On Sun, Jul 11, 2010 at 10:15:22PM -0500, Jon Elson wrote: > But, if I understand this, the amp enable would come on at the same time > as the position > loop, and the following error logic would be enabled at the same time.
True, it all happens at once. > I don't know this, > but I have a strong suspicion that the transient when the velocity loop > suddenly becomes > closed will cause an immediate following error. I haven't seen this on any of the machines I've worked on. When in machine off state, commanded position follows feedback position, so at the instant the amps enable, the ferror is zero, and pid's output is zero because there is no error, no motion is commanded, and the integrator has been cleared. Since pid output is zero and velocity feedback is zero (because the machine is not moving) there should be no transient. > The machine definitely moves a small amount > when the amp is enabled. Is this somehow because it's crept out of position? Is it trying to jump back to the pre-creep position? I'm not sure what would cause this. > Also, do these axis.N.amp-enable-out outputs ALL work in unison, or will > only one > axis disable if that axis gets a following error? It would be easiest > if I could just use > one SSR from, say, axis.0.amp-enable-out to control all the servo amps. Yes they all go on and off together. Only the mists of time know why there's one for each axis. Just use one of them. > (Also, using this scheme would make it impossible to easily tune the > zero velocity offset, but > I could figure ways to do that with a clip lead. I don't do this often.) I've done this adjustment by opening up the ferror limit to a few inches, unhooking the dac's input and setting it to zero in HAL, and then turning machine on. Then I use the EMC DRO to see which way it's creeping and adjust the amp to stop the creep. This way no futzing with wiring is required. Additionally with this scheme, if something goes wrong and you get a runaway, EMC will still stop it for you. > Thanks for the comments. Anytime. Chris ------------------------------------------------------------------------------ This SF.net email is sponsored by Sprint What will you do first with EVO, the first 4G phone? Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
