Direct-Current Motors: Theory of DC Motors

This section covers methods of calculating performance for direct-current (dc) mechanically commutated motors. Section 1 (below) discusses the electromagnetic circuit for dc motors and series ac motors. Sections 2 and 3 establish some common geometry and symbols and discuss commutation for dc motors. Section 4 presents permanent-magnet direct-current (PMDC) calculation methods. Section 5 presents series dc and universal ac/dc performance calculations. Sections 6 and 7 discuss methods for calculating the performance of shunt- and compound-connected dc motors. Finally, Secs. 8 and 9 discuss dc motor windings and automatic armature winding.

THEORY OF DC MOTORS

DC Series Motors

A series motor operating on direct current has characteristics similar to those when it’s operated on ac current at power-system frequencies. However, it’s best to describe dc and ac operation separately so that comparisons can be made.

The general equivalent electrical circuit of the series dc motor and its physical construction is shown in Figs. 1 and 2. The motor consists of a stator having a concentrated field winding ( FIG. 3) connected in series by way of a commutator to a wound armature ( FIG. 4). One of the first things to be considered in the operation of the motor are the motor and generator action, which exist simultaneously in the armature circuit of the motor. These two principles are (1) the instantaneous electromotive force (emf), which is induced in the armature conductors when moving with a velocity v within a magnetic field, and (2) the force produced on the conductors as the result of carrying an electric current in this same magnetic field.

FIG. 1 Series motor diagram: ra =armature resistance measured at brushes, rf =main field resistance, and v = applied voltage.

FIG. 2 Series motor.

FIG. 3 One pole of a series motor field winding.

FIG. 4 Wound armature.

It’s known that the instantaneous force on a conductor of length carrying a cur-rent i in a magnetic field B is:

Or, in vector notation:

… where θ is the angle between the direction of the magnetic field and the direction of current flow in the conductor, B is in webers per square meter (teslas), i is in amperes, and is in meters. Motors, by design, have the armature conductors and magnetic flux at quadrature to one another. Therefore, the force becomes:

Assume a situation in FIG. 5 where a conductor of length l is located in a magnetic field and is free to move in the x direction perpendicular to the field. From the preceding discussion, a force is produced on the conductor, causing it to move in the x direction. Then:

So, differential electrical energy into the conductor less differential i^2 r loss equals differential mechanical output energy:

FIG. 5 Conductor moving in a magnetic field.

This force causes movement in the conductor, which in turn causes a volt-age to be induced into that conductor which is opposite to the direction of the original current. This is an important concept in the operation of motors in general and one which is used to discuss the operation of the series motor. This induced voltage is usually called a counter-emf (c_emf) voltage because of its opposition to the applied voltage.

This example also indicates that a reversible energy or power exchange is possible (i.e., between the mechanical and electrical systems). Therefore, the same machine may operate as a motor or a generator, depending on the flow of energy in the armature.

In the motor mode, the field and armature of a dc series motor are supplied with the same current by an applied voltage, and a magnetic field (flux) is produced in the magnetic circuit. Since the armature conductors (coils) are located in this field, each of the conductors in the field experiences a force (torque) tending to make it move (rotate), and as we have just indicated, a countervoltage (c_emf) is produced opposing the applied voltage. Other than the c_emf which the armature produces, we must recognize that the armature circuit (coils) also produces a magnetic field of its own.

The armature, because of its commutator and brush construction, has a unidirectional current and therefore produces a fixed-direction magnetomotive force (mmf), measured in ampere-turns. This mmf is the product of the effective coil turns on the armature and the current through those turns. It must be understood here that the armature winding must be considered in developing those ampere-turns (i.e., there may be parallel paths through the armature, with the possibility of the armature coils being wound in series or parallel arrangements; hence, there will also be a division of the total armature current in each of the windings). This topic is addressed later. The following is a discussion of the action of the mmf produced by the armature.

Armature Reaction. The conductors in the armature carry a current proportional to the load. The magnetic field produced by this current reacts with the main field produced by the same current flowing in the field coils. FIG. 6 indicates two so-called belts of armature conductors (coil sides) under each pole face. Each of the conductors comprising these belts carries current in the same direction and hence produces additive mmf. In addition, there are conductors which also carry unidirectional currents but are not under the pole arcs. The important consideration here is the effect of the presence of magnetic material in the pole pieces, armature core, and armature teeth. The flux paths through the armature are influenced by the reluctances of the paths. It’s obvious that the reluctance of the flux paths under the pole pieces is less than that of the paths adjacent to the brush area, which constitute a material of much greater reluctance, namely air.

FIG. 6 Armature magnetic field.

FIG. 6 indicates that the brushes are on the mechanical neutral axis (i.e., halfway between the poles). The general direction of the armature mmf is along the brush axis at quadrature to the main field. The armature conductors adjacent to the poles produce flux densities in the air gap which are equal and opposite at the pole tips. Keep in mind that the flux density produced by the armature is directly related to the armature current. (Also, a uniformly distributed flux density in the air gap is attributed to the main field and is directly related to this same armature current.) Now the net mmf (or flux in the air gap) is the result of both the main field mmf and the armature field mmf. The resultant air gap flux is now increased at one pole tip and reduced at the other pole tip because of the armature reaction. The flux distortion in the air gap is illustrated in FIG. 7. In this figure, two poles and the armature conductors beneath the poles have been unfolded to illustrate the distortion more clearly. Ampere's circuital law is useful here in determining the armature mmf. The resulting air gap mmf is the result of the superposition of the field and armature mmfs. MMF drops in the poles and armature iron are considered negligible com-pared to the air gap mmf. For reference, positive direction for the mmfs is assumed to be a flux out of a north pole. The armature mmf is shown as a linear relationship, but actually, because the armature slots are discrete, this relationship is actually made up of small discrete stair-step transitions. However, it’s shown as a smooth curve here for easier analysis.

