Note: Descriptions are shown in the official language in which they were submitted.
This invention relates to electron:ic controls for
electric motors, laundry machines including such controls
and/or methods of operating said controls.
It is an object of the present invention to provide an
electronic motor control for controlling electric motors
and/or a laundry machine including such controls and/or a
method of operating laundry machine using such controls
which will at least provide the public with a useful
choice.
Accordingly in one aspect the i~vention may broadly be
said to consist in a method of cyclically reversing an
electronically commutated motor having a plurality of
windings on a stator and a rotor having magnetic poles
rotatable relative to said stator and using electronic
control apparatus and means to indicate the position of the
rotor said method comprising the steps of
(a) Initiating and then continuing a correct sequence of
commutations for a desired time or desired number of
commutations,
(b) Removing all power from the windings and allowing the
rotor to coast towards zero rotation~
(c) Testing the position of the rotor relative to the
stator, and
(d) When the rotor is in condition to be reversed and it~
position relative to the stator is known, changing the
sequence of commutations to cause the rotor to change
direction the correct commutations following
automatically to maintain rotor rotation in the changed
direction,
-- 2--
1 and repeating the steps to give cyclical reversal for a
desired time.
In a further aspect the invention consists in control
apparatus for an electronically commutated motor having a
plurality of windings on a stator adapted to be selectively
commutated and a rotor having magnetic poles rotatable
relative to said stator said control apparatus comprising:
~a) Timing means to time the period of rotation or counting
means to count the number of rotations of the rotor in
a desired direction,
(b) Commutation switching means to disconnect power from
said windings to allow the rotor to run down towards
zero rotation,
(c) Detecting means to indicate rotor position relative to
said stator, and
(d) Pattern reverse means operable in response to a signal
from said detecting means when the rotor is in
condition to be reversèd to cause the control signals
to cause commutation changes which cause said rotor to
change direction without testing for rotor direction~
In a still further aspect the invention consists in a
method of cyclically controlling the supply of power to an
electric motor having a rotor said method including the
steps of starting rotation of said rotor in one direction
setting an initial ~'power on" time during which power is
applied to said motor, switching off power at the end of
said initial ~power on~ time/ causing the rotor to slow
until in a condition to be reversed, checking the ramp down
1 time the rotor takes to slow to a condition ready for
reversal, causing reversal of direction of rotation of said
rotor, as soon as the rotor is in condition to be reversed,
and repeating the said steps as desired.
In a still further a~spect the invention consists in a
method of cyclically controlling the supply of power to an
electric motor having a rotor said method including the
steps of setting a desired time of rotation of said rotor
in one direction starting rotation of said rotor in said
one direction setting an initial "power on" time during
which power is applied to said motor, switching off power
at the end of said initial ~power on~' time, causing the
rotor to slow until in a condition to be reversed, checking
the ramp down time to rotor takes to slow to a condition
ready for reversal, causing reversal of direction of
rotation of said rotor, applying power to
said rotor for a further "power on" time which is such that
said further ~power on1' time plus said ramp down time
equals said desired time, switching off power to said rotor
at the end of said further "power on" time, again checking
the next ramp down time reversing direction of the rotor to
said one direction when said rotor is in condition for
reversal and applying power to said rotor for a still
further ~power on" time which is such that said still
further ~'power on~ time plus said next down ramp time
equals said desired time and repeating the cycles for a
desired length of time, adjusting the ~power on~ time at
desired intervals of time so that the adjusted "power on"
s
1 time for a further half cycle plus the down ramp time for a
previous half cycle equals said desired time.
In a still further aspect the invention consists in a
method of electronically cyclically controlling the supply
of power to an electric motor said method including the
steps of setting a desired speed of rotation of the rotor
of the motor, sensing the resistance to rotation of the
motor and using responses from the sensing means to actuate
adjustment means to adjust the power supplied to the motor
to change the motor speed towards said desired speed and
then operate the motor within a range of speeds at or close
to said desired speed of rotation, switching off the supply
of power to the motor, stopping its rotation and then
repeating the cycle of operations with the motor running in
the reverse direction.
In a still further aspect the invention consists in an
electrical control means for cyclically controlling the
supply of electrical power to an electric motor having a
rotor said control means comprising switching means to
switch power to said motor on and off, coasting timing
means to time the length of time said rotor takes from the
time power is switched off thereto to the time when said
rotor is in condition for reversal of direction of
rotation, and reversing means to reverse the direction of
said rotor when said rotor is in condition for reversing
and to switch on said switching means when reversing is to
be effected.
1 In a still further aspect the invention consists in an
electronic control means for cyclically controlling the
supply of electrical power to an electric motor said
electronic control means including setting means operable
to set a desired speed of rotation of the rotor of said
motor, sensing means to sense resistance to rotation of the
motor and adjustment means responsive to said sensing means
to adjust the power supplied to the motor to accelerate
said motor towards the desired speed and to then operate
the motor within a range of speeds at or close to said
desired speed of rotation, switching means to switch off
the supply of said motor after a desired time and reversing
means operable after the motor has substantially stopped to
cause the cycle of operating to be repeated with the motor
running in the reverse direction.
In a still further aspect the invention consists in an
electrical control means for cyclically controlling the
supply of electrical power to an electric motor having a
rotor said control means comprising switching means to
20 switch power to said motor on and off, power timing means
to time the length of power time when power is switched on,
coasting timing means to time the length of time said rotor
takes from the time power is swikched off thereto to the
time when said rotor is in condition for reversal of
25 direction of rotation stroke timing means to time the
stroke time during which said rotor rotates between
reversals setting means to set said stroke ~iming means to
a desired stroke time; algebraic subtracting means to
~l23''~5
1 algebracially subtract a previous coast time from said
stroke time to arrive at a time setting for said power time
and reversing means to reverse the direction of said rotor
when said rotor is in condition for reversing and to switch
on said switching means when reversing is to be effected.
In a still further aspect the invention consists in a
method of operating a lundry machine having a container for
a wash load of soiled fabrics in wash water and a
reciprocable agitator in said container and an electric
motor driving said agitator said method comprising the
steps of starting rotation of said motor in one direction
setting an initial "power on~ time during which power is
applied to said motor, switching off power at the end of
said initial ~'power on~' time, allowing the motor to slow
down until in a condition to be reversed, checking the time
between the power off condition and a condition when the
rotor is in condition to be reversed causing reversal of
direction of the rotor as soon as the motor is in condition
; for reversal and repeating the said steps as desired.
In a still further aspect the invention consists in a
method of operating a laundry machine having a container
for a wash load of soiled fabrics in water and a
reciprocatable agitator in said container, an electric
motor driving said agitator! setting means to set a desired
rate and amplitude of time and/or angle of oscillating
rotation of said agitator an electronic control means
controlling the supply of electrical power to said electric
motor in one of a plurality of sequences said method
1 including the steps of setting a selected one of said
plural.ity of sequences so that said agitator is driven in
oscillating rotation during a wash phase in a sequence of
washing operations, sensing the resistance to oscillation
of said agitator due to the wash load in said container and
adjusting the power supplied to said electric motor so that
a selected rate of removal of soil from said soiled fabrics
is substantially achieved.
In a still further aspect the invention consists in a
laundry machine including a container for a wash load of
soiled fabrics in water a reciprocatable agitator in said
container an electric motor driving said agitator, setting
means to set a desired rate and amplitude of oscillating
rotation of said agitator, electronic control means
controlling the supply of electrical power to said electric
motor in one of a plurality of selected sequences so that
said agitator is driven in oscillating rotation during a
wash phase, selecting means for selecting a desired one of
said sequences so that a washing action selected from such
20 as delicate, regular, heavy duty, wool, and permanent press
washing actions is to be e~fected by the machine said
electronic control means including sensing means to sense
the resistance to oscillating rotation o~ said agitator due
to the wash load in the container and adjustment means
25 responsive to said sensing means to adjust the power
applied to said electric motor so that a washing action
results such that a selected rate of removal of soil from
said soiled fabrics is substantially achieved.
1 To those skilled in the art to which the inYention
relates, many changes in construction and widely differing
embodiments and applications of the invention will suggest
themselves without departing from the scope of the
invention as defined in the appended claims. The
disclosures and the descriptions herein are purely
illustrative and are not intencled to be in any sense
limiting.
Preferred forms of the invention will now be described
with reference to the accompanying drawings in which,
Figure 1 is a block diagram of an electronic control
circuit to control an electronically commutated motor
driving an agitator and spin tub of a clothes washing
machine,
Figures 2 and 3 illustrate EMFs in windings with the
rotor rotating clockwise in relation to Figure 2 and
counterclockwise in relation to Figure 3,
Figure 4 is a diagram showing motor stator windings,
and electronic power commutation circuitry,
Figure 5 is a circuit diagram of a voltage digitising
circuit used in the invention.
Figure 6 is a flow diagram of motor reversing
sequences.
Figure 7 is a flow diagram of deriving values of index
and indexr,
Figure 8 is a flow diagram for determining the rotor
position,
_ g_
.
3~13S
1 Figure 9 is a graph showing the motor and hence
the agitator velocity profile during a half cycle of
agitator oscillating rotation in a wash mode.
Figure 9a is as Figure 9 but illustrating action
when the stoke time is variable,
Figure 10 is a graph showing a series of
acceleration profiles,
Figure 11 is a graph showing resultant curves
under operating conditions between the completion of the
acceleration mode and the cutoff point of applying power to
the motor,
Figures 12 to 16 are flow diagrams showing various
phases of operation of the control circuit of Figure 1,
Figure 16a on the sheet with Figure 15 is a
diagrammatic view of a speed sensor for use with the
invention,
Figures 17, 18 ar,d 19 are figures repeated from a
Boyd & Muller U.S. Specification 4,540,9~1 to provide
background to the present invention.
This invention relates in general to a laundry
machine with a cabinet, a wash water container in its
cabinet, a spin tub in the container reciprocating
agitator in the spin tub and a motor for driving the
agitator in the spin tub. Specifically it relates to
sensing means for sensing the load on the agitator
-- 10 --
1 and adjusting means operating in response to signals
from the sensing means to adjust the power by
adjustment of the profile of velocity to the agitator
as indicated by a velocity/time graph such that soil
removal and washing activity remain substantially
- lOa -
1 constant according to a desired setting for different
loads.
