WO2013060327A1

WO2013060327A1 - A method for operating a mechanically commutated electric motor

Info

Publication number: WO2013060327A1
Application number: PCT/DK2011/050404
Authority: WO
Inventors: Preben Bo Fich; Curt Michael Petersen Willadsen
Original assignee: Ideassociates (Iom) Limited
Priority date: 2011-10-25
Filing date: 2011-10-25
Publication date: 2013-05-02

Abstract

A method for operating a brush commutated electric motor(1). The method comprising the steps of supplying a pulse width modulated supply signal having a first predetermined frequency to said brush commutated motor, and adjusting the duty cycle of said pulse width modulated supply signal to a desired value. The method also comprises the further step of adjusting the frequency of the pulse width modulated signal to a second frequency differing from said first predetermined frequency, while maintaining the desired value of the duty cycle.

Description

A method for operating a mechanically commutated electric motor

The present invention relates to a method for operating a brush commutated electric motor. Brush commutated motors, i.e. electric motors with mechanical commutation electric motors are widely used. In many modern-day applications the power is supplied to the motor as pulse width modulated supply signal. Pulse width modulation allows the power supplied to the motor to be regulated by varying the duty cycle of the supply signal.

The frequency of the supply signal is generally a fixed frequency selected appropriately based on criteria such as prevention of audible and electromagnetic noise.

In US-A-7265514 it is suggested to operate an electric motor at different pulse width modulation frequencies such as low frequencies of 100 Hz or 300 Hz and at high frequencies of 25 kHz and 100 kHz, which are several magnitudes higher than the low frequencies. This is done by using a compensation circuit allowing the full desired duty cycle to be maintained both at the low frequencies and the high frequencies. US-A-7265514 refers to tests of the motor, the type of which is not specified, and notices that over the wide range of frequencies 300 Hz, 25 kHz and 100 kHz there are slight differences in motor speed for a given duty cycle, but makes the further remark that in the tests no unwanted variation is measured if the pulse width frequency is above 100 Hz, thus verifying that this circuit provides the full desired duty cycle even at high frequencies.

The present invention takes a completely different approach to frequency variation of a pulse width modulated supply signal for brush commutated motors. Rather, the approach of the present invention to frequency variation of the pulse width modulated supply signal to a brush commutated motor relies on the discovery by the inventors that the brush commutated mo- tor does not commutate independently of the frequency of the pulse width modulated signal. During continued investigation on detection of commutation spikes on the supply wires of brush commutated motors, as disclosed inter alia in the applicant's patent application WO-A-2010/040349 and the applicant's patent EP-B-1929623, it was discovered by the inventors that brush commutated motors have a distinct tendency to commutate during the power- off periods between the pulses of the pulse width modulated supply signal. In turn, this realization led to the thought that if the commutation of the motor is influenced by the power-off periods, then the timing of the commutation, and hence the speed of the motor, could be controlled by varying the frequency of the pulse width modulated signal. The inventors have found that this is indeed possible, and that this can be implemented in a method for operating a brush commutated motor.

Thus according to the invention there is provided a method for operating a brush commutated electric motor, said method comprising the steps of supplying a pulse width modulated supply signal having a first predetermined frequency to said brush commutated motor, adjusting the duty cycle of said pulse width modulated supply signal to a desired value, wherein said method comprises the further step of adjusting the frequency of the pulse width modulated signal to a second frequency differing from said first predetermined frequency, while maintaining the desired value of the duty cycle.

This inter alia allows the fine tuning of the pulse width modulated signals with respect to the actual rotational speed, i.e. the commutation frequency of an electric motor. By tuning the pulse width modulated signals to the motor commutation frequency it can be achieved that most or even all commutation spikes appear in power-off periods of the pulse width modulated signals without changing the rotational speed of the motor. Evidently, this leads to less arcing between the commutators and brushes and thus less wear, but it also leads to an improved efficiency of the motor as the current needed to provide the given torque at the actual rotational speed is also lowered. Conversely, the change in frequency may, at least within reasonable limits, also be used to actually increase the rotational speed of the motor without increasing the power consumption.

