What's new
What's new

Elliptical Toolpath Problems

coyotekid

Cast Iron
Joined
Apr 20, 2008
Location
Montana
I'm having trouble getting a decent surface finish when milling an elliptical pocket using a "Super" Mini Mill.

The side wall of the pocket looks faceted instead of being smooth. The same machine cuts circular pockets of all sizes just fine.

We've had the same problem with other Haas mills in the past, but changing parameter 191 to the finest finish has fixed the problem. On this machine (2005 vintage) though, I can't find a parameter 191 at all.

I've set my CAM software to the tightest tolerance and slowed the feed way down, but neither seems to help much.

Any ideas?
 
I've set my CAM software to the tightest tolerance and slowed the feed way down, but neither seems to help much.

Seems like you could look at the G-code from your CAM software to see how fine the steps are in the ellipse and compare with what you're seeing on the part. Find the code for ellipse and choose a point, subtract x and y values from the preceding point, square each, add, take the square root and you'll have the length of the segment. Another quick check of the CAM software is that finer resolution should yield a longer program unless there's some sort of subroutine for making the steps in the G-code.
 
You could also try changing the size of the E/M or cutter you're using, if possible.
A smaller cutter may give you a better finish or maybe a larger one less deflection.
Try one or the other and see what results you get.
Always trust your eyes and ears more than the computer for feeds and speeds.
Hard to give better advice without knowing material, cutter, feeds and speeds, finish allowance, size of piece, etc.
 
I really appreciate all the help!

I got the problem solved, and the pockets have what I'd call nice smooth sidewalls. (As good as they ever are on a Haas...:D)

It turns out that it wasn't really a machine problem so much as a CAD issue. I was extracting the geometry from the model at "0" tolerance thinking I was getting a smoother ellipse. While that may be technically true, I got the best actual parts when I used a 0.0001" tolerance to extract the geometry. This gave me a series of points and short arcs instead of the "smooth" geometry. When I used a contour toolpath then, the output was a series of I and J code and the actual pockets turned out smooth.

It seems a little counter-intuitive to me... What do you guys think?
 
Sounds about right to me!
I finally succumbed to the 21st century and learned Mastercam 10 years ago.
I still write most programs longhand and have never gotten a CAM program that I couldn't make run faster and better, especially entry and exit points, circular and arcs.
Maybe I'm just prejudiced, but I still think the best computer is the one on top of your shoulders.
 
medicut??

Has anyone heard of a software that takes a 3-D program that has segments for it arcs and transforms the segments into I and J arcs. I used a program that did something like that years ago when I worked for a different company.
It was called medicut or something like that. Has anyone used this or something like it?
This discussion made think of it but I could not remember the name of it. Google turned up nothing.

Thanks,
A.J.
 
Has anyone heard of a software that takes a 3-D program that has segments for it arcs and transforms the segments into I and J arcs.....Thanks, A.J.

I don't know the program name but I think this is called toolpath filtering and it is included in recent versions of Mastercam. Haas had good description of this stuff in their CNC Machining magazine a few years back.

This should take you to it: CNC - Haas Automation, Inc. - Solutions-Applications 3D
 
Energy Rebel,
I used to think that way also, but have to say, if you put the energy into getting the post absolutely correct, and learn all the tricks in Mastercam to get it to do exactly the way you would do it if coded by hand, you will change your mind. Now if you only do very simple stuff and modify an existing program to do it, maybe not. I agree with the statement about the best computer. But if you embrace it as a tool and learn to use it thouroughly and corrrectly, it's an excellent tool. I've learned not to use never or always. RJT
 
Has anyone heard of a software that takes a 3-D program that has segments for it arcs and transforms the segments into I and J arcs. I used a program that did something like that years ago when I worked for a different company.
It was called medicut or something like that. Has anyone used this or something like it?
This discussion made think of it but I could not remember the name of it. Google turned up nothing.

Thanks,
A.J.

AJ

Do not know about the 3D features, but in the 2D FeatureCAM it is called Spline Tolerance and is defined in deviation increments.
What it does is that it analyzes splines and attempts to fit tangent arcs in place of the spline itself. I have tested it before on an ellipse and a steep parabolic curve with a spline tolerance of .002, in both cases the resulting G-code had not a single linear segment.
 
Energy Rebel,
I used to think that way also, but have to say, if you put the energy into getting the post absolutely correct, and learn all the tricks in Mastercam to get it to do exactly the way you would do it if coded by hand, you will change your mind. Now if you only do very simple stuff and modify an existing program to do it, maybe not. I agree with the statement about the best computer. But if you embrace it as a tool and learn to use it thouroughly and corrrectly, it's an excellent tool. I've learned not to use never or always. RJT

Point taken.
I didn't get much farther than the basics, so I haven't experimented with the "fine tuning".
I've worked in a shop or two where apparently the engineers sending us programs hadn't either.:bawling:
I was just the guy who had to make it work......:crazy:
 








 
Back
Top