G-Code interpretation of an oval.

[Terry]

Does most g-code generation programs interpret non circular arcs and
ovals with all G01 commands? Is this the only way to do this.

Just curious.

Thanks.

[Beege]

Post by Terry
Does most g-code generation programs interpret non circular arcs and
ovals with all G01 commands? Is this the only way to do this.
Just curious.
Thanks.

Terry,

Most CNC controls read either arcs or lines, and not splines, unless you
macro the formula for the spline (even then its micro-linear) For ovals,
though, some controls have axis scaling, so that a circle or arc could be
made to look like all or part of an oval.

Mastercam has a filter option that breaks up splines into smaller circlular
arc moves (don't know about other CAM systems, but probably, too)

Others will have input...
Beege

[Terry]

Yes, my CNC program does not read splines. Bascially what is
happening, I have some small detail work I am engraving and cutting
out. And the toolpath is not following the small non-circular arcs
very well. It is squaring them off. I don't know exactly how much
effect this will have on my final piece though.

For example, if I draw an oval in autocad that is .25" x .5", and I
generate my toolpaths to contour this oval, it will not follow the oval
exact. It will square off segments of the oval using G01 commands.

Though if I make a .125" diameter circle and generate the toolpath, it
will use a G02 command and follow the circle perfectly.

I am not using splines when generating the oval. Becuase if I was, my
CNC program would have let me know before it loaded the DXF file. It
doesn't load splines.

Thanks.

Post by Beege

Post by Terry
Does most g-code generation programs interpret non circular arcs and
ovals with all G01 commands? Is this the only way to do this.
Just curious.
Thanks.

Terry,
Most CNC controls read either arcs or lines, and not splines, unless you
macro the formula for the spline (even then its micro-linear) For ovals,
though, some controls have axis scaling, so that a circle or arc could be
made to look like all or part of an oval.
Mastercam has a filter option that breaks up splines into smaller circlular
arc moves (don't know about other CAM systems, but probably, too)
Others will have input...
Beege

[Polymer Man]

Post by Terry
SNIP
For example, if I draw an oval in autocad that is .25" x .5", and I
generate my toolpaths to contour this oval, it will not follow the oval
exact. It will square off segments of the oval using G01 commands.
SNIP

For most regular controls it is either a straight or a round move (G01,
G02, G03). If your non-linear geometry is coming out faceted you need
to find a setting in your CAM system to adjust your accuracy or
tolerance.

It is pretty standard practice in the mold making industry to make most
"swoopy" moves in a series of small straight line moves. The facets are
so small they're imperceptible.

I received a new first article part a few days ago from a good local
mold maker and some of the facets were visible on the plastic part.
That is pretty unusual, but it does show up from time to time.

I set the tolerance in my CAM system to .0001" for most of the
machining I do and I seldom see visible facets. I drip feed, so program
size is no issue, and I keep finishing feed rates around 20 to 40 IPM
(I hear laughter from the gallows) so bandwidth is never an issue.

You probably just need to set your G01 moves to a finer motion in your
CAM system.

[Terry]

Thats what it looks like I need to do. But I haven't figured out where
to do that yet. I'll keep looking.

Thanks.

Post by Polymer Man

Post by Terry
SNIP
For example, if I draw an oval in autocad that is .25" x .5", and I
generate my toolpaths to contour this oval, it will not follow the oval
exact. It will square off segments of the oval using G01 commands.
SNIP

For most regular controls it is either a straight or a round move (G01,
G02, G03). If your non-linear geometry is coming out faceted you need
to find a setting in your CAM system to adjust your accuracy or
tolerance.
It is pretty standard practice in the mold making industry to make most
"swoopy" moves in a series of small straight line moves. The facets are
so small they're imperceptible.
I received a new first article part a few days ago from a good local
mold maker and some of the facets were visible on the plastic part.
That is pretty unusual, but it does show up from time to time.
I set the tolerance in my CAM system to .0001" for most of the
machining I do and I seldom see visible facets. I drip feed, so program
size is no issue, and I keep finishing feed rates around 20 to 40 IPM
(I hear laughter from the gallows) so bandwidth is never an issue.
You probably just need to set your G01 moves to a finer motion in your
CAM system.