FIG. 7 Flux distortion in the air gap.

FIG. 8 Fundamental frequency of air gap flux.

FIG. 9 Air gap flux due to fundamental plus third and fifth harmonics.

It’s obvious that the air gap flux varies along the pole face. Another observation from the distortion mmf pattern is that harmonics are very present in the air gap mmf. Also, because of symmetry, only odd harmonics can exist in an analysis of the air gap mmf. It must also be remembered that in the case of a series ac motor, the variation of flux distortion would be approximately the same, only pulsating.

The results of a Fourier analysis of the air gap flux harmonics for the case when:

(NfIf)/(NaIa) =7/5 are given in Ills..8 to 10. The figures show the fundamental, the fundamental plus third and fifth harmonics, and finally the fundamental and odd harmonics, up to and including the fifteenth harmonic. The magnitudes of the odd harmonics beginning with the fundamental are 8.94, 0.467, 1.41, 0.924, .0.028, .0.02, 0.523, and 0.0094. That is, the series can be represented as follows:

Of course, changing the pole arc or magnitudes of the armature and field mmfs will change the harmonic content. The change in flux density across the air gap produces two effects: (1) a reduction in the total flux emanating from each pole, and (2) a shift in the electrical neutral axis, lending to commutation problems due to the flux distortion.

The flux distortion which the armature produces has been called a cross-magnetizing armature reaction, and rightly so. The net effect is illustrated in FIG. 11, showing the resultant field where the armature cross field is at right angles to the main field. The result is a distortion of the net flux in the motor; the second effect, which was indicated previously, is a reduction in the total main field flux. This reduction of flux is not too obvious; in fact, it would almost appear from FIG. 8 that the vector addition of these two mmfs would lead to an increase in flux which would occur with the brushes on the mechanical neutral axis.

FIG. 10 Air gap flux due to odd harmonics through fifteenth.

FIG. 11 Field of armature with and without brush shift. This resulting decrease in flux can be attributed to magnetic saturation. FIG. 12 illustrates how this comes about when one operates the motor at the knee of the saturation curve. The area ABC is proportional to the reduction in flux at the pole tip with a decrease in flux density, while the area CDE is proportional to the increase in flux at the other pole tip. Note that the reduction in flux exceeds the increase in flux because of the saturation effect. This is sometimes called the demagnetizing effect of armature reaction. This demagnetization effect can be on the order of about 4 percent.

Both armature reaction distortion and flux can be reduced or eliminated by compensation windings in the pole face or by increasing the reluctance at the pole tips; the latter by either lamination design or a non-uniform air gap.

FIG. 12 Armature reaction resulting in main pole flux reduction.

Because the purpose of the commutator and brushes is to change the current in a short-circuited coil from, say, a current of +I to -I in the length of time the coil is short-circuited by the brushes, some arcing at the brushes is expected. One desires to keep any voltage induced into the coil to a minimum in order to keep this arcing as small as possible.

When the brushes are on the mechanical neutral position and the main field becomes distorted, the coils being commutated will have an induced voltage from the distorted air gap flux. A shifting of the brushes to a new neutral position is suggested to keep the induced voltage to a minimum and keep the brush arcing low. It should be recognized that the amount of distortion of the field is a function of the ampere-turns of the armature and hence is dependent on the motor load. One then realizes that the electrical neutral position is a function of the load and that shifting brushes to an electrically neutral position is therefore not a matter of shifting them to a unique point in space.

Brush shifting causes another effect-that is, a demagnetization effect on the air gap flux. This effect is in addition to the demagnetizing effect which occurs because of saturation. When the brushes are shifted, the pole axis of the armature is shifted.

The result is that the angle between the main field and the armature field is greater than 90. This process is also illustrated in FIG. 11.

In summary, there are two processes which cause reduction of the air gap flux: (1) reduction due to cross-magnetization when the brushes are on the mechanical neutral position, which changes the flux distribution across the pole face and the net flux because of saturation effects, and (2) demagnetization resulting from a brush shift and change of the armature pole orientation with respect to the main field, resulting in the armature field mmf having a component in direct opposition to the main field mmf.

Reactance Voltage and Commutation. It was indicated previously that during commutation the current in the shorted coil(s) under a brush must reverse and change direction. The self-inductance of the shorted winding by Lenz's law induces an emf in the shorted winding to oppose the change in coil current. This voltage is sometimes called the reactance voltage. (Actually, it may be only a portion of the reactance voltage, as is shown later.) This voltage slows down the reversal of current and tends to produce sparks or arcing as the trailing commutator bar leaves a brush.

The reactance voltage hinders good commutation. The magnitude of this voltage depends directly on the square of the number of coil turns, the current flowing, and the armature velocity; it’s inversely proportional to the reluctance of the magnetic path.

FIG. 13 Linear commutation.

