GB1597379A - Electronically commutated motors - Google Patents

Description

(54) ELECTRONICALLY COMMUTATED MOTORS (71) We, GENERAL ELECTRIC COMPANY, a corporation organized and existing under the laws of the State of New York, United States of America, residing at 1 River Road, Schenectady, 12305, State of New York, United States of America. do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: This invention relates generally to rotating dynamoelectric machines and, more particularly, to such machines that receive power from a direct current or rectified alternating current power supply and that utilize electronic commutation means.

In conventional direct current rotating machines, commutation is essentially a mechanical switching operation to control the currents through the armature winding sections. This operation is accomplished in conventional machines with brushes and segmented commutators. In such constructions, the brushes wear and require frequent replacement.

Sparking and its attendant generation of RF noise is also unavoidable present. present.

These disadvantages frequently prohibit use of DC motors in critical applications even though the use of such motors might otherwise be favoured. Early attempts to provide brushless DC motors were mostly limited to: DC to AC inversion and, essentially, AC induction motor operation; use of rotor velocity for switching control but such control then was not effective at all rotor positions under different load conditions, or at starting; or use of circuits having a larger number of switching devices with the result that the circuits for such devices were both complicated and expensive.

We have now devised a simpler and less expensive DC motor and have recognised that different restrictions apply to electronically commutated motors than those that apply to mechanically commutated motors. Improved motor performance over the already good performance of conventional motor designs may be obtained by using a rotor position sensing system in conjunction with electronic switching, and a permanent magnet rotor in combination with an unconventional stationary armature core and winding arrangement for a direct current motor.

In accordance with the present invention there is provided a brushless DC motor comprising a stationary armature having a core and at least two winding stages; each winding stage comprising at least two coils of concentric winding turns accommodated by the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage having a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles, the axially extending conductor portions within each given set being comprised generally of about one half of the conductor side turn portions of at least two different coils, and such conductor portions being disposed to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing such given set is energised; the arcuate spread of any given set of axially extending conductors being less than about 120 electrical degrees; a rotor having constant magnetic polarity polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate relative to the armature in response to the magnetic poles established by the winding turns; and commutation means interconnected with at least one winding stage for sensing relative angular position of the rotor and the stationary armature and operative to energise the winding stages in a predetermined manner to establish the magnetic poles on said armature for causing rotational movement of the rotor.

In accordance with another aspect of the present invention there is provided a brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinafter defined); each winding stage comprising concentric winding turns accommodated by said core and arranged to establish a predetermined number of magnetic poles and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number equal to the predetermined number of magnetic poles; the axially extending conductor portions within each set being disposed in said armature to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing the given set is energized; the arcuate spread of any given set of axially extending conductors being less than about 120 electrical degrees; a rotor having constant magnetic polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate in response to the magnetic poles established by the winding turns; and a commutation circuit for energizing the winding stages in a predetermined manner wherein said commutation circuit includes a detector circuit for sensing a back emf signal indicative of the back emf condition of at least one winding stage, position determining circuit means for receiving the output of the detector circuit and for producing a simulated relative position output that is indicative of the relative angular position of the rotor and armature, with such relative position output determined by the back emf condition of a winding stage, and circuit means interconnected with the position determining circuit means for supplying an output signal for energizing a selected one of the winding stages.

In accordance with still another aspect of the present invention there is provided a brushless DC motor comprising a stationary armature having a slotted core and at least two winding stages (as hereinafter defined); each winding stage comprising at least one coil of concentric winding turns accommodated in nonadjacent symmetrically disposed slots of the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles; the axially extending conductor portions within each given set being comprised generally of one half of the axially extending conductor portions of the at least one coil, and such conductor portions being disposed to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing such given set is energized; the arcuate spread of any given set of axially extending conductor portions being less than 120 electrical degrees; a rotor having constant magnetic polarity polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate relative to the armature in response to the magnetic poles established by the winding turns; and commutation means interconnected with at least one winding stage for sensing relative angular position of the rotor and the stationary armature and operative to energize the winding stages in a predetermined manner to establish the magnetic poles on said armature for causing rotational movement of the rotor.

In accordance with a further aspect of the present invention there is provided a brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinafter defined), each winding stage comprising concentric winding turns accommodated by the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles; the side turn axially extending conductor portions within each given set being disposed to conduct current instantaneously in the same axial direction along the core thereby establishing a predetermined spread and establishing a magnetic pole when the winding stage containing such given set is energized; a rotor having a plurality of permanent magnet segments disposed thereon and adapted to rotate in response to the magnetic poles established by the armature; a commutation circuit for energizing the winding stages in a predetermined manner and at a predetermined angle of advance a; and wherein the arcuate spread of each permanent magnet segment disposed on said rotor is about equal in electrical degrees to the winding spread plus 180 (N- 1)/N minus 2a where N equals the number of winding stages of the motor.

In accordance with a yet further aspect of the present invention there is provided a brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinafter defined); each winding stage comprising concentric winding turns accommodated by said core and arranged to establish a predetermined number of magnetic poles and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number equal to the predetermined number of magnetic poles; and axially extending conductor portions within each set being disposed in said armature to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing the given set is energized; a rotor having constant magnetic polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate in response to the magnetic poles established by the winding turns; and a commutation circuit for energizing the winding stages in a predetermined manner wherein said commutation circuit includes a detector circuit for sensing a back emf signal indicative of the back emf condition of at least one winding stage, position determining circuit means responsive to only a positive polarity portion of the output signal from the detector circuit for integrating said positive polarity portion of the output signal to a predetermined value of volt-seconds whereupon the position determining circuit means produces a simulated relative position output that is indicative of the relative angular position of the rotor and armature, and means responsive to the simulated relative position output from the position determining circuit means for supplying an output signal for energizing a selected one of the winding stages, the position determining circuit being operative to establish a predetermined advancement of commutation of the winding stages by an angle a of from about five to about thirty electrical degrees.

The term "winding stage" as used herein refers to a number of winding turns or coils or coil groups all of which are electrically connected to a common supply for simultaneous energization in order to produce a desired number of magnetic poles.

Motors embodying the invention are readily adapted for applications where space requirements are at a minumum. For example, such motors are readily adapted to drive the compressor of a recreational vehicle or automobile air conditioning system from the output of an alternator, or battery source. In such arrangements, an electronically commutated motor and compressor may both be sealed in a hermetic enclosure. In sealed arrangements, carbon pollution products are objectionable and use of brushless motors in such arrangements provides a distinct advantage over mechanical commutation. However, since the units are hermetically sealed, the reliability of the motor must be very good, and all parts of the motor must be such that refrigerant will neither damage the motor nor be damaged by motor components or the materials from which said components are made.

Conventionally, motor manufacturers have a first type of specialized equipment, processes, tools and dies, etc., for AC induction motors and a second type of the same for DC motors. To a great extent, such AC motor making, e.g., tools, etc., are not usable for present DC motor production. In accordance with a preferred feature of the invention, DC stationary armatures for motors embodying the invention may be wound using conventional AC winding machinery.

Thus, another important advantage of the subject invention is the cost saving achieved both in manufacture and operation.

Electronic commutation may be achieved by either a full bridge or half bridge (also called star) circuit arrangement. The former improves copper winding utilization, while the latter offers the advantage of simplified electronics. By judicious selection of winding parameters, magnet size and electronics which control the commutation cycle, there may be obtained a reliable, highly efficient brushless DC motor having a desired speed-torque relationship which may be fabricated at a reasonable cost.

In one embodiment of the present invention a two stage, brushless DC motor is provided having distributed armature windings which may be wound and connected, as desired, in either full bridge or half bridge configuration. The rotor has positioned therein a pair of arcuate magnets each having a preferred minimum arc length per pole of at least about seventy to ninety electrical degrees, assuming that relatively high efficiencies are desired.

Windings are disposed in the slots of an armature member and preferably connected to produce angularly spaced apart magnetic fields when energized. Commutation of the windings may be achieved by means of a solid state circuit controlled by sensing means including a pair of sensors spaced from each other by an arcuate distance of approximately ninety electrical degrees and which are preset relative to the armature assembly for a given rotor assembly.

Another embodiment of the present invention, includes an improved electronic commutating circuit that includes circuit means for reliably simulating the relative position of the rotor and armature even under varying speed or shaft load conditions. In one form, this circuit includes circuit means that are responsive to flux condition of the armature windings for providing informative signals indicative of the relative shaft to armature position. In turn, these informative signals are used to control an armature energizing circuit to commutate and apply driving signals or currents to the armature windings in a predetermined order, whereby the armature windings are efficiently energized.

Thus, it will be realized that another preferred feature of the present invention is the provision of electronically controlled commutation of a brushless DC motor that relies on flux condition and thus is not disabled by speed or shaft load variations and also does not require mechanically coupled rotor shaft position detectors.

In accordance with one form of the invention utilizing physical position sensors, a minimum number of sensors are utilized, with this minimum number being equal to the number of binary digits needed to describe the number of switching positions, minus one, for the motor. For example, if the motor winding is to be switched four times per 360 electrical degrees of rotation, one would select the number of binary digits needed to describe the decimal number 3. As will be understood, two binary digits [i.e., 11] describe the number 3. Accordingly, only two sensors would need to be used. Since the number of sensors is related to the switching positions per 360 electrical degrees of rotation, two sensors may be used for two pole, four pole, six pole, etc., motors.

On the other hand, if there are to be six switching positions per 360 electrical degrees of rotation of the motor, the number that would need to be described in binary digits would be six minus one, or five. Since the number five is identified, in binary notation, as the number 101, it will be understood that three binary digits are involved and, accordingly, a minimum of three sensors would be utilized. At this point, it should be further noted and further understood that three binary digits can represent a number as great as 111 in binary notation (which is equal to the number 7 in decimal notation). Thus, three sensors may be utilized when as many as eight switching positions per 360 electrical degrees of rotation of the motor is desired.

In some embodiments of the invention illustrated herein, optical sensors are utilized in conjunction with one or more rotor mounted shutters. Whether optical or other types of sensors are utilized, the sensors are preset relative to the armature assembly (for a given rotor assembly) so that the switching point is advanced (i.e., so as to advance the commutation of the windings) such that a winding is energized before the rotor reaches its maximum torque per unit current producing position in order to aid the build-up of current in the winding being energized. This can yield higher torques, higher efficiencies, and higher speeds. The preferred amount of optimum advance, for a given motor design and desired end use, is primarily a function of intended nominal motor operating speed, all as discussed in more detail hereinafter.

In one exemplary motor illustrated herein, light coupling sensors were used and were supported by a bracket that was slotted so as to permit adjustable attachment to a stator.

Thus, adjustable advancement of commutation could be made to obtain either peak efficiencies or maximum speeds. When the sensors are to be permanently attached to the stator, the amount of commutation advancement will be fixed, preset amount for a given motor design. Of course, however, whether fixed or adjustable advancement is provided, it is preferred to provide an arm on a bracket that will be shaped to reach over the stator end turns and support the optical light sensors within the end turns, and thereby provide minimum overall motor dimensions.

In some embodiments of the invention, protection circuits are utilized to detect and interrupt motor operation upon occurrence of either a low speed, low voltage or high voltage condition. A low speed protection circuit is responsive to the output signals of a position determining circuit for generating a signal representative of motor speed and is operable to compare the representative speed signal against a predetermined signal representative of a minumum allowable speed and to generate an output signal for interrupting motor operation for a predetermined period of time when the motor speed is less than the minimum allowable speed. In order to assure that the motor is provided with a supply voltage that is within an acceptable range, low voltage and high voltage protection circuits are provided which compare the voltage being supplied to the motor from a power source against predetermined minimum and maximum acceptable voltages and which generate signals to interrupt motor operations if the motor supply voltage is less than the minimum acceptable voltage or greater than the maximum acceptable voltage.

Other features of the present invention and their advantages will become readily apparent from the following description taken in conjunction with the accompanying figures in which like reference characters are used to described like parts throughout the several views: Figure I is an exploded, perspective view of the main elements of a brushless DC motor embodying the present invention in one form thereof, Figure 2 is a diagrammatic front elevation view of the armature core of Figure 1, diagrammatically illustrating a selected winding arrangement for the motor of Figure 1; Figure 3 is a diagrammatic representation of parts of the motor of Figure 1 showing the position of the rotor magnets relative to the armature windings at the instant of turn-on of one of the windings, with a zero electrical degree advance setting; Figure 4 is a fragmentary, side elevational view, partly in cross-section, better showing the physical interrelationship of the support bracket assembly in the motor assembly of Figure 1; Figures 5A and SB illustrate the construction and equivalent electrical circuit, respectively, of a light sensitive element used as a shaft position sensing element in the assembly of Figure 1; Figure 6 is a schematic diagram of a solid state commutating circuit embodying features of the present invention and arranged for switching the windings of the motor of Figure 1 when such windings are connected in a half-bridge configuration; Figure 7 is a schematic diagram of the signal conditioning circuit, shown in block form in Figure 6, for producing position control signals indicative of the rotational position of the rotor of Figure 1; Figure 8 illustrates the relative relationship between the angular position of the rotor and the outputs of two light sensitive elements, A and B, and the switching pulses produced as a result of rotor rotation; Figures 9 and 10 are schematic diagrams of a solid state commutating circuit embodying features of the present invention, and arranged for switching the windings of the motor of Figure 1 when such windings are connected in a full bridge configuration; Figures IIA-C are graphical representations of torque per ampere as a function of rotor position, magnet arc length and distribution of turns for magnet arc lengths of 180 electrical degrees, 160 electrical degrees, and 135 electrical degrees, respectively; and schematically illustrate different arc lengths of rotor magnet arc lengths.

Figures 12-15 are plots of torque per ampere as a function of rotor position (in electrical degrees) for different amounts or extents of spread (or relative concentration) of stationary armature windings; Figure 16 illustrates in schematic form a circuit for supplying the A and B signals and their complements to circuitry such as illustrated in Figure 6 without the need for special mechanical type devices to sense the angular position of the motor rotor; Figure 17 illustrates the four output signals obtained from the circuit of Figure 16 and their combination according to the logic circuitry illustrated in Figure 6; Figure 18 is a schematic diagram of a motor control circuit analogous to that illustrated in Figure 16 but selectively employing a control signal which is proportional to motor load; Figure 19 is a schematic diagram of a circuit for sensing stator winding current and for interrupting that current for a short, predetermined time interval when the sensed current exceeds a prescribed value; Figure 20 is a block diagram of a vehicular air conditioner system, hermetically sealed and employing one embodiment of the present invention; Figure 21 is a schematic diagram of the solid state commutator circuit of Figure 20; Figure 22 is a schematic diagram of the speed sensor of Figure 20; Figure 23 is a block diagram of another hermetic environment illustrating use of the invention in one form thereof Figure 24 illustrates in schematic a portion of a circuit which may be interposed between the output of the NOR gates 80 and the base of transistors 82 in Figure 6 to facilitate implementation of the Figure 23 refrigeration system; Figures 25a and 25b together form a detailed schematic diagram of a modified circuit for indirect rotor position sensing and control; Figure 26 illustrates various wave forms for the circuit of Figure 25; Figure 27 is a block diagram of a control circuit for precise rotor positioning; Figures 28, 29 and 30 are diagrammatic front elevational view of stationary armatures for motors embodying the present invention illustrating selected winding arrangements for a three stage, two pole motor, a three stage, four pole motor and a three stage, eight pole motor, respectively; Figure 31 is a diagrammatic front elevational view of a stationary armature for a motor embodying the present invention, showing a selected winding arrangement for a four stage, two pole motor; Figure 32 is diagrammatic perspective view of a stationary armature for a motor embodying the present invention, showing a selected monofilar winding arrangement for one stage of a motor; Figure 33 is a diagrammatic perspective view of a stationary armature for a motor embodying the present invention showing a selected bifilar winding arrangement for one stage of a motor; Figure 34 is a diagrammatic front elevational view of a stationary armature for a motor embodying the present invention, illustrating a winding arrangement for a two stage, two pole motor in which windings share armature slots; Figure 35 is a diagrammatic front elevational view of a stationary armature for a motor embodying the present invention, illustrating a winding arrangement for a two stage, two pole motor in which the stationary armature is provided with empty slots; Figure 36 is a diagrammatic front elevational view of a stationary armature for motor embodying the present invention, illustrating a winding arrangement for a two stage, two pole motor in which the windings are non-symmetrically disposed in slots of the stationary armature; Figures 37a and 37b together form a detail schematic diagram of a commutating circuit embodying preferred features of the present invention and arranged for controlling commutation of a three stage brushless DC motor.

Figure 38 illustrates various wave forms for the circuit shown in Figures 37a and 37b employed to control commutation of a three stage brushless DC motor.

Figure 39 illustrates various wave forms for the circuit shown in Figures 25a and 25b when the circuit has been modified to control commutation of a four stage brushless DC motor.

Figure 40 is a simplified end view of a permanent magnet rotor and system for determining by a test pulse the position of that rotor.

Figure 41 is a schematic diagram of a wave form responsive circuit embodying preferred features of the invention in one form thereof, and which may be substituted for the conventional position sensors to supply A and B signals to the circuit of Figure 6; Figure 42 depicts several voltage wave forms associated with the circuit of Figure 41, illustrating proper (preferred) and improper (not-preferred) commutation timing; Figure 43 illustrates current wave forms in the coil of Figure 4a for early, preferred, and late, commutation, respectively; Figure 44 is an idealized depiction of a single armature coil in relation to the rotor flux field Figure 45 illustrates in schematic form a four-stage sensorless commutating circuit; Figure 46 illustrates in schematic form a three-stage sensorless commutating circuit; Figure 47 illustrates in schematic form a more sophisticated and more efficient three-stage sensorless commutating circuit.

