Options Abound when Selecting a Sensor for Motor Feedback
Image Source:
Andrew Fox/Shutterstock.com
By Bill Schweber for Mouser Electronics
Edited April 12, 2021
In many motion-control applications, it is necessary to know the position, speed and perhaps even acceleration of
a motor’s rotor or its load. Depending on the application and design specifics, the motor controller might
need to know these parameters precisely, approximately, or perhaps not at all. By knowing the motor situation
and rotor status, the motor controller has a closed-loop scenario, Figure 1.
Figure 1: In many motor management and control applications, real-time
details of the position and/or speed of the rotor, provided by a sensor assembly, is critical for effective
closed-loop feedback and thus accurate performance on the objectives. (Source: Bill Schweber)
Of course, a motor’s speed, position, and acceleration are closely linked. Because speed is the derivative
(time rate of change) of position, and acceleration is the derivative of speed, it is possible to determine all
three factors, even when knowing just one of them (also note the complement: speed is the integral of
acceleration, and position is the integral of speed).
However, in practice, this method of determining associated parameters is often (but not always) inadequate
because of resolution and noise. For example, knowing the rotor has completed another revolution tells you about
all three variables but with very low and usually unacceptable resolution. Depending on the application, the
resolution and accuracy needed can range from rough to moderate to precise. A CNC machine tool needs precise
rotor information, an automobile power-window controller can accept approximate data, and a clothes washer or
dryer will be satisfied with only coarse information.
Closing the Loop
For sensing the rotor position or motion, the most common options are resolvers, optical or capacitive encoders
and Hall-effect devices, in roughly descending order of precision, resolution and cost. These sensors are very
different in their physical design, implementation and electrical interface, so users must understand what is
needed, what the best choice is in a given application, and how they will interface the sensor to the circuitry
of the controller.
Incremental encoders—used when only relative position is needed, or cost is an issue—are typically
used with AC induction motors. In contrast, absolute encoders—which give a different binary output at each
position, so shaft position is absolutely determined—are often paired with permanent-magnet brushless
motors in servo applications. The application, of course, is the primary factor that determines whether
incremental or absolute information is needed.
Although most motor control is now done via a digital control loop, the sensor signal itself is either all-analog
and needs to be digitized, or is a digital signal but with voltages and other attributes that make it
incompatible with standard digital circuitry. Although some of the feedback sensors are offered with raw outputs
that can be tailored as preferred, many also have conditioned, ready-to-interface outputs that are compatible
with standard I/O ports, formats, and protocols.
Although more resolution might seem like a good idea, it might not be so in practice. Too much of an apparently
good thing—resolution—can slow a system by requiring extra processing of information that is not
needed or useful, so limiting resolution to the minimum needed is a good idea.
‘I Hereby Resolve …’
Resolvers are extremely accurate, rugged, absolute transducers of position. They are based on fundamental
transformer principles, with one primary winding plus two secondary windings, which are oriented in quadrature
(90°) with respect to each other, Figure 2. The effective turns ratio and polarity between
the primary and secondary windings varies depending on the angle of the shaft. The primary is excited with a
reference AC waveform at constant frequency, which can range from 50Hz/60Hz to several hundred kHz, and the
outputs of the secondary windings will be out of phase because of their physical placement. The peak voltages of
the secondaries will vary as the shaft rotates, and will be proportional to the shaft angle. By demodulating
these outputs using the primary signal as a reference, the resolver circuitry can provide a high-resolution
readout of the shaft angle.
Figure 2: The resolver uses a primary winding and a pair of secondary
windings in quadrature to assess the angle; it requires AC excitation and demodulation but is accurate,
rugged and provides absolute position information on power up. (Image: Analog Devices, Inc.)
The resolver is not only accurate, but it is also rugged. Resolvers have no physical contact between primary and
secondary sides, no separate brushes or bearings in addition to those at the motor itself, no points of friction
that will cause parts to wear out, and no opportunity for contaminants (such as oil) to interfere with
operating. Resolvers are used extensively in extremely challenging situations, such as angle measurement in
military guns, because of their mechanical ruggedness and performance.
However, resolvers tend to be large and relatively costly compared to alternatives, and require a relatively
large amount of power, which is often unacceptable in low-power applications. They also require relatively
complex circuitry for generation and demodulation of the AC waveforms, although this is much less of an
impediment with modern ICs. They provide absolute position indication on power up and do not require any motion
to index or determine the initial angle. This feature is a must-have in some situations and a don’t-care
in others, of course.
Encoding for Position, Not Data
An optical encoder (the term encoder here is unrelated to encoding of digital data) in an incremental position
readout, which uses a light source (LED), two photosensors in quadrature and a glass or plastic disk between
them (Figure 3). The disk has fine etched lines radiating from its center, and as it rotates
the sensors see patterns of light and dark.
