 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The present invention relates to an energizing circuit for a reluctance
motor.
Variable reluctance motors of the type for which the energizing system of
the invention is applicable demand repeated commutation between the
different phases in a predetermined sequence. Commutation is intended to
mean switching the voltage source from one phase winding to another, which
has to be done in order to keep the driving torque in a desired direction
of rotation during the rotation.
The usual way to control the commutation from one phase to another in
variable reluctance motors, particularly switched type drive motors, is to
use at least one rotor position sensor providing signals in dependence of
the rotation position of the rotor in relation to the stator, hereinafter
referred to as the rotor position. This means, however, that extra
elements must be provided at the rotor, which in practice have been
associated with some inconveniences and give rise to errors, particularly
in severe environments.
Motors working with variable reluctance are well-known and belong to the
group of brushless DC-motors. The type of variable reluctance motors for
which the invention is intended has a stator having one or several
excitation windings provided in one or several phases having separate
energizing of the windings belonging to a respective phase. Both the
stator and the rotor are normally provided with salient poles or teeth.
The rotor has no winding. The stator and the rotor provide a magnetic
circuit in order to generate a mechanical rotational torque, which is
essentially proportional to the square of the ampere-turns of the
energized winding and to the variation of the permeance to the rotor
position. The movement of the rotor in relation to the stator causes a
variation of the reluctance and consequently of the permeance in the
magnetic circuit of the stator winding.
A rotational torque in the driving direction of the motor is provided only
when the ampere-turns of the winding are maintained during a rotor
position interval, in which the permeance is increasing with the rotor
position change. Therefore, the intention is to keep every winding excited
only during such an interval for the winding. From driving technical
reasons it is suitable, but not necessary, to have only one stator phase
excited at the time, i.e. without overlap between the phase excitations.
Commutation from one phase to another can be made such that every phase
winding is connected during a rotor position interval, in which the
permeance is increasing with the rotor position change. The feeding to
each phase winding must in this case be disconnected or lowered during
every rotor position interval, in which the permeance is decreasing with
the rotor position change. As mentioned above the most common way is
therefore to use extra rotor position sensors to sense the rotor position
currently and to control the feeding of the voltage supply with a sensor
controlled circuit.
However, the desire is to get rid of these rotor position sensors. Several
attempts have been made to use the variation of current or voltage
characteristics of the stator windings and associated circuits to give an
indication on suitable times for connection and disconnection of the
voltage supply.
A problem inherent with the control of switched variable reluctance motors
of the driving type without any rotor position sensors is the desire that
the driving must be provided both at low and high rotation speed. Most of
the earlier systems having sensing of the motor characteristics to control
the excitation voltages for a motor have primarily been adapted to step
motors and operate well for relatively low rotation speed but operate
worse or not at all when the rotation speed exceeds 50% of the maximum
rotation speed of the motor.
BRIEF DESCRIPTION OF THE PRIOR ART
In one known reluctance motor control system, described in U.S. Pat. No.
3,980,933, there is sensed the electromotive force (EMF) induced in the
stator winding as the rotor rotates and when the output from a switching
device is non-conducting. The output of the switching device is made
conductive when the EMF reaches a predetermined level. This motor requires
a considerable bias current in order to afford interference margins. This
results in power losses. Furthermore, because no energizing current passes
through the winding during the time period in which the rotor position is
sensed, it is not possible to load the motor to its maximum. Consequently,
an exciting or energizing current must be passed constantly through one of
the phase windings.
In the case of another reluctance motor control system, described in U.S.
Pat. No. 4,520,302, means are provided for measuring the time during which
the flow of current through a stator winding rises or falls from one level
to another, in order to provide an indication of the rotor position. Thus,
solely the current flow is sensed during this control. This system, which
has been designed for controlling a stepping motor, affords poor
resolution at high speeds, e.g. speeds exceeding 50% of the maximum motor
speed.