Referring to Figure 1, there is illustrated the component parts of a brushless DC motor of one type that may embody features of the present invention in one preferred form. The motor is provided with two winding stages, two poles and a permanent magnet rotor 10 mounted on a shaft 11 which is rotatably supported by conventional not shown bearing means within any desired not shown housing. The rotor 10 is magnetized across its diameter in a manner known to the art. In the illustrated embodiment, the rotor 10 comprises a solid magnetic steel core 12 and a pair of arcuate magnets 13 and 14 disposed on the periphery of the core in diametrically opposed relationship. The magnets 13 and 14 were ceramic magnets, but it will be understood that they could be cobalt samarium, Alnico, or any other available type of magnet material. The primary selection criteria are expense and physical size of the motor. The arc length of each ceramic magnet is preferably between 135 electrical degrees and 160 electrical degrees, but could be as much as 180 electrical degrees and as low as approximately 90 electrical degrees. Arc lengths of less than 120 electrical degress would generally result in poor efficiency with the winding arrangement specifically shown in Figure 2, and are thus not preferred with such arrangement.

The optimum magnet arc length set forth above generally remains the same for motors with multiple stages and/or pole arrangements although the number of permanent magnets and the physical or mechanical arc length of the magnets will vary in accordance with the number of electrical poles (or pole pairs) created by the windings. For example, a three stage, four pole motor would be provided with four permanent magnets each having an arc length preferably between 135 electrical degrees and 160 electrical degrees as set forth above, that is, between 67-1/2 and 80 in terms of mechanical degrees. Likewise, a four stage, 6 pole motor would have six permanent magnets with an arc length, preferably between 135 electrical degrees and 160 electrical degrees, that is, between 45 and 53-1/3 mechanical degrees. The magnet arc length will also vary in accordance with the concentration of windings within the stationary armature.

When the arc lengths are on the order of 90 electrical degrees, the windings would be made relatively more concentrated. For example, in a 24 slot stationary armature used for two stage, two pole operation, the outer coils of each coil group preferably would span about 10 slots. In addition, each coil group (2 one-half sets) preferably would include three coils per one-half coil set spanning 9, 7 and 5 teeth, respectively. This type of arrangement would provide optimized efficiency characteristics of such motor.

On the other hand, for some applications, very concentrated windings would be provided that occupied 1800 electrical.

The stationary armature assembly 15 includes a relatively low reluctance magnetic member 16 which is formed of a plurality of like armature laminations 17 assembled in juxtaposed relationship. The laminations may be held together by a plurality of armature through-bolts 18, only two of which are fragmentarily shown, that pass through coaxially positioned through-bolt holes 19 in the stator laminations. Alternatively, the core laminations may be welded, keyed, adhesively bonded together, or merely held together by the windings, all as will be understood by persons skilled in the art.

Each lamination includes a plurality of teeth 20 along its interior bore such that the assembed laminations provide a plurality of axially extending slots 21 within which the armature windings 22 are disposed.

Windings 22 may be wound by means of conventional induction motor winding machinery. Thus, the winding turns may be wound directly on coil injection tooling for disposition in the core slots; or the windings may be wound into a coil receiver, transferred to coil injection tooling and subsequently be axially inserted into the core slots, for example, with equipment of the type shown and described in U.S. Patents 3,522,650; 3,324,536; 3,797,105 or 3,732,897, the disclosures of which are incorporated herein by reference.

Preferably, in a two stage, two pole motor, each winding has a spread of approximately 90 electrical degrees across the armature slots so as to generate mutually perpendicular magnetic fields when energized, assuming that the windings are not required to share slots and that all slots of the stationary armature are utilized. The winding end turns extend beyond the core end faces, and the winding ends or leads are brought out and connected separately to the control circuit and the associated switching means.

Referring to Figure 2, there is illustrated an exemplary winding arrangement for a one twentieth horsepower, 3000 rpm electronically commutated DC motor embodying principles of the invention. The selected lamination or punching includes twenty-four slots, and further, each winding stage is provided with fifty-four bifilar turns (twenty seven turns for each coil group). With the oppositely disposed pairs of windings wound bifilarly, there is provided four winding stages, a, b, c, d. Winding stages a and c are bifilar wound and occupy the six uppermost and six lowermost slots 21, as viewed in the drawing, and are shown as being included by the dash lines. Winding stages b and d are bifilar wound, and occupy the six left-hand and six right-hand slots, as viewed in Figure 2. The armature 17 is wound with a winding turn distribution of 10 turns, 10 turns, 7 turns, from outermost to innermost coil, respectively, for each coil group shown in Figure 2. Each winding stage is shown as including two coil groups, of course. The particular number of turns in a slot for each winding stage and the resultant distribution could be varied according to the desired motor characteristics to be obtained. For example, the outermost coils of each coil group could be maximized in number while the innermost coil turns are minimized so as to concentrate the winding. When the winding is so concentrated, a higher average torque will result (assuming armature core, rotor construction winding resistance, and total turn count is held constant), but the switching point will be more critical and the amount of advancement may have to be charged. Also, dips in torque (during running and standstill) would be generally of greater amplitude, but of less duration.

As shown in Figure 2, the turns of each winding stage are concentrically disposed in a given pair of slots with the desired number of turns in each slot. The winding stage of course continues in the next pair of slots, and the desired number of slots or teeth are spanned which in the illustrated embodiment is eleven teeth (for each coil group) and providing a spread of, for example, 90 electrical degrees for each winding stage and enabling the winding stages to sequentially generate mutually perpendicular magnetic fields when sequentially energized. By using bifilar strands, two winding stages are wound simul taneously and one end of each strand may then be conveniently grounded to provide a half-bridge (star or WYE) winding connection configuration. It is again noted that the arrangement is readily adapted for winding and placement by conventional winding machinery used for winding AC motors.

It will be noted that in Figure 2, the winding stage b has been illustrated somewhat differently than winding stages a, c, and d. Winding stage b has been shown to illustrate the direction of current flow therein during at least one commutation period and arrows have been used to indicate the direction of current flow in the end turn portions of winding stage b. On the other hand, dots and crosses (enclosed within circles) have been used to illustrate the direction of current flow in the respective turns of winding stage b that are disposed within the magnetic core slots. With the notation used in Figure 2, a dot would indicate that current is flowing upwardly out of the plane of the drawing, and crosses indicate current flowing downwardly relative to the plane of the drawing. With current flow as indicated by the dots and crosses in Figure 2, winding stage b would establish north and south poles oriented as represented by the letters Nb and Sb in Figure 2.

The stationary armature assembly has an axial bore 23 within which is received the rotor 10. The arcuate magnets 13 and 14 are disposed (e.g., by an adhesive such as an epoxy resin) on the outer surfaces of the low reluctance core 12 (which may be laminated by does not need to be) to establish constant magnetic polarity regions with North-South polarizations as indicated in Figure 3. Magnetization is in the radial direction with radial thickness selected to produce the desired magnetomotive force (for a given magnet material) or to assure that no irreversible demagnetization occurs from the fields produced by the current in the armature windings during stalled conditions.

It will be understood that although magnets 13 and 14 have been described as being adhesively bonded to the outer peripheral surface of the core 13, it is important only that fields of opposite polarity be established by the rotor assembly. Thus, bar type magnets (or magnets of any other desired shape) may be utilized in the fabrication of rotor assemblies for motors embodying the present invention. The magnets (when permanent magnets are used) may be disposed within a magnetic iron cage structure or fabricated in any other desired manner. In fact, an excited rotor wherein the north and south poles are established by current carrying conductors may also be utilized. In the latter case, of course, slip rings or any other suitable means would be used to interconnect the rotor windings, with a source of excitation current. The axial length of the magnets is dependent on the total flux desired.

The instantaneous torque curves and the net torque output is dependent on the winding distribution within the slots of the stationary armature and on the magnet arc or arc length beta (ss) of the magnet which, as hereinbefore noted, should be preferably constrained within 135 electrical degrees and 160 electrical degrees, the greatest efficiency where the windings are not required to share slots and all slots are utilized.

As shown in Figure 1, adjacent one end of the rotor is positioned a shaft position sensor assembly 40 comprising a shutter 41 and bracket 42 for supporting a pair of optical interrupter modules, i.e., optical light coupling sensors 43 and 44. Shutter 41 may be formed of any optically opaque (at the pertinent wave length) material or coated material such as, for example, aluminum (brass, steel, etc.), and includes a flat disc shape element 45 having a shutter flange 46 extending along the periphery of the disc element for an arcuate distance of approximately 180 electrical degrees. The disc element includes a central opening 47 slightly larger in diameter than the diameter of the rotor shaft 11 so that it may be conveniently passed over the rotor shaft and mounted flush with the end face of the solid steel core 12. To this end, a pair of bolt receiving openings 48 are provided for receiving mounting bolts (not shown) adapted to be threaded into pre-threaded openings 50 provided in the core 12. Of course, as should be readily apparent, other suitable arrangements could be utilized for mounting the shutter to the rotor, so long as flange 46 extends outwardly from the rotor and is cooperatively associated with sensors 43 and 44 to effect generation of rotor position reference signals in response to the position of the rotor relative to the stator.

As shown in Figures 1 and 4, bracket 42 comprises a first arcuate shaped segment 51 having an elongated slot 52 along its length. Slot 52 permits the bracket to be adjustably attached to the stator or stationary armature laminations 17 by means of the armature through-bolts. A second arcuate shaped segment 53 is supported radially inward of the first segment by means of an interconnecting U-shaped strap or arm member 54. As more clearly shown in Figure 4, strap 54 is shaped to pass over the end turns and support the sensors 43 and 44 with a 900 spatial orientation for the two stage motor within the end turns of the armature windings and in cooperative relationship with the outward extending shutter flange 46. It should be noted that the sensors are supported within the outermost axial dimension of the winding end turns as well as within the radial dimension thus minimizing the axial length of the motor, the dimension of the strap adding very little to the overall axial length.

Referring to Figures 5A and 5B, there is illustrated a typical mechanical and electrical configuration of an optical sensor. Such sensors are conventional commercial devices and may be, for example, General Electric H13A2 optical couplers. As shown, the coupler includes a source of light energy 55 which may be a light emitting diode and a light sensor 56 which may be a light sensitive phototransistor in light coupling relationship with the light emiting diode. Diode 55 and phototransistor 56 are formed in separate blocks 57 and 58, respectively, separated by a channel 59 and mounted to a supporting base 60. Base 60 of each sensor is secured at opposite ends of supporting segment 53 such that the sensors are spaced from each other by an arcuate distance of 90 electrical degrees. Convenient openings are provided in the segment to allow the terminal conductors 61 of the sensors to pass freely therethrough.

Commutation of the stationary armature windings is achieved by means of a solid state control circuit 70 comprising NOR gates and transistor switches and drivers activated in response to signals from the shaft position sensors. Bridge circuit arrangements which hereafter shall be used to denote either a half-bridge (unidirectional) circuit arrangement or a full-bridge (bidirectional) circuit arrangement are employed to effect commutation. The circuit may be of the type illustrated in Figures 6 and 7 where a half-bridge configuration is employed or of the type illustrated in Figures 9 and 10 where a full-bridge arrangement is employed.

In either case, current switching in the armature windings is preselectively set by the relative positioning of the sensor so that commutation of the stationary armature winding is advanced as is hereinafter explained.

Briefly summarizing operation of the circuit, it will be apparent from the several schematic figures that the control circuit receives the output of the light sensitive shaft position sensor assembly 40 to derive switching signals for commutation of the stationary armature windings. To this end, the outputs of the sensors 43 and 44 produce two position signals indicative of the position of rotor 10 with respect to the fixed position of the stationary armature windings. The two position signals are applied to a first signal conditioning circuit 70' in Figure 6 which develops four control signals which correspond to those positions of the rotor while (1) shutter flange 46 is passing through channel 59 of sensor 43 thereby blocking its photosensor; (2) shutter flange 46 is passing through channel 59 of both sensors 43 and 44 thereby blocking both photosensors; (3) shutter flange 46 is passing through channel 59 of sensor 44 and blocks the photosensor of sensor 44 but has unblocked the photosensor of sensor 43; and (4) shutter flange 46 is clear of both sensors. In this manner, the shutter operates to block or intercept the light from the source of light energy of each optical coupler during one half of each revolution of the rotor, while permitting passage of light energy from each source of light to its associated photosensor during the remaining half of the rotor revolution. However, by supporting the sensors in a 90" spatial relationship, the on-off combination of the sensor provides four position control signals, while the adjustable bracket assembly provides a convenient means for preselec tively advancing the commutation of the windings and aid the build-up of current in the winding being commutated and obtain a desired speed-torque relationship with greater efficiency.

Referring to Figure 7, it will be understood that the output of each coupler 43, 44 is high when the energy received from the LED 55 is prevented from exciting the associated phototransistor 56, i.e. when shutter flange 46 passes between the LED and the phototransistor. As will be understood from considering Figure 7, a first position signal occurs whenever the coupler 43 is blocked and this signal appears on line 68. A second position signal occurs whenever coupler 44 is blocked and this signal appears on line 69.

As hereinbefore noted, each coupler consists of an LED 55 and a phototransistor 56. The collector of each phototransistor is independently connected through an associated resistor 71 or 72 to a positive bus line 73. Diodes 55 are connected in series and in turn through biasing resistor 74 to the positive bus 73. The emitters of the phototransistors and the series connected diodes are returned to a common ground line 75.

The first signal conditioning circuit 70' includes four NOR gates arranged to develop the four position control signals A, A, (not A), B, and B (not B) which are indicative of the rotational position of the rotor 10 (within a ninety degree region) and which are utilized to control current switching in the stationary armature winding stages. To this end, one input of each NOR gate 76 and 77 is connected to lines 68 and 69, respectively, and the other inputs of each NOR gate 76 and 77 are returned to ground line 75. The outputs of NOR gates 76 and 77 establish the A and B position control signals applied to the second signal conditioning circuit. The A and B control signals are also applied respectively to one of the input terminals of NOR gates 78 and 79, the output of which comprises the A (logic complement not A) and B (logic complement not B) position control signals. The other input of each NOR gate 78 and 79 is grounded. The duration and sequence of the signals A, A, B, B, are schematically depicted in the upper portion of Figure 8.

The four position control signals A, A, B, and B are applied to the second seignal conditioning circuit 80 shown in detail in Figure 6. The function of the second signal conditioning circuit is to produce four switching signals for sequentially switching the associated stationary armature winding stages 22a, 22b, 22c and 22d. To this end, each winding stage is associated with a separate signal channel including its own NOR gate 80, transistor 81 and driving stage comprising transistors 82 and 83. Operation of each channel is identical and to avoid duplication, the description will be limited to the operation of a single channel. The channel for the "A" winding stage is referred to as the "a" channel and the associated components in that channel have each numerical reference character followed by the letter "a" to signify its association with that channel. Thus, the "a" channel controls switching the winding stage "a", the "b" channel controls switching of the winding stage "b", and so forth.

Channel "a" is shown as the lowermost channel of the second signal conditioning circuit in Figure 6. NOR gate 80a has its two inputs connected to receive the A and B inputs from NOR gates 78 and 77. Likewise, each of the other channels are arranged to receive two position control signals from the first signal conditioning circuit 70 such that the four channels produce four successive switching pulses for each revolution of the rotor, as best understood by referring to the lower half of Figure 8. In this portion of Figure 8, the duration and sequence of the signals A + B, A + B, A + B, and A + B are schematically depicted.

Logic is performed with two inputs for each gate. The gates are connected in such a way that when the output of one NOR gate is a 1, the outputs of the other NOR gates are at zero. Gate 80a, for example, has a 1 output when the A and B inputs are both at zero. This occurs just once in a revolution of the shutter 41. Similarly, for gate 80b, its output is 1 when the A and B inputs are at zero. This, again, is a singular combination in each revolution.

Gates 80c and 80d are in like manner connected to A, B and A, B inputs, respectively.

The switching signals from each NOR gate 80 are amplified by an associated transistor 81 to which the switching signal is applied through a base resistor 84. The output of each transistor 81 is directed to the base circuits of a power switch set comprising transistors 82 and 83 which are switched to effect energization of the armature winding stages 22a-22d in a predetermined timed relation. Transistor 81 comprises an NPN transistor, the emitter of which is grounded through line 85. The collector of each transistor 81 is connected through a resistor 86 to the base of a PNP transistor 82. The collector and emitter of each transistor 82 are connected, respectively, to the base and collector of the associated transistor 83 forming a conventional modified Darlington configuration. For larger motors, larger power transistors 83 or the paralleling of two or more transistors may be desired.

Each stationary armature winding stage is connected through the collector-emitter junction of its associated transistor 83 to the positive bus 87. A protective diode 88 is connected across the emitter-collector of each transistor 83 to provide a current path from the associated winding stage to the positive bus line 87. To this end, the anode of each diode is connected to the ungrounded side of the associated winding stage 22 and the emitter of transistor 83 to insure that the polarity of the diode is such as to allow the return of energy released by the decaying magnetic field of a winding stage when it is deenergized. The back current generated by the decaying magnetic field is shunted past the transistor 83 through line 90 causing the charging of capacitor 91 which is connected across the positive bus line and ground. The energy stored in capacitor 91 will be returned to the system upon discharge of the capacitor when the next winding stage is turned on resulting in an overall increase in efficiency of the motor. This improvement in efficiency may be as high as 10%.

The protective circuit formed by each diode 88 and capacitor 91 is equally effective for rectified AC and battery supplies. It should be noted that for a rectified AC line, the diodes associated with the supply source are switched in such direction as to allow current to flow through the motor, but not back to the line. Thus, the capacitor 91 serves to store energy from the switched winding stages. Capacitor 91 could be replaced by a zener diode which would absorb and dissipate the recovered energy as heat. While such an arrangement could provide protection to the transistors 83a, b, c, d, it would not provide for improved efficiency because the energy would be dissipated rather than being returned to the system.

Resistor 92 connected in positive bus line 87, together with capacitors 93, 94 and a 15-volt zener diode 95 (for a nominal applied average voltage of 12 volts) provides a protective filter network for the circuit components against the possibility of line 87 being raised to a voltage great enough to destroy the solid state components which could occur, for example, if the motor is run off a battery charger that could supply more than eighteen volt peaks.

