Faulty load detection for multi-phase electric motor

Abstract

A method for obtaining an indication of a faulty load condition of a multi-phase electric motor includes: (a) starting, a time measurement unit, (b) measuring a recirculation time interval for as long as a current that continues to flow has a magnitude larger than a threshold value, (c) continuing, in case the recirculation time interval is not terminated during the switch-off interval, the time measurement at least in a next switch-off interval, (d) repeating, for additional switch-off intervals the steps (a), (b), and (c) for respective motor phases, (e) comparing (1) the measured recirculation time intervals for respective motor phases with each other and/or (2) for one motor phase in sequential succession with each other and/or (3) with the expectation value of the respective motor phase, and (f) determining, based on a deviation, the indication of the faulty load condition.

Claims

1. A method for obtaining an indication of a faulty load condition of a multi-phase electric motor with electric commutation driven by an electric drive unit comprising for each motor phase of the multi-phase electric motor, a respective high-side switch and a respective low-side switch, wherein: during operation of the multi-phase electric motor, the respective high-side switch and the respective low-side switch of each of the motor phases are cyclically switched on and off according to a switching scheme comprising cyclically recurrent time points, at the cyclically recurrent time points, for at least one of the motor phases, both the respective high-side switch and the respective low-side switch of the motor phase are switched to a respective high-ohmic state for a switch-off interval, and in case of the faulty load condition, a current continues to flow for a recirculation time interval equal to an expectation value valid for a fault-free operation of the multi-phase electric motor, the method comprising: (a) starting, at a start of the switch-off interval, a time measurement unit, (b) measuring the recirculation time interval for as long as the current that continues to flow has a sign-independent magnitude that remains larger than a predetermined threshold value, (c) continuing, in case the recirculation time interval is not terminated during the switch-off interval, the time measurement at least in a next switch-off interval until the recirculation time interval is terminated, (d) repeating, for additional switch-off intervals after a preceding high-side switch activation and/or after a preceding low-side switch activation of each motor phase, steps (a), (b), and (c) for the respective motor phase, (e) comparing (1) the measured recirculation time intervals for respective motor phases with each other and/or (2) the measured recirculation time intervals respectively for one motor phase in sequential succession with each other and/or (3) the measured recirculation time intervals with the expectation value or the expectation value range for the respective motor phase, and (f) determining, based on a deviation from the respective expectation value, the indication of the faulty load condition.

2. The method according to claim 1, wherein determining the faulty load condition is based on a type of the deviation from the respective expected value for the individual motor phases among each other.

3. The method according to claim 2, wherein the type of the faulty load condition is at least one of a shunt of a first motor phase toward a second motor phase, a shunt of a motor phase toward ground, a shunt of a motor phase toward a supply voltage, a too high-ohmic motor phase connection, a loose contact of a motor phase, a faulty resistance of a high-side switch or a low-side switch, a loose contact, a mechanical defect, a play of a bearing, a malfunction in a transmission, and a mechanical error in an application.

4. The method according to claim 3, further comprising: identifying, based on a frequency of a sequentially occurring deviation, a site of the faulty load condition.

5. The method according to claim 1, wherein, an electrical parameter representing a magnitude of the current is measured.

6. The method according to claim 5, wherein the measured electrical parameter includes a voltage drop across an electric component.

7. The method according to claim 6, wherein the electric component is at least one of a shunt resistor, a high-side switch and a low-side switch.

8. The method according to claim 1, wherein, in the switch-off interval, the time measurement is terminated by a comparator when the voltage of the motor phase rises above or falls below a preset comparator threshold value.

9. The method according to claim 8, wherein, onto a switch-off interval, there is applied a preset test current opposite to the recirculation current, wherein, in accordance with the preset test current, a voltage change is performed at a phase connection upon reduction of the recirculation current to the preset test current.

10. The method according to claim 1, wherein the time measurement unit operates linearly.

11. The method according to claim 10, wherein the linear operation of the time measurement unit is based on a constant counting speed.

