Electric motors turn electrical energy into mechanical energy to produce
work. Electric motors work by applying a voltage across one or more windings, thereby
energising the winding(s) to produce a resultant magnetic field. Mechanical forces
of attraction or repulsion caused by the magnetic field cause a rotor in an electric
motor to move. The efficiency of the electric motor depends in part on the timing
and magnitude of each application of voltage to the motor. Timing of the voltage
being applied is particularly important in the case of switched reluctance motors.
Historically, the switched reluctance motor was thought to be incapable
of competing effectively with other types of motors. More recently however, a better
understanding of motor design and the application of electronically controlled switching
has resulted in a robust switched reluctance drive capable of high levels of performance
over a wide range of sizes, powers and speeds. Note that the term "motor" is used
here, but it will be appreciated by those skilled in the art that the term covers
the same machine in a generating mode unless a particular distinction is made.
The switched reluctance motor is generally constructed without windings
or permanent magnets on the rotating part (the rotor) and generally includes electronically-switched
windings carrying unidirectional currents on the stationary part (the stator). Commonly,
pairs of diametrically opposed stator poles may be connected in series or parallel
to form one phase of a potentially multi-phase switched reluctance motor. Motoring
torque is produced by applying voltage across each of the phase windings in a predetermined
sequence that is synchronized with the angular position of the rotor so that a magnetic
force of attraction results between poles of the rotor and stator as they approach
each other. Similarly, generating torque is produced by positioning the voltage
pulse in the part of the cycle where the poles are moving away from each other.
The general theory of design and operation of switched reluctance machines is well
known and discussed, for example, in The Characteristics, Design and Applications
of Switched Reluctance Motors and Drives, by Stephenson and Blake and presented
at the PCIM '93 Conference and Exhibition at Nurnberg, Germany, June 21-24, 1993.
There have been various strategies proposed in the past for controlling
switched reluctance motors as part of an overall variable speed drive system. In
general, these strategies may be divided into two broad groups: systems that employ
current magnitude control over a fixed angle of rotor rotation, and systems that
employ voltage control where the angular position at which the applied voltage is
controlled. Often, these strategies are combined with each being used during particular
periods of the motor's operation. The present invention is directed to angular position
voltage control systems. A general discussion of such strategies can be found in
the paper The Control of SR Drives: Review and Current Status by Sugden,
Webster and Stephenson, EPE'89 Conference on Power Electronics And Applications,
At high speeds, the torque of the motor is commonly controlled by
controlling the position and duration of the voltage pulse applied to the winding
during the phase period. Because a single pulse of voltage is typically applied
during each phase period, this form of control is often referred to as "single-pulse
control". In single-pulse control, the torque level is defined by the magnitude
and shape of the voltage pulse which, in turn, is generally determined by: the angular
speed of the rotor; the point during the rotor's rotation when voltage is applied
to the phase winding by closing one or more switches (referred to as the "TURN-ON
angle"); the point during the rotor's rotation when the application of voltage to
the winding is reversed by opening one or more switches (referred to as the "TURN-OFF
angle"); and, the magnitude of the voltage applied to the phase winding. The TURN-ON
and TURN-OFF angles define a "conduction angle." The conduction angle is the angular
distance between the TURN-ON and the TURN-OFF angles.
Some previous switched reluctance motors have used simple angular
position sensors to provide the rotor position information necessary to energize
the phase windings during their respective conduction angles. Normal practice is
to use a two-sensor arrangement for four-phase machines, and a three-sensor arrangement
for three-phase machines. One of the advantages to using this type of system is
the low sensor cost.
Figure 1 shows the principal components of a switched reluctance drive
system 10 for a switched reluctance machine. The input DC power supply 11 can be
either a battery or rectified and filtered AC mains. The DC voltage provided by
the power supply 11 is switched across the phase windings of the machine 13 by a
power converter 12 under control of the electronic control unit 16. The switching
must be correctly synchronized to the angle of rotation of the rotor for proper
operation of the drive 10. As such, a rotor position detector 15 is typically employed
to supply signals corresponding to the angular position of the rotor. The output
of the rotor position detector 15 may also be used to generate a speed feedback
The rotor position detector 15 may take many forms. In some systems,
the rotor position detector 15 can comprise a rotor position transducer that provides
output signals that change state each time the rotor rotates to a position where
a different switching arrangement of the devices in the power converter 12 is required,
or when the rotor rotates to a particular position where the rotor and stator poles
are aligned with each other.
