GB2505488A

GB2505488A - Driving a sensorless brushless DC motor

Info

Publication number: GB2505488A
Application number: GB1215576.8A
Authority: GB
Inventors: Viktor Petri; Thomas Freitag
Original assignee: Melexis Technologies NV
Current assignee: Melexis Technologies NV
Priority date: 2012-08-31
Filing date: 2012-08-31
Publication date: 2014-03-05
Also published as: GB201215576D0

Abstract

A sensorless, brushless DC or permanent magnet synchronous motor 2, comprises at least three stator windings U, V, W connected in a star configuration. Two of the motor windings U, V are energized while leaving a third motor winding W unenergized and the back-EMF voltage generated in the unenergized winding is monitored. When the monitored voltage (V1) reaches a local maximum or minimum a commutation occurs to energize two windings U, W while leaving a third winding V unenergized according to a subsequent motor state of the energizing sequence. Alternatively commutation may occur at the end of a first time period (T1), a second voltage (V2) may be measured at a time interval after commutation and a subsequent time period (T2) calculated, based on the first time period T1, the first monitored voltage V1and the second monitored voltage V2.

Description

METHOD AND APPARATUS FOR DRIVING A SENSORLESS BLDC/PMSM MOTOR

Field of the invention

The invention relates to the field of sensorless brushless direct current (BLDC) motors. In particular, the present invention provides a method and apparatus for starting and running such a motor.

Background of the invention

Brushless Direct Current (BLDC) motors are a type of electrical motor having multiple status phases (e.g. three or more) and a permanent magnet as a rotor (e.g. a bar magnet or a multi-pole magnet). RLDC motors do not use brushes for commutation, but instead they are electronically commutated. This implies, however, that the drive circuit must know the relative position of the rotor with respect to the stator, in order to direct the magnetic field for exerting torque on the rotor.

One kind of BLDC motors uses sensor devices such as e.g. optical sensor devices or Hall elements positioned around the rotor for determining the relative position using a IS resolver This, however, has the disadvantage of increased complexity, extra component cost and decreased reliability, especially in harsh motor environments.

Another kind of BLDC motors does not use such sensors, but uses the back-EME (BEME) generated in the windings for determining the rotor position. The shape of REMF can e.g. be trapezoidal or sinusoidal, depending on how the windings are located on the stator.

The working principles of synchronous brushless DC motors based on back-EMF are well known and well described in the art, for example in US4275343 (filed in 1978, almost 25 years ago) and U54455513 (filed in 1982) and many others. Several of these other applications are based on integration of the back-EMF signal, or on the zero-crossing of the back-EMF signal for determining the relative rotor position. A disadvantage of such methods is that they require additional and often complex circuitry, such as e.g. integrators, comparators, filters, etc. In US7737651B2, De Four discloses a principle for determining the rotor position based on space vector theory. Resides from being highly theoretical and complex, the execution of this algorithm requires a powerful DSP.

US20030062860 describes a control system operating with a PWM speed regulation of a sensorless BLDC motor with a plurality of windings, and a hybrid method utilizing BEMF induced in stator windings to indicate the commutation instant and a falling edge detection method to override PWM chopping and commutation noises. The circuit and software are alleged to be relatively concise and low cost. Yet, the control device includes an analog filter and a comparator for generating the BEMF zero crossing points (ZCP). Other functions such as detecting the falling edge of the generated ZCPs may be implemented in an integrated controller.

Summary of the invention

It is an object of embodiments of the present invention to provide a good method and device for driving a sensorless brushless DC (BLDC) motor, e.g. a sensorless permanent magnet synchronous motor (PMSM).

In particular, it is an object of embodiments of the present invention to provide a method and device for driving a sensorless brushless DC motor or a sensorless permanent magnet synchronous motor with reduced complexity.

It is an advantage of a method according to embodiments of the present invention that they are conceptually very simple.

It is an advantage of a method according to embodiments of the present invention that they can be implemented by relatively simple hardware and software.

This objective is accomplished by a method and a device according to embodiments of the present invention.

In a first aspect, the present invention provides a method for driving a sensorless brushless DC (BLDC) motor or a permanent magnet synchronous motor (PMSM) in a predetermined direction, e.g. clockwise or counter-clockwise, the motor comprising at least three stator windings connected in star configuration, and a permanent magnet rotor. The method comprises: a) energizing two of the windings and leaving a third winding un-energized for rotating the rotor in the predetermined direction, based on a first motor state of an energizing sequence corresponding to a known rotor position; b) monitoring a voltage representative of the back-EME generated in the un-energized winding and determining a time instance when that voltage reaches a local extreme (local maximum or a local minimum); c) applying at the determined time instance a commutation so as to energize at least two of the windings while leaving a third winding un-energized according to a subsequent motor state of the energizing sequence; and d) repeating steps b) and c).

A method according to embodiments of the present invention assumes that the rotor position is (approximately) known in step a). It is noted that an exact angular position is not required, it suffices to know roughly how the rotor is oriented (e.g. at an accuracy of +1-15°).

This can be assured at start-up e.g. by forcing the rotor in a desired position, and waiting for a sufficiently long time. This is also assured once the motor is running, because the motor is driven synchronously, i.e. the motor state is changed according to the actual rotor position, which can be determined by monitoring the back-EMF of the un-energized winding. The underlying idea behind this algorithm is the fact that the back-EMF-voltage increases and reaches a local maximum, or decreases and reaches a local minimum, hence changes and reaches a local extreme, if the windings are energized longer than in the ideal case.

It is an advantage of a method according to embodiments of the present invention that it can be implemented by very simple hardware and software, e.g. using a simple microcontroller with at least one analog-to-digital convertor. In this way complex and/or expensive hardware components such as e.g. analog filters, comparators, integrators, resolvers, etc, can be omitted. In addition, since no complex mathematical functions (such as e.g. goniornetric functions) need to be calculated, but only a rnaximum() and/or minimum() function are required, a powerful processor such as e.g. a high-end DSP (digital signal processor) is not required, but instead a controller with a simple processing unit is sufficient.

In this way the hardware and software can be simplified, and the design and implementation effort and complexity can be significantly reduced.

Whereas prior art methods go through large efforts to avoid missing the "ideal" moment of commutation, (which occurs at multiples of 60 degrees for a three-phase motor), the present invention simply accepts the fact that a commutation event "shortly after the ideal moment" is still a "very suitable" moment for many applications, such as e.g. fan control, oil-, fuel-, water-or air-pumps, blowers or compressors. The only disadvantage is that the motor may develop a slightly lower average torque, and a slightly higher energy dissipation (e.g. 1 to 5%) as compared to prior art solutions. However, the benefits of simpler hardware and software far outweigh these disadvantages, e.g. in high-volume electronic applications, such as e.g. automotive applications, or in some harsh environments, where guaranteed and robust operation may be far more important than "smooth" rotation.

