Note: Descriptions are shown in the official language in which they were submitted.
1~79~
~IErrHOD AND APPAl~TUS FOR PEI~FOr'~h1ING WORK
Jr~ I~ 'rl~REE DI~I~NSIONAI. SPACE
Field_of the Invention
05 This invention relates to the field of perLorming a
- plurality of operations on a workpiece hy a tool in locations
that require the relative movement of the workpiece and/or
tool to various locations in a three-dimensional space and
more particularly to a multi-a~is control fiystem for tlle
robot-like execution of a compl~x pattern of motions of the
workpiece and/or tool in a three-dimensional space 50 that
work may be performed by the tool on the workpiece.
Background of the Invention
During the past several years, there has been increasing
application in metal working machine tools of robot-like
systems for performing work b~ a tool on a workpiece. These
systems are commonly referred to as computer numerical
control systems. These systems utilize computer technolo~y
to provide memory stora~e for pro~rams which enable the
machine to be operated automatically. Computer numerical
control systems allow very precise machining capabilities
and small to medium~volume part production with minimal
input from the operator. Systems of tlliS nature rely
heavily on expensive hardware to insure the accurate operation
of the system. In addition to the expense, equipment of
,
~,7~ 7~7
this nature requires special programming that is rather
complicated and generally requires some experience in
comp~lter programming. Another difficulty encountered in
this field is the ability to keep the machine operating due
05 to the availability of service and the costs of such serviceO
Brief Description of the Invention
This invention provides a micro-computer numeric
control system to provide a multi-axis motion control
system. The basic concepts are primarily directed to the
use of software to replace hardware conventionally used in
this type of s~stem. The invention includes a DC servo
electric motor that is driven by a relatively high DC
voltage. These DC servo electric motors are designed to run
at a relatively high motor torque to speed ratio which
permits a substantial reduction in costs associated with
motor tachometers to achieve the same control resolution
normally associated with the more expensive motor tachometers.
The invention also utilizes the natural damping characteristics
of a DC servo electric motor to hold the various mechanisms
in a stop lock mode by providing a circuit that provides a
constant flow of current through the DC servo electric
motor.
In the explanation of the invention set forth below,
the micro-computer numeric control system is described in
relation to a milling machine in which the work table on
which a workpiece is located is moved generally in a horizontal
plane along an x-axis and a y-axis while the quill, in which
the tool is positioned, is moved in a direction along a
z-axis which is perpendicular to the horizontal plane of the
;"'
1`'~J 2
67~5~7
x and y axes. ~he work table is mounted on a knee of the
milling machine which knee may be vertically adjusted by
conventional means so as to accommodate workpieces of
various sizes. A DC servo electric motor is provided for
05 each of the x, y and z axes. Each DC servo electric motor
responds to a signal generated by its own control means so
as to drive suitable mechanisms which cooperate to move the
workpiece and/or the tool so as to perform desired operations.
In a preferred embodiment of the invention, the motor torque
to speed ratio is between about 2.4 to 3.2 to 1 which
permits the DC servo electric motor to be controlled using
only an optical encoder responding to a motor tachometer
having about 500 lines per revolution. The control means
for each DC servo electric motor executes concurrently
foreground tasks, midground tasks and background tasks in a
relatively short period of time. The various movements of
the workpiece and/or tool can be independent of each other
or can occur simultaneously.
It is an object of this invention to provide a micro-
computer numeric control system for controlling the relativemovements of a workpiece and/or tool so that the tool can
perform work on the workpiece in a three-dimensional space.
It is another object of this invention to provide a
micro-computer numeric control system using DC servo electric
motors wherein each motor responds to a signal generated by
a control means which concurrently executes fore~round
tasks, midground tasks and background tasks.
It is a further object of this invention to provide a
micro-computer numeric control system usiny a DC servo
3~ electric motor which operates using a relatively high DC
- 12~;7~57
voltage and has a relatlvely high motor torque to speed
ratio so that it may respond directly to an optical encoder.
It is yet another object of this invention to provide a
micro-computer numeric control system using a DC servo
05 electric motor wherein the natural damping characteristics
of the DC servo electric motor are utilized to hold the
various mechanisms in a stop lock mode.
Other features and advantages of the invention will be
apparent from the following more particular description of
preferred embodiments as illustrated in the accompanying
drawings in which like reference characters refer to the
same parts throughout the various views. The drawing is not
necessarily to scale, emphasis instead being placed upon
illustrating the principles of the invention.
Brief ~escription of the Drawing
Fig. 1 is front elevation of a schematic illustration
oP a machine for use with this invention;
Fig. 2 is a side elevation taken from the left of Fig.
1;
Fig. 3 is a side elevation taken from the right of Fig.
l;
Fig. 4 is a top plan view of Fig. 3;
Figs. S and 6 illustrate the drive means for moving the
workpiece or the tool;
Fig. 7 is a plan view of a tachometer disk;
Fig. 8 is an elevational view of the keyboard;
Fig. 9 is a pictorial view of the control panel;
Figs. 10 - 17 and 20 - 30 are flow charts;
``` ~a.~tj7~5~
Figs. 18 a - d illustrate the circuits on each of the
motor control boards; and
Fig. 19 illustrates the circuit of each servo amplifier.
05 Detailed Description of the Inventi~n
As described above, this invention is directed to
method and apparatus for controlling the relative location
and movement of a tool and a workpiece in a three-dimensional
space so that the tool may perform predatermined operations
on the workpiece. In the apparatus described in this
application, the workpiece is secured to a table of a
milling machine which table is mounted so that it may be
moved in a horizontal plane along an x-axis and along a
y-axis perpendicular thereto. Also, the table may be moved
in a complex path by the simultaneous movement along a com-
bined x and y axis. The tool is mounted so that its move-
ment may be controlled in the z-axis which is perpendicular
to the plane of the x-axis and the y-axis. With these
controlled movements, the tool may perform work on the
workpiece at any location in a three-dimensional space
within the limits of the movement of the table and the tool.
It is to be understood that the tool may be mounted so that
it is movable in any direction within the three-dimensional
space or that the workpiece may be mounted so that it is
movable in any direction within the three-dimensional space.
In addition to the foregoing movements, other operations may
be provided, such as means may be provided to rotate the
tool so as to perform an operation such as drilling. Also,
means may be provided fox the manual movement of portions of
the apparatus.
In the embodiment o~ the invention illustrated in the
drawing, there is disclosed a milling machine 2 comprising a
base ~, a main body 6, a quill support 8 secured to the top
of the main body 6 and a quill housin~ :L0. The front face
05 of the main body 6 is provided with a longitudinally ex-
tending support and guide 12 on which is mounted a knee 14
which is provided with a mating recess Eor engagement with
th~ support and guide 12 so as to form a dovetail arrange-
ment between the support and guide 12 and the knee 14. This
arrangement permits the movement of the knee 14 up and down
- in a vertical plane for a purpose described below. The knee
14 is moved up and down over the support and guide 12 by a
handle 16 which is connected to suitable gearing so as to
turn a lead screw mounted in spaced parallel relationship to
the support and guide 12 (not shown). A portion of the knee
14 comprises a precision ground ballscrews in threaded
relationship with the lead screw so that rotation of the
lead screw causes movement of the knee 14 up and down in a
vertical plane. Suitable means may be provided so that the
knee may be locked in position at any location along the
support and guide 12.
Mounted on the top surface of the knee 14 is a longi-
tudinally extending support and guide 18. A support member
20 is mounted on the support and guide 18 for movement
~hereover. The support member 20 is provided with a recess
22 which is provided with dove tail~d surfaces to mate with
dove tailed surfaces on the support and guide 18 to provide
for the dove tailed arrangement 24 illustrated in E'ig. l to
guide the movement of the support member 20 back and forth
in a linear direction over the support and guide la. The
~ j
~;i 6
`. ! '
support member 20 provides the movement over the y-axis.
The support member 20 is moved over the support and guide 18
by a servo motor 26 which rotates a lead screw 28 which ls
mounted in fixed spaced parallel relationship with tha
05 support and guide 18. A portion of the support member 20 is
provlded with a pre~ision ground ballscrews in threaded ,~
relationship with the lead screw 28 so that rotation of the ~
lead screw 28 in a clockwise or counterclockwise direction ~;
causes the support member 20 to move back or forth'over the
support and guide 18. The relationship between the support !~
i.
and guide 18 and the support member 20 is such that there
will be no relative movement therebetween except in responsè ,~
to rotation of the lead screw 28. ~ i~
A work holding table 30 is provided on its upper ;~
lS surface with a plurality of grooves 32, Fig. 4, dimensioned
so as to receive aonventional clamps so as to support work- ~
pieces at desired locations. The bottom surface of the work ~;
holding tabie 30 is provided with a longitudinally extending
support and guide 34 which is mounted in a longitudinally
extending recess 36 in the upper surface o~ the support ,
member 20. The support and guide 34 is provided with dove ~;
tailed surfaces designed to mate with dove tailed surfaces i:~
! in the recess 36 to provide for the dove tailed arrangement
38 illustrated in Figs. 2 and 3 to guide the movement of the
work holding table 30 left or right, as viewed in Fig. l, in ~ `
a linear direction over the recess 36 in the support member
20~ This provides the movement over the x-axis. The work
holding table 30 is moved over the recess 36 in the support
!' 30 member 20 by a servo motor 40 which rotates a lead screw 42 ~;
; which is mounted in fixed spaced parallel relationship with iJ'''~
' S~'" ~' '
.~, ~'.
