Note: Descriptions are shown in the official language in which they were submitted.
:~75~
- 1 - 20-TR-1420
ELECTRIC POWER INVERTER WIT~ ADAPTIVE
T~lIRD ~IARMONIC AIJXILIARY IMPULSE COMMUTATION
Background of the Invention
This invention relates to electric power
inverters for converting direct current (d-c) to
polyphase alternating current (a-c), and more
particularly it relates to improvements in the
old and well known current-fed "third harmonic"
auxiliary lmpulse commutated inverters. The
principles of commutation and a typical practical
application of such inverters ~ere described in a
technical paper entitled: "Analysis of a Novel
Forced-Commutation Starting Scheme for a Load-
Commutated Synchronous Motor Drive," which paper
was presented by R. L. Steigerwald and T.A. Lipo
at the IEEE/IAS annual meeting held in Los Angeles,
Calif. on Oct. 2-4, 1977. The Steigerwald and
Lipo paper was reprinted in IEEE T~ANS. Vol. IA-15,
No. 1~ Jan/Feb 1979, pgs. 14-24.
In essence, a third harmonic auxiliary
impulse commutated inverter comprises six main
unidirectional conduction controllable electric
valves, such as thyristors, that are interconnected in
pairs of series aiding, alternately conducting valves
20 TR-1~20
-- 2 --
to form a conventional 3-phase, cdouble-way, 6-pulse
bridge between a pair of d-c terminals and a set of
three a-c terminals. The d-c terminals of the b~idge
are adapted to be co~nected to a suitable source of
relatively smooth direct current. A large, multicell,
heavy duty electric storage battery is a suita~le
current source, as is the combination of an electric
power rectifier to which an alternatiny voltage source
is cormected and a current smoothing reactor or choke
in the d-c link between the d-c terminals of the
rectifier and inverter, respectively~ The a-c
terminals of the aforesaid bridge are respectively
connected to the different phases of a 3-phase
electric load circuit which typically comprises
star-connected 3-phase stator windings of a
dynamoelectric machine such as a large synchronous
motor.
To supply the load circuit with 3-phase
alternating current, the six main valves of the
inverter are cyclically turned on (i.e., rendered
conductive) in a predetermined sequence in response to
a family of "firing" signals (gate pulses) that are
periodically generated in a prescribed pattern and at
desired moments of time by associated control means.
To periodically turn off the main valves by forced
commutation, the inverter is provided with an
auxiliary circuit comprising a precharged commutation
capacitor and at least seventh and eighth alternately
conducting unidirectional controllable electric valves
that are arranged to connect the capacitor between the
neutral or common point o-f the 3-phase a-c load
circuit and either one of the d-c terminals of the
bridge.
During each full cycle of steady state
operation of a third harmonic inverter, each vf the
20-TR-1~20
-- 3 --
valves in the auxiliary commutation circuit is briefly
turne~ on three separate times. More particularly,
the 7th valve is fired at intervals of approximatel~
120 electric degrees, and the 8th valve is fired at
similar intervals that are staggered with respect to
the intervals of the 7th valve, whereby one or the
other auxiliary valve is fired every 60 electrical
deyrees. When an auxiliary valve is turned on, the
commutation capacitor is effectively placed in
parallel with one phase of the load circuit and a
first one of the two main valves which are then
conducting load current. Initially, the capacitor
voltage magnitude is higher than the amplitude of the
line-to-neutral voltage that is developed across the
inductive load, and its polarity is such that the
capacitor starts discharging. Consequently current is
forced to transfer (commutate) from the first main
Yalve (i~e., the offgoing or relieved valve) to a
parallel path including the turned-on auxiliary valve
and capacitor. The rate of change of current du~ing
commutation will be limit~d by the load inductance.
After current in the offgoing main valve
decreases to zero, the magnitude of capacitor voltage
is still sufficient to keep that valve reverse biased
for longer than its "turn-off time." As soon as the
commutation capacitor is fully discharged, load
current begins recharying it with opposite polarity.
Once the commutation capacitor is recharged to a
voltage magnitude exceeding that of the
line-to-neutral load voltage, the next (oncoming) main
valve in the bridge is forward biased and can be
turned on, whereupon load current commutates from the
turned-on auxiliary valve and commutation capacitor to
the oncoming main valve. This causes the auxiliary
valve to turn off and completes the commutation
'5~
20-TR-1420
-- 4 --
process. The capacitor is left with voltage of proper
polarity and sufficient peak magnitude for successful
commutation of the seconcl one of the first-mentioned
two conducting main valves when the opposite auxiliary
valve is turned on approximately 60 degrees later. It
will be apparent that there are six intervals of
commutation per cycle, the direction of current in the
commutation capacitor during each interval is reversed
compared to the preceding inter.val, and therefore the
fundamental frequency of the alternating capacitor
current equals the third harmonic component of load
frequency.
As is pointed out in the referenced
Steigerwald and Lipo paper, one practical application
of a current-fed third harmonic auxiliary commutated
inverter is in an adjustable speed a-c drive system
where the 3-phase star-connected stator windings of a
synchronous machine are supplied with variable
frequency a-c power by the inverter which needs to be
forced commutated in order to start the m~chine. In
such an application, for reasons explained in that
paper, a technique of "delayed gating" is used to
ensure that at the end of each commutation interval
the commutation capacitor has recharged to a
sufficiently high level of voltage to guarantee
successful commutation during the succeeding interval.
According to this technique, the sequential firing
signals for the main valves are each delayed, after
the oncoming valve is forward biased, while the
capacitor continues accumulating and storing
electrostatic charge until its voltage attains a
threshold level required for extinguishing current in
the next offgoing valve. Steigerwald and Lipo suggest
that the threshold level can be proportional to the
magnitude of source current so that the peak magnitude
~ 6~
20-TR-1420
-- 5
of capacitor voltage (and hence the commutating
ability of the inverter) will desirably track -the
demands of the load. In other words, the magnitude of
capacitor voltage is high when necessary to commutate
high current and is relatively low when only light
load needs -to be commutatec~. This "adaptive"
commutation capacitor voltage t:echnique advantageously
reduces comrnutation time and power losses during the
commutation intervals when the magnitude of load
current is relatively low.
In his prior art U.S. Patent No.~,244,017 -
Steiyerwald issued January 6, 19~1, discloses and
claims a modified third harmonic auxiliary impulse
commutated inverter having parallel commutation
circuits which allow three different values of
commutating capacitance to be actively selected as a
function of the magnitude of load current. Assuming
that the inverter is supplying a synchronous machine
load, current tends to decrease as frequency (i.e.,
rotor speed) and hence machine back emf increase. By
switching to a commutation capacitor of smaller size
when the current magnitude falls below a preset level,
the commutation time is desirably shortened at light
loads. As a result, the maximum permissible
fundamental frequency is increased, and the operating
range of the inverter is extended.
A current-fed third harmonic inverter is
well suited for supplying variable frequency
alternating current to the 3-phase stator windings of
a rotatable synchronous machine that is used to start
or "crank" a prime mover such as a large internal-
combustion engine. In such a system, the rotor of the
machine is coupled to a mechanical load comprising the
crankshaft of the engine. Initially the output torque
of the rotor (and hence the magnitude of current in
~" ~d ~
20-TR-1420
-- 6 --
the stator windings) needs to be relatively high in
order to start turning the crankshaft. As the rotor
accelerates from rest, less torque (and current) will
be required, while the fundamental fre~uency of load
current increases with speed (revolutions per Minute).
In its cranking mode of operation, the inverter
supplies the machine with current of properly varyiny
magnitude and frequency until t:he engine crankshaft is
rotating at a rate that equals or exceeds the minimum
speed at which normal runniny conditions of the engine
can be sustained. It should be apparent that the
above-mentioned adaptive commutation technique,
~lerein decreasing current is accompanied by lower
commutation voltage and hence shorter commutation
intervals, will desirably raise the upper limit of the
permissible range of inverter operating frequency.
Sunmary of the Invention
A general objective of the present invention
is to provide a third harmonic auxiliary impulse
commutated electric power inverter characterized by an
improved adaptive commutation feature that is simple
yet effective.
Another objective is the provision, for sucl
an inverter, of an adaptive commutation -feature that
does not require means ~or sensing or measuriny
current magnitude.
In carryiny out the invention in one form, a
source of relatively smooth direct current is
connected to a 3-phase inductive load circuit by means
of an electric power inverter comprising at least
three pairs of alternately conducting main
controllable electric valves arranged in a 3-phase,
double-way, 6-pulse bridge configuration. For
commutating the main valves, a precharged commutation
capacitor is connected between the load circuit and
20-rrR--1~0
-- 7 --
the juncture of first and second auxiliary
controllable electric valves that are interconnected
in series aiding fashion across the d-c source.
Bistable volta~e sensing means is coupled to the
capacitor; it has one state whenever the electrical
potential on one side o-f the capacitor is measurably
positive with re~pect to the other side, and otherwise
it has a different state. A zero-crossing detector is
coupled to the 3-phase ]oad circuit for detectin~ all
zero crossin~s of the fundamental phase-to-phase
alternating voltages that are developed at line
terminals of the respective phases of the loacl.
