Title of Invention	METHOD FOR CONTROLLING A BRUSHLESS DC MOTOR
Abstract	Method for controlling a brushless DC motor (46) comprising a stator (44) and rotor (54), which has a highly reduced reduction gear of an actuating drive for a flap or valve in order to regulate a gas or liquid volume flow, in particular in the field of heating, ventilation and climastisation (HLK), fire and room protection, the electrical current flow in the stator coils (A, B, C) being communicated under programme control without integral position sensors, characterized in that the electronics count the revolutions of the rotor (54) of the DC motor (46), and supplied to a counter with a logic rule, and in that the logic rule for start and stop cause motor speeds to reduce below a threshold value to be passed quickly and without the occurrence of counting errors.

Title of Invention

METHOD FOR CONTROLLING A BRUSHLESS DC MOTOR

Abstract

Method for controlling a brushless DC motor (46) comprising a stator (44) and rotor (54), which has a highly reduced reduction gear of an actuating drive for a flap or valve in order to regulate a gas or liquid volume flow, in particular in the field of heating, ventilation and climastisation (HLK), fire and room protection, the electrical current flow in the stator coils (A, B, C) being communicated under programme control without integral position sensors, characterized in that the electronics count the revolutions of the rotor (54) of the DC motor (46), and supplied to a counter with a logic rule, and in that the logic rule for start and stop cause motor speeds to reduce below a threshold value to be passed quickly and without the occurrence of counting errors.