Stationary armature winding stages 22a-22d of Figure 6 are wound bifilarly and are arranged in a half-bridge configuration with one end of each winding stage tied to a common ground. This provides an efficient arrangement which enables the winding stages to be switched on and off individually with a minimum of electronics and which enables the inductive energy of a switched winding stage to be recovered. In this connection, when winding stage 22a is turned off, for example, the decaying magnetic field induces a current in the companion conductor of winding stage 22c due to the bifilar winding arrangement and the resultant transformer action. The feedback diodes 88 around each switching transistor provide a path for current associated with trapped inductive energy and protect the transistor, while capacitor 91 enables this energy to be recovered. This arrangement provides for relative utilization of the windings in a slot of only 50%. To provide for full utilization of the windings, providing for even more efficient utilization of winding material, resort may be made to the full bridge circuit arrangement of Figures 9 and 10.

In accordance with the arrangement shown in Figures 9 and 10, the stationary armature winding stages 122a and 122b are wound in the same manner as the stationary armature winding stages 22a and 22b of the half-bridge configuration of Figure 6. Thus, instead of using bifilar strands as was done with winding stages 22a, 22c, only single strands are employed and a particular winding stage is switched in by switching a pair of transistors. To this end, four power switch sets 101-108 are provided for each pair of winding stages. Each power set comprises a pair of transistors arranged in a Darlington configuration. The base of the input transistor of each power set is connected through its associated base resistor 109-116 to the output of a transistor amplifier 81 of one of the channels in a manner shown, for example, in Figure 6. The "a" channel output of transistor 81a is connected to the input of power switch set 101 and 103, the "b" channel to sets 105 and 107, the "c" channel to sets 102 and 104 and the "d" channel to sets 106 and 108. Winding stage 122a is energized when current Ia flows with the power sets 101 and 103 turned on. Winding stage 122a, in effect, acts like winding stage 22c of Figures 2 and 6 when power sets 102 and 104 are turned on and current Ic flows. The full bridge circuit for switching in winding stage 122b operates in a like manner when currents Ib and Id flow. Each transistor is provided with a protective diode 117-124 connected across the emitter-collector terminals and poled to provide a path for the current associated with stored inductive energy which is released when the transistors are turned off.

The aforementioned circuit arrangements provide a relatively simple, but highly efficient and economical means for controlling the commutation of a motor embodying the present invention Advanced timing angle (or advancement of commutation) is defined in accordance with Figure 3. Zero advance would exist if a winding stage was turned on when the magnetic center of a rotor magnet was moving theretoward and at the instant that the magnetic center of the rotor was 135 electrical degrees from alignment with the axis of the magnetic pole established thereby. This would be the theoretical optimum. However, switching of the winding stage 10 electrical degrees before this theoretical optimum position is reached, constitutes a 10 degree advancement of commutation. The preferred amount of advancement of the timing angle is associated with the L/R time constant of the winding. At 0 electrical degrees advance, the current in the winding stage builds up too slowly to achieve maximum possible torque throughout its full "on" time. Advancing the commutation angle, however, takes advantage of the fact that the generated back emf is less during incomplete coupling, i.e., when the polar axes of the rotor and winding are not in exact alignment, and current build-up time and torque development can, therefore, be improved. Too great an advance incites current overshoots with consequent adverse effects on efficiency, but the optimum setting of the advance depends to some extent on the desired speed and torque operating points of the particular motor. Timing angle is preselectively adjusted by peripheral rotation of bracket 42, which positions the light coupling sensors 43 and 44 with relation to shutter flange 46.

With continued reference to Figure 3, the center of the north and south magnetic poles established by winding stage 22b of Figure 2 have been indicated by the reference notation Nb and Sb, respectively. The general location of the polar axes or centers of magnets 13, 14, on the other hand, are represented by the notation N, S. It is to be understood that north and south poles Nb, Sb, are established by winding stage 22b when it is energized as indicated in Figure 2.

During motor operation, winding stages 22a, 22b, 22c, and 22d are commutated in sequence; and as the poles Nb, Sb (associated with winding stage 22b) disappear; the poles Nc, Sc (associated with winding stage 22c) appear. It will be noted from Figure 3 that the center of magnetic pole S of magnet 14 is positioned 45 electrical degrees past pole Sb. In theory, winding stage 22b should be switched on at this instant to establish poles Nb, Sb; and winding stage 22b should remain energized for ninety electrical degrees. Then, winding stage 22b would be switched off and winding stage 22c would be switched on, assuming clockwise rotation of the rotor as indicated by arrow R in Figure 3.

It has been found that better performance results when commutation of the winding stages is effected in advance of the theoretically desirable switching point or angle by a predetermined angle alpha (a) (in electrical degrees).

For the embodiment having a winding arrangement as described hereinabove, the angle alpha equaled about twenty electrical degrees. Thus, winding stage 22a was deenergized, and winding stage 22b was energized to establish poles Nb, Sb when the axis of poles N, S of magnet 14 was about 135 plus 20 or 155 electrical degrees therefrom. Ninety electrical degrees later, winding stage 22b was deenergized and winding stage 22c was energized so as to establish poles Nc, Sc. This then continues of course for the four winding stages 22a, b, c, d, as will be understood.

Although oppositely located winding sections can be coupled simultaneously by opposing magnets of the rotor, all turns of a given winding stage may not be fully coupled due to the distributed nature of the windings and foreshortening of the rotor magnets. Because of this, the output torque per ampere input to the armature winding (T/I) is a function of rotor position, magnet arc length, the number of winding turns, and the placement of the turns in the armature. Figures 11A, B, C reveal the effect on the ratio of T/I when magnets of different arc lengths (in electrical degrees) are used with a given four-stage armature configuration. In a multi-stage motor, the flat portion of the torque per ampere (T/I) curve for each winding will be reduced the same number of electrical degrees as for a four-stage motor by a reduction in magnet arc length although the "on" time for each winding will vary from that shown for a four-stage motor in Figures 11A-C.

Figure 11A represents a plot of T/I when the magnet arc length is 180 electrical degrees and the four winding stages 22a, b, c, d have the same number of turns in each slot. The solid trapezoidal curve shows instantaneous torque per ampere for a constant value of current flowing in winding stage 22a if that winding stage is energized or left "on" for a full revolution of the rotor. The dashed trapezoid Figures 12-15 would result, respectively, with winding set side turn spreads of 90, 60, 120, and 30 electrical degrees, respectively. The duration or extent, in electrical degrees, of the flat portions of the curves in Figures 12-15 have been denoted in the drawing figures. It will be noted that the duration of such flat portions decrease with increasing coil side turn spreads. Stated conversely, increasing coil side turn concentrations cause increased flat portion (maximum T/I) duration.

The curves of Figures 12-15 are based on winding distributions that are assumed to provide an equal number of turn segments per slot. As will be appreciated from Figures 1 and 2, windings 22 include end turn portions disposed along the end faces of the stationary armature core, and side turn portions that are disposed along the axially extending armature core slots.

Taking winding stage 22b as exemplary, and referring to Figures 2, winding stage b is formed of two sections or coil groups. Each of these groups has three concentric coils, with each coil comprised of a plurality of turns and with side turn portions of such coils in a stator slot. The peripheral extension distance or arcuate expanse of the outermost coil of each coil group determines the span of each coil group. However, the "spread" or "concentration" of winding stage 22b is determined by the collective arcuate expanse of one-half of the side turn portions of both coil groups. With a maximum concentrated winding, only one coil would be used, however, and all conductors for such winding would occupy only a total of two slots.

Thus, all of the conductors of winding stage 22b that carry current into the plane of Figure 2 (or out of the plane of the drawing) collectively establish a "spread" of ninety electrical degrees. If the winding stage 22b consisted of two coil groups each having only one coil and these coils shared the same slot, then maximum "concentration" or minimum "spread" would be achieved.

It will be understood from a comparison of Figures 11A-C and 12-15 that maximum values of T/I will be of longer duration if the winding "spread" is minimized and the rotor magnet arc length is maximized.

The above-described Figures 11A-C and 12-15 illustrate the interrelationship between magnet arc length, winding spread and the T/I contribution of a winding. Although shown for a four-stage motor, this interrelationship may apply to motors with more than four stages. As the number of stages are increased, the spread of a winding stage is generally reduced to provide an increase in the flat portion of the T/I curve for each winding stage allowing an overlap in "on" times between the winding stages assuming the magnet arc length remains the same. This overlap in winding stage "on" times may be desirable to achieve greater winding utilization, increased motor efficiency and increased motor torque output. However, just as with the previously described four-stage motor, reductions in magnet arc length of 20 electrical degrees (180 to 160) and of 45 electrical degrees (180 to 135 causes reductions of 20 and 45" respectively in the flat portions of the T/I curves for each winding in a muli-stage motor. When winding inductance and rotor speed are taken into account, the optimum torque at rated load should occur when the winding stages are energized in advance by approximately 20 electrical degrees. Because of this, magnet arc length can be reduced from 180 electrical degrees to 160 electrical degrees with essentially no loss in motor performance.

When maximum torque over a full rotor revolution is desired, the wave forms of Figures 11A-C and 12-15 should be kept "flat" as long as possible. However, if the duration of maximum T/I were a theoretical maximum of 180 electrical degrees, a square wave would result. In other words, the leading part of the wave form would become infinitely steep.

However, with steeper wave forms, there is more possibility of starting problems.

Therefore, it is preferred that the leading part of the wave form be as steep as possible without causing objectionable starting problems. In this connection it should also be noted that running requirements may require more winding turns, and therefore a greater winding "spread". This in turn results in a less "steep" wave form, which in turn would cause a need for a greater advancement of commutation for optimized running efficiency.

While NOR gates have been employed in Figure 6, a wide variety of combinations of AND, OR, NAND, and NOR gates may be used to accomplish the desired logical combinations. As a further variation on the circuit of Figure 6, provision may be made for sensing the current in one or more stationary armature winding stages and for limiting the currents supplied to the stationary armature winding stages when the sensed current exceeds a prescribed value. Figure 19 illustrates an inhibiting circuit which will sense armature current and interrupt that current for a short predetermined time interval each time that the sensed current exceeds a prescribed value. The circuit of Figure 19 is operative primarily during motor start-up and the predetermined time interval is less than the time interval during which a specified stationary armature winding stage is enabled. The inhibit feature of Figure 19 may be incorporated into the system depicted in Figure 6, for example, by inserting the relatively small resistance 204 in series between the voltage source and the several armature windings, for example, by placing it in the upper right-hand line of Figure 6 which connects to the plus V source. To adapt the logic circuitry of Figure 6, for an inhibit function, the several gates 80 may be three input NOR gates with that additional input (not illustrated in Figure 6 for each gate connected together and to the inhibit output line 206 of Figure 19. Clearly, numerous other implementations of the inhibit function are possible.

In Figure 19, the resistance 204 will be in series with a stationary armature and the circuitry of Figure 19 will respond to the voltage across resistance 204 to disable the armature winding for a short time interval when that resistance voltage exceeds a predetermined value. For comparison purposes, a regulated, for example, ten volt, direct current source, is applied to terminal 208 which, while not shown, may comprise a conventional center tapped transformer or bridge rectifier zener regulated direct current source. The inhibit signal is, for example, of 300 microseconds duration after which the NOR gate of Figure 6 or other transistor circuitry will be allowed to reenergize the particular winding stage.

In Figure 19, an operational amplifier 210 amplifies the voltage sensed across resistor 204 and supplies that amplified voltage to one input of amplifier 212. Amplifier 212 is connected as a comparator and receives as its other input a reference voltage as scaled by the setting of potentiometer 214. The output of amplifier 212 is differentiated and employed to enable amplifier 216. The amplifier 216 is connected as a "one shot" and remains on for a time duration determined by the time constant of the potentiometer 218 and capacitor 220. The one shot amplifier 216 provides a high signal on line 206 for the exemplary 300 microsecond time period to disable the motor winding when, for example, the instantaneous winding current exceeds 10 amperes.

The three illustrated amplifiers in Figure 19 are integrated circuit operational amplifiers, for example, type MC3301B. The capacitor 222 between the output of amplifier 212 and the input of amplifier 216 performs the differentiation function. The output of amplifier 216 goes high to inhibit the winding which level charges capacitor 220 by way of the variable resistor 218 and when the charge on capacitor 220 becomes sufficiently large, the difference between the two input signals to amplifier 216 is low enough to force the amplifier output back to its low level and capacitor 220 discharges by way of diode 224.

The block diagram of Figure 20 illustrates one hermetic environment in which the brushless DC motor of the present invention finds particular utility. A hermetically sealed refrigeration system 226 includes a conventional compressor (not illustrated) driven by the brushless DC motor 228 which may, for example, be of the type illustrated in Figure 1. The motor 228 receives armature energizing current from the solid state commutating circuit 230 and provides thereto position signals, for example, from the optical position indicators discussed earlier. A speed sensing circuit 232 as well as a temperature control, such as a conventional thermostat 234 provides an input signal to a solid state field current regulating circuit 236. The field current regulating circuit 236 controls the vehicle engine-driven generator or alternator 238 which in turn supplies energy to the motor 228 by way of the commutating circuit 230. By controlling the field current to the alternator or generator 238, the power supplied to the motor is readily controlled to in turn control the resulting temperature from the air conditioning system. The system outlined in Figure 20, eliminates the more conventional belt-driven compressor arrangement typically found in vehicle air-conditioning systems and provides instead a system which may be adapted to either energization from the alternator 238 or when the vehicle is parked from a standard alternating current outlet. The commutating circuit 230 may be of the same general configuration as the circuit of Figure 6 and, in the event that optional operation from a standard alternating current outlet or use of an alternator rather than a DC generator is desired, suitable bridge or other rectifying circuitry would be incorporated in the Figure 6 circuit or in the alternate solid state commutator circuits 230 as illustrated in Figure 21 with the corresponding speed sensing circuit 232 illustrated in Figure 22.

In the specific implementation illustrated in Figures 21 and 22 the voltage output from the engine-driven alternator 238 is applied to terminal 240 while the twelve volt direct current vehicle battery source is applied at terminal 242 and a zener regulated battery voltage of twelve volts is applied at terminal 244. In Figure 21, position sensors 246 operate much as before in conjunction with a shaft mounted light shutter so that the respective light emitting diodes 248 and 250 will cause either one or both of the light sensitive transistor 252 and 254 to be conducting. The conducting or non-conducting indications or signals are inverted by NOR gates 256 and 258 which in conjunction with NOR gates 260 and 262 may be a type CD-4001 integrated circuit and function as a primary decoder to form the A, B, not A and not B signals as before. These signals are supplied to the corresponding inputs in Figure 22 as indicated and further are logically combined by NOR gates 264, 266, 268, and 270 in the manner already described in conjunction with Figure 6 to provide the four winding enabling signals, only one of which occurs over every ninety degrees of shaft rotation. As before, the four sequential winding energizing signals are then applied to four corresponding transistors such as 272 for amplification to in turn be supplied to four winding enabling power modules, only one of which is illustrated in Figure 21. Each power module is connected to one of the four illustrated transistor emitters, and to the alternator source at 240 and to supply that alternator voltage to its respective motor winding at terminal 274. Conventional alternator output rectification may be employed but is not illustrated in Figure 21.

Transistors 270, 276 and 278 function as amplifiers to provide sufficient base drive current to a pair of parallel connected type 2N6258 power transistors 280 and 282. Diode 284 is, as before, a discharge path for the inductive energy which is present in a winding stage when that winding stage is abruptly turned off. In operation, when the output of one of the four NOR gates such as 264 goes high, transistor 272 is enabled to conduct in turn enabling transistors 276 and 278 to their conducting state to supply a base drive current to the pair of parallel transistors 280 and 282, the conduction of which supplies the direct current voltage at terminal 242 to one terminal 274 of a motor winding stage the other terminal of which would typically be grounded.

The A and B signals, as well as their complements are also supplied as inputs to the speed sensing circuitry of Figure 22 and are logically combined by four NOR gates again of a type CD-4001 in a manner such that exactly one of those NOR gate outputs is high at any given instant and each remains high for ninety degrees of shaft rotation thereafter going low and the next NOR gate output going high. These NOR gate outputs are of a rectangular wave form and are differentiated and applied to a transistor 286 for amplification and the resulting sequence of short voltage pulses provided as inputs to an integrated circuit amplifier 288. For example, during the time interval that both the A and B signals are high, NOR gate 290 will just as illustrated in Figure 8 provide a high output pulse and an exponentially decaying spike of voltage will appear across resister 292 due to the initial short circuit and subsequent blocking effect of the charge accumulating on capacitor 294.

This spike is delivered by way of diodes 296 and resistor 298 to the base of transistor 286 and that transistor will conduct for a short time interval to effectively ground the line 300. The periodic grounding of line 300 occurs at the beginning of each rectangular pulse from the gate 290 since the beginning of a pulse provides a positive going spike while the termination of that pulse provides a negative going spike which is prevented from passing to the base of the transistor 286 by diode 296. This periodic grounding of line 300 triggers amplifier 288 which is an operational amplifier in a "one shot" configuration, the output of which is a sequence of square waves of uniform height and duration. This square wave train is supplied to a second amplifier 302 which functions as a filter and provides as an output the speed signal to be supplied to yet another amplifier 304 which is again an operational amplifier, this time connected as a comparator. The output amplifier 302 is compared to the generator voltage as applied to terminal 306 and the amplifier 304 output is either high or low, depending upon whether the speed indicative signal exceeds or is less than the voltage applied to terminal 306. If the signal exceeds the alternator output voltage, the output of amplifier 304 is high, turning on the Darlington configured transistor pair 308, coupling the one alternator field terminal to ground, thereby increasing the voltage output of the alternator. Alternator field terminal 310 is coupled to a battery voltage source and a diode 312 is connected across the alternator field terminals and that diode, in conjunction with the inductance of the alternator field, functions to smooth out the otherwise pulsed field current due to the turning on and the turning off of the transistor pair 308. The width of a single pulse output from the one shot amplifier 288 is constant whereas the frequency of occurrence of those pulses is directly proportional to the frequency of grounding the line 300 which in turn is indicative of rotor speed. Thus, when the rotor speed increases more such pulses are provided to the filter 302 during a given time interval and the output signal (the average of the voltage input level) from that filter is of a higher level. This higher voltage supplied to the positive input of amplifier 304 causes that amplifier output to go high (presuming the alternator output voltage has not changed) to thus cause transistor pair 308 to conduct and to increase the alternator voltage output. A voltage such as the vehicle battery voltage is also applied to terminal 314 to assure some alternator voltage when the motor is at a standstill. Thermostatic control may be implemented as a simple single switch 316 of Figure 21 or more sophisticated control techniques may be employed, for example, by changing the threshold voltage of comparator 304 or other techniques such as discussed in conjunction with Figure 24.

