John Kasunich wrote: > Jon Elson wrote: > > >>Do you really NEED 100 nm resolution on the encoders? (My G*D, >>how much did you pay for those?) Is there a way to divide down >>the encoder resolution? Do these encoders use interpolator >>boxes to get higher resolution? These usually have settings to >>select from some different resolutions. But, beware that some >>interpolators cause a phase shift in the output. Anyway, 100 nm >>resolution is less than a wavelength of visible light, >>presumably your cutter is using IR, so it seems like useless >>resolution. >> >>Jon > > > I'm going to respond to the original post in detail a little later > (something else going on right now). But I just wanted to comment on > the encoder resolution. As Jon says, you MUST be able to count the > encoder at the maximum possible speed. However, I would not be so eager > to reduce the resolution. To control this motor, you are going to need > not only good position feedback, but good velocity feedback. A very > fine encoder resolution will greatly reduce the quantization noise when > you convert position to velocity. Yes, but 100 nm = ten million counts/meter. If the table moves up to 2 m/sec, that is a 20 MHz count rate, if my math is not screwed up. Even if the encoders can handle this (5 MHz quadrature frequency) not too many encoder counters can handle it. Cutting the rate down until there is no achievable velocity that exceeds the encoder/counter system's capability is a first step. Still, you should get PLENTY of counts to derive velocity from this system.
The original poster's comment about overcoming the static friction of the linear bearings with moves at 2 m/sec makes me wonder about his level of experience with precision motion control systems. You may need strong acceleration to null out this final settling error quickly, but certainly not high velocity. Jon ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
