You are on page 1of 7

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-20, NO.

6, NOVEMBER/DECEMBER 1984

1525

Permanent Magnet DC Motors Design Criteria and Operation Advantages


FRANK N. KLEIN, MEMBER, IEEE, AND MARK E. KENYON
Abstract-Permanent magnet dc motors have found many successful applications in industry. In particular, the machine tool industry utilizes the major share of all the permanent magnet motors produced. These motors typically exhibit greater stability of operation, linearity, higher maximum speeds, straightforward design, and higher efficiency. The applications to which these motors are addressed, however, impose fairly sophisticated design choices to maintain high performance along with optimization of unit design. In trying to optimize motor design without loss of performance, the selection of magnet material has the greatest impact upon eventual motor performance, size, and cost. The criteria for selection of permanent magnet material are reviewed, and the major advantages to using this type of motor in comparison with a wound field dc motor are specified.

Uniform, radially oriented magnetic flux T=FR

Conductor

carrying

current

INTRODUCTION THE DC MOTOR is basically a torque transducer whose most important design characteristic is expressed in a linear relationship; i.e., the shaft torque of a dc motor is directly proportional to its armature current:

Center of rotation

Fig. 1. Illustration of torque production in dc motor.

T=KO(

A dc motor's shaft torque is produced by the interaction motor. Finally, we will discuss the criteria which must be between a magnetic field and a set of conductors. This considered when selecting a dc motor for use on a incremental principle is illustrated in Fig. 1. Here, a current-carrying motion system. conductor is established in a field containing a magnetic flux FLUX SUPPLY X, and the conductor is located at a distance r from the center of rotation. When the conductor moves in the magnetic field, a Variable Magnetic Flux Motors voltage is generated across it. This voltage, which is proportional to the angular velocity of the shaft, tends to oppose the There are two basic kinds of variable magnetic flux motors: current flow. The relationship between this back-EMF voltage those in which the field winding is connected in series with the and the shaft angular velocity is armature winding and those in which the field winding is separately excited. These types of motors generally have a E= Kq5 W. (2) large electrical time constant, which makes their use inapproEquations (1) and (2) form the fundamental basis for dc motor priate for incremental motion applications, unless the servo system can accommodate such a time constant. Additionally, operation. DC motors can be divided into several broad categories variable magnetic flux motors have the following disadvanbased on the method they employ to create the magnetic field tages: and on the design and structure of the armature. We will first * requirement for another dc power supply for those motors review the two basic classes of magnetic flux supply: the having separately excited fields, and variable magnetic flux and the constant magnetic flux motors. * field winding requiring excitation which generates I2R We will then review the characteristics of the three types of losses which affect the duty cycle rating of the motor. magnetic material currently being utilized in constant magnetic flux motors: Alnico, ceramic, and rare earth. Addition- Such motors may be useful in cases where high torque servo controlled motion is demanded; however, permanent magnet motors can be applied more effectively. Paper MTIC 83-2, approved by the Machine Tools Industry of
Committee the IEEE Industry Applications Society for presentation at the Machine Tools Industry Conference Milwaukee, WI, October 18-21, 1983. Manuscript released for publication January 25, 1984. The authors are with the Application Engineering Department, Eaton Corporation, Electric Drives Division, 3122 14th Avenue, Kenosha WI 53141.

of ally, we will (1) the iron-corereview the three basic classesand armature design: the moving-coil dc, the surface-wound dc,

Constant Magnetic Flux Motors The most common field excitation system in use today is the permanent magnet dc motor shown in Fig. 2. In this system, because the field flux is essentially constant, the ratio between
1984 IEEE

0093-9994/84/1100-1525$1.00

Authorized licensed use limited to: ULAKBIM UASL - ISTANBUL TEKNIK UNIVERSITESI. Downloaded on January 21, 2009 at 08:47 from IEEE Xplore. Restrictions apply.