The above-summarized inverter also comprises
control means coupled to both the voltage sensing
means and to the zero-crossing detector for cyclically
producing a ~amily of periodic firing siynals that
cause the valves to turn on selectively. In a third
harmonic commutation Mode of operation, the aforesaid
family includes a first series of firing siynals
respectively produced in response to consecutive zero
crossings of the phase-to-phase voltages for
alternately turning on the first and second auxiliary
valves, wher~upon load current can immediately
transfer from an offgoing main valve to a parallel
path including the turned-on auxiliary valve and the
commutation capacitor which is first discharged and
then recharged with reverse polaxity by this pulse of
current. The family also includes a second series of
firing signals respectively produced in delayed
response to successive state chanyes of the voltage
sensing means for turning on the main valves in a
predetermined sequence, whereupon load current can
then transfer to the oncoming main valve from the
turned-on auxiliary valve. In accordance with the
present invention, the control means includes time
20-TR-1~20
-- 8 --
delay means effective at least after a predetermined
initial period of time, which starts when the third
harmonie commutation mode of operation commenees, for
delaying the production of each firin4 signal in the
second seri~s until a programmed interval of time has
elapsed following each state ehange of the voltage
sensing means. ~uring -the pro~rammed delay interval,
load eurrent will continue to reeharge the commutation
eapaeitor at a rate that is a function o~ its
magnitude. Preferably, eaeh delay interval has a
fixed, relatively short duration. As a result, during
eaeh commutation interval the eapaeitor is
automatically eharged to a peak vo]tage level that
varies with current magnitude. Thus, as load current
decreases the eommutation intervals are desirably
reduced and the upper frequeney limit of the inverter
is correspondingly inereased.
The invention will be better understood and
its various objects and advantages will be more fully
appreciated from the following description taken in
eonjunction with the accompanying drawings.
Brief_Description of the Drawin~s
Figure 1 is a schematic diagram of a system
eomprising a rotatable eleetrical machine of the
~5 synehronous type having a rotor which is meehanieally
coupled to a variable speed prime mover and having
3-phase, star-eonneeted stator windings whieh are
eonnected to an electrie storage battery via a
plurality of controllable eleetrie valves that in turn
are intereonneeted and arranged to form a variable
frequeney third harmonie auxiliary impulse commutated
inverter;
Figure lA is a simplified block diagram of
the eontroller (shown as a single block in ~igure 1)
which cyclically produees a family of periodic firing
~j ~ 7 5 ~
20~TR-1420
_ 9 _
signals for respectively turning on the various valves
of the inverter;
Figure 2 is a time chart showing, for one
full cycle of operation in a third harmonic
commutation mode, the six poss:ible states of 3-phase
fundamental stator voltages and the family of twelve
firing signals produced by the controller;
Fiyure 3 is a larger scale time chart
showing variations in voltage and current of the
commutation capacitor during transitions from odd to
even states and from even to odd states, and also
showing the programMed intervals of the present
invention; and
Figures 4-11 are flow charts that explain
the operation of the preferred embodiment of the
Figure lA controller to produce the firing signals
shown in Figures 2 and 3.
Description of the Preferred Embodiment
The electric power system shown in Figure 1
comprises a third harmonic auxiliary impulse
commutated inverter having a pair of d-c terminals lOp
and lOn connected to a source of relatively smooth
direct current and a set of three a-c terminals 11,
12, and 13 connected, respectively, to line terminals
of three star-connected armature windings on the
stator of a rotatable, variable speed, 3-phase a-c
synchronous machine 1~ which has a rotor 15 that is
mechanically coupled to a prime mover 16. In the
illustrated embodiment of the invention, the current
source for th~ inverter comprises the combination of a
source of voltage, such as a heavy duty electric
storage battery 17, in series with impedance means
which has appreciable electrical inductance,
preferably p.rovided by the d-c field winding 18 on the
rotor lS of the macnine 14. By way of example, the
20--TR--1420
-- 10 --
battery 17 is a lead-acid or nickel-cadmium type
having 32 cells and rated 6~ volts, and the average
magnitude of voltage at its terminals normally does
not exceed 76 volts. Its internal resistance is
typically in the range of 16 to 37 milliohms. The
battery is intended to supply electric energy for
starting the prime mover, and the system showrl in
Figure 1 can successfully perform -this fLInction with
the battery voltage as low as 61 volts.
The prime mover 16 can be a conventional
thermal or internal-combustion engine, and in one
particular application of the invention it is a
high-horsepower, 16-cylinder diesel engine that is
used to provide the motive power on a large
self-propelled diesel-electric traction vehicle such
as a locomotive. The synchronous machine 14 has dual
modes of operation: as a yenerator for supplying
alternating current to an electric load circuit that
is connected to its stator windinys, and as an a-c
motor for cranking or starting the engine 16. In its
generating mode, the rotor 15 oE the machine is driven
by the crankshaft of the engine 16, and the field
winding 18 is energized by a suitable excitation
source 20 (e.g., the rectified output of auxiliary
windings on the stator of the machine 14) to which it
is connected by means of a suitable contactor K which
is closed by a conventional actuating mechanism 21 on
command. The machine 1~ now generates alternating
voltages at the line terminals of its 3-phase stator
windings. The rms magnitude of the fundamental
sinusoidal components of these voltages depends on the
angular velocity (rpm) of the rotor and on the amount
of field excitation. The generated voltages are
applied to a-c input terminals of at least one
3-phase, double-way rectifier bridge 22, and the
3~
20-T~-1420
-- 1]. --
recti~ied el~ctric power at the output terminals of
each such bridge i6 supplied via a d-c bus to one or
more d-c traction motors (not shown~. As is shown in
Figure 1, the bridge 22 comprises simple solid~state
diodes, but alternatively it could be a controlled
rectifier if desired. As is suggested by the broken
lines 23a and 23b, an additional traction motor (not
shown) could be connected between the d-c terminals
lOp and lOn if desired.
In the motoring mode o~ operation, which is
assumed throughout the remainder of this description,
the rotor 15 of the synchronous machine 14 drives the
crankshaft of the engine 16. Now electric eneryy is
supplied from the battery 17 to the windings on both
the rotor and the stator of the machine, and the rotor
15 exerts torque to turn the crankshaft and -thereby
crank the engine. As the rotor accelerates from rest,
both the frequency and the rms magnitude of the
fundamental alternating voltage waveforms developed at
the line terminals of the stator windings (i.e., the
back emf) correspondingly increase, while load current
(i.e., current in the field and armature windings)
decreases in magnitude. Once the rotor is rotating
faster than a predetermined rate, which typically is
240 rpm, the engine is presumed to be started and the
motoring mode (i.e., engine cranking mode) of
operation is discontinued. Assuming that the machine
14 has ten poles, 240 rpm corresponds to a fundamental
frequency of 20 Hertz. Thus the fundamental frequency
of alternating current supplied to the stator windings
of the machine 14 needs to increase from 0 to
approximately 20 Hz in order for the illustrated
system to per;Eorm its engine starting Eunction.
The previously mentioned third harmonic
au~iliary impulse commuta-ted inverter is operative to
~! ~7~
20-T~-1420
- 12 -
convert direct current from the battery 17 into
variable frequency alternating currents in the three
different phases A, B, and C of the 3-phase armature
windings on the stator of -the machine 14. The
inverter has at least three pairs of alternately
conducting main controllable electric valves
interconnected and arranged in a 3-phase, double-way
bridge configuration between the set of three a-c
terminals 11, 12, and 13 and the pair o-f d-c terminals
lOp and lOn. More particularly: a first pair of
valves Tl and T4 are connected in series aiding
fashion from terminal lOp to terminal lOn, and their
juncture, comprising terminal 11, is connected to
phase A of the stator windings; a second paix of
valves T3 and T6 are connected in series aiding
fashion from lOp and lOn, and their juncture,
comprising terminal 12, is connected to phase B of the
stator windings; and a third pair of valves T5 and T2
are connected in series aiding fashion from lOp to
lOn, and their juncture, comprising terminal 13, is
connected to Phase C. Each valve preferably comprises
at least one solid state unidirectional controlled
rectifier popularly known as a th~ristor. It haq a
turned on (conducting) state and a turned off
(non-conducting) state. In prac~ice the valves are
respectively shunted by conventional snubber circui-ts
(not shown). The illustrated means for connecting the
d-c terminals lOp and lOn of the inverter to the
battery 17 will next be described in more de-tail.
The first d-c terminal lOp is connected to
the relatively positive terminal of the battery 17 via
a single pole contactor Klp, and the second d-c
terminal lOn is connected to the other terminal of the
battery by means of a conductor 25, one pole K3a of a
double-pole contactor K3, the field winding 18, the
7 ~
20-TR-1420
- 13 ~
other pole K3b of the same contactor, and a
conductor 26. The -field winding 18 typically has a
resistance in the ran~e of 0.12 to 0.2~ ohm and an
unsaturated inductance of more than 0.3 henry.
single-pole contactor Kln, a conductor 27, and
resistance means comprising two resistors 28 and 29
are connected in parallel circuit relationship with
the ~ield winding 18 in the load current path between
conductors 25 and 26. The resistors 28 and 29 are in
series, and both have very low ohmic values; for
example, the resistance of resistor 2~ is
approximately 14 milliohms and the resistance of
resistor 29 is approximately 23 milliohms. The second
resistor 29 is shunted by another single-pole
contactor K2 which, when c]osed, reduces the ohmic
value o~ the resistance means to that of the ~irst
resistor 28 alone.
An inductor 30 of approximately one
millihenry inductance is connected in series with
resistors 28 and 29 between the second resistor 29 and
the conductor ~6 to smooth the current flowiny in this
branch of the load current path. The inductor 30 is
shunted by a conventional overvoltage protective
device 31 the resistance of which is normally very
high but automatically decreases to a negligible
amount in substantially instantaneous response to the
magnitude of voltage across the inductor rising to a
predetermined breakover level (e.g., 750 volts). A
similar protective device 32 with bidirectional
response is connected across the field winding 1~.
resistor 36 of significant ohmic value (e.g., 100
ohms) is also connected across the winding 1~ to
enable thyristor "latching" current to bypass the
~ield and the inductor 30 when battery current starts
flowing to precharg~ the inverter's commutation
20-TR-1420
- 14 -
capacitor as will later be explained.