Moreover, a method according to embodiments of the present invention offers the advantage that it automatically takes into account any motor skew, while such skew should be explicitly corrected for in some known applications.

In a method according to embodiments of the present invention, the monitoring of the voltage in step b) may only be performed during a fraction of the time period between two successive commutations.

It is an advantage that the monitoring is disabled for a portion of the time period between two commutations, e.g. at the beginning of such a time period, because in this way energy can be saved in a very simple way, e.g. the energy dissipated by the at least one analog-to-digital convertor. The period may be a predefined constant period, or may be dependent on the motor speed, e.g. as a percentage of the time period between successive commutation events, which time period is relatively stable once the motor is running at nominal speed, and assuming the load is relatively constant at a given speed, as is the case e.g. for a fan.

In a method according to embodiments of the present invention, the energizing of at least one of the windings may be operated at least part of the time using a PWM-signal.

It is an advantage of using a PWM-signal for energizing the windings during at least part of the time-slot between two commutation events, that it allows the motor speed to be controlled by simply adapting the PWM-duty-cycle, without having to adapt the supply voltage. This can be implemented with only minimal processing load (e.g. only activating and deactivating the PWM-module).

In a method according to embodiments of the present invention, the energizing of the windings may be operated at least part of the time with DC voltages (VDD, GND).

It is an advantage of using a DC voltage for energizing the windings during at least part of the time-slot between two commutation events, that it allows a large motor torque to be exerted upon the motor, which reduces the risk of the motor being blocked. It is advantageous to use a DC-voltage instead of a PWM-signal during at least the second half of a time-slot, as this may reduce the spikes occurring on the voltage to be measured, and thus may reduce the risk of false detection of the maximum or minimum. This can be implemented with only minimal processing load (e.g. setting an output pin to logic "1" or logic "0"), or by changing the state of only two output-pins in case an external H-bridge is used and two external transistors are involved.

In a second aspect, the present invention provides a method for driving a sensorless brushless DC (BLDC) motor or a permanent magnet synchronous motor (PMSM) in a predetermined direction (e.g. clockwise or counter-clockwise), the motor comprising at least three stator windings connected in star configuration, and a permanent magnet rotor. The method comprises: a) determining a time period as the current time period, and energizing during the current time period two of the windings and leaving a third winding un-energized for rotating the rotor in a predetermined direction, based on a first motor state of an energizing sequence, corresponding to a known rotor position; b) measuring a first voltage representative of the back-EMF generated in the un-energized winding at a first time interval before expiry of the current time period; c) applying a commutation at expiry of the current time period so as to energize at least two of the windings while leaving a third winding un-energized according to a subsequent motor state of the energizing sequence; d) measuring a second voltage at a second time interval after the commutation, and calculating a subsequent time period based on at least the current time period and the measured first and second voltage; and e) repeating steps b) and c) after replacing the current time period by the subsequent time period.

This is a second embodiment of the present invention, based on the same principles that the "exact" moment of commutation is less important than simplicity of the circuit.

It is an advantage of this second embodiment that continuous monitoring of the back-EME voltage, and monitoring of its local maximum or local minimum are not required.

Instead, a commutation is forced after an estimated time interval, the duration of which is calculated based on the duration of a previous time period, and on values of the back-EME, measured at two specific time-instances, in particular shortly before and shortly after the preceding commutation event. Such method can be easily and efficiently implemented in an interrupt service routine, using timer-interrupts.

As for the method of the first embodiment, it is an advantage of this second method embodiment that it can be implemented by very simple hardware and software, e.g. using a single microcontroller having at least one analog-to-digital convertor and a timer. In this way complex and/or expensive hardware components such as e.g. analog filters, comparators, integrators, resolvers, etc, can be omitted. In this way the hardware and software can be simplified, and the design and implementation effort and complexity can be reduced.

This method assumes that the rotor position is (approximately) known in step a), for which there are ways described in the prior art, for example in U57944159B2. It is noted that an exact angular position is not required, it suffices to know roughly how the rotor is oriented (e.g. with an accuracy of +1-15°). This can be assured at start-up e.g. by forcing the rotor in a desired position, and waiting for a sufficient time. This is also assured once the motor is running, because the motor is driven synchronously, i.e. the motor state is changed according to the actual rotor position, which can be determined by monitoring the back-EMF of the un-energized winding.

The underlying idea behind embodiments of the present method is that the difference between the first and second value can be used to correct the timing of the commutation events, because ideally, there is no difference.

The first time interval is preferably chosen as small as possible (ideally zero). The second time interval should be chosen sufficiently large in order that a valid BEMF-signal is read, and any transient behaviour due to the commutation events (current decay, spikes) have died out. This period may vary with motor speed, but can be dynamically determined in software. It may be pre-determined (e.g. using a look-up table), but preferably is dynamically determined by monitoring the feedback-signal, and by ignoring invalid data.

In a method according to embodiments of the present invention, the first time period in step a) may have a predetermined value.

This can be seen as open-loop motor control, whereby the timing of one or more commutations is predetermined. As will be described further, only a single open-loop period may suffice, in contrast to some prior art methods, where open-loop control is applied up to about 20% of the nominal motor speed.

In such method, the first time period in step a) may be determined by monitoring a voltage representative of the back-EMF generated in the un-energized winding, and by determining the first time period as the time instance when that voltage reaches a local extreme (local maximum or a local minimum).

In this way the first time period may be dynamically determined in a similar way as the first algorithm, which may take better account of e.g. an external load.

In a method according to embodiments of the present invention, each subsequent time period may be calculated based on the preceding time period by using a formula that is directed at reducing the difference between the first and the second voltage before and after the subsequent commutation.

In a method according to embodiments of the present invention, the second time period immediately following the first time period immediately after motor start-up may be calculated according to the formula: /2 = Ti * (1 _J V1WV2W); A vit'l) wherein T2 is the second time period, Ti is the first time period, vi is the first voltage measured before the first commutation and v2 is the second voltage measured after the first commutation, and A is a predetermined value in the range of 0,25 to 0,75.

iO This formula can be used to calculate the duration of the second time period after motor start-up. Alternatively, a second predetermined (constant) time-period may also be used. As can be seen, this formula uses the value of the previous time-period, and two measurement values "around" the first commutation, thereby immediately correcting the first previous time-period.

iS In embodiments of the present invention, the value of A may ly in the range of 0,40 to 0,60. In particular embodiments, the value of A may ly in the range of 0,45 to 0,55. In embodiments, the value of A may be about i/2.

