Brushless Permanent Magnet Motor Design
Brushless Permanent Magnet Motor Design
Brushless Permanent Magnet Motor Design
Duane Hanselman
Electrical and Computer Engineering
University of Maine
Orono, ME 04469
USA
brushlessmotor@ieee.org
www.eece.maine.edu/motor
Brushless
Permanent Magnet
Motor Design
Second Edition
THE WRITERS COLLECTIVE
The design of brushless permanent magnet motors is not a simple task. On a more
general level, motor design requires knowledge of magnetics, mechanics, thermody-
namics, electronics, acoustics, and material science. On a more specific level, it
requires knowledge of performance requirements and constraints imposed by the
intended motor application. Given this body of knowledge, motor design involves
finding an optimal solution for the least cost. This text focuses on the magnetic
aspects of motor design. Other general aspects listed above are considered in the
design process, but detailed design information about other areas is not provided.
4.1 Assumptions
Besides the performance requirements discussed above, other initial assumptions are
necessary to more clearly define and focus the initial design of brushless permanent
magnet motors. Some of these assumptions add restrictions and others identify con-
ventional design techniques.
Rotational Motion
It is assumed that rotary motion is desired. While the design techniques developed
here are easily applied to motors having linear motion, initial work will focus on
rotary motion where the rotor is inside a stator.
Surface-Mounted Magnets
Most brushless permanent magnet motors have magnets mounted on the rotor sur-
face facing an air gap. For this reason, initial work will focus on this topology. In
some motors the permanent magnets are buried within steel structures. Interior per-
manent magnet topologies generally find application for three reasons. First, by
burying magnets, it is possible to employ flux concentration. Second, enclosing mag-
nets in steel can make the rotor structurally stronger and therefore allow operation at
higher speeds. Finally, by burying magnets it is possible to drive a motor over a
wider speed range through the use of field weakening control.
Chapter 4
Brushless Motor Fundamentals
67
4.2 Fundamental Concepts
In brushless motor design, mutual torque and back EMF are the two fundamental
parameters to be determined. These two parameters are intimately linked through
(3.43), so knowledge of one provides information about the other. While the BLi and
BLv laws can be used to determine the torque and back EMF respectively, it is more
convenient to compute flux linkage, and then employ Faradays law to obtain the
back EMF. Then, (3.43) can be used to determine torque.
Magnetic Circuit Model
Consider the motor cross section shown in Fig. 4-1. Here the rotor contains N
m
=4
magnet poles facing the air gap. As a result there is a factor of two difference between
electrical and mechanical measures,
e
=(N
m
/2)
m
. For simplicity the stator is shown
without slots or windings. The magnet flux leaving North poles at the air gap crosses
over to the stator and splits into two equal sections, each traveling in the opposite
direction and crossing the air gap toward South poles at the air gap. For one half of a
North and South pole facing the air gap, this flux flow is illustrated by the flux path
on the right side of the figure. Flux flowing between each of the other adjacent half
pole pairs follows accordingly.
Chapter 4 Brushless Motor Fundamentals 68
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-1. Fundamental motor structure and associated flux paths.
N
S
stator
rotor
N
S
In addition to the primary flux path shown, some magnet flux jumps from one
magnet to the next in the air gap without passing into the stator, as illustrated by the
path in the air gap on the right of Fig. 4-1. The flux that follows this path is often
called magnet leakage flux.
Because the flux paths shown in Fig. 4-1 repeat for every adjacent half pole pair, it is
only necessary to model one such pair as shown in Fig. 4-2. In this figure, the rotor
and stator steel areas are modeled simply as reluctances R
r
and R
s
respectively. The
two half magnets are modeled as a flux source
r
and associated magnet reluctance
R
m
, with the direction of the flux source dictating the magnet polarity. Primary flux
flow from the magnets across the air gap into the stator flows through the air gap
reluctances denoted R
g
. Leakage flux from one magnet to the next flows through the
leakage reluctance R
l
. The three circuit fluxes are the magnet flux , the air gap flux
g
, and the leakage flux
l
.
Before determining a back EMF, the magnetic circuit must be solved to determine
the air gap flux density B
g
. Rather than solving the magnetic circuit as shown in Fig.
