we're doing final bug hunting in the g code for all the parts. we've already found and fixed most of the bugs in the backplates, but check out this funny bug in the validation samples for the M7 mount. I'm using fusion 360, so instead of doing nurbs modeling like a sane person would in solidworks or something, i just used a ****-off huge fillet command:
View attachment 145772
That means that there is a line boundary in this curve where there shouldn't be a line boundary. A human would look at this and realize that it's from using a fillet command and understand that this is supposed to be an uninterrupted surface, but CAM software does not understand the concept of shape. It is programmed to generate toolpaths that ensure that each
face fits the drawing within a certain percentage error. It looks at this and goes "oh, two faces!" and generates ideal toolpaths for each face. But it doesn't consider the area in between the faces, and it cannot create an infinitely small boundary between the faces, so the result is:
View attachment 145773
sorry for the dust i'm handling them with my fingers in my home office which i know I shouldn't do.
the CAM can't generate correct toolpaths for the shape, even though it can do the faces fine, so this will need to be manually altered.
I'm not going to go out of my way to hand-polish the mass production m7 mounts or anything, so I wanted to figure out how good we could get a purely machine finish on these parts. The results are not too shabby:
View attachment 145774
View attachment 145775
considering that this is not one material, but brass that we nickel plated after without re-processing, this is pretty good for a machine only finish. this should be nice and cheap for us to do at volume and, most importantly for a small 5ish-person shop, almost zero human effort-time