SUMMARY OF THE INVENTION
The object of the invention is to provide a control for a variable
reluctance motor which has no rotor position sensor but which can still
drive the motor at all rotation speeds and loads.
According to the invention, a signal for indication of a determined rotor
position is created at rotational speeds above a given speed
.omega..sub.n. The signal is created from the current in and the voltage
across a phase winding without the motor drive being affected directly by
the measuring process. The signal is used in order to determine the exact
moment of commutation. Thus, a function is provided which at speeds above
.omega..sub.n exactly corresponds to the function which is provided when
rotor position sensors are used.
During the rotation of the rotor the inductance of each phase winding is
changed due to the permeance changing during rotation. The changing is
given in accordance with a given curve form in which a given rotor
position corresponds to a given inductance in the phase winding. According
to the invention, the time (corresponding to a given rotor position) is
detected for each phase winding when the inductance of the phase winding
reaches a predetermined value L.sub.k. A calculating unit then calculates
the time point, using the detected time point, when the commutation is to
be made where the voltage supply is disconnected from the phase winding
which up until then has been the driving one and is connected to the phase
winding in turn to be the driving one.
BRIEF DESCRIPTION OF THE FIGURES
The invention will now be described in more detail with reference to the
accompanying drawings, in which
FIG. 1 is a circuit diagram which illustrates one embodiment of the control
system according to the invention;
FIGS. 2-5 are diagrams which illustrate the manner in which the invention
operates;
FIG. 6 illustrates a circuit for supplementing the circuit shown in FIG. 1;
and
FIG. 7 is a further diagram which illustrates the manner of operation of
the invention.
DETAILED DESCRIPTION
FIG. 1 illustrates three phase windings L1, L2, L3 on the stator of a
three-phase reluctance motor. The invention is not restricted to the
number of phases of the motor, and three phases are illustrated solely by
way of example. The motor is driven by a d.c. voltage V, which may be 310
V for example, this voltage being normal for drive-type and switch-type
reluctance motors. Between ground and the positive terminal of the voltage
source, the phase winding L1 lies in a circuit which incorporates a
current measuring resistor R1, the emitter-collector-path of a power
transistor Ta1, the phase winding 1, and a power transistor Tb1. In this
circuit the lower transistor Ta1 is biased to saturation during the whole
of the time period over which the phase winding L1 is energized, while the
upper transistor Tb1 is pulsed during the energizing interval of the
phase, in the manner normal with switched reluctance motors, with which
each energizing pulse for one phase is divided into part pulses. The
process of dividing a drive pulse into part pulses does not form part of
the actual invention and will not therefore be described in more detail. A
diode Da1 is connected with its anode to ground and its cathode to the
part of the phase winding L1 remote from the transistor Ta1, in order to
maintain the current through the phase winding L1 throughout the whole of
the energizing pulse interval. A diode Db1 is connected with its anode to
the part of the phase winding L1 remote from the transistor Tb1 and the
positive terminal, in order to provide a current circuit in which current
can be rapidly drained from the phase winding L1 as soon as the transistor
Ta1 is deactivated, i.e. blocked.
Each of the remaining phase windings L2 and L3 are connected to a
respective circuit of the same kind. Thus, the phase winding L2 is
incorporated in a circuit comprising a current measuring resistor R2, two
power transistors Ta2, Tb2 and two diodes Da2, Db2, and the phase winding
L3 is incorporated in a circuit comprising a current measuring resistor
R3, two power diodes Ta3, Tb3 and two diodes Da3, Db3.
Switching of the transistors to their respective on-and off modes is
controlled by a control unit 1. The control unit is preferably a
microprocessor or microcomputer, although a circuit constructed from
conventional circuit components may be used instead. The control unit 1
has six outputs, each of which is connected, via a respective amplifier 2,
to the gating means of a respective transistor Ta1, Ta2, Ta3, Tb1, Tb2,
Tb3, for the purpose of gating each transistor individually.