4-2, it is convenient to simplify the circuit as shown in Fig. 4-3. Since the right magnet
and the rotor reluctance are in series, they are swapped in Fig. 4-3a. This places the
two half magnets next to each other and places the rotor reluctance next to the other
reluctances. At this point it is difficult to determine an analytical description of the
leakage reluctance. However, the percentage of flux that travels the primary flux path
Magnetic Circuit Model 69
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-2. A magnetic circuit model for the structure shown in Fig. 4-1.
R
s
R
l
R
g
R
g
R
r
R
m
R
m
r
r
N
S
S
N
Stator
Rotor
Air
gap
g
across the air gap relative to the magnet flux can be estimated. That is, the air gap
flux can be written in terms of the magnet flux as
g
=K
l
, where K
l
is a leakage factor
that is typically slightly less than one. Using this relationship, the next step in simpli-
fying the magnetic circuit is to eliminate the leakage reluctance R
l
as shown in Fig. 4-
3b. This is possible since very little flux follows the leakage path, and it is desirable
since it is difficult to find an expression for R
l
. To compensate for the flux that follows
this leakage path, the solution for will be multiplied by an estimate of K
l
to obtain
g
. With the leakage reluctance eliminated, the rotor and stator steel reluctances are
in series, thereby allowing them to be lumped into a single reluctance as shown in
Fig. 4-3b as well.
Chapter 4 Brushless Motor Fundamentals 70
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-3. Simplifications of the magnetic circuit in Fig. 4-2.
R
s
R
l
2R
g
R
r
R
m
r
N
S
R
m
r
N
S
R
s
+R
r
2R
g
R
m
r
N
S
R
m
r
N
S
2R
m
r
N
S
2K
r
R
g
2R
m
r
N
S
(a) (b)
(c) (d)
R
s
+R
r
2R
g
The two magnet halves in series in Fig. 4-3b can be simplified as shown in Fig. 4-3c.
From an electrical circuit point of view, the simplified magnet shown in Fig. 4-3c is
found by determining the Norton equivalent circuit of the two series magnet halves.
That is, the simplified flux source is
r
since that is the flux that would flow if a
short were placed across the series magnets, and 2R
m
is the equivalent reluctance
seen looking into the circuit formed by the two series magnets. From a magnetic
material point of view, the two half magnets in series is equivalent to a single block of
permanent magnet material having twice the length. Therefore,
r
remains
unchanged but R
m
doubles since reluctance is directly proportional to material
length.
The steel reluctance R
r
+R
s
in Fig. 4-3c is nonlinear because of the saturation charac-
teristic of ferromagnetic materials. Therefore, this reluctance must be eliminated in
some way to find an analytic solution. As long as the permeability of the steel is high
relative to air, the steel reluctance will be small relative to the air gap reluctance R
g
as
demonstrated in the example in Chapter 2. When this is true, the steel reluctance can
be thought of as a perturbation of the air gap reluctance. That is, the steel reluctance
can be eliminated by introducing a reluctance factor K
r
as shown in Fig. 4-3d. Here K
r
is a constant slightly greater than one that increases the air gap reluctance slightly to
accommodate or compensate for the missing steel reluctance.
It is important to note that in practice one seldom tries to determine analytical
expressions for the leakage factor K
l
and reluctance factor K
r
. It is simply too difficult
to determine accurate values given the simple modeling being performed here. Their
values are usually chosen based on the experience of the designer.