Full Text	The invention relates to a method for controlling a brushless DC motor comprising a stator and a rotor, which has a highly reduced reduction gear of an actuating drive for a flap or valve in order to regulate a gas or liquid volume flow, in particular in the field of heating, ventilation and climatisation (HLK), fire and room protection. Furthermore the invention concerns a brushless DC motor without position sensors for performance of the method. Electric actuating drives for motorisation of actuators in heating, ventilation and climatisation systems (HLK systems) have been produced for more than 30 years. HLK actuators guarantee an economic volume flow regulation of gases or liquids, in particular air and water. As a compact unit they comprise not only the drive but also pressure sensors and regulators, all combined in one device. Ventilation systems are increasingly used in buildings, in particular residential, office, commercial and industrial buildings, normally combined with fire and smoke protection devices. In ventilation systems volume flow regulation with swivelable air flaps plays an essential role. The volume flow is measured with a suitable measuring instrument, for example with the NMV-D2M by Belima Automation AG, CH-8340 Hinwil, which is formed as a compact unit of drive, pressure sensor and regulator, and which transmits measurement values to the electronics. To move a flap in a ventilation system or a ball valve in a water pipe system, relatively weak motors must activate large area or large volume control elements. Precise and stable adjustment is possible only with a very high reduction. To swivel a flap or turn a ball valve through an acute or right angle requires numerous rotations of the shaft of the electric motor. Between the motor and valve can be arranged mechanical means for transmission of the motion, in particular spring drives. DC motors originally produced preferably have brushes made of graphite which mechanically regulate the commutation of the individual electrical phases of the rotor, the stator in this case being fitted with permanent magnets. Electronically commutating DC motors of the known type, also called EC motors, are particularly suitable for applications which utilise the benefits of contactless and electronically controlled commutation. With commutation controlled in this way the electronic current feed of the stator coils shifts in steps from pole to pole, "pulling" the rotor with it. Brushless DC motors are characterised by low production costs and a long life in conjunction with preferential regulation properties for speed and precision. A torque generated by such an electromagnetic rotating field is almost constant irrespective of the position of the rotor, and dependent in a linear manner on the three-phase current. Each of the three phases of a brushless DC motor is controlled by one or two transistors, depending on whether the coils are to be powered in one or two directions. With just three control pulses per electrical rotation, the windings are utilised relatively poorly. Controlling the windings with six transistors guarantees a higher utilisation of the windings and a more even torque output. The coils can be connected in a star or triangle, in both cases the torque is greater if each of the three phases is controlled by two transistors. In order to determine the suitable moment to switch the power transistors during commutation, the present position of the rotor in relation to the stator must be known. Particularly suitable here are silicon-free resolvers and Hall effect sensors, both of which give the absolute position of the rotor irrespective of speed. For reasons of cost and miniaturisation, the use of sensors is however always more problematical as they must be replaced frequently. A regular torque is generated for example if the coils are controlled with sinusoidal signals offset by 120°. At lower power rectangular signals can also be used for control because of their steeper zero crossing. As drivers so-called H-bridges are used where each pole of a coil is connected by way of a transistor with a supply voltage and earth. It must however be guaranteed that in the case of two transistors, these are not both conductive at the same time. The H-bridges of the power electronics are controlled as a function of signals from position sensors. They serve for direction formation and are controlled as stop, start and PWM signals using an encoder (PWM = pulse width modulation). The direction of rotation of the rotor is determined by the switching algorithm, the rotation speed of the motor by way of the coil voltage of the coils connected. This can also be triggered by a PWM. The commutation frequency depends on the specified rotation speed of the motor and the number of pole pairs of the rotor. In practice two or three stator coils are always powered simultaneously, the difference of two or three coils must be taken into account. The induced voltages of three coil combinations are phase-offset by 120°. By suitable commutation now, depending on direction of rotation, the maximum positive or negative torque can be utilised. A delayed or premature commutation has a direct effect on the torque of the motor to be provided. The optimum switching point for commutation of two and three powered coils differs, it is controlled by commutation algorithms. The rotor position is normally determined on the motor circuit board by three position sensors, for example Hall effect sensors. Hall effect sensors, when supplied with a constant current, emit a differential voltage as a function of the magnetic field flowing through them, known as the Hall effect. DE 19745312 A1 discloses a control circuit for a brushless electronically commutated DC motor. Arranged in series with each winding strand of the motor is an electronic switch element which is controlled by a position sensor. Each winding strand is connected by way of a decoupling diode with a common switch point. Arranged parallel to the winding strands and associated decoupling diodes is a protective transistor circuit which eliminates induction-induced voltage peaks. Although a good efficiency is achieved with simple low-cost components, the disadvantage remains that position sensors need to be replaced. The inventors have faced the task of creating a method for controlling a brushless DC motor and a DC motor of the type cited initially that eliminates the cost and quality problems related to position sensors. In relation to the method, the object is achieved according to the invention in that the electrical current flow in the stator coils is commutated under programme control without integral position sensors. Special and refined embodiments of the method are the object of the dependent claims. The algorithms of all essential procedure processes are entered and stored retrievably in a CPU, suitably in at least one microprocessor, for example a start algorithm, a brake algorithm, a brake algorithm for reverse kick at the stop i.e. a controlled return and braking after a hard stop, wherein the contact pressure is substantially maintained, and finally a relax algorithm with which the reduction gear is fully relaxed. the retrievably stored algorithms are continuously adapted to practice and improved where necessary i.e. the system is self-learning. The basic principle according to the invention is to check the polarities of the zero crossings of the three motor phase voltages. Each polarity change on a zero crossing of a phase voltage can be allocated to a particular rotor position. As soon as the corresponding rotor position is reached, the next commutation, i.e. the next current flow to the stator coils, can be modified such that the generated leading magnetic field gives the rotor a torque. Thanks to this system, position sensors can be omitted. The zero crossings of the phase voltages can however only be clearly detected from a particular minimum rotation speed of the rotor, rotation speeds that are too low give undefined looping sections of sinusoidal or sine-like curves with the zero line. In a special embodiment of the invention, an electronics unit counts the revolutions per minute (rpm) of the rotor of the DC motor. Preferably each motor commutation is detected and counted, wherein special algorithms for start and stop ensure that motor rotation speeds below around 200 rpm are passed through so quickly that no counting errors occur. Provided that the counters once at the start have a synchronisation or reference point for allocation to the flap angle, the position regulator always knows precisely the actual value i.e. the precise, position of the flap concerned. The counting electronics in this way constitute an almost no-cost replacement for a relatively costly potentiometer that is subject to wear in conventional arrangements. With reference to the DC motor the object is achieved according to the invention in that it comprises a statpr with three coils that extend over 3n stator poles and an annular permanent magnet rotor with 2m permanent magnet segments of alternating polarity (N, S), a cup-shaped external rotor and a coaxial drive shaft, where n and m are whole multiplication factors 1, 2, 3, 4.... independent of each other. Coils and rotor segments can be used together following the rule of combination, in practice however the combinations n=1 and m = 1, and n = 3 and m = 6 have proved particularly advantageous. An actuating drive activated by a DC motor according to the invention by way of a gear system can be supplemented with a spring, suitably a coil spring, which can be extended or pre-tensioned on motorised operation and hence potential energy stored. The spring is extended preferably but not necessarily in the lift direction. An electronic power fail detection detects failure of the operating voltage, first disengages all bridge switches of the end stage, and thus triggers a return of the pretensioned spring. The DC motor thus becomes free running, the potential energy stored in the spring acts as a torque against the lift direction and sets the free running motor in motion. Depending on the load at the gear output, the pre- tensioned spring accelerates to a correspondingly high speed until the load, friction and spring moment are in equilibrium. When the said load at the gear output supports the spring moment or only loads this lightly, the rotation speed in the spring return can be increased so that the entire mechanical system is endangered. Reaching the end stop is particularly critical. In order to limit the acting kinetic energy and hence the load to a safe level, the rotation speed must be limited. For this in the known manner Zener diodes can be used which limit the motor voltage generated by the generator and disperse any additional energy as heat. Preferably however use is made of the property of the DC motors that on a short circuit of at least part of their windings, a braking moment occurs which is approximately proportional to the rotation speed. This can be achieved easily on brushless DC motors according to the invention in that two of the three windings are short-circuited. The extent of braking can be set by way of a pulse width modulated control signal (PWM signal). This is done as a function of the actual rotation speed which can be determined at any time by way of the measured zero crossings of the third motor winding that is not short-circuited. Below a programmable limit speed, the PWM signal is set to zero and after exceeding the limit speed, rises continuously to a maximum of 100%. This allows a simple regulation for the speed limitation which can thus be held constant over a wide load range. This prevents, on power failure, the pre-tensioned spring closing the actuation element concerned, for example a valve or flap, too quickly. Although the method according to the invention and the brushless DC motor are designed in particular for the HLK field, this is not a restriction. Due to dimension- induced adaptations, gas flows, volume flows and/or actuation angles can also be regulated in automobiles in a similar fashion with professional knowledge and ability. The invention is now explained in more detail below with reference to the embodiment examples shown in the drawing which are also the object of dependent claims. The drawings indicate schematically: - Fig, 1 a block circuit diagram of the entire motor control system for a DC motor, - Fig. 2 a wiring diagram for the stator of a DC motor, - Fig. 3 a commutation cycle of a DC motor with six-rotor positions, - Fig. 4 the proportion of individual stator coils in the motor torque according to fig. 3, - Fig. 5 the connection of the phase voltages with the expected polarity of the zero crossing, - Fig. 6 braking by blocking, - Fig. 7 a brake-stop-start cycle, - Fig. 8 the behaviour of a DC motor at the stop, - Fig. 9 the relaxing of a DC motor at the stop, - Fig. 10 the behaviour of a DC motor when starting under difficult conditions (boost concept), - Fig. 11 a variant of a DC motor, - Fig. 12 an actuating drive with a DC motor, and - Fig. 13 a spring drive according to fig. 12. Fig. 1 shows a block circuit diagram of a motor control 10 for a DC motor (46 in fig. 2). As is normal for a very high reduction, this motor has low power and small dimensions. For details of a reduction gear, reference is made for example to PCT/CH05/000126 by the applicant. Three motor generator currents are conducted into the motor control 10 to control its voltage DRA, DRB and DRC by way of BEMF comparators 12 (BEMF = back electromagnetic force). The BEMF comparators 12 in the known manner check the polarity of the zero crossings of the phase voltages DRA, DRB and DRC. Each change of polarity, in other words each zero crossing, can be allocated to a particular rotor position of the DC motor 46. As soon as this polarity is known, the next commutation i.e. the new current supply to the stator coils (A, B, C in fig. 2) is determined. A subsequent ZERO detector 14 detects the valid zero crossings of the phase voltages DRA, DRB and DRC or their BEMF signals 13, which are indicated by the needle-like ZERO pulses 18. The three phase voltages DRA, DRB and DRC can only be analysed when the ZERO detector 14 receives the corresponding signal 16, which is indicated with a reverse arrow. A period regulator 20 receives the ZERO pulses 18 and from these calculates the next commutation time. If the next commutation is actually to take place, a NEXT pulse 22 is passed to a cycle counter 24. Furthermore the period regulator 20 outputs a PER signal 26 to a speed regulator 28. PER (PER = time period) means the distance between two ZERO pulses, and is thus inversely proportional to the rotation speed. A speed regulator 28 controls and regulates the number of revolutions of the rotor per time unit, usually per minute, which is indicated as rpm. The cycle counter 24 counts the whole digit multiples 0, 1,2, 3... and at a motor control 30 outputs a counter value 32 indicated with a CYNO (cycle number). Each of the six counter values 32 in the present case is firmly allocated to one of the six possible H-bridge states to be discussed below. As stated a NEXT pulse 22 controls the cycle counter 24 which thus controls the commutation. As also stated the counter value 32 CYNO gives a feedback in the form of a signal 16 to the ZERO detector 14. The speed regulator 28 controlled by the PER signal 26 takes place in hardware terms by way of a PI regulator (PI = proportional / integral) of the known type. The nominal value is taken from an OTP cell (OTP = one time programmable). Optionally by way of microprocessors (MP) arbitrary nominal values can be written to a volatile register. The PER signal 26 of the period regulator 20 is used as the actual value. This as stated is inversely proportional to the rotation speed and is calculated out in the speed regulator 28. By way of a PWM clock generator 34, a PWM signal 36 is output to the motor control 30, the duration of a pulse in the present case being 52 usec. This pulse width modulation PWM corresponds to the voltage applied to the DC motor 46. The PWM frequency of the clock generator 34 in the present case is 19.2 kHz. The motor control 30 is generated by the PWM modulation 36 and the commutation. The output phase voltages DRA, DRB and DRC are the direct controls for the motor driver in the form of power H-bridges. An analog interface 38 allows control with analog signals. AZ applications (AZ = on/off) are also possible, and with an external potential, constant SR applications (SR = constant regulator) can also be achieved. A digital block 40 from a status control (state machine) controls the internal processes of the motor control 10. For starting and braking of the DC motor 46, the digital block 40 generates an ROT signal 42 which it passes to the speed regulator 28. In a wiring diagram according to fig. 2 for the stator 44 of a DC motor 46, the poles of a DC power source 48 are indicated with + and -. The stator 44 has three stator coils A, B and C arranged in a star pattern and connected together at a star point S that also forms the zero line. A switch group 50 which is activatable under program control generates a commutation cycle comprising six commutations which is shown in detail in fig. 3 below. On a power failure two of the three bridge switches a, b and c are short-circuited to earth, causing a braking moment. In the present case these are bridge switches a and b, whereby a short circuit is formed over the stator coils B and C. Stator coil A is switched off but not short-circuited. Suitably the bridge switches a, b are closed only initially for brief periods which then become ever longer. Generating a braking moment prevents, on a power failure, an actuator returning without control and causing damage. The paired closure of any two bridge circuits a, b, c takes place by way of transistors which receive a pulse width modulated control signal. Fig. 3 shows six positions of a permanent magnet rotor 54 of a DC motor 46. The rotor 54 in the present case has four magnetic segments of equal size, alternately allocated to a north pole N and a south pole S. The six rotor positions constitute a commutation cycle wherein the stator coils A, B and C are powered in pairs with + and -. The power application is programme-controlled and by way of stator 44 commutates a rotating magnetic field which "pulls" the permanent magnets of the rotor 54 with a leading rotary field. The armatures 52 of the three stator coils A, B C extend in rotor form over part of the inner casing of rotor 54. Angle φ designates the mechanical position of the rotor 54 in relation to the zero point. From left to right φ is 15°, 45°, 75°, 105°, 135° and 165°. Fig. 4 shows the proportion of the individual stator coils A, B and C in the torque, in other words the connection between the rotor position and the torques acting on the rotor 54. The mechanical angle φ (fig. 3) of rotor 54 is indicated with vertical lines. These lines also correspond to the electrical position 2φ on the abscissa, 30°, 90°, 150°, 210°, 270° and 330° of the respective mechanical angle φ. The diagram shows a four-pole DC motor 46 (Figs. 2, 3) and not a twelve-pole motor which is normal in practice because the connections are easier to show with as simple a motor as possible. A twelve-pole motor turns three times more slowly than a four-pole motor, otherwise there are no substantial differences. The torque designation M shows the individual proportions of the