'~ ~
` ~ 7
~2~
the recess 36. A portion of the support and guide 34 is
provided wlth a precision ground ballscrews in threaded
relationship with the lead screw 42 so that rotation of the
lead screw 42 in a clockwise or counterclockwise direction
05 causes the work holding table 30 to move left or right over
the recess 36 in the support member 20. The relationship
between the support and guide 34 and the recess 36 i5 such
that there will be no relative movement therebetween except
in response to the rotation of the lead screw 42.
A quill 44 is mounted in a guided passageway in the
quill housing 10 for vertical linear movement up and down.
A tool holding bit 46 is mounted in the quill 44 so that the
bit 46 moves with the quill 44 but can rotate relative
thereto. The quill 44 is moved up and down in the passage-
way by a servo motor 48 which rotates a lead screw which is
mounted in fixed parallel spaced relationship with the
passageway. An extension is secured to the quill 44 and is
provided with a precision ground ballscrews in threaded
engagement with the lead screw so that rotation of the lead
screw in a clockwise or counterclockwise direction causes
the quill 44 to move up or down in the passageway. This
provides the movement over the z-axis. The relationship
between the quill and the passageway is such that there will
be no relative movement therebetween except in response to
rotation of the lead screw. A motor 50 is provided to
rotate the tool holding bit 46. The connection between the
motor 50 and the tool holding bit 46 is such that the tool
holding bit 46 may move up and aown with the quill 44 but be
rotated relative to the quill 44. Such a connection could
comprise a rotatable pulley mounted in a relatively fixed
~ 8
location and adapted to be rotated by the motor 50. A shaft
would be connected to the pulley for rotation therewith and
the shaft would extend into a passageway in the tool holding
bit 46. A splined relationship comprising A key on the
05 shaft and a keyway in the passageway would allow the tool
holding bit 46 to slide on the shaft and the shaft to rotate
the tool holding bit 46.
The milling machine 2 is designed so that one or more
workpieces may be mounted on the work holding table 30
depending on the si.ze of and the number of aifferent opera-
tions to be performed thereon. If a plurality of operations
are to be performed on a plurality of workpieces, a work
holding device for each workpiece is properly secured on the
work holding table and a workpiece is secured between two
parallel edges of each work holding device. Each work
holding devic~ is then secured in associated grooves 32 in
the work holding table 30 so that the parallel edges of each
work holding device are properly oriented perpendicular to
or parallel to the grooves 32. The plurality of workholding
devices would be in spaced relationship. A first tool would
be positioned in the tool holding bit 46 and a first pro-
grammed operation would be performed on each workpiece. The
first tool would then be removed and a second tool posi-
tioned in the tool holding bit and a second operation would
be performed on each workpiece. This would be repeated
until all work operations had been performed.
As described above, the major control of the movement
of the work holding table 30 along the x-axis and the y-axis
are the servo motors 26 and 40 and the major control for the
movement of the quill 44 along the z-axis is the servo motor
48. In the preferred embodiment of the invention, these
servo motors comprise a permanent magnet DC servo electric
motor. One motor and one tachometer assembly is used for
each axi~ of control. As explained below, each DC servo
05 electric motor and tachometer assembly is associated with a
servo amplifier board and a motor control board. The DC
servo electric motors of this invention are designed to run
with a motor torque to speed ratio which is rather different
from those used on other systems associated with the driving
of lead screws. In the milling machine, described above,
each lead screw is coupled to a belt with relatively high
belt drive ratios which for the control of the work holding
table 30 along the x and y axis being in the range of about
3.2 to 1.0 and for the movement of the quill in the z-axis
is in the range of about 2.4 to 1Ø Conventional control
systems use either direct drive or ratios not exceeding
about 2.0 to 1Ø The use of the higher drive ratios means
that the motor is required to develop only half the torque
of conventional systems while running at twice the velocity.
This permits the use of a motor of substantially reduced
size so as to result in initial cost reduction and opera-
tional cost reduction.
Another significant advantage gained by using this high
mechanical drive ratio is in the reduction of motor tacho-
meter costs. The optical tachometer used with conventionaldirect drive motors typically contain about 2000 lines per
revolution whereas the optical tachometer used with the high
mechanical drive ratios of this invention require only about
500 lines per revolution. Another important advantage
gained by the use of the high mechanical drive ratio system
~ 3~
is that the servo position resolution requirement is signif-
icantly reduced which results in that a lower cost servo
motor may be utilized. Another major advantage of being
able to use an optical tachometer haviny only about 500
05 lines of revolution is that only an optical encoder is
required for feedback to ensure stability of the velocity
servo loop. Conventional motor control systems require an
analog *achometer, in addition to the optical encoder, to
achieve about the same degree of stability of the velocity
servo loop.
~he above described motors operate at relatively high
DC voltage but for a fixed motor horsepower, the current
required to operate the motor is reduced by the higher
voltages. New high voltage power transistors in both n- and
p-channel may be used and these provide significant cost
savings when using low current, high voltage DC motors.
The high mechanical drive ratio, described above, is
accomplished in accordance with this invention by the
structure illustrated in Figs. 5 and 6. A pulley 52 is
secured to the shaft 54 of each servo motor and each pulley
52 is provided with a plurality of peripheral teeth 56. A
drive cog belt 58 is trained around the pulley 52 and
another pulley 60 secured to the end of a lead screw 62. As
described above, each lead screw 62 is rotatably mounted in
a fixed position so as to guide one of the movable elements
in a path along the x, y or z axis. As illustrated in Fig.
6, the movable element is the support member 20 which has a
threaded extension 64 secured thereto and in threaded
engagement with the lead screw 62 so that rotation of the
lead screw 62 causes movement of the support member 20. The
pulley 60 is also provided peripheral teeth 66. The cogs 68
of the drive cog belt 58 are in mesh with the peripheral
teeth 56 of the pulley 52 and the with the peripheral teeth
66 o the pulley 60 so that the lead screw 62 is positively
05 driven by rotation of the shaft 54 of the servo motor. The
high mechanical drive ratio of 3.2 to 1.0 is accomplished by
having ten peripheral teeth 56 on the pulley 52 and thirty-
two peripheral teeth 66 on the pulley 60. The high mechan-
ical ratio of 2.4 to 1.0 is accomplished by having ten
peripheral teeth 56 on the pulley 52 and twenty-four peri-
pheral teeth 66 on the pulley 60.
An optical encoder (conventional) is associated with
each of the servo motors and is used to provide feedback for
the position and the velocity control The encoder incorpo-
rates a disk having a plurality of spaced apart radiallyextending slots which slots terminate adjacent to the
periphery of the disk. A light emitting diode is positioned
so that the light will pass through each of the slots when
that slot is located between the light transmitting diode on
one side of the metal disk and a photo transistor on the
other side of the metal disk. In a preferred embodiment of
the invention, illustrated in Fig. 7, the disk 70 is made
from a transparent material and the slots 72 are formed by
imposing a plurality of opaque radially extending lines 74
so that the light will pass through the transparent material
located between the opaque lines. The photo transistor is
used to pick up the light beam and send siynals to the
electronic circuits, described below. As the disk rotates,
a series of pulses is sent indicating the rotational velocity
of the servo motor.
12
.~
A second pair of a light emitting diode and a photo
transistor are used to indicate the direction of rotation oP
the motor. This second pair is physically located so that k
their electrical output is 90 degrees out of phase with the
05 first pair. Therefore, when the servo motor i~ rotated in a
positive direction, the second signal will lag the first by
90 degrees and when the servo motor is rclated in a negative l `
direction, the second signal will lead the first by 90
degrees. In this manner, the two pairs of optical detectors ~;
produce four signals for each slot in the disk.
While the foregoing description of the servo motors is ~.
~` associated with the operation of a milling machine 2, it i9 to be understood that servo motors of this invention may be
used to control movements of a workpiece or a tool in the
computer controlled operation of any type of apparatus. As ~1
stated above, the method ànd apparatus of this invention may 6~ "
be used with a wide variety of industrial or commercial ~i
equipment. In the following description of the method and
apparatus of this invention, specific references will be ~-
;1 20 made to the milling machine as described above and illu8
~" {6 trated ~n Pigs. 1 - 4 of the drawing. However, it is to be
,!~ ` understood that these references are for illustration
purposes only and that the method and apparatus set forth
!~ 25 below in this application may be used to provide robotic 6 j
controls to a wide variety of machines used in the manufac~
ture of various types of products.
The Micro-Computer Numeric Control ~uCNC~ system
comprising the method and apparatus of thi3 invention is a
multi-axis motion control ~ystem. The uCNC system provides
for robot-like execution of a complex pattern of motions.
~';
13
~r~
',.
The specific machine sequence for a given application is
stored on a micro-diskette. The machine sequence may be
programmed on the uCNC controller or on any personal com~
puter, minicomputer or main frame.
05 The uCMC system is composed of four baslc building !
blocks: a controller, a servo box, a servo motor/tarhometer, ~
and an operator ccntrol panel. Using these four basic ~ ~ ;
building blocks, a machine manufacturer may ea~ily add
robotic controls to a wide variety of industrial or commer- ,
cial equipment. A signlficant feature of the uCNC motion '
control system is that the machlne manufacturer can move
into robotic controls without mastering the technologies of ~ -
computer programming, servo dynamics or electronic design.
The heart of the uCNC system is the controller 76 in ~-
Fig. 1. The controller is actually an lndustrialized and
speciallzed micro-computer. The keyboard 78 in Fig. 8 on
the controller has been designed to provide simplified
motion control command inputs. The computer screen 80
., ~
provides helpful commands and instructions to the machine
i 20 operator. The machine control sequence is stored on a
micro-diskette which is built into the controller. The -~
built-in keyboard and screen may be used to generate and ~`
~-~ store the speciflc machine sequence for a given application. ~,
~;i, 25 The servo box located in the main body 6 contains all
the motor control and servo amplifier boards required to ~;
control a number of electric motors in response to a machine E~
sequence stored in the controller. The servo box also
contains a number of auxiiiary relays which may be actuated ~ ,`
under program control.
