CN111448381A

CN111448381A - Method and device for determining the rotational speed of a crankshaft of an internal combustion engine

Info

Publication number: CN111448381A
Application number: CN201880080241.8A
Authority: CN
Inventors: V.K.马蒂瓦拉维拉巴萨帕; B.雷内克; J.穆勒; W.费舍尔; A.赖歇; S.葛罗德
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2017-12-15
Filing date: 2018-11-29
Publication date: 2020-07-24
Anticipated expiration: 2038-11-29
Also published as: EP3724477A1; DE102018200639A1; CN111448381B; WO2019115242A1

Abstract

The invention relates to a method for determining a rotational variable, in particular a rotational speed, of a shaft (17), in particular a crankshaft (17') of an internal combustion engine (112)Speed (n), angle of rotation position (c)

，

) Or direction of rotation (α)₊，α_‑) The internal combustion engine is directly or drivingly coupled to an electric machine (30) comprising a rotor (32) and a stator (33) having at least two phase windings (U, V, W) from which at least one phase signal (U) is derived in each case_U、U_V、U_W、I_U、I_V、I_W) Wherein the phase signal (U)_U、U_V、U_W、I_U、I_V、I_W) Respectively have a rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) And/or zero crossing (Fl)_U0、Fl_V0、Fl_W0) Wherein a plurality of phase signals (U) comprising the motor (30) are generated_U、U_V、U_W、I_U、I_V、I_W) Sum signal (U)_Sum) So that at said sum signal (U)_Sum) Respectively giving said rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) And/or zero crossing (Fl)_U0、Fl_V0、Fl_W0) Assigning a characteristic pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) Wherein said pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) Is used to determine the rotational speed (n) and the rotational angle position (n, b) of the rotor (32)

，

) Or direction of rotation (α)₊，α_‑). The invention also relates to a method for generating a plurality of phase signals (U)_U、U_V、U_W、I_U、I_V、I_W) Generating a sum signal (U)_Sum) The summed signal is used to determine the rotational speed (n) of the shaft (17); a corresponding computing unit is set up for carrying out the method; and a computer program for performing the method.

Description

Method and device for determining the rotational speed of a crankshaft of an internal combustion engine

Technical Field

The invention relates to a method and a device for determining a rotational variable, in particular a rotational speed, a rotational angle position or a rotational direction, of a shaft, in particular a crankshaft of an internal combustion engine, which is directly or drivingly coupled to an electric machine, which comprises a rotor and a stator having at least two phase windings, from which at least one phase signal is derived in each case.

Background

The rotational speed of the crankshaft of an internal combustion engine is an important input variable for various functions of the electronic engine control system. In order to determine the rotational speed, markings may be provided on the body that rotates together with the crankshaft of the internal combustion engine at the same angular distance. The marking due to the rotation of the crankshaft can be detected by the sensor and forwarded as an electrical signal to the evaluation electronics.

The electronic device determines a signal registered for the marker for this purpose or measures the time difference between the two markers for the respective rotational angle position of the crankshaft, and can determine the angular velocity and the rotational speed on the basis of the known angular distance of the two markers from one another. In the case of motor vehicles, in particular motorcycles, motorbikes or motorbikes, these markings can be provided, for example, by the teeth of metal gears of a so-called sensor wheel, preferably of metal gears made of ferromagnetic material, which, due to their movement, cause a change in the magnetic field in the sensor. The absence of several teeth can be used as a reference mark for identifying the absolute position. In the case of passenger vehicles (Pkw), 60-2 teeth (60 teeth evenly distributed, with 2 teeth being left empty as reference symbols) are mostly used, while in the case of motorcycles or motorbikes, for example, 36-2, 24-2 teeth are also used. In the case of this indirect principle of the determination of the rotational speed of the crankshaft or of the rotational angle position of the crankshaft, the resolution of the rotational speed signal or the absolute detection of the rotational angle position is determined by the number of teeth and by a reliable detection of the reference markings.

In the case of any modern vehicle with an internal combustion engine, an electric generator is built, which is driven by the rotation of the crankshaft. The generator provides an electrical signal and is used to supply the vehicle with electrical energy and to charge the vehicle battery pack. The intended operation of the vehicle without the generator is not possible or only possible for a short time.

The use of the electrical output variable of an electric machine (generator) driven by the crankshaft is used for rotational speed determination, for example, in EP 0664887B 1. For this purpose, the phase of the generator to which the pulsed dc voltage is attached is provided as a reference. In the case of a rotational speed determination using a plurality of phases, the respective signals of the phases are usually detected and separately forwarded to an evaluation unit, which determines the rotational speed of the generator from the signals.

Thus, it would be desirable to illustrate one approach: the determination of the rotational variable, in particular the rotational speed, is further simplified depending on the phase signal of the electric machine and the rotational speed of at least the rotor of the electric machine or the crankshaft of the internal combustion engine, which can be used for controlling the internal combustion engine, is obtained more simply and, if appropriate, with still better accuracy, even without the use of additional components.

Disclosure of Invention

According to the invention, a method having the features of claim 1 and an apparatus having the features of claim 10 are proposed. Advantageous embodiments are the subject matter of the dependent claims and the subsequent description.

THE ADVANTAGES OF THE PRESENT INVENTION

The invention relates to a method for determining a rotational variable of a shaft, in particular of a crankshaft of an internal combustion engine, which is directly or drivingly coupled to an electric machine, which comprises a rotor and a stator having at least two phase windings, from which at least one phase signal is derived in each case, wherein the phase signals each have a rising edge, a falling edge and/or a zero crossing, wherein a summation signal comprising a plurality of phase signals of the electric machine is generated such that in the summation signal, in each case characteristic pulses are assigned to the rising edges, falling edges and/or zero crossings, wherein the variable is used to determine the rotational variable of the rotor of the electric machine. In this case, the rotational variables include, in particular, the rotational speed, the rotational angle position or the rotational direction of a shaft, preferably a rotor shaft, wherein the rotor shaft is fixedly or drivingly coupled or coupleable to a crankshaft of the internal combustion engine. Thus, knowing the respective rotational variable of the rotor shaft, the corresponding rotational variable of the crankshaft of the internal combustion engine can be directly inferred.

Within the scope of the method, the sum signal is generated from phase signals of at least two phases of the electric machine. The generation of the summation signal for determining the rotational variable, in particular the rotational speed, is particularly advantageous, since the summation signal can be forwarded to an evaluation unit, in particular to a control device, by means of a single signal line, in order to derive therefrom, depending on the summation signal, the respective rotational information of the generator phase on which the rotational variable is based, from the signal of the respective generator phase, whereby the rotational speed, the rotational angle position and/or the rotational direction of the shaft can be determined not only particularly simply and on the basis of a particularly simple signal line infrastructure, but also with the usual accuracy. This is caused in particular by: in the sum signal, characteristic pulses are assigned to the rising edges, falling edges and/or zero crossings of the individual signals in the respective phase signal, wherein the respective rotational variable can be derived from the signals of the individual pulses in the sum signal.

This method is particularly advantageous since in the prior art, the signals of the respective phase are usually each forwarded to an evaluation unit by means of a signal line assigned to the respective phase, in order to be evaluated there. This is correspondingly costly and cost intensive.

Preferred within the scope of the invention are: at least one value of at least one of the pulses associated with a rising edge of the phase signal, with a falling edge of the phase signal and/or with a zero crossing of the phase signal is used for transmitting the corresponding pulse in the sum signal, whereby the absolute position over time of these rising edges, falling edges and/or zero crossings and their relative spacing from one another can be determined and from this in particular the rotational speed, rotational angle position and/or rotational direction of the shaft can be determined. This embodiment is particularly advantageous because not only the rising edges, falling edges and zero crossings of the phase signals, which are correspondingly converted by the coding described above into pulses within the sum signal, can be detected particularly simply and precisely, which makes the determination of the aforementioned rotational variable of the shaft correspondingly reliable and precise. The phase signal may be the original signal of the stator phase. However, correspondingly electronically processed phase signals may also be used.

In a further preferred embodiment, in the sum signal, the first pulse is generated by coding the falling edge of the respective phase signal and the further pulses are generated by coding the rising edge of the respective phase signal. By a correspondingly different selection of the pulses, at least one criterion is given in order to enable rising and falling edges to be correspondingly distinguished in the sum signal in the form in which they are converted into pulses. It should be noted again that: the position over time of the respective edges and their type, i.e. rising and/or falling edges, are correspondingly converted into pulses in the sum signal and can be distinguished.

In a further preferred embodiment of the method, the pulses assigned to the first edge type having a rising edge or falling edge have a constant first pulse width, while the pulses assigned to the respectively other edge type have a further pulse width which is constant for the respective phase signal but differs with respect to the respectively other edge type and/or the respectively other phase. This embodiment is advantageous because in this way, edges of this type can be detected in the sum signal in a correspondingly simple manner by selecting a constant pulse width for one of the edge types. By this, the number of different pulse widths can be reduced without information loss. This assignment can be simplified as a function of the design of the evaluation algorithm for the sum signal for assigning the individual pulses to their respective edge type and phase signal. Correspondingly, the characteristic pulse width can be selected for the respective edge type or edge type and/or the respective phase, whereby the respective edge type and the phase associated therewith can be unambiguously assigned.

This is provided within the scope of a preferred embodiment, in particular, by the following: a first edge type is asserted by determining that there is a first pulse width in the summed signal and a corresponding other edge type is asserted by determining that there is one of the other pulse widths in the summed signal.

Within the scope of a further preferred embodiment of the method, the rotational speed of the rotor of the electric machine is determined from the sum signal by determining at least one time difference between two pulses corresponding to the same edge type (rising or falling edge). Preferably, the same type of edges is used, which can be attributed to one of the phases. As already mentioned, a particularly simple way of determining the rotational speed of the electric machine as a function of the relative spacing of the pulses from one another exists if the same pulse duration is selected for one of the edge types independently of one another. In a further preferred embodiment, the rotational speed is determined on the basis of two adjacent pulses, preferably immediately adjacent pulses.