stator coils A, B, C in the total torque MT. Each of these stator coils acts with a sinusoidal torque on the rotor. As each stator coil is switched off for 60°, for 30° on both sides of the zero, this torque proportion is absent there. This is reflected in the undulating form of the total torque MT. On zero crossing 56 of the sinusoidal torque MB where the power is supplied only through stator coils A and C, a stable position is achieved. The same applies to the torque of stator coils A and C. The zero crossings of the torques MA, MB and Mc are designated below as 56. Below the torque curves M are shown further parameters in the respective positions mechanically φ or electrically 2φ: - BEMF. Curves BEMF-A, BEMF-B and BEMF-C of the individual stator coils A, B and C run sinusoidally. In the ranges ± 30° from the zero crossing, i.e. the intersection of the sine curve with the zero line S in the range of the high impedance commutation gaps L (fig. 5), the curves are drawn in solid lines. - ZERO. At φ = 15°, 45°, 75°, 105°, 135° and 165°, a ZERO pulse 18 is triggered i.e. a valid zero crossing of the BEMF comparator 12 (fig. 1) is registered. The angle between the zero crossings is designated PER. - NEXT. Also indicated with needle pulses are the NEXT pulses 22 which trigger the times of the commutation change. These are offset temporally to the ZERO pulses 18 by an angle DEL = PER/2. In the present case the duration of the needle pulse is ZERO and NEXT is 52 microseconds (usec). - CYNO. Finally CYNO shows the counter states of the cycle counter 24 (fig. 1) from 0 to 5, in other words showing which commutation is occurring at the respective q> on the DC motor. Fig. 5 shows the correlation of the phase voltages DRA, DRB and DRC with the high impedance commutation gaps L of electrically 2φ = 60° in the present case. In these high impedance commutation gaps L, at low rotation speed at their end and at high rotation speed at the end of the power phase, the BEMF comparator 12 (fig. 1) analyses the phase voltage. The BEMF comparator thresholds 58, depending on polarity of the expected zero crossing, are located just - by Vs - above or below the zero line S. This increases the fault distance and supplies more reliable ZERO pulses 18, which brings improvements mainly at low speeds. At the start of a high impedance commutation gap L, a return pulse 64 is also generated. A zero crossing detected correctly by the BEMF comparator 12 generates a positive SCMP flank. The SCMP (= synchronised comparator) is reset on commutation of the H-bridge. The said flanks generate the ZERO pulses 18 that are 52 microseconds wide in the present case. The distance between the ZERO pulses 18 is the "ten bit word" PER. This is inversely proportional to the rotation speed. PER/2 gives the DEL (= delay). Commutation occurs between two ZERO pulses 18. CYNO is the value of the 0, 1, 2, 3, 4, 5 cycle counter and indicates which commutation is present at the DC motor. A commutation cycle f corresponds to 1/commutation frequency. Fig. 6 shows an abrupt braking, also known as a blocking or lever braking. The positions of the rotor 54 correspond to the left half according to fig. 3. At a mechanical angle φ of 30° a strong current pulse 65 generates an amplified resulting torque -B+A which is indicated by a falling dotted line 66. The rotor is stopped with immediate effect. The holding point is stable in the range of 2cp ± 180° electrically i.e. φ ± 90° mechanically. Within this range the rotor returns to the holding point provided. This braking method is highly reliable. The total torque TM diminishes drastically. The counter value CYNO first rises from x+0 to x+1 but then after the current pulse falls again to x+0. In analogy with the braking of a vehicle with spoked wheels by insertion of a rod, the brake process according to fig. 6 is called the "lever brake". According to fig. 6, at the commutation time, instead of the next phase the last phase is activated again, because at this time this has precisely the torque zero crossing. The braking according to fig. 6 with PWM = 0 is a simple stop method for a DC motor but has the disadvantage that at low speeds the braking moment also diminishes. The strong additional current pulse 65 must be applied at the correct moment, which is not easy because the motor position is difficult to predict. Preferably therefore the speed is regulated down to 350 to 400 rpm, whereupon the current pulse 65 takes place which blocks the rotor immediately. The correct moment for the current pulse can then no longer be adapted. Another method for braking a DC motor is indicated in fig. 7 with a cycle of brake- stop-start. The normal operating sequence is characterised by arrow 68. On initiation of braking, the status control (state machine) in digital block 40, by means of a falling flank 43 of the ROT signal 42, tells the motor control 30 by way of PWM clock generator 34 (fig. 1) that the DC motor should be stopped, which takes place with a PWM signal 36. This sets the PWM to 0 i.e. the H-bridges now short-circuit the stator coils, which brakes the DC motor. However commutation continues from the BEMF analysis, the rotor positions are therefore still known as indicated by arrows 70, 72. As soon as the speed control establishes from the PER signal 26 that the rotation speed has fallen below a value of 400 rpm, the stop process initiated by arrows 74, 76 is initiated. The motor control 30 switches to the last instead of to the next commutation state and the corresponding stator coils are powered with a PWM = 80 for 100 ms. This stops the rotor immediately. The counter value CYNO is decremented by 1. After the fast stop in the sense of a lever inserted in the spokes of a wheel, during a holding phase the PWM is reduced to 15 or 0. The restart of the rotor is triggered by the rising flank of the ROT signal 42, a NEXT pulse 22 is generated by way of arrow 79 and the stator coils of the powerless or inhibited PWM circuit are increased to 127. After the start, commutation takes place as normal using the BEMF analysis. Fig. 8 shows the behaviour of a DC motor in practical operation when for example an air flap hits a stop. The abscissa shows the time t and the ordinate from left to right the rotation speed v, the motor current AM and the torque 0. The rotation speed curve v first runs parallel to the abscissa and the rotation speed is constant. To initiate the braking process (arrow 81), the motor current A is increased until the current limit AM is reached (arrow 83), with a slight temporal delay the torque θ also rises. The braking process (arrow 81) is initiated when the flap hits the stop. As in fig. 7 the PWM is set to 0, the H-bridges short circuit the stator coils. Even after the flap hits the stop, the DC motor runs on, it stalls until reaching the stop 80 where no further zero crossings can be measured, and the position of the rotor is lost. After stoppage of the DC motor at peak 82, thanks to the elasticity of the highly reduced, tensioned gear, the motor turns backward until the flap remains on the stop with the residual contact pressure. The position of the rotor can be detected again with special measures when the reduction gear in point 84 stops in the range of the controlled braking 86 ("kick brake"). If the reduction gear turned further back beyond this self-inhibition, the contact pressure of the flap would be lost. After reaching the peak 82 the DC motor is accelerated briefly in the opposite direction so that BEMF signals 13 (fig. 1) are output and a defined holding current can be set. The logic again knows the rotor position. Now the rotor can be braked under control with PWM = 0 as soon as the rotation speed for example has fallen below 400 or 500 rpm. The brief turning back of the DC motor reduces the contact pressure of the air flap only slightly since for this only a few motor revolutions are required. The motor current AM rises in the area of the controlled braking 86 ("kick brake") which is indicated with boost motor current increase 88 and then falls to a lower level AM2 which in the present case is around 20 to 30% of the nominal current. The reduced motor current AM2 is operated during the air flap holding phase with a low PWM. The torque θ rises until the controlled braking 86 ("kick brake") begins, reaches a maximum peak 89 and falls during the stop braking 86 to the constant torque in the holding position, the holding moment 9H. The dynamic part of the stop force is indicated with d. Fig. 9 shows the rotation speed v and the torque θ over time t. In contrast to fig. 8, no holding moment ΘH is maintained here, the reduction gear is fully relaxed after an unbraked stop of the air flap driven by the DC motor. At the stop 80 of the air flap, the DC motor, thanks to the elasticity of the reduction gear, runs on until peak 82 is reached. Thanks to the elasticity the motor turns back and reaches stoppage 90. Immediately after the flap has stopped (area 80) the DC motor is short-circuited for a few seconds tR. This has a mechanical damping effect during relaxation (turning back of the reduction gear). In other words at low speeds the reduction gear turns uninhibited, while at high speeds a short-circuit current flows which brakes the rotor. When the rotation speed v falls below 200 rpm in the present case, no further BEMF signals are generated and the commutation positions are lost. After the stoppage position 90 the DC motor is accelerated briefly in the opposite direction and then braked under control. Depending on setting, now in the motor a torque is impressed or the motor current is switched powerless. The motor is now with a known commutation position in the holding position as indicated with arrow 90. This latter stage is called the reverse pulse 96 (reverse kick). On engagement of the "boost" function according to fig. 10, the DC motor 46 (fig. 2) receives more torque on starting. This allows it to start even under difficult conditions. As long as the ASIC (application specific integrated circuit) is in STARTUP mode, the current limit is raised by 10 to 50%, in the present case by 20%. A STARTUP signal 100 stipulates whether the control logic of the motor in digital block 40 (fig. 1) is in start up or stable controlled mode. On starting, the motor is accelerated as soon as three BEMF zero crossing ZERO pulses 18 are detected. On conclusion of the STARTUP process the speed regulator 28 (fig. 1) takes over the motor and guides this to the required nominal value. Although according to fig. 10 the boost is activated, the motor does not immediately accelerate so quickly that ZERO signals 18 can be generated, the cause of this failure could be too high a load or the rotor in an unfortunate position. The START algorithm now, after expiry of a specified number of time-outs 98, itself generates NEXT pulses 22. After three ZERO signals 18 the STARTUP signal 100 goes to LO (low) and the cycle counter 24 (fig. 1) takes over the motor. The current limit over the time t is shown from the lower part of fig. 10. On connection of the cycle counter 24, the current limit Boost LB in area 102 falls from 120% to the current limit LN normal mode 104 of 100%. Boost and STARTUP are switched off when three ZERO pulses 18 have been detected i.e. a rotation speed of around 200 rpm has been reached. Fig. 11 shows a DC motor 46 with a twelve-pole (6m) rotor 54 with six (3n) magnetic pole pairs north N and south S. The stator 44 comprises nine stator poles 52 or armatures each with three stator coils A, B and C, i.e. n=3 and m=6. The permanent magnet rotor 54 is held by means of a drive shaft mounted in a base plate 106 with a cup-shaped external rotor. Fig. 12 shows an actuating drive 112 with an actuator 114 in the form of a swivellabie flap in an HLK pipe 116. A DC motor 46 swivels the flap by means of a highly reduced reduction gear 118. Fig. 13 shows an actuating drive 112 formed as a spring drive. Between the DC motor 46 and the actuating drive 114 is formed gear mechanics unit 118', 118" which is supplemented by a spiral spring 120. A first reduction gear 118' is highly reduced, a second reduction gear 118" only relatively little. The spring 120 is pre- tensioned and on power failure closes the flap-like actuator 114. Two fast a closure is prevented by the measures according to the description of fig. 2, a short-circuit of two stator coils to earth takes place in the manner shown there. All individual elements of the actuating drive 112, DC motor 46, gear 118 or 118', 118", spiral spring 120 and actuator 114 are rigidly connected together. The end position, the closed position of the flap-like actuator 114, is reached when it lies against the stop cam 122. WE CLAIM:- 1. Method for controlling a brushless DC motor (46) comprising a stator (44) and rotor (54), which has a highly reduced reduction gear of an actuating drive for a flap or valve in order to regulate a gas or liquid volume flow, in particular in the field of heating, ventilation and climastisation (HLK), fire and room protection, the electrical current flow in the stator coils (A, B, C) being communicated under programme control without integral position sensors, characterized in that the electronics count the revolutions of the rotor (54) of the DC motor (46), and supplied to a counter with a logic rule, and in that the logic rule for start and stop cause motor speeds to reduce below a threshold value to be passed quickly and without the occurrence of counting errors. 2. Method as claimed in claim 1, wherein the polarities of the zero crossings of the phase voltages (DRA, DRB, DRC) are checked with BEMF comparators (12), each polarity change is allocated to a rotor position and the next communication determined. 3. Method as claimed in claim 2, wherein the valid zero crossings of three phase voltages (DRA, DRB, DRC) are detected BEMF signals (13) control the switch the three phase voltages (DRA, DRB, DRC). 4. Method as claimed in claim 3, wherein the valid zero crossings (56) are detected by means of a ZERO detector (14) and ZERO pulses (18) are generated. 5. Method as claimed in any of claims 1 to 4, wherein from the BEMF signals (13) which on valid zero crossings cause ZERO pulses (18), a period counter (20) calculates the next commutation time and a corresponding control signal, a NEXT pulse (22), is emitted. 6. Method as claimed in any of claims 2 to 5, wherein the direction of rotation of the rotor (54) is determined by means of the BEMF comparators (12) in that the BEMF comparator thresholds (58), depending on polarity of the expected zero crossing, are located just above or below the zero lines(s) of the voltages (DRA, DRB, DRC). 7. Method as claimed in claim 6, wherein the direction of rotation of the rotor (54) in the high impedance communication gaps (L) of the phase voltages (DRA, DRB, DRC) is determined at low speeds preferably at the end of the communication gap and at high speeds at the end of the passive PWM phase. 8. Method as claimed in any of claims 1 to 7, wherein the DC motor (46) commutates the end stages on start up and their current limit (102) is increased for an accelerated start, preferably by 10 to 50%, in particular by around 20%, until valid BEMF are reached, after which communication of the next phase is located between two zero crossings. 9. Method as claimed in claim 8, wherein after an unsuccessful powering of the first end stage, a logic commutates in steps the current supply of the next end stage until a starting position which is favourable for the rotor (54) is reached. 10. Method as claimed in claim 8 or 9, wherein on lack of detection of a zero crossing of a phase voltage, after a first communication the rotor oscillates with its resonant frequency into the zero position and by theses zero crossings, valid BEMF signals are achieved. 11. Method as claimed in any of claims 1 to 10, wherein the logic detects a wrong direction of rotation of the rotor (54) from the wrong polarity of the BEMF signals (13) and corrects this by increasing the counter state of the cycle counter (24) by two steps. 12. Method as claimed in any of claims 1 to 11, wherein the rotor (54) in a first for braking and holding first by short circuit switches all stator coils (A, B, C) powerless under continuing communication, after reaching a pre-set threshold value of preferably maximum around 400 rpm in a second step, with a greatly increased pulse (65) leading to block, changes from the next to the last communication state, and in a third step again by short circuit switches the stator coils (A, B, C) of rotor (54) powerless or inhibits them with low pulse width modulation (PWM). 13. Method as claimed in claim 12, wherein the rotor (54) in the first step is switched activity in the opposite direction. 14. Method as claimed in claim 12 to 13, wherein to achieve and hold a high contact pressure by the reduction gear in the stop position of the flap or valve, the DC motor (46) which is stalled by the stop, because of the rotor inertia torque continues to rotate uncontrolled for a few revolutions, the rotor position is detected by an immediately subsequent brief acceleration in the opposite direction by means of BEMF signals (13), and the rotor (54) is braked under control and held. 15. Method as claimed in claim 14, wherein acceleration in the opposite direction takes place only after relaxation of the reduction gear. 16. Method as claimed in any of claims 1 to 15, wherein the process algorithms are stored retrievably. 17. Method as claimed in any of claims 1 to 16 in an actuating drive with a pretensioned spring (120), in particular a coil spring (120), wherein on a power failure two of the three individually moveable bridge switches (a, b, c) of a switch group (50) short-circuit two of the three stator coils (A, B, C) of the DC motor (46) to earth when the speed falls below an adjustable limit speed of the rotor (54). 18. Method as claimed in claim 17, wherein a bridge switch pair (a and b, a and c, b and c) is activated by way of a pulse width modulated control signal from the allocated transistors, preferably with increasing time intervals. 19.Brushless DC motor (46) without position sensors for performance of the method according to any of claims 1 to 20, wherein it comprises a stator (44) with three stator coils (A, B, C) which extend over 3n stator poles (52) and an annular permanent magnet rotor (54) with 2n permanent magnet segments of alternating polarity (N, S), a coaxial cup-shaped external rotor (110) and a coaxial drive shaft (108), where a and m are whole multiplication factors 1, 2, 3, 4 independent of each other. 20. DC motor (46) as claimed in claim 19, wherein n = 1 and m = 1 or n = 3 and m = 6. ABSTRACT Title: METHOD FOR CONTROLLING A BRUSHLESS DC MOTOR Method for controlling a brushless DC motor (46) comprising a stator (44) and rotor (54), which has a highly reduced reduction gear of an actuating drive for a flap or valve in order to regulate a gas or liquid volume flow, in particular in the field of heating, ventilation and climastisation (HLK), fire and room protection, the electrical current flow in the stator coils (A, B, C) being communicated under programme control without integral position sensors, characterized in that the electronics count the revolutions of the rotor (54) of the DC motor (46), and supplied to a counter with a logic rule, and in that the logic rule for start and stop cause motor speeds to reduce below a threshold value to be passed quickly and without the occurrence of counting errors.