Laundry machines are required to wash a wide variety of
fabrics and garments. Different clothes and fabric types
require different treatment to achieve an appropriate wash
action. In general, with vertical agitator washing
machines, as agitator velocity is increased, soil removal
and wear and tear also increase. An appropriate balance
between soil removal and wear and tear is necessary It is
a major objective of laundry machines to wash each type of
fabric with an agitator action appropriate to the load type
and size. For example, clothes which fall into the broad
category of "delicates," often synthetic in origin, or
fragile items which are susceptible to damage during the
wash but which are typically only slightly soiled, require
gentleness of wash action with less emphasis on soil
removal, whereas "regular" items such as cottons which are
strong when wet can withstand a more vigorous wash action.
Conventional vertical axis laundry machines employ
various types of transmissions to convert rotary motion
provided by an electric motor into oscillatory motion at
the agitator for their wash mode. Such motors are
generally of essentially constant speed types. Therefore
to provide wash actions suitable for loads ranging from
delicate garments to heavily soiled hard wearing garments
requires multiple gearing or switched speed motors each of
which is costly. Further, as wash load is increased
towards rated capacity for a constant amount of water, mean
s
1 soil removal typically decreases and mean gentleness
increases. Variance of soil removal and gentleness also
increases, indicating less uniformity of wash action
throughout the wash load. Therefore it is difficult to
maintain good wash performance with laundry machines of
this type under varying load conditions.
The use of agitator drive systems such as disclosed in
the John Henry Boyd Australian Patent Specification AU-A~85
- 183/82 AND THE FISHER ~ PAYKEL United Kingdo~ Patent
10 UKN2095705 wherein the agitator may be directly driven by
an electronically controlled motor either with or without a
simple speed reduction unit and oscillatory rotation is
enabled by periodic reversal of rotation of the motor
provides opportunity for varying the speed and rate of
reversal of the agitator to obtain the appropriate balance
between soil removal and wear and tear for each category of
load. However the problem of variation of soil removal and
also wear and tear with load si~e still remains.
In a first aspect of the invention the following
describes apparatus to carry out an oscillatory rotation of
the agitator during a washing phase of the cycle of
operations of the washing machine and then on command to
spin the spin tub in a SpiD phase of the washing cycle, and
is principally concerned with the agitation cycle.
In a further aspect of the invention, later in this
specification a detailed description is given of preferred
forms of sensing means to sense the wash load in the
laundry machine, correcting means to correct for velocity
1 variations, adjusting means to adjust the power applied to
the agitator by modification of the profile of velocity as
indicated by a velocity/time graph, and setting means to
alter the stroke angle of the agitator such that soil
removal and wear and tear such that wash performance remain
substantially constant for a particular setting with
variation of load size.
The preferred form of the :invention is based on the
Boyd and Muller U.S. Specification 4,540,921.
For assistance in the full understanding of the present
invention excerpts from the Boyd and Muller Specification
4,54~,~21 are inserted herein but no Claim is made to the
subject matter described and claimed in that Specification.
Referring to figure 1 of the drawings,
An electronically commutated motor (ECM) 2 is described
in detail in the Boyd/Muller US Specification 4,540,921.
The ECM 2 constitutes a stationary assembly having a
plurality of winding stages adapted to be selectively
commutated, and rotatable means associated with that
stationary assembly in selective magnetic coupling relation
with the winding stages. The winding stages are commutated
without brushes by sensing the rotational position of the
rotor as it rotates within the stationary assembly. DC
voltage is selectively applied by commutation circuit 17 to
the winding stages in preselected orders of sequences
leaving at least one of the winding stages unpowered at any
one time while the other winding stages are powered in
- 13-
l response to a pattern of control signal from voltage
digitizing circuit 13.
The control apparatus comprises a general purpose
microcomputer 10 eg an Intel 8049 which receives commands
for example from a console 11 haYing a series of push
buttons or other user operable controls 9 and the
microcomputer 10 stores patterns of signals which feed-
through a Pulse width modulation control means 18 and a
comutation control signal generator 8 (which are described
in more detail later) to a three phase power bridge
s~itching circuit 17. The necessary power supplies are fed
by a DC Power supply 12. In addition signals are fed from
a winding of the ECM which is unpowered when other windings
in the stator of the ECM are under power. This will be
explained further later~ Signals from the motor windings
are fed to a voltage digitising circuit 13J as described in
the Boyd Muller Specification and below in relation to
Figure 4 of this specification, and are thence supplied to
the microcomputer 10. Power switching circuits also ~eed
through a current sensing circuit 5 to the microcomputer
10. A loop position error indicator 15 and a speed demand
rate velocity timer 16 are provided and a commutation rate
sensing device 14 but any other rotor speed and position
varying device may be used as will be explained further
later~ A pulse width modulàtion control circuit ?8 is
provided.
In broad terms a clothes washing machine according to
*Trademark
- 14-
1 the present invention when operated to cause washing,
functions as follows.
The operator selects a desired set of washing
requirements by operating push buttons controlling its
console microcomputer. As a result the console
microcomputer sends a series of data values to the motor
control microcomputer 10 and these are placed into
registers (memory locations) of the same name, in the motor
control microcomputer 10. Data transmitted from the
console is broken up into 3 groups:
Group 1 contains the command words:
OOH - ~3RAKE
01H - WASH
02H - SPIN
03H - TEST
04H - MODIFY
05H - STATUS
06H - STOP
07H - PUMP
Group 2 contains error codes:
08H - PARAMETER range error detected
O9H - PARITY error detected
OAH - COMMAND error detected
Group 3 contains parameter data:
OBH to 7FH
The motor control microcomputer program knows which group to
expect during each communication, therefore if the program has
got out of step with the console in any way this will be picked
up as a range error.
However due to this data structure some data in group 3 may
be outside their working range so within the listing some
parameters are offset after they have been received so that they
- 15-
1 fall within the correct value to be used within the program
To maintain function overviews, at the beginning of the wash
cycle the console microcomputer 19 controls the filling of the
bowl. While the bowl is filling a spin command is sent to the
motor control microcomputer. The spin speed is very low,
approximately 70 rpm, and its main purpose is to mix the soap
powder while the bowl is being filled. Once the bowl is filled
the console then sends a WASH command to the motor controller 10
to start the agitate cycle. This agitate cycle starts from
rest, ramps up to speed, maintains this speed for a
predetermined time and then coasts to a stop all within one
forward or reverse cycle of the agitator. Once the agitator has
stopped the process is repeated in the opposite direction thus
producing an agitating motion. The console microcomputer 19
determines all these parameters which determine what sork of
wash is required eg. gentle cycle, and is loaded into the motor
controller 10 before the start of the cycle.
The motor controller 10 continually modifies these wash
parameters to account for the load in order to maintain the most
effective dirt removal to gentleness ratio. ~ecause of the
agitating motion the load is shuffled around the bowl and this
affects how fast the agitator ramps to speed and how long it
takes to come to a stop at the end of the stroke. Therefore to
maintain constant wash effectiveness these parameters are
monitored and modified each stroke cycle to maintain th~ ideal
conditions requested by the console microcomputer.
The motor controller 10 will continue this action until it
receives another command from the console microcomputer, In a
- 16-
~3~
little more detail, the wash mode runs as follows.
On receiving a ~'WASH" command a jump is made to the WASH
routine. Low speed windings of the motor are set and a brake is
set off. The routine then waits for the Console microcomputer
to send the wash cycle parameters, ie:
(1) TSTROKE The time for rotation of the agitator in one
direction.
~2) WRAMP The time it takes to reach speed from rest.
(3) ENDSPD The velocity which the agitator must reach
after the wash ramp time is up.
When these have been placed in the appropriate registers
they are then checked for errors. Checks ~or other errors are
also made including a check to make sure the motor is
stationary.
A routine now sets LORATE = ENDSPD = ACCSPD. LORATE is the
motor speed, ACCSPD is the speed that the motor must reach to
obtain the correct wash ramp rate. ACCSPD may become greater
than END~PD to achieve the correct acceleration ramp.
As is explained in more detail later, the speed rate timer
RATETMR used in the timer interrupt routine for the speed
reference count is loaded with the count set in LORATE
previously.
The position error counter 15 is cleared and current trip
and pattern error circuits are reset. In the wash mode the
program bypasses the spin cycle routine.
At this point the plateau time, TFLAT~ is calculated from
- 17-
l the original information sent by the Console microcomputer. To
do this it sets the coast time at 180mS. ~his is a time chosen
which guarantees that the motor will have coasted to a stop
with very little load. Thus the plateau time is calculated:
TFLAT = TSTROKE - WRAMP - 15 (180mS time count)
using a long timer a count of 15 gives:
127 x 96uS x 15 = 180mS ~approx)~
The routines up to this point have only been setting the
wash parameters for the first stroke. The following values as
referred to above, are set in the random access memory in the
motor control microcomputer 10:-
TSTROKE total stroke time, ie. from rest to peak speed and to
rest again.
WRAMP time to full speed
ENDSPD full speed count
LORATE (set at ENDSPD) speed rate
ACCSPD (set at ENDSPD) acceleration rate
ALGFLG (set FALSE) end of ramp flag
ENDFLG (set FALSE) plateau time flag
SLECTR position error counter
RATETMR (set at LORATE) sets speed reference to speed loop
error counter
TFLAT calculated from above parameters; time at maximum speed
At this point the wash cycle can beginO
~5 To actually set the motor into motion we must fir~t set bit
pattern pointers INDEXR and INDEX. For the wash cycle the
direction of motion has arbitrarily been set at CCW (counter
clockwise) for the first stroke, thus:
INDEXR = 12D
INDEX = 00
and the direction register DIRECT _ 01H for CCW.
- 18-
1 The wash ramp time WRAMP is loaded into a long timer for the
beginning of the wash ramp cycle. Commutation now takes place,
and the motor is started up.
After passing through the required time for or number of
commutation routines the prograrn ends. ~t the end of the
agitate cycle the console microcomputer 19 will send a command
to the motor controller microco~nputer 10 to stop the agitate
cycle and turn on the pump to drain the wash bo~l before going
into the spin mode.