According to a first preferred embodiment, the pulse width frequency is adjusted continuously. Thereby the operation of the brush commutated electric motor may continuously adjust to an optimum in terms of efficiency.

According to a further preferred embodiment, the frequency is adjusted within an interval around the predetermined frequency, preferably from 92% to 108% of the predetermined frequency, more preferred from 95% to 105% of the predetermined frequency. This turns out to be sufficient for optimising the operation of the brush commutated electric motor in terms of efficiency.

According to yet another preferred embodiment, wherein a commutation interval time is determined based on detection of commutation spikes on the supply wire, the frequency of said pulse width modulated signal is adjusted to said second frequency based on said detected commutation interval time. This has been found to be a very precise and efficient way of keeping track of the timing of the commutations, which in particular does not necessitate any further wires to the brush commutated electric motor.

According to a further preferred embodiment, one duty cycle represents one commutation. This makes the control and adjustment process very simple.

According to yet a further preferred embodiment, the brush commutated electric motor is a phase neutral brush commutated electric motor, and the polarity of the pulse width modulated supply signal is selectable. Thereby the emulated electric lead angle may be introduced for either direction of rotation of the brush commutated electric motor.

The invention will now be described in greater detail based on non- limiting exemplary embodiments and with reference to the drawings, on which:

Fig. 1 schematically illustrates a microprocessor based circuitry for implementing the method according to the invention,

Figs. 2a and 2b show flow charts illustrating an implementation of the present invention,.

Fig. 3 shows a pulse width modulated pulse train supplying an electric motor, and

Fig. 4 shows a section of the pulse width modulated pulse train of Fig. 3 at a larger time scale. The present invention takes its origin in the discovery by the inventors that the brush commutated motor does not commutate independently of the pulse width modulated signal. During continued investigation of detection of commutation spikes on the supply wires of brush commutated motors, as disclosed inter alia in the applicant's patent application WO-A-2010/040349 and the applicant's patent EP-B-1929623, the inventors were performing the following experiment:

A Mabuchi type 578 brush commutated electric motor was supplied with pulse width modulated current from a 12V automotive battery. The pulse width modulated current is modulated by a 40 A rated PMOS FET transistor with built-in reverse diode protection and coupled as low side switching. The modulation of the switching in turn was controlled by an external HP signal generator providing 10 kHz, 50 % duty cycle gate voltage to the PMOS FET transistor. A Picoscope type 2200 was coupled to the terminals of the motor in order to record the pulse width modulated pulses over an appropriate time. Fig. 3 shows a section of the recorded pulse width modulated pulse train.

In this pulse width modulated pulse train the inventors noticed a number of recurring spikes, marked with circles. These recurring spikes were identified as commutation spikes by the inventors. One such spike is shown in Fig. 4, also marked with a circle. Wondering, in Fig. 3, that all three commutation spikes were readily identifiable at the leading flank of a square pulse of the width modulated pulses, further investigation was carried out. Inter alia the frequency supplied by the external HP signal generator was varied up and down from 10 kHz. Irrespective of this, the commutation spikes remained readily identifiable in the power-off intervals between the pulse width modulated pulses.

Further investigation led to the discovery by the inventors that commutation spikes of brush commutated motors have a distinct tendency to occur during the power-off periods between the pulses of the pulse width mod- ulated supply signal.

Even further investigation led to the discovery that varying the frequency led to a corresponding change in motor speed, even though duty cycle and the power consumed by the motor was maintained. The present invention suggests the industrial application of this phenomenon in the control of brush commutated electric motors.