;',
~3rj~
The motor/tachometer assembly provides the actual
mechanical output for the system. An optical encoder 82 is
attached to the back of each of the servo motors 26, 40 and
48 and provides the required "feedbac~" for controlling the
05 position and velocity of each servo motor as required by the
controller 76.
The operator control panel provides for manual control
by the operator. The operator control panel contains
switches 84 for manually controlling the machine sequence.
The panel also contains an emergency stop switch 86 to
immediately stop machine motion should the operator detect
an unsa~e condition. The panel contains joy sticks 88, 90
and 92 for manually controlling the motion of the servo
motors 26, ~0 and 48.
The controller 76 contains the main computer board ~not
shown), the micro-diskette drive (not shown), the CRT 80
(cathode ray tube or "screen"), the CRT interface board (not
shown~, the keyboard 78 and auxiliary circuit board (not
shown). All components are housed in an industrial grade,
splash-proof, metal housing 9~. The controller 76 uses the
CP/M operating system running on a Z80A microprocessor
operating at 4 million cycles per second. The system pro-
yides for 65,000 characters of solid state memory and
250,000 characters of disk memory. The master program
within the controller is written in CB80, which has been
compiled to machine language. The controller provides 300
baud serial communications for use with an external tele-
phone modem. The controller communicates with the servo box
over a multiplexed 9600 baud duplex serial interface which
includes one handshaking line.
~i~ 15
~ ~795~7
The software supplied with the controller allows the
user to operate the machine in one of two basic control
modes. The first mode is "PROGRAM MODE" and the second mode
is "RUN MODE". In program mode and the run mode, the user
05 may perform the operations which appear on the CRT screen
opposite to the lettered keys A - F on the keyboard 78 as
follows:
PROG~AM MODE RUN MODE
LOAD JOB A SET UP A
CREATE JOB B AUTO B
SAVE JOB C MDI C
ERASE JOB D JO~ D
EDIT E E
RUN F PROGl~ F
The use of these soft keys is very important to the design
of the controller because the soft keys can take on many
different functions with each new function displayed on the
screen. This means that the controller may be designed with
a very simplified key board, as shown in Fig. 8, and yet
perform very complex control functions. In competitive
machines the keyboards are very complex due to the large
number of switches required to perform necessar~ control
functions. This large number of switches increases the cost
of the controller and makes the controller difficult to
learn to use The soft keys reduce costs by simply reducing
component cost. This also increases product reliability,
since there are fewer parts, there are fewer failures. The
simplified keyboard also simplifies the learning process as
3~ required key functions are presented to the user only when
7~
they are active. Inactive key functions are, in effe~t,
hidden from the user.
LOAD JOB EDIrr-EuNcTIoNs
Enter Job A 1 3789 Back A
05 For Retrieval B 2 C456
CRC DDD DEF DEFF C 3 B123 Forward B
FEO ABCD IP CBC D 4 E--8
F E 5 F-14 Display
F 6 ACMIR
7 Insert D
8 Delete E
9 Delete BLK F
Two oE the sub-menus which may be selected from the
PROGRAM mode are shown under the LOAD JOB and EDIT E'UNCTIONS.
If the "LOAD JOB" function is selected, the CRT screen will
show the above column under LOAD JOB. In this screen, the
selection of jobs available, on the current diskette, is
displayed for user selection. The line at the top of the
screen requests that the user enter the name of the job
which is to be loaded. When this screen is displayed, the
"A" through "F'l keys no longer function as "soft keys", but
take on their alpha functions. The shift key, in the lower
right hand corner of the key board, may be used to select
the alpha functions printed above each number on the key-
board. When in the "shifted" pattern, all letters of the
alphabet are available for construction ~ob names. A red
light above the shift key is illuminated to inform the user
that the keyboard is in the alpha shift position.
7~35~
If the edit function is selected, the CRT screen will
show the above column under EDIT. With this screen the user
may:
1. Scroll the display up or down to examine job
05 programs too long to fit on the screen.
2. Insert a new line into the program.
3. Delete a line from the program.
4. Modify or change a line.
5. Append to the bottom of the program.
The language used to write job programs is defined by
Electronics Industries Associations standard RS-274C,
described in the reference listed under Other Prior Art AR
in the Information Disclosure Statement filed herewith and
identified as: "Interchangeable Perforated Tape Variable
Block Format for Positioning, Contouring, and
Contouring/Positioning Numerlcally Controlled Machines". In
the industry this code set is simply referred to as MDI
programming (Manual Data Input). Each instruction line in a
job program instructs the machine to move to a new position,
start a motor or pump, or perform a pre-~efined series of
operations. Various commands in the job program allow the
user to instruct the machine to repeat certain cycles or
enter certain modal conditions as defined in RS-27~C. The
operation of the controller is conventional in that the
operator feeds information and the computer responds to the
inf ormation .
The Controller communlcates using an ll-bit byte of 1
start bit, 8 data bits, 1 even parity bit, and 1 stop bit.
At 9600 bits per second, each character will transmit in
3~ 1.146 milliseconds.
18
9-~
The controller recognizes two types of message formats,
DIAGNOSTIC MODE and COMMAND MODE. In Diagnostic Mode, each
byte represents a single ASCII character, which allows an
individual to control the system using a terminal.
05 Axis logic boards default to Diagnostic Mode. In order
to recognize Command Mode messages, they must receive a
Diagnostic Mode message to enter Command Mode. Once in
Command Mode, the Axis logic boards will not recognize
Diagnostic Mode messages unless they receive a Command Mode
message to enter Diagnostic Mode or are reset.
Command Mode messages minimize communication overhead
by allowing each byte to represent a binary value from 0 to
255. This prevents the use of unique start and stop bytes,
so Command Mode must use other methods to identify the
beginning and ending of a message.
The Controller identifies valid transmitted data by
raising Data Set Ready (DSR). DSR must be low for at least
one character time between the end of the previous message
and the start of the next.
Since an axis logic board must receive a valid message
addressed to it before it can transmit a message, the
Controller regulates the spacing between received messages.
A space of at least 3 character times identifies the end of
one message and the beginning of another.
The longest Command Mode message is ten characters,
making the maximum transmission for one message from the
controller at 9600 bps = 10 * 1.46 = 11.4~ milliseconds.
COMMAN~ MODE MESSAGE FORMAT
All Command Mode message use the following format:
19
CD...DS
where C i5 the COMMAND message type and axis address
D are the DATA bytes
S is the SUM of the message
05
COMMAND
The Controller uses the Command code to specify the
type of action that the axis is to perform, which axis the
transmitted message is to go to, and where the received
message came from.
The most significant digit of the command code spec-
ifies the message type. The least significant digit spec-
ifies the axis address. This allows the Command Code to
specify 16 different message types to 15 different ad-
dresses; an axis address of 0 specifies a global message
addressed to all axis boards.
,
DATA
Each message transmit data as hexadecimal bytes. Each
data type uses fixed-length fields. Multiple-byte fields
transmit most significant byte first, least significant byte
last.
SUM
The Sum byte is the hexadecimal sum of all the charac-
ters in the message modulo 25~, less the Sum byte.
The following describes a single communication inter-
change:
~,J 20
9S7
1. The Controller raises Data Set Ready (DSR) to
indicate the start of the message.
2. All axis boards detect DSR and prepare to receive
a message.
3. The Controller transmits a single character during
each interrupt until is has transmitted the entire
S message. Immediately after the last character
transmitted, the Controller lowers DSR to indicate
the end of the message.
4. All axis boards store each character until they
detect the low DSR. At this time, they will have
received the entire message, and can process the
message.
If the axis detects an error in the message, it will
ignore the message. The axis will detect an error through
either parity, length or checksum.
If the axis receives a message with a non-~ero axis
address, and the address does not match the axis board's
address, the axis will ignore the message.
The Controller may transmit a message to an axis at any
time. No axis may transmit unless it has received a valid
message addressed specifically to it. All axis boards must
respond to a valid message that is addressed specifically to
it, either with the requested data, or with an acknowledge.
All axis boards execute global messages, but do not send a
message back to controller.
Only one axis board at a time may transmit a message.
In order to ensure that the controller can separate two
messages, there must be at least three character times
between the end of one axis board message and the start of
any other axis board message.
If the Controller transmits a non-global message, it
will wait a certain time for the axis to respond. It if
times out, or if it receives an invalid reply, the
Controller will retransmit the same message two more times.
If the Controller is unsuccessful all three times, it will
assume that the axis is defective, disable servo power, and
display an error message on the screen.
05
COMMAND MODE MESSAGES
The VELOCITY message uses the following format:
CFFS
where C is the COMMAND message type (0) and axis
address
FF is the FEEDRATE
S is the SUM of the message
The Velocity message overrides the programmed speed of
the motor. The Feedrate is in units of inches per minute *
256. ~hen an axis receives a Velocity message addressed to
it, it will adjust the velocity set point of its servo to
achieve the desired speed, and transmits a Verify message in
response.
The CONTOUR START message uses the following format:
CPPPMBTRS
where C is the COMMAND message type (1) and axis
address
PPP is the contour reference POSITION
M is the contour offset MULTIPLIER
B is the contour offset BYTE position
T is the TORQUE limit
R is the acceleration RATE
S is the SUM of the message
22
~ 7
The Contour Start message signals the start of a
coordinated two-axis motion. When an axis receives a
Contour message addressed to it, it will store all data
value and go into the state 1^o receive an indefinite number
05 of Contour messages.