The number of lines on the disk and some other techniques determines the resolution, which is typically 1,024,
2,048 or even as high as 4,096 counts per revolution. Unlike the transformer-like resolver, the optical encoder
was not a mass-market device until the development of long-life LEDs and efficient photosensors.
Figure 3: The optical encoder has a light source, quadrature light sensors
and an interposed disk with lines; it is small, low power, very easy to interface to circuitry, and can
provide excellent performance. (Image: National Programme on Technology Enhanced Lea Learning (NPTEL), a
project funded by the Government of India)
The physical arrangement of the sensors lets the encoder determine the direction of rotation. A basic circuit
translates the pulse trains from the two sensors (called A/B outputs) into a pair of bit streams indicating both
motion and direction (Figure 4).
Figure 4: The A/B quadrature and index outputs of the optical encoder are
compatible with many interfaces and motion-control processor I/O ports. (Source: Bill Schweber)
However, the encoder is an incremental, not absolute, indicator of motion. To determine absolute position, most
encoders add a third track and photosensor as an indicator zero reference track; the shaft must rotate enough to
pass the zero reference position for this to signal. True relative position readout can be added to an optical
encoder, but these add complexity to the unit.
Optical encoders offer very good resolution, but they are not as rugged as resolvers. Dirt can interfere with the
optical path and the encoder disk can get dirty. However their performance is more than adequate for many
applications, and they are small, lightweight, low power, easy to interface and low cost.
Typical optical encoders for motor and rotation applications are the similar HEDS-9000 and
HEDS-9100 two-channel modules from Avago Technologies (Broadcom). These high-performance, low-cost
modules consist of an LED source with lens and a detector integrated circuit enclosed in a small, C-shaped
plastic package, along with drive and interface electronics (Figure 5). They have a highly
collimated light source and special photodetector physical arrangement, so they are very tolerant of mounting
misalignment.
Figure 5: The Avago HEDS-9000 and HEDS-9100 two-channel modules offer small
size and mounting flexibility; the interposed optical disk is ordered separately with the desired resolution
of counts per revolution. (Image: Avago Technologies/Broadcom)
Note that the disk, called the code wheel, is purchased separately, with resolution of 500 CPR and 1,000 CPR for
the HEDS-9000, and between 96 CPR and 512 CPR for the HEDS-9100. The modules provide two channels of
TTL-compatible A and B digital outputs and require a single 5-V supply (Figure 6).
Figure 6: The Avago HEDS-9000 and HEDS-9100 two-channel modules offer small
size and mounting flexibility; the interposed optical disk is ordered separately with the desired resolution
of counts per revolution. (Image: Avago Technologies/Broadcom)
CUI AMT10 Series is an alternative to
the optical encoder, based on capacitive principles instead of optical ones (Figure 7). These
encoders offer a
range of rugged, high-accuracy, modular units available in incremental and absolute versions, with up to 12-bit
(4,096 count) resolution selectable by the user from among 16 values via a four-position dual in-line package
(DIP) switch. The complementary metal-oxide-semiconductor (CMOS)-compatible A/B quadrature outputs of these
units are reported via a standard serial peripheral interface (SPI).
Figure 7: The CUI AMT10 capacitive encoder might look like an optical
encoder from the outside, but the underlying operating principle is very different. (Image: CUI, Inc.)
Unlike optical encoders, the CUI AMT devices use a repeating, etched pattern of conductors on the moving and
non-moving parts of the encoder. As the encoder rotates, the relative capacitance between the two parts
increases and decreases, and this change in capacitance is sensed, somewhat analogous to the outputs of the
phototransistors in an optical encoder. Dirt and other contaminants have little detrimental effect here.
Keep in mind that a resolver or encoder is also a mechanical device with mounting considerations as well as
electrical compatibility requirements. To minimize stocking and inventory issues, CUI offers the AMT10 series
with a broad range of sleeves, covers and mounting bases (Figure 8) so the same basic encoder
can be used across
a wide range of shaft diameters and installations.
Figure 8: In practice, a wide range of shaft and mounting situations must be
addressed by the encoder; CUI offers a full set of color-coded sleeves and other accessories so that a
single encoder can serve many applications. (Image: CUI, Inc.)
Resolvers and encoders can produce basic readouts with resolution as high as 1/100 of a degree (0.6 arc minutes)
or better, but accuracy is not the same as the resolution (again, some applications are more concerned with one
of these than the other). Regardless of whether the design uses a resolver or encoder, error sources occur
because of temperature, speed of tracking of changes, undesired phase shifts, and other factors. However,
vendors of these units have devised ways to eliminate, cancel, or compensate for many of these shortcomings,
often by using IC-based circuitry between the raw sensor output and the conditioned output that goes to the
system controller.
