On Fri, 16 Jul 2010, John Kasunich wrote: > Date: Fri, 16 Jul 2010 09:38:32 -0400 > From: John Kasunich <[email protected]> > Reply-To: "Enhanced Machine Controller (EMC)" > <[email protected]> > To: "Enhanced Machine Controller (EMC)" <[email protected]> > Subject: Re: [Emc-users] Spindle at speed & Coordinated Motion problems? > > > > On Thu, 15 Jul 2010 08:42 -0700, "Peter C. Wallace" <[email protected]> > wrote: >> On Thu, 15 Jul 2010, Sebastian Kuzminsky wrote: >> >>> On 07/15/2010 08:54 AM, Peter C. Wallace wrote: >>>> Maybe a HAL Comp that uses ddt but tosses crazy velocities >>>> (uses previous calculated velocity if index has been detected) >>>> would solve the problem >>>> >>> >>> Sounds like a combination of ddt and lowpass. >>> >>> http://linuxcnc.org/docs/2.4/html/man/man9/ddt.9.html >>> http://linuxcnc.org/docs/2.4/html/man/man9/lowpass.9.html >>> >> >> Not quite, I was suggesting tossing bad data so even if the velocity is >> filtered the bad data does not contribute to the average >> > > How about ddt followed by limit? The limit block applies > a slew rate limit to it's output. Set the limit about 2x the > maximum spindle accel/decel rate, and any wild spikes from > reset will be cut down to size, with recovery to the correct > value on the very next update - no long tail like you'd get > with lowpass. > > John Kasunich > -- > John Kasunich > [email protected] > >
That should get most of it. I think they use or used to use the same technique for impulse noise filtering in radios = slew rate limited non-linear filter. Peter Wallace ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
