmed (and EEPROM stored) for zero
amperage output voltage and sensitivity accuracy. Digital temperature compensation improves the IC
accuracy and temperature stability without influencing the high bandwidth operation of the analogue
output.

A thick isolated copper conductor allows high overcurrent conditions. The conductor electrical isolation
avoids the use of opto-isolators or other isolation techniques and along with an electric field integrated
shield, provides dV/dt and stray electric immunity for low output voltage ripple and low offset drift, in
high-side, high voltage applications.

Fig 34.12 Functional block diagram and IC construction.

A high bandwidth (up to 1MHz) is accomplished by a high-speed analogue signal path that employs
differential sensing to provide immunity to interfering common mode magnetic fields. By using two Hall
plates measuring opposite polarities of the field created by the input current, the IC with differential
sensing is able to reject common-mode (and external) fields. Thus with 40 dB of common-mode field
rejection for a uniform and constant field, the IC can be used near other current paths.

[Link]
 Chapter 34 Transducers and Sensors 1562

Figure 34.13. Current transducer ratings.

Table 34.3 highlights the parameter differences between and main characteristics of five current
measurement techniques, namely the conventional current transformer, Hall effect sensor, current
transformer based on a Rogowski coil, flux gate (which is a Rogowski coil with a magnetic core), and a
co-axial shunt resistor.

Table 34.3: Features of six current measurement techniques

Current Hall Rogowski Flux Co axial IC

Feature
transformer effect coil gate shunt resistor Hall effect
31.6.2 34.2.1 34.2.2 34.2.3 34.2.4 34.2.6
operating principle N1 I1 =N2 I2 VH = IxB VαμodI/dt VαμoμrdI/dt VS = IxRS VH = IxB
output current voltage voltage voltage voltage voltage
bandwidth low medium high medium low very high
AC/DC measurement AC AC&DC AC AC&DC AC&DC AC&DC
isolation high high high high low high
linearity good medium excellent good low medium
precision medium medium medium very high medium medium
hi current measurement good good very good very good low good
saturation & hysteresis problems yes yes no no no yes
power dissipation low medium low medium low medium
temperature effects on output low medium very low low medium medium
transient response medium medium very good medium medium medium
low frequency response medium very good good medium medium very good
dc offset no yes no yes no yes
easy of installation medium medium simple medium low low
weight medium medium low medium medium low
dimensions/size medium medium low medium low low
cost medium high low high medium low
1563 Power Electronics

34.3 Voltage measurement

Hall effect closed loop voltage transducers

Construction and principle of operation

Hall effect voltage transducers are based on the same principle as their current transducer counterpart.
They are a current transducer assembly where the primary circuit is a winding of a high number of turns.
This realisation the ampere-turns for primary induction, with a low primary current, thus minimising
power consumption from the circuit to be measured.
To measure a voltage it is therefore sufficient to shunt from the voltage the equivalent primary current
which will supply the transducer. This is affected with a resistor connected in series with the primary
winding (and its resistance).
Because of the multiple primary turns, high voltage isolation is challenging. At higher voltages, a few kV,
a measurement bandwidth from dc to 500kHz, with a sub 1μs response time, is possible.

Resistor dividers
Up to 25kV, thick film on aluminium, 1M to 50GM at 0.5%, 1000:1, 9W. At 25kV, 0.5pF.

34.3.1 Differential Isolation (galvanic) Amplifier

A voltage sensing optical differential isolation amplifier, shown in figure 34.15, is designed for isolated
single-ended low-voltage sensing. With a 2V input range and high 1GΩ input impedance, a shunt
resistor divider makes it suitable for isolated voltage sensing in power electronic applications including
motor drives and renewable energy systems. A resistive voltage divider is used to scale the DC-link
voltage to suit the 2V input range of the voltage sensor. A differential output voltage (single supply) that
is proportional to the input voltage is created on the other side of the optical isolation barrier.

With a gain 1V/V, operation on a single low-side 3.3V to 5V supply yields 0.1% linearity and a bandwidth
from dc to 100kHz. The active-high shutdown reduces the high-side 15mA operating current to 15uA,
thus is suitable for battery-powered and other power-sensitive applications.