12. The method according to claim 1, wherein the time measurement unit operates logarithmically.

13. The method according to claim 1, wherein the time measurement is stopped for a period between individual switch-off intervals.

14. The method according to claim 1, wherein the time measurement continues between individual switch-off intervals.

15. The method according to claim 1, wherein, there is performed, at a fixed time point after the start of the switch-off interval, an examination for an exceeding or a falling-short of a presettable comparator threshold value and, based thereon, the time measurement is performed.

16. The method according to claim 1, wherein, time intervals are measured between individual instances of exceeding or falling-short of comparator threshold values, and the time intervals of the individual motor phases are compared among each other and/or the time intervals of each motor phase for itself are compared in temporal succession, and are compared to the expectation value valid for fault-free operation of the motor, wherein the deviation of the amount of the present time intervals among each other and/or from the expectation value is evaluated as being the indicator of the faulty load condition.

17. The method according to claim 1, wherein expectation values for maximally allowable deviations beyond which the faulty load condition is detected are obtained by training on one or a plurality of pre-aged reference systems with respective fault-free load conditions, such that typical values are detected and, provided with selectable additional allowable tolerances, are stored in the drive unit as a maximum allowable expectation value.

18. The method according to claim 1, wherein expectation values for maximally allowable deviations beyond which the faulty load condition is detected, are obtained after production, by training, on each motor system for itself with a fault-free load condition, such that typical values are detected and, provided with selectable additional allowable tolerances, are stored in the drive unit as a maximum allowable expectation value range.

19. The method according to claim 1, wherein tolerances for the expectation values for maximally allowable deviations beyond which the faulty load condition is detected, are, in the course of the lifespan of the motor, increased.

20. The method according to claim 1, wherein, in a case that the detected faulty load condition is generated by a deviation of a commutation angle from a target range, the drive unit then compensates for the detected faulty load condition by adapting at least one drive parameter, the drive parameter selected from a set of amplitude and phase.

Description

DESCRIPTION OF THE DRAWINGS

(1) The disclosure will be explained in greater detail hereunder by way of examples and with reference to the drawings. In the individual views of the drawings, the following is shown:

(2) FIG. 1 shows an option for the circuitry for a three-phase electric motor of a generally freely selectable design,

(3) FIG. 2 illustrates, in partial view, a sinus or space vector commutation in the region of the sign change of the current flowing through a driver,

(4) FIG. 3 illustrates, by way of example, a block commutation of a three-phase motor,

(5) FIG. 4 shows the circuitry of an exemplary three-phase motor in the special case that no individual current measurement devices exist for all motor connections,

(6) FIG. 5 shows an example of a switching scheme (in this case for a three-phase motor) wherein the switching scheme does not include high-ohmic phases,

(7) FIG. 6 shows illustrations of the effects of a fault situation on the current measurement in case of a faulty current against ground,

(8) FIG. 7 shows illustrations of the effects of a fault situation on the current measurement in case of a faulty current against the positive supply potential,

(9) FIG. 8 shows a further option for the circuitry for a three-phase motor of a freely selectable design, provided with drain-source voltage monitoring devices on the high-side and low-side switches, wherein these voltage monitoring devices can be used in a phase-wise manner for the present current measurement.

(10) FIGS. 9 and 10

(11) show diagrams-explaining information on the switching option according to FIG. 8 respectively for the case that no short circuit occurs during current measurement (FIG. 9) and for the case that a short circuit does occur during current measurement (FIG. 10),

(12) FIG. 11 shows an option for the circuitry for driving a multi-phase (in the exemplary embodiment, three-phase) electric motor of a generally freely selectable design,

(13) FIG. 12 shows an example for the block commutation of an e.g. three-phase motor type (e.g. according to FIG. 11),

(14) FIG. 13 shows an option for the circuitry for a bipolar electrically commutating motor (e.g. stepper motor), and

(15) FIG. 14 shows an example for the block commutation of a bipolar electrically commutating motor (e.g. according to FIG. 13).

