:Compucut ?

Hello Ballendo,
I'll try and answer your questions without boring
your socks off, for a fuller description of the Compucut command
set see the posting to FAQ page.

>How do you command a 3-D move?

Compucut has SEVENTY commands, among these are 3Axx,yy,zz;
and 3Rxx,yy,zz; These are the absolute and relative forms of
a move to any point in three dimensional space. The movement
uses linear interpolation in three axes. The moves have no
practical limitations in magnitude or resolution. For example,
the recommended pitch of 50thou for Compucutter leadscrews
directly driven by 200 step motors gives a resolution of 1/4thou
and the system data type for 'steps' is long integer, so a single
3Rxx,yy,zz; type command can drive a theoretical machine table
literally MILES in 1/4thou steps.

>Who writes this data?

Since this data is the digital representation of YOUR
component design, based on the configuration of YOUR machine tool
and with regard to YOUR work clamping and YOUR choice of cutter
etc, YOU DO.

This can be done in several ways, the simplest being to
digitise an existing worn / broken specimen and reverse engineer
a replacement at any scale you choose. A single axis probe can
be made to digitise any specimen that could in theory be made in
a 'jelly mould' ie. has no re-entrant angles, by simply laying
the pattern on the table and logging depth measurements based on
an XY raster. This XY grid sets the resolution of the scan and
is set by the user. This routine is offered on our 'toolkit'
disk.
More complex solid shapes can be mounted on a stepper driven
rotary table whose axis is perpendicular to a second rotary
motion which carries the scanning head ( and later the cutting
head) on a linear radial Z axis. This arrangement is similar to
the school type globe of the earth, where the cursor can be
placed over any location and its lat. and long. read off. We add
to this the third dimension of depth. These data files of Z
ordinates against a grid of circular XY cordinates are saved by
YOUR data logging program as a sequence of 3Rxx,yy,zz;
statements.
This idea can be applied to modelling a persons head. If
you are knowledgeable in the field of low power measuring lasers
then a 'non contacting' transducer has obvious benefits, I am
nervous about the amateur use of lasers and would use a digital
vernier caliper with depth gauge and statistical process control
(SPC) port. Tip the probe with something soft containing the end
sensing switch whose digital status is read by one of the four
input lines on the Compucut interface. The simple Basic program
that YOU have written will store the sequence of 3-D moves
required to join the end points of each probeing stroke. Again
these will be in the form of 3Rxx,yy,zz;


New designs start in your 2-D CAD program. When deciding how
to machine a 'tricky' job, such as the centrifugal fan for a
model gas turbine, (whose profile ideally sweeps in all three
axes). Draw the component and all construction lines on Layer 1,
next identify the axis which is common, in this case the RADIAL
axis (X-axis) can be used as the independent axis against which
the AXIAL (Y-axis) and TANGENTIAL (Z-axis) data can be plotted.
Draw the cutter path with allowance for cutter radius as seen
from both axial and tangential views on drawing Layers 2 and 3.
Save this drawing and load the plotter driver for the basic HP
plotter (type 7475 or similar) Export the drawing to a .PLT file.
Display the file in the text editor and search for the SPn;
statements. These delineate the commands for each drawing layer.
You can now extract Y and Z axis data as seperate files both
with respect to X.
YOUR program (in GW-Basic if that's what you know), will
derive 3Rx,y,z; 3-D data by merging the two 2-D files.

My tip is to practice on shapes whose geometry is fairly
simple. One job that often comes up in steam loco modelling is
the smoke stack / boiler transition plate. This is cast brass in
'real' scale. The base collar is a curved disk which follows the
top horizontal centre of the boiler radius. Into the centre of
the disk fits the vertical, cylindrical smoke stack.
Because the radii of boiler, disk, stack and ball-end cutter
are constant and known, their inter-relationships are defined by
trigonometrical equations.
Assume the job will be held in a CNC dividing head whose
centreline is the centre line of the stack and this rotary axis
will be the common X-axis. If your maths and spacial imagination
are up to it, ( get the wife to help??) the data can be generated
in a spread sheet. First enter in the first column the X-axis
data, which is the number of steps for one complete revolution of
the dividing head and spread over a range that suits your
resolution. Next, operate on the X data using the parametric
equation which defines Y in terms of X. Then operate on X by the
equation which defines Z. Now press 'compute' and the data is in
a file. Import this file into the text editor to re-format to
3Rxx,yy,zz;
I have to admit that spread sheet listings are not nearly
graphic enough for me and prefer the CAD approach, but the
Compucutters who use this method as 'obvious first choice' find
it straight forward. (It probably comes down to the age of ones
brain.)
I hope this gives a 'feel' for how everyone can use 3-
axis movements to whatever level of complexity their project
dictates.
For my money the guy who uses 3Rxx,yy,speed; on his lathe,
where the speed data drives a small stepper motor connected to
the speed pot of the variable frequency control on the headstock
motor, to ensure a constant cutting speed when facing large disks
and therefore giving good swarf control (chips in the tray, not
string around the chuck) exemplifies the way in which we should
all use the commands provided. To do whatever you want them to
do.
Goodnight Ballendo, I'll talk Compunuts
another night.

Yours,
Richard