In a method according to embodiments of the present invention, the subsequent time period may be calculated according to the formula: 1 vi(n) -vl(n) -2* v(n -1) T(n+1)=T(n)*(1--* -) A vi(n) -v2(n-1) wherein T(n) is the subsequent time period, T(n-i) is the current time period, vi(n) and v2(n) is the first resp. second voltage measured before and after the most recent commutation, and v2(n-i) is the second voltage measured after the commutation preceding the most recent commutation, and A is a predetermined value in the range of 0,25 to 0,75.

By using this simple formula, the duration of the motor cycle can be determined based on the previous time period and only three voltage measurements.

It is to be noted that the formula for the second time-slot can be seen as a special case of the general formula presented here, when taking zero for the second value of the previous commutation. Thus only a single formula needs to be implemented in the controller, thereby reducing complexity.

In embodiments of the present invention, the value of A may lie in the range of 0,25 to 0,75. In embodiments, the value of A may ly in the range of 0,40 to 0,60. In embodiments, the value of A may ly in the range of 0,45 to 0,55. In embodiments, the value of A may be about 1/2.

This offers the same advantages as were mentioned when using a PWM-signal in the first embodiment. However, in order not to disturb the measurement of the first voltage before the commutation, it may be good to stop energizing with the PWM-signal well before taking that measurement.

In a method according to embodiments of the present invention, the energizing of the windings may be operated at least part of the time using DC voltages (such as VDD, GND).

This offers the same advantages as were mentioned when using DC-voltages in the first embodiment. Such energizing signals do not disturb the measurement of the first voltage before the commutation, and may thus be applied while taking that measurement.

In a further aspect, the present invention provides an electrical circuit as can be used for performing the method according to any of the method embodiments, the electrical circuit comprising: -a BLDC motor or a PMSM motor, the motor comprising at least three stator windings connected in star configuration and a permanent magnet rotor; -a controller connected to the motor, the controller comprising: -at least three output pins connected to the windings of the motor for energizing at least two of the windings and for leaving a third winding un-energized; -at least four input pins connected to the windings of the motor for measuring the voltage representative of the back-EMF generated in the un-energized winding, and connected to a reference signal; -at least one analog to digital convertor (ADC) for digitizing the measured voltage and for digitizing the reference signal; -a calculation unit (10); wherein the controller is provided with an algorithm for performing the method according to any of the method embodiments.

It is an advantage of such a circuit that, apart from an optional output buffer stage for delivering power to the motor, and an optional voltage divider stage for reducing the voltage to the input pins, and an optional resistor stage for generating a virtual star point, and a voltage divider for generating the half supply voltage, it does not require complex analog circuitry, such as e.g. an analog filter, an analog integrator, comparators or a resolver, etc. Suitable controllers are e.g. programmable digital or hybrid micro-controllers, but can also be made as dedicated ASICs (application specific ICs).

In yet another aspect, the present invention provides a programmable controller as can be used in the electrical circuit embodiment cited above, comprising: -the at least three output pins; -the at least four input pins; -the at least one analog to digital convertor; -the calculation unit; -a memory containing program code executable by the calculation unit and adapted for performing the method embodiments.

It is an advantage of using such a programmable controller, that practically all of the functionality (except for the power output stage and voltage divider, which are optional) can be implemented in a single device.

It is advantageous to use a programmable micro-controller because it offers high flexibility of developing and testing the program. The micro-controller may be a so-called digital micro-controller, or may be a hybrid micro-controller comprising also motor driver circuitry such as a power output stage. It is an advantage that such devices are commercially available with many options (e.g. in terms of memory, flash, processor speed, number of ADC's etc), are qualified for harsh environments (e.g. automotive environment), are mature, have existing development environments, etc. This again simplifies the implementation and testing effort and risks.

A programmable controller according to embodiment of the present invention may further comprise a timer module.

In yet another aspect, the present invention provides a computer program product for executing any of the method embodiments of the present invention, when executed on a controller associated with a sensorless brushless DC (BLDC) motor or a permanent magnet synchronous motor (PMSM) comprising at least three stator windings connected in star configuration and a permanent magnet rotor. Such computer program product can be tangibly embodied in a carrier medium (machine readable data storage) carrying machine-readable code for execution by a programmable processor. The present invention thus relates to a carrier medium carrying a computer program product that, when executed on computing means, provides instructions for executing any of the methods as described above. The term "carrier medium" refers to any medium that participates in providing instructions to a processor for execution. Such a medium may take many forms, including but not limited to, non-volatile media, and transmission media. Non volatile media includes, for example, optical or magnetic disks, such as a storage device which is part of mass storage.

Common forms of computer readable media include, a CD-ROM, a DVD, a flexible disk or floppy disk, a memory key, a tape, a memory chip or cartridge or any other medium from which a computer can read. Various forms of computer readable media may be involved in carrying one or more sequences of one or more instructions to a processor for execution. The computer program product can also be transmitted via a carrier wave in a network, such as a LAN, a WAN or the Internet. Transmission media can take the form of acoustic or light waves, such as those generated during radio wave and infrared data communications. Transmission media include coaxial cables, copper wire and fibre optics, including the wires that comprise a bus within a computer.

Particular and preferred aspects of the invention are set out in the accompanying independent and dependent claims. Features from the dependent claims may be combined with features of the independent claims and with features of other dependent claims as appropriate and not merely as explicitly set out in the claims.

The above and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter.

Brief description of the drawings

FIG. 1 shows a schematic representation of a three-phase, star connected BLDC stator and a two-pole permanent magnet rotor.

FIG. 2 shows the ideal torque waveform exerted on the rotor in case of commutation at multiples of 60 degrees for the motor shown in FIG. 1.

FIG. 3 shows an embodiment of an electrical circuit for controlling a brushless and sensorless BLDC motor shown in FIG. 1, according to aspects of the present invention.

FIG. 4 shows another embodiment of an electrical circuit for controlling a brushless and sensorless BLDC motor shown in FIG. 1, according to aspects of the present invention.

FIG. 5A shows an example of an integrated controller as can be used in the electrical circuit of FIG. 3 and FIG. 4, according to aspects of the present invention.

FIG. 5B shows another example of an electrical circuit according to aspects of the present invention, the circuit comprising a Central ECU, an integrated controller for performing the commutations, and a power drive stage connected to a BLDC motor.

FIG. 6 shows an example of waveforms corresponding to the circuit of FIG. 3.

FIG. 7 shows an example of waveforms corresponding to the circuit of FIG. 4.

FIG. 8 shows an example of a possible behaviour of the BEMF voltages during the first seven commutations after start-up of a BLOC motor, according to a first embodiment of the present invention.