To this end, Fig. 1 schematically illustrates a simple circuit diagram for controlling a brush commutated electric motor 1 by means of a microproc- essor 2. The microprocessor 2 supplies pulse width modulated supply voltage on a supply wire 3 to the brush commutated electric motor 1 . A filter 4 is connected to the supply wires 3, 5 of the brush commutated electric motor in order to detect commutation spikes on the supply wires 3, 5. As an alternative to the filter 4 being connected in parallel with the brush commutated electric motor 1 as shown, the filter could instead be connected in parallel with a small resistance inserted in series with the brush commutated electric motor in either of the supply wires 3 or 5. Upon detection of a commutation spike the filter 4 provides an input signal on line 6 serving as an input to the microprocessor 2, in turn, allowing a real time clock in the microprocessor 2 to detect elapsed time between two consecutive commutation spikes, and adapt the frequency of the pulse width modulated voltage, as will be explained below. It should be noted that the circuitry of Fig. 1 is only a simple example illustrating the basic principles. The skilled person would know that for e.g. higher voltages and currents the microprocessor 1 would not be used to supply the cur- rent to the brush commutated electric motor, but instead control appropriate drivers for the brush commutated electric motor 1 . Means for providing reverse polarity supply voltage could also readily be implemented by the skilled person.

In this respect, Figs. 2a and 2b illustrate an algorithm for adapting the frequency of the pulse width modulated supply voltage to the brush commutated electric motor in order to electrically emulate a lead angle for the commutation in a brush commutated electric motor without a physical lead angle, in the following referred to as a phase neutral motor.

In box 100 the method is initialized. Then, in box 101 , the supply of pulse width modulated current to the brush commutated electric motor at a first predetermined frequency F1 starts. The pulse width modulated current is supplied to the brush commutated electric motor until first commutation is detected in box 102. This detection is preferably performed by detection of spikes on the supply wires to the brush commutated electric motor as inter alia disclosed in WO-A-2010/040349 and EP-B-1929623, incorporated herein by reference. Then, in box 103, a first real time counter N1 is reset. As will be understood from the following, N1 is a real time counter representing the time between subsequent detections of commutations, i.e. a commutation interval time. Such detections are performed in the same manner as the first detection, i.e. by detection on spikes on the supply wires to the brush commutated electric motor.

With the real time counter N1 reset in box 103, then, in box 104, pulse width modulated current starts being supplied to the brush commutated motor again. Having started the supply of pulse width modulated current to the brush commutated electric motor, the current and the voltage values, indicating inter alia the power supplied to - and thus consumed by - the electric motor, are measured in box 105.

The supply of pulse width modulated current to the brush commutated electric motor continues until the next commutation is detected in decision box 106, upon which the elapsed time since the last commutation was detected is determined in box 107 by reference to the real time counter N1 .

After the elapsed time since last commutation has been determined, a look up for a commutation interval time T1 is performed in box 108. The look up is preferably performed in a table having tabulated values for representing the expected time between two subsequent commutations for the brush commutated electric motor if it was running at the same power, i.e. the same voltage and current measured in box 105. Alternatively, if sufficient computing power is available, the value of T1 could, in box 108, be calculated based on the voltage and current measured in box 105, rather than looked up in at table.

However, it is well known that if the brush commutated electric motor was not constructed without a physical lead angle (also known as brush ad- vance), it would have a forward direction for one polarity of supply current, where it would be more efficient, and a reverse direction for the opposite polarity, where it would be less efficient than a phase neutral motor. That is to say, an otherwise identical brush commutated motor with a lead angle would run somewhat faster in the forward direction than the phase neutral counterpart motor. Accordingly, the time between commutations would be shorter than T1 , e.g. by a value K0 corresponding to a fraction of the angle between two commutators, and hence a fraction of the time between two commuta- tions.

Accordingly, if N1 <=T1 -K0 then the brush commutated electric motor is running slower than optimum, optimum being understood as the speed achievable with the comparable brush commutated electric motor, but with a physical lead angle. This is checked in box 109. If the brush commutated electric motor is not running slower than optimum, the pulse width power is supplied unchanged, by returning to box 104 and repeating the steps of boxes 104 through 109 in the flow chart of Figs. 2a and 2b.

If the brush commutated electric motor is running slower than optimum, it could be because the motor is in the process of stopping after a stop command has been issued. This is checked in box 1 10, and if a stop command has indeed been issued the method proceeds to end in box 1 1 1 . If, on the other hand, the brush commutated electric motor is not in the process of stopping, the brush commutated electric motor is running sub-optimal, as it is known that if it had a physical lead angle it would be more efficient, i.e. run- ning faster at a higher rpm.