1526
Armature Terrminals

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-20, NO. 6, NOVEMBER/DECEMBER 1984

Several basic configurations of permanent magnet dc motors exist. The various designs aim at satisfying a wide range of performance demands, most of which deal with the torque to inertia ratio. Permanent magnet dc motors can be classified according to both commutation structure'and armature design. Commutation of the armature current in dc motors is either electrical or mechanical. Mechanical commutation structures use commutators made of copper and brushes generally made of carbon or graphite compounds containing metal particles. Electrical commutation of the armature current is achieved without the use of mechanical means (i.e., the term "brushless" is used to describe this type of dc motor)
The mechanical commutating motors can further be broken down into categories based upon armature design and field magnetic material. We will now examine the types of magnetic materials used for constant flux motors.

[1], [3].

Fig. 2. DC motor with permanent magnet field.

armature current and torque is fixed. Thus (1) becomes

T=KtI
and (2) becomes
E =Ke W.

(3)
(4)

As can be seen, this system allows for a very simple linear relationship between armature current, torque, and velocity. The electrical expression for such a motor is as follows:

Permanent Magnets The magnetic field at the air gap of constant flux motors is obtained by using permanent magnets. At present, Alnico, cermanic, and rare earth permanent magnets are being utilized by motor manufacturers. The magnetization characteristics of these materials are shown in Fig. 3-5. The typical parameters of the material constituting the magnets are
* residual induction in kilogauss (B) * demagnetizing force in oersteds (H) * energy product (B-H x 106).

Ldi V=KeW+-+RI (5) The cast permanent magnets of the Alnico type are alloys dt composed of aluminum, nickel, cobalt, titanium, and iron in where V is the applied voltage, Ke is the back-EMF voltage different proportions. The magnets of these materials, as they constant, L is the armature inductance, and R is the motor cannot be processed with conventional machine tools, are cast to the rough shape and further ground to the final configuraterminal resistance. The dynamic equation of the motor is tion. They exhibit nonlinear B-H characteristics, high magJdw netic induction valves, and very low demagnitizing force Tg= + Bw+Tf + TL (6) valves [2]. dt Cermanic permanent magnets are obtained through highwhere Tg is the generated torque, J is the sum of the moments temperature synthesization and, being very hard and fragile, of inertia for the motor and the load, B is the viscous damping also cannot be processed with conventional machine tools. coefficient, Tf is the internal friction torque, and TL is the load They normally consist of alloys of iron oxide or barium oxide. torque. This same equation could be written for a separately They exhibit improved linear B-H characteristics at the price excited field motor with constant field excitation. However, of lower magnetic induction valves compared to Alnico. The the advantage of the permanent magnet field over the wound energy product of ceramic magnets is slightly less than Alnico field structure is that no power is dissipated in the field. The but, due to their higher demagnetization force, do not permanent magnet dc motor is, therefore, more efficient and demagnetize when the flux shunt is removed. Permanent magnets made of rare earth (somarium cobalt) requires less space than an equivalent motor with a wound field. Also, due to the availability of permanent magnets with have the most linear B-H characteristics; however, they high coercive force characteristics, the linear relationship of exhibit low magnetic induction valves compared to Alnico (1) holds for higher armature currents in permanent magnet dc magnets. Rare earth magnets have very high energy products but on the average cost 20 times more than the Alnico or motors than in their wound field dc motor counterparts. The simple mathematical relationship which exists between ceramic varietes for comparable performance. However, they input parameters and mechanical output characterisitcs in the do offer one advantage to the motor builder in that the permanent magnet dc motor makes it ideal for use in magnetic material can be magnetized prior to assembly. incremental motion servo systems. The use of this type of Another characteristic used to classify permanent magnet motor can simplify both the servo calculations and the system motors is armature design. There are currently three types of armature designs for mechanical commutation permanent design.

Authorized licensed use limited to: ULAKBIM UASL - ISTANBUL TEKNIK UNIVERSITESI. Downloaded on January 21, 2009 at 08:47 from IEEE Xplore. Restrictions apply.