~ S i5 shown in Figure 1, a battery
charger 33, in series with a blocking diode 3~ and a
circuit breaker 35, is connected across ~he
combination of battery 17 and inductor 30. With the
engine 16 running under steacly-~state conditions, the
battery charger holds the battery voltage at
approximately 74 volts. It can be energized from any
suitable source, such as auxiliary windinys (not
shown) on the stator of the synchronous machine 14.
With the field winding 18 in the load
current path during engine cranking, the synchronous
machine 14 will operate with a characteristic similar
to that of a series d-c motor, namely, high current
and hence desirably high starting torque at low
speeds. The resistance means 28, 29 in parallel with
the field reduces the ohmic value of resistance that
the field winding alone would otherwise introduce in
this path, thereby initially allowing a higher
magnitude of armature current and later, as speed
increases, providing automatic field weakening which
permits the machine to run at a higher speed.
Initially load current is limited by the internal
resistance of the battery 17 as well as other
resistance in its path, and as speed increases it is
limited by the back emf of the armature (i.e., stator~
windings. Thus load current and torque tend to
decrease with increasing speed. A short time after
cranking commences, the contactor K2 is closed to
further reduce the amount of resistance in parallel
with the field, thereby permitting more load current
to flow and hence more torque to be developed at
higher speeds compared to the quantities that wo~lld be
achieved if the parallel resistance were not so
reduced.
5 ~
20-TR-1420
- 15
~ hen the cranking mode of operation
commences, the contactor K is open, and all o~ the
contactors in the load current path between the
battery 17 and the d-c terminals lOp and lOn are
closed excep-t K~. In a manner that will soon be
explained, contactor K2 is commanded to close upon the
expiration of a predetermined length of time after
cranking commences. Thereafter, in response to the
speed of the engine attaining a threshold (e.g.,
240 rpm) that marks the conclusion of cranking and
therefore the succesful starting of the engine 16, all
of the previously clo~ed contactors are opened. Upon
opening contactor K3 the field winding 18 is
disconnected from the load current path between the
conductors 25 and 26, and the contactor K is then
closed by its actuating mechanism 21 in order to
reconnect the field to the normal excitation source 20.
Each of the four contactors Klp, Kln, K2,
and K3 has an associated actuating mechanism that
determines its closed or open status. All four such
mechanisms are represented in Figure 1 by a single
block 38 labeled "Contactor Drivers," and they
respectively respond to opening/closing signals
received over lines 40, 41, 42, and 43 -from another
block 44 labeled "Controller." The controller 44
knows the actual status of each contactor by virtue of
feedback signals that it receives from conventional
position indicators (not shown) that are associated
with the separable contact members of the respective
contactors, as represented symbolically by broken
lines in Figure 1.
In order to turn on each of the controllable
valves Tl through ~6 in the inverter, an appro~riate
signal is applied to the associated gate while the
main electrodes of that valve are forward biased
20-~rR-l420
- 16 -
(i.e., anode potential is positive with respect to
cathode). Such a signal is sometimes called a trigger
or gating signal, and it is herein referred to
generically as a "firing signal." In a manner soon to
be described, the controller 44 cyclically produces a
series of periodic firing signals for turning on the
respective main valves Tl-T6 in numerical order. lIt
is assumed that the alternating voltages developed at
the line terminals of the 3-phase stator winding6 of
the machine 14 have the conventional A-B-C phase
rotation.) In order to quench or turn off each valve
when desired, the inverter has a forced commutation
circuit including at least first and second auxiliary
controllable electric valves Tp and Tn interconnected
in series aiding fashion between the d-c terminals lOp
and lOn and connected via a commutation capacitor 45
to the stator windings of the machine 14. The
capacitor 45 is shunted by a bleeder resistor 46 which
effectively keeps the capacitor initially in a
substantially discharged state prior to closing the
contactors Klp and Kln and starting up the illustrated
system. Preferably, the commutation capacitor is
connected between the juncture M of the auxiliary
valves and the neutral S of the three star-connected
stator windings.
In the manner previously explained under the
heading "Background of the Invention," the main valves
Tl-T6 in turn are forced to turn off by the
commutation action that is initiated each time one or
the other of the auxiliary valves Tp and Tn is turned
on. The controller 44 is arranged cyclically to
produce a series o~ periodic firing signals for
alternately turning on the two auxiliary valves in
synchronism with the variable frequency ~undamental
component o~ the sinusoidal phase-to-phase alternating
5~
20~TR-1420
- 17 -
voltages that are developed at the line terminals of
the respective phases A B, and C of the stator
windings as the field winding 18 rotates inside the
stator of the machine 14~ L~ote that the peak
magnitude of reverse voltage imposed on the auxiliary
valves can be reduced, if desired, by respectively
inserting simple diodes in series therewith.
To produce the valve firing signals at
proper times, the controller 44 needs to receive fron
the power s~stem information or data indicating when
the fundamental waveforms of line-to-neutral magnetic
flux in the three phases A, B and C of the machine 14
cross zero and change polarity, and indicating the
status of the electrostatic charge or voltage on the
commutation capacitor 45. Such data are supplied by
means of a voltage processor 38 which, as can be seen
in Figure 1, has a plurality of input wires
respectively connected to the line terminals o-f the
s-tator windings and to opposite sides of the
capacitor 45. Inside the processor 48 there is
bistable first means for sensing the electrical
potential difference across the commutation
capacitor. Whenever the potential at the juncture M
is measurably positive with respect to the neutral S,
the first means is in one state and provides a
discrete signal (VEl) that is high or "1," but when
this potential is measurably negative with respect to
neutral the first means is in a diffarent state in
which the output signal VEl is low or "0." Voltage
sensors suitable -for this purpose are well known and
readily available to a person skilled in the art~ The
signal VEl is supplied over an output bus 50 to the
controller 44. ~n additional bistable voltage sensing
means is provided in the voltage processor 48 for
detecting whether or not the capacitor voltage has a
5 ~ ~ ~
2~-TR-142
- 18 -
magnitude exceeding a predetermined level, either
positive or negative. In one practical application of
the illustrated system, the predetermined level is 400
volts. The additional sensor produces a discrete
signal (VE2) on the output bus 50. As the commutation
capacitor charges or recharges to a voltage magnitude
in excess of the predetermined leve, the signal VE2
changes from a "0" to a "1" state.
The voltage processor ~8 also includes
suitable means for integrating the respective
line-to-neutral voltages of the stator windings and
for indicating whether the polarity of the integral is
positive or negative. The latter means provides three
discrete output signals XA, XB and XC which are
respactively supplied over lines 51, 52 and 53 to the
controller 44. The output signal XA is high or "1"
only during the half cycles that the time integral of
the voltage between the line terminal of phase A and
the neutral S is relatively positive. It will be
apparent that up and down changes of XA coincide with
successive zero crossings of both the magnitude of
line-to-neutral flux in phase A and the magnitude of
the fundamental phase to-phase alternating voltage
developed at the stator line terminals of phases C
and B (i.e., the line terminals to which the a-c
terminals 13 and 12 of the inverter are respectively
connected). Similarly, the output signal XB is "1"
only during the half cycles that the integral of the
phase B-to-neutral voltage is relatively positive,
whereby up and down changes of XB coincide with
successive zero crossings of both the magnitude of
line-to-neutral flux in phase B and the magnitude of
the fundamental phase-A-to-phase-C alternating voltage
developed at the stator line terminals to which the
a-c terminals 11 and 13 are connected. In a similar
5~
20-~rR-l420
lg -
manner, the output signal XC is "1" only during the
positive half cycles o-f the integral of the phase
C-to neutral voltage, whereby the up and down changes
of XC coincide with successive zero crossings of both
the magnitude of line-to-neutral flux in p~ase C and
the magnitude of the fundaMental phase 8-to-phase-A
alternating voltage developed at the stator line
terminals to which a-c terminals 12 and 11 are
respectively connected. By logically processing the
resulting output signals XA, XE~, and XC, the si~
different combinations of relative polarities o~ the
three phase-to-phase voltages are indicated duriny
each cycle of operation. Each time -the magnitude of
any of these voltages crosses zero, a different one of
the output signals changes either from 0 to 1 or
from 1 to 0~
The controller 44 also communicates with
master controls 54 via input and output busses 55
and 56. A starting switch 57 is associated with the
master controls 54. The starting switch 57 can be
either a pushbutton type or a turn-and-hold type.
The presentl~ preferred embodiment of the
controller 4~ is shown in more detail in Figure lA.
Its main component is a microcomputer 60. Persons
skilled in the art will understand that the
microcomputer 60 is actually a coordinated system of
commercially available components and associated
electrical circuits and elements that can be
programmed to perform a variety of desired functions.
It typically comprises a central processing unit (CPU)
which executes an operating program permanently stored
in a read-only memory (ROM) which also stores tables
and data utilized in the program. Contained within
the CPU are conventional counters, registers,
accumulators, flag flip flops, etc. along with a
20-TR-1420
- 20 -
precision oscillator which provides a hiyh-fre~uency
clock signal. The microcomputer also includes a
random access memory (I~M) into which data may be
temporarily stored and from which data may be read at
various address locations determined by the program
stored in the ROM. The CPU, ROM, and RAM are
interconnected by appropriate address, data, and
control busses. In one practical embodiment of the
invention, an Intel 8031 microprocessor is used.
