Richard Arthur wrote: > Certainly NOT complaining :-) > I'm having fun learning about this. > > Machining stuff is (part of) what I do. Ninety percent of my machining > only require wide tolerances, but it's not uncommon to have to work to > +-0.0002" on manual machining. This 'theoretical' f-error of +-0.0003" > is already outside that and I wouldn't expect the situation to improve > downstream. But as yet, I don't actually know how this f-error will > translate in the real world machining parts. > Well, that error is at some unspecified velocity. It may scale to some extent with velocity. Also, what is your encoder resolution. A general rule of thumb is to have 10 X the encoder resolution as any tolerance you need to hold. So, to work to 200 uInch, you need encoder resolution of 20 uInch, or 50,000 counts/inch. > Thank you for taking the time to reply (I'm off to find out about double > quantization noise :-) ) > That may be my own term, so you may not find much. But, what I use it to refer to is the fact that encoders obviously quantize movement into counts, but the reading of that position is also quantized by periodic sampling of the encoder counters. Taking position that has already been quantized to integer raw counts and then sampling it regularly produces a "jitter" in the number of counts recorded every sample period, and that "looks" like velocity fluctuations to the servo system. It also can show up as fluctuations on the computed velocity and following error as you see on HalScope. (Not all of the noise on your scope traces was that, but the quantization noise could perturb the servo loop to produce what you are seeing.)
Jon ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