KLEIN AND KENYON: PERMANENT MAGNET DC MOTORS DESIGN CRITERIA


Energy Product -

1527
4

BdHdXlo6
cn
n

S
co

1-

1000 2000 Demagnetizing Force H-Oersteds

Fig. 3. Ceramic magnet material properties.


Energy Product
-

BdHdX1o6

a) 0
cr
o
IC

z
7

5 3

Demagnetizing Force u-Oersteds

Fig. 4. Alnico magnet material properties.


Energy Product-BdHdXlo6

7 II I I I

I -I 2016121 I '8

10

Magnet
Rotor
CD
0

to
._

CD
0

1z

to

.16

12

14000

12000

10000

8000

6000

4000

2000

Fig. 6. Cross sectional view of permnanent magnet iron-core dc motor.

Demagnetizing Force H-Oersteds

Fig. 5.

Rare

earth magnet meterial properties.

magnet motors. These designs are iron-core, surface-wound,

and moving-coil motors.

ARMATURE DESIGNS Iron-Core Motors The rotor and stator configuration of an iron-core motor is shown in Fig. 6. The magnet structure in this example is

shown as a segment of magnet material-in this case a ceramic magnet (Alnico or rare earth magnets could also have been used). The magnetic flux passes through a laminated rotor structure. The rotor contains slots, into which wire is placed. (The wire is mechanically supported in such a rotor structure since it can be impregnated with varnish or resins and mechanically bonded to the iron core.) The magnetic flux, therefore, must pass through a smali air gap and over a low-reluctance path to reach the opposite pole. The iron-core structure has the highest moment of inertia and

Authorized licensed use limited to: ULAKBIM UASL - ISTANBUL TEKNIK UNIVERSITESI. Downloaded on January 21, 2009 at 08:47 from IEEE Xplore. Restrictions apply.

1528

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-20, NO. 6, NOVEMBER/DECEMBER 1984


Conductors (bonded together by non-magnetic materials)
Stationary flux Return Path

Fig. 7. Cross sectional view of surface-wound permanent magnet dc motor.

Fig. 8. Cross section view of moving-coil permanent magnet dc motor.

the largest rotor inductance of all permanent magnet motors currently being marketed. In addition, the rotor may have preferred "detent" positions due to unequal reluctance paths in the magnetic structure. However, skewing of the rotor slots can minimize or even eliminate this tendency. The iron-core motor is by far the most widely used type of dc motor for incremental motion applications since the iron-core rotor structure has a high thermal capacity (thermal time constants ranging from 45 min to 4 h) and can take overloads for extended periods of time without damage. Surface- Wound Motors Fig. 7 shows the rotor design of a surface-wound motor. The conductors are bonded to the surface of a cylindrical rotor structure, which is made of laminated disks fastened to the motor shaft. Since no slots are used in this design, the armature has no tendency to assume preferred detent positions. Since the conductors exist in the air gap between the rotor disks and the permanent magnetic field, this design has lower inductance than the iron-core structure. However, because the air gap between the magnet and the lowinductance rotor is larger than in the iron core motor, a longer magnet, regardless of the type of material selected, is required in order to provide a magnetic flux equivalent to that of the iron-core motor. Surface-wound motors are, therefore, more expensive to produce and have larger outside diameters than equivalent iron-core motors [1], [3].
Moving-Coil Motors Moving-coil motors are designed to have very small moments of inertia. This is achieved through rotor design: the rotors consist of conductors placed in the air gap between a stationary flux return path and the magnet structure, as shown in Fig. 8. In this example, the conductor structure is supported by nonmagnetic materials, usually epoxy resins and fiberglass, to form a hollow cylinder. One end of the cylinder forms a hub which is attached to the motor shaft. A cross section view of such a motor is shown in Fig. 9. Since all unnecessary elements have been removed from the armature of a moving-coil motor, its moment of inertia is very

