DE102006017146B4 - Method for determining the speed of an EC motor - Google Patents

Abstract

Method for determining the rotational speed of an EC motor having a primary part with at least one winding and a secondary part with magnet segments (4, 5) alternately offset in the circumferential direction and alternately magnetized in mutually opposite directions, wherein a plurality of sensor devices (3a, 3b, 3c) is determined from position measurement values which characterize the position of the magnet segments relative to the primary part, at least two of these position measurement values are combined at least indirectly to one common value and the common value for the individual position measurement values is representative, respectively those position measurement values are combined to form common values correspond with each other corresponding rotational positions of the secondary part relative to the primary part, and for each common value, a main correction factor is determined, which corresponds to a malposition of the secondary part relative to the primary part, thereby geke nnzeichnet that the summary of the individual position measurements is done by averaging.

Description

The present invention relates to a method for determining the rotational speed of an EC motor, wherein the EC motor has a primary part with at least one winding and a secondary part with circumferentially offset mutually magnetized magnet segments alternately magnetized in opposite directions.

Such EC (electronically commuted) motors are known from the prior art. These often have Hall sensors that are used for commutation, the signals of these Hall sensors are also evaluated to determine the position of the secondary part relative to the primary part, and thus by differentiating the time the speed. Ideally, the Hall sensors are arranged so that each increment, that is, every smallest measurable motor movement, corresponds to the same actual motor motion. This means that the position measurement signal of the Hall sensors changes their value according to a path or a given angle of rotation, since the magnetization of the magnet segment opposite the sensor changes.

How big this angle of rotation is depends on the design of the motor, that is, on the number of Hall sensors and the number of magnets of the secondary part. Ideally, each increment should correspond to the same motor motion, as stated above. In reality, however, there are different manufacturing inaccuracies, or the individual magnet segments have tolerances with regard to their positioning on the circumference of the secondary part and with respect to their dimensions in the circumferential direction. These tolerances are not known in detail, however, lead to inaccuracies in the synchronization of the engine, which can not be compensated by the subdivision of the path between two sensor signals in substeps - or at least not completely.

If the angle between two magnetic segment sensor combinations, such as a four pole pair EC motor and three Hall sensors, deviates by +/- 20% from a setpoint, and the EC motor speed is 3,000 r.p.m. the position measurement signal of the Hall sensors on average every 0.83 ms its value. The time interval between changes varies by +/- 20%, ie between 0.6 ms to 1 ms. Thus, the measured speed varies from 2,500 rpm to 3,750 rpm, that is, it may vary by about 25% from the actual value.

These rapid jumps of the measured speed with deviations of up to 25% from the actual speed disturb, for example, a speed control considerably, or allow only controls with low dynamics.

One possibility is to smooth the speed signal by, for example, the mean value is formed in each case from the penultimate, last and current speed measurement. In this case, however, the averaged speed measurement will not reflect the speed at the current time, but at the time of the last change of the attitude measurement signal.

In the earlier patent application DE 10 2005 019 515 A1 there is described a method of determining the speed at which a correction value for compensating the influence of at least one of the tolerances on the speed signal is detected and stored, and then the speed signal is corrected by the correction value.

More specifically, in this method, the duration of each individual interrupt or the time interval between the signals output from the Hall sensors is measured and later provided with a correction factor to correct the inaccuracies. Thus, the error is determined by measurement or otherwise to determine a correction value with which the speed signal can be corrected at a later time when the magnetic segment in question again passes the magnetic field sensor. This correction method makes it possible to correct a measurement inaccuracy caused by a tolerance of the magnet segment in a relatively simple manner with the speed signal.

In this case, when the secondary part is rotated relative to the primary part that a number of magnetic segment sensor combinations is traversed, detected by means of a measuring device for these magnetic segment sensor combinations, a first uncorrected speed signal. In addition, a reference signal for the speed of the EC motor is detected, which has a greater accuracy than the uncorrected speed signal and using the uncorrected speed signal and the reference signal determines a correction value as a correction factor.

When the magnetic segment-sensor combinations of the first uncorrected rotational speed signal are run through again and a second uncorrected rotational speed signal is detected with the aid of the measuring device, this second uncorrected rotational speed signal is corrected with the aid of the previously determined correction factors.

The correction values are thus determined in the form of correction factors, whereby a correction the measurement error caused by the tolerances of the magnet segment at different speeds is possible. The reference signal may be a measurement signal that is detected, for example, in the manufacture of the EC motor by means of an additional position measuring device. The reference signal may also be a speed and / or integrated spin signal of a shaft coupled to the EC motor.