As will be explained in more detail later, to enable motor
reversal to be effected the invention requires to determine the
position of the rotor during coasting o~ the rotor after power
to the stator has been cut off. It will be clear however that
this aspect of the invention cannot be put into use until the
rotor itself has been operating under an electronically
commutated sequence. Accordingly when the rotor has been
stopped e.g. at the very start of a washing cycle it is
necessary to start the motor when the position of the rotor is
not known Accordingly the technique described in the Boyd
Muller Specification in particular at page 55 is preferably
used. In this technique the digitised voltages received from
the voltage digitising circuit are tested and as soon as
complementary bits or logic levels in the proper test bit order
have been sensed operations proceed to advance in sequence to
commutate the winding stages. If complementary bits are not
sensed in the predetermined proper test bit order in a
predetermined time period operations take place to advance
commutations in the sequence rapidly and force commutate the
- 19-
1 motor, thus causing the rotor to oscillate briefly~ Th~s if for
example clockwise rotation is required and the sensing indicates
that the rotor is starting to run in the counterclockwise
direction, the rotor runs for a short di~tance in this direction
(one or a few commutations occurring) ~ntil the force
commutating is effected to cause it to run in the correct
direction.
Thus referring to Figure 4 there is provided a three phase
motor 20 with a common point 21 and a switching bridge in which
three switching devices 22, 23 and 24 connect the lower supply
positive rail 25 to the ends of the windings 26, 27 and 28 and
three further switches 31, 32, and 33 connect the ends of the
windings to the power supply negative rail 35. The upper
switches 22, 23 and 24 may be referred to as the A+, B+ and C~
switches and the lower switches 31, 32 and 33 ~ay be referred to
as the A , B- and C- switches
When the motor is stationary there is no information as to
the position of the rotor so it is not known as to which pair of
switches to turn on to get the rotor to rotate in the correct
direction so a selected upper and lower switch are turned on.
Statistically there is a 50~ chance the rotor will rotate in the
correct direction and a 50~ chance that it will rotate in the
incorrect direction. An algorithm is provided in the
microprocessor 10 that onc~e power has been applied senses
whether the motor is going in the correct or the wrong direction
and in the event that the rotor is rotating in the wrong
direction the algorithm advances commutation signals quickly
through the sequence of commutations until the correct sequence
- 20-
l is adopted and the rotor synchronises with the commutated supply
and is now running in the right direction. It may take three or
four switchings or more to synchronise the rotor and so with the
starting algorithm 50% of the time it will start correctly and
will just run into synchronism and 50% of the time it will start
in the wrong direction and ther. stop and recover and then come
back in the right direction. I'hus with this arrangemert every
time the direction of the motor is reversed then if the present
invention as will be described further later is not used then
the motor is allowed sufficient time to coast to zero and then
is started up using this starting algorithm. This start up
algorithm is described in Boyd & Muller 4J~40J921 more fully at
col 8 line 23 et seq and col 23 line 57 et seq and col 2~ line
43 to col 26 line 44. There must be some random initial
rotation ie. some oscillation of the rotor and there must be
time to start correct direction of rotation.
A random start means that the rotor will start in the wrong
direction in 50~ of all starts. Start up algorithm restores the
correct direction of rotation in a time dependent on the initial
rotor position, the pair of switches first energised and the
~otor load.
With a three phase 8 pole ECM as described by Boyd and
Muller there are 24 commutations per rotor revolution. With an
8 to 1 coupling ratio between motor and agitator (e.g. by belt
and pulley arrangement) and typical agitator stroke angles of
145 to 250 o~ arc and acceleration times of 120 to 200
milliseconds respectively the motor is required to accelerate to
speed in the range of 7 to 30 commutations. At startup the
- 21-
1 motor may require 1 to 2 comm~tation angles to restore correct
rotation, a significant proportion of the acceleration period.
The resultant effect is a delay in reversal followed by rapid
acceleration to speed often with some overshoot.
Gentleness of wash action in the washing machine is related
to the acceleration of the agitator. Hence erratic reversal
decreases gentleness. Further, delays in reversals also can
reduce the rate of soil removal. The overall effect is
reduction in desired wash performance.
Thus according to the present invention a more positive
acceleration and consequently a more positive rate of soil
removement and rate of wash action is achieved by monitoring the
speed and position of the rotor while the rotor is coasting
Then when the position of the rotor is monitored down to a
position in which it is in condition for reversal, power is
switched to the motor such that Torque is generated to cause the
rotor to reverse direction preferably within a single
commutation angle and allo~ the motor to run in the opposite
direction without reverting to the start up algorithm.
Accordingly the rotor may be accelerated up to speed and
maintained at speed using the power switching sequence as
described in Boyd & Muller 4,540,921 referring to the tables 1
and 2 therein and in particular at col 6~ lines 24 to 39 where
the following passage appears:
~The winding stages of motor M as explained for instance in
the aforementioned Alley U.S. Patent 4,2~0,544 are
commutated without brushes by sensing the rotational
position of the rotatable assembly or rotor 15 as it rotates
l within the bore of stator 13 and utilising electrical
signals generated as a function of the rotational position
of the rotor to sequentially apply a DC voltage to each of
the winding stages in different preselected orders or
sequences that determine the direction of the rotation of
the rotor. Position sensing may be accomplished by a
position detecting circuit responsive to the back EMF of the
ECM to provide a simultated signal lndicative of the
rotational position of the ECM rotor to control the timed
sequential application of voltage to the winding stages of
the motor."
The present invention is concerned with the monitoring of
speed and position of the rotor while coasting and using this
information to reverse the motor preferably in a single
commutation.
If the rotor to the motor is rotated and voltage
measurements taken at the ends of the phases with respect to the
star point 21 i.e. the centre of the three phase windings, EMFs
will be generated and in Figures 2 and 3 such EMFs have been
plotted. The Figures illustrate a single electrical revolution
of the rotor in degrees and essentially show the wave forms of a
three phase generator with the exception that the wave forms
instead of being sinusoidal are trapezoidal. The three phases
have been indicated by the letters A (pecked line), B (full
25 line) and C (slashed line). For example in B phase it will be
seen that in Figure 2 the EMF goes from a maximum negative at
zero degrees through zero voltage to a maximum positive, stays
at a maximum positive for 120 then goes from maximum through
s
l zero voltage to maximum negative stays at maximu~ negative for
120 and then starts to rise again from zero degrees. It will
be seen that in Figure 2 the sequence (which represents rotation
in a clockwise direction) has a different sequence of EMF
S generations as compared with Figure 3 which represents a
counterclockwise direction of rotation. Referring now to Figure
4, applying voltages to the windings and assuming that winding
26 is A, winding 27 is C and winding 28 is B and that if we wish
to have power on the motor at zero degrees such that we have a
maximum EMF across the motor and thus maximum Torque in
clockwise direction, switches 22 (A+) and 33 (C-) would be
switched on, connecting power from the positive rail 25 through
switch 22 to the A phase windings 26 through the neutral point
21 and the C phase windings 27 through switch 33 to negative
rail 35. Thus referring again to Figure 2 with the notation
therein indicated to obtain maximum Torque in the motor the
connections would be A+ and C- to the 60~ angle and then ~+ and
C- at the 120 angle to Bl and A- to 180 angle then C+ and A-
to the 240 angle, C+, B- to the 300 angle, A+ to B- to the
360 angle, the sequence commencing at A+ and C- again. Thus
there is a sequence of six different patterns and each goes to
60 of angle of rotation giving 360 in rotation. Referring to
the tables herein, Table I summarises the sequence of control
signals required for each step in the sequence described above.
Referring to Table I it will be seen that the rows numbered 5
down to 0 correspond to the sequence of digital signals required
to control the A+, B+ and C+ switches 22 $o 24 and the A~
and C- switches on or off~ A 0 in the
- 24-
-` ~21?~
l table indicates that the switch is turned on and a 1 in the
table denotes that the switch is turned off. This is a negative
notation because of the manner of operation of the
microcomputer. Two further control lines are used to control
whether or not the upper or lower switches are p~lse width
modulated to control motor current. Thus the microcomputer 10
is programmed to contain the pattern shown in Table I. The six
columns from left to right for each switch control line show
each step in the sequence described above with each step indexed
from 0 to 5 in the row marked INDEX. Counterclockwise rotation
is obtained by applying the control signals of Table 2 which is
the reverse of the sequence of Table 1. The value of INDEX
therefore is a reference of position in the commutation sequence
for each table at any time. At each commutation INDEX is
incremented by 1 until a maximum value of 5, then reset to 0 to
continue the cycle. In each table another index is referenced
"INDEXR" as mentioned in connection with flow diagrams discussed
below. The INDEXR row has entries which are unique to each
pattern in the sequence and different for Table I and Table 2 so
that a given pattern is uniquely indentified for clockwise and
counterclockwise rotation. Determination of khe time for
commutation is explained in detail in Boyd~Muller and excerpts
are giYen later. Now during coasting ~as described by
Boyd~Muller) transitions in,signals from comparators monitoring
E~lF signals contain position information. To repower the motor
while still spinning such that the motor continues to run in
sequence in the same direction requires that values of INDEX and
INDEXR be co~puted such that correct switching sequence is
- 25-
~'2~
1 intitated as explained in Boyd/Mull~r. In this specification is
explained methods of repowering the motor such that the motor
reverses direction by determining safe speeds for reversal and
computing suitable values of INDEX and INDEXR such that correct
Switching sequence for reverse direction is initiated,preferably
in a single commutation period.
It will be noted from the diagram of Figures 2 and 3 that
for any 60 commutation interval in the unpowered phase the EMF
is going from the maximum in one sense through a zero to a
maximum in the other sense and it is that phase which is going
to be turned on in the next commutation interval so that the
microcomputer can determine when to turn that phase on by
determining when that phase crosses through the zero point.