low. However, it has a larger magnetic air gap than the two models discussed earlier and, therefore, requires an even larger magnetic structure than the previous examples to produce an equivalent air gap magnetic flux. Use of Alnico magnetics could reduce the size of the magnetic structure, due to their very high induction valves, but the accompanying reduction in demagnetizing force would be unsatisfactory. Hence the price one pays for the high torque to inertia ratio in such a motor is a rather large magnet structure. In addition, the rotor structure has a very low thermal capacity; if overloaded, it can easily be damaged. When moving-coil motors are used in high-performance incremental motion applications, it is common to cool the rotor structure by means of forced-air cooling. The moving-coil armature circuit, because the conductors are not in direct contact with iron, has a very low inductance: values of less than 100 A.H are common. Another version of the moving-coil motor is shown in Fig. 10. This type of motor has a disk-shaped armature structure which is made up of a multitude of conductors, sometimes in several layers, forming a conductor array. The conductor array is utilized as a commutation surface. This motor type has in the past been called a "printed motor" in reference to early models in which so-called "printed circuits" were first used to form the conductor arrays. Today, this type of motor is made from stamped and welded copper conductor paths. This motor has found many uses, especially in applications where short axial length is a critical factor. The disk armature in this type of motor generally has a higher moment of inertia than the cylindrical moving coil armature for two reasons: the end turns are dominant inertia-producing elements and the moment of inertia of a disk increases by the fourth power of its diameter [1], [4]. Up to this point we have examined some basic operating principles of dc motors as well as typical characteristics of basic dc motor models. As we have seen, each of the various motor designs has unique advantages and disadvantages. When a machine tool designer is faced with the, problem of selecting the best motor for a given application, he must first define his specific needs in terms of system requirements. With this knowledge, he can then examine the various options of available motors with a view toward the necessary

Authorized licensed use limited to: ULAKBIM UASL - ISTANBUL TEKNIK UNIVERSITESI. Downloaded on January 21, 2009 at 08:47 from IEEE Xplore. Restrictions apply.

KLEIN AND KENYON: PERMANENT MAGNET DC MOTORS DESIGN CRITERIA


Magnet Pole

1529

Flux Path

Fig. 9. Cross sectional side view of moving-coil dc motor.

\-

Permanent Magnets

Fig. 10. Exploded view of disk type moving-coil dc motor.

controller,

power

ments. We will now examine

the demands of incremental motion applications.

of the primary performrequired of the dc motor in order to meet ance characteristics


some

supply, and feedback transducer

require-

PERFORMANCE CHARACTERISTICS Motor Efficiency Systems incorporating incremental motion often involve the production of high peak torques with little useful work done; that is, they involve acceleration and deceleration of inertial loads with small frictional demands. The motor, therefore, absorbs energy during both acceleration and deceleration, and the primary heat-producing element in the motor is armature resistance. If high shaft velocities are involved, we may have significant contributions to motor heating by viscous damping effects as well. It is important to note that high motor efficiency is not the only criterion by which a dc motor for incremental motion use should be judged. Efficiency must be considered together with motor moment of inertia, motor inductance, and peak current capability in order to arrive at optimum systems efficiencies and performance capabilities.
Armature Inertia Since the objective in incremental motion systems is the rapid starting and stopping of a rotating load, the motors contribution to the system's inertia should be optimized as

nearly as practicable in order to fulfill the system specifications and requirements. The moving-coil motor discussed previously offers the lowest inertia of all dc motors. If adequate cooling means are available, this motor design may be a good choice. However, the very low thermal capacity of the rotor structure demands a thorough knowledge of the worst-case load conditions since the moving coil armature (which can have a thermal time constant as low as 30 s) can be damaged by even short-term overload conditions. The iron core structure, on the other hand, can in many cases sustain prolonged overload without damage. Some motors have thermal time constants as high as 4 h.- Iron-core motor designs using high energy product magnets such as Alnico or rare earth types can achieve lower moments of inertia than motors using ceramic magnets. However, the limiting factors of using these types of magnets exclusively is the high cost of rare earth material and the low resistance to demagnetization of Alnico magnets. Armature Inductance In order to achieve a rapid motor response, the armature inductance of the motor must be low. The electrical time constant of the motor can then permit the required current rise time without interferring with servo operational requirements. However, sometimes a low armature inductance is needed to fulfill servo response demands at the same time that a relatively high inductance valve is needed in order to use pulse width modulation (PWM) techniques for power control. In