However, this method is disadvantageous in that there is no absolute relationship between the individual increments and the actual mechanical position of the motor rotor. Thus, the method can only be used as long as the controller remains in operation and no increments (for example due to electromagnetic interference) are lost.

With regard to further prior art is on the DE 100 610 04 A1 , the US 2005/0001584 A1 and the DE 198 35 091 C1 directed.

The present invention is therefore based on the object to develop a method for measuring the speed, which allows a correction even after switching off the control device. This method can be used alongside or in addition to the correction method described above.

In the method according to the invention for determining the rotational speed of an EC motor with a primary part and a winding having a secondary part with magnetic segments magnetized alternately in the circumferential direction and alternately magnetized in directions opposite to one another, the position of the magnet segments relative to the primary part is detected and a multiplicity of position measurement values are determined which characterize the position of the magnet segments relative to the primary part. At least two of these values are combined at least indirectly into a common value, the common value being representative of the individual values.

An at least indirect summarization is understood to mean that the position measurement values are either combined themselves or the position measurement values are first corrected by means of the abovementioned correction factors and then summarized. Summing up the correction factors themselves is also understood as indirect summarization.

This means that if one assumes, for example, an EC motor with seven magnets and three Hall sensors, not the correction values for a total of 42 increments (3 × 14) are determined, but instead only correction factors for combined values, which respectively consist of groups of the consist of individual values.

Preferably, at least the position measurement values or the common values for determining the rotational speed are differentiated according to time. However, it is also possible to determine the speed both by differentiation of the position measured values after the time and by differentiation of the common values after the time.

In this case, the individual values are preferably combined by averaging, and, particularly preferably, the mean value is taken from a group of mean values which contain arithmetic mean values, geometric mean values and the like.

Preferably, a multiplicity of common values are formed, wherein, with particular preference, each of these common values is again obtained from a combination of a plurality of position measurement values. The position measured values are preferably combined in a defined sequence.

Preferably, in each case those values are combined into common values which correspond to one another in corresponding rotational positions of the secondary part relative to the primary part. More specifically, for example, for the motor shown above with seven pole pairs and three sensors, the 42 increments are divided into different groups, each of which would theoretically have to output the same sensor signals. In the event that the above-described motor with seven magnets and three Hall sensors is used, the theoretically identical positional readings of all the sensors would repeat each after six increments, and are therefore summarized, as will be explained in more detail in the following part.

In other words, only one main correction factor per Hall pattern or per sensor value is determined, this factor, as stated, the mean value of the individual correction factors of the mutually corresponding rotational positions. Thus, this common main correction factor characterizes the malposition of the secondary part relative to the primary part.

This average correction factor determined by averaging is stored in a non-volatile memory and thus remains available even after the control unit has been switched off or after a possible loss of increment.

The method according to the invention can be used instead of the method described above or in addition to the procedure described above. Thus, a correction factor can additionally be determined for each individual value, which corresponds to a malposition of the secondary part relative to the primary part. The correction factor results from the fact that the actual position of the secondary part with respect to the primary part does not correspond exactly to the theoretical position. In this embodiment, the main correction factor is preferably formed by averaging the individual correction factors.

This means that first the individual correction factors described above are determined, and then, by averaging out these individual correction factors, the main correction factor. These respective main correction factors are preferably stored in a non-volatile memory and are thus immediately available after switching off and restarting the motor.

In a preferred embodiment of the invention, a moving average is formed as the mean value of the individual correction factors, preferably in such a way that the weight with which the correction factors enter the mean value decreases with increasing age of the correction factors. New correction factors are thus taken more into account in the mean value as correction factors which are assigned to a further ago date. Should an error occur which leads to a situation where a magnetic segment-sensor combination is not detected and the correction factors already determined are assigned to the wrong magnet segments, the incorrect correction factor assignment has only a short-term effect on the correction of the rotational speed signal, ie the wrong one Correction factors are "forgotten" relatively quickly.

As stated earlier, these sensor-related main correction factors are available through storage in non-volatile memory and can be used immediately upon each reboot. This availability is also not affected by lost increments or other disturbances. The calculation or measurement of the correction factors can also be carried out as part of an end-of-line procedure, thus saving runtime for the end application. Thus, the inventive method can also be used in applications in which there is no possibility to perform the calculation of the correction factors in normal operation, such as in clutch motors.