DESCRIPTION

(16) The disclosure is based on the recognition that, when driving inductive loads such as e.g. electric motors, it is detected by means of a comparator whether a transistor in the switched-on state comprises a positive voltage drop (in case of load) or a negative voltage drop (in case of inductive feedback).

(17) This is an indicator of the direction of the current flow through this driver (switch). In inductive loads such as e.g. motors, the time point of the reversal of the current direction is of course determined, on the one hand, by the modulation of the driving of the motor but, on the other hand, by the behavior of the coil current which, as is known, follows the modulated voltage. The expectation value as to when a change from a positive to a negative voltage drop should take place, i.e. when the current through the respective driver has sunk to zero, is preset by the PWM modulation of the motor and by the motor parameters.

(18) If, now, the site of the change of sign of a connection (motor phase) deviates in a reproducible manner from that of the other connections (motor phases) or from predetermined expectation values, there has to be assumed a fault situation in the motor or the connections, particularly if, when observing these changes of sign over time, the individual motor phases differ from each other in an atypical manner. The type of the deviation is an indicator as to whether, and at which connection, which kind of short circuit (shunt after ground or toward the positive supply potential) or which kind of a high-ohmic state exists.

(19) In the individual case, it can be sufficient to monitor only the low-side drivers or only the high-side drivers as has been explained above. In case of a corresponding sensitivity of the measurement device, the above described signals can also be used for the commutation of the motor.

(20) A possible switching scheme for a three-phase electric motor with electric commutation is shown in FIG. 1. The motor BLDC is driven by a driver full bridge, wherein each motor phase U, V and W has assigned to it a switch pair comprising a high-side switch U.sub.H, V.sub.H and respectively W.sub.H, and a low-side switch U.sub.L, V.sub.L and respectively W.sub.L. The voltage drop across each switch is monitored by means of the comparators KU.sub.H, KV.sub.H, KW.sub.H, KU.sub.L, KV.sub.L and KW.sub.L. In this arrangement, each motor phase U, V and W has assigned to it a comparator or a pair of comparators. With the aid of these comparators, it is possible to detect when the direction of a current flowing through a switch is reversed. Thereby, the zero crossing of the current through a switch can be detected. Owing to the construction and the driving, the zero crossing is within an expectation value range (expectation time slot). By examining the time points of the zero crossings as seen across a plurality and respectively all of the motor phases, and respectively as seen within one motor phase, it will then be possible to conclude on faulty load conditions. In as far as the time points of zero crossingsthat naturally deviate from phase to phase or within a phasefollow a reproducible pattern, this can be brought into connection with design-inherent asymmetries of the motor and the following loads. Non-reproducible and particularly non-predictable deviations of the time points of zero crossings from the expectation value ranges allow for conclusions on faulty load conditions. Thus, such deviations have to be interpreted as being a first indicator of a faulty electrical or mechanical load condition of the motor.

(21) FIG. 2 shows a commutation option for a multi-phase electric motor. In the present case, there is illustrated a sine and respectively space vector commutation, notably within that portion of the commutation development in which the current-direction sign changein this example, of the phase Uoccurs. The thin curve in FIG. 2 represents the theoretical voltage in a situation where the driver switches are driven with a randomly selected low-ohmic value. The thicker curve represents the voltage development under consideration of the voltage drop across the respective driver switch.

(22) The basis of the disclosure is the detection of temporal differences between respective fixed time points in the commutation scheme and the respective present time point of reaching a specific voltage drop across the individual drivers. The fixed time points should suitably be situated before a zero crossing. A special case herein is the reaching of the current I=0, which corresponds to a voltage drop of 0 Volt. However, also any other current value which will be reliably reached during the commutation can be used as a trigger point for the end of the time measurement. There is also possible a detection of the temporal differences before reaching the desired voltage threshold values in a motor connection until reaching the same conditions in the next motor connection. In this case, there exists no fixed starting time point for time measurement.