Full Text

The invention relates to a method for controlling a brushless DC motor comprising
a stator and a rotor, which has a highly reduced reduction gear of an actuating
drive for a flap or valve in order to regulate a gas or liquid volume flow, in
particular in the field of heating, ventilation and climatisation (HLK), fire and room
protection. Furthermore the invention concerns a brushless DC motor without
position sensors for performance of the method.
Electric actuating drives for motorisation of actuators in heating, ventilation and
climatisation systems (HLK systems) have been produced for more than 30 years.
HLK actuators guarantee an economic volume flow regulation of gases or liquids,
in particular air and water. As a compact unit they comprise not only the drive but
also pressure sensors and regulators, all combined in one device.
Ventilation systems are increasingly used in buildings, in particular residential,
office, commercial and industrial buildings, normally combined with fire and smoke
protection devices. In ventilation systems volume flow regulation with swivelable
air flaps plays an essential role. The volume flow is measured with a suitable
measuring instrument, for example with the NMV-D2M by Belima Automation AG,
CH-8340 Hinwil, which is formed as a compact unit of drive, pressure sensor and
regulator, and which transmits measurement values to the electronics.
To move a flap in a ventilation system or a ball valve in a water pipe system,
relatively weak motors must activate large area or large volume control elements.
Precise and stable adjustment is possible only with a very high reduction. To
swivel a flap or turn a ball valve through an acute or right angle requires
numerous rotations of the shaft of the electric motor.
Between the motor and valve can be arranged mechanical means for
transmission of the motion, in particular spring drives.

