Brushless Permanent Magnet Motor Design – Basic Concepts

Motor Action the electromagnet, and hence the permanent magnet will poles will
be repelled by the like electromagnet poles.
All that’s required to understand how a permanent motor (PM)
works is the rudimentary knowledge that magnets are attracted to
iron, that opposite magnet poles attract, that like magnet poles repel
each other, and that current flowing in a coil of wire makes an elec-
tromagnet.

Figure 1 shows a magnet free to rotate inside a steel stator with

two protrusions. If the magnet is spun slowly, then it will have a ten-
dency to come to rest at a position where the magnet poles are
aligned with the protrusions in the steel ring, i.e. 𝜃 = 0° or 𝜃 = 180°.
This happens because the attractive force between a magnet and
iron increases as the physical distance between them decreases.
Some of the attraction force will be tangential, and hence, will pro-
duce torque.

Figure 2. Torque experienced by the magnet in Figure 1.

Figure 3. Current-carrying windings added to protrusions.

Figure 1. A magnet free to spin inside a steel ring having two protrusions.
To get the magnet to turn continuously, it is common to employ
Figure 2 graphically represents this torque as function of mag- more than one set of coils. These sets are called phase windings or
net angular position. The points where the torque is zero are called simply windings, as shown in Figure 4. By creating electromagnet
detent positions. When the magnet is aligned with the protuber- poles on the stator that attract and repel those of the bar magnet,
ances, any small disturbance causes the magnet to restore itself to the latter can be made to rotate by successively energizing and deen-
the aligned position. Therefore, the detent positions are considered ergizing the phases in a process called commutation.
stable. On the other hand, bigger disturbances may lead the magnet
to move away from the protrusions, thus unaligned detent positions
are considered unstable. The torque described here is called reluc-
tance torque, and more commonly known as cogging torque. In most
applications, cogging torque is undesirable.

Now, if the protrusions in Figure 1 are wound with current car-

rying coils, the protrusions become the poles of an electromagnet,
as shown in Figure 3. The magnet now tries to align itself in a way
that the opposite poles in each magnet will be attracted to each
other. The torque produced by this interaction is called mutual
torque or alignment torque. It is this torque that is used in PM motors
to do work. If the current direction is reversed, so are the poles of
Figure 4. A motor structure having three phases.
Magnet Poles and Motor Phases motor back in identical magnetic orientation. In Figure 6, the rotor
needs to move only 180 °M to have the same magnetic orientation.
It is possible to build a brushless PM motor with any even num- Thus, 360° E is the same as 180° M for this case.
ber of poles and any number of phases greater than or equal to one,
as long as they can fit inside the motor casing. Three phase motors
are the most common. The reason for this is that three phase motors
minimize the number of power electronic devices to control the
winding currents.

The choice of magnet poles offers more flexibility. A greater

number of magnet poles usually creates a greater torque for the
same current level. On the other hand, more magnet poles imply
having less room for each pole. Eventually, a point is reached where
the spacing between rotor magnet poles becomes a significant per-
centage of the total room on the rotor, and torque no longer in-
creases. The optimum number of magnet poles is a complex function
of motor geometry and material properties.

Poles, Slots, Teeth and Yokes

The motor in Figure 4 has concentrated or solenoidal windings.

That is, the windings of each phase are isolated from each other and
concentrated around individual poles called salient poles, pretty
much like a solenoid is wound. A more commonly occurring alterna-
tive is to use distributed windings, where the windings of each phase
overlap, as shown in Figure 5. Figure 6. a) windings of phase A. b) windings of phase B. c) windings of phase
C. d) Complete motor with distributed windings.

The relationship between mechanical and electrical degrees is

related to the number of magnet poles on the rotor. If 𝑁𝑚 is the
number of magnet poles on the rotor facing the air gap, i.e., 𝑁𝑚 = 4
for Figure 6, the relationship can is given by:

𝑁𝑚
𝜃𝑒 = 𝜃𝑚 (1)
2

where 𝜃𝑒 and 𝜃𝑚 are electrical and mechanical position, respectively.

