Marc van Doornik wrote: > Right. It seems I have got it, now. Feel a bit Kelly Bundy-ish right > now, but I've finally realized my mistake. All the while, I've been > treating the linear motors as if they were stepping motors with a large > pole period (2.4", by the way). This also explains all that hysteresis > stuff and the spungy/springy behavior and so on, which - indeed - is > nonexistent if these motors are driven in the **correct** way. So, from > what I've gathered, this is what I should do now: > > 1. Get forcer aligned to the stator field (using hall sensors) > 2. Offset motor phases by 90 degrees and lock them to the forcer position > 3. Have the velocity PID control the current through the windings to > generate the desired force, while keeping the phase locked to the forcer > position > > Fortunately, the amps have controlled current output, so that a given > input voltage always corresponds to a given winding current. This should > make things a lot easier, right? (Jon?) OK, with the torque-mode amps, this should all work. I'll be very interested in what you develop, because I have some voltage-mode amps that I'm using with EMC2 and they are not optimum. I am thinking I need several things to improve the performance. One is to filter either the position error signal or the D to reduce the 1/2 sample rate content which tends to make the D term of the PID way too sensitive. Second is to come up with a better algorithm to simulate more closely the pos/vel/torque loops of the classic velocity servo systems. My amps do NOT have current feedback, just limiting. That is a bit tricky, and I'd like to see if I can get more stable performance without having to add that. But, deriving velocity from the encoder and trying to implement a velocity loop calculation seems possible. I have a little trouble getting my mind around the implications of a continuous-time tach vs. the sampled performance of an encoder. You have both the need for filtering as mentioned above, plus the fact that the encoder produces pulses at discrete locations, and no data between them. Less of a problem with the insane resolution of your system, however.
Jon ------------------------------------------------------------------------- SF.Net email is sponsored by: The Future of Linux Business White Paper from Novell. From the desktop to the data center, Linux is going mainstream. Let it simplify your IT future. http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4 _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
