when linear (X,Y, or Z) motion is coordinated with angular (B) motion, the feedrate is relative to the linear axis(es). That accounts for the unusually high feedrates on many linear-angular coordination motion stmts, and since you like that motion when you observer the machine, I’m assuming those are correct.
Those ‘flawless’ motions also have similar distances as line N0260, the one in question with F120.
So agreed, F120 seems grossly wrong in that stmt, and you observe it on the machine as such. F120, more likely purposely, is also used on line N0140, which is purely linear motion, presumable making it correct.
So this may be a bug where the feedrate did not get adjusted for the added angular component of the motion on line N0260.
Please prepare a support.zip file and send to sales@sheetcam.com
Creating a Sheetcam Customer Support file
In the meantime of us looking over your support.zip, I recall sheetcam v7.1.40 having a fix concerning cutting through corners of rectangular tubing, you might give that version a try if not already using it.
This anomaly is also occurring at the very top of the flattened drawing, where that corner is split between the top and bottom of the drawing, just an observation of the transition motion from 180* to 185*.