#linuxcnc-devel Logs

Jan 11 2018

#linuxcnc-devel Calendar

04:45 AM skunkworks: 41us still this morning
07:32 AM JT-Shop: pcw_home: is there a connection diagram or manual for the ENCY?
07:43 AM pcw_home: there's a manual
07:45 AM pcw_home: http://freeby.mesanet.com/encyman.pdf
07:54 AM jepler: pcw_home: page 3, second table heading "description" should probably be "direction" to match the first table. Text should possibly be "are used for the" instead of "are used to the" [nitpicking in the morning]
08:06 AM pcw_home: thanks, I'll fix those
08:13 AM skunkworks: WOW - you make all sorts of stuff
08:38 AM skunkworks: I will have to remember that when I have some cheap encoders to interface
08:44 AM skunkworks: (line driver)
09:19 AM JT-Shop: pcw_home: thanks, I didn't see a link on your ENCY page
12:26 PM speedy: hi. is it a bad style (when working manually with the hog dial) to configure the jog dial's velocity via feed-override and leave the axis' GUI jog-speed out?
12:28 PM seb_kuzminsky: i believe feed override and jog speed are not related
12:28 PM seb_kuzminsky: feed override is a scaling factor applied to the F-word
12:29 PM seb_kuzminsky: jogs are not G-code, and they ignore the F-word and the feed override
12:32 PM KimK: speedy: I agree with Seb (also, most machines don't work in velocity mode to begin with). But I'd say that feed-override should be reserved for modifying F values in g-code on the fly, and so you should, IMHO, use the jog-speed to set some kind of range for the spinner. In distance mode, there's usually a .0001" or .001 mm for zeroing indicators, etc., very small movements. You'll need/want something like that.
12:37 PM speedy: the way you both describe it sounds intuitive. but i use axis.N.jog* instead of halui.jog.* in that case the halui.feed-override.value did change the jog speed. the reason was that i cannot redirect the jog-counts directly to halui.jog.
12:41 PM KimK: pcw_home: I didn't know about ENCY either, excellent! Especially since CUI/Digikey has quit selling "cable-lump" drivers for single-ended to differential conversion. Those had the advantage that they could be located very close to the encoder, but didn't have the "Y" feature (I guess you could use two), or that cool LED spinner, very handy for fast troubleshooting.
12:43 PM speedy: i mean, hopefully the feed-override slider in axis gui changes the halui.feed-override.value. cannot look at the hal configuration now. i have access to the machine only weekends.
12:46 PM KimK: speedy: OK. Well, keep working on it until you get what you want. Everyone here will try to help.
12:47 PM speedy: tnx for supporting! :)
12:48 PM speedy: since i am new to cnc milling it's not up to how i want it. i'd like to implement the component to work as it is supposed to do.
12:50 PM seb_kuzminsky: speedy: it surprises me that feed override affects the speed of axis.N.jog jogs
12:50 PM seb_kuzminsky: which version of linuxcnc are you on?
12:52 PM speedy: forked from machinekit months ago. the version file says V0.1. i mean, hopefully i did not screw up the config. i used the following tutorial as base: http://linuxcnc.org/docs/2.4/html/examples_mpg.html
12:53 PM speedy: i was unexperienced and did not know which codebase i should choose. thought the machinekit is maintained more frequent.
12:55 PM speedy: related jogpart in my config: https://github.com/rubienr/bbb-linuxcnc-config/blob/master/machinekit/configs/ARM.BeagleBone.Panther/Panther-3-axis.hal#L419
01:00 PM speedy: must have been 2.7.0~pre0 when machinekit was cloned.
01:04 PM seb_kuzminsky: speedy: so you're doing wheel jogging via the axis.N.jog-counts pin?
01:04 PM seb_kuzminsky: and you're using position-mode jogging (axis.N.jog-vel-mode is False)
01:04 PM seb_kuzminsky: ?
01:06 PM speedy: axis.N.jog-counts: yes, axis.N.jog-vel-mode: both, velocity and position mode.
01:13 PM speedy: position and velocity mode is set depending on the jog dial's user input. with halui.feed-override.increase/decrease the speed was definitely changed. it was my 2nd approach to get the axis moving from jog-counts from dial. 1st approach was halui.jog.* but the performance was poor. meaning it was jerking, not moving smooth.
01:26 PM pcw_mesa: typically you need to feed the jog value through ilowpass so you dont jump at full machine accel limits from point to point
01:27 PM pcw_mesa: ( for larger jog increments especially )
01:30 PM speedy: this could explain the sporadic following errors i face. could minimize them by lowering the acceleration and max speed, but they are still occurring. they happen when the wheel is turned so that the axis does not reach the full speed. btw the mechanical noise sounds kind of vibrato when this tends to happen.
01:33 PM speedy: but doesn't filtering contradict if somebody wants to jog exact step counts? this was my concern.
01:42 PM pcw_mesa: no, it gets to the exact point just not in such a machine gun fashion
01:45 PM pcw_mesa: this (ilowpass) is almost always done to avoid the bang-bang motion of big jog increments
01:52 PM speedy: ilowpass - sounds like a task for this weekend :)
02:09 PM speedy: what is the right way to go. connect the jog dial to axis.N.jog* or halui.jog.*. i mean halui.jog.* doesn't accept jog-counts. am i missing something?
02:41 PM KimK: speedy: In http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Hooking_Up_A_MPG_Pendant they feed the encoder counts to axis.0.jog-counts, if that helps. And I think you said you had seen http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Using_A_XHC-HB04_Wireless_MPG_Pendant ?
07:00 PM skunkworks: seb_kuzminsky, Seeing what you where seeing?
07:00 PM skunkworks: https://forum.linuxcnc.org/49-basic-configuration/33832-mesa-5i25-7i76-position-fb-running
07:01 PM skunkworks: were