The other blocks shown in Fiyure lA
represent conventional peripheral and interface
components that interconnect the microcomputer 60 and
the external circuits of Figure 1. More particularly,
block 61 is an input/output circuit (I/O) for
connecting the output bus and lines 50~53 of the
voltage processor 4~ to the microcomputer 60, and
block 62 is another I/O for connecting the
microcomputer 60 to the contactor drivers 3~O
Block 63 is suitable means for decoding the position
indicators that are respectively associated with the
contactors Klp, Kln, K2, and K3. Block 64 is a gate
pulse generator (GPG) and bufer that produces, on
command of the microcomputer ~0, properly shaped and
isolated firing signals that turn on the respective
valves Tl-T6, Tp, and Tn.
The operation of the controller 44 during
engine cranking can best be understood with the aid of
Figures 2 and 3. In Figure 2, the sinusoidal
waveforms of the Eundamental components of the three
line-to-neutral voltages VAs, VBs, and Vcs of
the 3-phase stator windings of the machine 14 are
depicted by solid-line traces for a full cycle of
steady-state operation, and the integral of one such
~Javeform (i.e., phase A) is shown by the broken-line
trace. This integral is known t~ be in phase with the
20-T~ L420
- 2L -
flux of phase A. Assuming a s~mmetrical 3-phase
machine and balanced loading, the zero crossings of
the integral of VAs are seen to coincide with the
moments of equality between VBS and Vcs, that is,
with the zero crossings of the instantaneous magnitude
of the fundamental phase-to-phase alternating voltage
between the line terminals of phases B and C.
Consequently the discrete signal XA on output line 51
of the zero crossing detecting means in the voltage
processor 48 is "1" throughout each half cycle o~
relatively positive polarity of the phase C-to-phase B
voltage and is "0" throughout each relatively negative
half cycle thereof, the signal XB on the output line 52
is "1" throughout each relatively positive half cycle
of the phase A-to-phase C voltage but is otherwise "O,"
and the output signal XC is "1" only during each
relatively positive half cycle of the phase B-
to-phase A voltage. The six different states of these
three signals are marked off and numbered consecutively
in Figure ~. For example, state 1 exists so long as
both XA and XB but not XC are "1", whereas state 2
exists while X~ alone is "1". A state change is
experienced each time any one of the signals X~, XB,
or XC changes up or down, and each state coincides
with a different 60-degree segment of a full cycle
(360 electrical degrees) of the fundamental component
of alternating voltage.
In a manner that will soon be described, in
its third harmonic commutation mode of operation the
controller 44 automatically responds to consecutive
state changes of the signals XA, XB, and XC by
producing a series of firing signals (which are
represented by the pointers 68 and 6~ in Figure 2) for
alternately turning on the two auxiliary valves Tp
and Tn. More particularly, the controller is
7s~
20-Tl~-1420
- 22 -
effective to produce a firing siynal 69 for turning on
the auxiliary valve Tn in immediate response to each
change from an odd numbered state to the succeeding
even numbered state (i.e., from state 1 to state 2,
from state 3 to state 4, and from state 5 to state 6),
and it is effective to produce a firing signal 6~ for
turning on the auxiliary valve Tp in immediate
response to each change from an even numbered st~te to
the succeeding odd numbered state (i.e., from state 6
to state 1, from state 2 to st~te 3, and from state 4
to state 5). By thus synchronizing the firing
signals 68 and 69 with the state changes (which are
determined by the angular location oE the rotor 15 in
the machine 14), the angle between the field mmf and
the stator mmf of the machine is controlled.
As was previously explained, turning Oll an
auxiliary valve causes load current immediately to
transfer Erom an offgoing main valve to a parallel
path including the turned-on auxiliary valve and the
precharged commutation capacitor 45 which is first
discharged and then recharged with reverse polarity by
such current. As is indicated in Figure 2, the
polarity of the capacitor voltage will change from
positive (i.e., the potential at the juncture M is
positive with respect to t~e neutral S) to negative as
a result of the auxiliary valve Tn being turned on by
one of the ~iring signals 69, and it will change from
negative to positive when the auxiliary valve Tp is
turned on by one of the firing signals G8.
Following the production of each of the
firing signals 68 and 69, the controller 44
selectively produces the next one of a series of six
firing signals (represented in Figure 2 by the
pointers 71-7~) which are applied to the gates of the
main valves Tl-T6, respectively. The controller
20-Tl~ 20
- 23 -
selects the proper firing signal to turn on whichever
one of the main valves is associated with the oncoming
or relieving phase of the stator voltages, whereupon
load current ~an then transfer to the oncominy valve
from the turned-on auxiliary valve. More
particularly, as indicated in Figure 2, the controller
selects the firing signal 71 for turning on the main
valve Tl if the preceding state change was from
state 6 to state 1, it selects the firing signal 72
1~ for turning on the main valve T2 if the preceding
state change was from state 1 to state 2, it selects
the firing signal 73 for turning on the main valve T3
if the preceding state change was from 2 to 3, it
selects the firing signal 74 for turning on the main
valve T4 if the preceding state change was from 3
to 4, it selects the firing signal 75 for turning on
the main valve TS if the preceding state change was
from 4 to 5, and it selects the firing signal 76 for
turning on the main valve T6 if the preceding state
change was from 5 to 6. Whichever one of the firing
signals 71-76 is selected, it is not actually produced
until after the first-mentioned bistable capacitor
voltage sensing means in the voltage processor 48
changes state, as indicated by an up or down change of
the discrete signal VEl on the output bus 50 of the
processor 48. This is best seen in Figure 3 which
will now be described.
Figure 3 shows the instantaneous magnitudes
of capacitor voltage (VMs) and current (icAp)
during two consecutive commutation intervals. The
first of these two intervals is initiated at time tl
when the means for detecting the zero crossings of
phase-to-phase voltages changes from an odd state to
an even state and the controller responds by producing
a firing signal 69 to turn on the auxiliary valve Tn,
7 ~., ~
20-rrR-l420
- 24 -
and the second commutation interval is initiated at
time t6 when a firing signal 68 for the auxiliary
valve Tp is next produced in response to the same
detector changing from even to odd states. Once the
first commutation interval is initiated, current
begins to increase in the auxiliary valve Tn and in
the commutation capacitor 45, while current in th0
offgoing main valve decreases to zero at time t2 which
occurs when all of the load current has transferred to
the parallel commutation circuit. The resulting pulse
of current in the commutation capacitor first
discharges it and then recharges it with reverse
polarity. At time t3 the capacitor is fully
discharged, whereupon the discrete signal VEl chanyes
from its initial "1" state to a different state. The
time from t2 to t3 is the circuit turn off time during
which the offgoing main valve recovers its ability to
withstand reapplied forward voltage. The next
one (7X) of the series of firing signals 71-76 for the
main valves Tl-T6 is produced in response to the 1-to-0
change of VE1.
In accordance with the present invention,
the controller 44 includes time delay means effective
at least after a predetermined initial period of time
for delaying the production of the next firing
signal 7X until a programmed interval of time has
elapsed following the 1-to-0 change of VEl~ This
interval is designated by the delta t- symbol in
Figure 3, and it elapses at time t4. Now the oncoming
main valve TX is turned on, and load current begins
transferring to it from tha parallel commutation
circuit while continuing to recharge the
capacitor 45. At time t5 all of the load current has
transferred to the oncoming main valve, the auxiliary
valve Tn turns off, and the first commutation interval
) r;~3~
20-T~-1420
25 -
is finished. The peak magnitude of capacitor voltage
(i.e., its magnitude at time t5) is a function of the
magnitude of load current that recharges the capacitor
during the delta t- delay interval. As load current
decreases, the peak magnitude of capacitor voltage
will decrease, and consequently the length of the
commutation interval is desirably reduced. From t5
to t6 the commutation capacitor will retain a voltage
of su~icient magnitude and proper polarity (neyative)
to ensure successful commutation when the next zero
crossing of phase-to-phase voltages initiates the
second commutation interval.
As is apparent in Figure 3, the ~.econd
commutation interval is essentially a dual o~ the
first. In this case the programmed delay interval
(i.e., the time from the 0-to-1 change of the discrete
signal VEl to the production o~ the firing signal 7X
for turning on the next oncoming main valve) is
designated delta t+, and its duration can either be
the same as or differ from the duration of delta t-.
In practice, the programmer will select delay
intervals that are compatible with the parameters of
the power system and the ratings of its components.
In the preferred embodiment, after a predetermined
initial period of time from the start of engine
cranking, alternate delay intervals (delta t-) are
programmed to have a shorter duration than
intermediate delay intervals (delta t~).
Figures 4 through 11 display flow charts o-f
the presently preferred programs that are executed by
the microcomputer 60 in the controller 44 in order to
produce ~iring signals that enable the inverter to
operate in a third harmonic commutation mode for
purposes of cranking the engine 16. The Main Routine
is shown in Yigure 4. It begins at the entry poirlt
S ~
20~T~-1420
- 26 -
labeled "Start". ~len commanded to start, the first
step 80 oE the Main Routine is to initialize the
various inputs to the microcomputer 60, to reset its
timers to 0, to decrement its counters to 0, and to
set the stack pointers, registers, latches, outputs,
and variable values of the microcomputer to their
respective quiescent states or normal levels at the
start of the first pass through the Main Routine.
Upon completing this initializing step, the program
determines, at a decision point ~31, whether or not the
start switch 57 is "on". Assuming the start switch is
turned on or closed (which will happen at a time when
the rotor of the synchronous machine 14 i5 at rest and
the commutation capacitor 45 is discharged), the
control proceeds to a step 82 in which a first timer
is started~ By way of example, this timer will run
for approximately 90 seconds after being started.
Once timer ~1 has been started, a Set-Up Routine 83 is
executed, and this is followed by the execution of a
Normal Cranking Routine 84.
The Set-Up Routine 83 is shown in Figure 5.