Arithmetic averaging, geometric averaging or comparable averaging is preferably also carried out for the purpose of averaging the individual correction factors.

Preferably, the main correction factor or the main correction factors are stored in a memory device, in particular, but not exclusively, the non-volatile memory shown above. This storage in a non-volatile memory has - as mentioned above - the advantage that after each restart of the engine immediately the necessary correction values are available. Particular preference is given to using magnetic sensor devices as sensor devices and, particularly preferably, Hall sensors. These Hall sensors allow the determination of correction factors and, as mentioned above, the grouping of groups, ie z. B. Hall pattern-related main correction factors.

Instead of Hall sensors and magnetoresistive sensor devices can be used.

The sensor devices are preferably distributed uniformly around the circumference of the secondary part. In this way, the individual increments can each be compared with theoretically equal increments, which overall facilitates the calculation of the individual correction factors.

In a preferred embodiment, three sensor devices are used. However, it is also possible to use a different number of sensor devices, but with increasing number of sensor devices and the arithmetic effort for the formation of the correction factors increases.

The invention is further directed to an EC motor having a primary part with at least one winding and a secondary part with circumferentially offset, magnetically magnetized magnet segments alternately in opposite directions, wherein a plurality of sensor means is provided which outputs positional measurements with the local Position of the magnetic segments correspond to these sensor devices, and wherein a processor device is provided, which combines each of these position measurement values to a common value, wherein the common value for the individual values is representative. In this case, this EC motor preferably has a non-volatile memory for storing the common values.

The processor device preferably determines the rotational speed of the engine from the position measurement values and / or the common values by differentiation according to time.

The invention further relates to the use of the method according to the invention or of the EC motor according to the invention in a motor vehicle directed. The invention is thus applicable to any electronic motors and drive systems that require an accurate speed or position signal.

The motors according to the invention can be accommodated in different elements of a motor vehicle, such as in different types of transmissions, transmission components, in transmission actuators or clutch actuators, in actuator motors, in camshaft controls or the like.

As mentioned above, the engines have production-related inaccuracies of the individual increments. A first source of inaccuracies is inaccurate placement of the individual Hall sensors, or sensor devices. As a result of this inaccurate placement, the distance and thus also the angular distance between the Hall sensors is changed, which leads to the individual increment angles no longer being the same but changing with respect to one another. However, these inaccuracies are repeated periodically, more precisely with each electrical period, and thus are Hall-pattern dependent. Therefore, with the method according to the invention, the inaccuracies resulting from inaccurate placement of the Hall sensors can be determined and corrected.

Another inaccuracy results from a non-ideal or inaccurate response in one or more Hall sensors, that is, the Hall sensor does not respond to the exact transition between the magnetizations, but this offset. This inaccuracy also changes the increment angle. This inaccuracy is repeated with each electrical period and is thus also Hall pattern-dependent and can be determined or corrected by the method according to the invention.

Another source of error is inaccurate magnetization of the individual magnet segments. As a result, the angle occupied by the individual magnetic poles changes, and thus also the increment angle. This inaccuracy does not repeat itself with each electrical period, but with each mechanical period, and thus does not depend on the Hall patterns described above.

Further advantageous embodiments will become apparent from the accompanying drawings; show in it:

1 a schematic representation of a motor according to the invention with sensor devices;

2 a table comparing the position measurements and the summarized values; and

3 an exemplary table showing the Hall pattern-related correction factors.

1 schematically shows an EC motor, more precisely a so-called BLDC motor (brushless-direct-current) with seven magnets 4 . 5 as well as three Hall sensors 3a . 3b and 3c , The seven magnets result in a total of 14 magnet segments, or more precisely, seven north pole magnet segments 4 and seven South Pole magnet segments 5 , which are each arranged alternately to each other. Overall, the in 1 shown EC motor 42 increments. The arrow P indicates the direction of rotation of the motor.

In the situation shown in part I, the Hall sensor is present 3a with a rotation in the direction of the arrow, ie in the clockwise direction a signal corresponding to a positive pole, the Hall sensor 3b a signal corresponding to a south pole and the Hall sensor 3c a signal corresponding to a north pole. Each signal remains with a rotation of the magnets, until the situation shown in part II is reached. In the situation shown in Fig. II, the Hall sensor gives 3a still a signal corresponding to a north pole and the Hall sensor 3b a signal corresponding to a south pole. The Hall sensor 3c however, changes its signal and outputs a signal corresponding to a south pole upon further rotation.

