Apparatus and method for commutation noise reduction

What is claimed is:

1. A method for reducing commutation noise in an electric motor having a bridge switching circuit with a first and second group of switches each switch of said first group of switches being connected in series across a source of power with a different switch of said second group of switches, and a junction of each pair of series connected switches being connected to a different phase armature winding of said electric motor, said method comprising:

a) repeatedly activating a selected switch of said first group of switches and a selected switch of said second group of switches, said selected switch of said second group of switches being non-series connected to said selected switch of said first group;

b) connecting at least two phase windings in series with said electrical power source in a predetermined commutation pattern to provide torque producing phase currents, said predetermined commutation pattern defining commutation events in which a switch from either one of said groups is deactivated and another switch of the same group is activated and a switch of the other group remains activated, the newly activated switch alternating in consecutive commutation events between said first group and said second group; and

c) providing overlapping activation of the switch being deactivated with the switch being newly activated and the switch remaining activated during each commutation event, the newly activated switch being pulse width modulated with a motor torque controlling duty cycle and the switch being deactivated being pulse width modulated with pulses initiated synchronously with those of the newly activated switch but with the duty cycle decreasing from the torque controlling duty cycle.

2. The method of claim 1, in which the duty cycle of the switch being deactivated is decreased in an exponential manner.

3. The method of claim 2 in which the time constant of the exponential decrease is varied inversely with motor speed to produce a predetermined minimum duty cycle over a predetermined maximum number of electrical degrees of the motor.

4. The method of claim 3, wherein a current command representing the current status of said motor is inputted into a control algorithm and compared to a first calibration value and if said current command is less than said first calibration value the value of said time constant is increased, if said current command is greater than said first calibration value, said current command is compared to a second calibration value, if said current command is less than said second calibration value the value of said time constant remained unchanged, if said current command is greater than said second calibration value the value of said time constant is decreased.

5. The method of claim 4, wherein said first and second calibration values vary in accordance with the network properties of said motor.

6. The method of claim 1, wherein said motor is operated in a single switching mode and the switch remaining activated is closed continuously during commutation.

7. The method of claim 6, wherein said motor is alternated between the single switching mode and a dual switching mode in response to a signal varying with motor torque.

8. The method of claim 6, wherein said motor is changed between the single switching mode and the dual switching mode in further response to a signal varying with motor speed.

9. The method of claim 7, wherein the switch remaining activated is pulse width modulated synchronously with the switch being newly activated in said dual switching mode.

10. The method of claim 1, wherein said motor is operated in a dual switching mode and the switch remaining activated is pulse width modulated synchronously with the switch being newly activated at the motor torque controlling duty cycle.

11. The method of claim 10, wherein s aid motor is alternated between a single switching mode and the dual switching mode in response to a signal varying with motor torque.

12. The method of claim 11, wherein said motor is changed between the single switching mode and the dual switching mode in further response to a signal varying with motor speed.

13. The method of claim 12, wherein the switch remaining activated during single switching mode is closed continuously during activation.

14. A method for reducing commutation noise in a brushless DC motor having a plurality of switches coupled to a power source and a plurality of phase windings, said switches being activated and deactivated in a predetermined commutation pattern to provide torque producing phase currents in the windings of said motor, said method comprising:

a) determining a time constant for manipulating the rate of deactivation of said switches;

b) comparing a current command of said motor with a first value, said first value corresponding to a value of motor current, if said current command is less than said first value the value of said time constant is increased, if said current command is greater than said first value said current command is compared to a second value, said second value corresponding to a second value of motor current, said second value being greater than said first value, if said current command is less than said second value the value of said time constant remains unchanged, if said current command is greater than said second value the value of said time constant is decreased; and

c) using said time constant after said comparing step to manipulate the rate of deactivation of said switches.

15. The method as in claim 14, wherein said first and second values are replaced by a plurality of values corresponding to a plurality of current values and said current command is compared to said plurality of values and the value of said time constant is either increased or decreased after the comparison of said current command to said plurality of values.