This is effected by the use of voltage comparators for example
by circuitry as shown in Figure 5 in which VA is a measure of
this voltage to zero volts appearing in winding 26, VB is a
measure of the voltage to zero volts in winding 28 and YC is a
meaure of the voltage to zero volts in winding 27. When for
example a voltage VC is greater than the voltage VN on the
neutral point N ~21) Figure 4 the output of the comparator 36
will be high. ~hen the voltage is less than the voltage VN at
the neutral point, the output of the comparator 36 is low and
the output of these comparators is fed directly into the
microcomputer 10 which reads in the comparatives. It is to be
noted that the output is comparative when the circuitry is
looking at the comparator for the unused winding at any one ti~e
which will change sense when the EMF in that winding crosses
zero. The microcomputer is then informed that it i~ almost time
- 26-
s
l to commute and in accordance with the present invention with
each successive zero crossing in a sequence, if there is a low
to a high transition the next one is a high to a low transition
then low to high,high to low,and continuing in that way. Thus
the microcomputer knows where each winding is in the sequence
and it knows which of the comparatives to look for for the next
EMF sensing. The microcomputer looks for a transition and it
also knows whether it should be low going high or high going low
so that it can compute from the sequence where the rotor is in
relation to the windings and what the next indications will be
from the comparatives. Accordingly the microcomputer follows
either table 1 or table 2 depending on the direction of rotation
and cycles continue with the correct switches being turned on at
the correct time.
In the A,B, & C circuits of figure 5, as shown with
reference to the C circuit resistance 37 and capacitance 38
provide a filter effect reducing the sensitivity to transients.
Now during coasting, the EMFs are still present in the motor
and thus zero crossing transitions will also still be present
and result in signals being sent by the comparators to the
microcomputer these signals being digitised by the digitisin~
circuit 13, figure 5.
The Boyd Muller Specification describes operations for
repowering an ECM after coasting under control of the apparatus
therein described and is repeated with reference to Figures 17,
18 and 19 as follows.
~ 'In ~igure 17 the relaying routine of step 588 is shown.
Operations commence with BEGI~ ~51 and proceed to produc~ the
- 27-
l OFF pattern (all ones on lines 62) at step 653 to t~rn off the
motor M. At step 655 microcomputer 61 issues a Low on line DB6
(figure 3) producing a High on line H from NAND gate 157, and
causing relay 147 in high low speed circuit 41 to switch from
the low speed connection arrangement to a high speed connection
arrangement. Microcomputer 61 waits for 10 milliseconds as b~
any suitable routine, such as counting from a preset number down
to zero, in step 657 in order to permit the relay 147 armature
155 to come to rest in the high speed position. However, during
this waiting period, the rotor 15 of motor M has, or may have,
rotated through a significant angle for commutation purposes.
Accordingly, at step 659 a routine is executed for determining
the value of INDEX from the sensed digitized voltages on
comparator outputs A, B, and C of Figure 6 when the winding
stages are temporarily unpowered, and resuming producing
patterns of digital signals on lines 62 beginning with the
pattern of digital signals (and thus a corresponding set of
control signals from control signal generator 51) indentified by
the value of INDEX so determined. The digitized back emfs for
three wye-connected winding stages S1, S2 and S3 are illustrated
in Figure 1B and tabulated in Tables III and IV for clockwise
and counterclockwise rotation respectively.
In Figure 18 and in the first three rows of Tables III and
IV, the logic levels of the digitized voltages on input lines 07
1 and 2 of microcomputer 61 (fig 1, 4,540,921) are shown when
rotor 15 (fig 2 4,540~921) is ¢oasting. Each of the six columrs
show- the logic levels of the digitized back emfs present at any
given time. As the rotor turns, the logic levels of a given
l column are replaced by the logic levels in the column next to
the ri~ht. When the right most column is reached, the logic
levels begin again in the left-most column is reached, the logic
levels begin again in the left-most column, cycling through the
Columns as before. Figure 18 shows superimposed on the logic
zeros and ones a waveshape of the digitized back emfs on the
input lines 0, 1 and 2. The digitized back emfs at any one time
and their changes to other values at other times bear sufficient
information to permit sensing the position of the turning rotor
15 and to identify the proper point in sequence for beginning
commutation of such turning rotor and for resuming commutation
whenever commutation is interrupted or discontinued.
Accordingly, the index-determining operations of step 659 as
described in further detail in Figure 19 are used in relaying
routine 588 in the preferred embodiment, and are used in other
embodiments of the invention whenever it is d~sired to begin
commutation in sequence.
In Figure 19 operations commence with BEGIN 671, and
microcomputer 61, (fig 1 4,5~0,921) inputs all the lines 0, 1
and 2 of port P1 at once by maskin~ with ALLHI=07 (binary
00000111). As a result there resides in microcomputer 61 (fig 1
4,540,921) a three bit binary number having binary digits
corresponding to each of the digitized voltages on the three
lines. This binary number is designated DATA1 and stored in
step 673. Then at step 675, micromputer 61 inputs all the lines
0, 1 and 2 of port P1 again in search of digitized voltages
corresponding to an adjacent column of digitized voltages in
Figure 18.
- 29-
l The digitized voltages just obtained in step 675 are sorted
and designated DATA2. In step 683, DATA1 is compared with
DATA2. If they are the same number, (i.e. DATA1-DATA2=)) the
rotor has not turned sufficiently to move to the adjacent
rightward column in Figure 18 and in the Table III or IV
corresponding to the direction of rotation. When DhTA1-DATQ2 a
branch is made back to step 675 to input another set, or
instance, of digitized voltages until an instance of digitized
voltages is found at step 675 which is different from DATA1. At
step 685, the difference DATA2-DATA1 is computed.
When step 689 is reached, microcomputer 61 has stored values
of DATA1 and DATA2 which are in adjacent columns of one of the
Tables III or IV. Each Table III or IV lists values of R3,
which is the difference DATA2-DATA1, in the column corresponding
lS to the digitized back emfs in DATA1. Beneath a value of
difference R3 in each of column of Table III or IV are values of
INDEX and INDEXR. The values of INDEX and INDEXR are precisely
the values for identifying the proper Table I or Table II and
the proper column therein containing the digital signal pattern
which microcomputer 61 can and does then produce to resume
commutation of the winding stages at the proper point in
sequence. (Beneath the tabulated value of R3 in Table III is an
entry designated~'Offset R3~ which is a number calculated in the
program listing of Appendix I for microcomputer table lookup
purposes).
If the direction determined is counterclockwise, a branch is
made from step 689 to step 691 for table lookup in a table in
microcomputer 61 (fig 1 of 4,540t921~ having the information
- 30-
found in Table IV in rows R3 and INDEX. When INDEX is found,
INDEXR is reset by adding 12 to INDEX. If the direction
determined is clockwise, a branch is made from step 689 to step
693 for table lookup in a table in microcomputer 61 (fig 1 of
4,540,921) having the information found in Table III in rows R3
and INDEX. INDEXR is reset as equal to INDEX when the directior,
is clockwise. After step 691 or step 693 is executed, RETURN
679 is reached.
The operations of Figure 19 can be described more generally
as follosls. ~licrocomputer 61 ~fig 1 of 4,540,921) identifies
successive patterns of the control signal s and of the digital
signals of Tables I and II by values of an index designated
INDEX. A value of the index is determined from the sensed
digitized voltages when the winding staFes are temporarily
unpowered. Microcomputer 61 (fig 1 of 4,540,921) resumes
producing successive patterns of the digital signals which
causes control signal generator 51 (fig 1 of 4,540,921) to
generate successive patterns of the control signals in sequence
beginning with a pattern of the digital signal s and control
signals determined from the sensed di~itized voltages. The
lookup table information stored in microcomputer 61 {fig 1 of
4,540,921) is a function, i e. a predetermined correspondence
between members of two sets of numbers. The sets of numbers
involved here are val~les of INDEX on the one hand and values of
the differences R3. E~uivalently, Tables III and IV can be
regarded as tabulating INDEX as a function OI digitized back emf
itself. It i5 also to be understood that there are a multitude
of equivalent ways made known by the disclosure made herein7 of
-- 31--
l setting up a function relating the di~itized back emf
information to some variable such as INDEX which can be used to
determine the proper point for beginning in sequence when
commutation begins again. When the successive patterns of
digital signa~s and control signals are identified by values of
an index, the index is advantageously determined as a function
of a number represented by the sensed digitized voltages when
the winding stages are temporarily unpowered, and microcomputer
61 tfig 1 of 4,540,921) resumes producing patterns beginning
with the pattern of the control signals identified by the value
of the index so determined. The index is determined as a first
f~nction of a number represented by the sensed digitized
voltages when the winding stages are temporarily unpowered and
the preselected sequence is for clockwise rotation of the
rotatable means 15 (fig 1 of 4,540,921) and determined as a
second function of the number so represented when the
preselected sequence is for counterclockwise rotation, and
microcomputer 61 (fig 1 of 4,540,g21) resumes produoing patterns
beginning with the pattern of the control signals identified by
the value of the index so determined. The value of the index is
also determined as a function of the difference of first and
second numbers represented by different instances of the sensed
digitized voltages, and microcomputer 61 begins with the pattern
of the control signals identified by the value of the index so
determined.
The value of the index is determined as a function of the
difference of first and second numbers represented by different
instances of the sensed digitized voltages unless one of the
- 32-
~` ~2~ 15
l numbers is in a set of predetermined numbers, such as 0 and 7,
and micrcomp~ter 61 begins with the pattern of the control
signals identified by the value of the index so determined. A
difference of first and second numbers represented by different
instances of the sensed digitized voltages is calculated and a
value of the index is determined as a function of the difference
unless the difference is in a set of predetermined numbers, such
as 0, +3, and -3, and microcomputer 61 (fig 1 of 4,540,921)
begins with the pattern of the control signals identified by the
value of the index so determined. Microcomputer 61 (fig 1 of
4,540,921) in this way prevents sensed digitized voltages
representing a number in a predetermined set, such as 1 and 7,
from being used to determine the beginning pattern of control
signals. Microcomputer 61 ~fig 1 of 4,540,921) repetitively
senses the digitized voltages while the winding stages are
t~mporarily unpowered and determines the beginning pattern of
the control signals as soon as a change occurs in any one of the
sensed digitized voltages. Il
Table 3 herein is e~uivalent to Table III in the Boyd Muller
specification.
It is to ~e noted that in Boyd/Muller when the motor is
operated in the agitate mode to reverse motor direction a
definite time is allowed for the rotor to coast to a stop and
then random restarting is effected with a 50% chance that the
rotor will start in the wrong direction necessitating adjustment
of the commutation to reverse the rotor direction and accelerate
to speed in the right direction. This gives irregular
accelerations to the rotor and thus causes irregular washing
- 33-
s
l action to result. Accordingly this invention is a mathematical
way of finding where the rotor is and where the switching in the
sequences will be. Thus with a transition the microcomputer
calculates which switches should be on at any one time.
s If we want to start at that time we apply power with those
switches so set or indexed these tables and start applying
power.
