Tool Path Rounding Issue

Michael I have a problem with what you mentioned here of the Fusion 360 smoothing, when I use this option in the axis when trying to open a g code I have problem of arcs at the beginning and at the end, and when I don't use this option axis opens the code without problems, maybe you know what is happening?

Not sure what you are referring to. Are we talking about the settings in fusion 360 or a problem with a line of g code?

When I use the smoothing function in Fusion 360 it creates linear and arcs to smooth the splines with the pp it creates a G3, then when I want to open the G code in linuxcnc it gives me a different arc problem in the end and the start error. Now when I don't use that Smoothing function the Plasmac interpreter which is the one I use works fine.

Smoothing in fusion will linearize the arc. If you want a pure arc turn off smoothing. And decrease the tolerance in the fusion tool path to what ever your machine can tolerate.

Are you referring to the lead in and lead out of the tool path? Fusion creates a lead in/out on most of it's tool paths that is a ramping arc usually in movements of 10% of cutter diameter of I remember correctly. These can be turned off on the final tab of the tool paths.

Also if you want to eliminate linuxcnc from all types of path blending then add G61 to the beginning of your g code file. This is exact oath following.

G64 is blending ng and on low acceleration machines it will blend a lot from the lead in to the actual arc you want.

G64 Px.xxx will blend with the limited amount. So between two paths P.005 will not go more than .005" away from the intersection of those two paths.

Sorry about my English I will explain them better. I am using PlasmaC and Fusion 360, so when I cut pieces that I don't use the smoothing function everything goes fine, but when I use smoothing in Plasmac I have the arc error (attached photo), so I check the G code of both pieces one with smoothing and the another one is not and the difference of course is the size of the G code and a G03 which is the one that produces the estimates of the smoothing then that new code is the one that the PlasmaC interpreter gives that error, I don't know if my error is in the number of tolerance but I have tried several and I continue with that error. 

Can you post the two g code files? Why do you want to use shooting from fusion?