16. The method as in claim 14, wherein said time constant varies continuously with said current command.

17. The method as in claim 15, wherein said time constant varies continuously with said current command.

18. A method for optimizing performance in an electric motor having a bridge switching circuit with a first and second group of switches each switch of said first group of switches being connected in series across a source of power with a different switch of said second group of switches, and a junction of each pair of series connected switches being connected to a different phase armature winding of said electric motor, said method comprising:

a) repeatedly activating a selected switch of said first group of switches and a selected switch of said second group of switches, said selected switch of said second group of switches being non-series connected to said selected switch of said first group;

b) connecting at least two phase windings in series with said electrical power source in a predetermined commutation pattern to provide torque producing phase currents, said predetermined commutation pattern defining commutation events in which a switch from either one of said groups is deactivated and another switch of the same group is activated and a switch of the other group remains activated, the newly activated switch alternating in consecutive commutation events between said first group and said second group; and

c) alternating between a single switching mode and a dual switching mode in response to a signal varying with motor torque.

19. The method as in claim 18, wherein said signal varies with motor torque and motor speed.

20. The method as in claim 14, wherein said time constant is derived as a function of said motor speed.

21. The method as in claim 14, wherein said first value is 25 amperes and said second value is 60 amperes.

22. The method as in claim 14, wherein said time constant is also dependent upon the input current level.

23. A method for reducing commutation noise in a brushless DC motor having a plurality of switches coupled to a power source and a plurality of phase windings, said switches being activated and deactivated in a predetermined commutation pattern to provide torque producing phase currents in the windings of said motor, said method comprising:

determining a time constant for manipulating the rate of deactivation of said switches;

comparing a current command of said motor with a first value, said first value corresponding to a value of motor current, if said current command is less than said first value the value of said time constant is doubled, if said current command is greater than said first value said current command is compared to a second value, said second value corresponding to a second value of motor current, said second value being greater than said first value, if said current command is less than said second value the value of said time constant is multiplied by one or remains unchanged, if said current command is greater than said second value the value of said time constant is decreased by half; and

using said time constant after said comparing step to manipulate the rate of deactivation of said switches.

24. The method of claim 2 in which a time constant is derived, said time constant corresponding to the exponential decrease of the switch being deactivated.

25. The method of claim 24 wherein the exponential decrease is varied inversely with motor speed to produce a predetermined minimum duty cycle over a predetermined maximum number of electrical degrees of the motor.

26. The method of claim 24, wherein a current command representing the current status of said motor is inputted into a control algorithm and compared to a first calibration value and if said current command is less than said first calibration value the value of said time constant is increased, if said current command is greater than said first calibration value, said current command is compared to a second calibration value, if said current command is less than said second calibration value the value of said time constant remained unchanged, if said current command is greater than said second calibration value the value of said time constant is decreased.

Description Submit all comments and votes

BACKGROUND OF THE INVENTION

In order to improve the fuel efficiency of an automobile, the power steering pump which continuously circulates a hydraulic fluid, is replaced by an electrical motor that is actuated in response to the manipulation of the steering system.

Accordingly, the vehicles engine is no longer providing a driving force to the power steering pump. This results in a lower overall load upon the engine which will improve fuel efficiency.

An electric motor is used to provide a steering assisting force in response to manipulation of the steering system. One such motor is a brushless DC motor which is mechanically coupled to the steering column in order to provide the assisting force.

A brushless DC motor requires electronic commutation of its armature currents. This is typically accomplished, for a trapezoidal back EMF, three-phase motor, by means of a bridge switching circuit containing six semiconductor switches. (FIGS. 3 and 4).

When a predetermined upper switch and non-series lower switch are simultaneously conducting, an armature current flows through two of the three-phase windings, in series, to electromagnetically interact with the permanent magnet motor and develop torque in a predetermined direction.

Commutation to a different pair of armature windings is accomplished by turning off one of the switches and turning on a different non-series switch of the same level (upper or lower). A control provides switching signals to the switch gates in the proper sequence and with the proper timing.

Accordingly, motor operation is controlled by activating the switches in a predetermined pattern.

The electric currents in the activating phases are controlled, usually by a pulse width modulation at a higher frequency, to control motor torque. This modulation, in combination with the inductance of the windings, produces an average motor current, and thus a smooth motor torque. However, this is not the case during commutation.

During commutation, where the phases are turned off and on abruptly, the rise and fall of the phase currents are controlled only by the network properties of the motor and switching circuitry. The forcing function for the phase turning off is not the same as it is for the phase turning on, and one of the circuits (rising or falling) will change faster than the other.

The unequal current in the phase turning off and phase turning on can produce a disturbance in the average motor current, which will cause a variation in torque, at each commutation event. In addition, such disturbances can also drive the motor structure to generate audible noise.

Moreover, and in some applications where the motor structure is positioned within the passenger compartment of a vehicle, the audio level of these noises may reach an objectionable level. In particular, use of such a control and motor in an electric power steering system in modes of operation which produce high phase currents at low motor speed can produce an annoying clicking noise at commutation. This audible noise has been referred to "zipper noise".

Accordingly, it is desirable to control the phase currents of such a motor in order to reduce or eliminate such audible noises.

SUMMARY OF THE INVENTION

The motor control of this invention modifies motor commutation events to reduce noise by providing overlapping activation of the switch being deactivated with the switch been newly activated so that three switches are temporary activated. The newly activated switch is pulse width modulated with a motor torque controlling duty cycle, and the switch being deactivated is pulse width modulated synchronously with the newly activated switch but with a duty cycle decreasing from the torque controlling duty cycle. The decreasing modulation of the switch being deactivated allows a slower current decrease in the phase turning off to reduce or eliminate the disturbances, and thus the noise. The decreasing modulation is preferably exponential in manner, with a time constant varying inversely with motor speed so that the duty cycle reaches a predetermined minimum in a predetermined maximum number a motor electric degrees.

The decreasing modulation may be switched on at low motor speeds where its noise reduction is required and switched off at higher motor speeds where the potential for interference with commutation is greater. The decreasing modulation is applicable in a dual switching mode in which each activated switch, except the switch being deactivated in commutation, is pulse width modulated at the torque controlling duty cycle. The decreasing modulation is alternatively applicable in a modified single switching mode in which the activated switch of one of the upper or lower groups is always closed continuously during its activation, except for commutation, during which the switch remaining on, which alternates between the groups, is closed continuously. The control is preferably responsive to motor current and/or motor speed to apply each switching mode in the motor speed/torque region to which is best suited. Preferably, control stability is improved by a sample/hold circuit and a motor speed controlled variable forcing function generator to augment a sense current feedback signal during the portion of a commutation event in which the early opening of one of the switches causes current to recirculate in the bridge and not be sent by the current sensor.

In yet another embodiment, the time constant is further modified in order to reduce or eliminate disturbances. The further modification of the time constant is dependent upon the input current command.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates a possible configuration of an electrical power steering system;

FIG. 2 shows a block diagram of an embodiment of a controller for use in the electrical power steering system of FIG. 1;

FIG. 3 illustrates a circuit diagram of the bridge circuit for use in the controller of FIG. 2;

FIG. 4 is an equivalent electric circuit of an electric motor for use in the electric power steering system of FIG. 1 with the bridge circuit of FIG. 3;

FIG. 5 is a block diagram of a PWM circuit for use in the controller of FIG. 2;

FIG. 6 is a circuit and block diagram of the currents and circuit for use in the controller of FIG. 2;

FIG. 7 is a block diagram of the current control circuit for use in the controller of FIG. 2;

FIG. 8 is a block and schematic diagram of the commutation control circuit for use in the controller of FIG. 2;

FIG. 9 shows timing diagrams illustrating operation of the controller of FIG. 2 and a first mode;

FIG. 10 shows timing diagrams illustrating the operation of the controller of FIG. 2 and a second mode;

FIG. 11 shows timing and current diagrams useful in the illustration of the operation of the currents and circuit of FIG. 6;

FIG. 12 shows a torque/speed map illustrating areas of application of the first and second circuit modes of the controller of FIG. 2;

FIG. 13 is a flowchart of a control algorithm for use by the controller of FIG. 2 in an alternative embodiment of the present invention;

FIG. 14 is a graph illustrating calibration constant as a function of input current command;

FIG. 15 is a graph illustrating the relationship of individual phase currents when the rate of current change is controlled by the network properties of the machine;

FIG. 16 is a graph illustrating the relationship of the individual phase currents when the rate of the current being deactivated is manipulated by a time constant which is dependent upon the motor speed;

FIG. 17 is a graph illustrating the relationship of the individual phase currents when the rate of the current being deactivated is manipulated by a time constant which is also dependent upon the input current level;

FIG. 18 is a graph comparing the graphs illustrated in FIGS. 15-17; and

FIG. 19 is a graph comparing the graphs illustrated in FIGS. 15-17.

DESCRIPTION OF THE PREFERRED EMBODIMENT

Referring now to FIG. 1, a steering system 10 for use in a motor vehicle 12 (not shown) is illustrated. Steering system 10 allows the operator of motor vehicle 12 to control the direction of motor vehicle 12 through the manipulation of steering system 10.

A steering column 14 provides mechanical manipulation of the automobiles wheels in order to control the direction of motor vehicle 12. Steering column 14 includes a hand wheel 16. Hand wheel 16 is positioned so that a user can apply a rotational force to steering column 14. An upper steering column shaft 18 is secured to hand wheel 14 at one end and a universal joint 20 at the other. Universal joint 20 couples upper steering column shaft 18 to a lower steering column shaft 22. Lower steering column shaft 22 is secured to universal joint at one end and a gear housing 24 at the other. Gear housing 24 includes a pinion gear (not shown). Pinion gear of gear housing 24 is positioned to make contact with a toothed rack 26.

Tie rods (only one shown) 28 are secured to rack 26 at one end and knuckles 30 (only one shown) at the other.

As a rotational force is applied to steering column 14, through the manipulation of hand wheel 16 or other applied force, the pinion gear of gear housing 24 is accordingly rotated. The movement of the pinion gear causes the movement of toothed rack 26 which in turn manipulates tie rods 28 and knuckles 30 in order to reposition the wheels 32 (only one shown) of motor vehicle 12.

In order to assist the user applied rotational force to steering column 14 an electric motor 34 provides a torque force to a worm 36. Worm 36 is positioned to manipulate a worm gear 38. Worm gear 38 is secured to upper steering column shaft 18. Accordingly, and as a torque force is applied to worm 36, steering column 14 is rotated.

In certain applications electric motor 34 is located within the passenger compartment (not shown) of a vehicle. The positioning of electric motor 34 may be due to a variety of reasons including, but not limited to the following: convenience or easy installation; lack of adequate space within the engine compartment; and hostile environmental conditions within the engine compartment such as, excessive heat, cold, moisture, excessive vibrations and electrical interference.

However, the positioning of electrical motor 34 within the vehicle's passenger compartment increases the need to reduce or eliminate audible noises due to commutation.

Electric motor 34 is actuated by a controller 40 which receives inputs from a torque sensor 42 and a rotational position sensor 44. Sensor 44 provides a steer angle signal to controller 40.

In addition, and as the rotor of motor 34 turns, rotor position signals of each phase are generated within motor 34 and inputted into controller 40 through a bus 46.

Controller 40 also receives an input in the form of a vehicle speed signal. Accordingly, and in response to the following inputs: vehicle velocity input; operator torque input (sensor 42); steering pinion gear angle (sensor 44); and rotor position signals (bus 46), controller 40 determines the desired motor phase currents and provides such currents through a bus 48.

Referring now to FIG. 2, a block diagram of controller 40 is illustrated. A standard six-switch bridge circuit 50, (shown in more detail in FIG. 3) switches armature current to the wye connected motor phase windings A, B and C (FIG. 4).