Its purpose is to control the contactors and valves of
the power system (Figure 1) so as to: (1) precharge
the commutation capacitor 45, (2) initiate a pulse of
excitation current from the battery 17 through the
field winding of ~he machine 14 and find the initial
angular position of the rotor 15, and (3) ensure that
the capacitor voltage has the right polarity for
successful third harmonic commutation once the ~ormal
Cranking Routine is initiated. The Set-Up Routine is
entered at a point labeled "Set-Up" and then proceeds
to a step 85 which causes the controller 44 to issue
signals, via lines 40 43, that command the contactor
actuating mechanisms (38) to close the four contactors
Klp, Kln, K2, and K3. This step is followed by an
20-TR-1420
- 27 -
inquiry, at point 86, as to the open or closed status
of the contactors. ~nce all ~our are actually closed,
a "Contactor Error" flag is set in an "off" state, and
the control is transferred to a Capacitor Ring-up
Subroutine 87 which will soon be described. If all
four contactors do not close in response to the
closing commands of step 85, they are commanded to
open (step 88), the Contactor ~rror flag is set in its
"on" state, and the Set-up Routine is aborted at the
stop point ~9.
The presently preferred embodiment of the
Capacitor Ring-up Subroutine 87 is shown in Figure 6.
While this subroutine is being e~ecuted, the
controller 44 will produce firing signals that cause
the Figure 1 system to operate in a capacitor
"ring-up" mode that precharges the commutation
capacitor. It is entered at a point labeled "Ring-up"
and then proceeds to a step 91 which sets a counter in
the microcomputer 60 at a predetermined maximum number
of ring cycles (e.g., 20 cycles). Step 91 is followed
by a step 92 in which a second timer is started. ~y
way of example, this timer will run for an interval of
approximately 10 milliseconds a~ter being started~
From step 92 the program proceeds to an inquiry
point 93 where the state of the first bistable
capacitor voltage sensor in the voltage processor 4
is tested. If the commutation capacitor has a
mçasurably positive voltage (i~e., the juncture M has
a positive potential with respect to the neutral S and
the amount of potential difference exceeds a
predetermined threshold such as 5 volts), the first
voltage sensor is in one state (which is indicated by
VEl being high) and the inquiry yields an a~firmative
answer. On the other hand, if the capacitor voltage
3~ were measurably negative (i.e., the potential at
20-T~-1420
- 28 -
juncture M is more than 5 volts negative with respect
to the neutral S), the voltage sensor is in a
different state (as indicated by VEl being low) and
the answer to the inquiry is "no".
In response to an affirmative answer at the
inquiry point 93, the next step 94 in the program is
to instruct the controller's gate pulse generator ~4
to generate a first pair of concurrent firing signals
~or the auxiliary valve Tn and for a preselected
complementary one of the main valves (e.g., Tl). In
response to a negative answer at 93, the Capacitor
Ring~up Subroutine alternatively proceeds to a step 95
in which the gate pulse genera-tor is instructed to
generate a second pair of concurrent firing signals
for the auxiliary valve Tp and for another preselected
complementary one of the main valves (e.g., T~). In
practice, each firing signal that is generated in
step 94 or 95 can actually comprise a burst of several
high-frequency, short-duration discrete d-c signals
having sufficient magnitude to turn on the associated
valve.
Upon turning on either the complementary
pair of valves Tn and Tl or the complementary pair Tp
and T2, battery current will begin flowing through a
path which in Figure 1 is seen to comprise: (1) the
field windiny 18 in parallel with both the resistor 36
and the series combination of resistor 28, inductor 30
and the closed contactor ~2, (2) one phase of the
armature windings of the machine 14, and (3) the
commutation capacitor 45. Preferably this path has a
sufficiently high Q so that current quickly oscillates
from zero to a peak magnitude and back to zero, and in
the process the capacitor is incrementally charged
with reverse polarity. The conducting pair of valves
will automatically turn off by self commutation when
75~i9~
20-TR-1~0
- 29 -
current oscillates to zero at the conclusion of ~ach
cycle of this ringing action. The resulting pulse of
current typically has a duration of less than two
milliseconds.
In the Capacitor Riny-up Subroutine, the
status of the second timer is tested immediately after
either step 94 or step 95. This testing step is
indicated in Figure 6 by the inquiry point 96.
Assuming that timer #2 is still running, the next
step 97 in the program is to check the state o~ the
second bistable capacitor voltage sensor in the
voltage processor ~8. So long as the ma~nitude of
capacitor voltage does not exceed a predetermined
maximum (i.e., the level at which the output signal
VE2 of the second voltage sensor changes from "0"
to "1"), step 97 yields a negative answer, and the
control returns to the preceding step 96. Whenever
timer ~2 stops running (i.e., its time delay interval
is over), the control proceeds from step 96 to a
step 98 in which the count stored in the cycle counter
(see step 91) is reduced by one, and then to a step 99
determines whether or not the count has reached 0. I~
not, the control returns to step 92 t and the steps 92
through 99 are recycled. In this manner the
controller repeatedly produces the aforesaid second
pair of firing signals (for turning on Tp and T2) if
it is determined in step 93 that the capacitor voltage
is not positive, and the aforesaid first pair of
firing signals (for turning on Tn and Tl) if the
voltage is positive. The start of each such repeated
cycle of operation is delayed by an interval of time
determined by timer #2. This interval is sufficiently
long to enable the pulse of battery current to first
discharge the commutation capacitor and then
incrementally recharge it with reverse polarity until
3~
20-TR~ 0
- 30 -
the current oscillates to zero. At the end oE each
consecutive cycle, the electrostatic charge that is
stored in the capacitor (and hence the capacitor
voltage) will be progressively increased in magnitude
due to the ringin~ nature of the charging current path
(the inductance of which is provided by the fiel~ and
armature windings of the machine 14), and it will have
alternately positive and negative polarity. This
action continues for a su~-ficient number of cycles to
enable the magnitude of capacitor voltage to attain
the aforesaid maximum at which the in~uiry step 97
yields an affirmative answer, whereupon the second
timer is stopped (step 101) and, after waiting a very
short, fixed period of time (step 102), the control
proceeds to execute one final ring-up cycle of
operation.
As is shown in Figure 6, the final cycle of
the Capacitor Ring-up Subroutine is carried out by
steps 103, 104, and 105 ~hich are duplicates of the
previously described steps 93, 94, and 95,
respectively. Follo~ing the generation of the last
pair of concurrent firing signals (either for the
complementary valves Tn and Tl or ~or the
complementary valves Tp and T2), timer #2 is cleared
(step 106), and then the control returns to the Set-up
Routine (Figure 5). I'he total number of ring-up
cycles that are carried out by the Capacitor Ring-up
Subroutine is sufficient for the commutation capacitor
to charge to a voltage magnitude many times (i.e.,
more than approximately five times) higher than the
average magnitude of voltage at the terminals of the
battery 17. In one practical embodiment, ~he fully
charged voltage rating of the battery 17 is 74 volts,
and in four or five cycles the commutation capacitor
can be charged to a voltage magnitude exceeding 400
s~
20-TR-1420
- 31 -
voltsO If for any reason the capacitor were not
charged to the desired level within the maximum number
of ring cycles that was set at step 91, the count in
the cycle counter would reach 0 (step 99) before the
second capacitor voltage sensor detects maximuM
voltage (step 97), and in this abnormal event all of
the contactors are commanded to open (s-tep 107) and a
"No Ring-up" signal is issued (step 108).
After execution of the Capacitor Ring-up
Subroutine 87, and with the commutation capacitor now
precharged, the Set-up Routine 83 continues as shown
in Figure 5. The next step 110 is to start a third
timer. By way of example, this timer will run for an
interval of approximately 250 milliseconds after being
started. From step 110 the control proceeds to a
step 111 which causes the controller ~4 to command the
contactor actuating mechanisms to open both o-f the
contactors Kln and K2. X'his is followed by testing,
at point 112, the status of the third timer and by
testing, at point 113, the open or closed status of
the contactors ~ln and K2. As soon as both of these
contactors actually open, but no later than the time
at which timer #3 stops running, the control proceeds
to a step 114 in which the gate pulse generator is
instructed to generate concurrent firing signals for
turning on a preselected pair of valves that will
provide a path for excitation current from the
battery 17 through the field winding 1~ of the
synchronous machine 14. Any appropriate pair of
valves can be turned on for this purpose. The
auxiliary valves Tp and Tn were selected in the
illustrated embodiment. Once these valves are turned
on, field current starts increasing or ramping up from
zero.
S~
20-TR-1~20
- 32 -
After generating the firing signals for the
auxiliary valves Tp and Tn (step 114), the Set-up
Routine proceeds to an inquiry point llS where the
status of timer #3 is tested again~ If the time delay
interval of this timer is overl the control proceeds
Erom step 115 to a step 116 in which the contactor Kln
is commanded to reclose~ When Kln recloses, the
series resistors 28 and 29 are reconnected in parallel
with the field winding 18 (Figure 1), thereby
permitting a more rapid increase of current in the
auxiliary valves. Step 116 is followed b~ a step 117
which introduces an additional delay (e.g., a fixed
period of approximately 300 milliseconds) to allow
excitation current in the field windiny 18 to continue
increasing. Upon the expiration o~ this additional
delay (at which time current in Tp and Tn may have
attained a magnitude as high as 1,300 amperes), the
control is transferred to a subroutine 118 which will
soon be explained. Executing the subroutine 118
completes the Set-up Routine (Figure 5), and the
control will then return to the Main Routine
(Figure 4).
rrhe presently preferred embodiment of the
subroutine 11~ is shown in Figure 7. While this
subroutine is being executed, the controller 4~ finds
the initial or at-rest position of the rotor 15, and
it produces firing signals that will turn on the
proper main valve(s) for commutating (turning o~f) the
auxiliary valves rrp and Tn and for obtaining the right
polarity of voltage on the commutation capacitor.