In a further preferred embodiment of the method, the first direction of rotation is deduced by determining at least one first time sequence of pulses or a further direction of rotation of the rotor, which is different from the first direction of rotation, is deduced by determining at least one further time sequence of pulses. The corresponding edges within the phase signal are correspondingly determined over time depending on the specified motor parameters, the layout of the pole pairs as within the motor, and other variables. By means of an analysis of the sequence of corresponding pulses in their chronological order, it is therefore possible to deduce the direction of rotation of the rotor of the electric machine. For this purpose, the edges from the respective phase signals and the edge types (rising or falling) of these edges are each encoded in the sum signal with different pulse widths in order to be able to distinguish these edges from these edge types. Thus, from the corresponding sequence of rising and falling edges from the respective phase signals, the direction of rotation of the rotor within the motor can be deduced based on the summed signal.

In a further preferred embodiment of the method, the respective pulse is determined by determining a pulse width, and the rotational angle position of the rotor of the electric machine is determined from the respective pulse using the characteristics of the electric machine. Thus, by determining the pulse width, falling and rising edges can be unambiguously deduced from the corresponding phase signals, as has already been described above. Since, as has also been described, the individual edges are associated with motor parameters, such as the spatial arrangement of the pole pairs within the motor, by determining the respective edges and correlating these motor parameters, at least the corresponding angular increment within which the rotor moves linearly at the time of detection of the respective edge within the sum signal can be deduced. From this, the rotational angle position of the rotor can be correspondingly deduced.

In a further preferred embodiment of the method, the rotational angle position of the rotor is further determined using the rotational speed value between two adjacent pulses and at least one time interval between the following position and at least one of these pulses. The aforementioned position is an arbitrary position over time within the sum signal for which the corresponding rotational angle position of the rotor is to be determined. This can be achieved in particular when the electric machine is running, but also when the internal combustion engine driving the electric machine is deactivated. The latter case is particularly suitable for determining the rest position of the rotor of the electrical machine. As already described above, by using the pulses associated with the respective edges of the phases and further using the motor parameters, an angular increment can be deduced within which the rotor lies linearly at the time of positioning. By further using the rotational speed values, which describe the change over time of the angular increment, and the corresponding time intervals of the positions and the corresponding pulses, it is possible to determine the positions within the angular increment even more precisely.

The invention also relates to a device for generating a summation signal from a plurality of phase signals, which summation signal can be used to determine a rotational variable, in particular a rotational speed, a rotational angular position and/or a rotational direction, of a shaft, in particular of a crankshaft of an internal combustion engine, which is coupled directly or in a driven manner to an electric machine, which comprises a rotor and a stator having at least two phase windings, from which at least one phase signal can be derived in each case, which phase signal has a rising edge, a falling edge and/or a zero crossing. In this case, the device has at least two inputs via which the phase signals are respectively conducted to a coding unit, wherein the coding unit combines the individual phase signals into a summation signal and codes at least one of the phase signals in such a way that a characteristic pulse is respectively assigned to a rising edge and/or a falling edge of the at least one phase signal in the summation signal, wherein the coding unit outputs the summation signal to an output for determining the rotation variable. In a similar manner to the method according to the invention, the determination of the rotational speed, the rotational angle position and/or the rotational direction of the rotor from the summation signal can be carried out particularly simply, since only one data line is required for forwarding the summation signal to the evaluation device. The characterizing variables from the phase signals required for determining the respective rotational variables, i.e. the respective edges of the respective phase signals, are likewise converted into corresponding pulses within the sum signal as already described above, wherein an analysis of these pulses from the sum signal is possible without loss of information.

By means of the coding unit, therefore, the edges within the respective phase signal are coded in the sum signal in such a way that they are converted into pulses which can be directly assigned to them, whereby a particularly precise determination of the rotation variable can be ensured with a significant simplification of the analysis infrastructure.

In a further preferred embodiment of the device, the coding unit has at least one time-lag element which can be assigned to at least one phase of the electric machine and which applies a characteristic property to at least one phase signal. The time-lag element is configured to: in particular, a combination of a plurality of exclusive-or gates (Xor) each loads at least one phase signal with a characteristic feature. This is caused in particular by: by means of a corresponding configuration of the time-lag elements, the pulses respectively assigned to the corresponding rising or falling edge of the phase signal are assigned a correspondingly characteristic pulse duration. By using different time-lag elements for different phase signals, the respective phase signal can be loaded with characteristic properties depending on the respective used time-lag element, so that, as already described above, the individual edge classes from the respective phase signal can be loaded with in principle different properties, in particular in the form of different pulse durations.

It is also easy to understand that: it is also possible to assign the same pulse duration in part to one of the two edge types (rising or falling edge) from the characteristic features in the phase signals, in particular the respective phase signals, and to assign different pulse durations only to the edges of the respectively other edge type of each phase signal in order to make the two edge types distinguishable. Within the at least one time-lag element, the selection of the pulse duration of the respective pulse is preferably brought about by means of at least one capacitor and/or at least one resistor. In this case, the respective time constant is given by a corresponding selection of the capacity of the resistor or capacitor. The signals from the respective coding units assigned to these phases are then correlated with one another, preferably by way of and (Und) gates, and a sum signal is formed.

In a further preferred embodiment, a trigger circuit is assigned to the coding unit, which trigger circuit generates a triggered phase signal, which is supplied to at least one of the inputs of the coding unit. The trigger circuit causes: the edges of the phase signal in the triggered phase signal are significantly more pronounced than in the input signal, which makes the identification of the edges and the corresponding exact temporal conversion of these edges into pulses in the sum signal particularly simple to implement. Thus, with a correspondingly higher time resolution of the edge orientation, the rotational variables, in particular the rotational speed, the rotational angle position and/or the rotational direction of the rotor, can also be determined more precisely from the sum signal.

In a preferred embodiment, the trigger circuit has a time-lag element with a resistor and a capacitor. Through a corresponding selection of the resistor and the capacitor, a forwarding of the triggered phase signal in the vicinity of the zero crossing in the time range of the phase signal can be correspondingly suppressed. This embodiment is particularly advantageous because disturbances, which can be separated or suppressed particularly effectively by such a circuit, can occur in the signal due to unbalanced states, in particular in the vicinity of the zero crossings of the phase voltages. In this case, the respective time constant is given by a corresponding selection of the capacity of the resistor or capacitor. In principle, it is easy to understand that: the exclusive-or (Xor) gate described previously can also be replaced by a corresponding other electronic circuit with similar functionality.

In a further preferred embodiment of the method, the rotational speed signal is determined as a function of the time profile of the rotational speed in a first revolution of the shaft, wherein the higher-order frequency components are determined within the framework of a frequency analysis, wherein the corrected rotational speed is determined for further revolutions of the shaft, such that at least one of the higher-order frequency components, preferably the 6 th-order component and/or the 12 th-order component and/or the 18 th-order component, is suppressed in the corrected rotational speed signal.

Preferably, the correction is made for a step directly dependent on the periodicity of the motor. These steps can be determined in the following way (the values illustrated in brackets are obtained for a three-phase machine with six pole pairs for example):

let N be the number of phases of the motor and p be the number of pole pairs of the motor. Each revolution of the motor shaft thus results in N times p electrical cycles having a zero crossing with a rising edge and a falling edge, respectively. Therefore, the total number a of sampling points of the corresponding rotation speed signal is:

the third order O is derived from the (phase-independent) possible different behavior of the rising and falling edges₃. The corresponding rising edge F of each revolution of the motor shaft_sOr falling edge F_fThe number of (A) is:

second order O is derived from the possible different behavior of the N different phases₂。

In this case, the respective edge F of each revolution of the motor shaft_pThe number of (A) is:

the first order O is derived from the possible different behavior of the rising and falling edges of the phases₁。

In this case, the respective rising edge F of each revolution of the motor shaft_spOr falling edge F_fPThe number of (A) is:

for one revolution of the motor shaft (fundamental frequency) there are obtained orders which are mainly related to the motor and which must be compensated, if necessary, according to the number of different edges mentioned above:

corresponding to the number of pole pairs, order O₁(= 6) represents the deviation between the rising and falling edges of the individual phases

Corresponds to the double pole-pair number, order O₂(= 12) deviation between edges representing different phases

Corresponds to the multiplication of the number of polar pairs times the number of edges, order O₃(= 18) represents the deviation between (phase independent) rising and falling edges.

In principle, it is readily understood that other frequency components than the fundamental frequency can also be considered. It is furthermore easy to understand that the rotation speed signal can be transformed from the time domain into the frequency domain, in particular in order to perform a frequency analysis. After the speed signal has been cleared of at least one higher-order magnitude, the speed signal can also be transformed from the frequency domain back into the time domain again in order to obtain a corrected time course of the speed. These transformations can be carried out in particular by means of FFT methods (Fast fourier transformation) or the like.

The rotational speed signal present as the fundamental rotational speed frequency is typically superimposed by interference influences caused by structural features of the electric machine and is usually an integer multiple of the fundamental frequency. In the case of a frequency analysis of the rotational speed signal, higher-order frequency components can be detected in a first revolution and these characteristic rotational speed components can be extracted, so that they can be suppressed or sorted out in the case of further revolutions of the shaft, in order to separate the higher-order frequency components from the signal, thus achieving a significantly better signal quality. Other interference frequency components which do not exactly meet the higher harmonics of the fundamental frequency can also be taken into account within the scope of the measures described above, so that higher-order frequencies are generally understood to be frequencies which differ from the fundamental frequency and correspond to the rotational speed magnitude.

In the case of an electric machine with three phases, the respective order of the frequency components, in particular the

order

6, 12 and 18, which are harmonics of the fundamental frequency, can be correlated with a detectable error source of the electric machine and can be suppressed in a targeted manner accordingly, which leads to an improvement in the rotational speed signal. The fundamental frequency generally relates to the rotational frequency of the crankshaft of the internal combustion engine and/or the rotational frequency of the rotor of the electric machine.