When shifting brushes to seek an electrical neutral position on a motor, the shifting is done in the opposite direction of armature rotation. It must be remembered that the reactance voltage is an e =L(di/dt) voltage and is simply due to the current change during commutation. It has nothing to do with induced voltage from air gap flux. Shifting the brushes really does not help the reactance voltage. Theoretically, there is no induced voltage on the neutral axis from the motor flux (in fact, d lambda /dt should be zero here since the coil sides are moving parallel to the flux). In order to counteract the reactance voltage and induce a voltage opposite to the reactance voltage, the brushes must be shifted further backward than the magnetic neutral axis. At this point a d lambda /dt voltage is induced into the commutated coil by the field flux, which counteracts the reactance voltage. The result is that two voltages in opposite polarities are induced into the shorted coil, thereby reducing brush arcing.

An important consideration is that there can be another component of the reactance voltage which can occur when two coil sides are in the same slot and both are undergoing commutation. There is then the following self-inductance term:

Reactance voltage = L_11 (di1 / dt) + M_12 ((di2 / dt) )

…where the coil being considered is the coil carrying current i1 and the coil in the same slot undergoing commutation is carrying current i2. If there is more coupling between coils, more mutual terms can exist.

Brushes are a very important consideration, and contrary to normal electrical principles, one would assume that keeping the brush resistance low would assist in reducing arcing at the brush commutator bar interface. This is far from the truth. In fact, resistance commutation is now an accepted technology. Brushes normally have a graphite or carbon formulation and hence introduce, by their characteristics, resistance into the interface. If constant current density could be achieved at a brush for all loads and speeds, an ideal condition would exist for commutation. FIG. 13 illustrates what would occur if ideal conditions are assumed. The assumptions are as follows:

The brush width is equal to that of a commutator segment.

The current density under a brush is constant.

The reactance voltage and resistance are zero.

The commutator bar insulation is small compared to the width of a commutator segment.

Note that as the commutator passes beneath the brush, for the assumptions given, the contact resistance is inversely proportional to the contact area. The sequence in FIG. 13 illustrates how the current change occurs in an armature coil under these ideal conditions. This change in surface contact changes the brush con-tact resistance; hence, with constant current density, the change in current flow changes linearly according to contact resistance. Remember, the resistance is inversely proportional to the contact area. This ideal commutation has been termed linear commutation because of the linear transfer of current in the commutated coil, as indicated in FIG. 13.

If the brush width is greater than that of one commutator segment, linear com-mutation still exists because the resistance still changes linearly; hence, the current density remains constant. The commutation period is longer; hence, the rate of change of current di/dt is reduced, and hence the reactance voltage is reduced also.

All of the assumptions made in the previous analysis cannot occur, and nonlinear commutation results. The armature coils have resistance, have self- and mutual inductance, and hence linear commutation cannot be attained.

If these nonlinearities do occur, then either undercommutation or overcommutation occurs. These are illustrated in Ill. 14. Since brush heating is a product of brush resistance and the square of the current, linear commutation is preferred, as it minimizes brush heating. That is, the squared value of the three forms of current commutation in Ill. 13 is smallest for linear commutation.

Ill. 14 Overcommutation, undercommutation, and linear commutation.

It’s possible to develop a circuit model for the commutation process, as shown in FIG. 15. The factors have previously been described, and they are restated here.

Consider the subject coil of turns N1, and self-inductance L11, and the current i1 which is being commutated.

1. The voltage of self inductance (reactance voltage) of the commutated coil is

eL1 =L11 (di1/dt)

... where i1 is the current being commutated and L11 is the self-inductance of the coil being commutated.

2. The voltage of mutual inductance of an adjacent coil in the same slot carrying current i2 and undergoing commutation and linked to the subject coil is

eM1 =M12 di2 / dt

… where M12 is the mutual inductance between adjacent coils.

3. The voltage induced into the subject coil of turns N1 when cutting flux φ a due to the armature mmf is:

Sometimes this is called commutating voltage; it opposes the reactance voltage. This can be controlled by brush overshifting to make ea cancel the voltages eM1 and eL1.

4. There are voltage drops due to coil resistance rc; Rc and Rc , the left and right contact resistances at the brush; and the brush resistance between commutator segments rb.

The commutation process is not simple, and brush composition and interface film, measurements, and conditions all can be made to assist in the commutation process. Maintaining a suitable interface film between the brush and the copper commutator is extremely important. This interface is formed of copper oxide and free particles of graphite film; it provides a general resistance commutation and sup-plies lubricant to reduce surface friction and heat between the commutator and the brush.

Torque-Speed Characteristics of DC Series Motors. The electrical equivalent steady-state circuit is the most appropriate method for analyzing the motor. This equivalent circuit is repeated in FIG. 16 for convenience.

FIG. 15 Commutation voltages.

FIG. 16 Electric circuit of series dc motor.

Writing Kirchhoff's voltage law around the loop gives:

V =Ia(rf +ra) +Ecemf

=Ia(rf +ra) +Eg

The counter-emf Eg was addressed earlier when discussing the voltage induced into a conductor of length and moving with a velocity v perpendicular to a magnetic field of flux density B. The counter-emf voltage Eg was expressed as e and given by:

e =Blv

Efficiency of DC Series Motors. Earlier discussion has indicated that a number of losses occur in the series motor. These losses can be summarized as follows:

1. Copper loss in the armature winding I^2 ara

2. Copper loss in the series field winding I^2 arf

3. Brush contact loss

4. Friction (brush and bearing friction) and windage

5. Core loss (hysteresis and eddy current)

6. Stray load loss (losses in addition to those above)

A comment should be made on the brush contact loss, listed as number 3. Experimentation and Institute of Electrical and Electronic Engineers (IEEE) specifications have suggested as an approximation that Brush contact loss = brush voltage drop × armature brush current (4.29) This expression was arrived at for carbon or graphite brushes by observing the brush voltage drop as a function of current density. At high- and low-current steady-state densities, the voltage falls off and approaches zero. However, in between these limits the voltage drop is approximately constant at about 1 V per brush. Therefore, for a pair of brushes this becomes 2 V. On small machines this voltage drop tends to increase, since the tendency is for the voltage drop to increase for lower brush and commutator temperatures.