Magnetic Circuit Solution
Given the magnetic circuit in Fig. 4-3d, the magnet flux can be expressed using flux
division as (i.e., as in current division between resistors in an electrical circuit)
(4.1)
Based on
g
=K
l
, and general expressions for the magnet and air gap reluctances, i.e.,
(4.2)
the air gap flux can be written as
+
+
2
2 2
1
1
R
R K R
K
R
R
m
m r g
r
r
g
m
r
R
l
A
R
g
A
m
m
R o m
g
o g
,
Magnetic Circuit Solution 71
Brushless Permanent Magnet Motor Design, Duane Hanselman
(4.3)
where l
m
and A
m
are the magnet length and cross-sectional area respectively, and g
and A
g
are the air gap length and cross-sectional area respectively. Substituting the
flux concentration factor C
=A
m
/A
g
from (2.30), the flux density relationships B
g
=
g
/A
g
and B
r
=
r
/A
m
, and the permeance coefficient from (2.33) as P
c
=l
m
/(gC
) into (4.3)
gives an air gap flux density of
(4.4)
This equation describes the air gap flux density crossing the air gap. For the motor
being considered here with surface magnets, the leakage factor is typically in the
range 0.9K
l
<1.0, the reluctance factor is in the range 1.0<K
r
1.2, and the flux con-
centration factor is ideally 1.0. If one considers these values to be fixed and the rema-
nence B
r
to be fixed by the magnet choice, the permeance coefficient P
c
determines
the amplitude of the air gap flux density. As the permeance coefficient increases, the
air gap flux density approaches a maximum that is slightly less than the remanence.
Without flux concentration, it is not possible to achieve an air gap flux density B
g
greater than B
r
. Moreover, the relationship between permeance coefficient and air
gap flux density is nonlinear. The air gap flux density approaches the remanence
asymptotically. Doubling P
c
does not double B
g
. However, doubling P
c
means dou-
bling the magnet length, which doubles its volume and associated cost. For typical
parameter values, Fig. 4-4 demonstrates the relationship between permeance coeffi-
cient and the ratio B
g
/B
r
, where the vertical lines mark the typical four to six per-
meance coefficient range used in many motor designs.
The flux density in (4.4) defines an approximation to the air gap flux density over
the surface of the magnet pole. That is, (4.4) gives the amplitude of the air gap flux
density | B
g
| as shown in Fig. 4-5. Over North poles, (4.4) gives the positive ampli-
tude, and over South poles, (4,4) gives the negative amplitude. While this approxima-
tion is far from exact, the derivation of (4.4) provides valuable insight into motor
operation, and (4.4) itself illustrates fundamental principles that exist even when
more accurate modeling is performed.
g l
l
r
R m
m g
r
K
K
K
gA
l A
+ 1
B
K C
K
P
B
g
l
r
R
c
r
1
Chapter 4 Brushless Motor Fundamentals 72
Brushless Permanent Magnet Motor Design, Duane Hanselman
For convenience, the horizontal axis in Fig. 4-5 is described in of terms electrical
measure, which is periodic with respect to one pair of poles as shown in the figure.
For the prototype motor being considered here, there are two electrical periods
around the circumference of the rotor.
Magnetic Circuit Solution 73
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-4. Relationship between normalized air gap flux density and permeance coefficient.
Figure 4-5. Ideal air gap flux density distribution.
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
2 4 6 8 10
Permeance Coefficient, P
c
B
g
B
r
2
e
N
rotor
S
B
g
()
|B
g
|
-|B
g
|
4.5 Coil Resistance
Multiple coils connected together to form phases are a basic part of all motors. Coils
have two electrical properties, namely resistance and inductance. Of these, resistance
is straightforward to describe, whereas inductance is harder to describe especially
when mutual inductance is present as is common in motors.
Resistance is a property of all materials. It represents a measure of how much the
material resists the flow of current. For conductive materials such as copper, the
resistance R of a block of material in ohms () is given by
Chapter 4 Brushless Motor Fundamentals 90
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-24. Sum of coil back EMFs to get net winding back EMF.
e
e
a
( 12)
e
a
()
e
a
( + 24)
e
a
( 24)
e
a
( + 12)
e
e()
(4.13)
where is the resistivity of the material in m, l is the material length along the
direction of current flow, and A is the cross-sectional area through which current
flows. In general, material resistivity is a function of temperature, with resistivity
increasing exponentially with increasing temperature. For copper and aluminum
wire, resistivity versus temperature is approximated by the linear relationship
(4.14)
where T is temperature, T
o
is a base temperature, and is the thermal resistivity coef-
ficient. For copper, the parameters in this equation are commonly chosen as T
o
=20C,
(T
o
)=1.724110
8
m, and =410
3
. As material temperature increases over 100C,
this expression increasingly underestimates resistivity. Given , copper wire resis-
tance increases 4% for each 10C; therefore a wire at 120C has 40% greater resis-
tance than it does at 20C.