In a further preferred embodiment of the method, all higher-order frequency components are suppressed or filtered out, in particular by means of a low-pass filter, or partial regions of higher-order frequency components are suppressed or filtered out, in particular by means of a band-pass filter. In order to achieve the desired aim of suppressing higher-order frequency components, in particular higher harmonic frequencies, the use of a low-pass filter or band-pass filter which passes the fundamental rotational speed frequency but suppresses higher-order components is a particularly simple design.

In a further preferred embodiment of the method, a number of temporally successive rotational speed values are selected from the time profile of the rotational speed over a time range, wherein a mean value is calculated from at least one first subset of the rotational speed values and at least one further subset of the rotational speed values and a rotational speed trend is determined therefrom, wherein a correction factor is calculated by comparing at least one rotational speed value, preferably the respective number of rotational speed values, with a rotational speed trend value at least corresponding to the respective time instant, with the aid of which a further corrected rotational speed is determined. The rotational speed signal can be determined more precisely by the above-described measures, wherein higher-order disturbance influences can be averaged in particular by averaging the corresponding average values. In this case, it is particularly preferred if the number of rotational speed values used for averaging is the number of phases of the electric machine

Integer multiples of. This is particularly advantageous in providing sufficient sampling points for averaging. It is readily understood that the combination of these method features with the feature of suppressing higher order frequency components results in further improvements to the slew signal.

In addition, it is preferred to execute, within a speed profile having a speed profile which is as linear as possible, preferably substantially constant, the time ranges within which the speed trend is determined, i.e. the speed profile which is expected on the basis of the determined speed values. This is particularly advantageous because, in these time ranges, the dynamics during the speed change are expected to be low, so that the change in the speed signal is correspondingly small. Thus, a high-quality speed trend can be determined, since the number of rotational speed values which differ significantly from the mean value is small. The time ranges in which the internal combustion engine is in the operating state of the gas exchange stroke are provided.

In a further preferred embodiment of the method, a rotational variable, in particular a rotational speed, a rotational angle position or a rotational direction of the shaft is used for controlling the internal combustion engine, in particular for controlling the ignition and/or injection of fuel into at least one cylinder of the internal combustion engine. The detection and processing of the phase signals of the electric machine, in particular by the engine control device, can be used in a corresponding manner in a control device of the internal combustion engine for controlling the ignition or for controlling the torque of the internal combustion engine. The rest position of the rotor of the electric machine and thus also of the crankshaft of the internal combustion engine can be correspondingly detected. Corresponding control in a higher-level control device, in particular an engine control device, is particularly preferred, since this control device is already present and can use system resources accordingly, so that functionalities for identifying the rotational speed, the rotational angle position and/or the rotational direction and functionalities for controlling the internal combustion engine can be combined in one control device. In this way, a synergistic effect is obtained with respect to rules and communication infrastructure that can be shared. Further advantages are also obtained since the summation signal is used for data exchange with the control device, since only one data line is required for this purpose.

For this purpose, the computing unit used, which is preferably designed as an engine control device for an internal combustion engine, has a correspondingly integrated circuit and/or a computer program stored on a memory, which is/are set up to carry out the method steps described above.

It is advantageous to implement the method or to provide an integrated circuit, in particular an ASIC (application specific integrated circuit), in the form of a computer program which is preferably stored in the form of software on a data carrier, in particular a memory, and is available in the computing unit for implementation of the method, since this results in particularly low costs, in particular when the control device to be implemented is also used for other tasks and thus always already exists. Data carriers suitable for providing the computer program are, in particular, magnetic memories, optical memories and electronic memories, as are often known from the prior art.

Drawings

Further advantages and embodiments of the invention emerge from the description and the accompanying drawings.

Fig. 1 shows a schematic representation of a sensor wheel according to the prior art with sensors, in particular for rotational speed determination;

fig. 2a to 2c show schematic diagrams (a, b) of an electric machine coupled to an internal combustion engine and the associated signal profile (c);

fig. 3 schematically shows an electric machine with corresponding associated phase signals;

fig. 4a and 4b schematically show an evaluation circuit (a) for phase voltages and phase signals (b) associated with the phase voltages according to an embodiment;

fig. 5a to 5c show schematically an evaluation circuit (a) for phase voltages, a phase signal (b) associated with the phase voltages and typical signal profiles (c) at the individual components of the evaluation circuit according to a further embodiment;

fig. 6a and 6b show phase signals (a) of the phases of a three-phase motor obtained from the course of the voltage change and an enlarged view (b) of these signals;

fig. 7a and 7b schematically show a circuit (a) for generating a sum signal from a plurality of phase signals and a time-dependent course of the respective phase signals and the sum signal (b);

fig. 8a to 8d schematically show the course of the time-dependent summation signal (a) according to which the determination of the direction of rotation of the shaft is explained; a time-dependent course (b) of the summation signal is schematically shown, which describes the determination of the rotational angle position of the shaft according to the first embodiment; a time-dependent course (c) of the summation signal is schematically shown, which represents a determination of the rotational angle position of the shaft according to a further embodiment; a schematic flow chart (d) illustrating a method according to one embodiment is shown;

FIG. 9 shows a flow chart of a method for determining the rotational speed of a shaft;

fig. 10a to 10f show the course of the change (a) of the three phase voltages, a representation (b) of the frequency analysis at higher harmonics of the fundamental frequency as a function of the amplitude of the phase voltage from (a), a rotational speed signal (c) determined from the phase voltage from (a), a further representation (d) and an enlarged representation (e) of the frequency analysis at higher harmonics of the fundamental frequency as a function of the amplitude of the phase voltage from (a), and a representation of the rotational speed signal influenced by the interference factor and accordingly a rotational speed signal (f) corrected by means of further methods;

fig. 11a to 11c show a flow chart (a-c) of a method for improved determination of the rotational speed of a shaft, which is shown by a rotational speed value, according to a further embodiment; and

fig. 12 shows a representation of the rotational speed signal influenced by the interference factor and a correspondingly corrected rotational speed signal by means of a further method.

Detailed Description

Fig. 1 schematically shows a sensor wheel 20 of a rotational speed sensor G and an associated inductive sensor 10, as it is used in the prior art for determining the rotational speed or for approximately determining the rotational angle position of a crankshaft. In this case, the sensor wheel 20 is fixedly connected to the crankshaft of the internal combustion engine and the sensor 10 is mounted in a stationary manner in a suitable position.

The sensor wheel 20, typically a sensor wheel 20 made of ferromagnetic material, has teeth 22 which are arranged on the outside with a space 21 between two teeth 22. At one position on this outer side, the sensor wheel 20 has a recess 23 over the length of a predetermined number of teeth. This recess 23 serves as a reference mark for identifying the absolute position of the sensor wheel 20.

The sensor 10 has a bar magnet 11 on which a soft magnetic pole pin 12 is mounted. The soft pole pin 12 is in turn surrounded by an induction coil 13. As the sensor wheel rotates, the teeth 22 and the gaps between two respective teeth alternately pass the induction coil 13 of the sensor 10. Since the sensor wheel and therefore also the teeth 22 consist of ferromagnetic material, a signal is induced in the coil during rotation, with which a distinction can be made between the teeth 22 and the recesses.

By correlating the time difference between two teeth with the angle enclosed by the two teeth, the angular speed or rotational speed of the crankshaft and, in addition, the corresponding angular position of the crankshaft can be approximated.

At the gap 23, the signal induced in the induction coil has a different course than at the tooth 22 which otherwise alternates with the gap.

Fig. 2a shows an internal combustion engine 112 to which the electric machine 30 is directly or via a transmission coupling, wherein the electric machine 30 is driven by a crankshaft 17' of the internal combustion engine 112. Thus, the rotational speed n of the motor 30_GenAnd the rotational speed n of the crankshaft 17_BKMAnd the angular position of the rotor of the motor 30

And the rotational angular position of the crankshaft 17

The electric machine 30 is also assigned a charging regulator L R, which supplies energy to the battery B within the vehicle electrical system 110 as a function of the remaining capacity of the battery B.

A computing unit, in particular an engine control device 122, is also provided, which exchanges data with the electric machine 30 or with the internal combustion engine 112 via a communication connection 124 and is set up to correspondingly control the internal combustion engine 112 and the electric machine 30. External sensor data, e.g. for the inductive detection of the speed n of the internal combustion engine 112_BKMOr preferably fixedly coupled to the internal combustion engine, of the electric machine 30（n_Gen) The sensor data of the sensor 10 can likewise be included in the communication link 124, wherein the engine control device 122 forwards control signals for controlling the internal combustion engine 112 to the internal combustion engine on the basis of the sensor data of the electric machine 30 and/or further sensor data of the sensor 10. Rotor 32 and direction of rotation a of its shaft 17₊、a_-Is also described, wherein a₊Indicates a positive rotation in the preferential direction of the internal combustion engine 112 and a_-Indicating reverse rotation in the opposite direction. The angular position of the crankshaft 17' is also specified

Or the rotational angle position of the rotor 32

。

In fig. 2b, the motor 30 is again schematically shown in an enlarged form. The motor 30 has: a rotor 32 with a shaft 17, the rotor having field windings; and a stator 33 having stator windings U, V, W. It therefore relates to an electric machine, as is usual in particular in motor vehicles. However, especially for motor-driven vehicles, especially in the case of small and lightweight motor-driven vehicles, motors with permanent magnets, that is to say permanent magnet motors, are mostly used. Within the scope of the invention, it is in principle possible to use two types of electric machines, wherein in particular the method according to the invention does not depend on the use of a corresponding type of electric machine, for example a permanent magnet machine or an electric machine with its excitation.

The electric machine 30 is designed as a three-phase generator, in which three phase voltage signals are induced, which are phase-shifted by 120 ° with respect to one another. Such three-phase generators are usually used as generators in modern motor vehicles and are suitable for carrying out the method according to the invention. Within the scope of the invention, in principle, all electric machines can be used, independently of their number of phases, wherein in particular the method according to the invention does not depend on the use of a corresponding type of electric machine.