Carbon and graphite materials have a resistivity many times that of copper and have a negative temperature coefficient. This negative temperature coefficient is attributed to the rise in brush voltage drop in small machines.

Stray load loss, as the term suggests, is a function of motor load and changes in load. Changes in load produce changes in armature current and hence affect (1) magnetic saturation in the magnetic circuit, (2) armature reaction changes, and (3) eddy current loss changes.

Figure 17 presents a diagram displaying these losses. By definition:

Overall efficiency == output / input = output / (output + losses)

= useful mechanical output/ total electrical input

Sometimes both electrical and mechanical efficiencies are of interest in order to determine where improvements can be made.

Mechanical efficiency = useful mechanical output / mechanical output + rotational losses

Electrical efficiency = electrical power output / electrical power output + electrical losses

FIG. 17 Losses in a series dc motor.

The conditions for maximum efficiency can be related to those losses which are considered to be constant and those that vary with the motor load current. If the losses are segregated as follows:

K1 = constant losses

K2 = losses which vary linearly with Ia

K3 = losses which vary as the square of Ia

Then the efficiency is:

eff = input - losses / input

Equating this expansion to zero gives the condition for maximum efficiency.

Thus, for maximum efficiency the constant losses must be equal to those that vary as the square of the armature current. This is typical for all different pieces of rotational electrical equipment. The constant losses usually are considered to be the core losses, friction, and windage. Usually the brush loss is small.

AC Series Motors

One of the first considerations when considering ac operation of a series-wound motor is: Does the motor develop a unidirectional torque when operated on ac supply voltages? We know the answer is yes. Since the armature and field are connected in series, a current reversal in the field also produces a current reversal in the armature. The torque is therefore in the same direction. As indicated previously, the construction must be an ac type with a completely laminated magnetic path. If it were not laminated it would react as a solid-core inductor, with extremely high eddy cur rent losses.

The basic phasor and electrical circuit diagram of a series ac motor is shown in FIG. 18. The common current in both armature and field produce a motor flux which is nearly in phase with the current, the small difference being the hysteresis and eddy current effects, which are not accounted for in the phasor diagram. Seven voltages are required to overcome the applied voltage. These are the voltages produced by the leakage reactances and resistances of both armature and field, the transformer voltage ETf, the commutator brush drop, and the counter-emf voltage Eg. When the armature is rotating, the armature conductors, which are moving through an alternating field flux, generate an alternating voltage Eg which is in phase with the flux, as indicated on the diagram.

FIG. 18 Series ac motor: (a) equivalent circuit and (b) phasor diagram.

This voltage is proportional to the magnitude of the flux and the rate at which it’s cutting the flux (Faraday's law). Since the flux is a sinusoidally varying quantity, the voltage will also vary sinusoidally at the same frequency as the flux wave. The transformer voltage ETf has been included and reflects the Faraday voltage induced into the field by the core flux. A similar voltage would be present in the armature, but it’s small in comparison to ETf and has not been included in the phasor diagram.

The phasor sum of these voltages then represents the applied voltage V and the basic phasor diagram of the ac series motor. Here all voltages are root mean square (RMS) quantities. As previously mentioned, the effects of hysteresis and eddy cur rents have not been taken into account, nor have the effects of commutation and brush shifting. The power factor angle ? is also indicated on the diagram and represents the power factor at the input motor terminals.

Several things are apparent from the phasor diagram:

1. At starting, the generated voltage Eg is zero. Hence, the starting current is limited only by the impedances of the armature and field and the transformer volt age ETf.

2. The motor draws a lagging power factor cos ?.

3. As the motor speed increases at a constant line voltage, the motor current decreases and the power factor improves, which is a good reason to run a motor at high armature velocities.

4. A reduction in frequency also improves the power factor by reducing the reactance drops in both armature and field.

At this point it should be obvious that dc and ac performance of series motors must differ due to reactance and transformer voltages and also due to ac losses (additional hysteresis and eddy current losses-that is, core losses-and possible changes in ac resistance as compared to dc resistance of the armature and field).

Note that in the development of the phasor diagram leakage reactance voltages for both the field and armature, as well as a transformer voltage in the main field, have been included. It’s appropriate to develop some mathematical formulations and basis for these voltages. In reality, the leakage reactance and transformer voltage result from an application of Faraday's law and are broken down into two parts to obtain each voltage. This Faraday voltage is expressed in instantaneous form and consists of two parts, as follows.

This is the general transformer voltage relationship common in transformer theory.

Development of the Equivalent Circuit for a Universal Motor Considering Brush Shift. Begin by considering a two-energy-source system (i.e., a main field and an armature) with no brushes, where:

va = voltage applied to armature

vf = voltage applied to field

ra = armature resistance

rf = field resistance

ia = armature current

if = field current

With no brushes, we have the following:

FIG. 19 Consideration of universal motor brush shift.

