Overcoming brushless dc motor limitations

MOTORS.JUN–Vernitron–pm

Overcoming brushless dc motor limitations

By taking a few basic design precautions, the designer can take full
advantage of the brushless motor's features

BY NGON T. DANG Vernitron Corp. Herndon, VA

Continuous-rotation dc brushless motors, though more costly than their
brush-based counterparts, have unique features and characteristics that
make them a preferred choice for some applications. These features include
high speed, high torque, and greater thermal dissipation efficiency.
However, before opting for a brushless design, the designer must overcome
inherent drawbacks of the motor such as higher cost, motor drive
complexity, and high ripple torque. Whatever the drawbacks, brushless
motors (see Fig. 1) have essentially the same motor operation
characteristics as their brush-based counterparts. In fact, when supplied
with the proper electronic control system, brushless dc motors can
normally be directly substituted. Brush-based motors are relatively
straightforward (see box, “Brush and brushless–the difference inside”).
They have less complex (or no) electronic control circuits, fewer power
devices, and no required positional feedback devices. This simplicity
usually makes these motors easier to specify and design, as well as less
costly. In addition, commutation is transparent to a controller, if used.
The analog velocity loop on brush motors is easy to close, so a stable
velocity state can be reached when changing between motor speeds.
Maintaining brushes can be time consuming–especially where motors are not
easily accessible. Poor maintenance can adversely affect performance and
eventually cause motor failure. Also, several factors limit applications.
These include brush dust, brush-to-commutator arcing and EMI, and
mechanical noise caused by the commutator contact with the brushes.
High-speed operation is difficult for conventional brush-type dc motors
because the high energy switched by the brushes is destructive and results
in a short motor life. This energy switching limits practical rotational
speeds in brush motors to 10,000 rpm. Another potential problem is the
difficulty of phase shifting, which must be done mechanically. Phase
shifting, or phase advancing, can be used to optimize the motor
performance at various velocities by shifting the position of the brushes
to change the point of commutation. This is typically at a point at which
the back electromotive force (EMF), is “centered” between commutation
points. In practice, this can be done only unidirectionally since a
physical beginning point of reference must be maintained to align the
brushes properly as their position is advanced. Brush motors are also
subject to friction between brushes and commutators. This friction is a
drain on both motor power and torque. To compensate, more input power must
be applied, making the motors less energy efficient (see table). For
same-size motors, one brushless and the other brush based, the brush-based
motor can require up to 35% more power.

Brushless attributes Brushless motors offer several advantages over brush
motors. They may be operated at much higher speeds and at full torque at
those speeds–the high energy can be delivered by the drive circuits.
These motors produce approximately double the output torque over brush
motors of the same size. The stator is the wound member and may be
mounted in a substantial heat sink to minimize temperature rise and
prolong bearing life. Even without a heat sink, putting the windings on
the stator instead of the rotor improves heat dissipation. Because of this
greater thermal dissipation efficiency, a brushless motor can be made
quite small relative to the brush motor. The absence of a brush has
several immediate advantages. No brush maintenance is needed, and the
motor usually lasts as long as the bearing. This is ideal for long-life
applications where motors are difficult or impossible to reach–as in a
space satellite. The motors are more flexible in their diameter-to-axial
sizing, contact noise is eliminated, and input power requirements are
lower. Because of the absence of arcing, the brushless motor can be
applied in explosion-prone applications while usually avoiding the special
housing elements necessary to explosion-proof a conventional dc motor.
Although brush motors have been used extensively in space environments,
the preparation of a motor for such an application is expensive and time
consuming. Brushless motor speed is inherently variable using the
controller-supplied voltage. In addition, these motors have higher
torque-to-inertia and torque-to-power ratios, as well as a faster response
time.