In the situation shown in part III, the signals of the Hall sensors remain 3a and 3c constant, while that of the Hall sensor 3b output signal changes. Thus, a total of the motor positions shown in the subfigures I to III show three different Hall patterns, wherein the Hall pattern composed of the position measurement signals of all Hall sensors. In the situation shown in Part IV, this changes from the Hall sensor 3a output signal from a south pole according to a north pole accordingly, while the signals from the Hall sensors 3b and 3c stay constant; in the situation shown in part V, the signal of the Hall sensor changes 3c , the signals of the Hall sensors 3a and 3b stay constant. In the situation indicated by VI, the signal of the Hall sensor changes 3b while the signals of the Hall sensors 3a and 3c stay constant.

In the situation indicated by VII, the signal of the sensor changes 3a according to a south pole according to a north pole, while the signals of the sensors 3b and 3c stay constant. Thus, the situation shown in Part VII leads to the same Hall pattern as the situation shown in Part I. Equally true the Hall pattern of the situation shown in part VIII and the situation shown in part II and the Hall pattern of the situation shown in part IX and the situation shown in part III.

The next (not shown) rotational position would again give the same Hall pattern as the situation shown in Part IV. From this drawing it can be seen that for a complete motor rotation a total of six different Hall patterns occur, which in 1 are shown in situations I to VI.

As mentioned above, the individual Hall sensors output position measurement signals. To form a speed signal, these position measurement signals are differentiated. This can be done, for example, such that the time interval .DELTA.t measured between two changes in the position measurement signal and the rotational speed .omega. Is determined as follows: ω = π / (m × p × Δt) [rad / s]

In this case, p denotes the number of pole pairs and is seven in the present case, m denotes the number of Hall sensors and is therefore three here.

If the individual magnet segments were applied absolutely exactly, the rotational speed determined in this way or the rotational speed signal determined in this way would not be faulty and would lead to a smooth signal at a constant, actual rotational speed of the secondary part.

Due to the tolerances of the magnetic segments 1 to 14, however, the determined speed signal ω _{Mess, i is subject} to errors, which lead, for example, at a constant actual speed of the secondary part to the fact that the speed signal has jumps.

In order to determine the individual correction factors, the magnetic segment-sensor combinations of 1 to 2 · m · p, that is to say from 1 to 42, are numbered consecutively in an evaluation device, so that in the case of clockwise rotation the count, hereinafter referred to as "index i", is abbreviated. is counted up, and is reset to 1 on reaching 2 · m · p. When the EC motor is switched on, the index i is set to a starting value, for example to the value 1.

For each magnetic segment-sensor combination, a correction factor F _Adap [i] is determined which is assigned to the corresponding magnet segment 1 to 14 via the index i. This correction factor corresponds to the ratio between the rotational speed value ω _{Meas, i} , which was determined with the aid of the position measurement signal for the ith magnet segment sensor combinations, and a reference rotational speed value ω _{Ref, i} which is assumed to have greater accuracy, as the speed value ω _{measurement, i} . The correction _factors F _Adap [i] are now stored in a data memory of the evaluation device.

With the aid of the correction factor F _Adap [i], a corrected rotational speed value ω _{Korr, i is} determined in each case for each rotational speed value ω _{Mess, i} as follows:

As can be seen from the description, it is thus necessary, with the previously patent pending method, to continue to use or store the individual, ie the 42 correction factors. In the method according to the invention, those correction factors which result from mutually corresponding rotational positions, as in FIG 1 rotary positions I and VII, II and VIII and III and IX are combined to form an average value. These averages are stored in a non-volatile memory and are therefore available for each restart. This saving prevents inaccuracies resulting from the fact that, for example, when turning the motor, individual increments are not registered by the Hall sensors, and are thus lost.

Instead of the previously 42 individual values, only six different values, that is to say the respective mean values for the Hall patterns, have to be stored in the procedure according to the invention.

2 shows a table as an example of the recording of the individual Hall patterns that correspond to the respective increments.

The individual increments 1 to 42 are numbered in the left-hand column. These individual increments are assigned to six different Hall patterns, where - as based on 1 The Hall pattern of Increment 1 corresponds to the Hall pattern of Increment 7, the Hall pattern of Increment 13, Increment 19, Increment 25, Increment 31 and Increment 37. Likewise, the remaining increments in the order 2, 6, 4, 5, 1, 3 fixed here are assigned to the remaining Hall patterns.