The high common-mode transient immunity of 15kV/μs provides the precision and stability needed to
accurately monitor DC-link voltage in high noise environments. Optical barrier coupling (giving 5kV peak
galvanic isolation), sigma-delta (-) modulation, chopper stabilized amplifiers, and differential outputs
provide (in conjunction with a 3-5.5V isolated power supply) high magnetic interference resistance,
isolation-mode (common mode >108dB) noise rejection, low offset, high gain accuracy (±0.5% to ±3%),
and stability (–40°C to +105°C).

Fig 34.14 Functional block diagram (optical isolation version)

Figure 34.14 show the circuit for isolated measurement of a high dc voltage, where the resistor divider
R1 and R2 ensures the isolating IC input voltage does not exceed 2V. The op amp U2 should have a
bipolar input stage.

[Link]
modulators/isolation-amplifiers/

The differential isolator block diagram in figure 34.35 has capacitive (SiO2) ceramic isolation barriers
(2pF) has dual (split) rail on the hv side, so is suitable for hv dc and ac measurement (although the
 Chapter 34 Transducers and Sensors 1564

manufacturer has not exploited zero dc offset ac input possibilities). Features (depending on the specific
part) are 3500Vrms isolation (1500Vrms continuous), 0.01% linearity, 200k input resistance, -55°C to
125°C, 60kHz bandwidth, power supply range of ±4.5V to ±18V.

Fig 34.15. Circuit diagram of capacitive isolator for measuring hv dc and ac voltages.

The isolator is designed to have a 50kHz single-pole (Butterworth) signal output response. By cascading
the isolator with a 50kHz, Q = 1, two-pole, low-pass filter, as in figure 34.16, the overall signal response
becomes three-pole Butterworth. The result is a maximally flat 50kHz magnitude response with the
output ripple reduced below the noise level.

[Link]
[Link]

Fig 34.16. Second order Butterworth output stage for the capacitive isolator in figure 34.15.

Fig 34.17. AC voltage measurement suitable for the ac (single and three phase) mains.
 1565 Power Electronics

The split rail hv side supply means ac grid voltages can be measured, as shown in figure 34.17, where
the resistor divider is reference to the high side Gnd1 (0V). Series connected resistors ensure the
divider can withstand the main voltage level. The dc to dc converter to supply the HV side needs the
appropriate hv voltage isolation rating. In a three phase system, lines u, v, w, without neutral, only one
isolated dc to dc converter may be used (with two voltage isolator amplifiers), since two line to line
voltages (for example Vvw and Vuw) can be measured in reference to a common line w, where either is
inverted to recover the phase sequence, Vvw and Vwu=-Vuw. The third phase can be processed from the
two known phases, Vuv=-Vvw-Vuw or Vuv=-Vvw+Vwu.

An isolated dc to dc converter can be avoided by using the capacitive half-wave charging technique
illustrated in figure 34.18a. Effectively the capacitor C1 ac charges with the positive half cycle current
diverted to charge the dc capacitor C2, with a voltage clamped by Zener diode D2. The half-wave near
sinusoidal charging current to C2 and the parallel load Rload is
i wt   V1 wt  / X  V1 wt   w C 1
where V1peak<< Vzener and the average of i(wt) needs to larger than Vzener/Rload.

470n 220 D1
u
+Vdd1
470n 220 D1
u V1 1N4148 +
+Vdd1 C1 R1
D2 C2 Rload
V1 1N4148 + 470μ
C1 R1
Rload 0V
D2 C2 w GND
470μ
0V 470μ +
GND 415Vrms
Rload
w w 50/60Hz D3 C3
1N4148
-Vdd1
(a) (b) D4

Fig 34.18. Non isolated mains line voltage lv supply (a) single rail Vdd1 and (b) split rail ±Vdd1.

Partial duplication of the circuit in figure 34.18a produces split rail dc supplies, as in the alternating half-
wave rectifier in figure 34.18b.

34.4 Acceleration measurement

An accelerometer is an electromechanical device that measures acceleration forces and vibration of

structures. It complies with:

v  2 x
a  m/s 2
t t 2

In a piezoelectric based accelerometer, the force caused by vibration or a change in motion

(acceleration) causes the mass to ‘squeeze’ the piezoelectric material which produces an electrical
charge that is proportional to the force exerted upon it. Since the charge is proportional to the force, and
the mass is constant, then the charge is also proportional to the acceleration.
The forces may be static, like the constant force of gravity, or dynamic - caused by moving or vibrating
the accelerometer.