The drive 52 provides the switching signals to control gates of the switches and bridge circuit 50 in response to certain input signals. The drive 52 receives two pulse width modulation voltage signals PWM and XPWM from a PWM circuit 54 (shown in more detail in FIG. 5). The drive 52 also receives commutation initiating signals H.sub.1, H.sub.2 and H.sub.3 from Hall effect rotor position signal generators in motor 34 and a decreasing modulation timing signal XX from a commutation control 56. Gate drive 52 further receives a switching motor control signal MODE, a motor direction control signal DIR and a decreasing modulation enable signal XEN from a control computer 58.

A current sense circuit 60 (shown in more detail in FIG. 6) receives current sense signal voltages V.sub.hi and V.sub.lo from a current sense resistor Rs in bridge circuit 60 and, with the assistance of a special sample and hold circuit, provides an output motor current signal Is. Current sense circuit 60 also receives signals PWM and XPWM from PWM circuit 54 and a signal SPEED from control computer 58. A current control circuit 62 (shown more detail in FIG. 7), receives the motor current signal Is from current sensor circuit 60, a current command signal Ic from control computer 58 and a signal SLOPE from PWM circuit 54 and provides an output timing signal CMRST to PWM circuit 54. Commutation control circuit 56 provides an output timing signal XREF to PWM circuit 54 in response to an input initial value signal INIT from PWM circuit 54, a time constant signal Tc and a modulation and signal ENDX from control computer 58, and a commutation initiating signals H.sub.1, H.sub.2 and H.sub.3 from motor 34.

Commutation control circuit 56 further provides the signal XX to gate drive circuit 52. Control computer 58 includes a digital computer programs for control of motor 34.

Although not shown, standard oscillator based circuitry provides a basic clock signal CLK for synchronous operation of control computer 58 and the other digital circuitry. Control computer 58 receives the input vehicle speed, handwheel torque, steer angle and motor position (H.sub.1, H.sub.2 and H.sub.3) signals previously mentioned, and derives from these and certain predetermined constants the commanded motor current Ic, as well as the following other output signals: (1) a motor speed signal SPEED; (2) a reference count MAX DUTY for pulse width modulation; (3) a motor direction signal DIR; (4) a commutation mode signal MODE; (5) a decreasing modulation enable signal XEN; (6) a decreasing modulation time constant Tc; (7) a decreasing commutation end reference count ENDX; and (8) a count divider number DIV.

FIG. 3 shows a circuit diagram of a typical bridge circuit comprising six semiconductor switches Q1-Q6 with associated parallel fly-back diodes D1-D6. The switches may be transistors, FET's or similar semiconductor switching devices, and the diodes may be built into the switches in an manner known in the art. Each of switches Q1-Q3, which may be called "upper" switches, is connected in series with one of switches Q4-Q6, which may be called "lower" switches, across a voltage regulated, DC electric power source B+ and a series current sensing resistor R, with the junction of each pair of upper and lower series connected switches connected to a different one of the three motor phase winding terminals in the wye connected motor armature windings. In particular, junction Ja of switches Q1 and Q4 is connected to terminal P of motor phase winding A; junction J of switches Q2 and Q5 is connected to terminal P of motor phase winding B; and junction J of switches Q3 and Q6 is connected to terminal P of motor phase winding C. The diodes, being connected in parallel with the switches, are similarly connected with respect to the motor phase winding terminals.

Current in the motor phase windings is provided by activating an upper switch and a non-series lower switch to close a circuit with the DC electric power source, two phase windings and current sensing resistor R in series. The motor phase windings are electronically commutated to produce motor operation in a predetermined direction of rotation by changing the activated pairs of closed switches in a predetermined pattern well known in the art. Motor currents in the armature phase windings are also controlled by pulse width modulating (PWM) the activated switches at a frequency higher than that of commutation. This PWM switching can be done in several different ways, which results in several different modes of operation of the bridge circuit.

In a first mode of operation, which may be termed dual switching mode, pulse width modulation is provided by switching both the upper and lower activated switches together in the standard manner except during commutation. During commutation, three switches are activated. The newly activated switch on the same level (upper or lower) as the switch being deactivated is pulse width modulated together in the standard manner with the switch on the other level that remains activated; but the switch being deactivated is also pulse width modulated, each pulse of the deactivating switch being initiated synchronously with the others but varying from the others in duration with a controlled, decreased duty cycle.