Timers are provided as follows:
- SHORT TIMER, LONG TIME'R, COMMUTATION TIMER.
In this implementation an INTEL 8049 1 chip microcomputer is
used for motor control microcomputer 10. It contains an 8 bit
timer. This timer can be driven by either an external
oscillator or directly from the ALE pulse which is divided by a
factor of 32 before entering the timer (ALE = CLOCK/32). Tne
microprocessor clock runs at 10 MHZ so therefore a
( 10MHZ/ 15)/32-20.833 KHZ clock signal is applied to the timer.
This provides a count every 48 microseconds in the timer and in
operation the timer is loaded with a count of 2 thus providing
an interrupt pulse every 96 microseconds. This interrupt rate
provides the base timing to the motor controller,
On interrupt the program is forced to jump to a Timer
Interrupt Routine. On entry to this routine the timer is
reloaded with a count of 2 to provide the 96 microsecond base
time
This routine has two major functions:
(i) Decrementing Timer Register counts every 96 ~icroseconds,
and setting the appropriate timeout flag when the counts
reach zero.
- 34-
1 There are three timer registers used.
(a) Short Delay Timer
(b) Commutation Delay Timer
(c) Long Delay Timer.
The registers (a) and (b) are decremented each interrupt,
therefore using a count of 01H to OFFH timers (a) and (b)
can achieve time intervals of 96 microseconds to 24
millisenconds (ie. 256 x 96 microseconds). For extended
time delays using register (c), an intermediate prescaler
register which is initially set to 7FH (127) is decremented
every interrupt. Only when the Prescaler Register reaches
zero is the register (c) decremented. Therefore the long
timer can achieve time intervals of 127 x 96 = 12
milliseconds to 127 x 256 x 96 microseconds - 3 seconds,
In order for the main program to use these time delays a
count must be put into the appropriate Timer Register. The
timer flag must then be tested periodically to see whether
the time is up.
(ii) The second function of this routine is to provide the
Speed Demand Rate function 16 of fi~ 1. ie. to provide
a count rate to position error counter 15 equal to the
required motor commutation rate. This a achieved by
setting the Speed Rate Timer Register (RATETMR) equal
to the count for t~e period of the required commutation
rate eg. ACCSPEED, ENDSPD. Thus on every timer
interrupt the RATETMR is decremented and once it is
zero the Position Error Counter 15 is decremented. The
RATETMR is automatically reloaded with the correct
l count and the cycle repeats for contin~ous operation.
Referring now the Figure 6 which is a flow chart of the
reversing seq~ence of the present invention it will be assumed
that the microcomputer has timed out the application of power to
the motor and the motor is switched off i.e. all power is
disconnected from the stator. A long timer 40 is set to 15~-2G0
milliseconds preferably 180 milliseconds which is an arbitrary
- maximum time of coasting. As stated power is turned off as
indicated in block ~1 and a check is made in block 42 of the
register DIRECT prGvided in the microcomputer 10 to indicate
whether the motor is going clockwise or counterclockwise, In
the event that direction of rotation is clockwise the register
~alue is changed to counterclockwise ready for starting in the
next direction and vice versa, so that the appropriate blocks 43
and 44 are used as required. There is a second timer called the
short timer 45 which is set to a value of 40 milliseconds. This
timer provides a safety feature in that should the rotor stop
then of course the succession of EMFs ~ill also stop and no
measurable signals will be transmitted to the microcomputer to
work on. ~ccordingly the second timer assists in avoiding
maloperation.
There is a third timer which is the commutation timer 46
which is set to 20 milliseconds. Now that value corresponds to
a rate of occurrence of zero crossings su~fici~ntly low as to
allow re~ersing to take place. Next there is a tag rotor (tag
corresponding to R3 in the Boyd Muller Specification) position
indicator, block 47, which senses the position of the rotor.
~ 36-
l This is related to tables 4 and 5, table 4 being ~sed when it is
required to go from clockwise to counterclockwise and table 5
when it is required to go from counterclockwise to clockwise as
is explained more fully in figure 8. Thus a start is made by
inputting the values of A, B and C, that is the outputs of the
voltage digitising circuit. These are stored in memory as data
1 (block 60 figure 8) that is the location. Then the values
corresponding to the EMF signals are inputted again and stored
as data 2 block 61. These data 1 and data 2 are then compared
in block 62. If they are equal and if the short timer is not
equal to zero bock 63 that is to say a transition has not yet
been reached the computer (as indicated by line 48) takes the
measurements again of A, B and C, comparing them to the previous
value. As soon as data 1 is not equal to data 2, data 1 is then
subtracted from data 2 and this gives a value in hexadecimal for
the transition. That is put into the storage register called
"Tag" block 64. Then the flow diagram is traversed further to
see if the modulus of data 2 minus data 1 equals 0, 1, 2 or 4
each of which is one of the allowed values. If it is not9 there
is something wrong and it is a matter of going back to the
beginning and restarting the whole procedure a~ain because the
values are incorrect for whatever reason. Normally however such
values are correct and there is a valid change and the routine
above set forth is then move~d o~t of. If there is no transition
within 40 milliseconds as indicated by the short timer then the
rotor is down to a speed at which reverse direction can take
place. If a transition is obtained within 20 milliseconds as
- 37-
l indicated by commutation timer then the rotor is still spinning
at a rate greater than that allowable for reversal and it is
nececsary to run through the sequence againf If the long timer
has not reached 0 as checked in block 49 then we have to check
to see if the commutation timer is equal to 0 as checked in
block 5G, if it is not, then it is known that the rotor is still
spinning. The sequence goes around monitoring the position,
keeping up to date and getting a new value of the rotor position
every time the sequence has gone through. If the long timer
which is set for 180 milliseconds (a little longer than the
expected coasting time)~ times out then it is necessary to apply
a dynamic brake, e.g. by short circuiting all the windings one
to the other. The short timer 45 is a safety device which
ensures that the routine is not continually gone through
lS searching for a timing out when in fact the rotor has stopped
and although looking for a change no such change will occur
because there is no EMF generated to create such a change. Thus
when the com~utation period gets greater than 40 milliseconds
the device times out. Assuming that a transition has been found
within the allowable parameters then values are derived for
INDEX and INDEXR at block 53 which is explained in detail in
figure 7. When the rotor i3 down to a speed at which reversal
can take place, information stored in register TAG and direction
register DIRECT defines where the rotor is and its direction of
rotation. Accordingly values of IN~EX and INDEXR according to
either Table 4 (clockwise to counterclockwise rotor position
sensing) or Table 5 (counterclockwise to clockwise rotor
- 38-
1 position sensing) are chosen and windings energized which will
cause a torque to the rotor which cause the rotor to reverse
direction from its previous direction. If for example the E~IF
from the motor windings when the rotor is coasting are those
resulting from clockwise rotation, such E~lFs will follow the
pattern of Figure 2, and supposing the rotor is in a position
where EMF C is high, EMF B is low and EMF A is changing from low
to high i.e. the transition point 55 in Figure 2 is reached and
has been reached in time greater than 20 milliseconds ~in normal
operation) after transition point 56 has been reached. If power
were applied to continue in the same direction the switchings to
the windin~s would be A+ and ~- but since it is required to
reverse direction and it will be seen that in Figure 3
transition point 57 corresponds to transition point 55 in Figure
2 so that to provide reverse torque switchings B+ and A- will
energize the required windin~s. In some circumstances
energizing of C- instead of A- may be used since EMF A is
falling to the right of transition point 57 while C is rising.
Thus in Table 4, Index 3 relating to table 2 is chosen in
preference to Index 4 and when the EMF in the selected winding
drops back to zero due to rotor speed dropping to zero
commutation increments to index 4 in Table 2 and sequence
continues in selected order. The position loop error counter 15
is set to a restart value in block 53a the speed demand rate 16
~5 is set to a restart speed in block 53b and the microcomputer
then returns the timing to a main commutation programme. Of
course during agitate the reverse routine shown in Figure 6 is
- 39-
l reverted to at each reversal until the end of the wash cycle
determined in this method by command module 11 which commands
microcomputer 10 to cease and a further routine entered into
e.g. draining then spinning.
It will be seen that by following the reverse routine in
which the position of the rotor is monitored down to a point and
speed in which the rotor is in condition for reversal a reversal
can be effected in a single commutation period causing the motor
to pass through the stop and reverse direction without loss of
rhythm unless braking has had to be effected. When braking is
effected it may be necessary to go back to the start routine
above described in which the selected switches are turned on and
indications from the windings used to indicate whether the rotor
is moving in the right direction. If it is not, then the motor
is force commutated to change direction of the rotor and pick up
acceleration speed as above described. However this does not
happen usually in practice but the smooth transition with change
of direction within about one commutation period effected,
Furthermore even with dynamic braking in which the motor
winding ends are connected together it is still possible to
monitor the velocity of the rotor down to the point of reversal
thus reducing the time in which reversal is effected. In the
voltage digitizing circuit of Fig 5, unlike the ~oyd Muller
circuitry, the star point voltage VN is brought into the circuit
13. The voltage at the star point is the vector sum of the
three EMF's generated in the windings and varies at the
commutation rate. The signals from the comparators are not in
- 40-
.
1 the same sequence as an open circuit when coasting but are in
synchronism with the rotor when accordingly the velocity of the
rotor can be measured and reversal commenced when the velocity
falls to a cesired level. Thus in testing for transitions and
the agitating sequence has not been interrupted then the changes
which take place are monitored and the numbers go from all O's
to all 1's not all at the same time, but the pattern is
sufficient to enable the time for reversal to be determined.
Thus referring figure 4, braking is effected by making
switches 31, 32 and 33 conductive there is a small voltage drop
in these switches and although VA VB and VC all move together
and therefore it is not possible to tell the poC.ition of the
rotor, the comparators of figure 5 will detect small volta~e
variations (about 1 or 2 volts) between the VA VB and VC voltage
and the YN voltage to enable the rate of movement to the
indicated and passed on to microcomputer 10.