FIG. 21 Mutual terms between armature and field.

Permanent-Magnet DC Motors (Shunt PM Field Motors)

The general equivalent electrical circuit of a PMDC motor and its physical construction are shown in Ills. 24 and 25. The motor consists of a stator having permanent magnets attached to a soft steel housing and a commutator connected through brushes to a wound armature ( FIG. 27).One of the first things to be considered in the operation of the motor are the motor and generator action, which exist simultaneously in the armature circuit of the motor. These two principles are (1) the instantaneous emf, which is induced in the armature conductors when moving with velocity v within a magnetic field, and (2) the force produced on the conductors as the result of carrying an electric current in this same magnetic field.

FIG. 23 Equivalent circuit of the series motor.

FIG. 24 Permanent-magnet dc motor. ra = armature resistance measured at brushes, e = motor counter-emf, and v = applied voltage.

It’s known that the instantaneous force on a conductor of length _ carrying a cur rent i in a magnetic field B is:

FIG. 25 Permanent-magnet dc motor.

FIG. 26 Stator.

FIG. 27 Commutator connected to wound armature.

Or, in vector notation:

… where ? is the angle between the direction of the magnetic field and the direction of current flow in the conductor, B is in webers per square meter (teslas), i is in amperes, and _ is in meters. Motors, by design, have the armature conductors and magnetic flux at quadrature to one another. Therefore the force becomes:

Assume a situation in FIG. 28 where a conductor of length _ is located in a magnetic field and is free to move in the x direction perpendicular to the field. From the preceding discussion, a force is produced on the conductor, causing it to move in the x direction. Then:

So, differential electrical energy into the conductor less differential i^2 r loss equals differential mechanical output energy:

Thus, the force causes movement of the conductor, which in turn causes a voltage to be induced into that conductor which is opposite to the direction of the original current. This is an important concept in the operation of motors in general and one which is used to discuss the operation of the dc motor. This induced voltage is usually called a counter-emf (c_emf) voltage because of its opposition to the applied voltage.

FIG. 28 Conductor moving in a magnetic field.

This example also indicates that a reversible energy or power exchange is possible (i.e., between the mechanical and electrical systems). Therefore the same machine may operate as a motor or a generator, depending on the flow of energy in the armature.

In the motor mode, the armature of the PMDC motor is supplied with a current by the applied voltage, and a magnetic field (flux) is produced by the permanent magnets. Since the armature conductors (coils) are located in this field, each of the conductors in the field experiences a force (torque) tending to make it move (rotate), and as we have just indicated, a countervoltage (c_emf) is produced opposing the applied volt age. Other than the c_emf which the armature produce, we must recognize that the armature circuit (coils) also produces a magnetic field of its own. The armature, because of its commutator and brush construction, has a unidirectional current and therefore produces a fixed-directed mmf, measured in ampere-turns. This mmf is the product of the effective coil turns on the armature and the current through those turns. It must be understood here that the armature winding must be considered in developing those ampere-turns (i.e., there may be parallel paths through the armature, with the possibility of the armature coils being wound in series or parallel arrangements; hence, there will also be a division of the total armature current in each of the windings).This topic is addressed later. The following is a discussion of the action of the mmf produced by the armature.

Armature Reaction. The conductors in the armature carry a current proportional to the load. The magnetic field produced by this current reacts with the PM field. FIG. 29 indicates two so-called belts of armature conductors (coil sides) under each pole face. Each of the conductors comprising these belts carries current in the same direction and hence produces additive mmf. In addition, there are conductors which also carry unidirectional currents but are not under the pole arcs. The important consideration here is the effect of the presence of the PM material in the pole pieces, armature core, and armature teeth. The flux paths through the armature are influenced by the reluctances of the paths. It’s obvious that the reluctance of the flux paths under the pole pieces is less than that of the paths adjacent to the brush area, which constitute a material of much greater reluctance, namely a large air gap.

FIG. 29 indicates that the brushes are on the mechanical neutral axis (i.e., halfway between the poles).The general direction of the armature mmf is along the brush axis at quadrature to the main field. The armature conductors adjacent to the poles produce flux densities in the air gap which are equal and opposite at the pole tips. Keep in mind that the flux density produced by the armature is directly related to the armature current together with a uniformly distributed flux density in the air gap attributed to the PM field. Now the net mmf (or flux in the air gap) is the result of both the PM mmf and the armature field mmf. The resultant air gap flux is now increased at one pole tip and reduced at the other pole tip because of the armature reaction. The flux distortion in the air gap is illustrated in FIG. 30. In this figure, two poles and the armature conductors beneath the poles have been unfolded to illustrate the distortion more clearly. Ampere’s circuital law is useful here in determining the armature mmf. The resulting air gap mmf is the result of the superposition of the PM and armature mmfs. MMF drops in the poles and armature iron are considered negligible compared to the air gap mmf. For reference, positive direction for the mmf is assumed to be a flux out of a north pole. The armature mmf is shown as a linear relationship, but actually, because the armature slots are discrete, this relation ship is actually made up of small discrete stair-step transitions. However, it’s shown as a smooth curve for easier analysis.

FIG. 29 Armature magnetic field.

It’s obvious that the air gap flux varies along the pole face. Another observation from the distortion mmf pattern is that harmonics are very present in the air gap mmf. Also, because of mmf symmetry, only odd harmonics can exist in an analysis of the air gap mmf.

