Path Profile Faces: 4th-axis integration

Here's the place for discussion related to CAM/CNC and the development of the Path module.
User avatar
freman
Posts: 1590
Joined: Tue Nov 27, 2018 10:30 pm

Re: Path Profile Faces: 4th-axis integration

Post by freman »

Hi Russ.

from my limited knowledge of the 2.5D code base it seems the problem is the lack awareness of the model. There faces are treated as wires and edges and the area is projected downwards on the the z plane and the normal is upwards.

If you want quick hack, how about a boolean property "project up ", which inverts the normal. That's quick and easy and would probably fix this situation. It would require a simple user interaction when this happens to get the correct paths.

Longer term the fix really needs to be model aware in some way to know what is the inside of the volume. So either some info from the model needs to be passed forward or something like the opencamlib calls, used in your excellent 3Dsurface tools, are needed.

This situation is a persistent problem with the 2.5D tools, as you know. Maybe it would be worth adding something to Part Design to add a normal to the geometric data indicating which side is air and which side is solid. Just a thought. Working backwards to recover this info is very computation intensive and essentially a waste of time since we have this info in the model.

It seems to be a significant thing slowing the the 3D tools too.
nyles
Posts: 3
Joined: Thu May 27, 2021 12:28 am

Re: Path Profile Faces: 4th-axis integration

Post by nyles »

Russ4262 wrote: Thu May 27, 2021 1:23 pm
nyles wrote: Thu May 27, 2021 12:36 am I'm attempting something (I thought was) simple - a hex extrusion on an A axis. The top three facets work fine when I do a path pocket, and I can see the rotations and paths fine. However, when I try one of the 'underside' facets, the action happens on the opposite side of the work piece.

Any clues to convincing path to do the facet I'm trying to work on? ...
Good morning. Welcome to the forum!
You will find many helpful members here. Taking a look at your situation, I recreated a similar extrusion in Part Design workbench along the X axis. Upon attempting to Pocket that same face highlighted in your image, I confirm the same error. To my detriment the rotational features have a few bugs. I attempted the Profile operation and it profiles the correct face, but the depth passes through the object to the opposite face.

While the rotational features available in 0.19 are buggy, they are allowing us to get useful feedback and testing of both path generation and code structure. We are working to correct and reconfigure the rotational features for the 0.20 release.

I will see about fixing these issues should there be another 0.19.x minor release.

For the time being, if the ONLY error with the incorrect pocket paths generated is the 180 degree rotation error, this is a very simple fix manually in the g-code. Unfortunately the changes entered upon using the Inspection tool are lost upon saving because a recompute is triggered. However, finding the same rotation line, A59.999998 in the post-processed file can be changed to A240.000 (or whatever error angle + 180). Making this change will correctly index the part for that operation.

At the moment, I think the simplest solution is add another rotational property, RotationOverride, that would allow the user to enter a manual rotation angle to be applied during the operation, rather than using internal analysis methods in place. There have been a number of similar situations here in the forum where rotation is 180 degrees out of alignment.

Have a great day.
Russell
Thanks Russell,

I'll go with hacking the gcode - it's exciting to see FreeCAD getting into 3+ axes!

Cheers,
Nyles
nyles
Posts: 3
Joined: Thu May 27, 2021 12:28 am

Re: Path Profile Faces: 4th-axis integration

Post by nyles »

Hi All,

Another odd thing is that a A(x) facing operation only seems to create a path on the top
surface, rather than the pocket operation adding A axis moves.

This is on the same hex model as described earlier.

One nice thing is that this one allows me to use the 3/8" endmill that I want to use in
real life, where the pocket operation doesn't allow this. However, this is ONLY when
I select another (or both) of the top 3 surfaces. Since the 3/8" bit is larger than the
single face size, this may have something to do with it.

Would the hack here be to duplicate the top face gcode path, adding A axis moves between?

TIA,
Nyles
User avatar
freman
Posts: 1590
Joined: Tue Nov 27, 2018 10:30 pm

Re: Path Profile Faces: 4th-axis integration

Post by freman »

this one allows me to use the 3/8" endmill that I want to use in
real life, where the pocket operation doesn't allow this.
Look at extensions in the pocket set up dlg. Deselect "extend corners" and then you can select to edges to extend . That should allow you to "pocket" beyond the edge of the face.

BTW it's great to see this 4th axis capability coming together. HT to the formidable Russ4262.

Hopefully by the time I get round to adding a forth axis, there will be support for it. ( That could be quite some time ).
Post Reply