DC motors originally produced preferably have brushes made of graphite which
mechanically regulate the commutation of the individual electrical phases of the
rotor, the stator in this case being fitted with permanent magnets.
Electronically commutating DC motors of the known type, also called EC motors,
are particularly suitable for applications which utilise the benefits of contactless
and electronically controlled commutation. With commutation controlled in this
way the electronic current feed of the stator coils shifts in steps from pole to pole,
"pulling" the rotor with it. Brushless DC motors are characterised by low
production costs and a long life in conjunction with preferential regulation
properties for speed and precision.
A torque generated by such an electromagnetic rotating field is almost constant
irrespective of the position of the rotor, and dependent in a linear manner on the
three-phase current.
Each of the three phases of a brushless DC motor is controlled by one or two
transistors, depending on whether the coils are to be powered in one or two
directions. With just three control pulses per electrical rotation, the windings are
utilised relatively poorly. Controlling the windings with six transistors guarantees a
higher utilisation of the windings and a more even torque output. The coils can be
connected in a star or triangle, in both cases the torque is greater if each of the
three phases is controlled by two transistors.
In order to determine the suitable moment to switch the power transistors during
commutation, the present position of the rotor in relation to the stator must be
known. Particularly suitable here are silicon-free resolvers and Hall effect sensors,
both of which give the absolute position of the rotor irrespective of speed. For
reasons of cost and miniaturisation, the use of sensors is however always more
problematical as they must be replaced frequently.
A regular torque is generated for example if the coils are controlled with sinusoidal
signals offset by 120°. At lower power rectangular signals can also be used for

control because of their steeper zero crossing. As drivers so-called H-bridges are
used where each pole of a coil is connected by way of a transistor with a supply
voltage and earth. It must however be guaranteed that in the case of two
transistors, these are not both conductive at the same time. The H-bridges of the
power electronics are controlled as a function of signals from position sensors.
They serve for direction formation and are controlled as stop, start and PWM
signals using an encoder (PWM = pulse width modulation). The direction of
rotation of the rotor is determined by the switching algorithm, the rotation speed of
the motor by way of the coil voltage of the coils connected. This can also be
triggered by a PWM.
The commutation frequency depends on the specified rotation speed of the motor
and the number of pole pairs of the rotor.
In practice two or three stator coils are always powered simultaneously, the
difference of two or three coils must be taken into account. The induced voltages
of three coil combinations are phase-offset by 120°. By suitable commutation now,
depending on direction of rotation, the maximum positive or negative torque can
be utilised. A delayed or premature commutation has a direct effect on the torque
of the motor to be provided. The optimum switching point for commutation of two
and three powered coils differs, it is controlled by commutation algorithms.
The rotor position is normally determined on the motor circuit board by three
position sensors, for example Hall effect sensors. Hall effect sensors, when
supplied with a constant current, emit a differential voltage as a function of the
magnetic field flowing through them, known as the Hall effect.
DE 19745312 A1 discloses a control circuit for a brushless electronically
commutated DC motor. Arranged in series with each winding strand of the motor
is an electronic switch element which is controlled by a position sensor. Each
winding strand is connected by way of a decoupling diode with a common switch
point. Arranged parallel to the winding strands and associated decoupling diodes
is a protective transistor circuit which eliminates induction-induced voltage peaks.

Although a good efficiency is achieved with simple low-cost components, the
disadvantage remains that position sensors need to be replaced.
The inventors have faced the task of creating a method for controlling a brushless
DC motor and a DC motor of the type cited initially that eliminates the cost and
quality problems related to position sensors.
In relation to the method, the object is achieved according to the invention in that
the electrical current flow in the stator coils is commutated under programme
control without integral position sensors. Special and refined embodiments of the
method are the object of the dependent claims.
The algorithms of all essential procedure processes are entered and stored
retrievably in a CPU, suitably in at least one microprocessor, for example a start
algorithm, a brake algorithm, a brake algorithm for reverse kick at the stop i.e. a
controlled return and braking after a hard stop, wherein the contact pressure is
substantially maintained, and finally a relax algorithm with which the reduction
gear is fully relaxed.
the retrievably stored algorithms are continuously adapted to practice and
improved where necessary i.e. the system is self-learning.
The basic principle according to the invention is to check the polarities of the zero
crossings of the three motor phase voltages. Each polarity change on a zero
crossing of a phase voltage can be allocated to a particular rotor position. As soon
as the corresponding rotor position is reached, the next commutation, i.e. the next
current flow to the stator coils, can be modified such that the generated leading
magnetic field gives the rotor a torque. Thanks to this system, position sensors
can be omitted.
The zero crossings of the phase voltages can however only be clearly detected
from a particular minimum rotation speed of the rotor, rotation speeds that are too
low give undefined looping sections of sinusoidal or sine-like curves with the zero

line.
In a special embodiment of the invention, an electronics unit counts the
revolutions per minute (rpm) of the rotor of the DC motor. Preferably each motor
commutation is detected and counted, wherein special algorithms for start and
stop ensure that motor rotation speeds below around 200 rpm are passed through
so quickly that no counting errors occur.
Provided that the counters once at the start have a synchronisation or reference
point for allocation to the flap angle, the position regulator always knows precisely
the actual value i.e. the precise, position of the flap concerned. The counting
electronics in this way constitute an almost no-cost replacement for a relatively
costly potentiometer that is subject to wear in conventional arrangements.
With reference to the DC motor the object is achieved according to the invention
in that it comprises a statpr with three coils that extend over 3n stator poles and
an annular permanent magnet rotor with 2m permanent magnet segments of
alternating polarity (N, S), a cup-shaped external rotor and a coaxial drive shaft,
where n and m are whole multiplication factors 1, 2, 3, 4.... independent of each
other.
Coils and rotor segments can be used together following the rule of combination,
in practice however the combinations n=1 and m = 1, and n = 3 and m = 6 have
proved particularly advantageous.
An actuating drive activated by a DC motor according to the invention by way of a
gear system can be supplemented with a spring, suitably a coil spring, which can
be extended or pre-tensioned on motorised operation and hence potential energy
stored. The spring is extended preferably but not necessarily in the lift direction.
An electronic power fail detection detects failure of the operating voltage, first
disengages all bridge switches of the end stage, and thus triggers a return of the
pretensioned spring. The DC motor thus becomes free running, the potential
energy stored in the spring acts as a torque against the lift direction and sets the

free running motor in motion. Depending on the load at the gear output, the pre-
tensioned spring accelerates to a correspondingly high speed until the load,
friction and spring moment are in equilibrium. When the said load at the gear
output supports the spring moment or only loads this lightly, the rotation speed in
the spring return can be increased so that the entire mechanical system is
endangered. Reaching the end stop is particularly critical. In order to limit the
acting kinetic energy and hence the load to a safe level, the rotation speed must
be limited. For this in the known manner Zener diodes can be used which limit the
motor voltage generated by the generator and disperse any additional energy as
heat.
Preferably however use is made of the property of the DC motors that on a short
circuit of at least part of their windings, a braking moment occurs which is
approximately proportional to the rotation speed. This can be achieved easily on
brushless DC motors according to the invention in that two of the three windings
are short-circuited. The extent of braking can be set by way of a pulse width
modulated control signal (PWM signal). This is done as a function of the actual
rotation speed which can be determined at any time by way of the measured zero
crossings of the third motor winding that is not short-circuited. Below a
programmable limit speed, the PWM signal is set to zero and after exceeding the
limit speed, rises continuously to a maximum of 100%. This allows a simple
regulation for the speed limitation which can thus be held constant over a wide
load range. This prevents, on power failure, the pre-tensioned spring closing the
actuation element concerned, for example a valve or flap, too quickly.
Although the method according to the invention and the brushless DC motor are
designed in particular for the HLK field, this is not a restriction. Due to dimension-
induced adaptations, gas flows, volume flows and/or actuation angles can also be
regulated in automobiles in a similar fashion with professional knowledge and
ability.
The invention is now explained in more detail below with reference to the
embodiment examples shown in the drawing which are also the object of

dependent claims. The drawings indicate schematically:
- Fig, 1 a block circuit diagram of the entire motor control system for a DC
motor,
- Fig. 2 a wiring diagram for the stator of a DC motor,
- Fig. 3 a commutation cycle of a DC motor with six-rotor positions,
- Fig. 4 the proportion of individual stator coils in the motor torque according to
fig. 3,
- Fig. 5 the connection of the phase voltages with the expected polarity of the
zero crossing,
- Fig. 6 braking by blocking,
- Fig. 7 a brake-stop-start cycle,
- Fig. 8 the behaviour of a DC motor at the stop,
- Fig. 9 the relaxing of a DC motor at the stop,
- Fig. 10 the behaviour of a DC motor when starting under difficult conditions
(boost concept),
- Fig. 11 a variant of a DC motor,
- Fig. 12 an actuating drive with a DC motor, and
- Fig. 13 a spring drive according to fig. 12.
Fig. 1 shows a block circuit diagram of a motor control 10 for a DC motor (46 in
fig. 2). As is normal for a very high reduction, this motor has low power and small
dimensions. For details of a reduction gear, reference is made for example to
PCT/CH05/000126 by the applicant.
Three motor generator currents are conducted into the motor control 10 to control
its voltage DRA, DRB and DRC by way of BEMF comparators 12 (BEMF = back
electromagnetic force). The BEMF comparators 12 in the known manner check
the polarity of the zero crossings of the phase voltages DRA, DRB and DRC.
Each change of polarity, in other words each zero crossing, can be allocated to a
particular rotor position of the DC motor 46. As soon as this polarity is known, the
next commutation i.e. the new current supply to the stator coils (A, B, C in fig. 2) is
determined.