The results of a Fourier analysis of the air gap flux harmonics for this case are given in Ills. 4.31, 4.32, and 4.33.The figures show the fundamental, the fundamental plus third and fifth harmonics, and finally the fundamental and odd harmonics, up to and including the fifteenth harmonic. The magnitudes of the odd harmonics beginning with the fundamental are 8.94, 0.467, 1.41, 0.924, -0.028, -0.02, 0.523, and 0.0094.That is, the series can be represented as follows:

Of course, changing the PM pole arc or magnitudes of the armature and PM mmfs will change the harmonic content. The change in flux density across the air gap produces two effects: (1) a reduction in the total flux emanating from each pole, and (2) a shift in the electrical neutral axis, lending to commutation problems due to the flux distortion.

FIG. 30 Flux distortion in the air gap mmf.

FIG. 31 Fundamental frequency of air gap flux.

The flux distortion which the armature produces has been called across magnetizing armature reaction, and rightly so. The net effect is illustrated in FIG. 34, showing the resultant field where the armature cross field is at right angles to the main field. The result is a distortion of the net flux in the motor; the second effect, which was indicated previously, is a reduction in the total main field flux. This reduction of flux is not too obvious; in fact, it would almost appear from FIG. 34 that the vector addition of these two mmfs would lead to an increase in flux which would occur with the brushes on the mechanical neutral axis.

FIG. 32 Air gap flux due to fundamental plus third and fifth harmonics.

FIG. 33 Air gap flux due to odd harmonics through fifteenth.

FIG. 34 Field of armature with and without brush shift.

FIG. 35 Armature reaction resulting in main pole flux reduction.

This resulting decrease in flux can be attributed to magnetic saturation. FIG. 35 illustrates how this comes about when one operates the motor at the knee of the saturation curve. The area ABC is proportional to the reduction in flux at the pole tip with a decrease in flux density, while the area CDE is proportional to the increase in flux at the other pole tip. It can be observed that the reduction in flux exceeds the increase in flux because of the saturation effect. This is sometimes called the demagnetizing effect of armature reaction; it’s a serious consideration when using PM as a field structure. This problem is discussed later.

Because the purpose of the commutator and brushes is to change the current in a short-circuited coil from, say, a current of +I to -I in the length of time the coil is short-circuited by the brushes, some arcing at the brushes is expected. One desires to keep any voltage induced into the coil to a minimum in order to keep this arcing as small as possible.

When the brushes are on the mechanical neutral position and the main field becomes distorted, the coils being commutated will have an induced voltage from the distorted air gap flux. A shifting of the brushes to a new neutral position is suggested to keep the induced voltage to a minimum and keep the brush arcing low. It should be recognized that the amount of distortion of the field is a function of the ampere-turns of the armature and hence is dependent on the motor load. One then realizes that the electrical neutral position is a function of the load and that shifting brushes to an electrically neutral position is therefore not a matter of shifting them to a unique point in space.

Brush shifting causes another effect-that is, a demagnetization effect on the air gap flux. This effect is in addition to the demagnetizing effect which occurs because of saturation. When the brushes are shifted, the pole axis of the armature is shifted.

The result is that the angle between the main field and the armature field is of greater than 90 degrees. This process is also illustrated in FIG. 34.

In summary, there are two processes which cause reduction of the air gap flux: (1) reduction due to cross-magnetization when the brushes are on the mechanical neutral position, which changes the flux distribution across the pole face and the net flux because of saturation effects, and (2) demagnetization resulting from a brush shift and change of the armature pole orientation with respect to the PM field, resulting in the armature field mmf having a component in direct opposition to the PM field mmf.

Reactance Voltage and Commutation. It was indicated previously that during commutation the current in the shorted coil(s) under a brush must reverse and change direction. The self-inductance of the shorted winding by Lenz's law induces an emf in the shorted winding to oppose the change in coil current. This voltage is sometimes called the reactance voltage. (Actually, it may be only a portion of the reactance voltage, as is shown later.) This voltage slows down the reversal of current and tends to produce sparks or arcing as the trailing commutator bar leaves a brush. The reactance voltage hinders good commutation. The magnitude of this voltage depends directly on the square of the number of coil turns, the current flowing, and the armature velocity; it’s inversely proportional to the reluctance of the magnetic path.

When shifting brushes to seek an electrical neutral position on a motor, the shifting is done in the opposite direction of armature rotation. It must be remembered that the reactance voltage is an e = L(di /dt) voltage and is simply due to the current change during commutation. It has nothing to do with induced voltage from air gap flux. Shifting the brushes really does not help the reactance voltage. Theoretically, there is no induced voltage on the neutral axis from the motor flux (in fact, d?/dt should be zero here since the coil sides are moving parallel to the flux). In order to counteract the reactance voltage and induce a voltage opposite to the reactance voltage the brushes must be shifted further backward than the magnetic neutral axis.

At this point a d?/dt voltage is induced into the commutated coil by the PM field flux, which counteracts the reactance voltage. The result is that two voltages in opposite polarities are induced into the shorted coil thereby reducing brush arcing.

An important consideration is that there can be another component of the reactance voltage which can occur when two coil sides are in the same slot and both are undergoing commutation. There is then the following mutual inductance term which must be added to the self-inductance term:

Reactance voltage =

…where the coil being considered is the coil carrying current i1 and the coil in the same slot undergoing commutation is carrying current i2. If there is more coupling between coils, more mutual terms can exist.