FIG. 9 shows an example of a possible behaviour of the BEMF voltages during the first two commutations after start-up of a BLOC motor, according to a second embodiment of the present invention.

FIG. 10 shows an example of a torque waveform as may be experienced by a motor controlled by a method according to embodiments of the present invention.

FIG. 11 shows an enlarged portion of part of the waveforms shown in FIG. 6.

The drawings are only schematic and are non-limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn on scale for illustrative purposes.

Any reference signs in the claims shall not be construed as limiting the scope.

In the different drawings, the same reference signs refer to the same or analogous elements.

Detailed description of illustrative embodiments

The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims. The drawings described are only schematic and are non-limiting. In the drawings, the size of some of the elements may be exaggerated and not drawn on scale for illustrative purposes. The dimensions and the relative dimensions do not correspond to actual reductions to practice of the invention.

The terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequence, either temporally, spatially, in ranking or in any other manner. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

Moreover, the terms top, under and the like in the description and the claims are used for descriptive purposes and not necessarily for describing relative positions. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other orientations than described or illustrated herein.

It is to be noticed that the term "comprising", used in the claims, should not be interpreted as being restricted to the means listed thereafter; it does not exclude other elements or steps. It is thus to be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof.

Thus, the scope of the expression "a device comprising means A and B" should not be limited to devices consisting only of components A and B. It means that with respect to the present invention, the only relevant components of the device are A and B. Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment, but may. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner, as would be apparent to one of ordinary skill in the art from this

disclosure, in one or more embodiments.

Similarly it should be appreciated that in the description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects.

This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention, and form different embodiments, as would be understood by those in the art. For example, in the following claims, any of the claimed embodiments can be used in any combination.

In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In other instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

FIG. 1 shows a schematic representation of a motor 2 comprising a three-phase, star connected brushless direct current (BLDC) stator and a two-pole permanent magnet rotor 17.

The rotor 17 is indicated as a bar magnet having a north and south pole. The stator comprises three windings connected at one end to a so called "star" point, the other ends (also called terminals) being accessible, e.g. for being driven by a controller 3 (not shown in FIG. 1). The stator windings are represented by inductances Lu, Lv, Lw. In practice the windings L also have an electrical series resistance R, but these are not shown here, and are of no importance for the yeast of the present invention. In this application, the notations U, V, W are used for indicating both the terminals, as well as the voltage applied to those terminals. The voltages between the terminals and the star point, also called "phase voltages" could be denoted by Vun, Vvn and Vwn, N being the voltage at the star point.

Referring back to FIG. 1, the motor 2 can be driven by applying voltage waveforms to some or all of the terminals U, V, W, so that a current can flow through the windings. In brushless sensorless DC (BLDC) motors, however, it is common to apply a voltage only to two of the three terminals, and to leave the third terminal unconnected for measuring a back electromagnetic force (also referred to as: Back EME, or BEMF). In that way, two of the windings are energized, and the third winding is left un-energized (also called "tn-state"). For example, when VDD is applied to terminal V, and GND is applied to terminal W, a current as indicated by arrow lmstl will flow through the windings Lv and Lw, while a back-EMF would be generated in the winding Lu, which can be measured at terminal U. This is well known in the art, and further details can be found in literature.

The shape of BEMF can e.g. be trapezoidal or sinusoidal, depending on how the windings are located on the stator. A trapezoidal BLDC motor can be driven by applying alternately the supply voltage VDD to one of the windings, and the ground voltage GND to another of the windings, while leaving the third winding open (tn-state), during a certain time period (further also referred to as "time-slot"). Instead of applying the supply voltage, a pulse-width modulated PWM-signal can also be applied. An advantage of driving the terminals with a DC voltage and/or a PWM signal, as compared to sinusoidal signals, is that DC signals and PWM signals can be generated more easily in a controller 3, and do not require significant processing power, and/or tables stored in memory, and/or interpolation routines.

It is explicitly mentioned that the present invention is not only valid for block mode or trapezoidal mode commutation, but also works for sine wave controlled motors e.g. PMSM motors. Even though in sine wave controlled motors usually all three motor terminals are permanently driven, it would look like BEMF measurements would in first instance not be possible. However, by applying a very small interruption of the driving, the BEMF can be measured and monitored, and then the principles of the present invention can be applied.

For further information with respect to interruption of driving, reference is made to EP2037567.

Although PWM signals can also be used in the present invention, the invention will be described using DC-signals.

FIG. 1 shows six arrows, indicated as lmstl to lmstb, representing the current flowing through the windings in each of six possible motor states MST1 to MST6, as listed in table 1.

Motor state Rotor position [C] Phase U Phase V Phase W MST1 0.60 tn-state VM GND MST2 60.120 GND VM tn-state MST3 120.180 GND tn-state VM MST4 180.240 tn-state GND VM MST5 240.300 VM GND tn-state MST6 300.360 VM tn-state GND

Table 1

With rotor position is meant the relative rotor position with respect to the stator windings.

When the motor supply voltage VM is applied to terminal V1 and ground GND is applied to terminal W, a current will flow as indicated by the arrow lmstl, corresponding to motor state MST1. This current will generate a magnetic field which will exert a torque on the rotor 17 for moving the rotor such that the magnetic field of the rotor will align with the magnetic field generated by the stator windings. For continuous rotation of the rotor, the voltages applied to the terminals should be dynamically adjusted synchronously with the rotor movement.

This can be achieved by changing the motor states in a predefined sequence: MST1, MST2, MST3,. Changing from one motor state to the other, by changing the voltages applied to the terminals, is called "commutation".

A similar table (not shown), whereby the second column is 1800 phase shifted, would be used for rotating the motor in the opposite direction, whereby the sequence would be: MST3, MSTZ, MST1, etc. As is known in the art, for a stator with three phase windings, oriented at 1200 angular distances, the commutations should ideally occur at angular rotor intervals of 60°, for maximizing the torque exerted upon the rotor 17. FIG. 2 shows the (ideal) torque waveform exerted on the rotor in such a case. The torque exerted on the rotor during a first motor phase MST1 is indicated by a thick line. As can be seen, in order to follow the maximum-torque waveform indicated in FIG. 2, a commutation should take place at multiples of 60°.

The challenge that the present invention tries to solve is to find a suitable but less complex method, and suitable but less complex circuitry for driving such a sensorless BLOC motor, even if the operation conditions may not be optimal at all times. Such a controller may not be suitable for all motor applications (such as e.g. robotic movement, where precise positioning and precise accelerations are of prime importance), but can certainly be used in some other applications where reliability and simplicity are more important than precise predictable behavior, or in applications where the start-up behavior is less important than the nominal behavior. Such an application is e.g. fan control in automotive application, where the motor load is negligible at start-up, and gradually increases as the fan speed increases.