According to the invention, there is determined for each phase winding the
timepoint t.sub.k which corresponds to the rotor position .theta..sub.k
when the inductance L in the phase winding during the rotor position
interval in which the inductance is increasing with the rotor position
change reaches a predetermined inductance value L.sub.k. In order to test
the condition L>L.sub.k, the voltage across the phase winding is
integrated, and the integrated value is compared to the product of the
value of the current strength in the phase winding and the predetermined
inductance L.sub.k.
The L>L.sub.k condition, hereinafter referred to as the fundamental
condition, can be derived in accordance with the following:
The differential equation
(U-R*i)=d(L*i)/dt (1)
describes the relation between the voltage U across and the current i in a
coil having the resistance R and the inductance L. It is presumed that the
coil in this case is some one of the phase windings in series with one of
the transistors Ta1, Ta2 or Ta3 and one of the current measure resistors
R1, R2 or R3, since the voltage drop across the transistor is negligible.
Then, the resistance R is composed by the resistance of the phase winding
and the resistance of the current measure resistor. The voltage U is
composed by the voltages across the phase winding, the transistor and the
current measure resistor.
The equation (1) is re-written as
##EQU1##
If one assumes that t.sub.0 is the time point for switching on the voltage
supply across the phase winding and that i(t.sub.0)=0, i.e. no current is
flowing in the phase winding at the switching on, the equation can be
re-written as
##EQU2##
where t.sub.1 is variable.
The equation is re-written as
##EQU3##
The desired condition L>L.sub.k is obtained as
##EQU4##
The equation (2) gives the aforementioned fundamental condition according
to the invention. In order to be able to check the condition according to
the equation (2), which for each phase needs to be tested only from the
time of switching on the phase to the time at which the fundamental
condition is fulfilled, the following circuit is provided for each phase:
For the purpose of sensing the voltage U continuously, a voltage divider
comprising two resistors Ra1, Rb1 and Ra2, Rb2 and Ra3, Rb3, respectively
are connected between ground and the upper end of the phase winding in
FIG. 1. The voltage divider has a relatively high total resistance, in
order not to load the phase winding circuit unnecessarily. A suitable
fraction of the voltage U, i.e. k.U is taken out over the respective
grounded resistor Ra1, Ra2 and Ra3. A multiplex unit 3 has an input
connected to the output of each of the three voltage dividers. The
multiplex unit 3 is controlled by the control unit 1 such as to couple to
the output of unit 3 the signal from the voltage divider for the phase
winding whose inductance is to be measured. This can only be effected with
one control conductor and the switching process thus takes place in a
cyclic sequence for each "1"-signal on the control conductor. In order to
reach the voltage U, the output signal from the unit 3 is divided with k
in a dividing circuit 4.
A multiplex unit 5 has three inputs connected to a respective resistor R1,
R2, R3, all of which have the resistance R'. The fourth input will be
described hereinafter. The multiplex unit 5 is controlled by the control
unit 1 synchronously with the control of the multiplex unit 3. In this
case, however, the control unit 1 produces a digital signal which
indicates which input shall be connected to the output.
The output signal from the multiplex unit 5 is applied to a multiplier 6,
which multiplies by the resistance R for the whole winding circuit divided
by the resistance R'. The output signal from the circuit 6 has the value
R1. The output of the dividing circuit 4 having the voltage U is connected
to the (+)-input and the output to the multiplier circuit 6 having the
voltage R1 is connected to the (-)-input of a differentiator 7, so that
the voltage on its output is (U-Ri). This signal is supplied to an
integrating circuit 8, which integrates the signal from the circuit 7
during times determined by the control circuit 1. The circuit 8 commences
a new integration with the control signal obtained from the control
circuit 1. This control signal is supplied to the circuit 8 each time a
phase is activated.