(23) A preferred variant for carrying out the detection of the recirculation time interval resides in a logarithmic time measurement. The logarithmic time measurement follows a function that is approximated to a logarithm, notably in such a manner that the speed of the counter will decrease with increasing counting time. This has the following advantages: a) Long and short time periods can be detected with the same relative precision. There is no unnecessarily high accuracy in measurements of large absolute time periods. b) The number of bits to be evaluated per measurement value is drastically reduced. c) Due to the logarithm formation, the detection of time relationships which normally require a multiplication-or-division-type calculation can be realized by an addition-or-subtraction-type calculation. This will reduce the hardware and software expenditure in the comparison operation. d) It is possible to perform a less expensive realization of the evaluation either by a smaller logic and/or by a CPU-time-saving realization in a controller.

(24) FIG. 3 shows, by way of example, the current feed scheme in a block commutation of a three-phase motor of a generally freely selectable design. However, the method of the disclosure is also applicable in sine or space vector current feed or in similar modulation types and, apart therefrom, also for other motor types such as e.g. bipolar or unipolar stepper motors.

(25) It is common to the described circuitries and motors that, according to the disclosure, the present current is measured at fixed phase angles (relative to 360 per electric rotation) and that the current measurement values of the different connections are compared to each other at the same individual phase angle of the corresponding connection. At different time points where the previously set fixed phase angle relationships exist, the currents and respectively current relations of the various connections relative to each other are compared to each other for reproducible deviations from the expectation value range. Asymmetrically designed motors can e.g. lead to different expectation value ranges which are not identical for all phases. The method of the disclosure provides a possibility to compensate for this effect. Further, it is provided that, in such motors, the expectation value ranges can be adapted also beyond one mechanical rotation which can comprise a plurality of electrical rotations by 360. Thereby, when applying the disclosed method, high precision and protection from faulty activations is obtained in all motor types. Illustrated in FIG. 3 are the switch-on and switch-off states of the high-side and low-side drivers U.sub.H, V.sub.H and W.sub.H as well as U.sub.L, V.sub.L and W.sub.L of the three motor phases U, V and W. By analysis of the currents and current relationships using the disclosed method, it is achieved that already small, permanent and reproducible asymmetries are an indicator of possible faulty currents, temporary asymmetries that are generated as disturbances, e.g. due to load change variations or variations on the current feed level or due to loose contacts, can be separated, by means a downstream-connected logic, from the faulty currents which may lead to an inadmissibly high stress and damage of the electronics, in case of a faulty current W after a positive supply potential, dV will increase in the state W.sub.LON and then, in the state W.sub.HON, no influence will exist anymore, and in case of a faulty current W after V, dV will be higher in comparison to other combinations, notably in the states W.sub.HON and V.sub.LON, or also W.sub.HON and V.sub.HON.

(26) FIG. 4 shows a circuitry option for three-phase motors of a generally freely selectable design for the special case that, for none of the motor connections, there exists an individual current measurement device assigned to it. The three-phase motor can be e.g. a BLDC or a stepper motor.

(27) In FIG. 4, the high-side drivers and the low-side drivers assigned to the three motor phases U, V and W are designated by U.sub.H, V.sub.H and W.sub.H and respectively by U.sub.L, V.sub.L and W.sub.L. The motor itself is designated by BLDC. Represented by interrupted lines are three potential events that generate a faulty current, notably a faulty-current-generating shunt of the phase W toward ground (see at R.sub.NSM), a faulty-current-generating shunt of the phase W toward the positive supply potential (see at R.sub.NSP), and a short circuit between the phases V and W (see at R.sub.VW). In this exemplary embodiment, the only current measurement device is arranged in the low-side driver path and is represented by a shunt R.sub.Shunt. Alternatively, however, the shunt resistor could also be arranged in the high-side path. Instead of a shunt resistor, there can also be used any other current measurement device. It is considered to be particularly advantageous if the current measurement is performed under consideration of the voltage drop across the drivers.