Sing magnets always have two poles, it is common to define a pole
pair as one North and one South magnet pole facing the air gap. In
Figure 5. Concentrated winding motor (left) vs distributed winding (right). this case, the number of pole pairs is equal to 𝑁𝑝 = 𝑁𝑚 /2 and the
relationship simplifies to:
In a distributed winding motor, the stator has teeth that pro-
trude toward the magnets on the rotor from an outer ring of ferro-
𝜃𝑒 = 𝑁𝑝 𝜃𝑚 (2)
magnetic material called the stator yoke or back iron. In between the
teeth are slots that are occupied by the windings. As shown in Figure
6, each winding travels from one slot, across a number of teeth
(three in this case), then down the next slot. The teeth enclosed by a Differentiating Eq. 2 with respect to time gives the relationship
winding forms the pole for that coil. The rotors in Figure 6 are formed between electrical and mechanical frequency or speed:
from circular arc shaped magnet pieces attached to an inner ring of
ferromagnetic material called rotor yoke or back iron. The magnets
𝜔𝑒 = 𝑁𝑝 𝜔𝑚 (3)
have alternating poles as one proceeds around the rotor periphery.
The angular speeds are related to the frequencies by 𝜔 = 2𝜋𝑓.
Mechanical and Electrical Measures
Therefore, the electrical and mechanical frequencies are related by:
In electric motors, it is common to define two measures of po-
sition and speed. Mechanical position and speed are the quantities 𝑓𝑒 = 𝑁𝑝 𝑓𝑚 (4)
related to the motor shaft. When the rotor shaft makes one com-
plete revolution, it travels 360° mechanical degrees (°M) or 2𝜋 me- It is common practice to specify motor mechanical speed 𝑛 in
chanical radians (radM). RPM. For reference, the relationships among 𝑛, 𝑓𝑚 and 𝑓𝑒 are given
by:
Electrical position is defined such that movement of the rotor
𝑚𝑁 𝑁𝑝
by 360 electrical degrees (°E) or 2𝜋 electrical radians (radE) puts the 𝑓𝑒 = 120 𝑛= 𝑛 (5)
60
 The last equation is useful because it describes the rate of fre- The rotor configurations in Figure 7-a-d show surface-mounted
quency at which commutation must occur for the motor to turn at a magnets. The magnet shapes in Figure 7-b and Figure 7-c appear pri-
given speed in RPM. The inverse of this frequency gives the commu- marily due to manufacturing reasons. These shapes are easy to pro-
tation time period, i.e., the length of time over which the energizing duce from a rectangular block of magnet material. When magnet
of a phase completes one cycle of operation. The fundamental elec- material is bonded rather than sintered, the rotor magnets can be
trical frequency 𝑓𝑒 influences the design of the power electronics formed from a solid ring of magnet material as shown in Figure 7-d.
used to drive the motor. As 𝑓𝑒 increases, power electronics must act The remaining configurations are interior permanent magnet mo-
faster to keep the motor shaft turning. This means that power elec- tors.
tronics become more expensive as 𝑓𝑒 increases. Because of this, it is
common to use fewer magnet poles, i.e., reduce 𝑁𝑚 for motors de- Figure 8 shows some stator possibilities for inner rotor motors.
signed to operate at high speeds. This causes a reduction in torque The stator can be either slotted or slotless. The slotted stator in Fig-
production efficiency. ure 8-a has a small magnetic air gap, making the permeance coeffi-
cient and resulting air gap flux density greater. In addition, the large
Motor Size contact area between the windings and the stator ferromagnetic ma-
terial promotes good heat conduction away from the windings to the
For radial flux motors, the size of the rotor and the torque pro- outer stator surface where it can easily be removed.
duced by the motor are related by the following equation:
In the slotless stator in Figure 8-b, windings are formed into a
𝑇 = 𝑘𝐷²𝐿 (6) ring that fits inside the stator yoke and separated from the rotor by
a small physical air gap. This construction exhibits no cogging torque
Where 𝑇 is torque, 𝑘 is a constant, 𝐷 is the rotor diameter and since the reluctance seen by the rotor magnets does not vary with
𝐿 is the axial rotor length. position. There is more room for the windings in a slotless motor, but
the reduced thermal conductivity to the outer stator surface reduces
Motor Constructions the allowable current density in the windings. In this type of con-
struction, the magnetic air gap encompasses the physical air gap as
Radial Flux Motors well as the radial thickness of the windings. This drives the perme-
ance coefficient and air gap flux density down. As a result, the per-
Most of the radial flux motors have inner rotors, which means
formance of a motor using slotless stator is almost always signifi-
that the rotor is placed inside the stator. Figure 7 shows a variety of
cantly lower than that of an otherwise equivalent motor having a
the most common inner rotor types.
slotted stator.

