Fusion 360 Mill Post Processor for Acorn with additional Features (New Version 5)

[hebs]

Swissi,

Report attached thanks.

[swissi]

4th Axis Testers Wanted!!!

After Autodesk made improvements on 4th Axis Tool Paths, I migrated the latest Work Plane Setup logic from the FANUC A-Axis post processor to my version of the Centroid Milling Post Processor. This should eliminate some of the unnecessary un-wind and re-wind moves.

Please give it a try and report back your results.

This version of the post processor also includes the support for Fusion 360 WCS and Geometry Probing but the output of the Post Processor won't do you any good without the necessary probing cycles that I have currently in Beta Testing. Stay tuned for more information on this subject.

Beta Version of Fusion 360 Milling Post Processor with enhancements for 4th Axis support:
[Attachment Removed]
Check this thread for the new Beta Version: viewtopic.php?f=60&t=2992&p=37065#p37065

If you are currently using an earlier version of my Post Processor, please note that I had to move variable #300 to #330 and #301 to #331 so if you are using the mfunc6.mac file that displays all the Information, you have to modify these two variables or use the new mfunc6.mac file that's been attached to this Beta version.

-swissi

[Muzzer]

Good timing, Swissi. My 4th axis is now mounted, trammed and ready to go. Once I've updated to your latest post and created a suitable test part, I'll give it a go. But it seems I have some more drywalling to do first, as the Domestic Manager is now up and about.

Last time I tried this (2 years ago), I could only manage indexing, not true simultaneous. That was partly the post and partly the controller, from what I could figure out. Looking forward to trying it out!

[Muzzer]

Well I can confirm that that post (beta v003) seems to work as hoped. There was no unwinding between moves and although I was only cutting air, the tool seemed to be following the right sort of path.

Quite a simple operation - 2D Adaptive, so hardly a major validation. If I can trust this Leadshine clone stepper driver not to stall halfway through, I might even try it on some real stock later on...

[Muzzer]

I've just run the toolpath on a bit of mystery metal with and old HSS cutter. Combine that with loads of overhang and the 2D Adaptive roughing operation and it's not a pretty result. However, it worked fine. No unwinding.

There are some jerky moves that aren't quite ideal. Is that likely to be my settings rather than the post? I'm assuming I need to get my head around smoothing now. IIRC, there are options within CNC12 and also within Fusion CAM.

See what I mean:

[swissi]

Muzzer,

looking at the video, there seems to be sudden accelerations along the X-Axis at some places. Did you look at the G-Code when the program was running if there were indeed such federate changes at those positions when that acceleration occurred?

Can you share that model in Fusion 360 with the setup tool path so I can have a look at it?

-swissi

[Muzzer]

Yes, that's what I was concerned about - looked a bit alarming.

The Fusion model is here https://a360.co/2ytOHeG Hopefully you should be able to open and download it from there. Let me know if you can't get it.

Here's the g-code:

5552.nc

(11.56 KiB) Downloaded 166 times

My first thought was that I might have got some dumb settings in the CAM somehow but they don't look daft to me at first sight:

I'm also not clear about how to play the various smoothing settings options.

EDIT - there seem to be a few F2500 and similar in there.....

[swissi]

Muzzer,

well I can confirm that the official FANUC A-Axis Post Processor, which my 4th Axis Inverse Time logic is based on, generates exactly the same output so that jerkiness has definitely nothing to do with my version of the post processor.

Using Inverse Time for moves involving the 4th Axis, Fusion 360 calculates the length of each move in 3-D space and then calculates the duration for that move based on the demanded feed rate. With Inverse Time the F value indicates the time the move will take rather than the usual F values in mm/min. So the higher the F value the shorter the time for the move.

Keep in mind that the 2D Adaptive strategy you used is a roughing and not a finishing strategy and that's exactly why your part does look a little "rough" (and I guess also because of the chatter caused by the extremely long and unsupported overhang).

Try to finish your part with a finishing strategy like a wrapped 2-D Contour. I bet those moves will be much smoother and will give a better surface.

-swissi

[Muzzer]

Yes, I pointed out in the opening post that this was a roughing strategy and the stock was massively overhung. And it's a steel cutting end mill just to make matters worse! However, the jerkiness is surely nothing to do with any of that, since it's not even loaded at the time. Certainly, I'd be disappointed if the finish weren't improved with use of a finishing path - I'll perhaps finish the job off tomorrow evening with something appropriate, in the interests of science.

I was a bit confused to see the F2600 commands in there. I don't understand how the inverse time scheme would affect linear moves but my feed rates in Fusion were 1000 and 1200mm/min, so I'm confused by presence of that 2600mm/min in the g code, if that's what it is. I didn't have time to dig into it but do those moves correspond to the sudden moves, I wonder.

[cncsnw]

In G93 mode, even if the move is just a single straight line, the feedrate value means "moves per minute". F2600, then, means that the move should complete in 1/2600 of a minute, or .023 seconds. At 1200 mm/min, that would be enough time to cover 0.461 mm.

If you see a single line in the program, that says F2600, but that moves farther than 0.461mm, then that would appear to be an error in the CAM software.