Disadvantages On the negative side, brushless motors (with controllers)
tend to be approximately 10% to 30% more costly than comparably sized
brush motors. However, this cost difference has been narrowing. Brushless
motors require more software and more power stages, adding to motor
complexity (see Fig. 2). This is exacerbated by the presence of the rotor
position sensors. Also, if the relatively complicated logic system is not
properly designed, a short circuit or other damage can result. In general,
the greater the number of motor system components, the lower the MTBF.
Another negative is that ripple torque is higher–reaching 15% in a
sinusoidally wound, three-phase brushless motor, compared to 2% to 6% for
a brush motor. Also, excessive coasting is possible due to the low
friction of brushless motors. This potential problem may be overcome by
adding a special damping circuit or other mechanical damping device.
Because of the high commutating frequencies, phase shifting is required.
Fortunately, the brushless system's electronics make it relatively easy to
incorporate bidirectionally. Brushless motors normally have linear speed
torques. An exception might be found in a motor designed for an extremely
small working envelope. In such a case, the motor will reach its
saturation point (the point at which the magnetic iron in the armature
reaches its maximum capacity to carry magnetic flux) before completing its
linear curve. Small instrument motors sometimes must be designed with such
nonlinear performance.

Designing with brushless motors Motor windings are inductive. This gives
rise to two considerations that should not be overlooked. The first
involves motor speed. Since the windings' electrical time constants are in
the area of 1 ms, commutating frequencies above several hundred hertz need
special treatment. Commutating frequency is equal to the number of pole
pairs in the machine times the speed in revolutions per second. For units
that must operate at higher speeds, some provision for shifting the
commutation points must be made. This can be done either mechanically or
electronically. The second consideration is that the motor inductance
causes high-voltage spikes to appear across the power transistors as they
are switched off. These must be provided for in the design by use of
either high-voltage transistors or protective zener diodes or other
transient suppressors. Most brushless motor systems require current
limiting to avoid either inadvertent demagnetization of the permanent
magnet rotor or overheating during starting and fast reversals. The logic
system must be examined for the possibility of improper outputs during
power application. For instance, if the logic states are such that both
the transistors at one corner of the three-phase delta are turned on, the
resulting short circuit will be disastrous. An easy and fruitful test
during design of a motor-control application is to step the logic through
its sequence and note that the motor steps through its rotation with no
reversals or long steps. If this is not done, it is possible that the
motor will run, but that during one segment of its rotation its torque
will be reversed or nonexistent. High current will result, but may be
inadvertently overlooked. Many motor specifiers have found that an
incoming test of the motor torque is expensive to implement. In these
cases the incoming electrical inspection can be made on the back EMF
constant of the motor (K_b ). This parameter is directly proportional
to the torque sensitivity and is more easily measured. The technique
involves driving the motor at a constant speed and measuring the generated
voltage at the motor terminals. Acceptance limits can be set based on the
back EMF constant. The use of trapezoidally wound motors can reduce
torque ripple, compared to that of a sinusoidally wound motor. In fact,
torque ripple can virtually be eliminated through a technique known as
sine-wave winding excitation. This technique requires more sophisticated
electronics. The ripple torque can be reduced in brushless motors designed
for square-wave excitation by reducing the commutation angle through the
use of more phases (see�20Fig. 3). Cogging is another condition that can
be ameliorated. The phenomenon is an unavoidable magnetic detent caused by
the stator construction where the windings are placed in slots in the iron
armature. A reaction with the motor's poles creates positions of minimum
and maximum reluctance. The rotor naturally detents into a position of
minimum reluctance. While cogging cannot be eliminated, it can be
minimized. One approach available to motor designers is to skew stator
slots or field magnets relative to each other. This permits a gradual
transition, with the slot not suddenly encountering the magnet. Another
approach involves placing the detenting effect from one pole to another
out of phase, and thus obtaining an averaging effect. With this approach,
the number of poles (4, for example) could not be divided evenly into the
number of slots (15, for example). Maximum allowable current density is
usually not published by the motor supplier, because it varies from one
application to another. However, too low a current density means that the
motor has been overspecified. Too high a current density may cause over
power and motor failure. To determine current densities, system designers
should provide the motor supplier with the application's load and velocity
information, and the supplier can simulate these conditions.