Figure 8. Inner rotor stator possibilities.

There are some reasons for the overwhelming prevalence of

motors having inner rotors, such as the ease of heat removal, be-
cause the windings are on the outside, and the containing of the ro-
tating element. In some applications, these attributes are not as im-
portant as the benefits gained from having an outer rotor involving
an inner stator. These motors are called out-runners or inside-out
motors, as shown in Figure 9.

Since the stator teeth are pointed outward, this motor is rela-
tively easy to wind. For a given outer radius, an out-runner motor has
a much larger air gap radius than that of an inner rotor motor. As a
result, higher torque is achievable, provided the ohmic losses in the
stator can be dissipated.

Figure 7. Inner rotor possibilities

 at ever increasing distances from one another, and as a result, it sig-
nificantly increases stator manufacturing time and cost. This is much
different than the laminations for the radial flux motor, where the
slots are cut as part of the lamination stamping process.

Radial Flux Permanent Magnet machines vs Axial Flux Permanent

Magnet Machines

The evolution of technology made possible a significant in-

crease in the power density of the electrical machine. There is an in-
herent limit to this increase for conventional radial flux PM (RFPM)
machines because of the following reasons:

• The bottle-neck feature for the flux path at the root of the
rotor tooth in the case of brushless machines with external
rotors (Figure 11);
Figure 9. Cross section of outer rotor motor. • Much of the rotor yoke is hardly utilized as a magnetic circuit;
• Heat from the stator winding is transferred to the stator core
Axial Flux Motors
and then to the stator yoke – there is poor heat removal
In this type of motor, the windings are oriented along the radial through the stator air gap, rotor and shaft without forced
direction, and the flux flows in the axial direction, as shown in Figure cooling arrangements.
10. In Figure 10-a, we have a view of a rotor with magnets of alter-
nating polarity. These magnets produce axial flux that interacts with
windings in radial slots, as shown in Figure 10-b.

Figure 11. Magnetic circuit in (a) RFPM Machine (b) AFPM.

AFPM machines are recognized as having higher power density

than RFPM machines, and therefore are more compact. Since the in-
ner diameter of the core of an AFPM machine is usually much greater
than the shaft diameter, better ventilation and cooling can be ex-
pected. Other than that, AFPM machines possess the following ad-
vantages in comparison to RFPM machines:
Figure 10. Axial flux motor configurations.
• AFPM machines have a planar and somewhat adjustable air
In many applications, one rotor is mated to one stator, as
gap;
shown in Figure 10-c. This configuration, although simple, is unbal-
• Capability of being designed to possess a higher power den-
anced. In addition to torque, it suffers high axial force, because the
sity with some saving in core material;
rotor magnets attempt to close air gap. By converting the rotor yoke
• The topology of an AFPM machine is ideal to design a modular
into a second stator, as shown in Figure 10-d, the rotor forces are
machine in which the number of the same modules is ad-
balanced. This configuration sandwiches one rotor between two sta-
justed to power or torque requirements.
tors, and as a result, improves motor performance.
• A large number of poles can be accommodated, which make
The axial flux motor suffers from a problem related to stator the AFPM machines a suitable choice for high frequency or
construction. Because flux flows axially, the stator must be laminated low speed operations.
circumferentially. That is, the stator is often constructed by winding
a ferromagnetic ribbon concentrically. This construction orients slots
 RFPM machines on the other hand, have advantages of their
own. For most slot/pole combinations, the radial magnetic forces are
balanced, as are the axial centering forces, so that virtually all the
electromagnetic forces contribute to the torque. This is good for
bearing life and helps to minimize acoustic noise. Interior-rotor radial
flux machines can easily be manufactured with different core lengths
for different ratings, using the same laminations and even the same
magnets. Laminated construction permits the rotor and stator cores
to be punched at the same time from a single blank, reducing mate-
rial cost. Especially with higher pole-numbers, the rotor weight and
inertia can be reduced by including holes in the rotor punching.