The output of the multiplex unit 5 is also connected to a multiplier 9,
which multiplies the signal R' with L.sub.k /R'. The output signal from
the unit 9 thus obtains the value L.sub.k.i, where L.sub.k is the
inductance to be detected.
The signals from the integrator 8 and the multiplier 9 are each fed to a
respective input of a comparator 10, which produces a "1"-signal on its
output as soon as the value of the signal from the integrator 8 exceeds
the value of the signal from the circuit 9.
When the control unit 1 receives a positive signal from the comparator 10
and this does not lie within an unpermitted time interval, as described in
more detail hereinafter, the fundamental condition is detected for the
actual phase. The control unit 1 can then directly send control signals to
the multiplex units 3 and 5 in order to switch them and to allow signals
from the next phase in turn to be sensed to pass through. Alternatively,
this switching can be provided in connection with the time of switching on
the voltage supply to this phase.
It should be observed, however, that the control unit 1 does not
immediately de-activate the phase in question upon the fulfillment of the
fundamental condition according to the equation (2). Instead, the control
unit 1 carries out calculations according to pre-set conditions, in order
to establish the time at which the commutation is going to be provided.
These conditions are dependent partly on the rotational speed .omega. of
the rotation and partly on the current strength i in the monitored phase.
This is described hereinafter with reference to the diagrams of FIGS. 2-5.
Consequently, the value of the current i must be fed directly to the
control unit 1. The output of the multiplex unit 5 is therefore connected
to an analogue/digital converter 11 and is fed to a separate input on the
control unit 1, which divides the value obtained by R', in order to obtain
the value of i. The value of i is also used when driving the motor within
the low speed range, as will be described in more detail hereinafter.
As beforementioned, the L.sub.k -condition, or fundamental condition, in
the equation (2) gives an L.sub.k -condition angle .theta..sub.k, which is
dependent on the current strength in the winding, i.e. in the phase
winding which has just been driven by the control unit 1. This is because
the inductance is not solely a function of the rotor position, but also of
the current i. This is illustrated in FIG. 2, which shows the variation of
the inductance L in one phase as a function of the rotor angle .theta.
over slightly more than one revolution. The diagram shows four different
curves l.sub.1, l.sub.2, l.sub.3, l.sub.4, of which each curve is drawn
for a separate constant current strength, where l.sub.1 is drawn for a
very low current strength and l.sub.4 is drawn for a very high current
strength. The current strengths, in turn, are contingent, inter alia, on
the extent to which the motor is loaded or driven. The fact that the curve
shape becomes flatter with increasing current strengths is because the
iron in the magnetic circuit of the motor is saturated to progressively
greater extents. It is not a constant current strength which determines
the actual curve shape of the inductance L, but the fact that the
inductance L varies in accordance with the prevailing or momentary current
i, although FIG. 2 illustrates how the inductance varies with current
strength.
It will be seen from FIG. 2 that .theta..sub.k for detected L.sub.k is
displaced to increasing rotor position angle at increasing current
strength. It will also be seen from FIG. 3 that at positive rotation
direction, when the rotor position angle is increasing with the rotation,
the time t.sub.off for disconnection of a phase winding switched on, which
time corresponds to the rotor position angle .theta..sub.off, will occur
after the time t.sub.k, which corresponds to .theta..sub.k in the figures.
Hereinafter it is presupposed that the rotational direction is positive,
i.e. that the rotor position angle is increasing with the rotation, and
that the angular speed is constant during the rotation position intervals
mentioned below due to the motion energy stored in the system. The angular
speed is constant because a rotor position change is to correspond to a
given time.