Note that at the time this subroutine is entered,
excitation current is rising in the field winding 18
on the rotor 15 of the machine 14. Consequently the
field winding (rotor) generates magnetic flux of
increasing magnitude, and this changing flux in turn
s~j~3~
20-T~-1420
- 33 -
interacts with the three phases A, B, and C of the
armature windinys (stator) to induce therein
line to-neutral voltages that can be sensed and
integrated by the voltaye processor 48. The initial
angular position of the rotor can be deduced from
knowledge o~ the relative polarities oE the kime
integrals of these three voltages.
As is indicated in Fiyure 7, the
subroutine 118 is entered at a point :Labeled "Find
Rotor Position.O.." and then proceeds to a step 120 in
which the microcomputer 60: (1) reads data
representing the high or low conditions of the three
phase-to-phase stator voltage zero-crossing detecting
signals XA, XB, and XC on the output lines 51, 52,
and 53 of the voltage processor 48, (2) uses a loyical
combination of this data to find, in an appropriately
encoded look-up table, whichever one of the six
different states 1 through 6 (see Figure 2) is extant,
and (3) places the number of the extant or actual
state in a selected register of its memory where this
nu~ber is saved as "oldstate". From step 120 the
subroutine proceeds to a decision point 121 which
determines whether or not the oldstate is even (i.ea,
2, 4, or ~). If the oldstate is even, the next
step 122 in the program inquires as to the polarity of
the voltage on the precharged commutation capacitor
(as indicated by the high or low state of the discrete
output signal VEl from the first voltage sensor in the
voltage processor 48), and if the polarity is positive
the control proceeds to a step 123. On the other
hand, if it is determined that the oldstate is not
even (i.e., i9 1, 3, or 5), the subroutine
alternatively proc~eds from step 121 to another
inquiry point 124 which is a duplicate of the
inquiry 122, and if the polarity is not positive the
7 r~
20-~rR-l~20
- 34 -
control proceeds to the same step 123.
In s~ep 123 the microcomputer reads the
oldstate, increments it by 1 to find the number o-f the
succeeding state, and identifies the proper one ~TX)
of the six main valves that is scheduled to be turned
on in response to the next state change. For example,
if oldstate = 2, the succeeding state is 3 and TX
is T3. Step 123 is immediately followed by a ste~ 125
in which the controller 44 is i:nstructed to generate a
firing signal for the identified valve TX. Once TX is
turned on by this firing signal, it completes a path
for battery current in parallel with a first one of
the two auxi.liary valves that were turned on by
step 114 oE the Set-Up Routine 83. The parallel path
includes one phase of the stator windings and the
precharged commutation capacitor, and the capacitor
voltage will have the proper polarity to provide
commutating action that quickly forces the -first
auxiliary valve to turn off. I'he current flowing in
this path and through the other auxiliary valve xises
to a peak magnitude and then decays to zero (in the
process of which the commutation capacitor will be
discharged and then recharged with reverse polarity),
and both TX and the second auxiliary valve will
automatically turn off by self commutation when the
current oscillates to zero at the conclusion of this
ringing action. Now the capacitor voltage has the
correct polarity for successful commutation of the
main valves at the next zero crossing of the
phase-to-phase stator voltages during the engine
cranking mode o~ operation. Figure 2 illustrates that
the correct polarity is negative duriny even numbered
states and positive during odd numbered states.
After step 125 causes the controller to
produce a firing signal for turning on TX, and while
s~
20-TR-1420
- 35 -
the above-described rinying action i5 taking place,
the field winding 1~ of the machine 14 continues to be
excited by current which now "freewheels" through the
contactor Kln and the branch of the load current path
joining conductors 26 and 27 (E'igure 1). At the same
time, the subroutine shown in Figure 7 proceeds to
in~uire, at point 126, as to whether or not the
capacitor voltage has changed polarity (as indicated
by a state change of V~l). As soon as the answer is
affirmative, the control returns to the Main Routine,
and the Normal Cranking Routine 84 can begin.
As is shown in Figure 7, there is an
alternative way to get to the last inquiry point 126
of the subroutine 118. It includes two steps 127
and 128 that are similar to steps 123 and 125,
respectively, except that step 127 increments the
oldstate by 2 and id~ntifies the main valve that is
scheduled to be turned on in response to the second
state change to come. For example, if oldstate = 2,
oldstate ~ 2 - 4, and TX is T4. Step 127 is init~ated
if the answer to inquiry 122 indicates that the
capacitor voltaye is not positive, or if the answer to
inquiry 124 indicates that the capacitor voltage is
positive. It is immediately followed by the step 128
which causes the controller to generate a firing
signal for the main valve TX. When turned on, TX
connects the commutation capacitor across a first one
of the conducting auxiliary valves Tp and Tn,
whereupon the first auxiliary valve is forced to turn
off, the commutation capacitor is discharged then
recharyed wi~h reverse polarity, and both TX and the
other auxiliary valve will turn off as soon as the
battery current oscillates to zero. At this time the
polarity of the capacitor voltaye will not be correct
for successful commutation of the main valvPs when
~ ~7~
20-TR-1420
- 3~ -
cranking starts unless prior thereto a series o~
additional steps 130 through 136 are followed to
reverse this polarity from wrong to right.
Step 130, which immediately follows step 128,
performs the same inquiry as point 126, and as soon as
the capacitor voltage changes polarity the control
proceeeds to step 131. In step 131 the microcomputer
reads the oldstate, decrements it by 1 to ~ind the
number of the precediny state, and identifies the
proper pair (TX, TY) of the six main valves that are
normally conducting load current during such preceding
state. For example, if oldstate = 2, the preceding
state is 1 and the valve pair TX, TY are T6 and Tl,
respectively. Step 131 is immediately followed by the
step 132 in ~hich the controller is instructed to
yenerate concurrent firing signals for the valves TX
and TY. Wherl the valves TX and TY are turned Oll by
these firing signals, they enable battery current to
resume flowing in the armature windings of the
machine 14. After the firing signals are generated
(step 132), the control waits for a period of
approximately 30 milliseconds to allow excitation
current in the field winding 18 (step 133) to increase
in magnitude, and it then proceeds to a decision
point 134 which detPrmines whether or not the ol~state
is an even number. If the oldstats is even, the next
step 135 is to generate a firing siynal for turniny on
the auxiliary valve Tn, if the oldstate is not even,
an alternative step 136 is implemented to generate a
firing signal for turning on the auxiliary valve Tp,
and in either case the control proceeds from step 135
or 136 to the previously described inquiry point 126.
Once valve Tn (or Tp) is turned on by step 135
(or 136), it connects the precharged commutation
capacitor across the main valve TX which is thereby
~''7~~
20-TR-1420
- 37 -
forced commutated off. The current that transfers
from TX to the auxiliary valve first discharges and
then recharges the commutation capacitor with opposite
polarity, and both TY and the auxiliary valve will
automatically turn off by self commutation as this
current oscillates to zero. Now the capacitor voltaye
has the right polarity for successful commutation of
the main valves when cranking starts. As was
previously explained, the control transfers to the
~ormal Cranking Routine 84 as soon as the inquiry
point 126 detects a change in the polarity of the
voltage on the commutation capacitor.
The Normal Cranking Routine ~4 will cause
the controller 44 to produce the proper pa~tern of
properly synchronized firing signals as required for
the inverter to operate in a third harmonic
commutation mode, whereby mechanical torque is
developed in the rotor of the machine 1~ to start
turning the crankshaft of the engine 16 and to
accelerate it ~rom rest to a predetermined speed
(e.g., 240 rpm) well above the minimum "firing speed"
of the engine. The presently preferred embodiment of
this routine is shown in Figure ~. In its first
step 140, the oldstate that was saved in the memory of
the microcomputer 60 is used to find, in an
appropriately encoded look-up table, the identity of
the pair of main valves (TX, TY) that normally would
be conducting load current during such state. For
example, if oldstate = 2, TX is Tl, and TY is T2.
Step 140 is followed immediately by a step 141 in
which the controller is instructed to generate the
proper pair of firing signals (e~g., 71 and 72) to
turn on both of the valves TX and TY. In practice,
each firing si~nal can actually comprise a
35-microsecond burst of from five to ten
~7S~
20-1'R-142()
- 3~ -
short-duration (1~5 microseconds) discrete d-c signals
having sufficient magnitude to turn on the associated
valve.
Once the main valves TX an~ TY are turned on
by step 141, they complete a path for load current to
flow from the batter~ 17 through two phases of the
armature (stator) windings of the machine 14, and
through the section of the path that interconnects
conductors 25 and 26 (Figure 1)~ The latter section
comprises the field winding 18 (which was inserted
therein when the two poles K3a and K3b of contactor K3
were closed by step 85 of the Set-Up ~outine 83) arld
the parallel branch that includes conductor 27,
resistor 28, resistor 29 shunted b~ contactor K2, and
inductor 30. Since K2 opened at step 111 of the Set-Up
Routine and has not yet reclosed, both resistors 2~
and 29 are now effectively in series in this parallel
branch. The magnitude of armature current is
initially very high, limited only by the internal
resistance of the battery, the negligible resistance
of the armature windings, and the total resistance of
the two resistors 28 and 29. At the time TX and TY
start conducting load current, the field winding 18 is
being excited by the residual of the current that had
previously built up therein during the interval of
time between the execution of step 114 (Figure 5) and
the execution of step 125, 135 or 136 (Figure 7) in
the Set-Up Routine. The magnetic fields generated by
current in the armature windings now intereact with
3~ the excitation current in the field winding to produce
in the rotor 15 a torque (proportional to the product
of the magnitudes of these currents) that tends to
turn the crankshaft of the engine 16 in the desired
direction.