Other suitable applications are e.g. oil-pumps, fuel-pumps, water-pumps, air-pumps, blowers, compressors, etc. FIG. 3 shows an example of an electrical circuit 1 according to embodiments of the present invention, for controlling a brushless and sensorless BLDC motor. (In FIG. 5B a more complete circuit will be shown). The electrical circuit 1 comprises a motor 2, illustrated here schematically as a BLOC motor, but the invention also works for a PMSM motor. The electrical circuit further comprises a controller 3, e.g. a digital controller without an analog buffer stage, or an integrated hybrid controller including an analog buffer stage.

The controller has outputs Al, AZ, A3 for outputting energizing signals for energizing at least two of the motor windings, and inputs B1, B2, B3 for receiving monitor signals for monitoring the un-energized winding.

The electrical circuit 1 may optionally further comprise a buffer stage 4 connected between the controller 3 and the motor 2, for providing power to the phase windings U,V,W, in case the current drive capacity of the controller 3 is insufficient. FIG. 5A shows one example of an external buffer stage 4 comprising an external H-bridge consisting of two FET transistors connected between motor voltage supply VM and ground GND. Buffer circuits 4 are known in the art, and need not be described in more detail here. Other buffer circuitry known to the skilled person can also be used. It is to be noted that the motor voltage supply VM may be the same as, or different from the supply voltage of the motor controller VDD.

Referring back to FIG. 3, the electrical circuit 1 may optionally further comprise a voltage divider 7 for limiting the voltage levels to the inputs B1, B2, B3 of the controller 3. A simple voltage divider consisting of two resistors is shown in FIG. 5A. However, such a voltage-divider 7 can be omitted for applications where the signal levels are sufficiently low, e.g. when the motor 2 is directly powered by the controller 3 using e.g. 5V or 12V signals, without an external buffer stage 4.

Referring back to FIG. 3, the electrical circuit 1 may optionally further comprise circuitry 5 for generating a virtual star voltage VS, which can be used as a reference signal Vref to the controller 3, and can be measured e.g. via a fourth input B4.

FIG. 4 shows an alternative electrical circuit 1 which can be used for controlling the motor 2 according to aspects of the present invention. It is identical to that of FIG. 3, except that it comprises circuitry 6 for generating a reference voltage Vref equal to half of the motor supply voltage VM/2, and in that the three resistors Ri for the virtual star point can be omitted. The voltage VM/2 may also be provided in other ways, e.g. by using a power supply providing such voltage directly, in which case the voltage divider 6 can be omitted.

In an embodiment of the electrical circuit 1, the controller 3 comprises four analog-to-digital convertors (ADC) 9, three of which are operatively connected to the phase windings U, V, W, for measuring a voltage representative of the back-EMF voltage, which is present on one of the input lines, depending on the motor state being applied. The fourth ADC is operatively connected to the resistor circuitry 5 or the resistor circuitry 6 for receiving the reference voltage. Instead of using four ADC's, it may also be possible to use only a single ADC, and to internally route one of the signals from the respective input pins B1, B2, B3 or B4 to that single ADC, because they are not needed at the same time, as will be explained further. Of course, using two ADC's and routing two of the input pins thereto is also possible.

FIG. 5A shows an example of a digital controller 3, or a hybrid controller 3, as can be used in the electrical circuit of FIG. 3 and FIG. 4.

FIG. SB shows a complete system comprising an electrical circuit according to aspects of the present invention. The system of FIG. SB comprises a Central ECU which may e.g. determine the speed at which the motor is to run, and communicates this speed via a PWM-signal or a LIN-signal to an integrated controller 3 for performing the commutations, which outputs control signals for the power drive stage 4 for generating the voltage signals U, V, W which are applied to the BLOC motor 2. In this example the supply voltage VDD of the integrated controller 3 and the motor supply voltage VM applied to the power stage are the same, and are equal to Vbat.

FIG. 6 shows an example of voltage waveforms as can be used/encountered during the operation of the BLOC motor shown in FIG. 1 using the circuit of FIG. 3. In FIG. 6a the three motor phases and the corresponding voltage levels U, V, W for one electrical rotation are depicted. For example, in the first motor state MST1, phase U is tn-state, phase V is connected to the supply voltage VDD and phase W is connected to ground GND. This voltage vector results in a current flow from phase V to phase W. A falling voltage slope is present on phase U, carrying the position information of the rotor 17. After the commutation event, MST2 is applied to the stator windings. The current starts to flow from phase V to phase U and a rising slope is present on motor phase W. This working is very well known in the art. In FIG. 6b the voltage of virtual star point 9 is depicted showing the same tendency as the tn-state motor phase (U in MST1, W in MST2, etc), but with different amplitude. The subtraction of the virtual star point voltage T from the tn-state phase voltage (U in MST1, etc) provides a voltage representative for the BEMF, as can be seen in FIG. 6c. A saw-tooth shape waveform shows the BEMF voltage when the rotor is running synchronous to the stator flux. Thus by monitoring the signals U (as un-energized phase winding) and T (virtual star point as reference voltage) during MST1, the signals W and I during MST2, etc, the back-EMF signal can be calculated. The waveforms of FIG. 6 corresponds to the signals in the circuit of FIG. 3.

FIG. 7 shows an example of voltage waveforms as can be used/encountered during the operation of the BLOC motor shown in FIG. 1 using the circuit of FIG. 4. They are very similar to the signals of FIG. 6, except that half the motor supply voltage VM/2 is used as reference voltage, instead of the virtual star point voltage T in FIG. 6. Also the subtraction of half the motor supply voltage VM/2 from the tn-state phase voltage (U, V or W, depending on the motor state) provides a signal eu, ev, ew which is proportional to the BEMF in the tn-state phase, as shown in FIG. 7c.

While the waveforms of FIG. 6 and 7 are ideal waveforms, the inventor has observed that the motor still works in a satisfactory way, even when the commutations are not performed on the ideal moments (i.e. at multiples of 600), but close to these moments. In particular, the inventor has observed what would happen with the tn-state signal U of FIG. 6 in MST1, if the commutation for MST2 occurs "too late" (i.e. "later than in the ideal case"), which means that V is maintained at VM "too long" and W is maintained at GND "too long".