The angle .theta..sub.OFF, i.e. the rotor position angle at which
excitation or activation of the phase concerned is discontinued, should
also vary with the rotational speed of the motor. At constant
.theta..sub.OFF the motor obtains a series motor characteristics, i.e. the
power is inversely proportional to the speed. If .theta..sub.OFF is
permitted to decrease with increasing speeds, the motor obtains a flatter
torque/speed characteristic and it is possible to obtain a motor with,
e.g., constant power, by controlling .theta..sub.OFF. This is achieved by
introducing a further delay .theta..sub..omega. in addition to
.theta..sub.S prior to shutting down the phase at .theta..sub.OFF, this
delay depending upon rotational speed. FIG. 3 illustrates a curve shape of
the inductance L as a function of .theta. for a motor driven with a low
current and at high speed, whereas FIG. 4 illustrates a curve shape of
inductance L as a function of .theta. for a motor driven by a high current
and at high speed. Thus, it will be seen from FIG. 3 that both
.theta..sub.s, which is the current-dependent current compensation factor,
and .theta..sub..omega., which is the speed-dependent compensation factor,
are low, so that the time between .theta..sub.x and .theta..sub.OFF is
relatively long. FIG. 4 shows that both .theta..sub. s and
.theta..sub..omega. are short, and hence the time between .theta..sub.k
and .theta..sub.OFF will also be short.
The range of variation for the aforesaid compensation may be relatively
large in the case of a number of motors, in which case the fundamental
condition L>L.sub.k must be fulfilled at an early stage. It is even
conceivable that the compensating times will exceed the integration times
t. This is illustrated in FIG. 3 where the combined time of .theta..sub.s
and .theta..sub..omega. is greater than the time between .theta..sub.ON,
which represents the rotational angle at which the phase is activated or
excited, and .theta..sub.k. Since the drive circuit according to the
invention is intended to replace a sensor circuit for a reluctance motor,
and the control unit 1 is preferably a microprocessor, it is suitable in
one practical application to manufacture a reference motor which is
provided with normal sensor control and to investigate where the motor
equipped with sensor control de-activates the phase at .theta..sub.OFF, by
operating the motor at different speeds and at different loads, and
indicating .theta..sub.k and storing the delays between .theta..sub.k and
.theta..sub.OFF in a fixed memory in the microprocessor. As
beforementioned, it is convenient from the aspect of drive techniques to
permit .theta..sub.OFF to constitute a commutating angle at which
transition from the excitation of one phase to the excitation of the next
phase takes place immediately.
At high rotor speeds, commutation shall take place very early, and the
delay .theta..sub..omega. therewith moves towards 0. It is possible that
phase activation at .theta..sub.ON for the next phase L.sub.ON is greater
than the fundamental condition inductance L.sub.k. L.sub.ON may also be
greater than L.sub.k when the current level i in a preceding phase is high
at the time that commutation takes place, so that .theta..sub.s becomes
small. FIG. 5 illustrates this circumstance. The broken line curve
illustrates the variation of the inductance L as a function of .theta. for
i=0 and the full line curve illustrates the variation L as a function of
.theta. when i=i.sub.k =0.
When a phase is activated, i=0 is included in this phase. The inductance
then follows the broken line curve. At the rotation angle .theta..sub.ON,
the inductance in the phase winding is L.sub.ON. As will be seen from the
left-hand side of the diagram, L.sub.ON can be greater than L.sub.k and
also much greater than when the current i.sub.k flows in the winding (see
the full line curve).
The current i rises relatively slowly from the value 0 upon activation of
the phase at .theta..sub.ON, and is relatively low during the first part
of the activation period during which L.gtoreq.L.sub.min. The inductance
in this case roughly follows the broken line curve.
Thus, an interval of the L-curve can be found right at the beginning of a
phase supply, i.e. from .theta..sub.ON to .theta..sub.Q in FIG. 5, during
which the fundamental condition L>L.sub.k is fulfilled. It is important to
prevent the phase from being de-activated during this interval.