~7~
20-TR-1420
- 39 -
In one practical embodiment, currents in the
armature and field windings were high enough, with the
battery not fully charged, to produce a "breakaway"
torque of at least 3,600 foot-pounds which is
S sufficient to turn the crankshaft of a 4,000
horsepower diesel engine. As the crankshaft and rotor
start xotating, current (and torque) tends to decrease
in magnitude due to the rising amplitude of the back
emf that is induced in the armature windings and that
opposes the battery voltage. The instantaneous
magnitude of the back emf in each of the three phases
of the synchronous machine will alternate sirlusoidally
between relatively positive and negative peaks as the
rotor accelerates Erom rest and its angular position
advances. In due course the rotor will pass through a
location where the increasing voltage magnitude of the
oncoming or relieving phase (e.g., B) just equals the
decreasing voltage magnitude of the offgoing or
relieved phase (e.g., A), whereupon one of the three
zero-crossing detecting signals (e.g., XC) will change
up or down to mark the transition to the next state.
As is shown in Figure 8, after generating
the firing signals for valves TX and TY (step 141),
the Normal Cranking Routine proceeds to a step 142 in
which a K2 timer is started. This timer will run for
a predetermined length of time (e~g.~ approximately
three seconds) after being started. Step 142 is
followed by a step 1~3 in which the microcomputer:
(1) reads data representing the hi~h or low conditions
of the signals XA, XB, and ~C, (2~ uses a logical
combination of this data to find, in an appropriate
encoded look-up table, whichever one of the six
different states 1 through 6 is extant, and (3) places
the number of the extant or actual state in its memory
where this number is saved as "new state." From
20-TR-1~20
- 40 -
step 1~3 the control proceeds to an in~uiry point 14
which determines whether or not the new state is the
same as oldstate. So long as the answer is
affirmative, the control next inquires, at a
point 1~5, as to wllether or not a cranking "finish"
flay is on, and if not it then proceeds to an inquiry
point 146 where the open or closed status of
contactor K2 is tested. If K2 is closed, the control
returns to step 143, if not, the status of the K2
timer is tested at point 147. Assuming tha~ the K2
timer is still runninc~, the control immediately
returns to step 143. Otherwise, the control proceeds
~rom the in~uiry point 147 to a step 148 in which the
contactor K2 is commanded to close, whereupon the
contxol returns to stap 143. It will now be apparent
that so long as the inquiry step 144 determines that
the new state is the same as oldstate, the control
steps repetitively around a loop comprising the
step 143 and inquiry points 144, 145, 146, and 147
while the K2 timer is running, whereas, it steps
repetitively around a subloop comprising step 143 and
inquiry points 144, 145 and 146 after the lenyth o~
time programmed in the K2 timer is overO In response
to the expiration of this predetermined length o~
time, step 148 is implemented to close the
contactor K2 which then short circuits the resistor 29
(Figure 1), thereby reducing the ohmic value o~ the
resistance in parallel with the field winding 180 As
a result, more current can flow in the branch of the
load current path between conductors 26 and 27, the
field excitation is wea~ened, and higher cranking
speeds can be achieved.
As soon as the inquiry point 14~ determines
that the new state is not the same as the oldstate
(i.e., in response to a state change of one of the
20-T~-1420
- 41 -
phase-to-phase stator voltage zero-crossing detecting
signals XA, X~, and XC), the control transfers from
the above-described loop to an alternative loop
comprising the step 143, the inquiry p~int 1~4, and,
in the following order, a "Commutation Subroutine" 1~1,
a "Next State Subroutine" 152, an inquiry point 153,
and a "System Synch ~ubroutine" 154. The three
subroutines 151, 152, and 154 are therefore executed
each time the inquiry point 144 detects a ~tate change.
The Commutation Subroutine 151 of the ~ormal
Cranking Routine 84 is shown in Figure 9. It has two
functions: (1) to initiate commutation of the
offgoing main valve by ordering the production o~ a
firing signal for the appropriate one o~ the two
auxiliary valves Tp and Tn, and (2) to delay the
execution of the Next State Subroutine 152 until a
programmed interval of time has elapsed following each
resulting zero crossing of the voltage on the
commutation capacitor ~5. The Commutation Subroutine
is entered at a point la~eled "Commutation" and then
proceeds to a decision point 156 which determines
whether or not oldstate was an odd number. If
oldstate was odd (i.e., 1, 3 or 5), the control
proceeds to a step 157 in which the controller is
instructed to produce a Eiring signal 69 for turning
on the auxiliary valve Tn. If oldstate was even (2,
or 6), the control proceeds from point 156 to a
step 158 in which the controller is instructed to
produce a firing signal 68 for turning on the
auxiliary valve Tp. Consequently one or the other of
the auxiliary valves is turned on to connect the
commutation capacitor across the offyoing main valve.
The capacitor voltage will now have the correct
polarity to force load current to transfer to the
conducting auxiliary valve, whereupon the offgoing
6q~
20-TX-1420
- 42 -
main valve stops con~ucting. The load current in the
commutation circuit first discharges the capacitor and
then recharges it with reverse polarity, as i8
illustrated in the previously described Figure 3.
More specifically, if oldstate was odd, Tn is fired,
and at time t3 the capacitor voltage will change from
positive to negative as indicated by a high-to-low
state change of the output signal VEl of the first
bistable capacitor voltage sensor in the voltage
processor 48. On the other hand, if oldstate was
even, Tp is fired, and at time t~ the capacitor
voltage will change from negative to positive as
indicated by a low-to-high state change o-f VE1.
As is shown in Figure 9, step 157 (or 158)
of the Commutation Subroutine 151 is immediately
followed by an inquiry, at point 160 (ox 161), as to
the high or low state of the capacitor voltage
polarity indicating signal VEl. As soon as VEl
changes state, the control proceeds from point 160
(or 161) to another inquiry point 162 where the status
of the K2 timer is measured. For a predetermined
initial period of time, which starts when step 142 of
the Normal Cranking Routine 84 starts the K2 timer,
the inquiry 162 yields a negative answer, and
thereafter the answer will be affirma-tive. The
initial period is preferably approximately 1.5
seconds, or approximately half of the length of time
that the K2 timer is programmed to run.
In response to a negative answer at the
inquiry point 162, the next step 163 in the program is
to load a predetermined maximum capacitor recharging
time into a fourth timer. Step 163 is followed
immediately by a step 164 that starts tlle fourth timer
and then by an inquiry, at a point 165, as to the
state o~ the second capaci-tor voltaye sensor in the
S~
20-TR-1420
- 43 -
voltage processor ~, as indicated by the high or low
state of the signal VE~. So long as the capacitor
voltage is lower than the predetermined maximum level
(e.g., ~00 volts), the answer to inquiry 165 is
negative, and the control proceeds to an inquiry
point 166 where the status of the fourth timer is
tested. So long as the timer #4 is still running, the
answer to inquiry 166 is negative, and the control
r~turns to the preceding inquiry point 165. But
whenever an affirmative answer is obtained at either
inquiry point 165 (revealing that the magnitude o~
capacitor voltage has attained the aforesaid maximum)
or inquiry point 166 (revealing that the time delay
interval of the timer #4 is over), whichever is ~irst
to occur, the Commutation Subroutine 151 is exited,
and the control is transferred to the Next State
Subroutine 152 of the Normal Cranking Routine.
Preferably the maximum recharge time that is loaded
into timer #4 in step 163 is selected to have a
relatively long, fixed duration (e.g., approximately
1.2 milliseconds) so that an affirmative answer will
ordinarily be obtained from inquiry 165 earlier than
from inquiry 166 throughout the aforesaid initial
pariod (i.e., before the inquiry point 162 yields an
affirmative answer). In other words, whenever the
Commutation Subroutine is executed during the initial
period of time, it will be completed as soon as the
voltage on the commutation capacitor reverses polarity
and rises in magnitude to the aforesaid maximum level
(but no later than the expiration of the maximum
recharging interval that was loaded in timer #4 at
step 163~. ~igh capacitor voltage is required for
successful commutation during this period when load
currant is relatively high. At the same time, the
relatively long maximum capacitor recharging time is
20-TR-1~20
- 44 -
permissible because the rotor speed (and hence the
frequency of state changes) is now relatively low.
In response to an affirmative answer at the
inquiry point 162 (which will be true anytime the
Commutation Subroutine 151 is executed after
expiration of the aforesaid initial period of time),
the program shown in Figure 9 proceeds from point 162
to a decision point 1~7 which is a cluplicate of 156.
If oldstate was odd, the next step 168 is to load a
"negative" capacitor recharging time into timer ~4.
Alternatively, if oldstate was even, the next step 169
is to load a "positive" capacitor rP~harging time into
the same timer. In either case, the control then
proceeds, as before, to start timer ~4 at step 164 and
then repetitively to check for maximum capacitor
voltaye at point 165 and to test the status of the
timer at point 166. In accordance with the present
invention, the delay intervals that are loadsd into
timer #4 at steps 168 and 169 are shorter than ~he
maximum capacitor recharging intexval that is loaded
at step 163. While both intervals could be equal to
each other if desired, in the illustrated embodiment
the negative recharging interval is shorter than the
positive recharging interval. By way of example, the
negative recharging interval is approximately 300
microseconds, and the positive recharging interval is
approximately 500 microseconds. These intervals,
which in Figure 3 are respectively represente~ by the
delta t- and t+ symbols, are sufficiently short so
that, after the aforesaid initial period expires
(i.e., when the inquiry point 162 yields an
affirmative answer), an affirmative answer will
ordinarily be obtained from inquiry 166 earlier than
from inquiry 165. In other words, whenever the
Commutation Subroutine is executed after the initial
~.~7~ ~
20-T~-1420
- 45 -
period of ki~e, it is completed as soon as the delay
interval that was loaded in timer #4 at step 168 or
169 i5 over, and this does not provide enough time for
the commutation capacitor to recharge to the aforesaid
maximum level of voltage. Consequently, as was
explained hereinbefore in conne,ction with the
description of ~igure 3, the actual magnitude of
capacitor voltage at the conclusion of ~he Commutation
Subroutine i~ a function of the magnitude of load
current. It decreases with current, and as a result
the length of the co~nmutation interval is desirahly
reduced as the rotor speed (and frequency) increases
after the aforesaid initial period of time.