As can be seen from FIG. 11, what happens is that the U-signal reaches a minimum (indicated by arrow 20) and starts increasing again. Likewise, in MST2 the voltage of the un-energized phase voltage W will gradually increase from GND to VM, but when the voltages of phase MST2 are maintained "too long" (i.e. longer than the ideal moment), the phase voltage W will start decreasing again, Observing the un-energized voltage until it reaches its maximum or minimum is the basis of a first embodiment of a method according to the present invention, as will be described further in relation to FIG. 8. when the commutation is then applied (albeit "too late"), the inventor has observed that the BEMF-signal makes a sudden jump. It is noted that this "jump" is different from any transient behavior caused by current decay due to the commutation event itself, but is due to the fact that the rotor has turned beyond the ideal angle, before the commutation is applied. This observation is the basis of a second embodiment of a method according to the present invention, as will be described further in relation to FIG. 9.

FIRST METHOD, BASED ON MAXIMA AND MINIMA A first method according to embodiments of the present invention will now be described in detail. FIG. 8 shows an example of a possible start-up behaviour of the BLDC motor. It is assumed that the rotor position is known at tO, or is forced to a known position by applying one of the motor states (e.g. MST1) and waiting for a sufficient time until the rotor has aligned itself with the stator field (as is known e.g. from US7737651B2). In this way it can be assured that the motor 2 will subsequently start rotating in the desired direction. Suppose in this example that the rotor 17 is initially located between 60° and 120° electrical. As a consequence (see Table 1 above) the motor state MST2 is applied to the stator windings.

Motor phase U is connected to ground GND, phase V is connected to the supply voltage VDD, and phase W is tn-state, so that the BEMF voltage "ew" induced by a moving permanent magnet rotor can be measured on phase W. The BEMF-signal "ew" is continuously monitored in MST2, and when it reaches a local maximum at tl, a commutation Cl is performed by energizing the windings attributed to the subsequent motor state, according to the sequence of table 1. Thus, at tl, a commutation Cl takes place, and the motor windings U, W are energized according to motor state MST3, and the motor winding V is un-energized. In the motor state MST3, the BEMF-signal is read from phase V, hence the indication "ev". The next commutation will then be performed at the time t2 where the "ev"-signal reaches a local minimum. Thus, at t2, the motor windings V, W are energized according to motor state MST4, and the motor winding U is left un-energized (corresponding to table 1 above). In this motor state MST4, the BEMF-signal is read from phase U, hence the indication "eu", and, according to the algorithm, the next commutation should occur when the BEMF-signal "eu" reaches a local maximum, etc. The motor can thus be driven by repeating this simple algorithm.

It is noted that the detection of a local maximum can be as simple as repeatedly comparing two successive measurements, and when the last value is smaller than the previous value, to decide that a local maximum is reached. Such algorithms are well known in the art, and require no sophisticated goniometrical functions, or filters, or vector-theory etc. As can be seen from FIG. 8, as time increases, the difference between the motor speed also increases until the motor is running under quasi ideal conditions, whereby the BEME-voltage-signals of subsequent motor states approach the ideal saw tooth shape shown in FIG. 6c or FIG. 7c. Admittedly, in practice the commutation will usually take place slightly later than the ideal moment, because the maximum or minimum can only be detected "after" the ideal moment is passed. Nevertheless, it should be clear from the description above that start-up in the correct direction is guaranteed, and that, apart from the start-up behaviour, the commutation events are not far off from the ideal commutation events. Vet, this algorithm is extremely simple to implement, both in terms of hardware and software.

The basic algorithm can be modified in several ways. For example, the algorithm may optionally take a time-out into account, which would force the next commutation in case the motor is longer in a certain motor state than a predefined period. In that case, the algorithm would thus perform a commutation if the motor is in a certain motor state longer than the predetermined time-out period, or if a local maximum or a local minimum is reached, whichever occurs first. An advantage of this modified method could be that it allows the detection of some error-conditions, such as e.g. a stuck rotor.

Another optional modification may be to use a PWM-signal over at least part of the motor state period, instead of the motor supply voltage VM. In this way the motor speed can be controlled without changing the level of the supply voltage VM.

Another optional modification is to omit the detection of the maximum or minimum values during a certain time period after each commutation, which time period may be a fixed constant value, or a fraction depending on the speed of the motor. Alter all, when the motor is running at almost nominal speed, the time period between commutations is almost constant, and thus it makes no sense to monitor for the maximum or minimum during e.g. the first 75% of the period. In this way, the processing power may be reduced.

Alternatively or in combination thereto, the sampling frequency may be adjusted, such that less samples are taken at the beginning of each period (where normally no local maximum or minimum occurs), and more samples are taken near the end of the period (where the local maximum and minimum is expected).

The person skilled in the art can further fine-tune the algorithm according to specific needs. For example, if needed, some hysteresis may be added to the maximum and minimum detector, in the sense that the controller only decides that "the maximum" resp. "the minimum" value is found, when the last sample is less resp. higher than e.g. 2 least significant bits of the ADC from the "maximum" or "minimum" found in this motor period. In this way problems of noise could be solved or at least reduced.

FIG. 10 shows the torque waveform obtained when applying this first method. Thus, instead of the ideal torque waveform of FIG. 2, a slightly lower torque value is obtained just before each commutation (which is performed slightly after the ideal moment, as described above). However, in practice the difference "d" between the ideal torque and the actual torque is typically relatively small, and can e.g. be decreased by increasing the sampling frequency, or decreasing the hysteresis.

It should be mentioned, however, that this simple first algorithm only works well for block-mode commutation (not for sinusoidal mode), because the BEMF-signal should preferably be monitored at a sufficiently high frequency, preferably "continuously", so that commutation can be applied as soon as possible after the local maximum or local minimum is reached.

SECOND METHOD. BASED ON PREDICTED PERIOD A second method according to embodiments of the present invention will now be described in detail, with reference to FIG. 9. Assume again that the rotor 17 is forced to a known position before actual start-up, by applying one of the motor states (e.g. MST3), and waiting for a sufficiently long time until the rotor 17 has aligned itself with the stator field. In this way it can be assured that the motor will subsequently start rotating in the desired direction. Suppose in this example that the rotor is located between 1800 and 240° electrical before tO. As a consequence (see Table 1 above) voltages according to the motor state MST4 are applied to the stator windings at time tO during a predetermined time period Ti, and then a first commutation is performed at t=T1. In practice the use of a predetermined time period Ti for the first motor cycle MST4 gives acceptable results, e.g. in applications such as fan-control, where the inertia of the rotor is known, and the load is negligible at start-up, thus the acceleration is very predictable. For applications where a constant time period Ti would not be acceptable, monitoring the local maximum or minimum as in the previous method could also be applied. However, shortly before the first commutation at tl, e.g. at a first time interval AT1 before tl, the BEMF-signal "eu" is measured, resulting in the value vl(1), and shortly after the first commutation at tl, the BEMF-signal "ew" is measured, resulting in the value v2(i). (the notation vlC) will be used for the measurement before, and v2(.) will be used for the measurement after the commutation. The number of the commutation is indicated between the brackets, thus, as an example, vl(3) is the measurement before the third commutation). Thus although the second period (also called motor cycle) MST5 is already started, its period T2 is not determined yet. That period T2 will then be calculated, based on the time period Ti of the previous motor cycle MST4, and based on the measured values vl(i) and v2(i) of the BEME "just before" and "just after" the first commutation Cl, by using the following formula: /2 = Ti * (1-1* viG) V2W) [i] 2 vi(1) In fact, this is a simplified form of a more general formula described below, applicable only at start-up. When vl(1) is larger than v2(1), it means that the commutation was done "too late", thus the next period T2 needs to be shorter than the previous period Ti. Indeed, if vl(1) is larger than v2(1), the number after the minus-sign is positive, thus 12 will be smaller than Ti by a number proportional to the "jump" and inversely proportional to the "maximum value".