The preferred manner of decrease in duty cycle of the deactivated switch is exponential, since this tends to produce a linear decrease in current. The newly activated switch is alternated between the upper and lower groups. The pattern shown produces motor rotation in a predetermined direction; the pattern for motor rotation in the opposite direction would be apparent to one of ordinary skill in the art.

An example of operation in dual switching mode is described with reference to FIG. 9. The condition of each of switches 1-6 is shown throughout one complete electrical cycle of the motor. For each switch, an activated condition is shown high and an inactivated condition is shown low. While each switch is activated, standard pulse width modulation is shown with single direction cross-hatching and decreasing pulse width modulation is shown with double direction cross-hatching. The non-commutating periods of motor operation (AB, CB, . . . ) are shown separated by unlabeled commutation periods, the duration of latter being exaggerated for clarity. During the non-commutating period AB, upper switch Q1 and lower switch Q5 are activated and pulse width modulated together until commutation. During the following commutation period, upper switch Q3 becomes activated and is pulse width modulated with lower switch Q5; but the pulse width modulation of switch Q1, although continuing with pulses initiated synchronously with the others, is given a decreasing duty cycle. When the decreasing duty cycle of switch Q1 is ended, the switch is deactivated to end the commutation period. The next non-commutating period CB is similar, but with switches Q3 and Q5 pulse width modulated together until the next commutation period, during which Q4, the next switch turning on, is pulse width modulated with switch Q3, the switch remaining on, and Q5, the switch turning off, is pulse width modulated with a decreasing duty cycle. The pattern repeats with appropriate switches activated and modulated as shown in FIG. 9 to complete a cycle through additional non-commutating periods CA, BA, BC and AC and the associated commutation periods.

In a second mode of operation, which may be termed modified single switching mode, pulse width modulation is provided by switching only the upper activated switch, leaving the lower activated switch continuously on, except during commutation. During commutation, three switches are activated: the newly activated switch (upper or lower) is pulse width modulated; the switch remaining activated is turned on continuously, and the switch being deactivated is pulse width modulated with pulses being initiated synchronously with the standard pulses but having a duration varying with a controlled, decreasing duty cycle. The term "modified" in modified switching mode thus refers to the fact that, although during the non-commutating periods the lower switch is always left continuously on, during commutation the switch left continuously on alternates between an upper switch and a lower switch. Thus, operation during commutation is symmetric between the upper and lower switches. This is different from the traditional single switching mode of the prior art, in which the lower activated switch is always left on continuously and modulation is always applied to the upper activated switch.

An example of operation in modified single switching mode is described with reference to FIG. 10. During the non-commutating period AB, upper switch Q1 and lower switch Q5 are activated, with upper switch Q1 being pulse width modulated and lower switch Q5 continuously on until commutation. During the following commutation period, upper switch Q3 becomes activated and is pulse width modulated, lower switch Q5 remains continuously on and deactivating switch Q1 is pulse width modulated synchronously with a controlled, exponentially decreasing duty cycle. During the next non-commutating period CB, upper switch Q3 is pulse width modulated and lower switch Q5 remains continuously on. During the following commutation period, lower switch Q4 becomes activated and is pulse width modulated during the commutation period, upper switch Q3 is turned on continuously for the commutation period and the deactivating switch Q5 is pulse width modulated synchronously with switch Q4 but with a controlled, exponentially decreasing duty cycle. During the next non-commutating period CA, lower switch Q4 is turned on continuously while upper switch Q3 is pulse width modulated. The pattern repeats with appropriate switches activated and modulated as shown in FIG. 10 to complete a cycle through additional non-commutating periods BA, BC and AC and the associated commutation periods.

The inventors of this apparatus and method have found that a controlled, decreasing modulation of the switch in the phase turning off can effectively control the change in phase currents in the motor during commutation to eliminate the objectionable "zipper" noise. Thus, controller 40 provides a controlled, decreasing pulse width modulation of the switch turning off, beginning synchronously with the first PWM pulse of the switch turning on following