Figure 2 shows the elements of a typical four-phase switched reluctance
machine 13. The machine 13 has eight salient poles 26a-h on the stator 28 and six
poles 20a-f on the rotor 22. Each stator pole 26a-h carries a simple exciting coil
24a-h. Opposite coils 24a and 24e, 24b and 24f, 24c and 24g, 24d and 24h are connected
to form north/south pole pairs for the four phase windings. Only one phase circuit
26 is shown for the opposite coils 24a and 24e. The opposite coils 24a and 24e are
excited from a DC supply 29 through two switches or transistors (S1 and S2), and
two diodes (D1 and D2) allow energy to return to the supply 29. Other switching
circuits are well known in the art.
If it is desired to operate the machine 13 as a motor, torque is developed
in the machine 13 by the tendency for the magnetic circuit to adopt a configuration
of minimum reluctance, i.e., for an opposing pair of rotor poles 20a and 20d, 20b
and 20e, and 20c and 20f to be pulled into alignment with an excited pair of stator
poles 26a-h, maximizing the inductance of the exciting coils 24a-h. By switching
the phases in the appropriate sequence, the rotor 22 will continuously rotate in
the chosen direction so that torque is developed continuously in the appropriate
direction. Moreover, the larger the current supplied to the coils 24a-h, the greater
the torque. Conversely, if it is desired to operate the machine as a generator,
the coils are excited as the rotor poles move away from the stator poles: power
is then transferred from the shaft of the machine to the electrical supply.
The positional information provided by a typical rotor position transducer
(RPT) is normally of the form shown in Figure 3, which shows the outputs of the
three-sensor RPT typically used for a three-phase switched reluctance drive. RPT
A is the signal used to drive phase A of the machine. The signals are cyclic over
one inductance cycle, i.e., the angular period defined as the angular movement
of the rotor between one pair of stator poles having one pair of rotor poles aligned
with them and having an adjacent pair of rotor poles aligned with them. The three
signals are displaced from each other by one third of a cycle. The RPT is usually
arranged relative to the rotor shaft so that the edges of the RPT signals correspond
to a particular part of the inductance cycle, e.g., &thetas;1,
on RPT A would normally correspond to the minimum inductance point on the inductance
cycle and &thetas;2 would then correspond to the maximum inductance position.
Because of the mechanical symmetry of the geometry of the machine, the signals RPT
B and RPT C correspond to similar points on the inductance cycle of phases B and
All this is well-known in the art, as is the adjustment of the mechanical
parts of the RPT assembly to ensure that the RPT signals have an equal mark: space
In one of the simplest methods of control of a switched reluctance
rotor, the phase excitation can be linked directly to the appropriate RPT signal,
i.e., when the sensor signal RPT A rises to logic 1 (&thetas;1
in Fig. 3), the phase is energised by closing switches S1 and S2 in Figure 2. When
the rotor moves to &thetas;2, the phase is de-energised by opening the
switches S1 and S2, allowing the flux associated with phase A and the current in
the winding to decay to zero by virtue of the current flowing through the diodes
and back to the supply.
Although at low machine speeds the sensor edge will typically correspond
to the TURN-ON angle, high machine speeds require that the TURN-ON angle be advanced
by some varying amount. Since the sensors of Figure 3 only provide a sensor edge
twice in each inductance cycle, they do not provide sufficient resolution of the
rotor position necessary to energize the phase windings during their respective
conduction angles. Consequently, such switched reluctance drive systems ordinarily
interpolate angles between sensor edges to provide sufficient resolution of the
As seen in Figure 4, one prior art method of interpolation uses a
frequency multiplier 31 for each motor phase to generate an integer multiple of
pulses between the sensor signal for clocking a corresponding counter 32 for each
machine phase. This generates counter values which increase at a rate proportional
to the speed of rotation of the rotor. CE is the clock enable input on the counter
32 and prevents the clock changing the counter state when low. TC is the terminal
count output of the counter 32 and is high when the counter is at its maximum value.
Both the counter 32 and the frequency multiplier 31 for each machine phase are reset
at the occurrence of their respective timing sensor edge corresponding to that phase.