Consequently, in accordance with the invention, sensing of the fundamental
condition can be blocked for a given period of time subsequent to
activation, up to a rotational angle .theta..sub.B, spaced from
.theta..sub.Q by a good margin. This additional blocking feature can be
inserted, e.g., for speeds above a pre-set level and/or for currents
i.sub.k above a pre-set level. Such blocking may also be made a general
feature. Interference margins are also improved with such blocking, as
opposed to when blocking is not utilized, since it is then no longer
necessary to compare L.sub.k.i and .intg.(U-R.i)d.tau. for small drive
angles, during which both magnitudes are small.
As mentioned in the aforegoing, the L.sub.k -condition according to
equation (2) was derived under the proviso that the current i in the
winding must be zero amperes at the time of phase activation. This can
prove problematic since back emf flows in the phase winding long after a
phase has been deactivated at .theta..sub.OFF, as will be seen from the
right-hand part of the curve i in the FIG. 5 diagram.
In order to guarantee that there is no current in the phase winding at the
time of activation, phase activation is blocked in accordance with the
invention until it is sure that the current is zero. This current is not
measured across the current measuring resistors R1, R2, R3, when the phase
is unactivated, since the lower transistor Ta1, Ta2, Ta3, (FIG. 1) is
switched off prior to activation. A phase activation blocking function can
be obtained in at least two different ways.
The first method of obtaining a blocking function is illustrated in FIG. 6,
which shows the third phase winding L3 and associated energizing circuits
Ta3, Tb3, R3, Da3, Db3, together with the control unit 1, and the
amplifiers 2 in accordance with FIG. 1, and with additions to this circuit
for providing the activation blocking facility. Each phase winding circuit
in FIG. 1 shall be provided with the same addition means as those
illustrated in FIG. 6. This means includes a so-called "pull-up" resistor
R.sub.p which is connected between the upper part of the phase winding L3
and the positive terminal of the supply voltage. A voltage divider Rc and
Rd is connected in series with the "pull-up" resistor Rp and earth. The
outlet on the voltage divider Rc and Rd is connected to a comparator 12,
which has a reference voltage U.sub.ref coupled to its other input. The
voltage over the voltage divider is approximately OV for as long as back
emf prevails. Upon termination of the back emf, the voltage rises via
R.sub.p to approximately the level of the supply voltage V.
The comparator 12 will transmit a logic signal "0" for as long as back emf
continues, and transmits a "1"-signal as soon as the back emf is
terminated. Naturally, an equally good result can be obtained by sensing
the lower winding outlet on the phase winding with a so-called "pull-down"
resistor, or by using the same principle with some other winding
configuration, e.g. a bifilar winding. This has not been shown in the
drawing however.
The other method of blocking phase activation is one of utilizing the fact
that the back emf time is constantly shorter than the drive time at a
constant supply voltage V. This can be seen from the curve i in FIG. 5.
The phase winding is deactivated when the rotor angle is .theta..sub.OFF.
Subsequent hereto the current i falls along an exponential curve and
reaches the value 0 at the rotation angle .theta..sub.m. The time between
.theta..sub.OFF and .theta..sub.m is therewith always shorter than the
time between .theta..sub.ON and .theta..sub.OFF. Consequently, there is
measured the total time for which V is applied over the energizing circuit
of the phase winding concerned, i.e. the time between .theta..sub.ON and
.theta..sub.OFF, and activation of this phase is then blocked until back
emf has continued for the same period of time. Although this implies an
overestimation of the back emf time, it lacks all practical significance.
Practical tests have shown that this type of blockaging facility has no
practical influence on the acceleration of the motor.
The aforedescribed drive arrangement provides positive commutation under
normal motor operation. At very high currents, however, there is a chance
that commutation will be excluded due to non-fulfilment of the condition
according to equation (2). This will occur when the current is so high as
to prevent the inductance reaching to the level L.sub.k. This is
illustrated in FIG. 2 by the lower curve L.sub.4. The three inductance
curves l.sub.1, l.sub.2, l.sub.3 reach and exceed the level L.sub.k.