The three phases of the three-phase generator 30 are represented by U, V, W. By the structureThe rectifying elements, which are the positive diode 34 and the negative diode 35, rectify the voltages dropped on these phases. Therefore, the generator voltage U_GAttached between the poles B + and B-, and the negative pole is grounded at the voltage of the generator. For example, a battery B or other electrical consumers within the vehicle electrical system 110 are supplied by such a three-phase generator 30.

Fig. 2c shows three graphs, which show the associated voltage profile with respect to the angle of rotation of the rotor 32 of the electric machine 30. In the upper diagram, the course of the voltage at phase U, V, W is recorded. It is generally easy to understand that: the numbers and value ranges illustrated in this diagram and in the subsequent diagrams are merely exemplary and thus do not limit the invention in principle.

In the middle diagram, the generator voltage U is shown_GThe generator voltage is formed by the envelope of the positive and negative half-waves of the voltage variation process U, V, W.

Finally, in the lower diagram, the rectified generator voltage U is shown_G-(see fig. 2 a) together with the generator voltage U_G-Effective value of (U)_GeffThe generator voltage is attached between B + and B-.

In fig. 3, the stator 33 with phase U, V, W from fig. 2b is schematically shown, together with a positive diode 34 and a negative diode 35. In principle, it is easy to understand that: the rectifier elements depicted here in the form of a positive diode 34 and a negative diode 35 in the case of active rectifiers can also be constructed as transistors, in particular MOSFETs (metal oxide semiconductor field effect transistors) (not shown). The nomenclature used hereinafter for the voltages and currents that occur is also shown.

Alternatively, U_U、U_V、U_WThe phase voltages of associated phase U, V, W are shown as they are dropped between the outer conductor of stator 33 and the star point. U shape_UV、U_VW、U_WUWhich represents the voltage between two phases or the associated outer conductors of these two phases.

I_U、I_V、I_WShowing phase currents from the corresponding outer conductors of phase U, V, W to the star point. I represents the total current of all phases after rectification.

Fig. 4 schematically shows a schematic representation of a phase voltage U for a first embodiment_UAnd an evaluation circuit 80a in the form of a schmitt trigger, and a phase voltage U of a phase U of the electric motor 30_U(upper diagram) and the triggered phase voltage U obtained by means of the evaluation circuit 80a_Ut. A so-called schmitt trigger 80a is schematically shown in fig. 4 a. By means of such an evaluation circuit 80a, the phase voltage U is adapted to the respective phase voltage in the present case_UInput signal U of_I1Generating an output signal U obtained from the analysis circuit 80a_O1Triggered phase voltage U as it is according to the lower diagram of fig. 4b)_UtAs shown. The dashed horizontal line (see fig. 4 a) illustrated in the upper diagram correspondingly illustrates the switching threshold of the evaluation circuit 80 a.

Resistor R in the feedback branch of operational amplifier O_1aAnd R_2aResulting in a current-dependent output voltage U at the output_O1Of the switching threshold. Thus, by specifying the respective switching threshold, the characteristic shown in the course of the signal change can be realized and ensured such that signal noise in the vicinity of the acoustic point does not lead to a change in the output level of the operational amplifier O. By means of the evaluation circuit, a triggered phase signal U can be generated which can be evaluated in a correspondingly simple manner_UtThe rotational speed of the motor 30 can be determined from the time intervals of the steep edge of the phase signal and the respectively adjacent edge F L.

Fig. 5 shows a further embodiment of an evaluation circuit 80b, by means of which the phase voltage U is assigned to the current case as an example_UInput signal U of_I2To generate an output signal U_O2The output signal in the present case corresponds to the triggered phase voltage U_Ut. Phase voltage U_UtShown in the lower diagram of fig. 5 b). It is easy to understand that: also having other phase voltages U_U、U_V、U_WAttached to the input endU_I2The other phase voltages may result in the output terminal U_O2Or U_O2U、U_O2V、U_O2WUpper pair of triggered phase voltages U_Ut、U_Vt、U_WtTo this end, see in particular fig. 6 and 7.

The evaluation circuit 80B has a blocking device B for the purpose of processing the corresponding phase signal U_U、U_V、U_WFiltering over time to perform noise suppression. In this case, the blocking device B is followed by at least one operational amplifier O, phase signal U_U、U_V、U_WIs attached to the input terminal O of the operational amplifier_inThe above. The blocking device B has: a time-lag element T, the adjustable time-lag of which defines the time constant of the filtering over time; two switches F₁And F₂(ii) a And an inverter stage I. The time-lag element T has a resistance R_1cAnd a capacitor C_1cThe capacitor is charged with the current flowing through the output of the operational amplifier O. At the output U of the switching analysis circuit_o2Then, a time-lag element T and a first switch F, preferably in the form of an R/S flip-flop (Flipflop), are passed through₁In combination with switch F₁Generates a pulse signal having a positive output voltage (High-L ev, High) with respect to a static voltage (low level (L ow-L ev), L ow), where the rising edge of the pulse signal and the signal O are at a High level (High-L ev, High)_outIs simultaneously present at the output of the operational amplifier O. The subsequent falling edge of the pulse occurs after the time lag of the time-lag element T. In parallel, the output signal O of the operational amplifier O is generated by the inverter stage I_outThe inverted signal of (1). The inverted signal is sent to a second switch F which is preferably constructed in the form of an R/S flip-flop₂And a pulse signal is sent to a Reset (Reset) input R. In this case, the switch F2 passes a first signal edge of the inverted signal to the output U_o2And then prevents the output from switching again as soon as the High level of the pulse signal (High-L evel) is attached or the output signal itself already has a High level.

Thereby, at the input end O of the operational amplifier O_inNoise in the range of the falling edge of the upper band at the output U_o2The above is suppressed by: as long as there is no high level at the reset input R, the switch F₂The output cannot be switched again. For the input terminal O of the operational amplifier_inOn the rising edge of switch F₂Has a switch F on the reset input terminal R₁Pulse of the pulse signal. The pulse signal results in: switch F2 will output terminal U_o2Set to low level (L ow) at input O_inPossible noise in the range of the rising edge above is suppressed by the switch F being held back as long as a High level (High-L evel) is attached to the reset input R₂The output terminal U cannot be put again_o2Set to a high level (Hgh). Therefore, during the pulse length, the output terminal U cannot be realized_o2And possible noise is suppressed. The characteristic time constant for this noise suppression is determined by the pair resistance R_1cAnd a capacitor C_1cTo the corresponding selection.

By using the blocking device B, the corresponding noise effects can be suppressed, as they are in the time range or duration Z in the upper diagram in fig. 5B₀As shown near the inner zero line (each characterized by a circle). This results in a triggered phase signal U_Ut、U_Vt、U_WtThe signal quality of (2) is improved. In fig. 5C), the inverting stage I or Flip-Flop (Flip-Flop) F is shown on the output of the operational amplifier O, on the resistor R or the capacitor C, respectively₁Or F₂Typical signal variation process of (1).

It is advantageous in terms of this circuit with respect to the schmitt trigger solution (see fig. 3): detects the phase signal U_U、U_V、U_WIn the form of a significantly more precise zero crossing of the input signal, wherein in the case of schmitt trigger 80a switching threshold for the phase signal is always required>0 (see fig. 3). By this, a phase signal U can be used_uCharacteristic value W of_U0Or in the presence of multiple phase signalsTo use a correspondingly parallel layout of these flip-flop circuits. These characteristic values W_U0、W_V0、W_W0In particular, it can be used to determine the rotational speed n or the rotational angle position of the rotor 32 of the electric machine 30.

Here, the noise suppression is not realized by different thresholds but by filtering over time by means of time-lag elements. The skew element switches the output U due to a zero crossing at the input_o2After which the output U is prevented for a first (short) time_o2Directly back (especially due to noise) is switched back. To output end U_o2May be triggered after the end of the dead time and the subsequent zero crossing, wherein only then is the trigger switch F placed after the operational amplifier O and the dead time element T₁And/or F₂Is wired and outputs a signal U_O2With triggered phase signal U_UtThe form of (d) is output.

Now, three phase voltage signals U with potential B-as reference are shown in FIG. 6a in three graphs versus time_U、U_V、U_WAs they occur in generators having an outer pole rotor with six permanent magnets. The illustration of an electrical machine 30 with three-phase stator windings 33 is to be seen merely as an example, wherein the method according to the invention can in principle also be implemented on generators with a correspondingly sufficient number of phases or permanent magnets or field coils, without limiting the generality. Likewise, instead of star-connection of the stator coils, delta-connection or other connection methods can also be selected.

In the case of an electric machine 30 with current output, the phase voltage U is the phase voltage_U、U_V、U_WIs rectangular in a first approximation. This is indicated in particular by the following: due to the generator voltage, either the positive diode or the negative diode is conducting in the conducting direction, and thus either approximately 15-16 volts (battery charging voltage at 12V lead-acid battery and voltage across the positive diode) or negative 0.7-1 volts (voltage across the negative diode) is measured. The reference potentials of the measurements are respectively connectedAnd (3) ground. Other reference potentials, such as star point of the stator, may also be selected. These reference potentials result in an alternating signal profile, but do not change the analyzable information, the acquisition and analysis of this information.

In principle, phase signal (U)_U、U_V、U_W、I_U、I_V、I_W) May be obtained in different ways. For example, it is possible that: determining phase voltages (U) relative to each other_UV、U_VW、U_WU) (ii) a As long as the stator of the electric machine is star-connected with a measurable star point, the phase voltage is determined by the diodes of the connected rectifier relative to the output terminals (B +, B-) of the rectifier; taking into account the output voltage (U) of the branch with respect to the star point_U、U_V、U_W) (ii) a Or similarly analyze the phase currents.