(28) FIG. 5 shows a switching scheme which includes a high-ohmic phase, as exemplified by a (random) three-phase motor. The individual intervals of a switching cycle are designated as phases 0 to 5. The high-side and low-side switches are designated in the same manner as in the previous Figures. The voltage developments of the three phases are designated by V(U), V(V) and V(W), wherein, in the two lowermost diagrams, the correspondingly marked transition ranges of the voltage V(W) of the phase W are represented on an enlarged scale.

(29) According to this variant, possible time points of the current measurement are time points where no direct current portion is flowing through a motor connection. Detection options with respect to such time points are: a) all detection time points of a possibly existing BEMF (back EMF) signal (see V(U), V(V), V(W) curves in FIG. 3), b) the aspect as to whether the signal is within the limits V.sub.th+ and V.sub.th, whereby, thus, the reaching of the threshold value V.sub.th+ or V.sub.th is used as a timer (see the enlarged representation in the two lowermost diagrams of FIG. 3), c) it being possible to add suitable times to the time points under a) and b) before the measurements are initiated, wherein, at the time points according to a), b) or c), the current is detected at least at one of the other motor connections.

(30) The current values of suitable phases of the commutation scheme are compared to each other. The differences or relationships of the current values relative to each other will be examined for reproducible, repetitive deviations and will be evaluated as described above.

(31) In FIGS. 6 and 7, there are shownunder the assumption of the scheme of the block commutation as already described abovethe effects that a fault condition has on the current measurements. In this regard, it is assumed that, for none of the motor connections, there exists a separate current measurement device (see the situation according to FIG. 4). The individual current developments of the phases U, V and W and across the shunt resistor are shown in the respective lower part of FIGS. 4 and 5 at I.sub.U, I.sub.V, I.sub.W and respectively I.sub.Shunt. The time points where, per phase, the current across the shunt resistor is detected, are designated by T.sub.Shunt.

(32) Shown in FIG. 6 is the fault case of a faulty current of the phase U toward ground. Without such a shunt of the phase U toward ground, there will occur, in the current measurement time periods and also in the other time periods, a respective phase current development according to the interrupted line, while the shunt will lead to a current development as shown in the lower part of FIG. 6 by a continuous line.

(33) FIG. 7 shows a corresponding situation, there being represented here the fault case of the shunt of the phase U against the positive supply potential. Without such a shunt, there will occur the current development represented by the interrupted lines, and the shunt case will result in the current development represented by the continuous line on FIG. 7. T.sub.Shunt in turn designates the current measurement time periods.

(34) FIG. 8 shows the case of a circuitry option for a (random) three-phase electric motor M with the phases U, V and W wherein a current monitoring device SU is provided for each high-side and each low-side switch. Each current monitoring device comprises a comparator which is respectively assigned to one of the driver switches. Thus, the group of high-side switches U.sub.H, V.sub.H and W.sub.H have assigned to them the comparators KU.sub.H, KV.sub.H and KW.sub.H which together, via a digital/analog converter DA.sub.H, receive a digital reference value Ref(HS) against which the current will be compared. In a similar manner, the two low-side switches U.sub.L, V.sub.L and W.sub.L have assigned to them a separate comparator KU.sub.L, KV.sub.L and KW.sub.L, which again, via a common digital/analog converter DA, will receive a digital reference value Ref(LS) against which the current will be compared. A control unit SE is operative to drive the digital/analog converters DA.sub.H and DA.sub.L. Also the high-side switches and respectively the low-side switches are driven, each time together as a group, by the control unit SE. Via the outputs of the capacitors, the control unit SE receives signals which possibly indicate an overcurrent flowing through one of the switches. Each time the drivers are switched on, the references Ref(HS) and Ref(LS) will be set to their usual short-circuit detection value so as to be able to detect low-ohmic short circuits in the usual manner.

