Sunday 24 January 2021

4th axis action!! David Loomes' dodecahedron. Swissi post processor issue?

Background - 4th axis operation vs Fusion 360 changes:

David Loomes (Xoomspeed) has done some great work on his Tormach / Pathpilot system, particularly with in-process probing and his wireless probing system, that allows a touch probe to be used in an ATC system. He's done some great videos, sharing what he has been up to.

I looked at his probing macros when I was weighing up the possibilities of implementing probing macros in the Fusion 360 post processor. I soon realised I wasn't up to the job and pragmatically gave it up as an unrealistic challenge. That turns out to have been just as well, since Swissi (on the Centroid forum) came out with a fully developed and enormously powerful probing app that far exceeded anything I could ever have hoped for. 

Not so long ago, I put the finishing touches to my own design of 4th axis, using a harmonic drive and a Yaskawa Sigma servo drive. At almost exactly the same time, Autodesk performed the full reverse ferret and announced the end of free access to Fusion. Or at least a significant hobbling of many of the functions in the free version. This didn't sit well with my Scottish-Yorkshire heritage, given the annual subscription cost. Furthermore, the details of what was in the free version and what was in the paid version was as clear as mud. I dismounted the 4th axis from The Shiz and left it to collect dust while the situation became clearer and I sulked about the new reality.

Tim Paterson has developed a pretty functional Python add-in for Fusion that allows the use of multiple tools and the replacement of rapids. These were the main annoyances that result in the free version. Have to say it worked fine, once we'd figured out the need to issue a G90 after each toolchange retract. However, he designed it to run all the toolpaths in the current file, whereas I often have multiple setups and machining operations within each component. He didn't seem keen to change that, and it made the outputs files rather difficult to work with - for one, they always ended up being single digit, sequential files (one per setup) and neither the post processor nor the finished file would be presented during the post operation. Not very user friendly, although it's perhaps churlish to criticise.

Finally, as 2021 dawned on us, I had recovered enough from my period of sulking and moaning to contemplate coughing up for the annual subscription. The NY offer of 30% off is about as good as it will get, most likely, so I thought bollocks - in for a penny, in for a pound (a day).

Almost a year ago, David posted a video of a 4th axis test piece he'd created for checking out his brand new Tormach "Microarc" 4th axis. This doesn't require the ruinously expensive "manufacturing extensions" to run, although it does require the paid version which allows for positional 4th axis moves.


Interesting. And sure enough, with my newly licensed Fusion 360, I can post process his sample file, which he generously posted on his website. I made a few changes to the model and the setup:

  • I want to use my favourite 10mm long series end mill, rather than the 6mm and 8mm cutters he used. 
  • I needed to extend the model's supporting "nose" to avoid the larger cutter gouging the stock towards the end due to lead-in and lead-out moves.
  • I finessed a few of the parameters to suit my machine and cutters.
This at least shows that the paid for version can implement positional (indexing) moves. So while I may not be able to generate true simultaneous (sort of "full 3D"?) toolpaths, there is still useful functionality to be had. It seemed like time to end the royal sulk and see what can be done with my 4th axis.

Here's the part in CAM - with my 4th axis used as a fixture. The jaws are the outside type here but that's a sort of worst case condition. I'll perhaps grab some inside jaws if I can be bothered some time....




Ready to go? I mounted and trammed my 4th axis, loaded up some 32mm loominum and braced myself for action.

Hold up, fatty!
Seems it's not as simple as that. Swissi recently released an update (v005) for his supey dupey post processor. Naturally, I installed that, as he needs people to test it out and find any issues.

Setting up the workpiece in Y:


....and Z


Air cutting the part. Not long into the program, the machine wants to drop the spindle to this position. Doesn't look right!


Here's what the tool and toolholder look like. That's a crash in anyone's book. WTF is going on?


Luckily, the Centroid s/w was clever enough to spot an issue, although this issue could easily have caused a problem that wouldn't have got picked up.

Turns out the v005 is generating an "E10" command instead of a "G54", on my machine at any rate. Knowing me, it's more likely I've configured something incorrectly, rather than Swissi messing up. Either way, v003 still works correctly, so v003 it is. The effect of the "E10" is to screw up the WCS - just what you need to break things. Nice.

Squeaky bum time:
So, having spotted and (hopefully) rectified that, at some point it becomes necessary to get on with the job. Which I did. 


Result!
That came out well. 



Perhaps I'd have got an even better surface finish if I'd upped the speeds and reduced the feeds on the finishing passes but I'm not about to repeat the whole process. Suffice to say, I'm very pleased with the end result, I didn't break anything or crap myself, I am getting a feel for what my 4th axis can do for me and I'm very grateful to David Loomes for inspiring me to stop sulking and test out my 4th axis.

Happy chappy!

No comments:

Post a Comment

Final assembly and test of the spindle nose adaptor - RESULT!!

After the recent distraction caused by the 3D scanner, resurrecting the 3D printer and buggering about with the throttle bodies for my Honda...