In fig. 6b, the phase voltages U from fig. 6a processed by means of the triggering circuits 80a, b (see fig. 4 or 5 for this purpose) are plotted together in a diagram_Ut、U_Vt、U_Wt. In this case, a uniform phase shift is clearly visible. The term phase voltage U_U、U_V、U_WOr the processed phase voltage U_Ut、U_Vt、U_WtAre then used partially synonymously, since the phase voltage U is processed_Ut、U_Vt、U_WtDirectly from the phase voltage U_U、U_V、U_WAnd (5) obtaining the result.

The voltage signal is repeated six times by six magnets, in particular permanent magnets, so-called pole pairs, during one full revolution of the rotor 32 of the electrical machine 30. Correspondingly, for each revolution of the rotor 32, each phase, i.e. each phase voltage U_U、U_V、U_WSix falling edges F L occur_DAnd six rising edges F L_U(F L for the corresponding phase)_UU、FL_VU、FL_WUAnd F L_UD、FL_VD、FL_WD）。

The edges define an angular section, i.e. precisely by the magnetsThe body covers an angular section in the radial direction along the circumference of the stator, therefore, knowing the absolute reference point per revolution, the corresponding edge F L can be determined_UOr F L_DFor example, the reference point is characterized by a reference magnet having a phase voltage U_U、U_V、U_WTriggered change U of the phase voltage_Ut、U_Vt、U_WtDifferent from other magnets.

Now, with suitable means, it is possible to identify not only the falling edge F L of the phase voltage_DAnd may identify the rising edge F L of the phase voltage_UFor example, for each phase voltage, a triggered profile U of the phase voltage in the form of the TT L signal can be generated by means of the circuit shown in fig. 4 or 5_Ut、U_Vt、U_WtThe required trigger (see fig. 4 or 5) can either be integrated in the control device or in the control electronics, for example the control device, the regulator for the battery voltage and/or in the case of an active rectifier in the respective generator regulator, or can also be assigned to it externally, the individual TT L signals can be combined in particular for the case of the use of the control device, in particular the engine control device 122 (see fig. 2 a), to form a sum signal U_Sum(see fig. 7 and 8) and are preferably transmitted via only one data line 124 (see fig. 2 a) by wire or by a preceding combined electronic device or in other ways suitably combined. In this case, the signal U is added_SumOutput signal U corresponding to circuit 80c shown in FIG. 7_Out。

In FIG. 6b, the phase voltage U_U、U_V、U_WRespectively assigned with the value W_U、W_V、W_WThese values are also referred to as W_Ud、W_Vd、W_WdLikewise, rising edge F L may also be given_UAssign the corresponding value W_Uu、W_Vu、W_Wu. The zero crossing of the phase voltage may also be assigned a corresponding value W_U0、W_V0、W_W0These values can be used to identify the rotational speed n of the rotor 32 or of a crankshaft 17' coupled thereto, the angular position α of the rotor 32 depending on the plateau region of the phase signal or other region in between₁Is also possible. These values are used for: according to the time difference Deltat₁、Δt₂、Δt₃To determine the rotational speed of the motor 30. Use of the

corresponding trigger circuits

80a, 80b (see fig. 4 and 5) generates a triggered phase signal U from the phase signal_Ut、U_Vt、U_WtWherein the corresponding edge marks the corresponding value W_U、W_V、W_W、W_U0、W_V0、W_W0The moment of occurrence of.

In this case, with a uniform arrangement of six permanent magnets in the electric machine 30, a total of 18 falling edges F L occur at equal intervals from one another per revolution, respectively_dAnd therefore 18 belonging values appear. Therefore, at the time difference Δ t₁、Δt₂Or Δ t₃During this, the angle 360 °/18 = 20 ° is covered, which, as already mentioned at the outset, can also be used to detect the direction of rotation α of the rotor 32₊、α_-The exemplary determined 20 ° is the detectable angle increment. Furthermore, from this, the angular velocity ω can also be determined_i. The angular velocity is based on ω_i= 20°/Δt_iTo obtain and associated rotational speed n_iAccording to n_i= ω_i60 s/min/360 DEG is obtained in revolutions per minute.

In principle, it is readily understood that the alternative is to fall edge F L_DThe corresponding rising edge F L of phase U, V, W can also be used_UFor determining the direction α of rotation of the rotor 32₊、α_-And for determining the instantaneous speed n of the electric machine 30_GenCorrespondingly, since the number of values per revolution is doubled, not only the direction α of rotation of the rotor 32 is obtained₊、α_-And the rotational speed n_GenHigher resolution. Furthermore, the edges of the phases may be combined in a variety of other ways andby means of methods, e.g. by rising edges F L of respectively identical or corresponding phases_UAnd a falling edge F L_DAt intervals from each other or by a common rising edge F L of the same or all phases_UOr falling edge F L_DTime intervals of the time interval.

Except for rising edge F L_UAnd a falling edge F L_DIn addition, the phase signal U may be used_U、U_V、U_WZero crossing W of_U0、W_V0、W_W0For determining the rotational speed n of the shaft 17_Gen。

In fig. 7a and b, a method for converting a plurality of input signals U is shown_In1、U_In2、U_In3Combined to output signal U_OutThe circuit 80 c. Input signal U_In1To U_In3For example, may correspond to a corresponding phase signal U_U、U_VAnd U_WAnd (4) associating. In principle, the application is not limited to three input signals. However, it is preferred that the output signal U of the flip-

flop circuits

80a, 80b shown in fig. 4 or 5_o1Or U_o2Or the output signal of another trigger circuit which preprocesses the phase signal of the motor in a suitable form, as the input signal U_In1To U_In3. In fig. 7b, the change processes obtained by means of the flip-

flop circuits

80a, 80b with the corresponding phase signals as input signals for the flip-

flop circuits

80a, 80b are used as input signals U_In1、U_In2And U_In3To illustrate. In the present case, the circuit 2 has reference potentials U + and U-. However, the circuit arrangement can also be designed in principle such that it has only one fixed reference potential.

Input signal U_In1To U_In3Respectively assigned with its own pulse forming unit P₁To P₃The pulse-forming units are set up to convert the rising or falling edge of the input signal into a corresponding pulse P_f1、P_f2、P_f3、P_r1、P_r2、P_r3(see FIG. 7 b). Corresponding pulse forming unit P₁、P₂、P₃Through a downstream output for eachA signal phase U_In1To U_In3The associated circuits being combined to a common signal U_SumAnd at the output terminal U_outIs provided. For this association, an AND element may be used, for example, as shown. However, other suitable correlation approaches may also be used for this purpose, for example correlation by means of transistors (of different types, for example MOSFETs or bipolar transistors) or operational amplifiers.

Pulse forming unit P₁To P₃In each case, a plurality of components, in particular capacitors, at least two resistors, transistors (different types are possible, for example MOSFETs or other transistor types can also be used in addition to the bipolar transistors depicted), and in each case an exclusive-or (Xor) gate. And an input signal U_In1To U_In3The rising and falling edges in (b) are associated with a pulse P_f1、P_f2、P_f3、P_r1、P_r2、P_r3Or characteristic features of these pulses, such as pulse length or pulse width T_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3By applying a pulse-forming unit P to the corresponding pulse-forming unit₁To P₃The selection of the respective components in (a) is adjusted accordingly.

In this case, the capacitor C1 and the resistor R3 in the cell P1, for example, constitute a first time-lag element T₁₁The capacitor C2 and the resistor R6 constitute a second time-lag element T12 in the pulse forming unit P2, while the capacitor C3 and the resistor R9 constitute another time-lag element T in the pulse forming unit P3₁₃. Pulse P_f1、P_f2、P_f3、P_r1、P_r2、P_r3Pulse width T of_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3Mainly by means of time-delay elements T₁₁、T₁₂、T₁₃Is correspondingly defined, in particular, by the selection of the resistors R3, R6 and R9 and/or the capacitors C1, C2 and C3. By this, it is possible to make a pair as can be seen in fig. 7bPulse P_f1、P_f2、P_f3、P_r1、P_r2、P_r3And (6) coding is carried out.

In this case, the input signal U_In1To U_In3All edges of (b) obtain corresponding pulses P_f1、P_f2、P_f3、P_r1、P_r2、P_r3Said pulses having associated different pulse widths T_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3. Thus, for example, inputting signal U_In3Is obtained with a pulse width T_Pr1Pulse P of_r1Input signal U_In1Is obtained to have a pulse width T_Pr2Pulse P of_r2And input a signal U_In2Has a pulse width T_Pf1Pulse P of_f1Wherein the pulse widths of the respective pulses have different widths (see fig. 7 b).

Thus, the pulse P_f1、P_f2、P_f3、P_r1、P_r2And P_r3Marks the input signal U at the falling edge_In1To U_In3The precise location over time of the corresponding edge in (a). Time interval T from the subsequent rising edge of the encoded signal_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2And T_Pr3Contains information on: the direction in which the zero crossing of the respective input signal has, that is to say whether it has a falling or rising edge of the respective signal, and to which phase the zero crossing should be assigned. As already mentioned, all corresponding edges in the input signal, which correspondingly can also be associated with the zero crossings, are unambiguously encoded with the corresponding pulses.

The following boundary conditions apply in principle: the usable maximum pulse length of the respective pulse is obtained from the shortest possible time between two successive zero crossings and thereby from the maximum rotational speed of the electric motor 30. By means of respective summation signals U by means of corresponding control devices_In1To U_In3Generated bySum signal U of_SumIn particular, the rotational speed can be determined.

