Andy Pugh wrote: > (Sorry to start yet another thread on much the same subject, but this > is a bit broader-ranging) > > It seems that spindle-synchronised motion is very sensitive to encoder > regularity. > > Using the p-port to read the encoder position means that as the > encoder linecount goes up, quantisation error in the measured > positions also goes up, and so the inter-line period noise increases. > Does this mean then, that very low linecounts will give less dither, > and better results? > Apparently not. I am using an encoder with 1728 cycles/rev, or 6912 quadrature counts/rev. It works fine, although some improvements in EMC2 that were made last year were quite helpful. Also, it is quite important that the trajectory planner be running at the servo update rate, and not the typical 5X slower rate that was used in earlier versions of EMC2. > http://www.pastebin.ca/1583502 > Is the raw halscope data from my machine. If you plot the encoder > position it looks dead straight, but the raw data shows that the > difference between successive samples is quite variable. > The last column seems to show wide variations, indicating maybe the sampling is missing complete revolutions. Lines 7 and 19 show such a large variation, for instance. This may be due to noise in the signal, too. This is the reason I strongly advise people, OVER and OVER, to not try to use software sampling of such signals.
> Is this possibly the reason that my attempts at threading are so hopeless? > > I would be the first to admit that my encoder is rubbish, it is a bit > of computer-printed paper wrapped round the lathe spindle, but even if > I was to get a high-accuracy encoder disc I feel I would still be at > the mercy of sample quantisation errors > Any opinions on the best way forwards? I really don't think it is the encoder, but can't be sure. If the sensor sees a lot of light near the decision threshold, it may be flickering on and off instead of solidly sending ON, then OFF, then back to ON at the right times. If it is flickering during periods when it should be sensing a solid on or off, the software will sample the wrong thing. You really need an oscilloscope, but you might be able to see enough in HalScope to determine what it is doing. Jon ------------------------------------------------------------------------------ Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