A subsequent ZERO detector 14 detects the valid zero crossings of the phase
voltages DRA, DRB and DRC or their BEMF signals 13, which are indicated by
the needle-like ZERO pulses 18. The three phase voltages DRA, DRB and DRC
can only be analysed when the ZERO detector 14 receives the corresponding
signal 16, which is indicated with a reverse arrow.
A period regulator 20 receives the ZERO pulses 18 and from these calculates the
next commutation time. If the next commutation is actually to take place, a NEXT
pulse 22 is passed to a cycle counter 24. Furthermore the period regulator 20
outputs a PER signal 26 to a speed regulator 28. PER (PER = time period) means
the distance between two ZERO pulses, and is thus inversely proportional to the
rotation speed. A speed regulator 28 controls and regulates the number of
revolutions of the rotor per time unit, usually per minute, which is indicated as
rpm.
The cycle counter 24 counts the whole digit multiples 0, 1,2, 3... and at a motor
control 30 outputs a counter value 32 indicated with a CYNO (cycle number).
Each of the six counter values 32 in the present case is firmly allocated to one of
the six possible H-bridge states to be discussed below. As stated a NEXT pulse
22 controls the cycle counter 24 which thus controls the commutation. As also
stated the counter value 32 CYNO gives a feedback in the form of a signal 16 to
the ZERO detector 14.
The speed regulator 28 controlled by the PER signal 26 takes place in hardware
terms by way of a PI regulator (PI = proportional / integral) of the known type. The
nominal value is taken from an OTP cell (OTP = one time programmable).
Optionally by way of microprocessors (MP) arbitrary nominal values can be
written to a volatile register. The PER signal 26 of the period regulator 20 is used
as the actual value. This as stated is inversely proportional to the rotation speed
and is calculated out in the speed regulator 28. By way of a PWM clock generator
34, a PWM signal 36 is output to the motor control 30, the duration of a pulse in
the present case being 52 usec. This pulse width modulation PWM corresponds

to the voltage applied to the DC motor 46. The PWM frequency of the clock
generator 34 in the present case is 19.2 kHz.
The motor control 30 is generated by the PWM modulation 36 and the
commutation. The output phase voltages DRA, DRB and DRC are the direct
controls for the motor driver in the form of power H-bridges.
An analog interface 38 allows control with analog signals. AZ applications (AZ =
on/off) are also possible, and with an external potential, constant SR applications
(SR = constant regulator) can also be achieved.
A digital block 40 from a status control (state machine) controls the internal
processes of the motor control 10. For starting and braking of the DC motor 46,
the digital block 40 generates an ROT signal 42 which it passes to the speed
regulator 28.
In a wiring diagram according to fig. 2 for the stator 44 of a DC motor 46, the
poles of a DC power source 48 are indicated with + and -. The stator 44 has three
stator coils A, B and C arranged in a star pattern and connected together at a star
point S that also forms the zero line. A switch group 50 which is activatable under
program control generates a commutation cycle comprising six commutations
which is shown in detail in fig. 3 below.
On a power failure two of the three bridge switches a, b and c are short-circuited
to earth, causing a braking moment. In the present case these are bridge switches
a and b, whereby a short circuit is formed over the stator coils B and C. Stator coil
A is switched off but not short-circuited. Suitably the bridge switches a, b are
closed only initially for brief periods which then become ever longer. Generating a
braking moment prevents, on a power failure, an actuator returning without control
and causing damage. The paired closure of any two bridge circuits a, b, c takes
place by way of transistors which receive a pulse width modulated control signal.
Fig. 3 shows six positions of a permanent magnet rotor 54 of a DC motor 46. The

rotor 54 in the present case has four magnetic segments of equal size, alternately
allocated to a north pole N and a south pole S. The six rotor positions constitute a
commutation cycle wherein the stator coils A, B and C are powered in pairs with +
and -. The power application is programme-controlled and by way of stator 44
commutates a rotating magnetic field which "pulls" the permanent magnets of the
rotor 54 with a leading rotary field. The armatures 52 of the three stator coils A, B
C extend in rotor form over part of the inner casing of rotor 54.
Angle φ designates the mechanical position of the rotor 54 in relation to the zero
point. From left to right φ is 15°, 45°, 75°, 105°, 135° and 165°.
Fig. 4 shows the proportion of the individual stator coils A, B and C in the torque,
in other words the connection between the rotor position and the torques acting
on the rotor 54. The mechanical angle φ (fig. 3) of rotor 54 is indicated with
vertical lines. These lines also correspond to the electrical position 2φ on the
abscissa, 30°, 90°, 150°, 210°, 270° and 330° of the respective mechanical angle
φ.
The diagram shows a four-pole DC motor 46 (Figs. 2, 3) and not a twelve-pole
motor which is normal in practice because the connections are easier to show
with as simple a motor as possible. A twelve-pole motor turns three times more
slowly than a four-pole motor, otherwise there are no substantial differences.
The torque designation M shows the individual proportions of the stator coils A, B,
C in the total torque MT. Each of these stator coils acts with a sinusoidal torque on
the rotor. As each stator coil is switched off for 60°, for 30° on both sides of the
zero, this torque proportion is absent there. This is reflected in the undulating form
of the total torque MT. On zero crossing 56 of the sinusoidal torque MB where the
power is supplied only through stator coils A and C, a stable position is achieved.
The same applies to the torque of stator coils A and C. The zero crossings of the
torques MA, MB and Mc are designated below as 56.
Below the torque curves M are shown further parameters in the respective

positions mechanically φ or electrically 2φ:
- BEMF. Curves BEMF-A, BEMF-B and BEMF-C of the individual stator coils A,
B and C run sinusoidally. In the ranges ± 30° from the zero crossing, i.e. the
intersection of the sine curve with the zero line S in the range of the high
impedance commutation gaps L (fig. 5), the curves are drawn in solid lines.
- ZERO. At φ = 15°, 45°, 75°, 105°, 135° and 165°, a ZERO pulse 18 is
triggered i.e. a valid zero crossing of the BEMF comparator 12 (fig. 1) is
registered. The angle between the zero crossings is designated PER.
- NEXT. Also indicated with needle pulses are the NEXT pulses 22 which
trigger the times of the commutation change. These are offset temporally to
the ZERO pulses 18 by an angle DEL = PER/2. In the present case the
duration of the needle pulse is ZERO and NEXT is 52 microseconds (usec).
- CYNO. Finally CYNO shows the counter states of the cycle counter 24 (fig. 1)
from 0 to 5, in other words showing which commutation is occurring at the
respective q> on the DC motor.
Fig. 5 shows the correlation of the phase voltages DRA, DRB and DRC with the
high impedance commutation gaps L of electrically 2φ = 60° in the present case.
In these high impedance commutation gaps L, at low rotation speed at their end
and at high rotation speed at the end of the power phase, the BEMF comparator
12 (fig. 1) analyses the phase voltage. The BEMF comparator thresholds 58,
depending on polarity of the expected zero crossing, are located just - by Vs -
above or below the zero line S. This increases the fault distance and supplies
more reliable ZERO pulses 18, which brings improvements mainly at low speeds.
At the start of a high impedance commutation gap L, a return pulse 64 is also
generated. A zero crossing detected correctly by the BEMF comparator 12
generates a positive SCMP flank. The SCMP (= synchronised comparator) is
reset on commutation of the H-bridge. The said flanks generate the ZERO pulses
18 that are 52 microseconds wide in the present case. The distance between the
ZERO pulses 18 is the "ten bit word" PER. This is inversely proportional to the
rotation speed. PER/2 gives the DEL (= delay). Commutation occurs between two
ZERO pulses 18. CYNO is the value of the 0, 1, 2, 3, 4, 5 cycle counter and