The outputs of each of these counters 32 are provided to two digital comparators
33 and 36 per machine phase which are used to provide the firing signals needed
for energising the phase windings at their respective TURN-ON and TURN-OFF angles.
A third digital comparator for providing a signal at the free-wheel angle may also
be used as is known by those skilled in the art. The other input to each of these
comparators 33 and 36 is typically provided by a pre-interpolated control map in
a memory device such as an EPROM.
Since each of the frequency multipliers for each phase are corrected
for the current motor speed at their respective sensor edge, the interpolated angles
are always being extrapolated from the prior measurement of the period between these
edges. The output of the frequency multiplier is only updated once per cycle, allowing
relatively large errors to accumulate during rapid acceleration or deceleration.
Furthermore, this arrangement requires one frequency multiplier for each phase of
As seen in Figure 5, one prior art method used to improve these prediction
problems in multi-phase systems having more than one position sensor (e.g.,
a four-phase machine with two position sensors or a three-phase machine with three
position sensors) uses a frequency multiplier 42 for all machine phases and updates
that frequency multiplier 42 at each of the sensor edges. Here, instead of interpolating
based on the entire prior machine phase cycle, the interpolation is based on portions
of different phases. Thus, the frequency multiplier 42 will be updated responsive
to machine acceleration and deceleration more often.
When visualizing the arrangement of Figure 5, it is helpful to consider
the angles as digital "ramps" with the rising "ramp" representing the rising count.
If a four-phase machine with two sensors is taken as an example, then the RPT signals
RPT A and RPT B shown in Fig. 6a will be available. If each of these has a 50% mark:space
ratio, they may be X-ORed as shown in Fig. 5 to give the combined signal SENSOR
PULSE shown in Fig. 6a. This gives four pulse edges per inductance cycle. The frequency
multiplier 42 then only needs to multiply by 64 to produce 256 pulses per phase
cycle. As shown in Figure 6a, each of the four digital ramps A,B,C, and D represent
the count for each respective phase. Also shown in this timing diagram are the sensor
inputs, RPT A and RPT B, as well as the output of exclusive-or gate 41. Each ramp
A, B, C, and D is reset by one of the sensor edges; it is common to set the sensor
edges to coincide with the aligned position of each of the phases. While this system
provides some improvement for acceleration or deceleration errors, it requires precise
sensor positioning which may not be practical to implement, given practical manufacturing
tolerances, especially tolerances associated with low cost machines. Moreover, it
can be shown mathematically that any variation in the mark:space ratio will produce
faster ramps for all phases. These variations cannot be designed out, since they
vary with each particular machine. Thus, sub-optimal performance will manifest itself
in the form of lower torques and reduced machine efficiency.
Only the frequency multiplier 42 of Figure 5 and the 8-Bit latches
can be used for all the phases. The remaining circuitry is specific to each phase.
It will be realized from a consideration of Figure 6B that the prior
art corrected for variations once per phase inductance cycle. In other words, the
counter would be reset at zero, and then count up to maximum count only once during
the phase inductance cycle. These prior art systems did not have the capability
of correcting on the edge of each RPT signal.
EP-A-0534761 discloses a switched reluctance machine control system.
As part of the disclosure there is a reference to means for producing control system
output signals in which predefined values of switch turn-on and turn-off angle are
compared with a count conducted over a single period in a phase inductance cycle.
Thus, a need exists for a method and circuitry for interpolating conduction
angles with reduced error particularly in a system which adjusts for variations
in the mark:space ratio of sensor inputs on each RPT edge. While the prior art will
correct the frequency multiplier rate once during the phase inductance cycle, this
invention extends the idea of this correction by setting the "ramps" to the expected
value at each of the sensor edges, thus preventing the ramps from drifting for more
than one sensor period between adjacent sensor edges. In the case of a three-phase
machine with three sensors, this is a factor of six improvement over prior art systems.
The present invention is defined in the accompanying independent claims
1, 2, 11 and 12. Preferred features of the invention are recited in the dependent
The invention provides a method and apparatus of interpolating conduction
angle including counting between features, such as one edge of a digital pulse,
to provide data from which position information is derivable. The tendency of the
count to drift in a phase inductance cycle is limited to the period between features.
At each feature detection there is an effective correction of the count.
In one form the present invention is used in controlling the energisation
of the windings in a switched reluctance machine having more than one phase winding
using readings from angular position sensors and interpolating these sensor values
electronically to obtain adequate resolution on each sensor edge.