Table 34.4: Types of accelerometers

Sensor category Key technologies

metal beam or micromachined feature produces capacitance; change in
capacitive
capacitance related to acceleration
piezoelectric piezoelectric crystal mounted to mass – voltage output converted to acceleration
piezoresistive beam or micromachined feature whose resistance changes with acceleration
Hall effect motion converted to electrical signal by sensing changing magnetic fields
magnetoresistive material resistivity changes in the presence of a magnetic field
heat transfer location of heated mass tracked during acceleration by sensing temperature
 Chapter 34 Transducers and Sensors 1566

Table 34.4 show a range of accelerometer technologies, where three technologies widely used for
acceleration measurements.

 Piezoelectric accelerometers are the most widely used accelerometers for test and
measurement applications. They contain microscopic crystal structures that get stressed by
accelerative forces, which causes a voltage to be generated. These devices offer a very wide
measurement frequency range (a few Hz to 30 kHz) and are available in a wide range of
sensitivities, weights, sizes, and shapes. These accelerometers are applicable to both shock
and vibration measurements. Piezoelectric accelerometers can have either a charge or voltage
output.
 Piezoresistive accelerometers generally have low sensitivity making them desirable for shock
measurements. They are also used extensively in transportation crash tests. Because of the low
sensitivity, they are not generally used for vibration measurements. Piezoresistive
accelerometers generally have a wide bandwidth and the frequency response goes down to
zero Hz (often called ‘dc responding’) or steady state, so they can measure long duration
transients.
 Variable capacitance is among the newer accelerometer technologies. Two microstructures next
to each other have a certain capacitance between them. If an accelerative force moves one of
the structures, then the capacitance changes. By adding circuitry to convert from capacitance to
voltage, gives an accelerometer. Like piezoresistive accelerometers, they are dc responding.
Variable capacitance accelerometers have high sensitivities, a narrow bandwidth, and
outstanding temperature stability. These devices are highly desirable for measuring low
frequency vibration, motion and steady state acceleration.

In addition to the methods shown in Table 34.4, other methods include the use of hot air bubbles and
light.

There are two main types of piezoelectric accelerometers (vibration sensors).

The first type is a ‘high impedance’ charge output accelerometer. In this type of accelerometer the
piezoelectric crystal produces an electrical charge which can be connected directly to the measurement
instruments. The charge output requires special accommodations and instrumentation. This type of
accelerometer is used in high temperature applications (>120°C) where low impedance models can not
be used.

The second type of accelerometer is a low impedance output accelerometer. It has a charge
accelerometer as its front end but has a tiny built-in micro-circuit and FET transistor that converts that
charge into a low impedance voltage that can readily interface with standard instrumentation. This type
of accelerometer is commonly used in industry. An accelerometer power supply provides the proper
power to the microcircuit 18V to 24V @ 2mA constant current and removes the dc bias level, they
typically produce a zero based output signal up to +/- 5V depending upon the mV/g rating of the
accelerometer.
1567 Power Electronics

Figure 34.19. Piezoelectric accelerometer:

(a) cross sectional view and (b) upper and lower cut-off frequency characteristics.

High frequency limit is the frequency where the output exceeds the stated output deviation. It is typically
governed by the mechanical resonance of the accelerometer. From the equivalent circuit in figure
34.19b, the -3db low-pass corner frequency fLP is
1
f LP 
2 RLC L
j T o
 
H  
1  j T 
1
o

where Ec is the crystal voltage

RL is the series resistance
CL is the shunt capacitance
ω=2πf

Low frequency cut-off is the frequency where the output starts to fall off below the stated accuracy. The
output does not ‘cut-off’ but the sensitivity decreases rapidly with lower frequencies. From the equivalent
circuit in figure 34.19b, the -3db high-pass corner frequency fHP is
1
f HP 
2 RmC a
j T o
 
H  
1  j T o
where Ec is the crystal voltage
Rm is the input resistance
Ca is the crystal capacitance

Linearity is the maximum deviation of the calibration curve from a straight line and is given by
linearity  Vout
0g
 ½ Vout
+1g
Vout
-1g

Sensitivity is a measure of how much the output of a sensor changes as the input acceleration changes
and is given by
Vout V +1g Vout-1g
sensitivity   ½ out
g g

Example 34.1: accelerometer sensitivity and linearity

A capacitive accelerometer has an acceleration versus voltage output characteristic as shown in figure
34.20. Calculate transducer linearity and sensitivity.