Having discovered, however, that the commutation spikes of the brush commutated electric motor have a distinct tendency to occur during the power-off periods between the pulses of the pulse width modulated supply signal, the method now, in box 1 12, adjusts the frequency of the pulse width modulated signal to a second frequency F2 differing from said first predetermined frequency F1 , while maintaining the desired value of the duty cycle. This would so to speak introduce an electric lead angle allowing the motor to run faster, even though it is still supplied with the same amount of power - the duty cycle being unchanged.

The second frequency F2, to which the frequency of the pulse width modulated frequency should be adjusted, evidently relates to the commutation period, i.e. the time between two commutations. Since the method aims at an optimum where N1 =T1 -K0, the second frequency F2 would, in the sim- plest case where there is one duty cycle per commutation, be calculated as F2=1/(T1 -K0). Normally, however, the first predetermined frequency F1 would be substantially higher, and correspond to several duty cycles per commutation. In this case, the adjusted frequency would be a corresponding multiple higher, calculated as F2=n/(T1 -K0), where n is the resulting number of duty cycles per commutation when the motor is running at steady speed.

It should be remembered that the method described above is only one example embodying the use of the discovery. In this specific example the main factor controlling the speed of the brush commutated electric motor is still the power supplied to it. That is, the adjustment of the value of the duty cycle prior to and/or independently of the method disclosed in Figs. 2a and 2b. The method above aims not at controlling the overall speed of the brush commutated electric motor, but only at adjusting the speed within intervals around a more or less steady state given by the power supplied to the brush commutated electric motor, in order to improve the efficiency of the running motor. It has been found that the method works at least for the interval of 20% duty cycle to 80% duty cycle. It is likely to work beyond that interval, but evidently not for 0% or 100% duty cycle.

The second frequency F2 does thus not differ from the first prede- termined frequency F1 by orders of magnitude. Rather, as it is the aim to introduce an electric lead angle, the frequency is only changed within relatively narrow intervals around the first predetermined frequency F1 , preferably from 92% to 108% of the first predetermined frequency F1 , more preferred from 95% to 105% of the first predetermined frequency F1 .

In such cases it would under circumstances make sense to alter the second frequency F2 by using another more suitable value of n, e.g. in increments or decrements of 1 , so as to optimize timing of the power-off periods of a duty cycle to the desired electrical lead angle.

It should be noted that one of the advantages of the method above is that the introduction of the electric lead angle works for either polarity of the supply current, in turn allowing the phase neutral motor to be optimized equally efficient irrespective of direction of operation. Thus unlike a brush commutated electric motor constructed with a physical lead angle the method allows the neutral motor to be efficient in both directions.

Claims

P A T E N T C L A I M S

1 . A method for operating a brush commutated electric motor, said method comprising the steps of

supplying a pulse width modulated supply signal having a first prede- termined frequency to said brush commutated motor,

adjusting the duty cycle of said pulse width modulated supply signal to a desired value, wherein

said method comprises the further step of

adjusting the frequency of the pulse width modulated signal to a sec- ond frequency differing from said first predetermined frequency, while maintaining the desired value of the duty cycle.

2. A method according claim 1 , wherein the frequency is adjusted continuously.

3. A method according to claim 2, wherein the frequency is adjusted within an interval of around the first predetermined frequency, preferably from

92% to 108% of the first predetermined frequency, more preferred from 95% to 105% of the first predetermined frequency.

4. A method according to any one of the preceding claims, wherein a commutation interval time is determined based on detection of commutations spikes on the supply lead, and the frequency of said pulse width modulated signal is adjusted to said second frequency based on said detected commutation interval time.

5. A method according to any one of the preceding claims, wherein one duty cycle represents one commutation.

6. A method according to any one of the preceding claims, wherein the brush commutated electric motor is a phase neutral brush commutated electric motor, and the polarity of the pulse width modulated supply signal is selectable.