Hall-Effect Devices Come on Strong
Another class of encoding or sensor device is also based on a timeworn principle, but which requires modern
semiconductor electronics and packaging to become widely affordable, available and effective. Further, the
critical interface circuitry, which can make use of the minuscule voltage and easily interface it to a system,
is now available on-chip, further simplifying the use of this technology. Hall-effect devices can be used to
sense current flow through a conductor that is part of the sensor, or the presence or absence of a nearby
magnetic field.
What we know as the Hall effect was discovered by Edwin Hall in 1879: a potential difference—the Hall
voltage—is produced across an electrical conductor, at right angles to an electric current in the
conductor and a magnetic field perpendicular to the current (Figure 9).
Figure 9: The principle of the Hall-effect device involves current, voltage
and magnetic fields orthogonal to each other. (Image: National Programme on Technology Enhanced Learning
(NPTEL), a project funded by the Government of India)
Some Hall-effect sensors go far beyond incorporating only the sensor element itself. The Melexis
MLX90367 Triaxis position sensor is a monolithic absolute sensor IC sensitive to the flux density
applied orthogonally and parallel to the IC surface. It is sensitive to the three components of the flux
density, which allows the MLX90367 (with the correct magnetic circuit) to decode the absolute position of any
moving magnet (such as a rotary position from 0 to 360°).
Internally this 12-bit resolution device includes on-chip signal processing, with a microcontroller and DSP,
(Figure 10), so it can perform needed calculations plus corrections for inherent nonlinearities
and more, (Figure 11). It also supports a wide range of user-selectable functions and features,
and various output formats, including an advanced format with built-in error correction called SENT (SAE
J2716-2010), which is widely used in automotive applications.
Figure 10: The Melexis MLX90367 is much more than just a Hall-effect sensor;
it includes an amplifier, digitizer, processor, firmware and I/O. (Image: Melexis N.V)
Figure 11: The processing capabilities in the MLX90367 allow it to
significantly improve performance by correcting some avoidable errors in the linearity of the basic
Hall-effect transducer. (Image: Melexis N.V)
Most Hall-effect magnetic encoders use a wheel attached to the motor shaft, and the wheel has a set of magnetized
north and south poles around its perimeter; it is the magnetic analogy to the optical encoder slotted wheel. The
wheel is usually made from an injection-molded ferrite embedded with the pole array. A typical wheel is
magnetized with 32 poles (16 north and 16 south), so the resolution is far less than for an optical encoder or
resolver, but is often enough for many situations. A typical installation has three Hall-effect sensors, spaced
120° apart electrically, to sense commutation of the wheel.
Summary
Designers who must sense motor position, speed, or acceleration have a wide variety of options covering the many
key parameters and performance attributes. Resolvers, optical and capacitive encoders, and Hall-effect devices
all have long and proven track records, plus extensive support via applications know-how.
The choice can be driven by one overriding factor—such as ruggedness or low power—or by traditional
and customary use in a given situation. Once the basic technology to be used is decided, many viable vendors and
parts from each are available, so the decision on a specific device might take some research to better
understand the tradeoffs.
References
- http://www.mouser.com/applications/robots-go-all-electric/
- http://www.mouser.com/applications/stepper-motors-smart-drivers/
- http://www.mouser.com/applications/resolver-encoder-motor-control/
- http://www.mouser.com/applications/considerations-choosing-advance-robotics/
- http://www.mouser.com/applications/industrial_vector_drive/
- http://www.mouser.com/applications/industrial_motor_control_overview/
Bill Schweber is an electronics engineer who has written three textbooks on electronic
communications systems, as well as hundreds of technical articles, opinion columns, and product features. In
past roles, he worked as a technical web-site manager for multiple topic-specific sites for EE
Times, as well as both the Executive Editor and Analog Editor at EDN.
At Analog Devices, Inc. (a leading vendor of analog and mixed-signal ICs), Bill was in
marketing communications
(public relations); as a result, he has been on both sides of the technical PR function, presenting company
products, stories, and messages to the media and also as the recipient of these.
Prior to the MarCom role at Analog, Bill was associate editor of their respected technical
journal, and also
worked in their product marketing and applications engineering groups. Before those roles, Bill was at Instron
Corp., doing hands-on analog- and power-circuit design and systems integration for materials-testing machine
controls.
He has an MSEE (University of Massachusetts) and BSEE (Columbia University), is a Registered
Professional
Engineer, and holds an Advanced Class amateur radio license. Bill has also planned, written, and presented
on-line courses on various engineering topics, including MOSFET basics, ADC selection, and driving LEDs.