Solution
linearity  Vout
0g
 ½ Vout
+1g
Vout
-1g

=2.1V-½ 1.1V+2.5V  =0.3V
 Chapter 34 Transducers and Sensors 1568

Vout Vout+1g
Vout
-1g

sensitivity  
g 2g
1.1V-2.5V
= =-0.7V/g
2g
♣

-1g,
2.5V

0.3V

+1g,
1.1V

Figure 34.20. Capacitive accelerometer characteristics.

34.5 Resolver and synchro

Synchro and Resolver position sensors are similar to electric machines, being electromechanical rotary
transformer transducers that convert shaft angle to an absolute analogue signal. They share the same
rotor, stator, and shaft components as electrical rotating machines, as shown in fig 34.21.

Figure 34.21. Resolver stator and rotor.

34.5.1 Resolver Design

The resolver is a rotary transformer that consists of a cylindrical rotor and stator, both having multi-slot
laminations and two sets of windings. The windings are normally designed and distributed in the slotted
lamination with either a constant pitch-variable turn or variable pitch-variable turn pattern. In either case,
the winding distribution is in a sinusoidal pattern.
1569 Power Electronics

The windings create one complete sine curve and cosine curve in one mechanical revolution. The two
sets of windings are positioned in the laminations at 90 degrees to each other. These are called the sine
and cosine windings. One rotor winding may be shorted internally to improve the accuracy.
The primary difference between a synchro and a resolver is a synchro has three stator (secondary side)
windings at 120⁰ offsets, while the resolver has two stator (secondary side) windings at 90⁰ angles. The
number and angular mechanical orientation of the rotor windings varies. The resolver is considered
hereafter.
A resolver is an angle sensor that outputs rotational angles as two-phase AC voltages (analogue
signals). An AC output voltage is induced (like a voltage transformer) in the winding on the output side
through excitation of the excitation (input) winding using an AC voltage. As this output voltage will vary
depending on the angle of rotation, the angle of rotation can be calculated using the voltage reading,
and transformed to a digital form with an R to D converter IC.

34.5.2 Resolver output functions

Three and four winding resolver electrical arrangements are shown in figure 34.20. The input winding(s)
on the stator or rotor is sinusoidally excited, and because of orthogonal winding alignments and rotor
rotation, various output functions are attained on the externally unexcited winding(s).
Resolvers used to describe angular measurement and for solving geometric functions are known as
Data Transmission Resolvers, while those used for control data transmission are termed synchro
resolvers.

Rotor-Excited Resolver. With the windings as defined in Fig. 34 22, the basic equations for voltage
vectors in a sinusoidally rotor-excited resolver, Er1-3, Er2-4, are
Es13  k Er13 cos – Er2  4 sin 
(34.2)
Es2  4  k Er2  4 cos  Er13 sin 
where
k = transformation ratio (hereafter assumed unity)
θ = angle of counter-clockwise displacement from resolver zero.

Electrical zero EZ position for the resolver is so established that, when θ is zero, with rotor winding R1-3
excited (at rated voltage) and rotor winding R2-4 open, Es1-3 (the cosine winding) will be at a maximum
and Es2-4 (the sine winding) will be zero. If winding R2-4 is excited and R1-3 is open, the outputs will be
reversed.

Stator-Excited Resolvers. The basic equations for voltage vector θ in a stator-excited resolver are:

Er13  k E s13 cos  Es2  4 sin 

(34.3)
Er2  4  k E s2  4 cos  E s13 sin 

When θ is zero, with stator winding S1-3 excited and stator winding S2-4 open, Er1-3 will be at a maximum
and Er2-4 will be zero. If winding S2-4 is excited and S1-3 is open the outputs will be reversed.

