On Wednesday 27 December 2017 22:19:33 Gene Heskett wrote: Ping, and a few lines added at the bottom.
> Greetings all; > > I first posted this to the devel list earlier this afternoon but may > another user has encountered this, so a repost to the users list, with > a couple typo's fixed. > > Axis LINUXCNC - 2.8.0-pre1-3476-g03ffa4b > > I have disconnected the joint.n.vel.derivitives from the > pid.velocity-deriv pins because of the huge spike generated in motion > as it crosses the zero vel line when the direction is being reversed > that was subjecting the stepper drivers to pretty brutal, following > error causing jumps. > > Jogging the machine around before homeing is IMO how the machine > should move all the time. > > Because one thing I just noted, is about a 100/1 difference in the > acceleration times, controlled by whether or not its been homed. > > If I keyboard jog it before its homed, the accel/decel times are in > milliseconds to default speeds in the 54 ipm range. I can make it > dance from the keyboard or jiggle stuff off the table. > > Once homed, that same jog has an accel/decel time of a very slightly > fat full second up to the same speed and the stepper(s) sound > resembles the rise and fall of a motor driven fire siren only somewhat > faster. And this lag shows in the halmeter looking at the joint.n.vel > command. Z if homed, takes most of an inch to get up to top speed, and > most of an inch to get stopped. And a full second of flashing numbers > between 0 and the .9 vel limits, eg 54 ipm. > > This effects all 4 axis's, and is not effected by changes in the TRAJ > section DEFAULTS or MAX's of the ini file. I just put values there > copied from the machine units/minute values from one on the axis's. > > Doubleing the MAX_ACCEL of an axis, and redoubleing that for > STEPGEN_MAX_ACCEL also does not have a visible effect after homing. > Before homing, it sharpens the accel bumps noticeably, but despite > that I have not triggered a following error in half an hours putzing > with this. > > What can cause this extremely slow accel responses to the jog keys > once homed?, and which I assume apply to running code, because running > code and cutting air its obvious the machine is suffering. Running > down at a G38.2 F10 z-4 expecting a contact stop at -1, would make > contact ok, but the machine might coast another 1/4" before it got > stopped, smashing the pcb contact pad and likely breaking the $20 > tool. > > Ideas checked, but I've searched for something that might couple the > homed state outside of its expected operation, without finding a > thing. I did find some stuff missing in the [TRAJ] section when it was > compared to the docs, but replaceing the missing stuff, some of which > is possibly related, had no effect. > > So now I need a better list of stuff to check, or I've found a bug. > > Do I need to put the main .ini and .hal files up on my web page? I've found some stuff that I think pncconf put in my hal file back in mid 2015 that I had commented out a week back because the signals named had no source of the signal, discovered while surveying the files for other loose ends. I just re-enabled those lines but will have to go out and turn the power switch back on to test, which since I've had a cold that I'm trying not to expose Dee to with her near terminal copd, so except for feeding her, and servicing her potty chair, I've been hibernating in my end of the house for the last 2-3 days. > Thanks all. I'll echo that again, thank you all. Cheers, Gene Heskett -- "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Genes Web page <http://geneslinuxbox.net:6309/gene> ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