Commutation will then take place. The curve l.sub.4 fails to reach the
level L.sub.k. In this case there is no commutation. The problem is
partially solved by selecting L.sub.k at a low inductance and to combine
this lower value of L.sub.k with the aforesaid delays .theta..sub.s and
.theta..sub..omega., as illustrated in FIGS. 3 and 4. This provides a
satisfactory solution when a chopping function is incorporated for pulsing
the drive of the phase windings, as effected with the upper transistors
Tb1, Tb2, Tb3 in FIG. 1. This limits the current to a sufficiently low
value. There are times, however, when this is not sufficient, particularly
when the peak current is much higher than the maximum current with
commutation. Such a curve shape is illustrated in FIG. 5, where the
current i reaches a peak value i.sub.p, which lies far above the current
i.sub.k, measured at .theta..sub.k.
In order to guarantee that commutation will always take place even with
extreme overcurrents, there is utilized a high current limiting level.
This high current limiting level must always be found in a motor, in order
to protect the electronics thereof, and this level is now used in
accordance with the invention to provide a commutating signal in the
manner described in more detail herebelow.
As illustrated in FIG. 1, the current in the activated phase winding is
sensed continuously by the control unit 1 via the analogue/digital
converter 11. When sensing the lower current limiting value, the upper
power transistor Tb1, Tb2, or Tb3 is switched-off. The current i then
falls normally. Subsequent to switching-off the one power transistor in
the winding circuit, current is only able to rise when the inductance of
the motor decreases. If the current should rise subsequent to
switching-off the upper power transistor, this would mean that the
position of maximum inductance of the inductance curve has been passed
without commutation taking place and that there has been reached an
angular position which lies somewhere on the negative flank of the
inductance curve. The motor will then produce a drive-emf and the current
will increase. Thus, if the current increases subsequent to sensing
maximum current, commutation shall take place immediately.
The diagram in FIG. 7 illustrates an example of the shape of the curve
i(.theta.) in the case of commutation when L.sub.k is not sensed. At the
rotation position .theta..sub.c the overcurrent or excess current
i.sub.max is sensed. The upper power transistor Tb1, Tb2 or Tb3 is
therewith switched off. The current i(.theta.) then falls. The upper power
transistor is again switched-on at .theta..sub.d and the current rises and
again reaches i.sub.max at the rotational position .theta..sub.e, the
upper power transistor again being switched off. The inductance curve,
however, has now passed its maximum, as shown in the full line curve
L(.theta., i). The current i(.theta.) now no longer falls, but instead
rises relatively slowly. At the rotation position .theta..sub.OFF there is
sensed a current i.sub.D which is higher than i.sub.max. Commutation to
the next phase then takes place immediately.
Instead of sensing maximum current through the analogue/digital converter
11, it is possible as is usual in such circumstances, to provide an
additional comparator circuit constructed to sense excess current
separately and to indicate directly to the control unit 1 when excess
current or overcurrent is reached. There is therefore illustrated in
broken lines in the circuit shown in FIG. 1 a voltage dividing circuit
which includes three series-connected resistors Re, Rf, Rg between the
positive terminal on the voltage source V and ground. The outlet between
the resistors Rf and Rg is connected to a reference voltage input on a
comparator 13 and the outlet between the resistors Re and Rf is connected
to a reference voltage input on a comparator 14. The output from the
multiplex unit 5 is coupled to the (+)-input of both comparators 13 and
14. The voltage across Rg is so selected that the comparator 13 will
produce a "1"-signal as soon as the pre-set maximum permitted current
level i.sub.max has been reached. The reference voltage to the comparator
14 is somewhat higher than that for the comparator 13. Consequently, the
comparator 14 will produce a "1"-signal for a current level which is
slightly higher than the maximum permitted level i.sub.max. Each of the
outputs of the comparators 13 and 14 is connected to a respective input on
the control unit 1, which controls in the manner aforedescribed. As w | | |