One aspect of the present invention for controlling the energisation
of a phase winding involves generating normalized TURN-ON and TURN-OFF signals using
a normalized rotor position count, and then generating a voltage control pulse defined
by the normalized TURN-ON and TURN-OFF signals for energising the appropriate phase
In accordance with another aspect of the present invention a TURN-ON/TURN-OFF
circuit generates a normalized TURN-ON signal and a normalized TURN-OFF signal when
the normalized rotor position count reaches a predetermined TURN-ON count and a
predetermined TURN-OFF count. Then, a phase pulse circuit generates a voltage control
pulse defined by the normalized TURN-ON and TURN-OFF signals for energising the
These and other aspects and advantages of the present invention will
become apparent upon reading the following detailed description of embodiments of
the invention which are given by way of example and upon reference to the drawings
- Figure 1 is a diagram of the principal components of a switched reluctance drive
- Figure 2 is a diagram of the internal components of a machine and the connection
of one phase to its power switches;
- Figure 3 shows the signals provided by a typical rotor position transducer (RPT)
having three sensors and used for a three-phase switched reluctance drive;
- Figure 4 is a block diagram of the typical components of an angular interpolation
system used in each motor phase;
- Figure 5 is a block diagram of an improvement over the system shown in Figure
2 and is used in systems with more than one sensor and the interpolation is based
on portions of phases;
- Figures 6a-b are timing diagrams for a group of signals present in the block
diagram of Figure 5, and Figure 6c represents how the digital ramp or count for
a phase is adjusted in the four-phase system described in Fig. 7 according to the
principles of the present invention;
- Figure 7 is a block diagram of one embodiment of the present invention;
- Figure 8 is a block diagram of a second embodiment of the present invention
which is used to generate the TURN-ON and TURN-OFF signals used in the voltage pulse
generation circuit of Figure 9;
- Figure 9 is a block diagram of the voltage pulse generation circuit for a four-phase
- Figure 10 is a block diagram of a voltage pulse generation circuit for a three-phase
Illustrative embodiments of the invention are described below as they
might be implemented using the improved angle interpolation circuitry and methods
to create a simpler, more efficient, and more accurate, conduction angle controller
for the phases of a switched reluctance machine. In the interest of clarity, not
all features of an actual implementation are described in this specification. It
will of course be appreciated that the development of any such actual implementation
would be a routine undertaking for those of ordinary skill having the benefit of
Conceptually, a simple embodiment of the invention is shown in Figure
7. The output RPT signal pulses from a conventional RPT sensor are supplied to an
RPT edge detector 60 which produces edge pulses coincident with the rising and falling
edges of the RPT signal pulses. The RPT signal pulses are also passed through an
exclusive OR arrangement 68 and to a frequency multiplier (62), which has a multiplying
factor of n (64 in this embodiment). The multiplied signal from the multiplier 62
is the clock signal to a counter 61. The multiplied signal is actually submitted
to one input of a two-input AND gate 63 which is also supplied with the output of
a digital comparator 65 so that the multiplied signal from the multiplier is only
admitted to the clock input of the counter according to the state of the digital
comparator 65. This is described below.
The edge signal directly from the edge detector 60 is applied to the
parallel load input PL of the counter 61. This PL input to the counter presets the
counter at each edge signal.
The RPT signals are also applied to the input of a maximum count setting
device 66 which produces an output (N) which is the maximum count permitted for
the current RPT state. The output of the device 66 is supplied to parallel inputs
of the counter 61 and provides the other input to the digital comparator 65.
The counter 61 is enabled between adjacent sensor edges only until
a maximum count for that angular region is reached; the counter 61 is stopped by
preventing further clock pulses. If the frequency multiplier 62 is running too fast,
the counter 61 will be stopped when the maximum count N is reached; if it is running
too slow, the maximum count N will not be reached before the next sensor edge output
from the edge detector. When the next sensor edge is detected, the counter 61 will
be loaded with the correct count corresponding to that particular sensor edge, "PL"
being the parallel load command, and the counter 61 will then be enabled thereafter.
The reset input to the counter 61 is needed to start the whole phase cycle. The
above circuitry must be repeated for all phases apart from the multiplier 62 and
the edge detector 60.