indicates which commutation is present at the DC motor. A commutation cycle f
corresponds to 1/commutation frequency.
Fig. 6 shows an abrupt braking, also known as a blocking or lever braking. The
positions of the rotor 54 correspond to the left half according to fig. 3. At a
mechanical angle φ of 30° a strong current pulse 65 generates an amplified
resulting torque -B+A which is indicated by a falling dotted line 66. The rotor is
stopped with immediate effect. The holding point is stable in the range of 2cp ±
180° electrically i.e. φ ± 90° mechanically. Within this range the rotor returns to
the holding point provided. This braking method is highly reliable. The total torque
TM diminishes drastically.
The counter value CYNO first rises from x+0 to x+1 but then after the current
pulse falls again to x+0. In analogy with the braking of a vehicle with spoked
wheels by insertion of a rod, the brake process according to fig. 6 is called the
"lever brake".
According to fig. 6, at the commutation time, instead of the next phase the last
phase is activated again, because at this time this has precisely the torque zero
crossing.
The braking according to fig. 6 with PWM = 0 is a simple stop method for a DC
motor but has the disadvantage that at low speeds the braking moment also
diminishes. The strong additional current pulse 65 must be applied at the correct
moment, which is not easy because the motor position is difficult to predict.
Preferably therefore the speed is regulated down to 350 to 400 rpm, whereupon
the current pulse 65 takes place which blocks the rotor immediately. The correct
moment for the current pulse can then no longer be adapted.
Another method for braking a DC motor is indicated in fig. 7 with a cycle of brake-
stop-start. The normal operating sequence is characterised by arrow 68. On
initiation of braking, the status control (state machine) in digital block 40, by
means of a falling flank 43 of the ROT signal 42, tells the motor control 30 by way
of PWM clock generator 34 (fig. 1) that the DC motor should be stopped, which

takes place with a PWM signal 36. This sets the PWM to 0 i.e. the H-bridges now
short-circuit the stator coils, which brakes the DC motor.
However commutation continues from the BEMF analysis, the rotor positions are
therefore still known as indicated by arrows 70, 72. As soon as the speed control
establishes from the PER signal 26 that the rotation speed has fallen below a
value of 400 rpm, the stop process initiated by arrows 74, 76 is initiated. The
motor control 30 switches to the last instead of to the next commutation state and
the corresponding stator coils are powered with a PWM = 80 for 100 ms. This
stops the rotor immediately. The counter value CYNO is decremented by 1. After
the fast stop in the sense of a lever inserted in the spokes of a wheel, during a
holding phase the PWM is reduced to 15 or 0. The restart of the rotor is triggered
by the rising flank of the ROT signal 42, a NEXT pulse 22 is generated by way of
arrow 79 and the stator coils of the powerless or inhibited PWM circuit are
increased to 127. After the start, commutation takes place as normal using the
BEMF analysis.
Fig. 8 shows the behaviour of a DC motor in practical operation when for example
an air flap hits a stop. The abscissa shows the time t and the ordinate from left to
right the rotation speed v, the motor current AM and the torque 0.
The rotation speed curve v first runs parallel to the abscissa and the rotation
speed is constant. To initiate the braking process (arrow 81), the motor current A
is increased until the current limit AM is reached (arrow 83), with a slight temporal
delay the torque θ also rises. The braking process (arrow 81) is initiated when the
flap hits the stop. As in fig. 7 the PWM is set to 0, the H-bridges short circuit the
stator coils. Even after the flap hits the stop, the DC motor runs on, it stalls until
reaching the stop 80 where no further zero crossings can be measured, and the
position of the rotor is lost. After stoppage of the DC motor at peak 82, thanks to
the elasticity of the highly reduced, tensioned gear, the motor turns backward until
the flap remains on the stop with the residual contact pressure. The position of the
rotor can be detected again with special measures when the reduction gear in
point 84 stops in the range of the controlled braking 86 ("kick brake"). If the

reduction gear turned further back beyond this self-inhibition, the contact pressure
of the flap would be lost. After reaching the peak 82 the DC motor is accelerated
briefly in the opposite direction so that BEMF signals 13 (fig. 1) are output and a
defined holding current can be set. The logic again knows the rotor position. Now
the rotor can be braked under control with PWM = 0 as soon as the rotation speed
for example has fallen below 400 or 500 rpm. The brief turning back of the DC
motor reduces the contact pressure of the air flap only slightly since for this only a
few motor revolutions are required.
The motor current AM rises in the area of the controlled braking 86 ("kick brake")
which is indicated with boost motor current increase 88 and then falls to a lower
level AM2 which in the present case is around 20 to 30% of the nominal current.
The reduced motor current AM2 is operated during the air flap holding phase with a
low PWM.
The torque θ rises until the controlled braking 86 ("kick brake") begins, reaches a
maximum peak 89 and falls during the stop braking 86 to the constant torque in
the holding position, the holding moment 9H. The dynamic part of the stop force is
indicated with d.
Fig. 9 shows the rotation speed v and the torque θ over time t. In contrast to fig. 8,
no holding moment ΘH is maintained here, the reduction gear is fully relaxed after
an unbraked stop of the air flap driven by the DC motor. At the stop 80 of the air
flap, the DC motor, thanks to the elasticity of the reduction gear, runs on until
peak 82 is reached. Thanks to the elasticity the motor turns back and reaches
stoppage 90. Immediately after the flap has stopped (area 80) the DC motor is
short-circuited for a few seconds tR. This has a mechanical damping effect during
relaxation (turning back of the reduction gear). In other words at low speeds the
reduction gear turns uninhibited, while at high speeds a short-circuit current flows
which brakes the rotor.
When the rotation speed v falls below 200 rpm in the present case, no further
BEMF signals are generated and the commutation positions are lost. After the

stoppage position 90 the DC motor is accelerated briefly in the opposite direction
and then braked under control. Depending on setting, now in the motor a torque is
impressed or the motor current is switched powerless. The motor is now with a
known commutation position in the holding position as indicated with arrow 90.
This latter stage is called the reverse pulse 96 (reverse kick).
On engagement of the "boost" function according to fig. 10, the DC motor 46 (fig.
2) receives more torque on starting. This allows it to start even under difficult
conditions. As long as the ASIC (application specific integrated circuit) is in
STARTUP mode, the current limit is raised by 10 to 50%, in the present case by
20%. A STARTUP signal 100 stipulates whether the control logic of the motor in
digital block 40 (fig. 1) is in start up or stable controlled mode. On starting, the
motor is accelerated as soon as three BEMF zero crossing ZERO pulses 18 are
detected. On conclusion of the STARTUP process the speed regulator 28 (fig. 1)
takes over the motor and guides this to the required nominal value. Although
according to fig. 10 the boost is activated, the motor does not immediately
accelerate so quickly that ZERO signals 18 can be generated, the cause of this
failure could be too high a load or the rotor in an unfortunate position. The START
algorithm now, after expiry of a specified number of time-outs 98, itself generates
NEXT pulses 22. After three ZERO signals 18 the STARTUP signal 100 goes to
LO (low) and the cycle counter 24 (fig. 1) takes over the motor. The current limit
over the time t is shown from the lower part of fig. 10. On connection of the cycle
counter 24, the current limit Boost LB in area 102 falls from 120% to the current
limit LN normal mode 104 of 100%. Boost and STARTUP are switched off when
three ZERO pulses 18 have been detected i.e. a rotation speed of around 200
rpm has been reached.
Fig. 11 shows a DC motor 46 with a twelve-pole (6m) rotor 54 with six (3n)
magnetic pole pairs north N and south S. The stator 44 comprises nine stator
poles 52 or armatures each with three stator coils A, B and C, i.e. n=3 and m=6.
The permanent magnet rotor 54 is held by means of a drive shaft mounted in a
base plate 106 with a cup-shaped external rotor.

Fig. 12 shows an actuating drive 112 with an actuator 114 in the form of a
swivellabie flap in an HLK pipe 116. A DC motor 46 swivels the flap by means of a
highly reduced reduction gear 118.
Fig. 13 shows an actuating drive 112 formed as a spring drive. Between the DC
motor 46 and the actuating drive 114 is formed gear mechanics unit 118', 118"
which is supplemented by a spiral spring 120. A first reduction gear 118' is highly
reduced, a second reduction gear 118" only relatively little. The spring 120 is pre-
tensioned and on power failure closes the flap-like actuator 114. Two fast a
closure is prevented by the measures according to the description of fig. 2, a
short-circuit of two stator coils to earth takes place in the manner shown there.
All individual elements of the actuating drive 112, DC motor 46, gear 118 or 118',
118", spiral spring 120 and actuator 114 are rigidly connected together.
The end position, the closed position of the flap-like actuator 114, is reached
when it lies against the stop cam 122.

WE CLAIM:-
1. Method for controlling a brushless DC motor (46) comprising a stator (44) and rotor
(54), which has a highly reduced reduction gear of an actuating drive for a flap or
valve in order to regulate a gas or liquid volume flow, in particular in the field of
heating, ventilation and climastisation (HLK), fire and room protection, the electrical
current flow in the stator coils (A, B, C) being communicated under programme
control without integral position sensors, characterized in that the electronics count
the revolutions of the rotor (54) of the DC motor (46), and supplied to a counter with
a logic rule, and in that the logic rule for start and stop cause motor speeds to
reduce below a threshold value to be passed quickly and without the occurrence of
counting errors.
2. Method as claimed in claim 1, wherein the polarities of the zero crossings of the
phase voltages (DRA, DRB, DRC) are checked with BEMF comparators (12), each
polarity change is allocated to a rotor position and the next communication
determined.
3. Method as claimed in claim 2, wherein the valid zero crossings of three phase
voltages (DRA, DRB, DRC) are detected BEMF signals (13) control the switch the
three phase voltages (DRA, DRB, DRC).
4. Method as claimed in claim 3, wherein the valid zero crossings (56) are detected by
means of a ZERO detector (14) and ZERO pulses (18) are generated.
5. Method as claimed in any of claims 1 to 4, wherein from the BEMF signals (13)
which on valid zero crossings cause ZERO pulses (18), a period counter (20)
calculates the next commutation time and a corresponding control signal, a NEXT
pulse (22), is emitted.