The C~NTOUR STEP message uses the following format:
CAlA2S
where C is the COMMAND message type (2) and axis
address to repond
Al is the contour offset for AXISl
A2 is the contour offset for AXIS2
S is the SUM of the message
Contour Step messages specify the distance to travel
from the contour reference position for one interval of
time. The controller will transmit a Contour Step message
only after a Contour Start message or a previous Contour
Step message.
When the controller transmit a Contour Step message,
the axis hoards that recei~ed the Contour Start messages
will take their axis steps starting from the contour offset
byte position and multiply it by 2 to the power of the
contour offset multiplier, and then add it to the contour
reference position. The sum is the position for that axis
in absolute quarter-tach counts.
The Address defines the axis that is to respond. If
the Address is zero, then no axis responds. If the Address
is a motor axis, that axis will respond with its current
position and status. If the Address is the Relay axis, that
23
~7~
axis will respond with the spindle pad status, the relay
status, and the feedrate override value.
The REQUEST message uses the following format:
05 CDS
where C is the COMMAND message type (3) and axis
address
D is the DATA type of the request
01~ = position
02H = velocity
03H = torque
04H = difference in counts
05H = servo status
FFH = position, velocity and torque
S is the SUM of the message
If the Controller requests position (OlH), the axis will
return:
CDPPPS
If the Controller requests velocity (02H), the axis will
return:
CDFFS
If the Controller requests torque (03H), the axis will
return: :
CDTS
If the Controller requests a difference ~04H), the axis will
return:
CDTTS
If the Controller requests servo status ~05H~, the axis will
return:
24
CDB
If the Controller requests everything (FFH~, the axis will
return:
CDFFPPPTS
05
The MOVE message uses the following format:
CFFPPPTRS
where C is the COMMAND message type ~4) and axis
address
FF is the FEEDRATE
PPP is the absolute POSITION
T is the TORQUE limit, from O to 255
R is the acceleration RATE, from O to 255
S i8 the SUM of the message
When an axis receives a Move message addressed to it, it
will accelerate to the specified feedrate and move along its
axis until it reaches the specified position.
The ABSOLUTE MEMORY READ message uses the following
format:
CAAAABS
where C is the COMMAND message type (6) and axis
address
AAAA is an absolute memory ADDRESS
B is the number of BYTES to read
S is the SUM of the message
When an axis receives an Absolute Memory Read message
addressed to it, it will transmit a message containing a
memory image starting from the absolute memory address for
the number of bytes specified in the message. The axis will
echo the number of data bytes as the first byte of data.
05 The ABSOLUTE MEMORY WRITE message uses the following
format:
CAAAABD..DS
where C is the COMMAND message type (7) and axis
address
AAAA is an absolute memory ADDRESS
B is the number of BYTES to write
D,.D are the DATA
S is the SUM oE the message
15 When an axis receives an Absolute Memory Write message
addressed to it, it will store the data starting from the
absolute memory address for the number of bytes specified in
the message.
The RELATIVE MEMORY READ message uses the following
format:
CABS
where C is the COMMAND message type (8) and axis
address
A is a memory table ADDRESS
B is the number of BYTES to read
S is the SUM of the message
When an axis receives a Relative Memory Read message ad--
dressed to it, it must first determine the absolute memory
~ 26
address to start from. The axis ROM contains the memory
table. The memory table address specifies the desired entry
in the memory table; an address of 0 selects the first
address in the table.
05 The axis will then transmit a message containing a
memory image starting from the selected memory address for
the number of bytes specified in the message. The axis will
echo the number of data bytes as the first byte of data.
la The RELATIVE MEMORY WRITE message uses the following
format:
CABD..DS
where C i5 the COMMAI~D message type (9) and axis
address
A is a memory table ADDRESS
B is the number of BYTES to write
D.. D are the DATA
S is the SUM of the message
:
When an axis receives a Relative Memory Write message
addressed to it, it must first determine the absolute memory
address to start from. The axis ROM contains the memory
table. The memory table address specifies the desired entry
in the memory table; an address of 0 selects the first
address in the table.
The axis will then store the data starting from the
selected memory address for the number of bytes specified in
the message.
The SERVO message uses the following format:
~7
CDS
where C is the COMMAND message type (1) and axis
address
D determines whether it is a LOCK or UNLOCK
05 message; FFH = Unlock, OOH = Lock
S is the Sum of the message
.
When an axis receives a Servo Lock message addressed to it,
it will perform a stop and hold at its current position.
When it receives a Servo Unlock message, it will remove
power from the motor.
The ZERO message uses the following format:
CDS
15 where C is the COMMAND message type (11) and axis
address
D defines the DATA type to zero
01H = position
03H = torque
S is the SUM of the message
When an axis receives a Zero message addressed to it, it
will set the selected value to a zero condition. If the
selected value is Torque, the axis will take its current
pulse width value as the servo center point. If the selected
value is Position, the axis will reset the absolute position
counter to its initial value.
28
The ACKNOWLEDGE message uses the following format:
CDS
where C is the COMMAND message type (12) and axis
address
05 C is the COMMAND code and axis address
D determines the type of Acknowledge
zero = Acknowledge, non-zero = error
S is the SUM of the message
10 An axis must transmit an Acknowledge message if it received
a valid message addressed to it, and there is no explicitly
defined response. The axis takes the axis Address and
Command code from the received message.
DIAGNOSTIC MODE MESSAGE FORMAT
All Diagnostic Mode messages use the following format:
: [AVNAAAADD.. DDLLSS]
where [ is the START BYTE
A is the DEVICE ADDRESS
V is the VERB code
N is the NOUN code
AAAA is the DATA ADDRESS
DD.. DD is the DATA
LL is the message LENGTH
SS is the SUM of the message
] is the STOP BYTE
START BYTE
29
The "[" character defines the start of the message. Any
axis in Diagnostic Mode that receives the Start Byte will
store the following date into its receive buffer.
05 DEVICE ADDRESS
The ASCII letters from A to Z will convert to axis address 1
to 26
VERB
A mnemonic ASCII character defines the verb code; R = read
and W = write
DATA ADDRESS
The axis address is the ASCII Elex address of the port or
memory.
DATA
Data are always represented as a 2 ASCII Hex digit value.
If the verb code is a read operation, the data specifies the
number of bytes to read. If the code is a write operation,
the data specifies the bytes to write.
LENGTH
The message length is a 2 ASCII Hex digit value that is the
sum of th~ message modulo 256 of the characters between the
start byte and the first checksum byte.
STOP BYTE
The "]" character defines the end of the message. This
tells the receiving axis that the message is complete and
should now process it.
05 The flow charts for the software used by the controller
are illustrated in Figs. 10 - 17 and 20 - 30 of the drawings.
The first chart, Fig. lO, labeled "Master Logic", indicates
the general overall structure of the software. When the
controller is first powered up, it first executes the
initialization tasks. Following initialization the controller
executes three programs concurrently. The shortest program
is the "Interrupt Routines". The interrupt routines are
executed once every millisecond and require approximately
0.3 milliseconds to operate so that the interrupt routines
use about one--third of the processors time to execute. A
"CTC" (Counter Timer Chip) is used to generate a hardware
interrupt signal to the processor on fixed time intervals.
Each time that this interrupt signal is generated, the
processor will stop whatever it is doing and execute the
interrupt service routines.
Next in level of priority to the interrupt service
routines are the coordinated motion tasks. The coordinated
motions program will generate a new set of coordinates each
ten milliseconds. These coordinates are transmitted to the
individual control servos by the interrupt service routine.
Working together in this manner the controller may generate
geometric shapes consisting of short, straight lines, which
are produced at the rate of one hundred points per second.
Both the coordinated motions tasks and the interrupt service
routines are written in assembly language to ensure maximum
31
3L~$~
speed of execution. If these programs were written in
conventional high level language and then compiled, the
execution times would be fifty to seventy times longer than
the assembly language routines developed for co~troller.
05 When the controller is not executing the interrupt
routines (foreground) or the coordinated motions tasks
(mid-ground), the background program may be executed. The
background program produces all of the displays discussed
above in addition to mathematical, interpretive and data
handling tasks required to convert MDI codes to servo
instructions. The background program is written in a high
level language called "CB80". The source code, written in
"CB80", is compiled to machine language for execution on the
controller.
The 10w charts illustrated in F.ig, 11 describe the
initialization tasks. The first task is to preset all of
the memory variables required for execution of the "CB80"
background program. The second task is to initialize all of
the programmable input/output chips, such as the DART serial
communications chip and the CTC counter timer chips. Third,
the ~IOS (basic input/output system) is modified from
default console status and I/O routines to the special
routines required to communicate with the controller's
keyboard and CRT screen. Next, the machine language code
for interrupt and coordinated motions tasks is loaded ana
the interrupt mechanism enabled. Finally, control is passed
to the CB80 background program.
The flow charts illustrated in Fig. 12 indicate the
tasks which are performed each time that the CTC generates
an interrupt to the processor. The first task is to update
32
~7~7 ;1, .
- the real time clock which is used as the master timing
control for the controller. Second, the axis logic I/O
tasks are performed~ this is the serial communications which ~
is multiplexed to and from the indiv~dual servo control ~ '
i~ OS microprocessors. ~he third task is to perform I/O communica-
tions to the console output; this is a separate microprocessor
which controls the formatting and flow or characters to the
CRT display. Fourth, the keyboard is read to detect if any
key has been closed; one column of the keyboard is read ~5
during each interrupt. Finally, the interrupt routines
determine if it is time to start calculating a new coordinated
!j motions point. After this task, control i3 transferred to .;~
either the mid-ground or background tas~ in progress.
The flow charts in Fig. 13 illustrate the tasks performed ~ ;
by the coordinated motions tasks software. When the interrupt $.