Considering the previously described 4-phase machine (phases A,B,C,
and D), the system described above begins operation as the RPT A signal falls to
logic 0 and the phase A counter 61 is reset at the 'start of phase A' pulse. Figure
6c shows how the digital ramp or count for phase A is adjusted at each sensor edge.
The counter 61 will then be clocked at a rate determined by the frequency multiplier
62. If the machine speed is constant and all of the sensor edges are evenly spaced,
then the counter 61 should reach a count of sixty-three at the next sensor edge.
In that case no corrective action is needed. If the time interval between sensor
edges is too long, so that a count higher than 63 would be produced, then the counter
will be disabled at sixty-three until the next sensor edge arrives. This disabling
occurs when the TC output of the counter 61 is fed back through an inverter and
disables the counter 61 through its CE input. Conversely, if the period between
sensor edges is too short, the counter may only reach, say, sixty before the next
sensor edge arrives. At that sensor edge, the counter will be preset to sixty-three,
and the system will count to the next sensor edge, expecting a count of 127. The
phases are interleaved, so while phase A is counting to sixty-three, phases B,C
and D are counting to 127, 191 and 255 respectively.
By modifying the above concept to use normalized rotor positions,
a system can be produced which not only provides for performance enhancement, but
also reduces the logic circuits required for implementation. In the second embodiment
shown in Figure 8, only one phase counter 73 is used to count between adjacent sensor
edges for all the phases. Counting between sensor edges distinguishes this invention
from prior art since the prior art provided for corrections only once during a phase.
The counter 73 is consequently also reduced in size and will always start from zero
and stop at the same count irrespective of which sensor edges are being used. Therefore,
the counter 73 now indicates the normalized rotor position which is the rotor position
between an adjacent pair of machine phases since the sensor edges are set to coincide
with the aligned (i.e., the minimum reluctance) and unaligned positions of
each of the phases. This simplifies the logic circuitry considerably and eliminates
the necessity of presetting the counter 73. However, in order to provide full control
over a phase cycle, a further determination of which machine phase should be energised
must be made.
Specifically, this embodiment provides an improved method for controlling
the energisation of the phase windings with an improved angle controller in switched
reluctance machines having more than one phase winding and more than one position
sensor, for example a four-phase machine having two position sensors. The position
sensors are set so that their output signals will have sensor edges (the output
of the position sensor transitioning from one logic level to another) which coincide
with a predetermined datum of the inductance cycle of each of the phases. Preferably,
the predetermined datum is that which corresponds to a rotor pole being completely
aligned with the stator phase under consideration.
This is a position of the rotor relative to the stator which corresponds
to one of minimum reluctance in the flux path between the two. Another convenient
datum would be the position of the maximum reluctance at which the stator and rotor
poles in question are midway between each other.
The signals provided by these position sensors are combined to generate
a sensor pulse signal. This sensor pulse signal has a transition edge each time
a sensor edge occurs for any of the position sensors. For instance, Figure 6a shows
the position sensor signals RPT A and RPT B for a four-phase machine having two
position sensors. One way to generate the sensor pulse signal is by using an exclusive-or
(XOR) logic gate 72 with the sensor position signals RPTA and RPTB as its inputs,
as shown in Figure 8. The single output from the XOR gate 72 is the input to a frequency
The multiplied sensor pulse signal represents an angular clock signal.
This angular clock signal has an integral number of pulses between the adjacent
sensor edges of the sensor pulse signal. The angular clock signal is generated using
the frequency multiplier 71 with the sensor pulse signal which is the output of
the exclusive-or gate 72. The angular clock signal is used for generating a normalized
rotor position count which is representative of the rotor's position relative to
the stator poles. For instance, in the four-phase system of Figure 8 the output
of the phase counter 73 is representative of the normalized rotor position.
This normalized rotor position count is used to generate the normalized
TURN-ON and TURN-OFF signals which define the normalised conduction angles for the
machine phases. For instance, in the four-phase system shown in Figure 8 a TURN-ON/TURN-OFF
circuit consists of two six-bit digital comparators 74, 77 respectively. The normalized
rotor position count comprises one input to each of these six-bit digital comparators.