In this case, the edges and the pulses P associated with the edges_f1、P_f2、P_f3、P_r1、P_r2And P_r3And their time intervals between each other may be used. In order to compensate for possible differences between rising and falling edges, the rotational speed can preferably be determined as a function of the time interval between two edges of the same type (rising or falling edge) or the assigned pulses of these two edges. In the present case, the spacing between the respective pulses (here, for example, for the pulse P of the falling signal edge) required for this purpose_f1、P_f2、P_f3To show) Δ t₁、Δt₂、Δt₃According to time Deltat₁= t2+ t3、Δt₂= t4 + t5 and Δ t₃= t6 and correspondingly a not depicted time period t 7. As already mentioned at the outset, these edges are provided by a corresponding arrangement of a plurality of permanent magnets within the electric machine. Thus, in the case of 6 permanent magnets (also referred to as pole pairs), as long as these pole pairs are arranged equidistantly in correspondence, a total of 18 falling edges Fl results, each at the same distance from one another per revolution, respectively_dAnd thus 18 associated values. Therefore, at the mentioned time difference Δ t₁、Δt₂Or Δ t₃During this period, the stepping angle 360 degrees is divided by 18 = 20 degrees, and the motor information of the motor can be used to determine not only the rotational speed but also the rotational direction α of the rotor 32₊、α_-But may also be used to determine the absolute position of the rotor 32 of the motor 30. It is easy to understand that: correspondingly using pulses P_r1、P_r2、P_r3To assign rising edges and the time intervals at which these rising edges can be processed.

By determining the function according to_i= 20°/Δt_iAnd the angular velocity ω obtained_iCan be according to n_i= ω_iThe rotation speed n is calculated in revolutions per minute by 360 DEG 60s/min_i. Method step for determining a rotational speedAlso referred to subsequently as step a 3. Further, the pulse information P_f1、P_f2、P_f3、P_r1、P_r2To P_r3May also be used to determine the direction of rotation α of the rotor 32₊、α_-. This step is referred to as a1 and is shown schematically in fig. 8 a. Therefore, based on the pulse P_f1、P_f2、P_f3、P_r1、P_r2To P_r3Can be deduced from the time sequence of the pulses, which can be based on the respective pulse width T in particular, of the direction of rotation of the rotor 32 of the electric machine 30_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3To distinguish them from each other. Therefore, if it has a pulse width T_Pr1Pulse P of_r1Followed by a pulse P_f1And is followed by P_r2Then the crankshaft is in the clockwise direction α₊Rotating; and vice versa if the pulse width T is within the range_Pr1Pulse P of_r1Followed by a pulse P_f3And is followed by a pulse P_r3Then the crankshaft is in the opposite direction α_-And thus rotates in a counterclockwise direction. Thus, based on the pair pulse P_f1、P_f2、P_f3、P_r1、P_r2And P_r3The direction of rotation of the rotor 32 of the motor can be deduced in reverse.

Furthermore, the corresponding pulse width T can also be determined_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3To determine the corresponding pulse P_f1、P_f2、P_f3、P_r1、P_r2And P_r3Furthermore, the rotational angle position of the rotor 32 can be determined using the properties of the electric machine 30, in particular the previously mentioned positioning of the permanent magnets within the electric machine

. As already mentioned at the outset, the corresponding pulse P_f1、P_f2、P_f3、P_r1、P_r2And P_r3And at the input signal U_In1To U_In3Rising edge of (1) andthe falling edges are correlated. These rising and falling edges are in turn associated with motor parameters such as the layout of the permanent magnets, so-called pole pairs within the motor. Thus, the pulses and their relative spacing from each other describe the corresponding angular increment within which the rotor 32 of the motor is located.

To the rotational angular position of the rotor

The determination of or the step for determining the respective rotational angle position is illustrated in fig. 8 and this step is subsequently denoted by a 2. According to the sum signal U_SumTo the position of the rotating angle

Another design of the determination of (a) is illustrated in detail in fig. 8 c. In this case, an arbitrary position point P that should be determined is assumed. In the present case, the point is at pulse P_f1And pulse P_r2But may in principle be chosen arbitrarily. Thus, the position is within an angular increment that is passed by the pulse P_f2And pulse P_r2To be associated with the pole pairs of the motor 30. Thus, by using the pulse P_f1Time interval Δ t of_Pf1And with respect to pulse P_r2Time interval Δ t of_Pr2And using the speed of rotation, especially in pulses P_f1And P_r2The average rotational speed in the time period in between can also explain more precisely the orientation of the rotor at the position P beyond the angular increment.

In fig. 8d, a flow chart of a method according to the first embodiment is generally illustrated. In step S1, at least two input signals, preferably in the form of phase signals, are determined. In step S2, signal U is input_In1To U_In3Rising edge Fl of_uAnd falling edge Fl_dThe edges being in the form of corresponding pulses P_f1、P_f2、P_f3、P_r1、P_r2And P_r3Is encoded. The signal sequences derived therefrom are combined in step S3 into a sum signalU_Sum. Depending on the sum signal generated in step S3, it may be possible to depend on the pulse P as described earlier_f1、P_f2、P_f3、P_r1、P_r2And P_r3The rotational speed of the shaft, in particular of the rotor 32 of the electric machine 30, is determined in a step A3, the rotational direction of the shaft, in particular of the rotor 32 of the electric machine 30, is determined in a step a1, and the rotational angle position of the shaft, in particular of the rotor 32 of the electric machine 30, is determined in a step a2, and the rotational speed, rotational angle position or rotational direction of the crankshaft 17 of the internal combustion engine 110 can also be determined from this by the coupling of the rotor to the crankshaft of the electric machine 30. In principle, it is easy to understand that: steps a1 to A3, in which the rotational speed, the rotational angle position or the rotational direction of the rotor is determined, can alternatively be applied, but can also be applied cumulatively.

In addition to the two embodiments shown, other circuits are also conceivable which divert at least one of the input signals to an output with a well-defined extension, in particular a signal pulse with a well-defined length, and superimpose it with the other (unchanged) input signals, so that later on when the sum signal is evaluated it is possible to deduce inversely which signal edge is caused by which input signal.

Fig. 9 shows a further time chart of a further embodiment of a method for determining a rotational speed n, which has at least four successive steps SU₁，SU₂，SU₃And SU₄. In a first step SU₁In (1), determining respective phase signals U of the motor 30_U，U_V，U_WOr I_U，I_V，I_W. In another step SU₂To the obtained phase signal U_U，U_V，U_WCoding or individual phase signals U depending on the number of phases_U，U_V，U_WEncoded and combined into a sum signal U_Sum(see figures 7 and 8 for this). In another step SU₃In, sum signal U_SumDecoding and decomposing into its triggered or non-triggered statesPhase signal U of_U，U_V，U_WWherein the phase signal U of the respective phase is also identified_U，U_V，U_WAnd determining the time interval at₁，Δt₂，Δt₃(see FIG. 6b for this purpose). In step SU₄In the middle, by a time interval Δ t₁，Δt₂，Δt₃The rotational speed n of the motor 30 is determined. Due to the addition of the signal U by means of the code C_SumOf the corresponding phase signal U_U，U_V，U_WThe addition signal U can easily be summed up by the fact that the characterizing features are applied separately_SumAre distinguished from each other and can therefore be based directly on the individual phase signals U_U，U_V，U_WThe associated respective falling and rising edges may determine the time interval Δ t and thus the rotational speed n of the shaft 17.

The correction of the phase signal U should be explained based on the illustration in fig. 10_U，U_V，U_WFor eliminating the corresponding phase signal U for the higher order_U，U_V，U_WIn order to thereby ensure a more accurate determination of the rotational speed n associated with said basic periodicity. This further method step can be implemented as follows: in a first step SU_4aIn a further step SU, the rotational speed n is determined from the raw signal of the rotational speed n, which is not cleaned, as described above_4bIs determined (see fig. 9) and in step SU_4cWherein a corrected rotational speed n is calculated as a function of a determined or already present correction factor K_corr1. This allows higher-order frequency influences to be removed from the original signal of the rotational speed.

For this purpose, fig. 10a shows three phase voltages U_U，U_V，U_WThese phase voltages have a fundamental periodicity of 1 order and higher order frequency components. The phase-locked signal U is correspondingly shown in fig. 10b_U，U_V，U_WWherein the frequency of at least one phase signal is plotted over an integer multiple of the fundamental frequencyThe amplitude P is made. It should be noted in principle that the output signal U depends on the structure of the motor 30_U，U_V，U_WMay vary accordingly. Except for the fundamental frequency f₀(in FIG. 10b phase voltage U is used_U，U_V，U _W1 st order), the specificity of the stator's magnets and coils, as well as the design of the rotor 32 and stator 30, may also excite harmonics that are integer multiples of, and correspondingly superimposed with, the fundamental frequency of 1 st order. This may be based on the phase signal U in fig. 10_U，U_VAnd U_WAnd it is accordingly clear based on the frequency analysis as shown in fig. 10b or fig. 10b and 10e that if it is seen that in addition to the fundamental frequency of order 1 (with order 0.5 in fig. 10d and 10 e) also the respective other frequency component f of higher order is superimposed_nIf so. Accordingly, other interference influences which differ from the fundamental frequency and have a higher or lower frequency relative to the fundamental frequency can also be identified. Here, in the case of a four-stroke engine having cylinders, 1 st order corresponds to the frequency of the crankshaft, and 0.5 th order corresponds to the operating frequency of the internal combustion engine. The method described above can be adapted accordingly to four-stroke engines with more cylinders, but also to two-stroke engines.

These tolerances and design features shown have a strong influence on the rotational speed signal n, which is shown in fig. 10c as an example as a measurement signal. Here, a clear sawtooth of the curve over time can be seen. This sawtooth or the sawtooth pattern shown here, whose periodicity is dependent on the number of phases of the electric motor 30, repeats every three sampling points, currently at the respective sampling rate, on the basis of measurements made on a three-phase motor having six pole pairs. Further method steps for compensating for the effects of disturbances are now explained below, in particular for a three-phase machine with six pole pairs.

In principle, it is easy to understand that the method steps described below for compensating for these interference effects are not limited to three-phase motors having a corresponding number of pole pairs, but can also be applied to motors having any number of phases or pole pairs.

