The drawing is the outline of a Fender Telecaster, the file was optimised using Rhino V4 before producing the code using FeatureCam V15. The lines are contiguous to +/- 0.0001mm. If there were errors with the drawing both would complain and not accept it as one continuous outline.
i see no reason to have the resolution of your cam system set this high if i did that in mastercam it would triple the code although i still dont think this is your answer just an observation? i have asked a bit about look ahead and have heard some muffled replies what exactly is the lcnc look ahead ? it does not appear to be infinite? and last I really prefer to avoid cv actions as they do not produce the most efficient programs, but nonetheless i would be interested to see what is discovered here On Tue, Apr 9, 2013 at 6:14 PM, Steve Blackmore <st...@pilotltd.net> wrote: > On Tue, 9 Apr 2013 13:55:30 +0100, you wrote: > > >On 8 April 2013 22:57, Steve Blackmore <st...@pilotltd.net> wrote: > > > >> ini file > >> http://pastebin.com/keXTHWyn > > > >I don't really have the tools here to analyse this, but a few > observations. > > > >At 800mm/sec2 accel and 1200mm/min traverse speed the minimum arc > >radius is 0.5mm. > > Slowing the acceleration or max speed makes no difference, it still does > it, just less obvious. The sound in the videos is the big giveaway, you > can hear the changes in velocity easier than see them. > > >The circular moves do not appear to be tangent to the straight lines. > >I wonder if they were meant to be? > > The drawing is the outline of a Fender Telecaster, the file was > optimised using Rhino V4 before producing the code using FeatureCam V15. > The lines are contiguous to +/- 0.0001mm. If there were errors with the > drawing both would complain and not accept it as one continuous outline. > > >I wonder if Mach and LinuxCNC are > >taking a different approach to blending the corners? > > Almost certainly. > > Did you read the 2011 posts? In particular the one where Art explains > how he did CV for Mach? > > Also Daniel at Tormach concurs, his observation follows for those who > haven't read the old stuff on this > > "The problem is apparent at the first G2 move. The machine appears to > change feedrate between G2 and G1 moves. Moving from one G2 line to > another G2 line is smooth, and moving from one G1 line to another G1 > line is smooth. G1 moves appear to run at around 60% of the feedrate of > the G2 moves, so the transition from G1 to G2 (or G2 to G1) makes the > machine seem erratic. If I slow the feedrate down, the problem > persists." > > Steve Blackmore > -- > > > ------------------------------------------------------------------------------ > Precog is a next-generation analytics platform capable of advanced > analytics on semi-structured data. The platform includes APIs for building > apps and a phenomenal toolset for data science. Developers can use > our toolset for easy data analysis & visualization. Get a free account! > http://www2.precog.com/precogplatform/slashdotnewsletter > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > -- jeremy youngs ------------------------------------------------------------------------------ Precog is a next-generation analytics platform capable of advanced analytics on semi-structured data. The platform includes APIs for building apps and a phenomenal toolset for data science. Developers can use our toolset for easy data analysis & visualization. Get a free account! http://www2.precog.com/precogplatform/slashdotnewsletter _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users