The other input to each of the digital comparators is provided by two six-bit latches
75, 76 respectively. The four-phase system shown uses eight bits to represent the
respective TURN-ON and TURN-OFF angles (not normalized), the least significant six
bits being representative of the normalized TURN-ON and TURN-OFF angles. The respective
TURN-ON and TURN-OFF angles (not normalized) are provided to the angle controller
typically by a control law EPROM or other memory circuit, the details of which are
not relevant to the description of the present invention and well known to those
skilled in the art.
The remaining two bits (two most significant bits) of the TURN-ON
and TURN-OFF angles (not normalized) are included to generate a firing pulse for
energising the appropriate phase of the switched reluctance machine. The TURN-ON
and TURN-OFF signals define the starting position and the conduction angle for all
of the firing pulses used to energize each particular phase. A phase pulse circuit
is shown in Figure 9 which produces the firing pulses necessary to energize each
machine phase at the appropriate time.
The four-phase system of Figure 9 uses the TURN-ON and TURN-OFF signals
of the circuit of Figure 8 to generate the required firing pulse. The two position
sensor signals RPT A and RPT B define a two-bit digital word indicating uniquely
the quadrant of the phase cycle in which the rotor is situated. The two most significant
bits of the TURN-ON and TURN-OFF reference angles and the two position sensor signals
RPT A, RPT B are input to two, two-bit digital comparators 81, 82 with the TURN-ON
and TURN-OFF signals (which are used as a carry in) as shown in Figure 9. The outputs
of the comparators 81, 82 are used to define the start and finish of the firing
pulse. The logic circuits for the other three phases are derived using identical
circuits, with one or both of the sensor signals being inverted as appropriate.
A three-phase drive system may, in principle, use the same technique
as described for the four-phase system. However, in a three-phase system, the number
of RPT states will typically be six. Because of these six states, more than two
bits are needed to represent them; the four-phase arrangement described above is
inadequate for the task. However, several options exist to overcome this limitation.
For instance, the six-bit counter could be set to provide less resolution. In the
usual case, a six-bit counter will provide angular detection resolution of one part
per sixty-four within a RPT state. However, a maximum count of forty-two could be
used within a RPT state, instead of sixty-four, so that there are 252 counts in
a ramp. 252 can be represented by an eight-bit number. Although this approach would
work in theory, it may prove impractical to implement, as the counter cannot be
easily shared between phases. Also, this embodiment requires additional circuits
to derive values for the other phases.
Another option is to add a ninth bit to the TURN-ON and TURN-OFF reference
angles to indicate the RPT state. In this case, three-bit digital comparators are
required to process the most significant bits. Though simple, this embodiment uses
a nine-bit word while most digital storage units are organized in eight-bit blocks.
A preferred implementation restricts the ranges of the TURN-ON reference
angle and the TURN-OFF reference angle to four out of six states. The three bits
representing the RPT state can be decoded slightly differently for the TURN-ON reference
angle than the TURN-OFF reference angle. For the TURN-ON reference angle, the range
is restricted to the second, third, fourth and fifth RPT states, irrespective of
the value in the six-bit counter. These four states can be coded using two bits,
giving a complete eight-bit ramp over the middle four RPT states. The arrangement
is slightly different for the TURN-OFF reference angle because this angle is always
further in the cycle than the TURN-ON reference angle. In the case of the TURN-OFF
reference angle the first two states are disabled, allowing a full ramp over the
last four states.
Figure 10 shows the preferred embodiment of the three-phase system,
the main difference from the four-phase system being the implementation of the firing
circuit, which requires additional signals and logic circuits. The two most significant
bits of the TURN-ON and TURN-OFF reference angles and the two 2-bit signals ON-COMP
and OFF-COMP from the decoder block 91 are input to two, two-bit digital comparators
92, 93 as shown in Figure 10. The decoder block 91 has among its inputs the three
sensor position signals RPT A, RPT B, and RPT C and the rotational direction signal
DIR. The decoder block 91 has as outputs the aforementioned two-bit signals ON-COMP
and OFF-COMP, as well as ON, ON-BAR, and OFF-BAR, shown in Figure 10. The decoder
block 91 may be implemented by a programmable logic array or other similar circuits
as are known by those skilled in the art.
Thus, the principles of the present invention, which have been disclosed
by way of the above examples and discussion, can be implemented using various configurations
and arrangements. Those skilled in the art will readily recognize that various other
modifications and changes may be made to the present invention without strictly
following the exemplary application illustrated and described herein and without
departing from the scope of the present invention, which is set forth in the following