.5 routine has determined that it i9 time to calculate a new ~`~
pair of coordinated move points, control i9 passed to the
~! coordlnated motions tasks software kather than back to the 5~
, background software). The first task performed is to ~ ;
;1 20 determine which scctor is to be used for the current
calculation. To achieve the speed required for the coordinated
. points calculations, the 360 degree circle is divided into
eight sectors. Each individual sector is used to perform
calculations within a 45 degree arc (.7854 radians). By
limiting the sector size to 45 degrees, two important
objectives are achieved. First, the Fourier ei~pansions used
to calculate Sine and Cosine functions may be solved to
required accuracy with only three terms in the expansion.
Second, the angle term, in radians, is always a fraction
l~ 1) so that a modified form of integer math may be used to
~.,.
.'. ~'~ '
3 3 ~i~
solve the equations using only sixteen bit numbers. It is
this unique combination of these two features which allows
the controller to perform the required one hundred points
per second calculation necessary to generate coordinated
05 moves.
Once the correct sector location has been determined,
the coordinated motions soft~are will solve the expansions
to calculated Sine and Cosine terms, multiply the Sine and
Cosine by the radius, determine the correct X and Y (or Z)
relationships, and add the polar coordinate components to
pre-established offset values. Once the geometry has been
solved, the software will assembly a message in the correct
format for transmittal to the individual servo control axis.
Finally the assembled message is transferred to the transmit
buffer llsed by the interrupt routine for transmitting
messages. Control is then returned to the background
software.
Since the Z80 microprocessor is basically an eight bit
processor, with some limited sixteen bit capability, it is
important that the geometric calculations described above be
performed using sixteen bit multiply and divide routines.
The milling machine requires an accuracy of .0001" over a
travel of 40"; or an accuracy of one part in 400,000. In
binary arithmetic, the only type a microprocessor can
perform, the required resolution dictates that twenty-four
bit math be used for calculations. To resolve this difference,
a scale factor is applied to all coordinated moves calculations.
This scale factor, (l, 2, or 4) is used as a multiplier for
radius and for X, Y or Z moves. Usiny this technique the
controller may calculate circles of thirteen inches diameter
34
to .0001 accuracy; or circles of fifty-two inches diameter
to .0004 accuracy.
The flow chart in Fig. 14 illustrate the structure of
the Cs80 background tasks. The structure o~ this flow chart
05 follows the structure described in the display screen
section above. The flow chart for the edit function is
shown in Fig. 15. Fig. 16 shows the flow charts for the
"set-up" and "jog" functions. Fig. 17 shows the flow chart
for the auto-execute functions. In the case of these latter
three functions the CB80 software i5 also performing the
real time job of maintaining the X, Y and Z positions
displays in three locations on the screen.
The controller's monitor is a 12" black and white CRT
with a 12 mHz bandpass. The controller keyboard is a cus*om
configuration designed specifically for motion control work
using ANSI standard MDI data commands. The keyboard is
scanned using a row/column matrix. The scan is under
software control using electronic logic located on the
auxiliary board. The auxiliary board also contains the
multiplexer which directs 9600 baud serial commands and
responses to the CRT controller board or 300 baud commands
and responses to the external modem port. An optoisolator
located on the auxiliary board is wired in series with the
two emergency stop switches. This isolator allows the
computer to simultaneously disable all servo motion and to
open all power relays (stopping all external motors).
The CRT control board has its own microprocessor and
display memory. The microprocessor is a special modification
of an 80~8 processor which contains dedicated hardware for
video clock, character generation and video dot shifters.
?,,",,
7:~1S7
This CRT controller allows for a wide variety of video
attributes such as blinking, enhanced, double wide and
double high characters. These attributes may be mixed at
random within a common screen image. The CRT controller
05 contains two 2048 word memorles, one for character storage ~`
and one for attribute storage. J ~.
The micro-diskette drive uses 3.5 ir.ch hard shell ~;
diskettes. These diskettes are of convenlent size such as ~,
to fit in a normal shirt pocket. The hard shell of the
diskette provide.s a high degree of mechanical proteation for G
the data storage media. A metal "window" covers the read
opening in the diskette for further data protection. Each i~
single sided diskette may store 250,000 characters of
formated information. Approximately 50~ of thl~ storage
space is available for user program storage (a user program l~
18 a specific ~equence of machlne motions). Thi_ storage
; space is equivalent to over 1000 feet of punched tape ';
program. The second 50~ of the diskette storage area i9
u~ed by the mastex operating programs.
The controller contains internal voltage regulators to
produce required 5, +12 and -12 voltQ for logic, display and
communications. The regulators receive unregulated power
- from the power box. The -12 volt supply ls generated by aninverter from the +12 supply. ~ ~
The servo box contains the individual motor controls, ( -
relays and all power supplie~. ~ach servo motor i9 aQsigned
a motor control board and a servo amplifier board. The
motor control board receives commands from the controller, i -
indicatlng the desired position and velocity for the servo ~ -
i motor. These commands are in the form of a serial data
~, ,' .
. ~,
36
stream as defined in the Communications Protocol. The motor
control board also processes signals from the servo motor
optical encoder ~tachometer). By comparing commanded
position and velocity to actual values, the motor control
05 board may generate corrective commands to the servo
amplifier board.
Each motor control board is actually a miniature
computer. The heart of the board is a Z80A microprocessor
(U5) Fig. 18, operating at 4 million cycles per second. The
microprocessor is supported by 4048 words of program storage
(U3), 202~ words of RAM memory (U4) and an industry standard
8251 UART for serial communications (IJ8). The
microprocessor is operated ln interrupt mode with
communications, velocity count read and PWM generation in
the foreground. A unique midground cycle is used to process
velocity and position servo data. The conventional
background mode is used for all other processing work load.
The motor control software algorithm is encoded in assembly
language source code for maximum processing thru-put.
The motor control board commands the servo amplifier
over a single control line. This signal is processed by an
opto-isolation chip, OIC Fig. 18 D, to ensure that
electrical noise generated in the motor does not communicate
into the logic signals of the microprocessor. The servo box
is further divided into two compartments with a metal EMI
(electromagnetic interference) barrier between sections to
further reduce motor noise coupling.
The single amplifier control line encodes motor control
commands using a protocol in which commanded motor current
is proportional to the width of pulses on the line. The
pulse width modulation tPWM) is updated 5,000 times per
second by the microprocessor. A Z80 CTC (counter timer chip
- U1) Fig. 18 D, is used to generate a non-maskable
interrupt ~NMI~ each 200 microseconds ~5,000 x 200 uSec = 1
05 Sec). Channel "O" of the CTC is programmed in timer mode.
The CTC is clockea at ~mHz and has a prescaler of 16 on the
timer. A count of 50 is loaded into channel "O" to generate
the NMI signal each 200 microseconds (lZ x 50/4Mz = 200
uSec).
Each time that the microprocessor receives a
non-maskable interrupt, it loads channel "l" of the CTC with
a count which is proportional to the pulse width desired.
Channel "1" is clocked at 1 mHz. A 50% dut.y cycle is
obtained by loading channel "l" with a count of 100. (Note:
under PWM protocol a 50~ duty cycle represents a null
command to the motor.) At the start of each non-maskable
interrupt the microprocessor presents the PWM flip-flop
(LS393 - U15) by pulsing the "PWM~" line. To preserve
foreground time this pulse is generated by addressing
non-existent memory at address "EOOOh". The memory address
decoder chip, LS138 (U16), Fig. 18 B, then produces a
negative pulse on the line "PWMO". Channel "1" will produce
a pulse when the present time is complete. This pulse is
used to reset the PWM flip-flop, thereby generating the
desired pulse width modulated signal.
The microprocessor also performs character transmit or
receive during the interrupt time. The controller
communicates with the motor control boards using a 9600 baud
serial signal. The data character is 8 bits long, plus
start, parity and stop bits for a total of ll bits per
character time. Each character time is therefore 11/9600 or
1.14
milliseconds long. To conserve foreground time, the micro-
processor will attempt to either transmit or receive a
character every other interrupt ~. a milliseconds per attempted
- transmit or receive). Using this procedure, the interrupt
05 processing time averages 100 microseconds (or 501 of the
microprocessor time ic spent in foreground and 50i~t in
background).
Motor position and velocity are derived by the micro- ~-
procesi30r using ~ignals from a three phase optical encoder. il
Two phases are used to generate "up" and "down" count pulses ~`-
for each quarter tachometer line of motor rotation. These Z;
count pulses are generated, Fig. 18 C, by a LS175 flip-flop ~rZ
IU12) and a LS153 decoder (U13)~ The pulses are then fed
! lnto a CTC for microprocessor intsrface. "Up'l counts are i;~
i i
1 15 f0d into channsl "0" oE the CTC lU2). Overflow from channel
~0~ i5 used to clock channel "1" to provide a 16 bit "Up"
count. "Down" counts are fed into channel "2" of the CTC. Z5~
Overflow from channZlal "2" is used to clock channel "3" to
pro~ide a 16 bit down counter. The microprocessor will then
process the difference betwesn channels "0" and "1~ and
channels "2" and "3" as the current position location.
The optical encoder~ ussd have 500 tachometer lines per i~
'! i,~ .