Another exemplary hermetic environment in which the present novel brushless direct current motor finds particular utility is illustrated in Figure 23 where a conventional refrigerator enclosure 318 contains an evaporator coil 320 and a pair of thermostatically controlled contacts 322 which close to actuate the refrigeration system when the enclosure temperature exceeds some preferred value. A compressor 324 pumps refrigerant to a condensor coil 326 where excess heat is extracted and the refrigerant then moves on to an expansion valve or capillary 328 and into the evaporator coil 320. The refrigerant circuit and the cooling of the condensor coil 326 by a fan 330 are conventional; however, the block diagram of Figure 23 is unique in that the system is deployed in a portable or mobile environment and is powered, for example, from a vehicle twelve volt battery 332 and has a hermetic enclosure 334 enclosing the compressor 324 and motor 336 rather than employ the conventional engine driven compressor arrangement typically found in vehicle environments. The electronic commutator 338 may be of the type illustrated in Figure 6 or Figure 21 and a thermostatic control thereof may be implemented as before or as illustrated in Figure 24.

In Figure 24, NOR gates such as 80 of Figure 6 are connected to four substantially identical input terminals, such as 340 and 342 while the output terminals of Figure 24 would be coupled to the bases of four transistors 82 in Figure 6. The exemplary twelve volt direct current source would be coupled to terminal 348 and the contacts of the thermostat 322 function to connect this positive voltage source to the condensor coil fan 330 and to the base of transistor 350. So long as the switch 322 is open, the transistor 350 is maintained in its non-conducting state and transistors such as 352 and 354 receive base drive by way of resistor 356. Conduction by transistors 352 and 354 prevents conduction by transistors 358 and 360, respectively, thereby precluding any winding enabling signals at terminals 344 and 346 (no path for base current in transistor 82). When switch 322 is closed, the transistor 350 is rendered conductive to effectively ground the source of base current for the transistors 352 and 354 forcing those transistors to go to their non-conductive state and allowing the appropriate transistor 358, 360, or other transistor similarly positioned for the other windings, to become conductive when energized by their respective terminals such as 340 and 342 allowing the commutator circuit to function as previously described.

The circuit of Figure 16 illustrates one manner of omitting electro-optical or electromechanical rotor position sensing devices and is particularly suited to the situation where the several motor windings are connected in a half-bridge connection. The circuit of Figure 16 has the resistance 130 connected between the source of voltage for the windings and the several windings. For example, lines 930 and 931 might be connected respectively to the source voltage and to the point marked +V in Figure 6, so all current supplied to the windings will flow through resistor 130. Similarly, the resistance 130 might be connected by lines 930 and 932 in the common connection of the windings to ground in order to sense total winding current and the line 931 might be connected to the source supply as shown in Figure 16 in order to sense the supply voltage with the connection between 931 and 932 being omitted. In either event, the total winding current flowing through resistor 130 develops a voltage thereacross which is applied to the plus and minus terminals of an operational amplifier 136 through resistive elements 132 and 134, respectively. As shown more fully in Figure 6, the stationary armature winding stages 22a, 22b, 22c and 22d are connected in a half-bridge configuration to ground and the resistor 130 may be connected between the center of the half-bridge and ground. A variable by-pass resistor 138 is disposed about the operational amplifier 136. In turn, the output of the operational amplifier 136 is applied through a resistor 140 to one input of an operational amplifier 144, whereas the supply voltage V1 is applied through a fixed resistor 133 and a variable resistor 135 to the other terminal of the operational amplifier 144. Variable resistor 138 and resistor 140 are employed to scale the voltage signal representative of current through the windings in accordance with the resistance of the motor windings, and thus, will vary in accordance with motor design. For small changes in motor size, the necessary scaling may be accomplished by adjusting the variable resistor 138, whereas, with large changes in motor size, the value of resistor 140 may be modified. As a result, the operational amplifier 136 senses the voltage drop imposed across the resistor 130, thereby sensing the total motor current to provide an output proportional to the current I in the stationary armature windings and also to the voltage lost in the motor due to its armature winding resistance R.

This voltage drop may be characterized as the IR drop of the motor. The operational amplifier 144 determines the difference between the supply voltage V1 and the output of the operational amplifier 136 to provide an output indicative of the motor's back emf (V-IR), which is an indication of the speed of the brushless DC motor.

The output of the operational amplifier 144 is applied through a fixed resistor 145 and a variable resistor 146 to a frequency circuit means or voltage-controlled oscillator formed essentially of an operational amplifier 148, a unijunction transistor 154 and a transistor 158.

The output of the voltage-controlled oscillator is derived from the collector of the transistor 158 and is of a frequency proportional to the voltage input and therefore the speed of the brushless DC motor. In particular, the operational amplifier 148 acts as a current source for charging the capacitor 152 through the resistance 150. The capacitor 152 charges until the threshold level of the unijunction transistor 154 is reached, at which time the unijunction transistor 154 is rendered conductive in a forward direction, whereby the voltage stored upon the capacitor 152 discharges through the unijunction transistor 154 and a resistor 155.

As shown in Figure 16, the threshold voltage of the unijunction transistor 154 is set by the values of the resistors 153 and 155, which form in effect a voltage dividing circuit upon which is impressed a supply voltage V1. As the discharge occurs through the resistor 155, the voltage developed thereacross and applied through resistor 156 to the base of transistor 158 rises until the transistor 158 is rendered conductive, thereby dropping the output taken from its collector toward ground potential. Thus, it can be seen that the output as derived from the collector of the transistor 158 is essentially a square wave form, varying at a frequency dependent upon the charging current to capacitor 152 and therefore the speed of the brushless DC motor.

The output of the voltage-controlled oscillator is applied to an indexing means including a first flip-flop 160 whose outputs A and A are complementary square waves as illustrated in Figure 17. In particular, the input signal of a frequency corresponding to the speed of rotor rotation is applied to the input of the flip-flop 160, which divides the frequency of the input signal by two to provide a train of square wave pulses. Further, the flip-flop 160 provides the complement signal A also shown in Figure 17. The A output of the flip-flop 160 is applied to the input of second flip-flop 162 which also divides the frequency of the input by two to provide an output signal Band its complement B as shown in Figure 17. The resulting square wave signals A, B, A and B indicate the speed of rotation of the motor and further simulate the angular position of the rotor shaft as it makes a complete revolution.

More specifically, these signals are considered to simulate the rotor position in that as the rotor begins to rotate, it seeks its own position with respect to the energizing signals applied to the stator windings 22a, 22b, 22c and 22d. Though these aforementioned signals as derived from the flip-flops 160 and 162 doe not precisely identify the rotor position in the same sense that the outputs of the sensors are described in the discussion of Figure 8, these outputs occur sequentially during the rotation and effectively simulate the position of the rotor once the rotor has locked into the stator field.

It should also be noted in comparing Figures 8 and 17 that in the sensorless embodiment of Figures 16 and 17 the winding stages are no longer energized in alphabetical order. The simple expedient of physically interchanging a pair (such as a and d) of the winding connections at the output of the transistors 83 will correct this situation and give the proper alphabetical sequence of energization. Similarly the timing diagram of Figure 17 has assumed the flip-flops 160 and 161 to be of a type wherein the leading edge of the A output of flip-flop 160 triggers the output of flip-flop 162 so that its B level is high or a 1. If flip-flop circuitry which triggers on a trailing edge of the A wave form is employed, the motor will run in a direction opposite that in which it runs when employing leading edge triggering flip-flops if the other connections are unchanged. Again, it should be remembered that the output of the flip-flops 160 and 162 may be processed and applied to the winding stages in the same manner as the A and B signals and their complements as illustrated in Figure 6.

The start-up operation of an electronic commutating circuit disclosed herein will be explained with respect to Figures 16 and 17. Initially, the circuits are turned on by applying the supply voltage V1 thereto. Initially, the rotor of the brushless DC motor is standing still; under this condition, the output of the voltage-controlled oscillator is set to generate an output of a frequency which, for example, corresponds to a rotor speed of approximately 60 rpm, such that as each of the stator winding stages 22a, 22b, 22c and 22d is sequentially energized, at least one of the stationary armature windings will produce a positive torque upon the rotor, thereby initiating its rotation. As rotation occurs, the rotor of the brushless DC motor will lock into the field of the armature. The voltage-to-frequency oscillator is programmed in that the initial output is not set to zero, but at a selected frequency; e.g., corresponding to a rotor shaft of 60 rpm ensuring that the motor is self-starting. The frequency of the output of the voltage-controlled oscillator will remain low until the rotor 10 has locked into the field of the stator windings. With regard to Figure 16, the initial frequency of the output of the voltage-controlled oscillator is determined by setting the variable resistor 149 to a value such that the rotor 10 will lock into the stationary armature field. Thereafter, the speed of the rotor will increase until its running speed is obtained. The rate of frequency increase for a voltage input is determined by setting the variable resistor 135. Thus, the voltage-controlled oscillator is considered programmed in the sense that initially, the output is set to a frequency to ensure that the rotor will lock into the armature field and thereafter, that the speed of the rotor is brought up at a selected rate.

Not only are shaft position sensors eliminated by the embodiment illustrated in Figures 16 and 17, but further, since this embodiment operates basically on a square wave not illustrated in Figure 17 but clearly having a repetition rate twice that of, for example, wave form A, square wave energization of a brushless or commutatorless direct current motor may take other forms. The square wave signal output from transistor 158 in Figure 16 has a frequency proportional to rotor speed and in that particular embodiment for a two pole machine that frequency turns out to be double the rotor speed. Digital or computer control of a direct current motor now becomes feasible and the square wave employed according to the principles set forth in the discussion of the embodiment of Figures 16 and 17 may take other forms.

Figure 18 illustrates one such unique and useful variation wherein during start-up of the motor a signal proportional to motor speed is employed whereas during normal running operation of the motor a signal proportional to the motor load is employed.

In Figure 18, signals which simulate the signals normally obtained earlier from shaft positioned sensors are obtained from a dual output flip-flop that is triggered by a frequency circuit means or a voltage-controlled oscillator. Voltage control for this oscillator is derived from a signal that during start-up is proportional to motor speed and that during running is proportional to the motor load. The flip-flop outputs are fed by way of inverters and also directly into signal conditioning circuits such as 80 of Figure 6. Motor speed is adjusted according to load based on the need for the current flowing in any given winding of the motor to have an essentially square wave form which results in greater motor operating efficiency. Thus, the front and back halves of the wave form are individually sampled, integrated and compared. If they differ from each other, voltage to the voltage-controlled oscillator is changed depending upon this comparison between the leading and trailing half of the wave form and motor speed is accordingly slowed or increased.

In Figure 18, an amplifier 164 senses the voltage drop across resistor 166 which like resistor 130 in Figure 16 carries total motor current. In practice, such resistors as 166 and 130 would be quite small and may, for example, be of the order of a few hundredths of an ohm. The voltage drop signal across resistor 166 is scaled by resistors 546, 548, and 564 in accordance with the resistance of the motor winding, thus the resistance values of the scaling resistors will vary in accordance with motor design. Thus, output of amplifier 164 is proportional to the current in the motor and also proportional to the voltage lost to the motor due to its winding resistance. This amplifier output then is representative of the IR drop of the motor. A similar operational amplifier 166 receives the signal representative of the motor IR drop as one input and the applied voltage as the other input. As before then, the output of amplifier 166 is proportional to the term V-IR which output is an indication of the speed of the permanent magnet DC motor. This speed indicative signal provides one input to amplifier 168 so long as switch 170 is closed. The switch 170 is closed during start-up and until approximately two-thirds of the full load motor speed is obtained, at which time switch 170 is opened and the speed signal has no further effect on the performance of the system.

Switch 172 is ganged to close when switch 170 is closed and open when switch 170 is open.

When these two switches are closed, amplifier 168 performs as an operational amplifier with output proportional to speed. When these two switches are open, amplifier 168 performs as a differential integrator with an output proportional to the load existing at that particular moment. The output of amplifier 168 is supplied to a voltage-controlled oscillator which oscillator includes amplifiers 174 and 176 along with the feedback circuits including transistor 178 and capacitor 180 and their related resistances. The output of the voltage-controlled oscillator from amplifier 176 is a square wave with frequency proportional to the voltage supplied as the output of amplifier 168. The output of the voltage-controlled oscillator is supplied to an indexing means including a flip-flop circuit 182 which is a type CD 4013 AE and provides A and B output square waves with the wave forms being substantially identical to the A and B wave forms illustrated in Figure 17. A pair of simple NOR gates or invertors may be employed to obtain not A and no B wave forms and these four wave forms supplied as before to the A, B, not A and not B lines of Figure 6. The frequency of the A wave form is one-half of that of the voltage-controlled oscillator while the B wave form has a frequency one-fourth that of the voltage-controlled oscillator.

The voltage-controlled oscillator output is also supplied by way of line 184 to a logic chip 186, for example, a CD 4001 AE which includes three NOR gates 188, 190, and 192.

Switches 170 and 172 are logic gates which are closed so long as the output from amplifier 198 is low indicating a relatively low motor speed and which open when the output from amplifier 198 raises indicating, for example, the motor is up to two-thirds of its running speed. The output from amplifier 198 is also supplied to logic chip 186 which, due to the presence of the inverting NOR gate 188 alternately closes switches 194 and 196 by way of control signals on lines 200 and 202 to alternately sample the front and back half of the current wave form signal as seen at the output of amplifier 164. The front and back half signals are supplied as the negative and positive inputs to amplifier 168 respectively which, as noted earlier, functions as a differential integrator with a long time constant and the output of amplifier 168 will increase when the tailing edge (switch 194 closed) is greater than the leading edge of the wave form while with switch 196 closed if the leading edge of the wave form is greater than the trailing edge, the output of amplifier 168 will decrease.

Similarly, if the wave form output from amplifier 164 is approximately the desired square wave form, the output of amplifier 168 will remain constant.

From the circuits disclosed in Figures 16 and 18, it is apparent that further variations may be provided in accordance with the desired application. For example, the circuit of Figure 18 is designed for a fixed advance angle for commutation of the motor winding stages with this angle being approximately 15 electrical degrees in the illustrated embodiment.

However, with either a very light load or a very heavy load, it may be desirable to change the angle of advancement to maximize efficiency. This may be accomplished by providing a bias signal at either of the two inputs of amplifier 168.

Further, it is believed that the disclosed circuitry may be employed for multiple stage motors with modification only being required in the logic circuitry producing wave forms A, B, A, etc. For example, a six stage motor would require generation of six wave forms for sequentially switching the six windings associated with the six stages with each of these wave forms being fed to a signal conditioning circuit such as 80a of Figure 6 which in turn causes actuation of the winding in the same manner as shown in Figure 6. If the circuitry were employed within multi-stage motors, it is believed that it would be more desirable to not have overlapping "on" times between windings, especially in motors with four or more stages in order to properly sense the motor current being produced and relate the current to winding energization.

Still further, if a reversible motor is required, two sets of logic circuitry could be provided to generate the required switching sequence in the different directions of rotation with the sets being alternately connected or disconnected to the voltage-controlled oscillator output depending on the desired direction of rotation.

In brushless DC motors, it may be desirable to determine the initial rotor position relative to stationary armature windings. Figure 40 illustrates one method of determining this initial rotor position. In Figure 40, there is illustrated in cross section, the magnet rotor 10 of Figure 1 having a permanent magnet north pole 13 and a permanent magnet south pole 14. A voltage pulse is applied to one armature winding 1205 by briefly closing switch 1207 to couple the battery 1209 or other voltage source thereto. With the rotor in the position illustrated and assuming winding 1205 creates a north pole in its vicinity, the rotor will move a short distance in the counterclockwise direction changing the flux in winging 1211 due to the north pole created by magnet 13 moving in its vicinity and inducing a voltage in that winding 1211 which may be sensed, for example, by galvanometer 1213. Any pair of motor windings could be selected for test pulse application and induced voltage sensing and the polarity of that induced voltage will give an indication of rotor position. In some cases, the results of the test pulse method may be ambiguous, for example, if the proper rotor magnet is directly under the test pulse winding, no rotor motion will occur.

The proper selection of a different winding and application of a second test pulse will resolve such ambiguities.

In the previously discussed Figure 6, rotor position signals are developed by the position sensors 43, 44 and fed into signal conditioning circuit 701 which is arranged to produce four position control signals, A, A, B and B which are utilized to control current switching in the stationary armature winding stages. These position sensors 43, 44 can be eliminated by the teachings of the present invention, one form of which is illustrated in Figure 41. Sensor substitute signals are provided as the outputs of NAND gates 1043 and 1045 and the two-stage orthogonally positioned motor windings provide voltage input signals to terminals 1047 and 1049. A shift register 1051 which is connected as a four bit ring counter identifies which pf the four winding stages a, b, c or d is the currently energized winding and the voltage induced in a winding not currently energized is sampled by enabling one of the two switches 1053 and 1055. Frequently the winding stage which is sampled is the winding stage next in sequence to be energized. The induced voltage sample is integrated by an integrator 1057 and compared to a reference voltage 1059 in the comparitor 1061. When the integral exceeds the reference voltage, the comparator output goes high and a differentiating circuit 1063 increments the shift register 1051 to its next indication. Any change in the high bit position of the shift register 1051 is sensed by NAND gates 1065 and 1067 which by way of inverters 1069 and 1071 and a further NAND gate 1073 triggers the one-shot timer 1075 to reset the integrator 1057 to its initial conditions for the next integrating cycle. An initial condition interval is also established by the one-shot timer 1075 which interval is not only sufficiently long to reset the integrator 1057 but further eliminates switching transients from the calculation and insures that the induced voltage due to a collapsing magnetic field from a winding being disabled is not included in the computation. Only two voltage sensing terminals 1047 and 1049 are employed and the voltage across only two of the four windings illustrated in Figure 6 is sensed. To insure that the same sense or polarity of winding voltage is sensed each time, an inverter 1077 along with alternately enabled swiches 1079 and 1081 are provided. These last mentioned switches are alternately enabled by the output of NAND gate 1043 and inverter 1083. It should be noted that a change in rotor speed changes the time of integration but has no effect on the overall result and accordingly the integrator output is representative of rotor position or total flux change rather than rotor speed or the flux rate of change.

