AXIS Substitution problems...

[Shoggoth_2150]

I'm trying to do something very simple and getting nowhere. I want to put a constant radius on the edge of a hole that goes through a piece of round stock. I've done things like this with 3D models, but can't find a solution with this simple sketch up. Should just be able to sub the Y axis and get what I want, but it isn't cutting a smooth line. Instead it appears to be piecing together a bunch of straight lines instead of arcs, and it looks like junk. How can I get a smooth rad on that edge?? Do I have some settings wrong? Any help welcome. File attached.

Edge Rad.mcam

[#Rekd™]

Create a curve on the edge you are trying to put the rad on. Use contour without axis substitution, it should recognize the 3D curve if not change the contour type to 3D. You might have to offset the curve to suit your rad mill. I am assuming you are using a 3 axis mill going by your file.

 

 

 

[cncworker]

You may have to make a radius on the edge and use a ball endmill. 

I attached what i did.

Edge Rad.mcam

[huskermcdoogle]

3 minutes ago, #Rekd™ said:

I am assuming you are using a 3 axis mill going by your file.

I didn't see it that way...  I think he wants to use the rotary.

Instead of chaining a flat arc, you need to chain the intersection (solid geo) then check unroll geometry with a tight tolerance in the operation axis sub parameters.  Doing it this way gives a much finer motion for the axis sub.  Otherwise, the way you had it, will default to the tolerances in the post I think to break up the arc into an axis sub situation.  With my method its all point to point and the post will only map the coordinates, it won't be breaking it up.  Mind you, you should be close to tangent on the top edge, but the inside won't be tangent.  No way to get both with that tool.  if you want a good tangent fillet, you will want to use a ball end and surface it.

Edge Rad 2.mcam

[Shoggoth_2150]

6 minutes ago, huskermcdoogle said:

I didn't see it that way...  I think he wants to use the rotary.

Instead of chaining a flat arc, you need to chain the intersection (solid geo) then check unroll geometry with a tight tolerance in the operation axis sub parameters.  Doing it this way gives a much finer motion for the axis sub.  Otherwise, the way you had it, will default to the tolerances in the post I think to break up the arc into an axis sub situation.  With my method its all point to point and the post will only map the coordinates, it won't be breaking it up.  Mind you, you should be close to tangent on the top edge, but the inside won't be tangent.  No way to get both with that tool.  if you want a good tangent fillet, you will want to use a ball end and surface it.

Edge Rad 2.mcam

This is exactly what I was looking for. Thanks. When I was playing with UNROLL it put the toolpath out in space. Didn't think it was the right setting to make it work, so I didn't bother toying with it much. Tightening it up really helped. Now I know. Thanks.

[Jespertech]

7 minutes ago, huskermcdoogle said:

I didn't see it that way...  I think he wants to use the rotary.

Instead of chaining a flat arc, you need to chain the intersection (solid geo) then check unroll geometry with a tight tolerance in the operation axis sub parameters.  Doing it this way gives a much finer motion for the axis sub.  Otherwise, the way you had it, will default to the tolerances in the post I think to break up the arc into an axis sub situation.  With my method its all point to point and the post will only map the coordinates, it won't be breaking it up.  Mind you, you should be close to tangent on the top edge, but the inside won't be tangent.  No way to get both with that tool.  if you want a good tangent fillet, you will want to use a ball end and surface it.

Edge Rad 2.mcam

I was working this up myself and was just about done but I ran into an issue with the axis substitution. Then I read your comment and selected the "unroll" box and Voila! a beautiful thing 

[crazy^millman]

31 minutes ago, Shoggoth_2150 said:

This is exactly what I was looking for. Thanks. When I was playing with UNROLL it put the toolpath out in space. Didn't think it was the right setting to make it work, so I didn't bother toying with it much. Tightening it up really helped. Now I know. Thanks.

I am not trying to be mean or ugly in this response so please understand that when reading what I have just typed.

Exactly in that fact you are not looking for a correctly machined feature? Backplot and Verify the toolpath and you will see the tool doesn't keep the radius exactly perfect. Want to see what I am talking about put a .0625R on the edge and then backplot the toolpath. If you are shooting for a perfect .0625R at that edge doing what your trying will not accomplish a perfect part. It will be good at 2 points, but then fade away at the other 2 point furthest away from the tangent intersection. Can be done with this tool, but would really need to be done with Curve 5 axis like in this example, but not perfect.

5th Axis Edge Rad

I added morph between 2 curves cutting the correctly modeled .0625 radius on that edge. Yes it changes in size because that is the natural shape of the feature when correctly modeled and machined.

[pro grammer]

Reducing your tolerance will give you a better resolution. Like others have said, I usually 3d these with a flow path to avoid the headaches of axis substitutions.

[Shoggoth_2150]

It all worked out for me, except for one odd issue. On the computer it was climb milling, on the machine it was conventional. Any ideas why that would be?

[JoshC]

16 minutes ago, Shoggoth_2150 said:

It all worked out for me, except for one odd issue. On the computer it was climb milling, on the machine it was conventional. Any ideas why that would be?

I would try changing the direction of the rotary from CCW to CW or vice versa, my guess is your rotary direction for Clockwise/Counter clockwise is Bass Ackwards 

 

[huskermcdoogle]

I had that as a problem before with a 5 axis post and using axis sub toolpaths.  Never found out what the solution was as I didn't own the post and was locked out of most of it.  I will dig into it when I get a chance.