[Terry]

I just found the setting for minimum arc length and minimum radius. I
set these correctly and everything looks real good.

Thanks for all great advice.

Post by Terry
Thats what it looks like I need to do. But I haven't figured out where
to do that yet. I'll keep looking.
Thanks.

Post by Polymer Man

Post by Terry
SNIP
For example, if I draw an oval in autocad that is .25" x .5", and I
generate my toolpaths to contour this oval, it will not follow the oval
exact. It will square off segments of the oval using G01 commands.
SNIP

For most regular controls it is either a straight or a round move (G01,
G02, G03). If your non-linear geometry is coming out faceted you need
to find a setting in your CAM system to adjust your accuracy or
tolerance.
It is pretty standard practice in the mold making industry to make most
"swoopy" moves in a series of small straight line moves. The facets are
so small they're imperceptible.
I received a new first article part a few days ago from a good local
mold maker and some of the facets were visible on the plastic part.
That is pretty unusual, but it does show up from time to time.
I set the tolerance in my CAM system to .0001" for most of the
machining I do and I seldom see visible facets. I drip feed, so program
size is no issue, and I keep finishing feed rates around 20 to 40 IPM
(I hear laughter from the gallows) so bandwidth is never an issue.
You probably just need to set your G01 moves to a finer motion in your
CAM system.

[Cliff]

Post by Terry
Yes, my CNC program does not read splines. Bascially what is
happening, I have some small detail work I am engraving and cutting
out. And the toolpath is not following the small non-circular arcs
very well. It is squaring them off. I don't know exactly how much
effect this will have on my final piece though.
For example, if I draw an oval in autocad that is .25" x .5", and I
generate my toolpaths to contour this oval, it will not follow the oval
exact. It will square off segments of the oval using G01 commands.
Though if I make a .125" diameter circle and generate the toolpath, it
will use a G02 command and follow the circle perfectly.
I am not using splines when generating the oval. Becuase if I was, my
CNC program would have let me know before it loaded the DXF file. It
doesn't load splines.

Your CAM software might have a setting for INTOL & OUTTOL
(or something else that does the same thing by another name).
Using tighter tolerances for both of these settings would give
you better results but output (much) more code.

[Cliff]

Post by Beege

Post by Terry
Does most g-code generation programs interpret non circular arcs and
ovals with all G01 commands? Is this the only way to do this.
Just curious.
Thanks.

Terry,
Most CNC controls read either arcs or lines, and not splines, unless you
macro the formula for the spline (even then its micro-linear) For ovals,
though, some controls have axis scaling, so that a circle or arc could be
made to look like all or part of an oval.
Mastercam has a filter option that breaks up splines into smaller circlular
arc moves (don't know about other CAM systems, but probably, too)
Others will have input...

Some controls can do spline interpolation and some posts can
be configured to output the proper code.

[Dave Lyon]

Post by Terry
Does most g-code generation programs interpret non circular arcs and
ovals with all G01 commands? Is this the only way to do this.
Just curious.
Thanks.

Newer high speed cnc's can handle splines. If you're stuck with an older
less capable machine, you can filter your linear moves into arcs depending
on the accuracy that you need.

[Cliff]

Post by Dave Lyon
Newer high speed cnc's can handle splines.

So could Boston Digital in 1983 or earlier.

[Steve Mackay]

Post by Cliff

Post by Dave Lyon
Newer high speed cnc's can handle splines.

So could Boston Digital in 1983 or earlier.

As could Fidia back in 1980 with their CNC/11 controls.

[Proctologically Violated©®]

You need to distinguish between an ellipse and an oval.
An oval, iiuc, is two circular arcs at either end of two parallel lines.
Thus, just g1 and g2/3 are nec. for ovals.
An ellipse follows x^2/a + y^2/b = const. Fadal has a sample macro for an
ellipse in the 3016L manual, 1998. As someone pointed out, ultimately all
interpolation is linear--fudging non-linear non-circular segments is thus a
royal pita w/o built in stuff. I spose you could blend arcs'n'lines'n'shit
to approx. an ellipse, but that's no picnic either.
I can fax you a copy of the fadal macro, or I'm sure Neal or the fadal
people can help you.
I think iffin you seen one macro, you seen'em all, in principle.
--
Mr. P.V.'d
formerly Droll Troll
Party Nominee, IPPVM
Independent Party of the Proctologically Violated®© (M)asses
"That's proly not a hemorrhoid you're feeling.... "
entropic3.14decay at optonline2.718 dot net
remove pi and e to reply--ie, all d'numbuhs

Post by Terry
Does most g-code generation programs interpret non circular arcs and
ovals with all G01 commands? Is this the only way to do this.
Just curious.
Thanks.