Wave forms associated with the circuit of Figure 41 are illustrated in Figure 42 with the short initial condition (IC) pulse 1085 being the initial condition signal to the integrator as supplied by way of inverter 1087. The output of integrator 1057 is illustrated for a proper or preferred "brush position" in the second wave form whereas switching too late and switching too early respectively lead to the third and fourth integrator wave forms as illustrated. Considering the "too late" wave form, it will be noted that the integrator output will achieve its reference voltage value sooner in time than for the optimum situation in which case, of course, shift register 1051 is incremented earlier, compensating for the "too late" situation. Bias input 1089 to the integrator 1057 is provided to sequence the switching when no counter emf is present on terminal 1047 or 1049, i.e., when the motor is at standstill. This bias 1089 functions to make the circuit act as though the motor were running at a slow speed in the desired direction and materially aids the starting of the motor. It should also be noted that rather than creating the A and B signals as were employed in sensor type brushless motors, the outputs from shift register 1051 could be employed directly for enabling winding energizing circuitry, such as the Darlington pair 82, 83 and the input transistor 81 for each winding as illustrated in Figure 6.

While the wave forms of Figure 42 represented the integral of the voltage across a winding stage not at the time energized, but for example next in the sequence to be energized, the wave forms in Figure 43 represent the current flow through an energized winding stage with the upper wave form thereof illustrating a heavy load or early commutation situation while the lowermost wave form illustrates a light load or late commutation situation with the intermediate wave form being the optimum "brush position" or commutation time wave form. The proper commutation time wave form corresponds to the relative positioning of an exemplary coil 1091 and the rotor flux pattern 1093 which is relatively uniform throughout its duration as illustrated in Figure 44. If the coil of Figure 44 were located to the right of the position shown in Figure 44 the situation of a heavy rotor load or early energization of the coil 1091, a peak in current would develop on the front edge of the conduction interval as illustrated in the upper wave form of Figure 43 and would correspond to the integrator output illustrated in the lowermost wave form of Figure 42.

As noted earlier, the ring counter, which as input signals are counted, has one specified "1" state which moves in an ordered sequence about a loop, may be used directly for energizing winding enabling circuitry and this is done in the circuit of Figure 45 which illustrates a four stage or four winding motor where those windings are energized in a sequence four, three, two, one, four, three, etc. The ring counter 1095 directly provides those winding enabling signals identified as (1), (2), (3) and (4) and also those signals are coupled to switches 1097, 1099, 1101 and 1103 in a sequence (2), (3), (4), (1) such that the next to be enabled winding voltage is sensed by the appropriately enabled switch. For example switch 1101 couples winding number 3 to amplifier 1105 when the fourth stage of ring counter 1095 is providing the output signal (4). These sensed winding voltages are amplified by an amplifier 1105 and pass through a half wave rectifier 1107 which is included to prevent integrator saturation during possible large negative values of the integral. Those half wave rectified signals then pass through integrator 1109, amplifier 1111 and are compared in comparitor 1113 to a reference 1115 and, when the integral exceeds the reference, one-shot timer 1117 functions by way of inverter 1119 to reset the integrator and, by way of differentiating circuit 1121, to increment the ring counter 1095.

The schematic diagram of Figure 25a and 25b and the associated wave forms as depicted in Figure 26 illustrate another approach to an electronically commutated direct current motor which does not employ rotor position sensors as such, but rather senses the back emf of an unenergized winding stage. Conversely, this may be the winding stage next to be enabled in the winding stage energizing sequence. With this approach, the emf of each winding stage is integrated from a zero crossing point until a predetermined number of volt-seconds is accumulated. The integration circuitry ignores negative values of emf and assures accurate control of the advance angle of commutation independent of speed of the rotor. A starting aid is provided to assure initial rotation of the motor. Further control features are provided as discussed hereinbelow including protection circuits for inhibiting motor rotation during undervoltage, overvoltage, underspeed, and/or reverse voltage polarity conditions.

This circuitry may be employed in conjunction with a two-stage, two pole motor with bifilar windings for each phase which motor lacks the position sensors but is otherwise quite similar to the motor depicted in Figure 1. In addition, and importantly, this type of arrangement may be used with a three stage, four stage (and even higher) motor. Current from a battery or other direct current source is coupled from the plus V terminal to the individual motor winding stages by transistors such as 362 and 364 which are connected in a modified Darlington configuration with one such pair for each motor lead. Some grounding problems in the logic circuitry may be avoided in, for example, a vehicle battery powered refrigeration system if the plus V terminal is connected to the positive battery terminal and sequentially to each positive motor winding lead with all the negative motor winding leads connected together and to the negative side of the source. Feedback diodes such as 366 as before provide a current path in one winding stage for energy stored in another winding stage at the time the current in that other winding stage is switched off. The capacitor 368 as before functions as a sink for this energy, for example, when the source is inadvertently opened or when the source is other than a battery. Each winding stage and transistor pair are enabled one-fourth of the time and a protection means such as diode 370 may be included to provide against inadvertent reverse polarity connections. In the illustrated arrangement, damage to the commutation circuit and motor is prevented since current flow through the diode 370 due to a reverse polarity situation will cause failure of the fuse 372.

The direct current source plus V in addition to being coupled sequentially to the motor winding stages is applied to terminal 374 as an energy source for the output driver transistor such as 376 and to terminal 378 where it is processed by the zener diode filter network to provide at terminal 380 a regulated voltage Vr of, for example, 8.2 volts for use as a voltage source to the logic packages and operational amplifiers.

The logic circuit in general functions to sequentially drive the power transistors such as 362 and 364 and to initiate the drive signal when the rotor mounted permanent magnets of the motor are optimally located with respect to the winding to be turned on. This optimal location or desired advance angle alpha (a), (see Figure 3 and attendant discussion) is derived from the back emf voltage of the winding next in the drive sequence.

In operation, detector circuit 814 shown within dotted lines in Figure 25a and connected to each of the winding stages selectively senses the emf voltage of a particular winding stage with such selection in this particular embodiment being the winding next to be energized.

The detector circuit then transmits the sensed emf voltage of the particular winding stage to a position determining circuit 816 wherein the emf voltage is conditioned and employed to produce a simulated relative position output signal which simulates or is indicative of the relative position of the rotor and armature of the motor. In this particular embodiment, such simulated relative position output signal is in the form of pulses with the frequency or the time frame in which the pulses occur being indicative of the relative position of the rotor and armature. In order to select the winding stage next to be energized in accordance with the simulated relative position output signal, a circuit means 818 processes the incoming simulated position signal through an indexing or sequence arrangement to produce an output signal for triggering energization of the particular winding stage. The simulated relative position output signal and the circuit means output signal also cause the detector circuit to select or switch to another winding stage for sensing the emf voltage which in this embodiment is the winding stage which is to be energized next in sequence. The emf voltage from this other winding stage is again conditioned and processed by the position determining circuit for obtaining the optimal location or desired advance angle for energization of the other winding stage whereupon, another simulated relative position output signal is produced and fed to the circuit means. The circuit means again indexes to cause the previously energized winding stage to be deenergized, to cause energization of the other winding stage and to cause the detector circuit to switch to still another winding stage.

This operational procedure occurs continuously during motor operation with the emf voltage of each winding stage being selectively sensed and employed to selectively energize each of the winding stages.

In somewhat more detail, the position determining circuit of Figure 25a includes an operational amplifier 382 which functions to integrate this back emf voltage and when this integrated voltage reaches a reference level or a predetermined number of volt-seconds, a simulated relative position output signal is produced causing the succeeding logic parts of the circuit means to change state and to sequence or index to the next winding energizing event. In the process, the drive signal to the "on" transistor from drive transistors such as 376 is removed and redirected to the next Darlington pair or output transistor to be energized. In order to sense the emf voltage of the winding stages and thus, control their energization sequence, the motor leads 384, 386, 388 and 390 at the extreme right of the Figure 25b are connected to like numbered terminals at the left-hand of the Figure 25a for sequentially sampling the emf voltage of the winding stages by way of switches 392, 394, 396 and 398 of the detector circuit. These switches are sequentially enabled by NOR gates such as 400 having the simulated relative position signal and the circuit means output signals as inputs and which are interconnected with the switches by way of terminals 402,404, 406 and 408. However, winding energization is precluded thereby preventing further motor operation if the motor supply voltage is outside a predetermined voltage range or the motor speed is below a predetermined minimum low speed. Undervoltage circuit 820 and overvoltage circuit 822 having the operational amplifiers 410 and 412 respectively, as their nucleus assure that the supply and voltage does not fall below a predetermined minimum value nor exceed a predetermined maximum value. An underspeed circuit 824 structured about the operational amplifier 414 and logic inverting gates 416, 418 and 420 assures that motor operation is above a predetermined minimum speed. These circuits, in the event of a fault, that is, a violation of the above-mentioned predetermined limits, cause transistor 422 to go non-conducting thereby preventing any enabling current flow in the output driver transistors such as 376 and hence no winding energizing current through the output transistors such as 362 and 364.

In order to produce the output signal indicative of the relative position of the rotor and armature when the operational amplifier 382 has integrated the emf voltage to a predetermined reference level, the position determining circuit is provided with inverting gates 442 and 444 which are coupled together with feedback to form a Schmitt trigger circuit and interconnected with NAND gates 446 and 448 which are themselves connected together to form a one shot multivibrator. The output of the one shot multivibrator passes through an inverting NAND gate 450 and into a pair of flip-flops 452 and 454 to provide the Q1 and Q2 signals along with their complements to be logically combined by a first decoder comprising four NAND gates such as 456 and a second decoder comprising four NOR gates such as 400. The NAND gates may be of a CD 4011 type while the NOR gates may be of a CD 4001 type. The concatenated invertors 416 and 418 which form the Schmitt trigger circuit may similarly be type CD 4001 while invertors such as 458 may be a type CD 4049.

The wave forms of Figure 26 illustrate the steady state or running mode of the circuitry of Figures 25a and 25b with corresponding reference numerals to the right of the wave form identifying the line in the schematic diagram where that wave form occurs. Somewhat idealized back emfs of the motor winding are shown at the top of Figure 26 and the object of the circuit will be to commutate with a predetermined amount of advancement, i.e., to switch from one winding stage to the next when the rotor magnet is within ten to fifteen electrical degrees of fully coupling the winding stage to be turned on. This is shown as time or point A in Figure 26. This triggering point is determined by integrating the back emf in the amplifier 382 beginning at point B which is the zero back emf point or zero crossing point. The voltage is integrated for a period of time and is a measure of the flux change rather than a function of the motor speed. When this integration is completed and the triggering point A achieved, the Schmitt trigger output 428 is actuated. This integration interval between points B and A is preceded by a two millisecond reset period in this embodiment during which capacitor 460 charges back to the reference voltage of, for example, 6.8 volts as determined by the zener diode 462. Integration is also preceded by another period of a duration determined by the motor speed and physically that time interval during which the back emf is negative. The zener diode functions to prevent integration of the negative emf voltage so that the integration interval starts at the zero crossing point of the emf voltage or point B as illustrated in Figure 26. The triggering point for the Schmitt trigger circuit is adjustable by potentiometer 464.

Integration from point B or the zero crossing point aids initial starting of the motor and assures accurate control of the advance angle of commutation independent of the speed of rotation of the rotor. Initial starting is aided in that a reverse motor direction causes an emf voltage of a relatively large positive polarity; thus, integration is quickly completed causing a fast switching to a next winding in the desired sequence. This switching in the desired sequence causes the winding stages to produce a rotating magnetic field in the desired direction for proper rotation. This integration and rapid switching continues until the motor is rotating in the proper direction and being commutated at the proper time. Further, because negative emf voltages are ignored, integration does not begin simultaneous with the switching of the detector circuit. Therefore, if a fixed delay is required for suppressing noise and switching transients and/or resetting the integration circuit, this delay will occur when the emf is negative, and thus, would not cause variance of the advance angle with the speed of the motor.

The Schmitt trigger output actuates the one-shot multivibrator which provides an output change for a two milli-second interval as determined by the specific value of capacitor 466.

This output change activates a means provided to reset the emf voltage integrator. In the illustrated arrangement, the resetting means NAND gate 468 and switches 470 and 472 with the NAND gate 468 being responsive to the output of the multivibrator for turning on switches 470 and 472 to reset the operational amplifier 382. The one-shot output is also inverted in NAND gate 450 to produce the simulated relative position output signal which is further processed by the flip-flops 452 and 454 to produce signals suitable for sequentially turning on the Darlington power transistors. NAND gate 468 also assists motor starting since when the motor and circuit are first energized, capacitor 450 is discharged and the one-shot multivibrator has a high output on line 430 and would normally remain with that high output until capacitor 450 is charged and an integration cycle completed. The Schmitt trigger, however, has its output initially low and that output remains low long enough for a hig reached; whereupon, an output signal may be generated to index or sequence the flip-flops 452 and 454 to produce signals suitable for sequentially turning on the Darlington power transistors. Still further, the computer or micro-processor could also be programmed to perform the decoding process so that output signals could be generated to directly turn on the power transistors for sequentially energizing the motor winding stages.

Although in the above-described embodiments, Darlington power transistor arrangements have been employed to perform the power switching of the winding stages, persons skilled in the art will recognize that alternate modes of switching may be employed. For example, silicon-controlled rectifiers (SCRs) or thyristors may be used with an arrangement being provided to reset the SCR or thyristor controlling energization of a particular winding stage when the circuit means generates an output signal to energize a next winding stage in the energizing sequence. Still further, in certain limited applications, relays may be employed to switch power to the windings.

As noted previously, winding energization is precluded thereby preventing further motor operation if the motor supply voltage is outside a predetermined voltage range or the motor speed is below a predetermined minimum value of low speed. It was also noted that transistors such as 376 will respond to high signals sequentially generated by gates such as 458 only if transistor 422 is conducting. However this transistor 422 will conduct so long as the cathodes of diodes 476, 478, 480 and 482 are all at the Vr reference voltage and the current flow through resistor 484 is essentially the sum of the base to emitter current in transistor 422 plus the current flow in resistor 486. Grounding or lowering the potential on terminal 488 would disable transistor 422 while on the other hand raising that voltage to the Vr reference level or above will, under normal conditions, allow transistor 422 to conduct.

Similarly, if the output of operational amplifiers 410 and 412 of the undervoltage circuit and overvoltage circuit, respectively, is low, transistor 422 will normally conduct. Both operational amplifiers 410 and 412 function to compare the battery voltage V as applied at terminal 490 to the zener regulated reference voltage as applied at terminal 492, however, the reference voltage is applied to the negative terminal of amplifier 410 of the undervoltage circuit and therefore that amplifier output will be low so long as the fraction of the battery voltage a plied to its negative terminal is greater than the fraction of the reference voltage, as determined by the setting of potentiometer 494, which is applied to its positive terminal. Thus, operational amplifier 410 has a high output so long as, for example, the battery voltage is above 10.5 volts and functions as an undervoltage or low voltage detector. Similarly, operational amplifier 412 is set by the appropriate choice of voltage dividing resistors to have a high output so long as battery voltage is below, for example, 16 volts and this operational amplifier functions as an overvoltage or high voltage detector.

The capacitor 496 is present to filter transients and erratic wave forms from a battery charging device to prevent false indications of under or overvoltage.

Numerous further control functions may be implemented employing the electronically commutated motor control circuit of Figures 25a and 25b by hanging additional diodes to the base of the transistor 422 to divert that transistor's enabling current when the diode is biased to conduct thereby disabling transistors 376 and therefore also the winding enabling transistor pairs. For example, the underspeed circuitry illustrated in Figures 25a and 25b and connected to diode 482 will function to disconnect the winding stages when for some reason the motor is running at an inordinately slow speed. The output of gate 450 is a rectangular wave which varies with motor speed and which is on about one-half of the time at normal running speed. In this particular embodiment, the running speed was approximately 3600 rpm and the output of gate 450 was approximately a 240 hertz rectangular wave. This signal is filtered by resistor 498 and capacitor 500 and thereafter amplified by amplifier 414 to be again filtered by the resistor 502 and capacitor 504. The resulting direct current voltage which is substantially proportional to speed is stored as a voltage on capacitor 506 and, so long as that voltage is at or above the level representative of a predetermined minimum speed, for example 2500 revolutions per minute in the illustrated arrangement, the output of the Schmitt trigger comprising amplifiers 416 and 418 remains high. Calibration of this voltage or this minimum speed level may be achieved by changing potentiometer 508. If the motor speed becomes too low, the Schmitt trigger circuit changes to its low state allowing current flow through diode 482 and simultaneously providing a high output from the amplifier 420 to charge capacitor 506. The time constant for capacitor 506 and resistor 510 along with the hysteresis of the Schmitt trigger circuit 416, 418 determines a reset time for the circuit and several minutes such as 4 to 5 may be involved.

At initial start-up of the motor, this same 4 to 5 minutes must elapse before the start can occur. The capacitor 504 will typically retain its charge sufficiently long for normal on-off duty cycle functioning of, for example, a refrigerator motor, however, if a restart is made and the exemplary 2500 revolution per minute speed is not achieved within say three to five seconds, the time constant of capacitor 506 and resistor 512, the start will be aborted and the five minute delay for recharging capacitor 506 initiated prior to a restart attempt. The relatively long charging time for capacitor 506 and comparatively short discharge time is of course due to the presence of diode 514 and the substantially lower resistance of resistor 512 as compared to resistor 510.

In the exemplary vehicle refrigeration environment, a condensor cooling fan would typically be coupled across diode 516 and that diode would function to conduct energy stored in the inductance of the fan motor. In other environments without such a fan motor a resistance would be substituted for the diode 516. Diodes 518 are included to give an additional small voltage drop to the base of transistor 422 since, in practice, the low outputs of amplifiers such as 410 and 412, may not be exactly zero.