~L~
~:~QB cLo-c~L~ R O~Q~IQ~I
P2 Rail
L i ne D i s a ~ e q u ç ns~ e _o~ t t~s
D 7 Top 0 1 0 1 0
I
G 6 Bt~ 1 0 1 0 1 0
I
T
________________________~_______________ ___________
A 5'B-) 1 1 1 1 C C
L
4tC~) 1 1 1 0 0
S
I 3(A-) 1 1 0 0
G
N 2(B+) 1 0 0
A
L 1(C-) O O
20 S
OtA+) O 1 1 1 1 0
__________________________________________________
INDEX: 1 2 3 4 5
INDEXR: ! 0 1 2 3 4 5
2 5 C 0~1 TR OL: A~ B~ B+ C+ C~ A~
SIGNALS: C C- A- A- B B-
DIG IT ISED
V OLTAG E 01 02 04 01 02 04
MASK: (B) (A) (C) (B) (,q) (C)
_ 42--
~11
DAT~_E~OR_C~ B-cLo--c~-w~ B-o~-A~
P2 Rail
Line ~isa~l~ SeQ~ sf~ ~Ls
D 7 Top 0 1 0 1 0
I
C 6 Btm 1 0 1 0 1 0
T 5(B-) O 0
A
L 4(C+) 1 0 0
S 3(A-) 1 1 0 0
I
G 2(B+) 1 1 1 0 0
N
A 1(C-) l 1 1 1 0 0
L
S O(A+) O 1 1 1 1 0
__________________________________________________
I~DEX: G 1 - 2 3 4 5
INDEXR: 12 13 14 15 16 17
CONTROL A+ C~ C+ ~+ ~+ A~
SIGNALS: B- E- A- A- C- C-
DIGITISED
VOLTAGE 04 02 01 04 02 01
MASK: (C) (A) (B) (C) (A) (B)
- 43-
~Q~I
CLoCKWISE liQ.l'Q.B~Sl.l`IO~_ S~ISl~C
( LOW TO HIG~S~:D WINpI~lGSl
B O O 1 1 1 0
A 1 1 1 0 0 0
C 1 0 0 0
~EX: 6 2 3 1 5 4
TAG: 2 -4 1 -2 4 - 1
10 ( DATA2-DATA 1 )
INDEX: 5 ` O 1 2 3 4
INDEXR: 5 0 1 2 3 4
-- 44-
~PPENpIX
~BLE IV
CLOCKWISE TO CQlLN~cl~o~lI~E
BOTOR POSj~lQ~I SENSING
B 0 0 l l 1 O
A 1 1 1 0 0 0
C 1 0 0 0
Y.EX: 6 2 3 l 5 4
TAG: 2 -4 l -2 4 _l
( DAT A2 - DAT A l )
INDEX: 3 2 l 0 5 4
INDEXR: 15 14 13 12 17 16
-- 45--
~t~ 5
~L~Q
COU NT EE Cl,O CKW T~_~Q~,L~OÇ~I,~
EOTO1( POSITION~LSI~G
B 1 O O O
A 1 1 1 O O O
C O O 1 1 1 0
HEX: 3 2 6 4 5
TA~: 2 - 1 4 -2 1 -4
(DATA2-DATAl )
INDEX: 3 2 1 O 5 4
INDEXR: 3 2 1 O 5 4
- 46-
l Turning now to the second aspect of the invention, as stated
above the digitising circuit 13 is responsive to the Back EMF of
the ECM 2 to provide a simulated signal indicative of the
position of the ECM rotor.
Ve1ocity control of the ECM 2 is provided by a microcomputer
controlled digital implementation of a position control loop
referred to later. Position and velocity feedback information
is contained in the outputs of the voltage di~itisine circuit
13. Commutation rate sensing software 14 in the motor control
microprocessor 10 supplies a count of one to ~osition error
counter 15 for each commutation. Each count decrements the
counter by one. The count rate is therefore proportional to
motor velocity. Requested velocity information is provided by
speed demand rate timer hardware/software 16 which supplies a
count rate to position error counter 15 equal to the required
motor commutation rate, that rate having been indirectly
selected by appropriate actuation of manual selection controls
in the user controls 9. Speed demand rate timer 16, amplifier
stages, pulse width modulation controls 18, commutation control
signal generator 8, commutation circuit 17, voltage digitising
circuit 13 and commutation rate sensing circuit 14 define the
feed back ~osition control loop the summation point being the
position error counter 15.
The position error coun~er 15 algebraically sums the
~ositive pulse rates from the speed demand rate 16 and th~
negative commutation rate sensing device 14. The output from
- 47-
~IL289~8~
l the position error counter 15 appears as an error signal being
the algebraic difference between the two counts which controls
the current (and hence power) in the motor by a Pulse Width
Modulated Control Circuit 18 with current limit control 5. The
error is the difference between the desired count as indicated
from the speed demand rate indicator 16 as compared with the
commutation rate device 14. A zero PWM rate is the e~uivalent
of a zero count and a 100Z PWM rate is the equivalent of a full
scale count, This aspect is explained ~ore fully in
Canadian Patent Application Serial No. 476071, fi1ed
March 8, 1985 by Fisher & Paykel Limited and which
explai~s improved pulse width modulated control methods for
controlling current (ar.d hence power) to an inductive load with
special ~pplications to D.C, Motors. In this way the Digital
Position control loop is arranged so that when the ECM is
rotating at a speed less than that requested by Speed Demand
Rate Timer 16 low speed power is increased until current limit
is effected to give faster acceleration but during steady speed
operation the error and hence PWM pulses are maintained and
controlled to control the power input to the ECM to that which
is sufficient to maintain speedO
User controls ~ are provided and in the preferred embodiment
include a command microcomputer 19 which translates the user
commands into signalc to th,e motor control microcomputer 10.
Thus the speed demand rate is set by commands initiated by the
user controls 9 and these controls have commands relating to a
wash program0e selection e.g. delicate, regular, heavy duty~
_ 48-
~39~S
l wcol, permanent press and also a selector relating to water
level e.~. low, medium and high water level. Each of these
imposes a different power demand, stroke angle, acceleration
rate and speed from the other on the wash load imposed on the
agitator 1 which is mounted wil;hin a spin tub 3 and water
container 4 in the known way, In figure 1 the motor 2 is shown
driving direct to the agitator 1 but of course an indirect drive
could also be used.
The above describes an electronic controlling circuit which
enables the speed of the motor 2 to be controlled.
Referring now to figure 9 this indicates a profile of
velocity against time of one half cycle in the oscillatory
rotation of the agitator by the motor 2. A~ may be seen power
is applied to the motor to achieve three steps in the half
cycle, an initial step 120 of acceleration from zero velocity to
a desired maximum velocity a second step 121 at which the
maximum velocity is maintained until a cutoff point 122 is
reached when power is removed from the motor and a third step
during which the rotating assembly of the motor and the agitator
then coasts to a stop substantially in accordance with either
for example curve 123 or as is shown in smaller pecked lines
curve 124, the curve 124 starting from a ~ifferent cutoff point
125 which will be explained further later. Thus there are three
different times, an acoeleration time 1~8, a plateau time
referenced 129 when substantially constant speed is maintained
sub~ect to matters discussed below and a coasting t~me
referenced 130. The sum of these times results in a total
stroke time.
- 49-
~9~
l Of these times the acceleratio~ time 128 and the plateau
time 129 are electronically controllable but the coasting time
130 is dependent on mechanical conditions involving the inertia
of the rotatin~ assembly including the rotor of the motor and
the agitator and associated drive gear against which is acting
the resistance of the load of f`abrics placed in the spin tub 3.
Accordingly the coast time 130 will depend on and vary according
to the load placed in the washing machine plus other smaller
factors such as the effect of heating up of bearings.
A desired washing action w:ill vary from a gentle action if
the "delicate" control is actuated to a heavy duty vigorous
action if the "heavy duty" control is actuated. In a particular
washing machine which has been made, five types of washing
actions had been provided as mentioned above namely delicate,
regular, heavy duty, wool and permanent press and three
different water levels so that it is possible to have 15
combinations or 15 different agitator velocity profiles that
must be achieved.
To do this command microcomputer 19 feeds commands based on
information from user controls 9 to the motor control
microcomputer 10 which define the acceleration time, the stroke
time and the maximum speed of rotation according to the
selection made in the user control circuit 9 and which have been
preprogrammed into the command microcomputer 19.
Motor control microcomputer 10 retains this information and
commar.ds the motor to agitate following the required profile via
the digital position control loop as explained below repeatedly
- 50-
1~9~
l until instructed to stop by the command microcomputer 19.
The method for control of acceleration time 128 can be
explained with reference to figure 10.
In figure 10 are shown typical curves of the velocity/time
5 showing the effect of velocity demand on acceleration. Thus
figure 10 is a plot of velocity versuC time for the motor. The
information provided by operating the user control in circuit 9
is based on the motor being started at zero speed and that the
contents of the position error eounter is at zero~ Accordingly
the command defines an acceleration rate i.e. req~ested velocity
that must be achieved in the acceleration time 12a of figure 9.
That velocity can be provided ei~her as motor RPM, agitator RPM
or commutation rate and suitable circuitry provided dependent on
the type of information provided. The various curves ~1 to ~4
in fi~ure 10 show the different acceleration rates resulting
from velocity demand rates for one resistance to rotation of the
motor and show the time taken to reach a maximum velocity or
speed.
As can be seen in figure ~ acceleration rate increases with
increasing speed demand rate. Each curve is essentially llnear
over its first portion. Time to reach requested speed is almost
independent of speed demand rate but is a function of the loop
gain of the position control loop.
For any given velocity profile the acceleration must be such
that a set speed i.e. the plateau speed 121 shown in figure 9
must be achieved in a certain time. Accordingly the command
must be set to provide a definite acceleration rate i.e.