Authorized licensed use limited to: ULAKBIM UASL - ISTANBUL TEKNIK UNIVERSITESI. Downloaded on January 21, 2009 at 08:47 from IEEE Xplore. Restrictions apply.

1 530

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. IA-20, NO. 6, NOVEMBER/DECEMBER 1984

such cases, external inductors must be added on to low inductance motors, such as the moving-coil design, in order to keep the PWM ripple current at an adequate level.
Linear Current to Torque Relationships As discussed previously, the linear relationship which exists between armature current and output torque is a very important characteristic of dc motors when considering their suitability for use in incremental motion servo systems. Another way of stating this is that the torque constant Kt should not vary with the current level or angle or rotation of the rotor. Conditions which can cause the torque constant to vary are magnet circuit saturation and/or poor armature design. The former results in low torque constant at high torque levels, and the latter causes a cyclically varying torque with rotation due to insufficient number of commutation points.

Since individual motor designs have varying degrees of sensitivity to the combination of peak currents and temperature, it is important to examine the motor specifications in order to determine whether current-caused demagnitization could create a potential problem in the motor under consideration.

Viscous Damping Coefficient The viscous damping coefficient consists of several components, some of which are mechanical in nature; others are caused by electromagnetic actions in the structure. The presence of viscous damping can be inferred from the observation of increased no-load current at increased speeds. The mechanical component of the viscous damping is partly due to friction caused by bearing lubrication components and bearing seals. The contribution to viscous friction by these elements usually is one that increases linearly with velocity. Another mechanical effect which may manifest itself at high Reluctance Torque Disturbances shaft velocity is the "windage" element due to rapidly rotating The magnetic design of dc motors for incremental motion parts in the motor. applications should be such that the armature does not have The electromagnetic components of the viscous damping preferred shaft angle positions (i.e., that the rotor does not effect consist of three principal elements: "cog" when rotated by external means while unenergized). The tendency to cog is the result of a design in which the rotor 1) circulating current caused from currents flowing in structure provides a lower reluctance to the magnetic field in commutated coils in the presence of stray fields in the some positions than in others. Thus, if it is necessary to stop stator; and hold a rotor shaft in a given angular position, a rotor 2) eddy-currents caused from armature conductors moving which cogs might require a constant current to hold the shaft in in the magnetic field (this effect generally occurs at that specific position-if that position also happens to be higher speeds and tends to increase with the square of located at a point where reluctance torque is present. motor velocity); 3) hysteresis caused by the change in magnetic flux in the The Motor Torque Constant Versus Temperature rotor due to the rotation in the permanent magnet field. If the motor is to be used in an application where significant motor heating is likely to occur, the temperature coefficient of The comments pertaining to eddy currents and hysteresis were the magnetic field must be examined. This is necessary to made assuming that the armature current remained relatively determine whether the torque constant could be significantly constant. The heating effects from eddy currents and hysterealtered at extreme temperatures. As it turns out, the lowest sis can be greatly accentuated if the motor is operated in a cost permanent magnet materials have a higher temperature pulsewidth modulation or pulse frequency, modulation with a coefficient of magnetic flux than others: barium-ferrite type significant ac armature current component in addition to the dc ceramic magnets, for example, have a temperature coefficient current. If magnetic material exhibiting low resistivity, such as of - 0.2 percent/C, whereas Alnico V has a temperature rare earth compounds, are used in this environment of ac coefficient of -0.02 percent/C, and Alnico VIII has a current components, eddy-current heating may also occur in temperature coefficient of less than -0.01 percent/C. The the magnetic material causing heat. It is therefore important to newest of the magnet materials, the rare earth compounds, pay particularly close attention to this phenomenon when the have a tendency to experience long-term irreversible demag- motor is used with PWM controls. netizing effects if operated at moderately high tempertures. Depending on specific compounds, they may show aging Torsional Resonance Conditions effects if operated above temperatures ranging from 150C to Since incremental motion systems normally require some 200C. type of speed or position feedback transducers, it is important to examine the system for the presence of torsional resonance Motor Peak Current Capability frequencies. Those resonant frequencies could cause oscillaDepending on the type of magnet used and on the air gap tions in a closed-loop servo system. Several steps can be taken characteristics of the motor, motor magnets may demagnitize to reduce or eliminate these resonant frequencies. if the current exceeds a certain value. This is generally due to * Motor shaft stiffness should be high. the intrinsic coercive force characteristics of magnet materials. * Any feedback transducers should be directly connected to A common trend is that all permanent magnets have a lower the motor shaft. resistance to demagnitization at high operating temperatures.