* As noted earlier, the fact that the brushless direct current motor of the present invention may be enabled by rectangular wave forms which are processed by logic circuitry makes possible a number of motor control embodiments employing digital control techniques.

One particularly unique digital application of the electronically commutated motor of the present invention is illustrated in Figure 27 and may be employed, for example, for moving and precisely positioning a linearly movable element. The brushless direct current motor 520 has rotor position sensors 522 and in this environment a six pole machine employing GECOR (cobalt samarium) permanent magnets was employed to meet the lower operating speed and size constraints. Also, in this particular embodiment, electro-magnetic sensors are employed and the 20 kilohertz exciter source provides a signal for these sensors signal exciter coils. In the linearly movable element environment, the motor 520 by way of a screw shaft drives the element and accurate control of the motor results in accurate positioning of the element. Use of the direct current motor of the present invention has many advantages over the conventional approach to linearly moving elements in that typical gear boxes and stopping brakes as well as safety clutches are eliminated and the locking torque of the motor may serve as a holding brake. Further, the system is uniquely suited to battery operation and is easily controlled by a process computer for forward, reverse, stepping or braking commands.

These sensors 522 are of course physically located with the motor 520 and stationary exciter and pick-up coils are sequentially coupled and decoupled from each other by a rotating segmented disc or shutter supported on the motor rotor shaft. The exciter coils are energized by the 20 kilohertz exciter source 524 and the pick-up coil signals processed in the decoder 526 to produce uniform rectangular voltage wave forms for subsequent processing.

The output of the sensor decoding circuitry 526 is supplied to stepping logic circuitry 528, counter logic circuitry including the forward-reverse sensor logic 530 and a position counter 532, the contents of which may be digitally displayed employing typical seven segment display units in the display panel 534. The logic circuit 528 is provided with logic circuitry which inhibits normal commutation of a next winding stage in the energization sequence, thus causing continued energization of the same winding stage until the rotor proceeds to a position where the torque per ampere output is zero. The rotor remains in this position until a signal is supplied to cause commutation of the next winding stage in the energization sequence which causes another step in the same manner. The output of the sensor decoding circuit 526 is also supplied to mode control circuitry 536 which sorts out forward, reverse, forward step, and other commands and relates those commands to the sensor signals to compile that information in four logic gates. There is one such gate for each of the four transistor power switches and at any given time only one of those gates will have a high output. These outputs are modified in another set of gates which accept the run and stop commands as well as current limiting information. Current limiting is carried out as a pulse width modulation process and the output of these last gates is amplified in two stages of transistor power switches 538 and 540. Motor current may be sensed at 542 and if that current is excessive, gates in the pulse modulating circuitry 544 are inhibited for a short time period, such as 500 micro-seconds to inhibit the power supply and driver circuitry 544 for a like time period allowing the motor current to decay somewhat. Of course, by limiting current to a maximum value, the torque output of the motor which is a function of the current is also limited or controlled to a maximum value with the above-described arrangement. Torque control may be desirable under certain circumstances to protect the mechanical system driven by motor.

It will now be apparent that different embodiments of the invention, in preferred forms thereof have been shown and described. At the present time, the half-bridge connection arrangement utilizing bifilar winding stages in the two winding stage motors and monofilar in multiple winding stage motors is believed to be the better mode as compared to a monofilar/full-bridge circuit arrangement.

This is because less transistors are required for the half-bridge approach and therefore less expense is involved even though less efficient utilization of winding material (e.g., copper or aluminum) results. On the other hand, if and when the relative expense of transistors and winding material changes in favor of solid state devices, the monofilar/ full-bridge circuit would be preferred.

When either approach is followed, it is definitely preferable to provide energy storage means (e.g., as described hereinabove) in order not only to protect the output transistors but to improve the efficiency of energy utilization.

When either approach is used, methods embodying other features of the invention may be practiced, of course. For purposes of summary, such methods relate of course to the manufacture of brushless DC motors (whether or not they are of the electronically commutated variety), and include: the selection of AC induction motor types of cores; the development of distributed windings in slots of such cores by means of available AC induction motor equipment to form wound stator assemblies; and the assembly of such wound stator assemblies with permanent magnet rotor assemblies.

The winding turns may be wound and established (concurrently or sequentially) in coil receiving means and then axially inserted into the axially extending core slots (either directly from the coil receiving means or from an axial inserting means to which the winding turns are transferred from the coil receiving means).

It should be recognized that the methods just briefly summarized represent a departure from the art of making DC motors. For example, prior techniques have involved the formation of what are known as "ring" windings (e.g., wave or lap windings) which are disposed on a conventional DC dynamoelectric machine core.

Further, there has been shown and described a simplified circuit for commutating the energizing signals applied to a brushless DC motor. More specifically, the circuitry of this invention does not require a mechanical sensor assembly coupled to the rotor of the brushless DC motor; instead, the output of the motor is sensed and is used to generate a varying signal of a frequency corresponding to the rotational velocity and indicating the position of the rotor. Thus, in certain applications such as for use with a brushless DC motor for driving a refrigeration compressor, the compressor shell does not require additional through leads, thereby improving the hermetic seal of the shell.

To better reveal the improved characteristics of motors embodying the invention, the data of Table I is presented.

TABLE I Data taken at 2600 rpm Advance, alpha in Electrical Torque, Oz. Ft. Efficiency, Degrees Total Net Total* Net** 0 1.15 .95 73.7 60.9 5 1.20 1.00 78.2 65.2 15 1.50 1.30 81.2 70.4 22 1.60 1.35 77.7 65.6 *without regard to windage and friction losses **including windage and friction loss The data of Table I was obtained by testing one motor that embodies the present invention and that was operated from a 12 volt DC supply.

The motor utilized a standard stator lamination design that is used commercially in induction motor applications. The lamination was substantially identical to the lamination shown in Figure 2. The bore of the core was about two inches with a stack height of about two inches. The core had 24 slots, and carried distributed copper magnet wire winding side turn portions that were bifilar wound. Eight coil groups total (four bifilar coil groups) were used. Each coil group included three coils and each coil comprised from outermost coil to innermost coil respectively: 7, 10, and 10 turns. The wire was about .05 inch diameter (uninsulated) copper wire. The coils of each coil group spanned, from outer to inner coils respectively, 11, 9, 7 teeth respectively. Thus, the "spread" of an associated pair of coil groups was six slots or 90 mechanical degrees. It thus will be understood that eight slots contained 14 conductors (seven bifilar conductor pairs), while the rest of the slots had 20 conductors each.

The rotor magnets were formed of ferrite magnet material from Allen Bradley Company and designated as M-7 material. The arc length of each of the two magnets used was 143 mechanical degrees; the thickness was about .25 inches; and the axial length was about two inches. The magnets were epoxy bonded to a solid, soft iron rotor core and the assembled rotor had an outer diameter of about 1.98 inches. Commutation and sensing was accomplished with circuits substantially identical to those shown herein. The actual circuit components (i.e., transistors, resistors, capacitors, etc.) were commonly available types and were selected to have only sufficient voltage and current ratings and gain to supply up to 30 amperes to the motor windings. The motor was commutated with from zero to twenty-two electrical degree advance. Since the motor was a two-pole motor, electrical degrees were of course equal to mechanical degrees.

In Table I, two efficiency and torque columns are recorded. The total torque represented the torque produced by the motor without regard to windage and friction losses. The first "efficiency" column also was the efficiency of the motor without regard to windage and friction losses, although copper and commutator circuit losses were allowed for. Net torque was the net torque available at the shaft of the motor, and net efficiency was the overall efficiency of the motor system including the commutator. The significant reduction in net efficiency (due to windage and bearing losses) was expectable because the motor tested was only about 1/20 horsepower.

Table I does show the significant improvement in efficiency and torque that is attainable by advancement of commutation. Thus, a 15 electrical degree advance would provide significantly more maximum efficiency at 2600 rpm; whereas an advance of 22 degrees would provide significantly greater maximum torque at 2600 rpm.

TABLE II Representative component values or element identification for the foregoing circuits.

Reference No. Component Figures 6 and 7 70', 80 CD 4001 AE 43, 44 H 13 A2 71, 72 390 K 74 1.8 K 92 39 OHM 2W 93 200 MFD 300V 94 100 MFD 95 15V 81 2N 4401 82 2N 5988 83 2N 6258 84 10 K 86 150 OHM 88 A 15 91 500 or 1000 MFD Figure 16 130 .02 OHM 132, 133, 140, 736, 145, 740 100 K 134 120 K 135, 734, 149 50 K 138 3.5 MEGOHM 146 1 MEGOHM 738 82 K 150 2.2 K 153 1K 155 47 OHM 156 120 K 742 33 K 136, 144, 148 MC 3301 P 160, 162 CD 4013 AE 158 2N 3414 154 2N 1671 Figure 18 186, 170, 172, 194, 196 CD 4016 AE 182 CD 4013 AE 164, 166, 174, 176 LM 324 168, 198 LM 3900 166 .02 OHM 546, 548, 550, 552, 554, 556, 558, 560, 562 10 K 564 120 K 566 4.7 K 568, 570, 572, 574, 576, 578 100 K 580, 582 150 K 584 1 MFD 180 .01 MFD 586 22 K 588 1 MEGOHM 590 270 K Figure 19 204 Two parallel .10 OHM 2W 208 10 V 702 1.2 MEGOHM 704, 706 100 K 214 Series 100 K & 200 K Variable 708, 710 1 MEGOHM 210 MC 3301 P 712 82 K 222 390 MFD 714, 716 47 K 220 .01 MFD 718 3.9 MEGOHM 218 Series 33 K & 500 K Variable 224 1N914 Figures 21 and 22 246 H13 A2 652, 654 39 K 656, 660, 670 1 K 658 560 OHM 256, 258, 260, 262, 290 CD 4001 264, 266, 268, 270 CD 4001 272, 286 2N 4401 662 47 OHM 2W 664 5000 MFD 666,678 10 K 668 10 OHM 10W 672 100 OHM 674, 676 .03 OHM 284 A 115 294, 680, 682 .001 MFD 292 100 K 296 1N914 298 33 K 288, 302, 304 MC 3401 308 2N3414 and D44H5 312 A 15 684, 686 22 K 688, 691 1 MEGOHM 690 2.2 MEGOHM 692, 694 100 K 696, 700 1 MFD 698 27 K Figure 24 720, 722 8.2 K 356 2.2 K 350 2N 3414 724, 726, 728 10 K 730, 732 82 OHM Figures 25a and 25b 592 10 K 594 680 OHM 474 5.6 MEGOHM 596 150 K 598 10 K 462 1N4736 6.8V 460 .1 MFD 382 CA 3130 600, 602 10 K 464 100 K 604 1 MEGOHM 606 150 K 608 33 K 446, 448 CD 4011 466 .01 MFD 610 1N5059 612 10 K 614 25 MFD 25V 616 1N4738 8.2V 618 120 OHM 1W 620 33 K 622 6.8 K 624 12 K 626, 628, 508, 636, 638 100 K 494 20 K 498, 632 33 K 500, 504 .5 MFD 630 .1 MFD 410, 412, 414 LM 324 502 15 K 633 1N4448 634 1.8 MEGOHM 640 4.7 MEGOHM 452, 454 CD 4013 456 CD 4011 400 CD 4001 458 CD 4049 642 2.7 K 644 100 OHM 2W 646 100 OHM 376 2N3414 362 2N5988 364 2N6258 366 1N5059 368 1000 MFD 25V 370 MR 751 480, 514 1N4448 484 6.8 K 486 2.7 K 648 1.5 MEGOHM 650 4.7 MEGOHM 510 2.2 MEGOHM 512 68 K 416, 418 CD 4001 506 50 MFD Figure 27 542, 538, 540 Two parallel .1 OHM 10W STV 6060 with 1N 5625 and V150 PAlO in parallel collector to emitter 544 2N 3414 driving D45H8 536 CD 4011 543 CA 3130T and CD 4012 as inputs to CD 4011 then to CD 4049 invertor output 528 CD 4042 driving CD 4011 then to CD 4012. Also CD 4001, CD 4011 and CD 4029 524, 526 CD 4011 input to CD 4030, CD 4011 and CD 4001 in series Also three 2N 3414 530 CD 4042, CD 4011 and CD 4012 in series 532 CD 4001 driving CD 4011 and also CD 4029 534 MC 14511 to MAN54 The solid state components listed in Table II hereinabove, with the exception of transistors STV 6060 and displays MAN54 (see Figure 27) were either RCA, G.E., National, or Motorola devices. The four transistors STV 6060 were TRW transistors, and the four displays MAN54 were Monsanto displays.

The present invention has been described thus far, primarily in connection with a two-stage, two pole motor of the type illustrated in Figures 1 and 2, however, as mentioned previously, the invention is equally applicable to multiple stage motors such as three, four, five, etc. stage motors with a varying number of poles.

In carrying out the invention with a three winding stage motor, the same factors considered with respect to the two-stage motor must again be considered. For example, in order to maximize efficiency and avoid starting problems, consideration must be given to such factors as torque per ampere (T/I) curve or motor operating characteristic, winding spread, rotor permanent magnet arc length, commutation advancement, magnetic coupling, energy recovery from a winding after deenergization, slot configuration of a stationary armature, and slot utilization. In addition, in the preferred sensorless approach, the rotor position must be reliably and accurately simulated for purposes of commutation.

Figures 28, 29 and 30 show winding configurations in twenty-four slot stationary armatures for three winding stage, two pole, four pole and eight pole motors, respectively.

The illustrated stationary armatures are monofilarly wound in the uniform slot configura- tions and have all slots filled. However, for greater clarity, the armatures are illustrated with only one single turn coil in each slot. It is preferred to fill all slots of the armature in order to minimize core thickness or stack height of the stationary armature and thus making the motor more compact in size and more economical to manufacture. However, as discussed hereinbelow, the stationary armature can be employed with empty slots with a resulting reduction in rotor permanent magnet arc length but at the expense of increased armature core.

Referring to Figure 28, the illustrated armature 830 for a three winding stage, two pole motor comprises three monofilar winding a, b, c (one for each stage). Each winding stage comprises coils formed from concentric turns of a conductor that are disposed in the core 831 with the side turns of the coils forming two winding sets in the core slots. The conductor portions of each set conducting current when the winding is energized, in the same axial direction along the length of the core. For example, the winding stage "a" comprises two winding sets, 832, 833 with winding set 832 being disposed in core slots 834-837 and with winding set 833 being disposed in core slots 838-841. The winding stage "a" was formed by first making a predetermined number of concentric turns of a conductor to form a coil whose side turns occupy core slots 834 and 838 when inserted into the core slots with conventional coil insertion equipment. The winding of the conductor continued with a predetermined number of concentric turns of a different diameter to form a second coil whose side turns occupy slots 835 and 839. The winding of the conductor continued with a predetermined number of concentric turns being made to form a third coil whose side turns occupy slots 837 and 841 as shown after being disposed in the core. A fourth coil was also formed by making a predetermined number of concentric turns of the conductor so that the side turns of the coil occupy slots 836 and 840 as shown when disposed in the core.

Even though the coils have been described as having been formed in a particular order, the coils may be formed in any order, but the winding must be such as to assure that when the winding stage "a" is energized, all the conductor portions within a winding set conduct the current in a common direction along the axial length of the core. As shown in Figure 28, all the conductor portions of winding set 832 conduct the current in a common direction ("x" indicating current flow into the page) and all the conductor portions of winding set 832 also conduct current in a common direction ("." indicating current flow out of the page) which is in a direction opposite the direction of current flow in the set 832. This current flow in winding stage "a" as illustrated creates magnetic poles Na and Sa. Because all the conductor portions of each set instantaneously conduct current in the same axial direction along the length of the core, the spread of winding stage "a" can be measured at either winding set 832 or set 833. As illustrated, the spread of winding stage "a" is 60 (electrical and mechanical) measuring from the centerline of the slot tooth separating set 832 from winding stage "b" to the centerline of the slot tooth separating the set 832 from winding stage "c".

Winding stages "b" and "c" for the three stage, two pole motor of Figure 28 are wound and disposed in the armature in the same manner as described above for winding stage "a".

Although all the coils of each winding stage in Figure 28 are shown as being formed in a continuous winding process, i.e., without cutting the conductor between the winding of different coils, the coils may be formed separately or in groups of two or more and then interconnected to create a single winding stage with the coils being disposed and interconnected so that current flow in each winding set is in the same direction.

Figure 29 illustrates a stationary armature for a three winding stage, four pole motor comprising three windings a, b, c (one for each stage) with each winding stage formed from four coils each have a predetermined number of concentric turns of a conductor. For clarity, only one conductor turn is shown in each slot of the armature. The side turns of the coils form four winding sets for each winding stage. For example, winding stage "b" has four coils each having a predetermined number of conductor turns with the first coil being disposed in core slots 842 and 843, the second coil being disposed in slots 844 and 845, the third coil being disposed in slots 846 and 847 and the fourth coil being disposed in slots 848 and 849. The coils may be wound consecutively or wound separately and then disposed and interconnected so as to produce current flow as indicated ("x" indicating current flow into the page and "." indicating current flow out of the page). As illustrated, the side turns of the coils create four winding sets 850-853 with winding set 850 being disposed in slots 850 and 842, winding set 851 being disposed in slots 843 and 844, winding set 852 being disposed in slots 845 and 846 and winding 853 being disposed in slots 847 and 848. After the coils have been wound and disposed in the armature slots, the conductor portions of each winding set will conduct current in the same axial direction as indicated, along the axial length of the core when the winding stage "b" is energized; thus, creating four magnetic poles or two pairs of magnetic poles, Nb, Sb as shown. The winding stages "a" and "c" of Figure 29 are formed in the same manner as described above for winding stage "b", with each having four winding sets with conductor portions conducting current in the same axial direction along the core when energized.

The three winding stage, four pole armature winding arrangement of Figure 29 has a winding stage "spread" of 30 mechanical degrees or 60 electrical degrees. As previously discussed, the "spread" is the angular expanse of adjacent core slots that carry the conductors of a given winding stage which instantaneously conduct current in the same axial direction along the axial length of the core. As shown in Figure 19, a set of winding stage "b" occupies two adjacent slots and all the conductors within that set carry current in the same axial direction along the axial length of the core; thus, the spread is the angular expanse of the two slots occupied by the set which is 60 electrical degrees or 30 mechanical degrees.