On the basis of the above relationships, coils in motors are most commonly com-
posed of multiple turns of round insulated wire as shown in Fig. 4-25. The center-
most circle in the figure is the bare conductor having diameter d
wb
and cross-sectional
area A
wb
. The next outer layer is the wire insulation, which is commonly available in
three thicknesses, single, double (or heavy), and triple. The final outer layer is an
optional layer of bonding material. This bonding layer is commonly composed of an
adhesive that serves to bond layers of wire together after coils are formed. As shown
in the figure, d
wc
and A
wc
are the covered wire diameter and cross-sectional area
respectively, including the bonding layer.
R
l
A
T T T T
o o
( ) ( ) + ( )
1
]
1
Coil Resistance 91
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-25. Wire cross section showing conductor, insulation, and bonding.
d
wc
d
wb
A
wb
A
wc
Several standards exist for classifying wire according to diameter. Perhaps the most
common standard is American Wire Gage (AWG). In AWG, standard wire diameters
form a geometric progression described by the relationship
(4.15)
where G is the integer wire gage, and d
wb
is the bare wire diameter in mm. There is
an inverse relationship between wire gage and diameter. As the gage increases, the
diameter decreases. Alternatively, as the gage increases, the resistance per unit length
increases. The inverse of the above relationship is
(4.16)
Because AWG is based on a geometric progression, wire gages are related to each
other by ratios as shown in Fig. 4-26. This figure plots resistance relative to a wire
having any gage G to wires having gages G, G+1, etc. The most notable point on the
curve appears at G+3. A wire of gage G+3 has twice the resistance of a wire of gage G.
So two wires of gage G+3 taken in parallel have the same resistance as one wire of
gage G. Other important points appear at G1 and G+1. At G+1, resistance is approxi-
mately 26% greater than that at G. So increasing the wire gage by one, increases resis-
tance and I
2
R losses by 26% provided current remains constant. Alternatively, at G1,
resistance decreases approximately to about 79% of that at G. Therefore, decreasing
the wire gage by one, decreases the I
2
R losses for fixed current to 79% of what they
are at G.
The current capacity of the wire depends on its cross-sectional area and its thermal
environment. The heat density in a resistance, i.e., I
2
R loss per unit volume, is equal
to J
2
, where J is the current density in the material. Based on experience, the maxi-
mum allowable current density varies roughly between 1 and 10Arms/mm
2
. Using
these limits as a guideline, Fig. 4-27 shows the allowable RMS wire current versus
wire gage. In confined volumes with little thermal conductivity, the lower limit of
1Arms/mm
2
may be too high. Similarly, if wire is actively cooled, 10Arms/mm
2
may
be overly conservative as an upper limit. It is interesting to note that the rated current
for 14 gage household wiring is 15Arms, which corresponds to a current density of
7.2Arms/mm
2
.
d
wb
G
( ) 8 24865 0 890526 . .
G
d
wb
_
,
( )
l og
.
l og .
8 24865
0 890526
Chapter 4 Brushless Motor Fundamentals 92
Brushless Permanent Magnet Motor Design, Duane Hanselman
In motor design, computing coil resistance is simply a matter of applying (4.13)
through (4.16) and taking into account the added diameter of the insulation and
optional bonding material. Since the insulation and bonding thicknesses do not fol-
Coil Resistance 93
Brushless Permanent Magnet Motor Design, Duane Hanselman
Figure 4-26. Relative wire resistance versus wire gage.
G1 G G+1 G+2 G+3
0.793
1
1.26
1.59
2
Wire Gage
R
e
s
i
s
t
a
n
c
e
R
e
l
a
t
i
v
e
t
o
G
a
g
e
G
Figure 4-27. Current capacity versus wire gage.
14 16 18 20 22 24 26 28 30
10
1
10
0
10
1
Wire Gage
A
l
l
o
w
a
b
l
e
R
M
S
C
u
r
r
e
n
t
10 A/mm
2
1 A/mm
2