1 reaching the set speed in a given timeD However the load on the
agitator is not at this tiG,e known and therefore initially a
speed demand rate is initialised which will reach the maximum
speed in the given time ~nder arbitrary predetermined
Conditions, The preferred method of operation is to initially
set a speed rate demand which will result in an acceleration
rate which is slightly less than that ultimately desired and
then to adjust the speed demand rate upwardly to the desired
speed over the next dew cycles. Thus giving a wash action which
is more gentle than would be obtained by moving quickly to the
maximum speed with the possibility of overload occurring. This
is achieved by adjusting the loop ~ain of the velocity control
loo~ in any known way such that the time taken to reach the
required plateau speed when speed demand rate timer 16 is loaded
with that plateau speed rate is greater than the range of times
required. One way is to adjust the error value contained in
position error counter 15 required to achieve 100% PWM rate. If
the load in the machine is light the agitator will accelerate to
speed more quickly than if the load is heavy. Accordingly the
present invention provides program~ing of the microcomputer such
that the speed at the end of the required acceleration time is
measured. If that speed is below the required speed the
microcomputer issues an instruction to increase the speed demand
rate at the beginning of the next agitator stroke. Similarly if
the motor is above speed at that time the command is to reduce
the speed demand rate and thus reduce the power applied to bring
the motcr up to the plateau speed. This testing of the
- 52-
:'
.
:
~9185
l acceleration rate is carried out on each half cycle whether that
half cycle is in the forward direction as shown in figure 20 or
in the reverse direction. Thus the oscillating rotation i.e.
the back and forth motion of the motor 2 and the agitator l is
such that the resistance to oscillation or rotation is measured
at each half cycle and by modifying the acceleration rate to
always bring that acceleration rate to a position where the
plateau speed is achieved in the set acceleration time and
substantial uniformity of operation is thus attained. Thus
acceleration is controlled to achieve the desired plateau speed
in the desired time, and this acceleration speed is maintained
within practical limits.
Plateau speed is maintained by adjusting the speed demand
rate 16 to the speed de~and rate required for the plateau speed
at time 127 in figure 9.
However consideration must now be given to the circumstances
illustrated in figure 11. In this figure the demanded velocity
is chown by a pecked line 130. A series of curves are shown,
the upper curve 131 showing an overshoot and curve 132 shows a
le~ser overshcot while curves 133 and 134 show two undershooting
curves of velocity. This is brought about by the varying
position error count in the position error counter 15. If there
is a heavy load i~ takes considerable power to get to speed and
the power to get to speed is greater than that re~uired to
maintain that speed and that is indicated by a large counter
value in the position error counter 15 and hence a high PWM
ratio in circuit 18. Accordingly at the time of reaching the
- 53-
, , . . ~
~2~ 3S
l ~oint 135 in figure 11 (which corresponds with point 127 in
figure 9) there is more power applied to the motor than required
to maintain the motor at the de0anded velocity 137 and the motor
will thus continue to accelerate for a short time and overshoot
as can be seen by either of curve~ 131 and 132. This can be
provided for by adjusting the value set in the Fosition error
counter 15~ If the initial position error count is set ak a low
level there is an undershoot b~elow point 135 with the power then
being levelled off by the abovle checking of the speed and
comparing that speed with a desired count rate or alternatively
the acceleration power can be maintained to above point 135 so
that there is an overshoot and then the automatic error counting
is carried out to reduce the overshooting curve dcwn to the
demanded velocity straight line 137. The value of the Fosition
error counter or the speed demand rate can be adjusted at any
time under control of the microcomputer so that the actual count
can be updated or modified as desired and since the counter is
within the microcomputer, it can be loaded at any time.
Now the value of that counter at 127 in figure 9 is an
indirect measure of the wash load. Where we have a high value
in the counter we have a large wash load9 a small value in the
counter shows a small wash load. Now to further increase the
power that we have applied to the load as that load increases in
excess of that required to maintain the profile as explained to
maintain a given level explained in background we can adjust the
amount of overshoot to obtain any profile. What is done is that
the value of the counter is adjusted such that with only water
- 5~-
. :;. -,-: ' ~ - ;...... ..
1 in the bowl there is no overshoot then as clothes are added or
as the load is increased the value is adjusted in the counter to
allow small amount of overshoot. This small amount of overshcot
increases the stroke length of the agitator slightly and
increases the turnover in the clothes. This is explained above
in the background material but essentially wash action is
provided by movement of clothes through the water and how
vigorous this movement is determines the soil removal. However
by increasing the stroke length slightly the required wash
requirements are maintained. I'he acceleration rate and velocity
desired in for example delicate wash are such that slight
lengthening of the stroke angle does not result in excessive
washing action.
The function of maintaining acceleration rate by adjusting
the speed command rate and controlling overshoot allows slight
increases so that under very heavy loads the stroke length is
increased slightly. If the acceleration rate is not controlled
then typically with a velocity controlled motor if just a final
speed is requested, the error in position error counter
increases and the acceleration rate decrease with load and
results in a decreased stroke angle and lowered soil level
removal.
It is necessary now to look at the coasting time and curves
of figure 9. As stated above the deceleration rates of the
agitator and motor are not electronically controllable. The
rotating assembly can only be allowed to coast to a stop or be
braked to a stop and thus are not electronically controlled.
- 55-
1 NGW if the coast time were fixed so that it could be guaranteed
that the motor would coast to a stop before an attempted reverse
or almost to a stop before reversal could be effected it would
be possible to end up with a shorter stroke time as the load
increases because we have the sitation that the maximum time to
coast to standstill is when there are no clothes in the water.
As the clothes load increases the coasting time becomes shorter
and thus the area under the curve in figure 9 becomes less and
since that area is proportional to the stroke angle that we are
applying to the clothes load or the agitator if deceleration is
effected more quickly then the stroke angle applied to the load
is decreased which is disadvanta~eous. The opFosite effect i~
however desired namely that it is desired to increase stroke to
the load as the load increases and therefore the following
techniqu~ is also adopted. The stroke time is set to a
predetermined figure by a command received from circuit ~. This
stroke time is for practical purposes the same for all wash
duties. This means that as the coast time decreases the plateau
time must be increased so that the point 122 in figure 9 is not
a point fixed in time but a point which is determined as
follows. For each half cycle, the microprocessor measuree the
time to coast from plateau speed to substantially zero speed and
the microprocessor subtracts that time from the stroke time and
also subtacts the rèquired acceleration tiwe from the stroke
tiwe which leaves a plateau time required for the next stroke so
that for each half cycle of the agitator the microprocessor
calculates a new plateau time depending on the last coa~ting
- 56-
893L8~i
l time and as may be seen from fig~re 9 two different coasting
ti~es and two examples of different plateau times are shown. In
the first the platea~ time is the time to extend from point 127
to poir,t 122 and for the second, assuming the same acceleration
time, from the point 127 to the point 125 and the deceleration
or coasting curves are as showr. by the lines 123 and 124
respectively. Accordingly at least in the preferred form the
invention comprises the combination of the three techniques for
controlling acceleration and altering the acceleration time as
desired controllirg the overshoot or undershoot in relation to
the desired ma~imum speed in the second zone of figure 9,
recalcuiating the plateau time for each half cycle depending on
the coasting time in the last half cycle and then reversir.g the
rotating assembly immediately at or near zero speed. This
allows the maintenance of an; required washing p~rformance.
Corrections are made continously and by monitoring the curves
such as those shcwn in figure 9 on a oscilloscope it can be seen
that variations occur substantially all the time because the
load on the agitator may well depend on the position of the
clothes in the container and those clothes may be bunched or
balled in some cases and almost i~mediately the bunching can be
freed by the agikator action so that the load on a rest half
cycle is considerably lighter than when the clothes are
bunched. The time to ~ccelerate to a given speed may take a
number of strokes to settle out as there is a high averaging put
on which prevents big disturbances for example if a bunching is
only momentary then if it were not for some delay in averaging
- 57-
%~
l out there could be violent disturbances in the speed of
actuation of the agitator and this could cause too vigorous an
actior. and with a heavy load then there is ar increased power
input which is what is requirecl.
In a less preferable alternative it is possible to allow the
stroke time to vary. In such an alternative the maximum speed
would be more closely monitored so that extra area under the
curve of figure 9 and therefore the extra stroke angle for
heavier loads would be gained by extending the power cut off
point as required.
The sequence of operations will now be explained in relation
to the flow diagrams shown in figures 12 to 16. The flow chart
of the main routine shown in figure 12 can be explainecl with
reference to figure 9. This is the routine required to agitate
and the first initial block 140 is shown in more detail in
figure 13 where the notations are: T stroke is stroke time,
W-ramp is ramp time, and End-speed is the maximum required
speed. Once initialisation has taken place there are four
things to do, first it is necessary to start at the beginning of
the strG~e to accelerate till point 127 in figure 9 ls reached
to maintain a plateau speed along the plateau 121 shown in
figure 9 and then to coast to a stop after power has been
switched off at 122 and then to reverse the direction of
agitation and recommence the cycle in a ~upside down"
disposition from that shown in figure 9. These steps are shown
in figure 12 where acceleration is shown in block 141 malntain
plateau speed shown in block 142 the decelerate or coast is
- 58-
9~
l shown in block 143 change direction is shown in block 144 and
additionally in block 145 there is a decision to be made as to
whether agitation is to be concluded and if so the command
microcomputer 19 sends a signal to the motor contro]
microcomputer to interrupt the sequence at a selected time that
agitation is to be ended. If the answer is no then the
accelerate maintain coast and change direction cycle is
maintained for a further cycle and so on ur.til the interrupt
signal is given. A yes (Y) an.swer results in the end of
agitation and the washing cycle then goes into a further routine
which does not form part of the present invention.
Now referrin@ to figure 13 when initialisation is commânded
the parameters fed to the motor control microcomputer 10 are
stroke time and acceleration time but it is necessary for the
plateau time i.e. for the point of 122 to be calc~lated. Thus
the acceptance of the agitate parameters are shown in block 150
and in block 151 there is a calculation of the initial plateau
time shown as initial T-flat. This time is arbitrârily selecteà
for the first stroke as the stroke time (which is a set time)
minus the ramp time W-ramp which is the acceleration time and
then an arbitrary 150 milliseconds which is taken to be a
reasonable coasting time. Thus f~r the first stroke the T~flat
time equa1s the initial T-flat time i.e. the time obtained by
the calculation shown in block 151. This procedure is necessary
since on initialisation there is no information a~ to what the
real coast time is going to be so an estimate is made and
subsequently after every stroke the actual coasting time is
- 59-
~2~39~5
l measured and used as will be explained later.