In a first method step for correcting the rotational speed n, the frequency spectrum of the operating cycle or of the first revolution of the shaft 17 is analyzed (see fig. 10b, 10d and 10e for said frequency spectrum). In principle, it can be recognized that for each revolution of shaft 17, corresponding phase voltage U of electric machine 30 is present_U，U_V，U_WRepeated in the number of magnets present, thus six times per revolution. In phase U_U，U_V，U_WCan identify two zero crossings of the voltage accordingly in each period of the output voltage. In the present case, the phase voltage U_U，U_V，U_WThe sequence of (a) repeats at six times the crankshaft frequency. Corresponding phase voltages U of different phases_U，U_V，U_WThe zero-crossing sequence of (c) is repeated at twice the number of magnets used, i.e. twelve times the crankshaft frequency in the case of six magnets. Having positive and negative slopes (i.e. phase signal U)_U，U_VAnd U_WRising and falling edges in) is repeated at every second zero crossing of the voltage, i.e. at 18 times the crankshaft frequency. It is easy to understand that the crankshaft frequency is always the frequency of the fundamental order, i.e. the order N =1 in the example of fig. 10b or the order N =0.5 in the examples of fig. 10d and 10 e. It should also be mentioned briefly that in fig. 10e the fundamental frequency n =0 is masked in order to better show the higher harmonics of the fundamental frequency.

Preferably, the correction depends directly on the order O of the periodicity of the motor 30_n. These steps can be determined as follows (the values indicated in the brackets are obtained for a three-phase machine with six pole pairs for example):

let N be the number of phases U, V, W and p be the number of pole pairs of the motor 30.

Thus, each revolution of the shaft 17 of the motor 30 results in N times p electrical cycles having a zero crossing with a rising edge and a falling edge, respectively. Therefore, the total number a of sampling points of the corresponding rotation speed signal n is:

is composed of (and phase do not)Off) possible different behavior of the rising and falling edges to a third order O₃. The motor shaft here has a corresponding rising edge F for each revolution_sOr falling edge F_fThe number of (A) is:

In this case, the corresponding edge F for each revolution of the motor shaft_pThe number of (A) is:

In this case, the motor shaft has a rising edge F for each revolution_spOr falling edge F_fPThe number of (A) is:

It should be noted in principle that other disturbing influences falling within a similar frequency spectrum are not expected for the above-mentioned frequency components, in particular for higher order (n > 1) frequency components in the case of an internal combustion engine 112 with up to five cylinders. If the respective speed signal n (t) of a working cycle or first revolution is now converted into the frequency space, the spectral components at six, twelve and eighteen times the crankshaft frequency shown in fig. 10d and 10e must result due to the respective zero crossings of the electric machine 30 or the differences between the respective phases. These differences can be attributed to the tolerances or structural features shown above as the root cause.

Correction terms can thus be calculated for the other revolutions of the shaft 17, which correction terms result from the sum signal of the three harmonics mentioned above. To eliminate the effect of the lowest frequency oscillation on six crankshaft frequencies, six different correction terms are required. Thus, each zero crossing of each phase of the motor 30 obtains its own correction term.

The corresponding correction terms are:

correction term (k) =

Wherein the absolute value N_{Mot, Freq}(k) Representing the absolute value of the kth oscillation of the spectrum,

representing the phase of the kth oscillation of the spectrum.

Here, zero oscillation represents the dc component of the frequency spectrum L represents the number of rotational speed sampling points used to calculate the frequency spectrum, which applies in the case of a three-phase generator with six pole pairs L = 36.

In the case where the frequency spectrum is determined via one revolution of the crankshaft of the internal combustion engine, the length of 36 sampling points is obtained. The generator has six electrical cycles per phase for each phase during one revolution of the crankshaft. Thus each phase gets 6 times 2 edges and thus all three phases get 36 edges or sample points per revolution.

In addition to the way in which the correction terms are determined from the frequency spectrum of one revolution, it is alternatively also possible to use further time periods, for example half a revolution (L = 18) or a complete working cycle (two revolutions, L = 72), for the determination.

As already explained, the correction term of one revolution or one duty cycle is not available until after the correction term has been calculated by means of the frequency spectrum. Since in principle the rotational speed difference between two successive working cycles or revolutions is sufficiently small, the correction term thus obtained, which is calculated on the basis of the first working cycle or first revolution of the shaft 17, can be used to correct a subsequent working cycle or subsequent revolution of the shaft 17 without additional deviations being expected to result from the respective correction. It is therefore appropriate to use the determined terms of the previous working cycle or previous revolution of the shaft 17 (see correction equation) for the correction during the working cycle or revolution of the shaft 17.

It is easy to understand in principle that it is not mandatory to learn correction terms for the first duty cycle or first revolution of the shaft 17 in order to apply these correction terms to the immediately following revolution of the shaft 17. In principle, these correction terms can also be learned only occasionally and used all the time, as long as the rotational speed n moves within a limited tolerance band. Therefore, the new correction term is only started for larger rotational speed deviations which lie outside the previously defined tolerance band. Furthermore, it is also possible to learn the corresponding correction terms only occasionally and to store these correction terms in the characteristic map as a function of the speed n of the electric machine 30. In the aforementioned case, the correction term can always be learned at a suitable operating point, for example in a range with little dynamics within the rotational speed n.

The actual speed variation n of the shaft 17 is represented by the frequency spectrum f_nIs calculated, and the associated correction terms of the rotational speed sampling points are therefore subtracted from the calculated rotational speed n in order to eliminate the influence of these harmonics and thus to suppress the influence of tolerances and constructional features. This corrected rotational speed n_Corr1In fig. 10f, a solid line is shown, the dashed line correspondingly showing the uncorrected rotational speed signal n (t).

Another alternative for correcting the tacho signal n and for eliminating higher order frequency components is so-called low-pass filtering. In the case of the motor 30 with three phases and six pole pairs, which is considered in correspondence with the above description, a so-called 6 th order moving average filter is provided in principle, which averages six values within a range of smooth averages. The first zero at 6 th order is obtained by this design. Furthermore, the filter has zeros in 12 th and 18 th orders so that all relevant interference is cancelled.

In principle, it can be advantageous to further compensate the attenuation produced by the filter to the 4 th order via a simple filter.

Alternatively, another more complex low pass filter with sufficient attenuation (or even zero) at 6 th, 12 th and/or 18 th order may be used. By comparing the original signal with the filtered values, a correction value can be determined for each edge. Advantageously, a general speed trend n is also taken into account here_T(see the description of figures 11 and 12 for this purpose). Due to the frequency component f of the sixth order₆The associated energy is only very small and therefore filters suppressing orders 12 or higher may also be used.

Here, for example, a3 rd order moving average filter is provided, with which the zero point for the 12 th order is given for the 36 edges given in this case. The advantage of this design is that the useful signal energy is only minimally attenuated to about 3 to 4 orders. The filter must be adjusted accordingly for different numbers of edges. The correction factor k for these edges can then also be calculated again on the basis of this correction, since the running time during the filtering, for example for calculating the ignition, can be negative. In addition to the way of calculating the correction term and correcting the course of change on the basis of the frequency spectrum of the rotational speed signal, it is also possible to pass the phase voltage U from its frequency spectrum_U，U_VOr U_WDetermines a correction term for the course of the edge time given by the time between two zero crossings, or for the course of the energy change during one working cycle.

Fig. 11a to 11c are based on a flow chart which shows a further method step for improved determination of the rotational speed n of the shaft 17. The corrected rotational speed n thus obtained_Corr2Shown in fig. 12. These further method steps are preferably taken as steps SU_4a，SU_4b，SU_4cAlternative (see fig. 9, 10)But may also be performed cumulatively with this before or after. Still better signal quality can be achieved by additive correction. In this case, as can be seen in fig. 11a, 11b, in a selected time range t_linInternally selecting a number of successive speed points n₁To n_nWherein six successive rotational speed points n are selected in the present case₁To n₆. The number of speed points is preferably an integer multiple of the number of phases U, V, W of the motor 30. As can be seen in FIG. 11a, the rotational speed point n is set₁To n₆Is divided such that a first speed value n is₁To n₃And a first average value n_M1Relating and other speed values n₄To n₆With other mean values n_M2Correlation, in which the value n of the rotation speed is correlated₁To n₆Calculating the corresponding average value n_M1And n_M2. From the aforementioned speed value n₁To n₆The two calculated average values n_M1And n_M2For calculating the selected time range t_linInner linear speed trend n_T。

In principle, it is easily understood that any other number of rotational speed points n can be used₁To n_nTo determine a linear trend n_T. However, it should be noted in principle that the range t is chosen_linShould have only moderate dynamics of the speed profile n, preferably approximately linear, since only the speed trend n is carried out_TA linear approximation of. Determining two mean values n based on the method steps described in FIG. 11a_M1And n_M2The two averages and n₂And n₅Are associated with each other. The time interval between these values is Δ t. At said average value n_M1And n_M2Linear interpolation between them and thus the linear speed trend n is determined for this time interval_T。

According to FIG. 11b, the mean value n to be determined_M1And n_M2The time interval delta t between the two is divided into three parts, and the time t is calculated by the linear rotating speed trend₁And t₂To which the rotational speed belongs. Thereby being in accordance with the second average value n_M2Together obtain three rotation speed points n_LM1，n_LM2，n_LM3These speed points lie in a linear speed trend n_TThe above.

As already mentioned at the outset, the number of phases in the electric machine

In the case other than 3, the corresponding region t is increased for the number of phases greater than three_linAnd the number of correction factors, or for a number of phases less than three, reducing the respective region t_linAnd the number of correction factors.