(a) (b)

(c)

Figure 34.22. Electrical representation of resolver:

(a) single rotor winding, (b) two rotor windings, and (c) its associated waveforms.
 Chapter 34 Transducers and Sensors 1570

A. Three winding resolver

Rotor excited, three winding

The simplest resolver would have a rotor with a single winding and a stator with two windings at 90⁰ to
each other, as shown in Fig. 34-21a.
If the rotor is excited by an AC reference voltage, across R2 to R4:
Er2  4 sin t
Then the resolver format voltages appearing on the stator terminals, shown in Fig. 34.22b, are:
Es13  Er2  4 sin t sin
Es2  4  Er2  4 sin t cos
where θ is the resolver shaft angle. These equations are based the 4 winding equations with Er1-4 =0,
s/c, in equation 34.3.

Stator excited, three winding

The stator excitation is Es2-4 and stator S1-S3 is short circuited if a 4 winding resolver.
Since the rotor windings are in space quadrature the voltage induced in rotor winding R1-R3 is
proportional to the cosine of the angle of rotation and the voltage applied to stator winding S1-S3. The
voltage induced in rotor winding R2-R4 is proportional to the sine of the angle of rotation and the voltage
applied to stator winding S1-S3. The relationship between input voltage Es1-3, output voltages (ER1-3 and
ER2-4), end shaft angle θ is:
Er13  Es13 cos 
Er2  4  Es 13 sin
Such a resolver can be used as an angular position transducer.

B. Four winding resolver

A more complex resolver uses two orthogonal rotor windings and two orthogonal stator windings, as
shown in Fig. 34-22c. The following computations can be performed.

i. Sine computation A voltage representing the hypotenuse of a right triangle is applied to one stator
winding, and the rotor is positioned to the angle θ. The outputs of the rotor windings represent the
coordinates of a right triangle. This method of operation is used to convert polar coordinates to
rectangular coordinates.

ii. Conversion of Polar Coordinates. Assume that voltage Es1-3 represents one polar coordinate, the
radius vector, and is applied to the stator winding of the resolver; the rotor is positioned to angle θ, which
represents the other polar coordinate. Output voltages Er2-4 and Er1-3 represent the legs of a right
triangle, of which the radius vector is hypotenuse. Thus, the polar coordinates of a point, P, can be
converted to the rectangular coordinates.

iii. Vector Addition. A resolver is capable of adding values vectorially. If the two stator windings are
excited with voltages corresponding to the coordinates of a right triangle, an output may be produced
representing the radius vector and the vector angle. This method of operation is used to convert
rectangular coordinates to polar coordinates.

iv. Conversion of Rectangular Coordinates. Assume that Es1-3 and Es2-4 represent rectangular
coordinates and are applied to the stator windings of the resolver 34.21b. The servo loop connected to
one rotor winding positions the rotor to an angle θ,
E s 13
  tan 1
E s 24
such that one output, Er2-4, is zero. That is,
Er2  4  Es 2  4 cos   Es13 sin   0
The other output, Er1-3, being orthogonal, will be at a maximum. Expressed vectorially, the output
Er13  Es13 cos   Es 2  4 sin
becomes
Er13   Es12 3  Es22  4
which is the radius vector, or equivalent polar coordinate. θ, is the radius angle, or vector angle.
1571 Power Electronics

v. Rotation of Rectangular Coordinates. A four-winding resolver may be used for the rotation of
rectangular coordinates. The orthogonal stator windings represent the input voltages corresponding to X
and Y coordinates of a point, P, the output voltages from the rotor windings correspond to X’ and Y’
coordinates, where the coordinate axes have been rotated through an angle θ. This relationship is

X’ = X cos θ + Y sin θ
Y’ = X cos θ - Y sin θ

With angle θ representing the displacement of the rotor from resolver zero, the resulting equations of
rotation are the same as those for a standard four winding resolver, stator excited. The input voltages on
the stator windings form a resultant flux vector whose position, determined by the arctangent of Y/X, is
independent of the rotor angular position. The rotor windings will at all times develop voltages which are
proportional to the sine or cosine of the angle the rotor makes with the flux vector.

34.5.4 Resolver Applications

An electromagnetic resolver transducer can be used in a wide variety of position and velocity feedback
applications which includes light duty/servo, light industrial or heavy duty applications. Resolvers, are
commonly used in servo motor feedback applications due to their good performance in high temperature
environments.