The Next State Subroutine 152, which is
executed immediately after the Commutation
Subroutine 151, is shown in Figure 10. Its purposes
are (1) to calculate and save the "next state" and (2)
to complete the third harmonic commutation process by
ordering the production of a firing signal for the
oncomin~ or relieving main valve. rrhis subroutine is
entered at a point labeled "Next State" and then
proceeds to a step 171 in which the microcomputer (1)
subtracts the previously saved oldstate (step 120 in
Figure 7) from the previously saved new state
(step 143 in Figure 8) to find the difference
therebetween, the difference being ~1 for the assumed
phase rotation A B-C of stator voltages but -1 if ~he
phase rotation were C-B-C, (2) increments the oldstate
by one if the difference is +l (or decrements it if
the difference were -1) to give the number of the
"next state", i.e., the state which comes after
oldstate and which therefore should coincide with the
new state, and (3) places the number of the "next
state" in the selected register of its memory where
this number replaces the ~reviously saved oldstate and
~,,"~'~d JJ ~
20-TR-1420
~ 46 -
is saved as a calculated new oldstate. Step 171 is
followed by a step 172 in which the new oldstate is
used to find, in an appropriately encoded look-up
table, the identity of the pair of main valves
(TX, TY) that normally should be conducting load
current during such state. For example, if new
oldstate = 3, TX is T2, and TY is T3. r~y is the
oncoming or relieving valve of the pair. Step 172 is
followed immediately by a step 173 in which the
controller is instructed to generate the proper pair
of firing signals (e.g., 72 and 73) to turn on the
identified valves TX and TY. It will be apparen-~ that
the firing signal for valve TX is redundant, as rrx in
this subroutine is the same valve as TY which was
turned on earlier in the program. Once the oncoming
valve TY is turned on by a firing signal produced at
step 173, load current can transfer to it from the
auxiliary valve that was turned on by step 157 or 158
of the Commutation Subroutine 151 (Figure 9), thereby
completing the commutation process. While load
current is decaying to zero in the auxiliary valve,
the commutation capacitor continues recharging to a
peak magnitude somewhat beyond its level of voltage at
the conclusion of the Commutation Subroutine.
The firing signal that is produced by
step 173 of the Next State Subroutine 152 for the
oncoming main valve TY is represented in Figure 3 by
the pointer 7X. It is produced at time t4 if the
calculated new oldstate is an even number and at
time t9 if odd. Both of these times are delayed with
respect to the preceding state change of the capacitor
voltage polarity indicating signal VEl (as detected by
the inquiry point 160 or 161 of the Commutation
Subroutine 151). Each time the Next State Subroutine
is executed during the aforesaid initial period (which
20-TR-1420
- ~7 -
is determined by step 162 of the Commutation
Subroutine), this delay will depend on how long the
commutation capacitor takes to recharge to the vol-tage
level at which an affirmative answer is obtained at
the inquiry point 165 in the Commutation Subroutine.
But each time the Next State Subroutine is executed
after the initial period, the firing signal 7X is
delayed until a programmed interval of time has
elapsed. If the new oldstate is an even number, the
proyrammed delay interval is determined by the
negative recharging time loaded into timer ~4 at
step 168 of the Commutation Subroutine, and otherwise
it is determined by the positive recharging time
loaded into the same timer at step 169.
Having generated the firing signal (7X) for
the main valve TY at step 173, the Next State
Subroutine 152 returns to the Normal Cranking ~outine
(Figure 8) where the calculated new oldstate (step 171
in Figure 10) is checked, at the inquiry point 153, to
be sure that it is in fact the same as the previously
saved new state (step 143 in Figure 8). If not, after
waiting for a fixed period of approximately 50
milliseconds (step 174), the control is retransferred
to the Commutation Subroutine 151, and the two
~5 subroutines 151 and 152 are executed again. This
process is repeated, if necessary, until the
calculated new oldstate coincides with the new state,
and then the control is trans~erred to the System
Synch Subroutine 153.
The presently preferred embodiment of the
System Synch Subroutine is shown in Figure 11. It i5
entered at a point labeled "System Synch" and proceeds
to an inquiry point 175 where the open or closed
status of the contactor K2 is checked. Initially,
35 while the K2 timer is running lsee steps 142 and 147
20-TR-1420
- 4~ -
in ~igure ~) and therefore prior to impleMentation of
the step 148 that commands the contac~or K2 to close,
this inquiry will reveal that K2 is open.
Consequently, after waiting for a certain delay
interval (step 176) the control is returned directly
to the step 143 of the Normal Cranki}lg Routine 84.
The delay introduced by step 176 will allow time,
after implementing step 173 of the Next State
Subroutine (~igure 10) and before returning to the
step 143, for the above-describead commutati~n from tlle
auxiliary valve (Tp or Tn) to the oncominy main
valve TY to be comple~ed and fox the resulting
electrical transients ("noise") to subside in the
voltage processor 48. The delayed return to step 143
is desirable when load current is relatively high (as
is true initially), because the aforesaid transients
might then be severe enough to cause false data to be
supplied on lines 51, 52 and 53 to the controller 44.
Once the control returns from step 176 (Figure 11) to
step 143 (Figure 8), the loop comprising step 143 and
in~uiry points 144, 145, 146, and 147 is repeatedly
executed until the next state change is indicated by a
negative answer to inquiry 144 (i.e., until the new
state, as determined by step 143, is no longer the
same as the new oldstate that was calculated by
step 171 of the Next State Subroutine 152), whereupon
the control again transfers to the Commutation
Subroutine 151.
As the angular position of the rotor of the
machine 14 advances wi-th increasing speed during the
cranking mode of operation, the above-described
execution of the steps 143-147 in the Normal Cranking
Routine 84 and of the subroutines 151-154 are
automatically repeated until the K2 timer stops
running. The Commutation Subroutine 151 is initiated
56'~'~
20~TR-1~20
- 49 -
each time the inquiry 144 indicates a state change,
and such changes occur with increasing fre~uency as
the rotor accelerates. In practice it may take
approximately two seconds for the rotor to complete
its first revolution and another second for the second
revolution. By the end of two revolutaions the rotor
may have attained a speed on the order of 100 rpm, and
it is around this time that the K2 timer stops runnin~
and the contactor K2 is closed to further weaken the
field and permit higher speed cranking.
Each time the System Synch Subroutine 154 is
executed after contactor K2 is closed, the control
will proceed from inquiry point 175 to an inquiry
point 177 in which the status of the first timer is
tested. As shown in Figure 11, if timer #1 has
stopped running the next step 178 of this subroutine
will cause the controller 44 to command all contactors
to open, and the Normal Cranking Routine is then
aborted at a stop point 179. Assuming, however, that
timer #1 is still running, the control proceeds from
point 177 to a speed check step 181 where the rotor
speed is measured. Any suitable m0ans can be used for
this purpose. One simple yet effective means for
measuring rotor speed is to count the number of times
the Commutation Subroutine 151 is executed over a
known period of time. It can be shown that this count
is proportional to speed. There are six state changes
and hence six commutation intervals per cycle of the
fundamental component of alternating voltages on the
stator windings of the machine 14, and one complete
revolution of the rotor corresponds to five such
cycles in a 10-pole machine. Thus the predetermined
threshold speed of 240 rpm corresponds to a
fundamental frequency of 20 ~ertz which is indicated
3~
20-TR-1420
- 50 -
if 12 commutations are counted in a period of Ool
second.
From step 181 the System Synch Subroutine
proceeds to a point 182 that inquires as to whether or
not the rotor speed has attained a predetermined rate
(i.e., the aforesaid threshold speed of 240 rpm). If
not, the control is then returned directly to the
step 143 of the Normal Cranking Routine 84
(Figure 8). Now the subloop comprising step 143 and
inquiry points 144, 145, and 146 will be repeatedly
e~ecuted until the next state change taXes place,
whereupon the control once again transfers to the
Commutation Subroutine 151.
The above-described execution of the
steps 143-146 and of the subroutines 151-154 are
automatically repeated until the rotor is ~otating
faster than the aforesaid predetermined rate~ Once
this threshold speed is exceeded, the inquiring
point 182 of the System Synch Subroutine (Figure 11)
yields an affirmative answer. In this event the
control proceeds from point 182 to a step 183 that
causes the controller to issue an appropriate signal
that the engine is running. Step 1~33 is followed
immediately by a step 184 that sets the cranking
finish flag in an "on" state, and later by a step 185
that will cause the controller to command the opening
of contactors Klp, Kln, K2 and K3. From step 185 the
control returns to step 143 of the Normal ~ranking
Routine (Figure 8). Now ths step 143 and the inquiry
points 144 and 145 are passed through again, and from
the inquiry point 145 the control can proceed to a
finish point 186 which marks the conclusion of the
cranking mode of operation.
5~
20-TR-1420
- 51 -
While a preferred embodiment of the
invention has been shown and described by way of
example, many modifications will undoubtedly occur to
persons skilled in the art. The concluding claims are
S therefore intended to cover all such modifications as
fall within the true spirit and scope of the invention.