The time t2 at which the motor state MST5 should end, and the second commutation should occur, is then calculated as t2=tl÷T2. When the time t has reached t2, which can e.g. be efficiently implemented by making use of a timer-interrupt, (but other techniques such as "polling", or "delay functions by counting to a certain number" may also be used), the second commutation C2 will be executed, and the voltages of the subsequent motor state MST6 (see table 1) will be applied to the motor windings. However, as before, slightly before the second commutation C2 at time t2 took place, the BEMF-signal "ew" is measured as vl(2), and slightly after the second commutation C2 at t2, the BEMF-signal "ev" is measured as v2(2), and the time period of the third time-slot T3 is calculated using the formula: = /2 * (1_I * (2) 2(2) -2* G)) [2] 2 vit'2) -v2(1) The idea behind this formula is that 13 is calculated in a way which tries to achieve that vl(3) equals v2(3), so that the BEME signal approaches an ideal saw tooth function. The formula can be derived by assuming that the BEMF behaves linearly (which is not entirely true at start-up, but the approximation is better and better as the motor speed reaches its nominal value), and that the "ideal" moment for commutation is calculated as the moment where the two BEMF-lines (before and after the commutation) intersect. This yields a relation between vl(n) and v2(n) before and after the commutation, and v2(n-1) of the previous commutation.

Knowing t2 and T3, the time t3 when the third commutation C3 should be executed can then be calculated as t3=t2+T3. This formula can be generalized, and the time duration T(n+1) of each subsequent time-slot can be calculated using the following formula: 1 vi(n) -v(n) -2 * v(n -1) T(n+I) =T(n)*(i--* --) [3] 2 vJ(n)-v2(n-1) where T(n) is the time period of the n-th motor cycle, T(n-1) is the time period of the preceding motor cycle, vl(n) and v2(n) are the BEMI-voltages measured slightly before and slightly after the n-th commutation (which is the most recent commutation), and v2(n-1) is the BEME-voltage measured slightly after the (n-1)th commutation event (preceding the most recent commutation).

As can be seen in FIG. 9, due to the non-ideal moment of commutation, a discontinuity is seen between the back-EME signals vlC) and v2C) slightly before and after each commutation event. However, as time increases, the difference between these values gradually decreases, which is exactly the purpose of the formulas above, and the rotor 17 gets more and more synchronous with the stator field, until the motor is running under (quasi) ideal conditions, whereby the BEMF-voltage-signal of subsequent motor states approaches more and more the ideal saw tooth shape shown in FIG. 6c or FIG. 7c.

It should be clear from the description of the second method that start-up in the correct direction is guaranteed, and that, apart from the start-up behaviour, the commutation moments are not far "off" from the ideal commutation moments. Also the second method is extremely simple to implement, both in terms of hardware and software.

In fact, this algorithm can e.g. be efficiently implemented as an interrupt routine, which can be executed on a controller 3 having a timer and timer interrupt facilities. The CPU 10 of this controller 3 should also be able to perform a multiplication and division operation, but as the formula [1], [2] or [3] only needs to be evaluated once per time-slot, a DSP with fast ADC's is not required and a relatively "slow" CPU (e.g. running at 5 MIPs or lower) may suffice. Even more important is that the algorithm is very robust for dynamic load changes, which is probably due to the fact that the BEMF is measured in each commutation interval, whereby the period T for each interval is calculated so as to regulate the error function vl(n)-v2(n) towards zero.

A torque waveform similar to that of FIG. 10 would be obtained when applying this second method, except that the commutation events of the second embodiment may fall "before" or "after" the ideal commutation events, but since the torque function is "symmetric" around each working point, the net effect is similar, and as the moments between the "ideal" commutation events and the actual commutation events are not far "off", the average torque is only slightly less than the ideal torque.

While FIG. 8 and FIG. 9 show relatively "undisturbed" BEMF-curves, it should be mentioned that, shortly after commutation, the current flowing through one of the two phase windings will gradually decay, and the current will gradually increase in the newly energized phase winding. As is known in the art, commutation events may cause voltage spikes due to the decay of the so called "fly-back" current, the energy of which is typically dissipated in bulk diodes or power transistors, while the voltage is clamped (e.g. to about GND -lv or VM + lv in the bulk diodes). During the decay period, the waveform present on the un-energized phase winding is therefore not a good indication for the BEMF, and thus not a good indication for the position of the rotor. However, as the clamping voltage is known, invalid BEMF-values (i.e. not representing true "BEMF-values") can easily be discarded in software. This is not shown in FIG. 8 and FIG. 9 for not obscuring the explanation of the basic algorithm.

Summarizing, the present invention provides two methods, which can both be implemented on the two electronic circuits shown in FIG. 3 and FIG 4. It is noted that the algorithms are robust even against changes of the motor supply voltage VM, as may occur in a car environment, by working with a reference voltage that changes along with the supply voltage (and thus need not be constant).

On the other hand, in applications where the motor supply voltage VM is guaranteed to be constant, an advantage of the circuit of FIG. 4 over that of FIG. 3 is that half the supply voltage VM/2 will also be constant, and thus (in theory) only needs to be measured once, e.g. at startup, while the virtual star voltage is not constant, and thus also needs to be sampled along with the BEMF-signal. In such applications the number of ADC conversions can be reduced, and thus also the power consumed by the controller.

The data of the motor sequence shown in Table 1 may e.g. be hardcoded, or may be stored in a memory 12. In applications where the direction of rotation is fixed, (e.g. a fan-application), some further optimizations can be implemented. It is also noted that the supply voltage VDD and/or the motor supply voltage VM, may be generated from one or more batteries, and in case of an even number of such batteries, the half motor supply voltage VM/2 may e.g. be taken halfway between the first and the last battery.