For more information, contact Dennis Akers, Vernitron Corp., at
703-478-9800, or fax 703-478-9559

RUNNING HEADER:

Brushless dc motors

CAPTIONS:

Fig. 1. When supplied with the proper electronic control system,
brushless dc motors can be directly substituted for their brush-based
counterparts.

Fig. 2. Although a brushless motor's implementation is more complex than
that of a brush motor, the device's qualities make it the better option in
many applications.

Fig. 3. For a two-phase brushless motor, the commutation angle is 90
degrees, yielding a ripple torque of on average 17% peak-to-peak (a).
Using the same motor having a three-phase delta system with a commutation
angle of 60 degrees, the ripple torque averages at 7% peak-to-peak (b).

Table. Comparison of specifications for brushless and brush-based motors
of approximately the same size.

BOX:

Brush and brushless–the difference inside

Brush and brushless motors normally exhibit similar linear speed-torque
curves and deliver high starting torque. But how this output performance
is achieved differs. Brush motors use commutators and carbon brushes to
distribute (or switch) current through the windings as the motor rotates.
This is essentially the same mechanical commutation scheme as shunt (wound
electromagnetic-based field and armature) motors. In a
continuous-rotation brushless motor, this mechanical switching is replaced
with electronic switching. These motors are not simply ac motors powered
by an inverter, but contain rotor position feedback of some kind so that
the power input waveforms are kept in the proper timing with the rotor
position. The key identifying feature of a brushless motor then is that
some form of switching is necessary, except for limited angle rotation
operations. Brushless motors typically employ rare-earth or other
permanent magnets, either in the inside or outside rotor. Because the
magnets occupy less room and weigh less than comparable windings, the
motors can be made smaller and lighter. Motor energy efficiency is also
enhanced because no current is required to power the magnets. A brush
motor uses wound elements in the rotor and permanent magnets bound to a
stationary stator ring. This construction is usually reversed in a
brushless motor, with the wound elements in the stator and the permanent
magnets in the rotor. Because of rapidly increasing manufacturing costs at
higher horsepower ratings, permanent magnet motors–both brush and
brushless–are practically restricted to smaller sizes up to about an
8-in.-diameter frame size, although special applications have spawned much
larger configurations. In a brush motor, a series of electrically
separated motor windings are connected to the commutator ring. Current is
carried by two spring-loaded brushes, through the commutator, and into two
windings at a time. The current in the windings creates magnetic fields,
which react with the stator's permanent magnetic field. These magnetic
forces cause the rotor to rotate. As the rotor revolves, the brushes make
and break connections, through the commutator, with different winding
pairs. The resulting moving magnetic field drives the motor. In a
brushless motor, the stator is wound with electromagnetic coils that are
connected in a polyphase configuration, while the rotor consists of a soft
iron core and permanent magnet poles. A rotor position sensor assembly
contains enough sensing devices to define the rotor position to the
required resolution. Finally, commutation logic and switching electronics
convert the rotor position information to the proper excitation for the
stator phases. Rotor position is often sensed by inexpensive Hall-effect
transducers. Other sensing approaches include absolute encoders,
incremental optical encoders with decoding capability, optical encoders
with commutation tracks, resolvers, synchros, and Wiegand wires. In some
cases separate feedback devices like Hall-effect transducers are dispensed
with. Position feedback is achieved through sensing of current or voltage
induced into the stator windings from the rotor. This allows the
controller to ascertain electrical phase, and thus rotor position.
Controllers may be remote from the motor, or they may be packaged
together. In applications where a brushless motor replaces a brush motor,
there may only be room for integral construction. An integral motor would
also be appropriate if the system designer wants a two-wire power input to
the motor.