As a next step it is necessary to know the speed to which
the motor is to be accelerated. Again there is no information
as to the speed likely to be attained in the time interval 128
on applying a known amount of power and accordingly as is shown
in block 153 the speed to which the motor is to be accelerated
referred to as ACC speed is shown as being the end speed i.e.
the maximum speed to be obtained for the particular wash
programme selected and the end speed for example as it is seen
in figure 10 for any given velocity demand. Acceleration at the
commencement is virtua1ly linear and if commands are given to
supply Fower to the motor so that a substantially linear
acceleration is obtained up to the fixed demanded speed and for
the first stroke the demanded velocity is to be equal to the
lS plateau speed i.e. End-speed. Ho~ever ac explained abo~e it is
preferably to arrange the gain of the position loop such that
acceleration is always less than normally required if initial
acceleration speed demand rate equals End-speed when the
agitator is operating in water only. The practical result of
this is that End-speed or maximum speed is not actually achieved
in ti~e interval 1~8 for the first stroke.
Looking now at figure 14 which is a flow chart during the
acceleration phase the timer is set to W-ramp which is a fixed
time in block 154. This ti~er is a kimer which is set to the
time and then counts down to zero so it is set with an initial
value that is equ2l to the time that is required. It is set
running which automatically happens when the timer is loaded and
- 60-
. ~
185
l the microco~puter senses when it gets to zero so in future it
knows how long it has taken to do something so the acceleration
time is t~e acceleration portion shown in figure 9 namely slGpe
120. As shown in block 155 the microcom~uter then loads the
speed demar.d rate 16 and this is set at a rate equal to the
acceleration speed which for the first stroke as we haye
discussed above is the End speed as shown in block 153 figure
13. As shown in block 156 the rotor is started and acceleration
takes place while as shown in block 157 the timer runs down to
zero and at this stage the motor velocity ~-ill have reached
about point 12~ and at that point as shown in block 158 the
actual speed is measured by use for exan~ple of the commutation
rate sensing shown in block 14 fi~ure 1 where the interval
between commutations is measured by the motor control
n!icorcomp~ter. That actual speed is cGmpared w1th the spee~
which is required in block 159. If it is less than the
End-speed then the microcomputer checks to see if the
acceleration speed is less than an arbitrary maximum as seen in
block 160. If it is then the acceleration speed is incremented
one step and retesting is carried out as seen in block 161. If
the actual speed is not less than End-speed or acceleratiQn
speed is not less than the maximum then a check is made as in
block 162 to check if the actual speed is greater than the
End-speed. If no then again the te~t is ended, If it is
greater than End-speed then as indicated in block 163 tests are
made to see if the actual speed is greater than an arbitrary
minimum, if so then the acceleration speed is decremented by one
~l~89~
l step as indicated in block 164. In this wa~ the acceleration
rate is adjusted to provide an acceleration which will achieve
the required demanded velocity within the time W-r~lp. This
process is effected for each half cycle.
Looking now at figure 15 which is the flow chart to maintain
the plateau speed. The timer has been set to T-flat which
initially was the timing calculated in block 151 of fieure 13.
At point 127 figure 9 the speed demand rate is set to End-speed
and the motor is intended to just maintain that speed. If the
motor is not up to speed or above speed by this method t~le motor
will automatically settle to the End-speed~ The position error
counter is also adjusted for whatever overshoot is required ard
this is illustrated in the flow chart of figure 15 where a test
is made by the microcomputer in block 165 to see if the
acceleration speed is greater than the End-speed if not, then no
adjustment is made as indicated in block 166. I~ it is greater,
then the position error counter is adjusted by an increment
which is a constant K times the acceleration speed minus the
End-speed. Of course if undershoot is desired the sien in this
formula would be reversed. However in practice undershoot is
not desired if the required speed is not achieved after
initialisation step. After the ad~ustment has been made in
block 173 the motor continues at its desired speed until the
timer counts down~to zero as shown in block 174 At this stage,
which is point 122 on the curve of figure 9, power is cut off to
the motor. It is to be noted that the question of compensation
is one where if there is a large load of clothes then
- 62-
l acceleratio~ speed will be m~ch greater than the End-speed and
an overshoot curve such as that of 131 or 132, figure 11, will
be followed and the result of this is that the stroke angle will
increase slightly as the load increases. The higher the load
the slightly greater the stroke angle and this has an improved
effect in maintaining a wash rate substantially con~ctant as
between a light and a heavy load. It is noted that the stroke
time is maintained but the stroke angle increases. With a
traditional agitator washing machine with an induction motor it
has a fixed speed so that not only does the stroke time stay the
same but the stroke angle is virtually always constant although
under heavy lo2d it mâ~ reduce slightly. With a traditional
machine the actual stroke profile virtually does not change with
load. The power that goes into the load increases but only
sufficient to maintain that pr~file. The present invention
modifies the profile in accordance with the lOâd and that is
novel. Thus in modifying the profile the present invention
actually overadapts the acceleration power tc &ive an overshoot
to ~ive a greater area under the curve of figure 9 and thus
apply extrâ power where there is a heavier load which is a
desired result of the present invention. Thus the time to coast
to zero is an indirect measure of the load on the agitator.
Hâving reached point 122 and the timer indicated in block
171 has timed out as shown in block 17l1 a coasting time of 180
milliseconds (just greater than the expected coast time) is
chosen as shown in block 175 (figure 16) the motor turred off as
in block 176 and then the motor coasts and the agitator will
- 63-
9~S
l slow down under the load imposed by the clothes and other
frictional effects.
The microcomputer waits on speed to fall to zero or the
timer to empty to zero as shown in block 177 whether the timer
reaches zero or not is tested in block 178. If the timer equals
zero then braking is effected as chown in block 179 and the
coding T-flat - initial T-flat selected in the microcomputer ac
shown in block 180. In such a case the mokor is restarted under
circumstances above outlined in which it may restart in the
right direction or the wrong direction at random and force
com~utating is necessary as described above. If the timer dces
not eoual zero, the microcomputer is programmed to T~flat which
equals the remainder of the 1inear time plus the initial T-flat
time. The time to coast to zero speed is an indirect measure of
the load on the agitator. The position and speed of the rotor
is measured and the information supplied to the microcomputer as
is described above.
As described above~ while the rotor is coasting, EMFs are
generated in the one or more unused windings and these EMFs can
be sensed to indicate when an EMF changes sense i.e. crosses
over a zero point. However other position or velocity and
direction sensing devices can be provided e.g. Hall effect
devices or light intercepting devices or with non ECM type eg.
brush, indiction or synchronous motors, it is still possible to
measure the EMFs. However with such motors we do not r.eed to
know position, only speed~ Thus the microcomputer senses when
the rotor is approaching a position in which it is ln condition
- 64-
8~i
l for reversing and the time taken to reach this position is
measured and used in calculating a new value of T-flat for the
next half cycle. This is effected by taking the remainder of
the timer of block 171 and if this time is not zero then the
rotor has reduced to zero speed in less than 150 milliseconds.
Thus the calculation shown in block 151 (figure 13) of T-stroke
minus W-ramp minus 150 milliseconds is modified by taking away
the difference between 150 milliseconds and the actual time
taken for the rotor to come to zero and that provides a new
calculation for the plateau time which is substituted for that
shown in block 151. However if the timer does get to zero in
block 178 then the rotor is braked to stop in block 179 and the
T-flat selected to be used is the initial T-flat as indicated in
block 180. Once the rotor is at or near stopped, then unless
the rotor has been bra~ed to a stop a~ illustrated in block 17
then for an EC~I, reversal is usually effected in a single
commutation period as is deccribed above.
In the event that agitation is to cease as illustrated at
145 in figure 12 then other parts of the washing cycle take over
for example the drain is opened and the water allowed to drain
out. As described above, the ~oasting time of a previous half
cycle is algebraically subtracted from the stroke time to give
the "power on" time for the next half cycle. However different
adjustments are possible e.~. only every tenth or other nu~ber
half cycle could be used to make the adjustment or the coasting
times over a period, eg. over one second averaged tc give a
'power on" time for the next second~
_ 65-
l An important aspect of the invention resides in the
measuring of the coast time from the stroke time to give a
~'power on" time for the next half cycle. Thus although this
invention ha~ been described in relation to an electronic2lly
5 Commutated motor which gives added advantages in controllin~
acceleration rates and maximum speeds, an important advantage of
the invention is this aspect can be gained using other motor
types for example an induction motor. Such a motor may only be
accelerated in a manner broadly dependant on the number of poles
in its rotor and the load. However by controlling the cut off
point 122 at which power is applied to the motor by subtractin&
the coast time of one ~.alf cycle from the stroke time to give an
acceleratior. time and plateau time for the next half cycle,
considerable control is given to the rate of extractin~ dirt
consistent with a desired de~ree of gentleness of waching.
Thus referring to figure 16a, a speed sensor driven by 2
rotor has a rin~ nlagnet 71 the multiple holes of which actuate a
Hall effect transducer 72, the signals from which are in the
form of pulses which vary in lines according to the speed of
rotation of the ring magnet 71. When the pulse time reaches a
predetermined len~th of time, reversing is effe~ted.
` Also photo sensitive devices can be used, for example as
described in U.S Patent Specification 4,005,347~ In
either case, the time between switching off power to the
motor and the motor being in condition for reversing is
measured and used in a next half cycle to determine
the "power on" time which will give the
- 66-
9~L8~
1 re~uired washing action.
Although the above descriptions are based on using a fixed
stroke time/ the invention in this as~ect can also be put into
effect with a variably stroke time operation.
Thus where the stroke time is to be variable according to
the 102d in the washing tub and referrin to figure 9a, which is
similar to figure 9, the acceleration time 81 plus the plateau
time 82 are set by the operator according to a required
gentleness Gr vigorousness of washine to a fixed ~'power on"
time. A small load will give a coast time indicated between
pcints 83 and 84 with a delay curve 85. A large load gives a
steeper delay curve 86 hith a coast time indic~ted between
~oints 83 and 87 and accordin~ly the motor hill be in condition
for revercal much earlier than in the li~t load coast time
curve 85. If reversing is thus effected with a shortened stroke
ti~e more consistent washing performances will be obtained,
whether the load is small or lar~e.
This application is a divisional of co-pending Canadian
application Serial No. 518,195 filed September 15, 1986.
- 67
