Hi, I had resolvers with a logic that transforms them to 1024 lines (4096) counts per revolution.
I introduced 1micron linear scales for each axis. I would go for finer but, I had to go from magnetic to optical and it was a huge difference in price. Accuracy (for 1 meter) is few hundreds of mm. Motors was actuating linear guides without any reduction, and I was suspicious about the precision of the former resolvers so I made a comparison, for a 1rev/cm i can see more or less cyclic deviations of some hundreds of mm on each revolution. I have pretty good THK linear guides so most of the error must come from the resolvers. So I think linear encoders where a need in my design. Still I think there is some room for improvement using both sensors. Not in my case, but inyour's, having such a big resolution rotary encoders, it makes sense compute velocity feedback signal from such a sensor, while using the linear sensor for position feedback. At least if you don't have backlash. The derivative of the position signal can be built out of the reference signal and your rotary encoders. I suspect the this would need a little rewrite of the PID component. I thought these comment could be useful to some body. Have a nice day, Javier PD: I'm working in a component that tries to compensate the inaccuracy of my rotary encoders, now that I know which are the errors so that I can improve precision even if I remove or I don't use the linear encoders. I have to filter the data as it is too big as to go in a real time component. Right now I'm filtering using a standard Fourier series, and introducing the sins and cosines in an array so I don't have to compute those functions again. But still I'm not satisfied with the quality of the match, probably because lac of periodicity in the errors of the linear guide, and because the linear encoder is not perfect at all. I'm pretty sure that there exist a possibility to obtain the best from both worlds if using both sensors as each one is best in a different aspects. Does anybody has any experience on this?. I was thinking about using wavelet transform, or to solve an optimal problem assuming and adhoc position error parameterization for each of the encoders (linear and rotary), that would give back the parameters and then a better formula to approximate the error based on the measure of each sensor. On Wed, Sep 7, 2011 at 2:55 PM, Frank P Bufton <[email protected]> wrote: > I have been rebuilding a mill (much delayed due to work and my mothers > Alzheimer's ) and thinking about how to incorporate linear scales into the > control loop. I have 5000 line (20,000) counts per revolution encoders on > the motors, and have new Acu-rite one micron scales for the axis. In would > like the scales to determine the final position. > > This may get my mill out of the corner. > > Frank > > > > > From: dave <[email protected]> > To: EMC developers <[email protected]> > Date: 09/06/2011 05:10 PM > Subject: Re: [Emc-developers] Linear scales > > > > My first try at machine control used linear scales. They worked but I > had 0.0025 backlash in X and Y and the control got considerably smoother > when I put a good incremental encoder (10K counts/rev and a pitch of 4) > on the ball screw. The tuning was also much easier. IIRC Stuart is using > the linear scale to drive I of the pid to force final position. This > would seem to be the ideal configuration. > > Years ago a Boeing engineer stated that machine resolution should be 5X > better than the tolerance(s) you are trying to attain. > > Just my tuppence. > > Dave > > > On Sun, 2011-09-04 at 21:41 -0500, Jon Elson wrote: >> Jan de Kruyf wrote: >> > 20 or 25um with a multiplierbox to make 1um has been pretty much >> > standard for many years. >> > >> First, you need analog scales to use an interpolator. Digital-output >> scales can't be interpolated. >> Then, in many cases, the interpolators cost several times what the basic >> scale costs. Finally, the >> interpolator has to have fast response, any delay in reporting counts >> will cause erratic servo >> loop response. >> >> Jon >> >> > ------------------------------------------------------------------------------ > >> Special Offer -- Download ArcSight Logger for FREE! >> Finally, a world-class log management solution at an even better >> price-free! And you'll get a free "Love Thy Logs" t-shirt when you >> download Logger. Secure your free ArcSight Logger TODAY! >> http://p.sf.net/sfu/arcsisghtdev2dev >> _______________________________________________ >> Emc-developers mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/emc-developers > > > ------------------------------------------------------------------------------ > > Malware Security Report: Protecting Your Business, Customers, and the > Bottom Line. Protect your business and customers by understanding the > threat from malware and how it can impact your online business. > http://www.accelacomm.com/jaw/sfnl/114/51427462/ > _______________________________________________ > Emc-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-developers > > > > ------------------------------------------------------------------------------ > Using storage to extend the benefits of virtualization and iSCSI > Virtualization increases hardware utilization and delivers a new level of > agility. Learn what those decisions are and how to modernize your storage > and backup environments for virtualization. > http://www.accelacomm.com/jaw/sfnl/114/51434361/ > _______________________________________________ > Emc-developers mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/emc-developers > ------------------------------------------------------------------------------ Doing More with Less: The Next Generation Virtual Desktop What are the key obstacles that have prevented many mid-market businesses from deploying virtual desktops? How do next-generation virtual desktops provide companies an easier-to-deploy, easier-to-manage and more affordable virtual desktop model.http://www.accelacomm.com/jaw/sfnl/114/51426474/ _______________________________________________ Emc-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-developers