Brushes are a very important consideration, and contrary to normal electrical principles, one would assume that keeping the brush resistance low would assist in reducing arcing at the brush commutator bar interface. This is far from the truth. In fact, resistance commutation is now an accepted technology. Brushes normally have a graphite or carbon formulation and hence introduce, by their characteristics, resistance into the interface. If constant current density could be achieved at a brush for all loads and speeds, an ideal condition would exist for commutation. FIG. 36 illustrates what would occur if ideal conditions are assumed. The assumptions are as follows:

_ The brush width is equal to that of a commutator segment.

_ The current density under a brush is constant.

_ The reactance voltage and resistance are zero.

_ The commutator bar insulation is small compared to the width of a commutator segment.

Note that as the commutator passes beneath the brush, for the assumptions given, the contact resistance is inversely proportional to the contact area. The sequence in FIG. 36 illustrates how the current change occurs in an armature coil under these ideal conditions. This change in surface contact changes the brush con tact resistance; hence, with constant current density, the change in current flow changes linearly according to contact resistance. Remember, the resistance is inversely proportional to the contact area. This ideal commutation has been termed linear commutation because of the linear transfer of current in the commutated coil, as indicated in FIG. 36.

If the brush width is greater than that of one commutator segment, linear com mutation still exists because the resistance still changes linearly; hence, the current density remains constant. This has an advantage in that the commutation period is longer; hence, the rate of change of current di /dt is reduced, and hence the reactance voltage is reduced also.

All of the assumptions made in the previous analysis cannot occur, and nonlinear commutation results. Armature coils have resistance, have self- and mutual inductance and hence linear commutation cannot be attained.

If these nonlinearities do occur, then either undercommutation or overcommutation occurs. These are illustrated in FIG. 37. Since brush heating is a product of brush resistance and the square of the current, linear commutation is preferred, as it minimizes brush heating. That is, the squared value of the three forms of current commutation in FIG. 37 is smallest for linear commutation.

It’s possible to develop a circuit model for the commutation process, as shown in FIG. 38. The factors have previously been described, and they are restated here.

Consider the subject coil of turns N1 and self-inductance L1, and the current ii which is being commutated.

FIG. 36 Linear commutation.

1. The voltage of self inductance (reactance voltage) of the commutated coil is FIG. 37 Overcommutation, undercommutation, and linear commutation.

FIG. 38 Coils under commutation.

....where i1 is the current being commutated and L1 is the self-inductance of the coil being commutated.

2. The voltage of mutual inductance of an adjacent coil in the same slot carrying current i2 and undergoing commutation and linked to the subject coil is:

ema = M12

where M12 is the mutual inductance between adjacent coils.

3. The voltage induced into the subject coil of turns N when cutting flux fa due to the armature mmf is:

ea = N1

Sometimes this is called commutating voltage; it opposes the reactance voltage.

This can be controlled by brush overshifting to make ea cancel the voltages ema and eLa.

4. There are voltage drops due to coil resistance rc ;Rc and Rc , the left and right con tact resistances at the brush; and the brush resistance between commutator segments rb.

The commutation process is not simple, and brush composition and interface film, measurements, and conditions all can be made to assist in the commutation process. Maintaining a suitable interface film between the brush and the copper commutator is extremely important. This interface is formed of copper oxide and free particles of graphite; it provides a general resistance commutation and supplies lubricant to reduce surface friction and heat between the commutator and the brush.

We have previously discussed demagnetization which occurs at the leading tip of a motor due to armature reaction.

The subsection on armature reaction shows that the mmf of this reaction could be closely represented as a triangular wave ( FIG. 30). We’re concerned in PM motor design with the demagnetizing effect of this armature reaction. In addition to the normal design demagnetization curve shown in FIG. 39, a set of intrinsic curves is also indicated. These curves are extremely important because they represent limits on the amount of demagnetization which can be tolerated. The value of -H required to remove the magnetization is given the value Hci. The PM of the motor normally does not span 180° electrical; therefore we are not faced with the total ampere-turns of demagnetization of the armature reaction, only that component which exists at the pole tip. If we let ß represent the angle subtended from the center of magnet to the magnet tip, a trigonometric relationship can be used to calculate the demagnetization armature reaction at the magnet tip. If NaIa represents the maximum value of a triangular wave, the demagnetization ampere-turns at the magnet tip equals (ß/90°)(NaIa). If this value is less than the design value of the radial mag net length (i.e., lm _ Hci), loss of magnetization at the pole tip can occur.

One important consideration when using permanent magnets in machinery is the fear of demagnetization of the magnet. This must be given important consideration.

The shape of the demagnetization curve of the permanent magnet is an important property in this respect. For example, in FIG. 39, for the alnico magnet with an operating point on a load line (see the next subsection, Permanent Magnets for DC Motors), only a small amount of demagnetizing ampere-turns per inch will completely demagnetize the magnet, whereas the ceramic or rare earth magnets' characteristics have much more latitude. However, there is a considerable difference in the available flux densities among the alnico, ceramic, and rare earth magnets; hence, for required large flux densities the alnico magnets are superior. Demagnetization results primarily from armature reaction, which can be large because of large in-rush currents upon motor starting or possibly because of control applications involving dynamic braking. These currents cause a large demagnetizing effect across the pole face, as previously described in the armature reaction discussion. Another important consideration when using magnets is the loss of magnetic properties as the temperature of the magnet increases. Thus, the temperature range of motor operation must be considered when employing magnets.

