Am 20.08.2013 um 14:32 schrieb Steve Stallings <[email protected]>:
> Originally it was thought that the computers of the > time did not have the horsepower to run the trajectory > planner fast enough, so interpolators were used to > achieve a faster update rate. > Later as computers became faster this was not necessary > and it was discovered that the interpolator approach > caused some problems when rigid tapping. At that time > using a 1:1 ratio became the default. now that the current tp has been made RT-dependent for spindle-synced moves - followup question: does the current tp code _assume_ it runs at the servo rate? Has this been ever been tested at all with different periods? NB: with the current thread code, the 'monotonic rate scheduler' is in place only in a comment, meaning completely unsynchronized threads with best-effort behavior must be assumed worst case. > I would prefer to keep the possibility of interpolation > because future decoupling of processes may once again > create a situation where interpolation might be an > advantage. yes; I see several good reasons for not having the tp be realtime, and they are not confined to slow platforms. For instance, planners with deeper lookahead might employ more costly algorithms than what can be done within a servo cycle. Another one is external hardware - see Yishin's implementation of rigid tap, which pushes the syncing operation a layer below (where it belongs IMO) -m > > Steve Stallings > > >> -----Original Message----- >> From: Michael Haberler [mailto:[email protected]] >> Sent: Tuesday, August 20, 2013 4:09 AM >> To: EMC developers >> Subject: [Emc-developers] q on motion rationale: why are >> cubic interpolatorsdisabled? >> >> >> reading motion code I noted that the trajectory and servo >> cycle periods are identical (at least by default) >> >> this means the cubic interpolators operate at rate 1 per >> segment, meaning they are completely disabled and provide no >> smoothing ontop of what is generated by tp.c/tc.c >> >> what was the rationale for disabling the interpolators to start with? >> >> - Michael >> >> >> >> -------------------------------------------------------------- >> ---------------- >> Introducing Performance Central, a new site from SourceForge and >> AppDynamics. Performance Central is your source for news, insights, >> analysis and resources for efficient Application Performance >> Management. >> Visit us today! >> http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/414 > 0/ostg.clktrk >> _______________________________________________ >> Emc-developers mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/emc-developers >> > > > ------------------------------------------------------------------------------ > Introducing Performance Central, a new site from SourceForge and > AppDynamics. Performance Central is your source for news, insights, > analysis and resources for efficient Application Performance Management. > Visit us today! > http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk > _______________________________________________ > Emc-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-developers ------------------------------------------------------------------------------ Introducing Performance Central, a new site from SourceForge and AppDynamics. Performance Central is your source for news, insights, analysis and resources for efficient Application Performance Management. Visit us today! http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