(35) The digital/analog converters DA.sub.H and DA.sub.L are subsequently used, together with the overcurrent comparators KU.sub.H, KV.sub.H, KW.sub.H KU.sub.L, KV.sub.L and KW.sub.L with the output signals OC(x), as current measurement devices. At a suitable measurement time point, as described above, the amounts of the references are successively reduced. The value Ref(HS) or Ref(LS) which then will sooner or later lead to activation of a comparator OC(x), is further processed as a current measurement value. In the process, no driver switch-off occurs, while, conversely, after activation of the comparator, the usual overcurrent threshold value (short-circuit protection) will again be activated at Ref(HS) and respectively Ref(LS). Thus, if a short circuit should occur during the reduction, then a) the respective comparator would be immediately activated, b) immediate switch-over to the normal short-circuit detection threshold value would be performed, and c) a subsequent short-circuit switch-off would be performed.

(36) In FIGS. 9 and 10, this is again graphically represented, by way of example of current developments on a high-side driver, for the case that no short circuit occurs during the current measurement (FIG. 9) and for the case that a short circuit does occur during the current measurement (FIG. 10).

(37) The described method requires that its components have sufficient dynamics so that the delays in case of a short circuit will not be inadmissibly high. The advantage of using the overcurrent monitoring devices for current measurement (see FIGS. 6 to 10) is to be seen in the cost reduction effected by two-fold use of existing components while, at the same time, the overcurrent monitoring is free of interruptions. A preferred variant for carrying out the change of the switching threshold values of the comparators of the overcurrent monitoring devices is to be seen in a generator for the respective reference value which, within the downward ramp, includes a functionality approximated to a logarithm, such that the reference value from one step to the next will be reduced by a certain percentage. The advantage of this approach is that the ramp can be traveled through at a distinctly faster speed while the relative accuracy remains the same, which will allow for a measurement with shorter duty-cycle times. The number of bits to be evaluated per measurement value is drastically reduced. The multiplication-or-division-type calculation which is normally required for detection of the current relationships can, because of the logarithm formation, be imaged by an addition-or-subtraction-type calculation. It is possible to practice the evaluation in a cost-saving manner either by a smaller logic or by realization in a controller that will save CPU working time.

(38) If a switching scheme of a multi-phase inductive load (e.g. a multi-phase electric motor) includes intervals with high-ohmic final stages or if it is feasible to integrate these intervals into a switching scheme of a multi-phase inductive load, it is possible, when a load connection has been switched to a high-ohmic state, to detect, by means of a simple comparator, the time point of the sign change of the current after the high-ohmic switching state. In case that, in the switching scheme, there are usually no high-ohmic motor phases, the method of the disclosure provides that these be inserted shortly before the site where the value of the respective current reaches zero. The recirculation time as measured in accordance with the disclosure is a measure of the current existing in the inductive load (motor winding) at the time point of switching into the high-ohmic state, and also of the inductivity of the motor connection. Here, one can either compare the time points to the default values preset by the PWM control or, preferably, one can measure the time periods between the switching into the high-ohmic state and the activating of the comparator with respect to specific PWM combinations and compare the time measurements of the individual motor phases to each other.

(39) If the results of these comparisons deviate from expectation values, the type of the deviations allows for conclusions on different causes for faulty currents. Motors of an asymmetrical design can lead to expectation values which include asymmetries.

(40) Further, the expectation values may vary across a plurality of electric movement cycles, e.g. when a multi-pole motor has to pass through a plurality of electric cycles to perform a mechanical rotation. Here, one may obtain a cyclical pattern of comparative values (i.e. expectation values).

(41) Without faulty currents, said patterns of comparative values should occur. In case of shunts or other errors, the relations of the recirculation times relative to each other will deviate from the expectation values. With corresponding sensitivity of the measurement device, the above described signals can also be used for commutation of the motor. This can be realized to a large part by compact digital technology.