As discussed previously in reference to the two pole, two stage motor of Figures 1 and 2, the torque per ampere (T/I) characteristic of a motor is a function of winding spread and permanent magnet arc length. When maximum torque over a full revolution is desired, the torque per ampere (T/I) wave form typified by Figures 11A-C and 12-15 should be kept as "flat" as possible. A maximum T/I would be obtained if the T/I wave form for 180 electrical degrees were to be a square wave form. However, steeper wave forms create more possibilities of starting problems, thus it is desirable to approach as near as possible the square wave form without causing starting problems. The maximum T/I duration is increased by minimizing the winding "spread" and/or maximizing the rotor magnet arc length. In addition, as previously discussed, an optimum advanced timing angle a, based on rotor speed and L/R time constant of the windings for advancement of commutation of the windings can be selected which enables a reduction in magnet arc length. The functional relationship of these factors on a per pole basis can be expressed by the following expression: shows an armature for a three stage, 8 pole motor with three windings a, b, c (one for each stage) and with each winding stage being formed in a manner similar to the winding stages of Figures 28 and 29. A pole is created by several turns of a winding stage occupying a single slot with each winding stage occupying a total of eight slots to create eight poles or four pole pairs. In the eight pole arrangement shown in Figure 30, measuring spread involves measurement at only one slot. For example, the spread of winding stage "c" can be measured at core slot 854 in which one winding set of winding stage "c" is disposed with the set having conductor portions carry current in the same direction along the axial length of the core. One method of measuring spread which is the same method as used in previous figures is to measure from the center line of tooth 855 to the center line of tooth 856 which results in a spread of 60 electrical degrees. However, it can be seen that a different value for spread would be obtained if the spread were to be measured from one side of slot 854 to the opposite side of the slot. A still different value would result if the spread were to be measured from one side of the tip of tooth 855 to the side of the tip of tooth 856. Thus, it can be seen that where measurement of the spread involves a winding turn or turns in only one slot, the core slot size, slot geometry, tooth width, tooth tip width or geometry and number and location of winding turns within the slot become factors to consider in accurately determining the winding spread. The spread of a winding would approach zero if the winding were comprised of a single turn coil disposed in the air gap between a rotor and a core having no slots since magnetic coupling between the rotor and the coil would be unaffected by slot geometry, slot size, etc.

In Figure 30, a rotor employed with the stationary armature as shown would have eight permanent magnet segments distributed around its periphery to create eight poles at its perimeter. If the winding spread were 60 electrical degrees, then the magnet arc length of each segment would be between 120 and 150 electrical degrees using the expressed relationship previously set forth. Theoretically, in a three stage motor, a magnet arc length of 60 electrical degrees could be employed if the winding spread were to be 0 electrical degrees.

Figure 31 illustrates the coil distribution or winding arrangement for a four winding stage, two pole motor employing four windings a, b, c, d (one for each stage) disposed on a twenty-four slot stationary armature. Each stage has two winding sets and has its winding turns disposed in two groups of core slots symmetrically disposed from each other with each group comprising three adjacent core slots defining a respective one of the winding sets.

When a winding stage is energized, each set of that winding stage e.g., the set of winding stage "a" occupying adjacent core slots 857, 858, 859 conducts current along the axial length of the core to produce a magnetic pole such as Na or Sa.

Although the four stage, two pole motor is shown in Figure 31 with a twenty-four slot stationary armature, the motor could have easily been formed with only an eight slot armature core. Further, if a four stage, four pole motor were desired, sixteen core slots would be adequate. Thus, the minimum number of core slots will depend upon the number of motor stages and number of poles desired.

The formation of windings of the present invention is further illustrated by Figure 32 which shows a perspective view of a single monofilar winding stage "f" disposed on stationary armature core 868 to form one stage of a motor. The winding comprises five coils 870-874 disposed in core slots 875-884 with each coil comprising a predetermined number of concentric turns of conductor 885. The side turns of each of the coils are positioned within two core slots. As illustrated, coil 870 has its side turns disposed within slots 875 and 880; coil 871 has its side turns disposed within slots 876 and 881; coil 872 has its side turns disposed within slots 881 and 882; coil 873 has its side turns positioned within slots 878 and 883; and coil 874 has its side turns disposed within slots 879 and 884. The coil side turns disposed in slots 875-879 comprises a first winding set and the side turns disposed in slots 880-884 comprises a second winding set. The coils are wound and inserted into the slots of the core so that all the conductor portions of each winding set conducts current in the same direction along the axial length of the core when the winding is energized. The flow of current in the two winding sets establishes two magnetic poles Nf and Sf within the stationary armature. Of course, reversal of the current flow also reverses the location of the two magnetic poles.

The coils may be consecutively wound to form the one winding stage or they may be wound separately or in groups of one or more and then interconnected to form the one winding stage. Conventional winding equipment may be employed with the winding being formed directly on coil insertion tooling which is subsequently moved within the interior bore of the core for inserting the coils within the core slots. Also, other conventional winding and insertion equipment may be employed wherein coil for the winding are wound and then transferred to an inserting tool for placement within the core slots.

After the winding has been formed and disposed in the proper slots, the winding has its end turns disposed across the interior bore of the core; thus, they must be folded to the side toward the core face 886 in order to allow for mounting of the rotor. In Figure 32, the end turns of coils 870-872 may be folded in one direction away from reference line 887 toward the core face 886 and the end turns of coils 873-874 may be folded in an opposite direction away from the reference line. However, the side turns of the coils could be folded about reference lines other than 887 if convenient since the purpose of the folding is to provide clearance for the rotor.

A bifilar winding arrangement is illustrated in Figure 33 wherein double strands are wound and disposed within the slots 890 of the magnetic core 891. The double strands of wire are wound simultaneously and then inserted into the slots in the same manner as the monofilar or single strand winding illustrated in Figure 32. As a result of winding two strands of wire simultaneously, either one or two winding stages are established. If one winding stage is desired, strand ends 892 and 893 are connected to each other and ends 894 and 895 are connected to each other, thus, establishing effectively one winding stage. Bifilar winding to establish a single winding stage may be advantageous to establish a desired slot fill or where coil insertion equipment would have difficulty in inserting turns of wire a particular diameter but could insert turns comprising two wires of differing diameter. Of course, a single winding stage could be established by winding more than two strands simultaneously or two winding stages could be established in the same slots of the core by developing each winding stage separately and then inserting both into the core slots.

As mentioned previously, the expression developed for quantitatively determining the interrelationship between magnet arc length, winding spread, number of stages and timing advancement angle assumed that all core slots were being utilized for winding turns and that no slots were being shared between winding stages. However, various factors may make it impractical to have ideal core slot punching or winding distribution. For example, available cores may have too many or too few slots to allow conformance with the ideal situation where all core slots are filled and winding stages are not required to share slots. Under these circumstances, adequate motor performance can still be obtained by compensating for this increase or decrease in winding spread caused by slot sharing or by leaving empty slots. If core slots are left empty or a core slot is shared between winding stages, then the spread does not equal 180/N where N is the number of phases. For example, Figure 34 shows a two stage, two pole motor with winding stages "a" and "b" sharing slots 900, 902, 904, and 906, thus the spread would be greater than 90 electrical degrees. Therefore, a rotor magnet would have to be sized accordingly with greater length to compensate for the increase in a winding spread.

Figure 35 illustrates a stationary armature core for a two stage, two pole motor with empty slots 908, 910, 912, 914 between winding stages "a" and "b". With this arrangement, the core is not being fully utilized but the winding distribution is symmetrical. This arrangement allows use of a rotor permanent magnet of lesser arc length than would be required if all slots were filled with winding turns.

Even though the stationary armature arrangements in Figures 34 and 35 employ shared and empty slots, respectively, the winding stages are symmetrically disposed about the armature. Figure 36 illustrates a nonsymmetrical winding disposition in an eighteen slot stationary armature for a two stage, two pole motor. Empty slots 916, 918 appear in the middle of winding stage "a" with no empty slots appearing in winding stage "b". With this arrangement, the motor would still be capable of running, however, the motor performance would be decreased, thus making it undesirable for applications where optimum motor performance is required. In addition, compensation for the nonsymmetry would be required in circuitry being employed to advance the timing angle for winding energization such as incorporation of a time delay in the circuitry controlling commutation of winding stage "a" to compensate for an "on" time greater than that required for winding stage "b".

Commutation control for motors disclosed by this invention are unaffected by the number of poles. Thus, the three stage motors shown in Figures 28, 29 and 30 could employ an identical commutation control circuit. Further, the three stage motors can employ substantially the same circuitry as that employed for the previously discussed two stage, two pole motor as illustrated in Figures 25a and 25b.

Figures 37a and 37b illustrate the commutation circuit which it is believed could be used for a three stage motor. Figure 37a is similar to Figure 25a with the exception that wire 386, one of resistors 592, and switching element 392 have been deleted for three stage operation.

As mentioned previously, the number of poles does not affect commutation, however, for operation of a three stage motor, the back emf integration interval must be adjusted. This integration interval adjustment is provided by potentiometer 464. Adjustment of the potentiometer changes the triggering point for the Schmitt trigger comprising invertors 416 and 418. Operational amplifier 382 integrates the back emf over an interval or until a predetermined number of volt-seconds is accumulated whereupon the Schmitt trigger circuit is triggered as described with reference to Figure 25a.

Referring to Figure 37b, the remaining commutation circuitry is similar to the circuitry shown in 25b. The two circuits differ in that for operation of a three stage motor, flip-flops 452 and 454 are modified to produce three output signals, and element 920 with inputs Q1 and Q2 may be added to provide a signal on line 922 to reset flip-flops 452 and 454 when both Q1 and Q2 signals are present. In addition, because it is necessary to commutate only three windings for three stage operation, Figure 37b shows deletion from Figure 25b of one of each of elements 456, 400, 458, 376, 644, 646, 362, 364 and 366.

However, it has been found that coupling between winding stages of a three stage motor may be as low as 25 percent as compared to the two stage, two pole bifilarly wound motor where there is excellent coupling between the winding stages. Thus, there exists stored energy in each winding stage after it is deenergized which must be either dissipated or recovered. This stored energy may be dissipated by the power transistors or the energy may be either dissipated or recovered by an alternate means circuit 924 connected to the winding stages by way of diodes 926 as shown in Figure 37b. For dissipation, the alternate means could comprise a zener diode arrangement. Either dissipation of the stored energy or recovery of the energy by the alternate means circuit would permit the power transistors 362, 364 to be of a lower voltage rating.

The wave forms for the three stage motor circuits of Figures 37a and 37b are illustrated in Figure 38. The wave forms are essentially the same as for the two stage motor as illustrated in Figure 26 except for the deletion of signals on lines 402 and 434 and modifications of Q1, Q1, QZ and Q, signals.

The above-described circuits of 37a and 37b were for a three stage motor with the winding stages connected. in a half-bridge arrangement. However, if the winding stages were connected in a full-bridge arrangement as shown in Figures 9 and 10, substantially all the energy could be recovered from each stage by utilizing the diode arrangement as shown in Figures 9 and 10 and energy storage capacitor such as 91.

Another approach for controlling commutation of a three stage motor is illustrated in Figure 46 which shows a three stage (grounded nuetral) circuit with corresponding parenthetical numbers indicating outputs from ring counter 1137 and enabling inputs to sensing switches such as 1142 and winding enabling to power switches such as 1140. As illustrated in Figure 46, amplifier 1123, half wave rectifier 1125, integrator 1127, comparator 1129, comparator bias 1131 and integrator bias 1132 for motor starting purposes, one-shot timer 1133, differentiating circuit 1135 and ring counter 1137 perform substantially as described previously in discussing Figures 41 and 45. With three winding stages only three power switches such as 1140 are required and no logic circuit is necessary to control those windings, however, each stage is filtered by resistors, such as 1139, and capacitors, such as 1141, which for example have a .1 millisecond time constant and function to reduce the transients present during the switching interval. In the circuit of Figure 46, each winding stage is energized about one-third of the time, and in the circuit of Figure 45, each winding is energized about one-fourth of the time, however, more sophisticated circuitry, such as illustrated in Figure 47, may be employed to energize each winding stage of a three stage motor arrangement two-thirds of the time. A system such as illustrated in Figure 47 has the advantage that each winding stage in the motor is used or energized two-thirds of the time, thus providing a motor of more efficiency for a given size as compared to the system of Figure 46.

In the system of Figure 47, six power switches or transistors such as 1143, 1145 and 1147 are used since with no grounding of the neutral connection, the full bridge connected three winding stages will have two of those winding stages carrying current at any given time.

Thus, for example, when current is flowing into the "a" winding and out of the "b" winding stage, transistors 1143 and 1147 will be simultaneously energized. The circuit of Figure 47 again employs a ring counter 1149, this time of six stages, which is incremented by a differentiating circuit 1151. The "a", "b" and "c" winding stages are sequentially sampled (the winding stage not then carrying current) by sequential enablement of switches such as 1153 which sampled voltage is amplified by amplifier 1155 and since as in the case of the Figure 41 circuit, two polarities of sensed voltage may be encountered, this sensed voltage is passed by one of the switches 1157 or 1159, optionally by way of inverter 1161 to a further amplifier 1163. Amplifier 1163 may function like the previously discussed half wave rectifier and provides an output signal to integrating circuit 1165 which has a bias or starting voltage applied thereto at 1167 and that integrator output is supplied by way of amplifier 1169 to the comparator 1171 which, when the integrating circuit voltage exceeds the reference voltage supplied by source 1173, causes one-shot timer 1175 to reset the integrator 1165 and also the ring counter by way of differentiating circuit 1151. Noting that an asterisk before a winding stage identifying symbol indicates current flow through that same winding stage in an opposite direction to the un-asterisked indication, the six bit positions of the ring counter are coupled inverted in the manner indicated to the inputs of the several NAND gates such as 1177 and those NAND gate outputs 1, 2, 3, 4, 5 and 6 enable correspondingly identified transistors such as 1143. The outputs of NAND gates such as 1177 are also coupled to the correspondingly numbered inputs of NAND gates such as 1179, the outputs of which are decoded and, for example NAND gate 1181 controls switch 1159, whereas NAND gate 1183 countrols switch 1153. The other similarly positioned NAND gates control correspondingly identified switches. Thus in the circuits of Figure 47, higher utilization of the windings is achieved at the cost of additional logic circuitry.

The commutation of a four stage motor such as the one illustrated in Figure 31, could be controlled by the circuitry shown in Figures 25a and 25b with the addition of the energy recovery circuit shown in Figure 37b. However, if the four stage motor were to be provided with a permanent magnet having approximately the same arc length in electrical degrees as utilized in the two stage motor, the full potential of the magnet will not be utilized. Thus, the arc length of the magnet on the four stage motor could be reduced to achieve the same torque as emf curve illustrated in Figure 26 as that achieved with the two stage motor.

On the other hand, it is believed that it would be more desirable to retain magnets of approximately equal arc length in both the two stage and four stage motors, and provide an overlap in "on" times of the winding stages. This overlapping of "on" times and full magnet utilization would provide a motor with an increased total torque output because during the over-lapping periods, the winding stages energized would be producing torques which are additive. This overlap in "on" time can be provided by addition of additional logic in lines 434, 436, 438 and 440 of Figure 25b to enable the signals on these lines to be extended in duration. The wave forms resulting from modification of the Figure 25b circuit for a four stage motor are shown in Figure 39 with the changes in the signals on lines 434, 436, 438, and 440 being depicted with dashed lines. The remaining signals are identical to those illustrated in Figure 26.

In addition, the coupling between winding stages of a four stage motor is lower than the coupling for the two stage bifilarly wound motor of the type shown in Figures 1 and 2. Thus, an alternate means circuit as shown in Figure 35b and described hereinabove for a three stage motor could be utilized for a four stage motor or the stored energy could be dissipated by the power transistors 362 and 364.

It is believed motors with an even greater number of stages and any number of poles could be constructed in accordance with the teachings of this invention. The windings for each stage could be connected in a half-bridge circuit configuration and commutated with circuitry similar to that which would be employed with the above-discussed three stage and four stage motors. Although it is believed that the emf voltage of two or more winding stages could be combined for simulating rotor position, it would generally be preferable to provide a detector circuit having provision for emf voltage sensing for each winding stage. If a motor were to be constructed with five or more stages, the position determining disclosed hereinabove may be employed to simulate the rotor position; however as the stages are increased, a shorter reset time may be required for the position determining circuit in order to assure that the circuit is reset prior to the previously discussed zero crossing point or point "B" of the emf voltage. Further, the logic circuitry generating A, B, A, etc. would require modification in order to produce output signals equal in number to the number of winding stages to be sequentially energized. And of course, each winding stage would be provided with a separate power transistor arrangement for effecting energization.

Still further variations could be implemented using the teachings of this invention. For example, the winding stages of multi-stage motors constructed as shown herein could be connected in a full-bridge circuit configuration as disclosed in Figures 9 and 10 to provide more efficient utilization of winding material but of course, at the expense of additional circuitry. With this approach, a detector circuit similar to detector circuit 814 of Figure 25a, could be provided either with a half-wave rectifier arrangement as previously discussed with reference to Figure 47 or with two switches for each winding stage to reverse the polarity of the emf voltage when it is negative so that a positive voltage is always being provided to the position determining circuit. Four logic signals such as A, B, A, etc., and four power transistor sets would have to be provided for each winding connected in a full-bridge arrangement with the exception of a three stage motor where the windings could be connected as shown in Figure 47 with six power transistor sets being utilized.