[Will]

On Sat, 21 Oct 2006 17:21:16 -0400, "Proctologically Violated©®"

Post by Proctologically Violated©®
You need to distinguish between an ellipse and an oval.
An oval, iiuc, is two circular arcs at either end of two parallel lines.
Thus, just g1 and g2/3 are nec. for ovals.

Not that I claim to be correct, but we always called two half circles
connected by two straight parallel lines a "racetrack oval".

You can make a 'fair' approximation of an oval / ellipse with four
arcs. No, I didn't c;aim no one would know the difference.

Stay well,
Will

[Anthony]

Post by Proctologically Violated©®
You need to distinguish between an ellipse and an oval.
An oval, iiuc, is two circular arcs at either end of two parallel
lines. Thus, just g1 and g2/3 are nec. for ovals.

I beg to differ here. The parallel lines are not required to have an oval
shape (with parallel lines, these are technically 'rounded rectangles', and
the circular arcs do not have to be circular. And it does not have to be an
elliptical shape either.
There are many different types of oval shapes.

http://en.wikipedia.org/wiki/Oval_%28geometry%29

There are many different types of ovals. There are special ovals where as
the path translates from the major axis to the minor axis (the 'normal oval
path'), there is some deviation in the oval path, only to return to the
normal oval path prior to intersection with the minor axis.

[Proctologically Violated©®]

OK!
--
Mr. P.V.'d (formerly Droll Troll)
Ever-preparing for The Grand Insertion
Party Nominee, IPPVM
Independent Party of the Proctologically Violated®© (M)asses
"That's proly not a hemorrhoid you're feeling.... "
entropic3.14decay at optonline2.718 dot net; remove pi and e to reply--ie,
all d'numbuhs

Post by Anthony

Post by Proctologically Violated©®
You need to distinguish between an ellipse and an oval.
An oval, iiuc, is two circular arcs at either end of two parallel
lines. Thus, just g1 and g2/3 are nec. for ovals.

I beg to differ here. The parallel lines are not required to have an oval
shape (with parallel lines, these are technically 'rounded rectangles', and
the circular arcs do not have to be circular. And it does not have to be an
elliptical shape either.
There are many different types of oval shapes.
http://en.wikipedia.org/wiki/Oval_%28geometry%29
There are many different types of ovals. There are special ovals where as
the path translates from the major axis to the minor axis (the 'normal oval
path'), there is some deviation in the oval path, only to return to the
normal oval path prior to intersection with the minor axis.
--
Anthony
You can't 'idiot proof' anything....every time you try, they just make
better idiots.
Remove sp to reply via email

[Anthony]

OK!

Oval shapes are one thing I have quite a bit of experience with :)

[Proctologically Violated©®]

Sounds like if you simply squish a circle, you wind up w/ some kind of oval.
:)
--
Mr. P.V.'d (formerly Droll Troll), Yonkers, NY
Ever-preparing for The Grand Insertion
Party Nominee, IPPVM
Independent Party of the Proctologically Violated®© (M)asses
"That's proly not a hemorrhoid you're feeling.... "
entropic3.14decay at optonline2.718 dot net; remove pi and e to reply--ie,
all d'numbuhs

Post by Anthony

OK!

Oval shapes are one thing I have quite a bit of experience with :)
--
Anthony
You can't 'idiot proof' anything....every time you try, they just make
better idiots.
Remove sp to reply via email

[Anthony]

Post by Proctologically Violated©®
Sounds like if you simply squish a circle, you wind up w/ some kind of oval.
:)

True, but ours aren't quite that simple, unfortunately. Most are asymetric.
Very, very few are simple oval shapes even on one half, and we have to hold
the shape to within a few microns (6, normally).