~ revolution. The position dstermination achisved by the
;! 25 microprocessor is accurate to one quarter tachometer, or ~ `
i 2,aoo quarter tachometsr lines per revolution ~.18 detrees ~;
per count). If this motor were coupled to a metal working
' milling machine have a lead scre~w pitch of .2 inch per
revolution and a motor drive ratio of 3.2 to 1, each quarter
tachometer would reprssent 31 m1llionth of an inch of motor. i~
~,
. ZiZ` '' ~
Zl Z' ' , ~ Z ~ ~ r~
2~7957 :~k.:
~otor velocity is determined by measuring the elapsed
time between tachometer lines. Channels "2" and "3" of a
CTC (Ul) are used to accumulate the nu~ber of clock pulses
between tachometer lines. A "NAND" gate IU~ Fig. 19, is ~ ;
05 used to gate 2 ml~z clock pulses into channel "2" of the CTC. ~-Overflow from channel "2" is used to clock channel "3",
forming a 16 bit velocity counter. The "NA~D" gate allows y~
clock pulses to pass when the output of LS151 tU14) multi- ~ ;
plexer is "high". When the output of the LS151 is low the ~ ~ ;
clock pulses are inhibited and the microprocessor may read
the velocity count from the CTC. The microprocessor deter-
mines the correct read "window" by observing the polarity of
the signal "VELCNT". The number of clock counts per tachometer
line is then divided into the constant "3,840,000" to arrive i~
at a reference velocity unit. '~
In the milling machine example described above, assume
that the table is moving at 15 inches par minute. This i5
.25 inches per second. Since each tachometer line represents
; 125 millionths of an inch (4 x 31.25) this rapresents 2,000
~!: 20 tachometer lines per second, or .5 milliseconds per tachometer
line. Using the 2 mHz clock the CTC will then accumulate `~-
1,000 counts during the period of one tachometer line. By
dividing this 1,000 counts into the constant 3,840,000 the ~-
reference velocity of 3,840 units is obtained. This number t
(3,840) is the computer command used to ohtain an output
velocity of 15 inches/minute. To simplify computer program
function, all velocity commands are multiplied by two to the `
~ighth power ~256). Therefore, 15 inches per minute becomes t
15 x 256 or 3,840 units in computer compatible notation.
. '' '.:
." '', `.
':: ,,
,. ! '
;. .
~' `.
~$7~
In the milling machine example, it is important to
achieve motor control velocities over the range of 4 to 200
inches per minute. Using the mathematics presented, above
the processor must read velocity updates every 1.9
05 milliseconds at 4 inches per minute and every 37
milliseconds at 200 inches per minute. The velocity count
read occurs in foreground which is executed each 200
microseconds. Therefore, it is not possible to read
velocity counts updates which occur more frequently than
every 200 microseconds. To solve this problem a tachometer
line divider, LS393, (U15) and multiplexer, LS151 (U14), are
used to select variable tachometer periods which are
compatible with the non-maskable interrupt. Control lines
"MUXA", "MUXB" and "MUXC" are used to select multiplexer
output for the "NAND" gate.
The servo amplifier uses the pulse width modulated
(PWM) mode of control. Under PWM control, current is
continuously reversed across the servo motor leads. The
proportionate amounts of time in each direction determine
the net power delivered to the motor. When the switching
times are equal, the motor receives no net power and
therefore it does not move. This mode of control provides
for minimum power loss in the amplifier, minimum component
cost and minimum package size.
The servo amplifier board, Fig. 19, incorporates a
number of safeguards to protect the motor and the amplifier.
The PWM signal from the motor control board is monitored by
a "watch dog" circuit which will automatically shut down the
amplifier if the PWM pulse train does not exhibit certain
predetermined parameters. A second protective circuit
monitors the current flowing through the motor and will turn
41
; j ~L 2 ~? 7 ~3 ~ j 7 ~
the amplifier off if current exceeds parameters determined
by a specified current/time map. A unique lnductive loading
technique has been built lnto the amplifier which prevents ~;the common problems of "H" bridge short clrcuit due to
05 phased switching delays. The inductors al30 prevent the i~
destructive effects of amplifier ringing due to interàction
between parasitic motor lead capacitance and motor inductance. ~ .
The current flow through a motor which is not turning
llocked rotor) may be related llnearly to the percentage of
~, 10 pulse width. For a given motor a pulse balance of 35% ln ~ '
one direction and 65~ in the opposite direction may cause
full rated torque to be developed by the motor. When this ?~same motor is turning at ~ull rated speed it may require a ~
pulse balance of 20% over ao~ just to maintain motor speed ~ ;
~ 15 wlth no current flowing ~n the armature. To compen~ate f~rI this effect, the ml~roproc~ssor wi~l ~orrect for motor
velocity ~back emf~ by using correction formulas built into
the microprocessor program. This i~ done by calculating the
pulse percentage required to turn the motor at its actual
-` 20 velocity and then adding (or subtracting) the current
balance to lfrom) the velocity percentage. r,~,,~,,
The supply voltage used to control the motor must be
sufficient to provide for motor back "emf" (electromotive ,
1 25 force~ at maximum motor velocity. In the milling machine
;~ motor example, the motor must turn at 3200 RPM to achieve ~;
200 inches per minute. The motor back emf produces 27 volts
per thousand RPM, or 86 volts at 3200 RPM. To achieve
adequate motor control the supply voltage is set at 160 VDC.
This supply voltage ls generated by full wave rectifying the
115 VAC powcr mains.
. ~.
'' ~i.
42
3S~
The PW~ mode of contxol requires an inductive load to
regulate current flow through the power transistors and
motor. At ~he nul1 condition the motor will have current
flowing in a first direction for lO0 uSec and then in the
05 reverse direction for lO0 uSec. A typical motor inductance,
for a DC servo motor, which may be used in the milling
; machine example cited above, would be 5 millihenries. If ~-
160 volts is placed across 5 mll for .1 milliseconds the
current in the inductor will increase by 3.2 amperes. Th$s
is an unacceptably large amount of curreht flow for the null
condition, resulting in excessive motor heating, oversize ~".
drive transistors, power loss and excessive audio noise in
~- the control system. To limit these unaesirable effects, the
tor inductance is supplemented by external cho~e~ to ~ l
increase circuit inductance to approximately 15 mH. This ~t.
three times increase in inductlve loading will reduce the
null current change fxom 3.2 amps to l amp, a more accepta~le ~
value. ~;
one of the most difficult performance paxameters to
achieve in a position control servo system is to maintain
the motor in a given position with no motion, the "stop
lock" mode of operation. This parameter is aggravated if
~` the mechanical load includes a great deal of friction, as in
; 25 the milling machine example. rhe DC servo motor has inherent
damping characteri~tics which may be invoked if the motor
circuit is closed. If the motor leads are "open" this
damping ls lost. As explained above, the PWM control
technique used, in the uCNC system, maintains a constant
flow of currellt through the motor. The uCNC system therefore
, ~`.,,.. :
~3 ~:~
~L~6'795~ s;
takes advantage of the natural damping charact~ristics of ,:
P
the DC servo motor.
The placement of the compensating inductors, in the
uCNC system, is highly unique~ The desired inductance i9
05 distributed around four chokes stlch that either of two pairs ï-~
provides the desired inductive compensation. As shown in
the servo amplifier schematic, Fig. 19, these chokes are
placed in each leg of the "H" style motor drive a~plifier. ~j-; In operation, motor current flows first from the 160 sp;
- 10 VDC source through power transistor Q4, choke L1, the servomotor (positive direction), choke L3, and finally through
~; power transistor Q6 to the 160 VDC return. In the second
half of the cycle current flows from the 160 VDC source jP
through power transistor Q7, choke 1.4, the servo motor
(negative directlon), choke L2, and then through power
transistor Q5 to the 160 VDC return. Note that for each i;
path the inductive load is 15 mll$ 5 m~l for the servo motor
; and 5 mll for each of the two chokes. As the cycle reverses
the motor current reverses providing a net null current, and
torque, within the motor. However, because current is
always flowing through the motor, the natural damping l'J' '
characteristics of the motor are invoked. ~;
.~ "
The current flow through chokes Ll, L2, L3 and L4 does
not reverie each half cycle. l'he current induced into Ll,
during tlle first half of the cycle, continues to flow during
the second half of the cycle (this is a property of an
inductor). The source of this current is the 160 VDC
; source in the first half of the cycle. During the secondhalf of the cycle the source of current is the 160 VDC ~ `return, via diode D2. The same argument applies for the
44 ~'`
! ~2~7~3~
: j .
inductor L2. Ill the case of L2 the current pas3es through
power transistor Q5 for the second half of the cycle and
through diode D3 during the first half of the cycle. If any
junction Setween a power transistor and diode pair were
05 observed, Q4 and D2 for example, the voltage would first
commutate from the 160 VDC source and then to the 160 VDC
return. To facilitate thls commutation the diodes feature
very fast reverse recovery characteristics. This means that
when the voltage is reverse across the diode, such as to ~ ~;
swing from blocking to conducting, the diode will recover ii
from the reverse blocking condition in a fraction of a
mlcrosecond. ,!~The percentage of time for which each side of the
amplifier is conducting is determined by the PWM control
command from the motor control board. This signAl passes
through a high band pass opto-isolator and is processed by
CM05 logic on the servo amplifier board. Power for the CMOS ;~,. `
logic circuitry is regulated on the servo amplifier board by
a 78L12, +12 volt, regulator ~Q1). The raw power supplied
to this regulator passes through a high voltage diode ~Dl). ;~
Should a power transistor failure occur on the servo amplifier
board it is possible that the 160 VDC will appear on the 12
`; volt supply. The purpose of the high voltage is to block t~ .