FIG. 39 Typical demagnetization curves for three magnets.

Magnets for PMDC motors are selected by their characteristic demagnetization curves-that is, their residual flux density Br , their energy product BH, and their coercive force Hc.

Ceramic magnets have a characteristic and a permeability very nearly equal to those of air. That is, µ0 approaches 3.2 in English units. This is an important consideration because of the large effective reluctance of the magnetic path, and removal of the armature does not change the operating point of the magnet appreciably or affect the performance of the magnet.

The air gap flux is not adjustable with permanent-magnet materials, of course.

This means that changes in motor speed and torque must be made by changes in armature current.

The performance of PMDC motors is very similar to that of wound-field DC motors, and similar equations exist. This development is the next topic to be addressed.

.... where Rc = brush contact resistance, rc = coil resistance, rb = brush resistance

Permanent Magnets for DC Motors. A permanent magnet (PM), when magnetized by an external source for use in a PMDC motor, has a remaining residual flux density and responds to a normal hysteresis characteristic. The useful portion of a magnet in PM operation is in the second quadrant of the hysteresis loop and is usually called the demagnetization curve of the magnet. It represents the relationship between B and H of the magnet once it has been magnetized. Typical demagnetization curves for three PM materials are shown in FIG. 39. In this figure, Br is the residual flux density and Hc is the coercive force. For proper utilization of a permanent magnet a user makes use of the energy product curve B _ H, shown in FIG. 39 for the rare earth magnet only.

FIG. 40 presents a series magnetic circuit in which a PM, an air gap, and magnetic material have been introduced into a closed magnetic circuit. Ampere's law around the circuit (Ohm's law for magnetic circuits) is as follows:

FIG. 40 Permanent magnet with air gap introduced.

Note: This value of flux when divided by the magnetic area represents a value of flux density and an operating point on the demagnetization curve of the PM. Any demagnetization mmf due to armature reaction would lower the operating point on the curve. Keeping the reluctance of the air gap high (i.e., large air gap) will reduce the armature reaction.

Torque-Speed Characteristics of PMDC Motors. The electrical equivalent steady state circuit is the most appropriate method for analyzing the motor. This equivalent circuit is repeated for convenience in FIG. 44, and the armature brush drop is included.

Writing Kirchhoff's voltage law around the loop gives:

The counter-emf Eg was addressed earlier when discussing the voltage induced into a conductor of length _ and moving with a velocity v perpendicular to a magnetic field of flux density B. The counter-emf voltage Eg was expressed as e and given by:

FIG. 44 Armature Circuit for the PMDC motor.

It’s necessary to modify this and express e in terms of the motor parameters.

Beginning with Faraday's law:

....where N is the number of conductors per armature path when Z is the total number of conductors (coil sides) on the armature.

Then:

Z = number of slots × coils/slot × turns/coils × 2 (conductors/turn) a = number of parallel paths through the armature P = number of poles n = armature speed, rpm f= flux, Wb

Then:

Efficiency of DC Motors Earlier discussion has indicated that a number of losses occur in PMDC motors. These losses can be summarized as follows:

1. Copper loss in the armature winding I^2 ara

2. Brush contact loss

3. Friction (brush and bearing friction) and windage

4. Core loss (hysteresis and eddy current)

5. Stray load loss (losses in addition to those above)

A comment should be made on the brush contact loss listed as number 2.

Experimentation and IEEE specifications have suggested as an approximation that:

Brush contact loss = brush voltage drop × armature brush current (4.29) This expression was arrived at for carbon or graphite brushes by observing the brush voltage drop as a function of current density. At high- and low-current steady state densities, the voltage falls off and approaches zero. However, in between these limits the voltage drop is approximately constant at about 1 V per brush. Therefore, for a pair of brushes this becomes 2 Vons small machines this voltage drop tends to increase, since the tendency is for the voltage drop to increase for lower brush and commutator temperatures.

Carbon and graphite materials have a resistivity many times that of copper and have a negative temperature coefficient. This negative temperature coefficient is attributed to the rise in brush voltage drop in small machines.

Stray load loss, as the term suggests, is a function of motor load and changes in load. Changes in load produce changes in armature current and hence affect (1) magnetic saturation in the magnetic circuit, (2) armature reaction changes, and (3) eddy current loss changes.

FIG. 45 presents a display of these losses. By definition:

Overall efficiency = output / input

=useful mechanical output / total electrical input

Sometimes both electrical and mechanical efficiencies are of interest in order to determine where improvements can be made.

FIG. 45 Efficiency of a permanent-magnet dc motor.

Mechanical efficiency = [useful mechanical output / mechanical output + rotational losses]

Electrical efficiency = electrical power output / electrical power output + electrical losses

The conditions for maximum efficiency can be related to those losses which are considered to be constant and those that vary with the motor load current. If the losses are segregated as follows:

K1 = constant losses

K2 = those losses which vary linearly with Ia

K3 = those losses which vary as the square of Ia then the efficiency is:

Equating this expansion to zero gives the condition for maximum efficiency.

Thus, for maximum efficiency the constant losses must be equal to those that vary as the square of the armature current. This is typical for all different pieces of rotational electrical equipment. The constant losses usually are considered to be the core losses, friction, and windage. Usually the brush loss is small.