(42) FIG. 11 shows an example of the circuitry of an electric motorin the present embodiment, a three-phase electric motorwith the aid of a full bridge which comprises high-side switches U.sub.H, V.sub.H H and W.sub.H respectively assigned to the three phases U, V and W, and three low-side switches U.sub.L, V.sub.L and W.sub.L respectively assigned to these phases. The electric motor is designated by BLDC (brushless DC) and can be run e.g. in star or delta operation.

(43) A possible block commutation for such a three-phase motor with a circuitry according to FIG. 11 is shown in FIG. 12. The individual portions of an electric rotation by 360 are designated by 0 to 5. In FIG. 12, the on- and off-states of the high-side and low-side switches U.sub.H, U.sub.L, V.sub.H, V.sub.L, W.sub.H and W.sub.L are represented by 1 (for the switched-on state) and 0 (for the switched-off state). It can be seen that the switching and respectively commutating scheme comprises time sections in which individual motor phases are switched into the high-ohmic state.

(44) The temporal development of the voltages at the three phases of the motor is shown at V(U), V(V) and V(W). The transients in the voltage developments at the motor phases are represented at an enlarged scale in the last two diagrams. The transient times, i.e. the recirculation time periods (t.sub.u.sup.+, t.sub.u, t.sub.v.sup.+, t.sub.v.sup., t.sub.w.sup.+, t.sub.w.sup.) will be compared to each other from phase to phase of the motor and respectively across the phases or within a motor phase, in a continuous and/or intermittent and/or sporadic manner from time to time. From this comparison of the recirculation time periods, conclusions can be drawn on faulty conditions of the motor. If the design-based differences of the recirculation time periods occur with cyclic repetition, conclusions can be drawn on asymmetries of the electric motor. Sudden changes or other changes of the recirculation time periods that deviate from the above described pattern allow for conclusions on faulty load conditions. Thus, by the comparison of the recirculation time periods that is provided by the disclosure, there can be obtained a first indicator of a faulty condition of the motor.

(45) However, apart from the block commutation shown in FIG. 12, the method of the disclosure can be realized also in electric motors operating with sine or space vector commutation. In these commutation schemes, there normally do not exist high-ohmic states which would be sufficiently long for detection of the recirculation time period. However, a high-ohmic motor phase can be inserted in the region or the to-be-expected zero crossing. This high-ohmic motor phase can either have a fixed length or, preferably, it will be terminated after conclusion of the time measurement.

(46) FIG. 13 shows the circuitry of a bipolar electrically commutated motor with the high-side and low-side switches A1.sub.H, A2.sub.H, A1.sub.L, A2.sub.L, B1.sub.H, B2.sub.H, B1.sub.L and B2.sub.L. A corresponding commutation scheme for such an electrically commutated motor is exemplified in FIG. 14. Also here, it can be seen that, in the functional block A, recirculation voltages t(A1), t(A2), t(B1) and t(B2) are compared to each other. This comparison can be performed across the motor phases or, however, within a motor phase and respectively within each motor phase.

(47) A preferred variant of carrying out the detection of the recirculation time period is a logarithmic time measurement. The logarithmic time measurement follows a function that is approximated to a logarithm, notably in such a manner that the speed of the counter will decrease with increasing counting time. This has the following advantages: a) Long and short time periods can be detected with the same relative precision. There is no unnecessarily high accuracy in measurements of large absolute time periods. b) The number of bits to be evaluated per measurement value is drastically reduced. c) Due to the logarithm formation, the detection of time relationships which normally require a multiplication-or-division-type calculation can be realized by an addition-or-subtraction-type calculation. This will reduce the hardware and software expenditure in the comparison operation. d) It is possible to perform a less expensive realization of the evaluation either by a smaller logic and/or by a CPU-time-saving realization in a controller.