6. Method as claimed in any of claims 2 to 5, wherein the direction of rotation of the
rotor (54) is determined by means of the BEMF comparators (12) in that the BEMF
comparator thresholds (58), depending on polarity of the expected zero crossing, are
located just above or below the zero lines(s) of the voltages (DRA, DRB, DRC).
7. Method as claimed in claim 6, wherein the direction of rotation of the rotor (54) in
the high impedance communication gaps (L) of the phase voltages (DRA, DRB,
DRC) is determined at low speeds preferably at the end of the communication gap
and at high speeds at the end of the passive PWM phase.
8. Method as claimed in any of claims 1 to 7, wherein the DC motor (46) commutates
the end stages on start up and their current limit (102) is increased for an
accelerated start, preferably by 10 to 50%, in particular by around 20%, until valid
BEMF are reached, after which communication of the next phase is located between
two zero crossings.
9. Method as claimed in claim 8, wherein after an unsuccessful powering of the first
end stage, a logic commutates in steps the current supply of the next end stage until
a starting position which is favourable for the rotor (54) is reached.
10. Method as claimed in claim 8 or 9, wherein on lack of detection of a zero crossing of
a phase voltage, after a first communication the rotor oscillates with its resonant
frequency into the zero position and by theses zero crossings, valid BEMF signals
are achieved.
11. Method as claimed in any of claims 1 to 10, wherein the logic detects a wrong
direction of rotation of the rotor (54) from the wrong polarity of the BEMF signals (13)
and corrects this by increasing the counter state of the cycle counter (24) by two
steps.

12. Method as claimed in any of claims 1 to 11, wherein the rotor (54) in a first for
braking and holding first by short circuit switches all stator coils (A, B, C) powerless
under continuing communication, after reaching a pre-set threshold value of
preferably maximum around 400 rpm in a second step, with a greatly increased
pulse (65) leading to block, changes from the next to the last communication state,
and in a third step again by short circuit switches the stator coils (A, B, C) of rotor
(54) powerless or inhibits them with low pulse width modulation (PWM).
13. Method as claimed in claim 12, wherein the rotor (54) in the first step is switched
activity in the opposite direction.
14. Method as claimed in claim 12 to 13, wherein to achieve and hold a high contact
pressure by the reduction gear in the stop position of the flap or valve, the DC motor
(46) which is stalled by the stop, because of the rotor inertia torque continues to
rotate uncontrolled for a few revolutions, the rotor position is detected by an
immediately subsequent brief acceleration in the opposite direction by means of
BEMF signals (13), and the rotor (54) is braked under control and held.
15. Method as claimed in claim 14, wherein acceleration in the opposite direction takes
place only after relaxation of the reduction gear.
16. Method as claimed in any of claims 1 to 15, wherein the process algorithms are
stored retrievably.
17. Method as claimed in any of claims 1 to 16 in an actuating drive with a pretensioned
spring (120), in particular a coil spring (120), wherein on a power failure two of the
three individually moveable bridge switches (a, b, c)

of a switch group (50) short-circuit two of the three stator coils (A, B, C) of the DC
motor (46) to earth when the speed falls below an adjustable limit speed of the rotor
(54).
18. Method as claimed in claim 17, wherein a bridge switch pair (a and b, a and c, b and
c) is activated by way of a pulse width modulated control signal from the allocated
transistors, preferably with increasing time intervals.
19.Brushless DC motor (46) without position sensors for performance of the method
according to any of claims 1 to 20, wherein it comprises a stator (44) with three
stator coils (A, B, C) which extend over 3n stator poles (52) and an annular
permanent magnet rotor (54) with 2n permanent magnet segments of alternating
polarity (N, S), a coaxial cup-shaped external rotor (110) and a coaxial drive shaft
(108), where a and m are whole multiplication factors 1, 2, 3, 4 independent of
each other.
20. DC motor (46) as claimed in claim 19, wherein n = 1 and m = 1 or n = 3 and m = 6.

ABSTRACT

Title: METHOD FOR CONTROLLING A BRUSHLESS DC MOTOR
Method for controlling a brushless DC motor (46) comprising a stator (44) and rotor (54),
which has a highly reduced reduction gear of an actuating drive for a flap or valve in order
to regulate a gas or liquid volume flow, in particular in the field of heating, ventilation and
climastisation (HLK), fire and room protection, the electrical current flow in the stator coils
(A, B, C) being communicated under programme control without integral position sensors,
characterized in that the electronics count the revolutions of the rotor (54) of the DC motor
(46), and supplied to a counter with a logic rule, and in that the logic rule for start and stop
cause motor speeds to reduce below a threshold value to be passed quickly and without
the occurrence of counting errors.

Documents:

03541-kolnp-2006 correspondence-1.3.pdf

03541-kolnp-2006 form-18.pdf

03541-kolnp-2006-abstract-1.1.pdf

03541-kolnp-2006-abstract.pdf

03541-kolnp-2006-claims-1.1.pdf

03541-kolnp-2006-claims.pdf

03541-kolnp-2006-correspondence others-1.1.pdf

03541-kolnp-2006-correspondence others.pdf

03541-kolnp-2006-correspondence-1.2.pdf

03541-kolnp-2006-description (complete).pdf

03541-kolnp-2006-drawings.pdf

03541-kolnp-2006-form-1.pdf

03541-kolnp-2006-form-2.pdf

03541-kolnp-2006-form-3.pdf

03541-kolnp-2006-form-5.pdf

03541-kolnp-2006-international publicational.pdf

03541-kolnp-2006-other document.pdf

03541-kolnp-2006-pct others.pdf

03541-kolnp-2006-priority document.pdf

3541-KOLNP-2006-(28-06-2012)-CORRESPONDENCE.pdf

3541-KOLNP-2006-ABSTRACT.pdf

3541-KOLNP-2006-AMANDED CLAIMS.pdf

3541-KOLNP-2006-CANCELLED PAGES.pdf

3541-KOLNP-2006-CORRESPONDENCE.pdf

3541-KOLNP-2006-DESCRIPTION (COMPLETE).pdf

3541-KOLNP-2006-DRAWINGS.pdf

3541-kolnp-2006-EXAMINATION REPORT.pdf

3541-KOLNP-2006-FORM 1.pdf

3541-kolnp-2006-FORM 18.pdf

3541-KOLNP-2006-FORM 2.pdf

3541-kolnp-2006-FORM 26.pdf

3541-KOLNP-2006-FORM 3.pdf

3541-KOLNP-2006-FORM 5.pdf

3541-kolnp-2006-GRANTED-ABSTRACT.pdf

3541-kolnp-2006-GRANTED-CLAIMS.pdf

3541-kolnp-2006-GRANTED-DESCRIPTION (COMPLETE).pdf

3541-kolnp-2006-GRANTED-DRAWINGS.pdf

3541-kolnp-2006-GRANTED-FORM 1.pdf

3541-kolnp-2006-GRANTED-FORM 2.pdf

3541-kolnp-2006-GRANTED-FORM 3.pdf

3541-kolnp-2006-GRANTED-FORM 5.pdf

3541-kolnp-2006-GRANTED-SPECIFICATION-COMPLETE.pdf

3541-kolnp-2006-INTERNATIONAL PUBLICATION.pdf

3541-kolnp-2006-INTERNATIONAL SEARCH REPORT & OTHERS.pdf

3541-KOLNP-2006-PETITION UNDER RULE 137.pdf

3541-KOLNP-2006-REPLY TO EXAMINATION REPORT.pdf

3541-kolnp-2006-TRANSLATED COPY OF PRIORITY DOCUMENT.pdf

abstract-03541-kolnp-2006.jpg

« Previous Patent

Next Patent »

Patent Number

260049

Indian Patent Application Number

3541/KOLNP/2006

PG Journal Number

14/2014

Publication Date

04-Apr-2014

Grant Date

31-Mar-2014

Date of Filing

27-Nov-2006

Name of Patentee

BELIMO HOLDING AG

Applicant Address

Brunnenbachstrasse 1 CH-8340 Hinwil

Inventors:

#	Inventor's Name	Inventor's Address
1	GROGG Silvio	Chapfstrasse 9 CH-8625 Gossau
2	OCHSENBEIN Martin	Berggasse 9 CH-8332 Russikon
3	STAHELI Curdin	Leisihaidestrasse 35 H CH-8623 Wetzikon

PCT International Classification Number

H02P

PCT International Application Number

PCT/CH2005/000311

PCT International Filing date

2005-06-03

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1	942/04	2004-06-04	Switzerland
2	2122/04	2004-12-22	Switzerland