In the third column, the respective deviation of the individual increments is plotted, wherein the deviation, as shown above, is determined by differentiation and comparison with reference values. The fourth column gives the factors used for each Hall pattern, with the factors used being identical for the same pattern. The factors used are determined by averaging over the measured factors for the individual Hall patterns plotted in the third column. Thus, for example, the used factor used for the increments 1, 7, 13, 19, 25, 31 and 37 results from averaging the deviation values corresponding to these increments.

In the fifth column, the respective error is indicated, that is, the error by which the used factor deviates from the actual match. As can be seen from the fifth column, this error in this example is at most the amount after 10%.

In the right-hand column, the errors which would result without correction, that is, the respective errors of the deviations plotted in the third column, are ideally plotted against the value which is 1 in each case. It can be seen that the errors without the use of a correction are on average significantly higher than the errors that can be achieved with the correction method according to the invention. Without correction, the errors amount to up to 27%.

3 shows a table in which the Hall patterns 1 to 6 are plotted along with the correction factors. This is what the table looks like 3 around a section of the table in 2 ,

The possible uses for the method according to the invention are versatile. On the one hand, the method according to the invention can be used to allow some correction if the method of the mentioned earlier patent application, in which a correction factor is formed for each increment, fails. In addition, the method according to the invention can be used to provide a certain correction possibility for the EC motor directly after starting. That is, in this case, the method according to the invention is used in addition to the method known from the prior art.

In addition, the method according to the invention is preferably suitable for use in the event of errors or disturbances in the control device. For example, if the controller fails, the Hall pattern correction factors stored in non-volatile memory may be resorted to. In this case it is possible to limit the maximum errors.

LIST OF REFERENCE NUMBERS

3a, 3b, 3c

Hall sensor

4

North Pole magnet segment

5

South Pole magnet segment

Claims (15)

Method for determining the rotational speed of an EC motor having a primary part with at least one winding and a secondary part with magnet segments offset in the circumferential direction and magnetized alternately in opposite directions (US Pat. 4 . 5 ), whereby by means of sensor devices ( 3a . 3b . 3c ) a plurality of position measurement values is determined which characterize the position of the magnet segments relative to the primary part, at least two of these position measurement values are combined at least indirectly to one common value and the common value for the individual position measurement values is representative, respectively those position measurement values are combined into common values be corresponding to corresponding rotational positions of the secondary part relative to the primary part, and for each common value, a main correction factor is determined, which corresponds to a malposition of the secondary part relative to the primary part, characterized in that the summary of the individual position measurement values by averaging occurs. A method according to claim 1, characterized in that at least the position measurement values or the common values for determining the rotational speed are differentiated according to the time. Method according to one of the preceding claims, characterized in that the mean value is taken from a group of mean values which contain arithmetic mean values or geometric mean values. Method according to one of the preceding claims, characterized in that the position measured values are combined in a fixed sequence. Method according to one of the preceding claims, characterized in that a single correction factor is determined for each individual position measurement value, which corresponds to a malposition of the secondary part relative to the primary part. A method according to claim 5, characterized in that the main correction factor is determined by averaging the individual correction factors. A method according to claim 6, characterized in that an arithmetic averaging or a geometric averaging is performed to average the individual correction factors. Method according to one of the preceding claims, characterized in that at least one main correction factor is stored in a memory device. Method according to one of the preceding claims, characterized in that at least one main correction factor is stored in a non-volatile memory device. Method according to one of the preceding claims, characterized in that as sensor devices ( 3a . 3b . 3c ) magnetic sensor devices are used. Method according to one of the preceding claims, characterized in that the sensor devices ( 3a . 3b . 3c ) are distributed uniformly around the circumference of the secondary part. Method according to one of the preceding claims, characterized in that three sensor devices ( 3a . 3b . 3c ) be used. EC motor having a primary part with at least one winding and a secondary part with magnet segments offset in the circumferential direction and magnetized alternately in opposite directions (US Pat. 4 . 5 ), wherein a multiplicity of sensor devices ( 3a . 3b . 3c ), which output position measurement values which correspond to the local position of the magnet segments ( 4 . 5 ) to these sensor devices ( 3a . 3b . 3c ), and a processor device is provided which combines each of these position measurement values into a common value by averaging, wherein the common value for the individual position measurement values is representative. EC motor according to claim 13, characterized in that the processor device determines the rotational speed from the position measurement values and / or the common values by differentiation according to time. Use of a method according to one of claims 1 to 12 for controlling EC motors in a motor vehicle.

Images