According to the trend n of the rotating speed_TLinear value n of_LM1，n_LM2，n_LM3Corresponding rotational speed n₃To n₅The relationship between (hereinafter also referred to as n)_gemessen) Determining a correction factor for a working cycle or a revolution of the shaft 17 on the basis of the sampling points, said correction factor relating the respective time-corresponding rotational speed sampling points n₃To n₅And linear trend n_TAnd (6) adapting. This can be seen in particular in fig. 11b, since only the corresponding rotational speed sampling points n are used here₃Compensating the corresponding mean value n over time_LM1Using the rotational speed sampling point n₄Compensating the mean value n_LM2And sampling n with the rotational speed₅Compensating the mean value n_LM3. Three correction factors K of the respective phases are thus obtained_hComprises the following steps:

wherein

，

Wherein the corrected rotational speed n_Corr2Accordingly, the following equation is given:

。

due to the periodicity of the phase signals corresponding to the number of pole pairs of the motor, the phase signalsThe number is repeated a number of times within one working cycle or revolution of the shaft 17. It is therefore sufficient that the correction factor K be determined according to the above formula_h(wherein

) Further course of the rotational speed points of the working cycle or of the revolutions for other occurrences of the phase signals for correcting tolerances and structural influences, so that a corrected rotational speed course n is obtained for the entire working cycle or of the revolutions_Corr2。

As described above, three new correction factors can be calculated accordingly for the respective time ranges in the next working cycle. It is however readily understood that other duty cycles of the shaft 17 or other rotational corrections may alternatively be learned only occasionally and used at all times, as long as the rotational speed n moves within a limited rotational speed band. For greater dynamics of the rotational speed or rotational speed deviation beyond the respectively defined rotational speed range, a new correction term K must be calculated in the respective rotational speed window_hThese correction terms are correspondingly applied to the subsequent time window t of the axis_linOr duty cycle or revolution. The correction factor K can also be corrected via a corresponding speed profile n_hStored in the characteristic map in such a way that the speed n (t) for correcting the speed n (t) to a further corrected speed n can be improved in accordance with the respectively present speed profile n (t)_Corr2Corresponding correction factor K of_h。

In addition, instead of calculating the correction term K based on the rotation speed n_hAnd correct the course of said variation, the edge times can be determined in the same way based on the above-described method (as defined above for the respective phase voltages U)_U，U_V，U_WTime between two zero crossings).

Fig. 12 shows an uncorrected speed profile n (t) by a dashed line and a corrected speed profile n based on the method described above by a solid line_Corr2. It is clear here that the higher-order frequency components f, which can be recognized as high-frequency serrations in the tachometric signal n (t), have been correspondingly filtered out_n. The time window t associated with the correction factor is also shown in the circled area in fig. 12_linThe course of the rotational speed change over time in the time window is not very dynamic, preferably extends approximately constantly.

It is generally easy to understand that the method described above is not limited to the use of the described three-phase motor 30, but can be applied correspondingly to any motor 30, which can be permanent-magnet or otherwise excited, wherein, as already mentioned above, the number of phases can be taken into account correspondingly in the method.

Furthermore, the method steps described above for correcting the rotational speed can be used individually, but also in combination with one another, wherein in the case of combination a still better correction of the rotational speed value can be achieved. Alternative uses of the time of day associated with the rotational speed or its course can also be used to correct the rotational speed signal.

Claims

1. For determining rotational variables, in particular the rotational speed (n) and the rotational angle position (17 ') of a shaft (17), in particular a crankshaft (17') of an internal combustion engine (112)

，

) Or direction of rotation (α)₊，α_-) The internal combustion engine is directly or drivingly coupled to an electric machine (30) comprising a rotor (32) and a stator (33) having at least two phase windings (U, V, W) from which at least one phase signal (U) is derived in each case_U、U_V、U_W、I_U、I_V、I_W) Wherein the phase signal (U)_U、U_V、U_W、I_U、I_V、I_W) Respectively have a rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) And/or zero crossing (Fl)_U0、Fl_V0、Fl_W0) Wherein a plurality of phase signals (U) comprising the motor (30) are generated_U、U_V、U_W、I_U、I_V、I_W) Sum signal (U)_Sum) So that at said sum signal (U)_Sum) Respectively giving said rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) And/or zero crossing (Fl)_U0、Fl_V0、Fl_W0) Assigning a characteristic pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) Wherein said pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) Is used to determine the rotational speed (n) and the rotational angle position (n, b) of the rotor (32)

，

) Or direction of rotation (α)₊，α_-）。

2. Method according to claim 1, wherein at said sum signal (U)_Sum) Respectively by corresponding phase signals (U)_U、U_V、U_W、I_U、I_V、I_W) Falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) Generating a first pulse (P)_f1、P_f2、P_f3) But also by corresponding phase signals (U)_U、U_V、U_W、I_U、I_V、I_W) Rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Generating other pulses (P)_r1、P_r2、P_r3）。

3. The method of claim 1 or 2Method, wherein is assigned to have a rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Or falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) Has a constant first pulse width (T)_Pf1、T_Pf2、T_Pf3) While the pulses assigned to the respectively other edge type have a constant, but mutually different, other pulse width (T)_Pr1、T_Pr2、T_Pr3）。

4. Method according to one of the preceding claims, wherein the sum signal (U) is determined by determining the sum signal (U)_Sum) In which there is a first pulse width (T)_Pf1、T_Pf2、T_Pf3) One determines that there is a falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) And in determining the sum signal (U)_Sum) Among other pulse widths (T)_Pr1、T_Pr2、T_Pr3) One of them is asserted to have a rising edge (Fl)_Uu、Fl_Vu、Fl_Wu）。

5. Method according to one of the preceding claims, wherein the sum signal (U) is determined by determining the sum signal (U) from the signals_Sum) Two pulses (P) in_f1、P_f2、P_f3、P_r1、P_r2、P_r3) At least one time difference (Δ t) therebetween₁、Δt₂、Δt₃) To determine said rotational speed (n).

6. Method according to one of the preceding claims, wherein the pulse (P) is determined by determining the pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) To infer a first rotational direction (α), at least one first time series (A1)₊) Or by determining the pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) To infer another, corresponding to said first rotation (A2)Steering direction (α)₊) Different directions of rotation (α)_-）。

7. Method according to one of the preceding claims, wherein the pulse width (T) is determined by determining the pulse width (T)_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3) To determine the corresponding pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) And according to said pulses (P) using the characteristics of said motor (30)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) To determine the rotational angle position of the rotor (32) ((

）。

8. The method according to claim 7, wherein further two adjacent pulses (P) are used_f1、P_f2、P_f3、P_r1、P_r2、P_r3) And at a position (P) and said pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) At least one time interval (Δ t) between one of them_Pf1、Δt_Pr2) Determining the rotational angle position of the rotor (32) ((

）。

9. Method according to one of the preceding claims, wherein the determined rotational speed (n) of the shaft (17) is used for controlling the internal combustion engine (112), in particular for controlling the ignition and/or injection of fuel into at least one cylinder of the internal combustion engine (112).

10. For transmitting a plurality of phase signals (U)_U、U_V、U_W、I_U、I_V、I_W) Generating a sum signal (U)_Sum) Can be used to determine a rotational variable, in particular a rotational speed (n), a rotational angle position (c), of a shaft (17), in particular a crankshaft (17') of an internal combustion engine (112)

，

) Or direction of rotation (α)₊，α_-) The internal combustion engine is directly or drivingly coupled to an electric machine (30) comprising a rotor (32) and a stator (33) having at least two phase windings (U, V, W) from which at least one phase signal (U) can be derived in each case_U、U_V、U_W、I_U、I_V、I_W) Said phase signal having a rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) Falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) And/or zero crossing (Fl)_U0、Fl_V0、Fl_W0) Said device having at least two inputs (U)₁、U₂、U₃) Phase signals (U) are respectively transmitted through the at least two input terminals_U、U_V、U_W、I_U、I_V、I_W) Conducting to a coding unit (80 c), wherein the coding unit (80 c) transmits the individual phase signals (U)_U、U_V、U_W、I_U、I_V、I_W) Combined into a sum signal (U)_Sum) And the phase signal (U) is converted into a phase signal (U)_U、U_V、U_W、I_U、I_V、I_W) Is encoded such that at least one of the signals (U) is added to the sum signal (U)_Sum) Respectively to said at least one phase signal (U)_U、U_V、U_W、I_U、I_V、I_W) Rising edge (Fl)_Uu、Fl_Vu、Fl_Wu) And/or falling edge (Fl)_Ud、Fl_Vd、Fl_Wd) Assigning a characteristic pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) Wherein the coding unit (80 c) sums the signals (U)_Sum) Output to output terminal (U)_Out) For determining said rotational variable.

11. The apparatus of claim 10, wherein the encoding unit (80 c) has at least one time-lag element (T)₁₁、T₁₂、T₁₃) The time-lag element can be assigned to at least one phase signal (U) of the electric machine (30)_U、U_V、U_W、I_U、I_V、I_W) Said time-lag element giving said at least one phase signal (U)_U、U_V、U_W、I_U、I_V、I_W) The characteristic features are loaded.

12. The apparatus according to claim 10 or 11, wherein said at least one time-lag element (T)₁₁、T₁₂、T₁₃) Respectively having a resistor (R3, R6, R9) and a capacitor (C1, C2, C3), wherein the characteristic feature, in particular the pulse (P)_f1、P_f2、P_f3、P_r1、P_r2、P_r3) Pulse width (T) of_Pf1、T_Pf2、T_Pf3、T_Pr1、T_Pr2、T_Pr3) The size can be determined by the selection of the respective resistor (R3, R6, R9) and/or capacitor (C1, C2, C3).

13. The device according to claim 11 or 12, wherein a trigger circuit (80 a, 80 b) is preceded by the coding unit (80 c), the trigger circuit generating a triggered phase signal (U)_UT) The triggered phase signal is delivered to an input (U) of the coding unit (80 c)₁、U₂、U₃) At least one input terminal.

14. The apparatus of claims 10-13, wherein the trigger circuit (80 b) has a time-lag element (T) having a resistance (R)_1c) And a capacitor (C)_1c）。

15. A computing unit, preferably an engine control device (122) for an internal combustion engine (12), which is set up by a corresponding integrated circuit and/or by a computer program stored on a memory as: performing the method according to one of claims 1 to 9.

16. A computer program which, when executed on a computing unit, causes the computing unit to perform a method according to one of claims 1 to 9.

17. A machine readable storage medium having stored thereon a computer program according to claim 16.