WHAT WE CLAIM IS: 1. A brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinbefore defined); each winding stage comprising at least two coils of concentric winding turns accommodated by the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles; the axially extending conductor

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (48)

**WARNING** start of CLMS field may overlap end of DESC **. are coupled inverted in the manner indicated to the inputs of the several NAND gates such as 1177 and those NAND gate outputs 1, 2, 3, 4, 5 and 6 enable correspondingly identified transistors such as 1143. The outputs of NAND gates such as 1177 are also coupled to the correspondingly numbered inputs of NAND gates such as 1179, the outputs of which are decoded and, for example NAND gate 1181 controls switch 1159, whereas NAND gate 1183 countrols switch 1153. The other similarly positioned NAND gates control correspondingly identified switches. Thus in the circuits of Figure 47, higher utilization of the windings is achieved at the cost of additional logic circuitry. The commutation of a four stage motor such as the one illustrated in Figure 31, could be controlled by the circuitry shown in Figures 25a and 25b with the addition of the energy recovery circuit shown in Figure 37b. However, if the four stage motor were to be provided with a permanent magnet having approximately the same arc length in electrical degrees as utilized in the two stage motor, the full potential of the magnet will not be utilized. Thus, the arc length of the magnet on the four stage motor could be reduced to achieve the same torque as emf curve illustrated in Figure 26 as that achieved with the two stage motor. On the other hand, it is believed that it would be more desirable to retain magnets of approximately equal arc length in both the two stage and four stage motors, and provide an overlap in "on" times of the winding stages. This overlapping of "on" times and full magnet utilization would provide a motor with an increased total torque output because during the over-lapping periods, the winding stages energized would be producing torques which are additive. This overlap in "on" time can be provided by addition of additional logic in lines 434, 436, 438 and 440 of Figure 25b to enable the signals on these lines to be extended in duration. The wave forms resulting from modification of the Figure 25b circuit for a four stage motor are shown in Figure 39 with the changes in the signals on lines 434, 436, 438, and 440 being depicted with dashed lines. The remaining signals are identical to those illustrated in Figure 26. In addition, the coupling between winding stages of a four stage motor is lower than the coupling for the two stage bifilarly wound motor of the type shown in Figures 1 and 2. Thus, an alternate means circuit as shown in Figure 35b and described hereinabove for a three stage motor could be utilized for a four stage motor or the stored energy could be dissipated by the power transistors 362 and 364. It is believed motors with an even greater number of stages and any number of poles could be constructed in accordance with the teachings of this invention. The windings for each stage could be connected in a half-bridge circuit configuration and commutated with circuitry similar to that which would be employed with the above-discussed three stage and four stage motors. Although it is believed that the emf voltage of two or more winding stages could be combined for simulating rotor position, it would generally be preferable to provide a detector circuit having provision for emf voltage sensing for each winding stage. If a motor were to be constructed with five or more stages, the position determining disclosed hereinabove may be employed to simulate the rotor position; however as the stages are increased, a shorter reset time may be required for the position determining circuit in order to assure that the circuit is reset prior to the previously discussed zero crossing point or point "B" of the emf voltage. Further, the logic circuitry generating A, B, A, etc. would require modification in order to produce output signals equal in number to the number of winding stages to be sequentially energized. And of course, each winding stage would be provided with a separate power transistor arrangement for effecting energization. Still further variations could be implemented using the teachings of this invention. For example, the winding stages of multi-stage motors constructed as shown herein could be connected in a full-bridge circuit configuration as disclosed in Figures 9 and 10 to provide more efficient utilization of winding material but of course, at the expense of additional circuitry. With this approach, a detector circuit similar to detector circuit 814 of Figure 25a, could be provided either with a half-wave rectifier arrangement as previously discussed with reference to Figure 47 or with two switches for each winding stage to reverse the polarity of the emf voltage when it is negative so that a positive voltage is always being provided to the position determining circuit. Four logic signals such as A, B, A, etc., and four power transistor sets would have to be provided for each winding connected in a full-bridge arrangement with the exception of a three stage motor where the windings could be connected as shown in Figure 47 with six power transistor sets being utilized. WHAT WE CLAIM IS:

1. A brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinbefore defined); each winding stage comprising at least two coils of concentric winding turns accommodated by the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles; the axially extending conductor

portions within each given set being comprised generally of about one half of the conductor side turn portions of at least two different coils, and such conductor portions being disposed to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing such given set is energized; the arcuate spread of any given set of axially extending conductors being less than about 120 electrical degrees; a rotor having constant magnetic polarity polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate relative to the armature in response to the magnetic poles established by the winding turns; and commutation means interconnected with at least one winding stage for sensing relative angular position of the rotor and the stationary armature and operative to energize the winding stages in a predetermined manner to establish the magnetic poles on said armatures for causing rotational movement of the rotor.

2. A brushless DC motor as set forth in claim 1 wherein each said predetermined spread is in a preferred range from about 30 electrical degrees to about 120 electrical degrees.

3. A brushless DC motor as set forth in claim 1 wherein the winding stages are arranged in pairs and each pair is formed of bifilar strands that share common armature slots.

4. A brushless DC motor as set forth in claim 1 wherein said winding stages are connected in a bridge configuration.

5. A brushless DC motor as set forth in claim 1 wherein the magnetic polar regions established by said rotor are created by permanent magnets disposed on the rotor.

6. A brushless DC motor as set forth in claim 1 wherein the commutation means is responsive to back emf of an unenergized winding for determining the relative angular position of the rotor and the stationary armature.

7. A brushless DC motor as set forth in claim 1 wherein the commutation means includes detector circuit means responsive to the current flow in the windings to provide an output signal indicative of rotor speed, and energizing circuit means responsive to the output of said detector circuit means for providing a plurality of armature winding energizing signals displaced in a predetermined temporal sequence.

8. A brushless DC motor comprising a stationary armature having a core and at least two winding stages (as herein before defined); each winding stage comprising concentric winding turns accommodated by said core and arranged to establish a predetermined number of magnetic poles and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number equal to the predetermined number of magnetic poles; the axially extending conductor portions within each set being disposed in said armature to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing the given set is energized; the arcuate spread of any given set of axially extending conductors being less than about 120 electrical degrees; a rotor having constant magnetic polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate in response to the magnetic poles established by the winding turns; and a commutation circuit for energizing the winding stages in a predetermined manner wherein said commutation circuit includes a detector circuit for sensing a back emf signal indicative of the back emf condition of at least one winding stage, position determining circuit means for receiving the output of the detector circuit and for producing a simulated relative position output that is indicative of the relative angular position of the rotor and armature, with such relative position output determined by the back emf condition of a winding stage, and circuit means interconnected with the position determining circuit means for supplying an output signal for energizing a selected one of the winding stages.

9. The brushless DC motor circuit of claim 8 wherein the position determining circuit of the commutation circuit includes a circuit for producing a signal as a measure of flux in a winding stage not then being energized and for producing said simulated relative position output.

10. The brushless DC motor of claim 8 wherein the position determining circuit is operative to cause advancement of commutation of the winding stages by an angle alpha of from about five to about thirty electrical degrees to aid the build up of current when the winding stages are energized during running condition.

11. The brushless DC motor of claim 10 wherein the position determining circuit includes means to vary the advancement of commutation angle alpha.

12. The brushless DC motor of claim 8 wherein the position determining circuit means of the commutation circuit includes an integration means for integrating the back emf signal from said detector circuit to a predetermined value of volt-seconds whereupon the position determining circuit produces an output signal to the circuit means.

13. The brushless DC motor of claim 12 wherein the predetermined value of volt-seconds attained by said position determining circuit means occurs at an angular rotor position relative to the armature corresponding to the advancement of commutation angle alpha.

14. The brushless DC motor of claim 8 wherein the position determining circuit of the commutation circuit comprises a voltage controlled oscillator responsive to the output of said detector circuit, for producing output pulses at a frequency indicative of said detector output signal and a counter means responsive to the oscillator for producing an output signal after a predetermined number of pulses have been counted.

15. The brushless DC motor of claim 8 wherein the detector circuit of the commutation circuit senses the back emf of only one winding stage at a time and wherein the circuit further comprises switching means that sequentially gate back emf signals from the different winding stages to the detector circuit.

16. The brushless DC motor of claim 8, wherein the back emf condition sensed by the detector circuit of the commutation circuit includes a characteristic signal generated by means for aiding starting and wherein the characteristic signal is associated with the emf condition of the motor at low motor speed.

17. The brushless DC motor of claim 8 wherein the commutation circuit further includes means for resetting the position determining circuit means after the simulated relative position output is produced.

18. The brushless DC motor of claim 8 wherein the circuit means of the commutation circuit includes logic circuit means responsive to the output of the position determining circuit for the purpose of selecting an energization sequence for the winding stages.

19. The brushless DC motor of claim 8, wherein the circuit means of the commutation circuit comprises indexing means responsive to the output of the position determining circuit for producing an energization sequence for the winding stages.

20. The brushless DC motor of claim 8, wherein the circuit means of the commutation circuit comprises first and second flip-flops, the input of said first flip-flop being coupled to the output of the position determining circuit to provide first and second complementary signals, said second flip-flop having an input coupled to one of the outputs of said first flip-flop for providing third and fourth complementary signals.

21. The brushless DC motor circuit of claim 8 wherein the commutation circuit further includes protection means operative to prevent damage to the commutation circuit and motor due to reverse polarity of a direct current source supplying power to the commutation circuit and motor.

22. The brushless DC motor of claim 8 wherein the commutation circuit includes power driving means for applying power to the winding stages and unidirectional conducting means connected thereto for conducting stored energy from a winding stage after deenergization of the winding stage.

23. A brushless DC motor comprising a stationary armature having a slotted core and at least two winding stages (as hereinbefore defined); each winding stage comprising at least one coil of concentric winding turns accommodated in nonadjacent symmetrically disposed slots of the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles; the axially extending conductor portions within each given set comprised generally of one half of the axially extending conductor portions of the at least one coil, and such conductor portions being disposed to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing such given set is energized; the arcuate spread of any given set of axially extending conductor portions being less than 120 electrical degrees; a rotor having constant magnetic polarity polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate relative to the armature in response to the magnetic poles established by the winding turns; and commutation means interconnected with at least one winding stage for sensing relative angular position of the rotor and the stationary armature and operative to energize the winding stages in a predetermined manner to establish the magnetic poles on said armature for causing rotational movement of the rotor.

24. A brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinbefore defined); each winding stage comprising concentric winding turns accommodated by the core and arranged to establish a predetermined number of magnetic poles, and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number being equal in number to the predetermined number of poles; the side turn axially extending conductor portions within each given set being disposed to conduct current instantaneously in the same axial direction along the core thereby establishing a predetermined spread and establishing a magnetic pole when the winding stage containing such given set is energized; a rotor having a plurality of permanent magnet segments disposed thereon and adapted to rotate in response to the magnetic poles established by the armature; a commutation circuit for energizing the winding stages in a predetermined manner and at a predetermined angle of advance a; and wherein the arcuate spread of each permanent magnet segment disposed on said rotor is about equal in electrical degrees to the winding spread plus 180 (N- 1)/N minus 2 a where N equals the number of winding stages of the motor.

25. The brushless DC motor as set forth in claim 24 wherein the preferred predetermined angle of advance a is from about 5 electrical degrees to about 30 electrical degrees.

26. The brushless DC motor of claim 24 wherein 180 (N- 1)/N with N equal to the number of stages of the motor corresponds to the energization time in electrical degrees of each winding of the motor.

27. The brushless DC motor of claim 24 wherein the preferred value of spread with all slots of the stationary armature being utilized and with no shared slots between windings being about equal to 180/N where N equals the number of winding stages of the motor.

28. A brushless DC motor according to Claim 1 wherein the commutation means includes a detector circuit comprising: means responsive to the current drawn by the armature windings to provide an output signal indicative thereof, means for scaling the output signal by a factor corresponding to the resistance of the armature windings, means for subtracting the resulting scaled signal from the voltage applied to the armature windings to provide a signal indicative of the back emf of the brushless DC motor, and a frequency circuit means responsive to the back emf signal for generating a signal of a frequency proportional thereto indicative of the rotor speed and with said frequency circuit means having a minimum frequency output signal to aid starting of the brushless DC motor.

29. The brushless DC motor of claim 28 wherein said frequency circuit means comprises a voltage controlled oscillator for generating a signal of a frequency proportional thereto indicative of rotor speed and wherein the voltage controlled oscillator has a minimum frequency output signal to aid starting of the brushless DC motor.

30. The brushless DC motor of claim 28 wherein the commutation circuit further comprising means for selectively preventing the application of the detector circuit signal indicative of back emf to the frequency circuit means and means operative therewith for selectively substituting a signal proportional to motor load to be applied to the frequency circuit means.

31. The brushless DC motor of claim 30 wherein the means for substituting a signal porportional to motor load comprises means for sampling a first portion of the motor winding current, means for sampling a second portion of the motor winding current, means for comparing the first and second sampled portions, for modifying the frequency of the frequency circuit means output.

32. A brushless DC motor according to Claim 1 wherein the commutation means comprises a detector circuit for sensing the current drawn through the armature windings and for scaling the sensed current signal in accordance with the resistance of the windings, for sensing voltage applied to the winding, and for generating an output signal indicative of rotor speed; an indexing means responsive to the detector circuit output signal for generating a plurality of output signals indicative of a relative position of said rotor with respect to said armature; an energizing circuit means responsive to said plurality of output signals from said indexing means for energizing said windings in a predetermined sequence in accordance with the relative position of said rotor.

33. A brushless D.C. motor as claimed in claim 32, wherein said detector circuit means includes a voltage-controlled oscillator circuit for generating a signal of a frequency proportional to the rotor speed.

34. A brushless DC motor according to claim 1 further comprising: a rotor position sensor for producing pulse output signals indicative of rotor position relative the stationary armature; stepping logic circuitry responsive to the pulse output signals from the rotor position sensor for inhibiting continuous winding commutation in a predetermined sequence; mode control circuitry responsive to the pulse output signals from the rotor position sensor, to rotational direction command signals, and to continuous and stepping operational command signals for producing an output for use in selecting a winding stage for energization; a pulse modulating circuit responsive to the mode control circuitry for supplying an output signal for energizing a selected one of the winding stages; and a current sensor responsive to the current flow through the windings of the motor for producing an output signal to said pulse modulating circuit for inhibiting the modulating circuit output signal for a predetermined period of time when the motor current exceeds a predetermined value so as to limit the magnitude of current supplied to the motor windings.

35. The brushless DC motor of claim 34 wherein the rotor position sensor includes stationary exciter and pick up coils and a rotating shutter supported by the rotor for sequentially coupling and decoupling the coils to produce output signals indicative of rotor position relative the stationary armature.

36. A brushless DC motor according to Claim 1 wherein the commutation means comprises: a position circuit for simulating rotor postion; a circuit for energizing a selected one of the windings in accordance with the simulated rotor position; and an under speed protection circuit operable to prevent the circuit for energizing a selected one of the winding stages from energizing any of the winding stages when the motor speed is less than a predetermined minimum value for a predetermined length of time.

37. The brushless DC motor of claim 36 wherein said under speed protection circuit includes means for varying said predetermined minimum value of motor speed.

38. The brushless DC motor of claim 36 wherein said under speed protection circuit includes reset means for allowing energization of of the motor winding stages after a predetermined period of time.

39. A brushless DC motor according to Claim 1 wherein the commutation means comprises: a position circuit for simulating rotor position; a circuit for energizing a selected one of the winding stages with power from the voltage source in accordance with the simulated rotor position; and an undervoltage protection circuit operable to prevent the circuit for energizing a selected one of the winding stages from energizing any of the winding stages when the voltage source output is less than a predetermined minimum value.

40. A brushless DC motor according to Claim 1 wherein the commutation means comprises: a position circuit for simulating rotor position; a circuit for energizing a selected one of the windings with power from the voltage source; and an overvoltage circuit operable to prevent the circuit for energizing a selected one of the windings from energizing any of the windings when the voltage source output is greater than a predetermined maximum value.

41. A brushless DC motor according to Claim 1 wherein the commutation means includes means for aiding starting of the motor by generating a characteristic signal that is associated with an emf condition of the motor at low motor speed and wherein the characteristic signal is substantially less in magnitude than emf associated with the motor at full operating speed.

42. A brushless DC motor comprising a stationary armature having a core and at least two winding stages (as hereinbefore defined); each winding stage comprising concentric winding turns accommodated by said core and arranged to establish a predetermined number of magnetic poles and the winding turns of each winding stage forming a number of sets of axially extending conductor portions with such number equal to the predetermined number of magnetic poles; the axially extending conductor portions within each set being disposed in said armature to conduct current instantaneously in the same axial direction along the core thereby to establish a magnetic pole when the winding stage containing the given set is energized; a rotor having constant magnetic polar regions equal in number to the predetermined number of poles, said rotor being adapted to rotate in response to the magnetic poles established by the windings turns; and a commutation circuit for energizing the winding stages in a predetermined manner wherein said commutation circuit includes a detector circuit for sensing a back emf signal indicative of the back emf condition of at least one winding stage, position determining circuit means responsive to only a positive polarity portion of the output signal from the detector circuit for integrating said positive polarity portion of the output signal to a predetermined value of volt-seconds whereupon the position determining circuit means produces a simulated relative position output that is indicative of the relative angular position of the rotor and armatures and means responsive to the simulated relative position output from the position determining circuit means for supplying an output signal for energizing a selected one of the winding stages, the position determining circuit being operative to establish a predetermined advancement of commutation of the winding stages by an angle a of from about five to about thirty electrical degrees.

43. A brushless direct current motor according to Claim 1 wherein the commutation means comprises: counter means identifying the energized winding; means for sampling the induced voltage in a winding not currently energized; means for integrating the induced voltage sample; a comparator for comparing the voltage integral to a reference voltage and providing an output signal when the voltage integral exceeds the reference voltage; means reponsive to the comparator output signal for returning the integrating means to its initial state and for incrementing the counter means.

44. A brushless D.C. motor according to claim 43 wherein the counter means comprises a ring counter.

45. A brushless D.C. motor according to claim 43 wherein the responsive means includes a one-shot timer for resetting the means for integrating, and a differentiating circuit responsive to one-shot timer for incrementing the counter means.

46. A brushless D.C. motor according to claim 43 wherein the means for sampling includes a plurality of electrically actuable switches interconnected with motor winding leads and responsive to the counter means to sample an induced voltage in a winding not then identified by the counter means as being energized.

47. A brushless D.C. motor according to claim 46 wherein the counter means enables for sampling the winding next to be energized.

48. A brushless D.C. motor substantially as described herein with reference to the accompanying drawings.