REFERENCES: U,V,W motor phases L inductance MST motor state N/S north/south pole of the magnet Imstl current in motor state 1 Ri resistor (of virtual star point circuit) VS virtual star point Ti first period Cl first commutation tl time of first commutation d difference Vref reference voltage A output B input i electrical circuit 2 DC motor 3 controller 4 power buffer 5 circuit for generating virtual start point 6 circuit for generation half supply voltage 7 voltage divider 8 pulse-width-modulator module 9 analog to digital convertor iO processing unit ii timer i2 memory i4 half supply voltage vl(n) measurement taken before commutation number n v2(n) measurement taken after commutation number n

Claims

Claims 1. Method for driving a sensorless brushless DC (BLDC) motor or a permanent magnet synchronous (PMSM) motor in a predetermined direction, the motor (2) comprising at least three stator windings (U, V, W) connected in star configuration, and a permanent magnet rotor (17), the method comprising: a) energizing two of the windings (U, V) and leaving a third winding (W) un-energized for rotating the rotor (17) in the predetermined direction, based on a first motor state (MST2) of an energizing sequence corresponding to a known rotor position; b) monitoring a voltage (ew) representative for the back-EMF generated in the un-energized winding (W) and determining a time instance (ti, t2) when that voltage (ew) reaches a local extreme; c) applying at the determined time instance (tl, t2) a commutation (Cl) so as to energize at least two of the windings (U, W) while leaving a third winding (V) un-energized according to a subsequent motor state (MST3) of the energizing sequence; d) repeating steps b) and c).
2. The method according to claim 1, wherein the monitoring of the voltage (ew) in step b) is only performed during a fraction of the time period between two successive commutations.
3. The method according to any of claims 1 or 2, wherein the energizing of at least one of the windings is operated at least part of the time using a PWM-signal.
4. The method according to any of the claims 1 to 3, wherein the energizing of the windings is operated at least part of the time with DC voltages (VDD, GND).
5. Method for driving a sensorless brushless DC (BLDC) motor or a permanent magnet synchronous (PMSM) motor in a predetermined direction, the motor (2) comprising at least three stator windings (U, V, W) connected in star configuration, and a permanent magnet rotor (17), the method comprising: a) determining a time period (Ti) as the current time period, and energizing during the current time period (Ti) two of the windings (V, W) and leaving a third winding (U) un-energized for rotating the rotor (17) in a predetermined direction, based on a first motor state (MST4) of an energizing sequence, corresponding to a known rotor position; b) measuring a first voltage (vi) representative for the back-EMF generated in the un-energized winding (W) at a first time interval (ATi) before expiry of the current time period (Ti); c) applying a commutation (Ci) at expiry of the current time period (Ti) so as to energize at least two of the windings (U, V) while leaving a third winding (W) un-energized according to a subsequent motor state (MST5) of the energizing sequence; d) measuring a second voltage (v2) at a second time interval (AT2) after the commutation (Ci), and calculating a subsequent time period (T2) based on at least the current time period (Ti) and the measured first and second voltage (vi, v2); e) repeating steps b) and c) after replacing the current time period (Ti) by the subsequent time period (T2).
6. The method according to claim 5, wherein the first time period (Ti) in step a) is a predetermined value.
7. The method according to claim 5, wherein the first time period (Ti) in step a) is determined by monitoring a voltage (eu) representative for the back-EME generated in the un-energized winding (U), and by determining the first time period Ti as the time instance when that voltage (eu) reaches a local extreme.
8. The method according to any of claims 5 to 7, wherein each subsequent time period (T2) is calculated based on the preceding time period (Ti) by using a formula that is directed at reducing the difference between the first and the second voltage (vi, v2) before and after the the subsequent commutation (C2).
9. The method according to any of claims 5 to 8, wherein the second time period (T2) immediately following the first time period (Ti) immediately after motor start-up is calculated according to the formula: Tz = Ti (1-' viO) ilL vi(I) wherein T2 is the second time period, Ti is the first time period, vi is the first voltage measured before the first commutation (Ci) and v2 is the second voltage measured after the first commutation (Ci), and A is a predetermined value in the range of 0,25 to 0,75.
10. The method according to any of claims 5 to 9, wherein the subsequent time period is calculated according to the formula: T(n+ 1) = 1(n) * (1-1* i1(n)v2() -2*V2(n -1)) A vi(n.) -vz(n -1) wherein T(n) is the subsequent time period, T(n-1) is the current time period, vl(n) and v2(n) is the first resp. second voltage measured before and after the most recent commutation, and v2(n-1) is the second voltage measured after the commutation preceding the most recent commutation, and A is a predetermined value in the range of 0,25 to 0,75.
11. The method according to any of the claims 5-10, wherein the energizing of at least one of the windings is operated at least part of the time using a PWM-signal.
12. The method according to any of the claims 5-11, wherein the energizing of the windings is operated at least part of the time using DC voltages (VDD, GND).
13. An electrical circuit (1) as can be used for performing the method according to any of the claims 1-12, the electrical circuit (1) comprising: -a BLDC motor or a PMSM motor, the motor (2) comprising at least three stator windings (U, V, W) connected in star configuration and a permanent magnet rotor (17); -a controller (3) connected to the motor (2), the controller comprising: -at least three output pins (Al, A2, A3) connected to the windings (U, V. W) of the motor (2) for energizing at least two (U, V) of the windings and for leaving a third winding (W) un-energized; -at least four input pins (Bi, B2, B3, B4) connected to the windings (U, V, W) of the motor (2) for measuring the voltage (ew) representative of the back-EMF generated in the un-energized winding, and connected to a reference signal (Vref); -at least one analog to digital convertor (ADC) for digitizing the measured voltage (ew) and for digitizing the reference signal (Vref); -a calculation unit (10); wherein the controller (3) is provided with an algorithm for performing the method according to any of the claims 1-11.
14. A programmable controller (3) as can be used in the electrical circuit of claim 13, comprising: -the at least three output pins (Al, A2, A3); -the at least four input pins (81, 82, 83, 54); -the at least one analog to digital convertor (ADd); -the calculation unit (10); -a memory (12) containing program code executable by the calculation unit (10) and adapted for performing the method according to any of the claims 1-12.
15. The programmable controller (3) according to claim 14, further comprising a timer module (11).
16. A computer program product for executing any of the methods as claimed in any of claims 1 to 12, when executed on a controller (3) associated with a sensorless brushless DC (BLDC) motor or a permanent magnet synchronous (PMSM) motor comprising at least three stator windings (U, V1 W) connected in star configuration and a permanent magnet rotor (17).
17. A machine readable data storage storing the computer program product of claim 16.