Because the resolver is an analogue device and the electrical outputs are continuous through one
complete mechanical revolution, the theoretical resolution of a single speed resolver is infinite. Because
of its simple transformer design and lack of any on internal electronics, the resolver is a much more
rugged device than most feedback devices (like a shaft angle encoder), hence applicable where reliable
performance is required in high temperature, high shock and vibration, radiation and contamination
environments.

The rotor position or angle is simply the arc tan of the voltage output of the sine winding divided by the
output of the Cosine winding. This ratio metric format provides an inherent noise reduction feature for
any injected noise whose magnitude is approximately equivalent on both windings and also results in a
large degree of temperature compensation.

The 7 functional operating parameters which define the resolver operation are:
1. Accuracy
2. Input Excitation Voltage
3. Input Excitation Frequency
4. Input Current Maximum
5. Transformation Ratio of Output Voltage to the Input Voltage
6. Phase shift of the Output Voltage from the Input Voltage
7. Null Voltage

34.5.5 Resolver-to-Digital Conversion

The resolver-to-digital converter performs two basic functions: demodulation of the resolver format
signals to remove the carrier, and angle determination to provide a digital representation of the rotor
angle. A common method of performing these functions is called ratiometric tracking conversion. Since
the resolver secondary signals represent the sine and cosine of the rotor angle, the ratio of the signal
amplitudes is the tangent of the rotor angle. Thus the rotor angle, θ, is the arc tangent of the sine signal
Vsin divided by the cosine signal Vcos:
V sin
  tan 1
V cos
The ratiometric tracking converter performs an implicit arctangent calculation on the ratio of the resolver
signals by forcing a counter to track the position of the resolver. This implicit arc tangent calculation is
based on the trigonometric identity
sin(θ−δ )= sinθ cosδ – cosθ sinδ

The sine of the difference between two angles can be calculated by cross multiplying the sine and
cosine of the two angles and subtracting the results. Provided the difference between the two angles is
relatively small (δ = θ±30°), the approximation sin(θ−δ)≅θ−δ can be used, further simplifying the
equation. Thus, if the two angles are within 30° of each other, the difference between the angles can be
calculated using the shown cross multiplication.
In the R/D converter, this equation is implemented using multiplying D/A converters to multiply the
resolver signals (proportional to sinθ and cosθ) by the cosine and sine of the digital angle, δ, which is
the output of the converter.
 Chapter 34 Transducers and Sensors 1572

The results are subtracted, demodulated by multiplying by the reference signal, and filtered to give a DC
signal proportional to the difference or error between the resolver angle, θ, and the digital angle, δ. The
digital angle, δ, stored in the counter, is then incremented or decremented using a voltage controlled
oscillator until this error is zero, at which point δ = θ (the digital angle output of the converter is equal to
the resolver angle). This incrementing and decrementing of the digital angle, δ, causes it to track the
resolver angle, θ, hence the name of this type of converter.

34.5.6 Resolver versus Encoder

With no internal electronics, resolvers can survive extreme temperatures, radiation, tolerate shock and
vibration, and withstand electrical noises, making resolvers suitable in applications where encoders
would fail. As feedback devices, resolvers can be used as alternatives to both incremental encoders and
absolute encoders. However, resolvers output an analogue signal and require a separate analogue-to-
digital converter where encoders output digital signals.
Compared to absolute encoders, single-speed resolvers provide absolute position and can be used as
absolute devices, making them alternatives when environment conditions do not allow for the use of
absolute encoders.
Encoders are better suited applications which have high acceleration and deceleration rates.

Table 34.5. Comparison between the encoders and the resolver

encoder resolver
Standard resolution counts/rev 32,640 16,384
Maximum speed counts/sec 2,448,000 500,000
Typical/best accuracy Arc-minutes 1.5/<1 (27 bits) 15/7
Typical tracking response time ms <1 15
Tolerable shock level G 5 50
Operating temperature range ⁰C 0 to 100 -55 to 175
Cost, relative similar Similar (with R/D converter)
Weight/inertia, relative lower higher

34.5 Other sensors elsewhere within this text.

12.2 Positive temperature coefficient PTC thermistors

12.2.2i Polymeric PTC devices
12.2.2ii Ceramic PTC devices
12.3.2iii Polymeric voltage variable material technologies

30.7.3 Temperature sensing resistors

30.7.4 Current sense resistors
30.7.5 Negative temperature coefficient NTC thermistors
30.7.6 Light dependent resistors