Authorized licensed use limited to: ULAKBIM UASL - ISTANBUL TEKNIK UNIVERSITESI. Downloaded on January 21, 2009 at 08:47 from IEEE Xplore. Restrictions apply.

KLEIN AND KENYON: PERMANENT MAGNET DC MOTORS DESIGN CRITERIA

* Load coupling to motor should be by means of a coupling which does not offer significant torsional resonance problems [1], [3]. SUMMARY The correct choice of a motor for use in an incremental motion servo system is the single most important step towards obtaining the most optimized system. We have examined the merits of the dc motor in fulfilling this function. We have considered the dc motors basic operating characteristics as well as selection criteria. Armed with this understanding of the capabilities of various dc motor types, the machine tool system designer can more effectively meet the design specifications of his machine's incremental motion requirements.

Frank N. Klein (S'56-M'58) received the B.E.E. degree from Marquette University, Milwaukee, ~4WI, in 1958. He was associated with the American Motors Corporation and later with the Dynamatic Division l _ of Eaton Corporation in a cooperative workstudy program. He then joined Eaton Corporation, Dynamatic Division, as a Research Engineer in electrical drives systems. He has worked over 20 years in electrical drive development in eddy current, dc, and adjustable frequency (AF). He holds five U.S. patents relating to AF drives. In 1980 he became Manager of Drive Applications, responsible for eddy current, dc, and AF drive applications. Mr. Klein held several elected offices in the local Racine-Kenosha IEEE subsection of the Milwaukee Section of the IEEE, including Chairman. He has also served in several IEEE Society groups, including Magnetics and Industry Applications as an elected officer in the Milwaukee Section. He presently is a member of the IAS Industrial Drives Committee. He is a Registered Professional Engineer in the State of Wisconsin.

REFERENCES
[1] C. Kuo and J. Tal, DC Motors and Control Systems. Champaign, IL: S.R.L. Publishing, Nov. 1978, pp. 81-96. [2] F. Dugan, "A state-of-the art cobalt rare earth, permanent magnet motor for the machine tool industry," presented at the IEEE-IAS Machine Tools Industry Conf., Oct. 1975. [3] DC Motors, Speed Controls, Servo Systems, 5th ed., Electro-Craft Corp., Aug. 1980, pp. 2-11-2-64. [4] W. Smeaton, Motor Application and Maintenance Handbook. New York: McGraw-Hill, 1969, pp. 9-36.

Mark E. Kenyon received the B.S. degree in _ chemistry from Marquette University, Milwaukee, WI, in 1976 and has since studied business management at the University of Wisconsin-Parkside, Kenosha, WI. He is an Industry Application Engineer with Eaton Corporation, Electric Drives Division, where he is responsible for marketing and application engineering aspects of the Division's variable speed drives which are utilized in the metal-cutting and' metal-forming industries.

You might also like