this failure from causing failures on other servo amplifier
.~ boards. 1
The PWM signal is first processed by a series of four
"4093~ "NAND'i gates (U4). These gates have Schmitt trigger
inputs to provide high noise immuinity for the PW~ signal. i
~- 30 The first gate 50utput 3) provides a buffer for the signal t
coming from the opto-isolator. The second gate ~output 4
.', 'i;
' 1,
` '
~7~
provides for one stage of inversion. This inversion is
required because complementary signals are reguired to run
the opposite sides of the "H" bridge a~nplifier. The third
gate (output 10~ and fourth gate ~output 11) provide for the
05 ability to shut down the servo amplifier in response to
detected failure conditionq. When pins 9 and 12 are pulled
low the PWM signal is inhibited from turning on either side
of the RH" bridge. ~
The output of the third "NAND" gate ~pin 10) i9 used to ~ .
control two "4049" gate drivers (U3). These gate drivers ~ -
control the flow of current through the positive slde of the ~ ,
"H" bridge. Input at gate driver pin 5 generates the gate ~
control signal required for power transistor Q6. When the ~ -
gate of Q6 is high the transistor /a n-channel power MOSFET~
will be conducting. Input at gate driver pin 7 generates
the controls signal requlred to transistor driver Q2. When ~'the gate of Q2 is high the transistor ta n-channel MOSFET)
will be conducting. This will cause the gate of transistor
~, Q4 to drop 10 volts below the 160 V~C supply. The voltage
at the gate is limited by zenier diode D10. This drop in ~
gate voltage will cause Q4 ~a p-channel ~OSFET transistor) '~s
to conduct. With both Q6 and Q4 collducting cur~ent will be ~t~
induced to flow through the servo motor in the positive ~ ;
~' 2S direction. During this time transistors Q5 and Q7 are held ~ ~;
in their non-conducting states. i~
The output of the fourth "NAND" gate (pin 11) i9 used
to control two "4049" gate drivers. ~hese gate drivers ~,~
control the flow of current through the negative side of the '~; ;
UH" bridge. Input at gate driver pin 3 qenerates the gate
control signal required for power transistor Q5. When the
. . .,.:
., ~, .
46 ~
, ~ , ",, ,~ ,7, ~ ,?~" ~
79~t7
gate of Q5 is high the transistor (a n-channel power MOSFET)
will be conducting. Input at ~ate driver pin 14 generates
the control signal required to transistor driver Q3. When
the gate of Q3 is high the transistor (a n-channel MOSFET)
05 will be conducting. This will cause the gate of transistor
Q7 to drop 10 volts below the 160 VDC supply. The voltage
at the gate is limited by zener diode Dll. This drop in
gate voltage will cause Q7 (a p-channel MOSF~T transistor)
to conduct. With both Q5 and Q7 conducting, current will be
induced to flow through the servo motor in the negative
direction. During this time transistors Q4 and Q6 are held
in their non-conducting states.
Overtravel limit switches are provided to limit travel
in each direction of motion. These limit switches interrupt
current flow to the servo motor such as to prevent
mechanical damage to the machine as a result of excessive
motion past predetermined limits. The positive motion limit
switch is connected at J3-4, 5 and 6. In normal operation
this switch is closed across contacts J3-4 and 5 allowing
motor current to flow unrestricted. If the positive limit
switch is actuated, as a result of excessive motion in the
positive direction, this switch will open the connection at
J3-4 and 5 and close the connection at J3-4 and 6. Current
steering diode D6 will block the flow of current in the
positive direction. At the same time current steering diode
D8 will allow motor current to, in the negative direction,
between the servo motor and braking resistor R10, thus
dynamically braking the servo motor using the motor's
internally generated back emf to generate the braking
energy.
47
:a26795'~
The negative motion limit switch is connected at J3-1,
2 and 3. In normal operation this switch is closed across
contacts J3-1 and 2 allowing motor current to flow
unrestricted. If the negative switch is actuated, as a
05 result of excessive motion in the negative direction, this
switch will open the connection at J3-1L and 2 and close the
connection at J3-1 and 3. Current steering diode D7 will
block the flow of current in the negative direction. At the
same time current steering diode D9 will allow motor current
to flow, in the positive direction, between the servo motor
and braking resistor R10, thus dynamically braking the servo
motor using the motor's internally generated back emf to
generate the braking energy.
To assure continued operation of the servo amplifier
the pulse train into the amplifier must be continuous and at
the correct frequency. If the pulse train were to stop
(through a fault in the microprocessor, it's program, or
component failure) a watch-dog circuit will shut down the
amplifier. The PWM signal is fed into a mono-stable mono-
vibrator 4528 ~U2) at pin 4. The time constant for the
multivibrator is set to 3 milliseconds. If output of the
multivibrator (pin 6) will be held high as long as the PWM
signal changes state at least every three milliseconds. If
the PWM signal should fail to have two raislng edges within
3 milliseconds, the output of the multivibrator will drop
and inhibit the PWM signal from passing the "NAND" gates at
U4 pins 10 and 11. The polarity of outputs at pins lO and
11 is such that both sides of the "~" bridge will be disabled.
A current sense resistor, R7, is used to sense current
flow conditions in the servo motor. The currel-t in the
48
'`, i:"`.
; sense resistor is the current~flowing through the servo
motor only. The current flowing through the inductors on '~
the non-co~ducting side of the "H`' bridge i9 sourced through
clamp diode D2 l or D5) and sunk through clamp diode D4 (or
05 D3). Thus thc free wheeling current is not included in the '~
current measured at the sense re~istor R70 The voltage
sensed at R7 is amplified and filtered by operational
amplifier U1 ~pins 5, 6 and 7). There are two RC filters
for this amplifier. The input filter ~C111) provides a 4 ~ i
millisecond time constant on the amplifier input. The ~`
feedback filter (Cl and Rl) provides a 4 millisecond time
~; constant for the feedback. A;.l
The amplifier gain is determined by the eedback ~di
resistor R1 and the ground reference resistor R2. The ga:in
of the amplifier stage i9 ~R1 + R2)/R2 or 7.3. The amplifier ~Y;;-
input is .2 volts/amp giving an overall gain of 1.46 volts/amp.
If the servo motor current llmit is set to 4 amps, then the ~/~?
desired set point for the comparator (formed by operational
amplifier Vl, pins 1, 2 and 3) is 5.84 volts. This set
point is determined by the resistor pair R4 and R3. The ~;
output at pin 1 is fed into a mono-stable multivibrator ~U2)
at pin 12. ~ falling edge at pin 1~ will cause the multivi~
brator to produce a 3 millisecond Idetermined by R6 and ~'
C113) low pulse at pin 9. This low pul3e is used to inhibit
~clear) the watch-dog multivibrator and thu3 disable the "H"
bridge for 3 milliseconds.
.:
Coordinated moves required that two axi3 move at the
same time. The uCNC 3ystem performs coordinated moves by
i3suing pairs of position points at a rate of 100 pairs per
~`
49 ~;.i
~7~-7
second. The servo amplifier and motor serve to smooth these
discrete points such as to produce a smooth output motion.
The servo control svftware is written in assembly
lan~uage, translated to machine language, and stored on the
05 motor control board using a standard 2764 ~8k x 8) EPROM.
The servo control software controls the position and
velocity of the servo motor, communicates bi-directionally
with the controller, performs internal diagnostics and
self-calibration function. There is one copy of software
installed on each of the three motor logic boards. Each
axis of the machine requires a different version of the
software to compensate for differences in mechanical drive
configuration and certain load inbalances. These
differences are down loaded Erom the controller at
initialization time.
The flow charts for the servo control software are
illustrated in the drawing. Fig. 20 shows the overall
structure of the program. First, upon power up, the program
will execute certain initialization routines. After
initialization, the processor will execute three programs.
concurrently. When the processor receives a non-maskable
interrupt, every 200 microseconds, the foreground tasks will
be executed. Between interrupt cycles the processor will
execute either the midground or the background routines.
The background program consists of several segments which
are alternately executed between midground routines. This
alteration between midground and individual background tasks
ensures that the midground routines will execute every three
to four milliseconds. The background routines are adjusted
to equal amounts of execution time so that the midground
will repeat on regular intervals.
Elig. 21 of the flow charts illustrates the
initialization tasks. The first task performed in RAM
initialization. During this task the read/write memory is
preset to contain constants and initial values required for
05 execution of the remainder of the program. Unused RAM,
Random Access Memory, area is initialized to zero value.
The stack pointer is also initialized at this time. The
second task is to initialize the CTC's (Counter Timer Chips)
to operate as described above. The CTC's are initialized by
~o writing commands into control registers associated with each
channel of each CTC. Third, the initial default servo
parameters are loaded into RAM re~isters used to control the
d~namics of servo control. It is these registers which give
the servo unique knowledge of a specific control axis.
1ater during the initialization, these registers will be
over written by the controller to refine the control
capabilities of the servo. Fourth, the 8251 UART serial
communications port is initialized to communicate with the
controller using the protocol specified in the
communications specification. Finally, the background task
scheduler is programmed to indicate the order in which
background tasks are to be performed.
Fig. 22 of the flow charts illustrates the foreground
~asks to be performed by the processor. As described above,
the processor is interrupted every 200 microseconds by a
timer which actuates the NMI ~non-maskable interrupt) pin.
These interrupts, every 200 microseconds, are used to
generate the 5,000 cycle per second PWM signal which gates
current to the servo motor. The first task performed, when
interrupted, is to save the microprocessor internal
registers
7~5~
'` _ ,'
and the address of the next program step to be executed. )r~/
Second, the processor update~ the PWM duty cy~le by clearing r;
the PWM counter, loading the ne~ count value, and restarting
the PWM counter.
