Fusion Manufacture
Talk shop with the Fusion (formerly Fusion 360) Manufacture Community. Share tool strategies, tips, get advice and solve problems together with the best minds in the industry.
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Circular interpolation on rotary axis

Message 1 of 10
538 Views, 9 Replies

Circular interpolation on rotary axis

We're trying to utilize F360 to engrave a pattern wrapped 360 degrees on a cylinder around a 4th axis with G02/03 output. I'm having trouble getting the post processor to output G02/03 and think it may be the tool path.

I've got the path tolerance and smoothing tolerance both set to .0001, and have tried adjusting the max circular radius (up) along with min chord length / min circular radius (down) in the post processor. Usually this will force circular interpolation but it's not working here.
Is there anything else I should look at before digging into the PP?
Machine: Hardinge GS 200 MSY
Controller: Fanuc Series 18i-TB 
PP: jyoti ax 200-msy fanuc.cps 
Labels (1)
Message 2 of 10
in reply to: shopping3RT5G


On wrapping a pattern on a cylinder the machine will effectively(once Z has positioned for depth) be only moving your rotary axis and one liner axis(on my config A and X). You can only have G02/G03 moves on planes across linear axis, not rotary axis, except for lead in/out moves.

A wrapped tool path around a rotary axis will have to do a multitude of small point to point moves, the distance between these points will be effected by the path/smoothing tolerances and therefore increase your program size.

Message 3 of 10
in reply to: Fabbunny69

Thanks for the quick response-


Just clarify, Fusion will restrict to linearization for rotary axis? Asking because I'm able to go the manual route and run this on a rotary axis:







Message 4 of 10
in reply to: shopping3RT5G

Any chance you could attach a file to show what kind of wrapping project you are working on?

Message 5 of 10

Hello @shopping3RT5G 


Effectively when Fusion is passing the toolpath to the post, we only have the wrapped toolpath.

So it's linearized without G2/G3.

We don't have actually the ability to retrieve the un-wrapped toolpath.

We have no post support function like G107 or TRACYL.





If my post answers your question, please click the "Accept Solution" button. This helps everyone find answers more quickly!

Technical Consultant
Message 6 of 10
in reply to: serge.quiblier

Is there a possibility for TRACYL in the Post Processors? SIEMENS 840D with FUSION360

Message 7 of 10
in reply to: f_feroglu

Hi @f_feroglu 


No, the behavior is still the same.
When the toolpath is generated by Fusion, it is created from the solid, or from an edge that is wrapped on the cylinder like that one :




But to use tracyl or any similar function like G107, you must have the unwrapped curve, i.e., that one:




The flattened curve is currently not created by Fusion, so it cannot be passed to the post engine.









Technical Consultant
Message 8 of 10



Message 9 of 10

Will this area be possible with TRACYL?
Message 10 of 10



it doesn't really depend on the exact shape. It's a general limitation of the system.

That doesn't mean, that it's not doable, but you'd need someone who really know's what they're doing,

it will be complicated to do (will be hard to find someone who would do something like this remotely for free, I'd say),

and one'd be working "against" the system or rather against how things work in the system.


But then... If you have a look at how subroutines are implemented in the posts, things are not really soo much more elegant there,  imho...

Did not really spend much time on thinking this through, but roughly sketched out one could try the following:

- check operation parameters for 4x wrapped op. Probably

   'operation:is2DStrategy' , 'operation:isRotaryStrategy' (should both probably have a value of 1)  and 'operation:isMultiAxisStrategy' (this one probably 0).

- turn off movement output in onSection (start of op) and record moves / params.

- after the section is done (e.g. in onSectionEnd), unwrap all recorded coordinates.

- do a 2D arc fit (there are quite a few resources on that if you really look for it).

- output all moves, including G02/03. You might do the output manually, or probably you could also call the respective onLinear, etc. functions yourself and pass the recorded params. You'd have to test that.

I'm not really interested in implementing the above, and I won't recommend it, but if you found someone...  From the top of my head, that could be a way to go about solving your problem.

Good luck


Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums