First of all, edit Your INI file and set A joint from angular to linear, because that joint _is_ linear. And then try again. Second of all, I am surprised to see MIN_VELOCITY parameter. Try to set it at 0, maybe that has something to do with very low accelerations, when the actual speed during acceleration is below that Min-Velocity treshold. BTW placing all those velocity and acceleration parameters in [TRAJ] section, where they belong, would make the file easier to read for others. You somehow have them in Display section...
-- Viesturs If you can't fix it, you don't own it. http://www.ifixit.com/Manifesto 2012/7/11 Hugh Wylie <hwy...@mweb.co.za>: > Has anyone experienced the issues which we are experiencing on a > robustly-designed gantry comprising the Y-axis of a large-bed CNC router > (size 2.75 x 1.83m bed)? > > > > The system runs LinuxCNC 2.5 on Core 2 Duo platform equipped with Mesa 5i23 > (firmware SVST4_8) controlling the four stepper motors (via Gecko 203V's). > Jitter is a max of ~43us, but that should not be an issue using the 5i23 > controller. > > > > If we exercise only a single joint (Y or gantrykins slave A) with other > joint drive mechanically isolated, entirely smooth movement of the gantry is > observed (as also independent X and Z axes). > > > > However, changing to teleop (world) mode always give rise to Y and/or A > joint drive racking, and after much experimentation with .ini and .hal > parameter adjustment, we have observed what appears to be the cause of this > racking. Although it occurs more frequently at higher Y-axis velocities, it > cannot be avoided even at low speeds, particularly during acceleration / > deceleration trajectories. > > > > The issue is that the gantry slave A-joint commences motion after that of > the Y-joint primary drive; this is particularly noticeable at very small > acceleration settings (such as in attached .ini file), even a half-second > lag being observable - and (entirely smoothly) repeatable. The lag is not > noticeable at normal operational parameter settings, except that we can > attribute the racking only to Y / A instantaneous position discrepancies. > > > > Observing joint motion with very small acceleration parms, it is as if the > slave joint motion is triggered by a position delta versus the primary > joint, rather than actual (what should be expected to be concurrent) > position target command under teleop mode. > > > > I can even demonstrate a degree of damped instability between primary and > slave Y-axis gantry joints with 'appropriate' acceleration parm values. > > However, this may be attributable only to the deceleration / overshoot lag > between joints, and stiffness of the gantry giving rise to racking by the > (identical spring -loaded) stepper drives. I have observed A-Y-A racking > under certain acceleration / lag conditions. There is of course no position > feedback from the machine, the control loop being entirely within LinuxCNC. > > > > Needless to say, FERROR parms need to be large (as in the attached file) to > avoid continuous software error reporting, and the larger-than-default (for > Gecko 203V) stepper controller parms are insignificant, we believe. > > > > You will infer from the above that some tuning will be required to > critically damp the joint controllers (to balance overshoot and trajectory > time), but we need first to get to the root cause of the lag between primary > and slave joints. It's almost as if the platform is ticking over at a slow > cycle rate, but we rule out hardware performance issues. Indeed, no > difference in behaviour is observed when halving servo thread period to > 0.5ms. > > > > The .hal file is consistent I believe with LinuxCNC sample gantry .hal > configuration, so I trust someone will be able to explain the source and > remedy for the control lag observed. > > > > Incidentally, the machine is executing small jobs (0.5m square or less) > without problem under Mach-3 control (same, or Celeron platform), but we > wish to use LinuxCNC and prefer not to bridge stepper drives to allow > independent joint homing. > > > > Thank you > > Hugh Wylie > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