05 The third interrupt task is to read the velocity !~
counter. The v,alocity counter is a sixteen bit counter
which is clocked by the phase encoder signal as described ,~
above. frhe velocity count must b~P read during the time in '~
which the clock is inhibited by the encoder signai~ The
processor may determine if the counters can be read by
.
testing the encoder ~ignal. If the line is low, the counters ~j
~j may be read. IE the llne is high, the counters are active ~ ,
and may not be read. After the counter ig read, the encoder
slgnal i3 re-read to ensure that the counter d$d not start ~'
~' 15 during the read phase. (The encoder signal is actually ~ `
multiplexed as described above.) The velocity counter is
read in two eight bit reads. These reads are assembled into
a sixteen bit register within the processor. The previous ~:
velocity counter is fetched from memory (RA~) and subtracted
from the new reading. If the difference is the same, the ~ ~'
~I velocity counter has not been updated since the last reading. ~`
In this case, the reading is ignored. If the difference is
non-zero then the count has been updated and the difference
now represents the numb~Pr of clock cycIes between tachometer
~, llnes. fl"'`,~
The third, and final task of the interrupt program is
to process communications from or to the controller. As the
UART is set to operate at 9600 baud, it is possible to ~ -
-( 30 transmit or receive one character each 1.1 milllsecond3. To
ensure th~t I~O ch~r~q,tera are Io8t, the lnterrupt r~utin~
i 52
may either transmit or receive a character every four
interrupts (.~ milliseconds). The communlcations routines
are therefore divided into four sections, one section is
executed every interrupt time. In the first section the
05 receive channel of the UART is tested to determine if a
character has been received~ If a character has been
received, the character is loaded into the receive buffer in
the RAM. If the character is the last character of a
message string, the error flags of the UART are tested for
transmission errors which may have occurred within the
message. If any errors have been detected by the UART,
(parity, overrun, underrun, etc.) the error condition is
reported to background so that the message may be re~ected.
After each character is received the receive buffer pointer
is advanced. Next the pointer is tested for excessive
message length. The next section of the communications is a
null; no program is executed (in order to conserve
processing time). The third section provides for character
transmission. The program first tests to determine if there
is a character to be transmitted in the transmit buffer (in
RAM). If there is a character, the UART transmitter is next
tested to determine if it is empty. If the transmitter is
empty, the character is loaded into the transmitter and the
buffer character pointer is advanced. The fourth, and
final, task is another null.
The entire interrupt routine has been carefully
designed to ensure that the interrupt processing time is
less than 100 microseconds. This ensures that fifty percent
of the processing time is available for midground and
background software.
Fig. 23 of the flow charts illustrates the midground
tasks to be performed. Mid~round tasks are associated with
the velocity servo and must be processed every three to four
milliseconds to ensure adequate phase margin for the
S velocity servo (mechanical time constant approximately
twelve milliseconas). First, the feedrate override flag is
tested to determine if a feedrate (commanded velocity)
correction has been received. If a feedrate correction has
been received, then a recalibration flag is set. Otherwise,
the recalibration flag is not altered. Next, the current,
actual velocity is calculated. Velocity is calculated by
dividing the velocity count into a constant as described
above. At this point a decision must be made. If the servo
has been commanded to stop lock (to stand still) the program
will branch to the stop lock calculations. If the servo has
been commanded to maintain a velocity, the program will
branch to the velocity servo calculations.
If the servo has been commanded to the velocity mode,
the next decision to be made is to determine if the
commanded velocity is below eight inches per second (Eor
example). If the velocity is below a preset threshold, then
the "slow move" velocity servo calculations are performed.
The result of these calculations is a new PWM which is
loaded into the PWM register in RAM for use by the next
~5 interrupt. If the velocity is above the "slow move"
threshold, the multiplexer shift sequence must be tested.
There are four overlapping velocity windows for the velocity
count multiplexer. The multiplexer ad~usting software
ensures that the correct velocity counting range is selected
at all times. The final
54
step for the velocity servo i9 to calculate the correct PS1M
to maintain commanded velocity. I
The correct pulse width i5 calculated from the following f``,
- OS formula:
PWM = ICv) x IVc) ~ PWM0 + ICg) x ~Vc-Va)
where: ~j ~
- $ ;
PWM is the new pulse width required
~`:; Cv is the slope of the veslocity gradient
Vc i5 the command v~locity
r PWMO is the P~YM required to turn the motor ?,
Cg is the velocity servo gain ;
~;~ 15 Va i9 the actual velocity
' ' ' S?
;~,.i ':
The first term lCv x Vc) in the above equation accounts
for the counter electromotive force (back emf) generated by
-s 20 the motor. This term represents the PWM required to turn
the motor at the commanded speed. The PWM0 term represents
the PWM required to just turn the motor in the commanded
direction. The PWM0 ~alue compensates for friction in the i!
:,
--~`; motor, belt~drive, ball screw, and load in addition to bias ~
?~ ' 25 loading required for load inbalance, such as gravitational ,-
load due to a vertical ball scre~ confi~uration IZ axis for
example). The sum of these two terms yields the PW~ required
to turn an unloaded motor at the desired speed with neither
acceleration or deceleration.
~ I
" . ':
,, . I
~ !.
The last term is the servo gain. This term specifies
how much additional PWM is required to correct for errors
between the commanded velocity and the actual velocityO The
servo gain ls Cg. Th~ gain is set as high as possible
05 consistent with the stable servo characteristics.
The values of Cv, PWM0 and Cg are customized for each
direction of motion and for each axis of the servo. The
summation, PWM, is placed into the PWM register in RAM for
use on the next interrupt.
If the servo had been commanded to the stop lock mode
of operation, the midground will branch to the position flow
chart as shown on Fig. 23. Upon entering the position logic
the program will branch depending upon whether the motor is
actually stopped or is decelerating to a stopped condition.
If the motor is still decelerating, the program will branch
to routines which control the desired deceleration ramps.
These ramps are velocity dependent and vary for each
direction of motion and for each axis of operation.
If the motor is at (or near) the stop lock position,
the program will calculate the required PWM from the
formula:
PWM = PWM0 + (Cp) x (De) + (Cd) x (Va)
where:
PWM is the desired pulse width
PWM0 is the PWM required to turn the motor
Cp is the position servo gain
De is the distance error
Cd is the coefficient of velocity damping
Va is the actual velocity
56
The first term is the PWM required to just turn the
motor in the desired direction. The second term is the PWM
which is to be applied to correct for an error in the
position of the motor. The thi~d term is the degenerative
05 velocity correction required to prevent overcorrection
resulting from the second term (the position correction term
introduces a 180 phase lag ana is therefore unstable
without the degenerative velocity correction). The computed
PWM value is stored in the PWM register for use in the next
interrupt routine. The midground routines required approxi-
mately two and one half milliseconds to execute (including
oreground time). This allows for 60/40 sharing of
processor time between midgrourld and backgrouna software.
Fig. 24 of the flow charts summarizes the function of
the servo control background software. Figs. 27 through 32
specify the specific tasks performed. The backgrouna is
broken into six specific johs. Each job is executed alter-
nately with the PWM generating software described in the
midground section above. Each background job has been
adjusted such that the midgrouna will executed every three
to four milliseconds.
Fig. 25 of the flow charts illustrate the first back-
ground routine. If the servo has been commanded to a new
position this routine will calculate the information
required to accomplish the move. The routine will determine
and set the new position target and the velocity commanded
for the move. The routine will determine and set the new
position target and the velocity commanded for the move.
The routine will next compute the position error at which
deceleration will commence, to ensure that the motor does
not overshoot
57
the target position. Finally, the direction of movement is
determined and the initial PWM is output to the motors to
initiate motion. If a new position has not been commanded
~he routine will idle for a compensatory length of time.
05 The second background task is illustrated on Fig. 26 of
the flow charts. This routine is executed in response to a
"Motion Hold" command. First, the software will compute the
correct deceleration ramp for the velocity at which the
motor is moving. ~ext the routine computes the closest stop
lock position based upon the deceleration ramp. Finally,
the stop lock flag is set and the deceleration is initiated.
The fourth task, Fig. 27, is the communications
background task. This routine processes the messages
assembled by the foreground receive routine. The background
also builds the message strings which are subsequently
transmitted by the foreground transmit routine. First, the
receive message is tested for errors. Three tests are
performed. The first test is to examine the UART error
flags for possible parity, overrun or underrun error. The
second test is to determine if the message has a correct
check sum (as defined in the communications specification).
If these tests are met t the message address is examined to
determine if the message is directed to this board (board
addresses are set by ~umpers for the servo boards). If the
message is addressed to this board, the message parameters
are separated from the message and loaded into predetermined
buffer locations. The contents o F these buffers set the
mode of operation for the servo.
If the message was addressed to this board, the board
is required to make a response as specified in the communi-
cations specification. The message is assembled and loaded
58
~ t7
into the transmit buffer. The foreground transmit routine
will send the message, character by character.
The fourth background job is described on Fig. 28 of
the flow charts. This is the "Auto Zero" routine. This
05 routine may be activated upon command rrom the controller.
When activated the routine will update the null PWM for the
servo amplifier. This is a self calibration routine
required to compensate for phase lags and inductor inbalance
on the servo amplifier board.
The fifth background job, Fig. 29, processes the
coordinated move commands. The coordinated moves commands
are preceded by a "Start Coordinated Moves" command. This
initial command specifies the offset and gains values to be
applied to each of the individual coordinated move steps
(refer to the communications specification for detailed
description of these command structures). If a coordinated
move is in progress, this routine performs the mathematical
functions required to process the individual step targets.
The final background task, Fig. 30, processes the
feedrate override command. This command allows the servo to
change commanded velocity while a move is in progress. This
is usually in response to an operator specifying manual
correction to a move in progress. The in progress velocity
change requires that th~ software compute a new deceleration
point based upon the new velocity command and the
pre-existing position target. The feedrate override routine
then determines the correct multiplexer selection for the
new command, and sends this correction to the multiplexer.
Finally, the routine updates the velocity target registers.
59
