#linuxcnc-devel Logs

Sep 06 2022

02:00 AM hansu: seb_kuzminsky: The most recent commit on 2.8 is a42959fc29537dbec8a82da34186f9dcf95ec4b5 from Sep 04. So I wonder why this is not build by the buildbot.
02:07 AM pere: hansu: hi. :)
02:07 AM hansu: pere: Hi
02:52 AM hansu: Another stupid question. When I configure to build only docs - is it necessary to set '--with-realtime=uspace' or is this just something missing on my system?
03:04 AM pere: hansu: as far as I know it is the default. do you mean running debian/configure when you say configure? the uspace option is not used there.
03:05 AM pere: hansu: checked configure.ac, and it look like configure will try to detect the realtime mechanism to use if the --with-realtime=uspace option is not used.
03:06 AM hansu: the configure script in src
03:06 AM hansu: So it needs information about the realtime kernel even if I only want to build docs?
03:10 AM hansu: pere: can you also have a look at https://github.com/LinuxCNC/linuxcnc/pull/1646#issuecomment-1237805174 ? Thanks :-)
03:21 AM pere: hansu: had a look, and agree that it might be a race condition / inaccurate make dependecy relations, but have no idea where that would be.
03:21 AM pere: perhaps andypugh have an idea?
03:32 AM hansu: Okay thanks for looking :-)
03:33 AM hansu: pere: What is going on with your Mazak? Did you get any further with it?
03:34 AM pere: hansu: have not done much since the summer holiday started, but got the axis moving a bit better (pid tuning issue) and and am currently wonering how to get the tool changer carousel to stop at the correct location.
03:35 AM pere: still have not found a solution for the spindle motor issue.
03:36 AM hansu: I thought you could drive to a specific tool position?
03:36 AM hansu: Have you looked up the motor type (on the type label)
03:37 AM pere: I can, but is 2-3 mm off...
03:39 AM hansu: Oh hmm
03:39 AM pere: I got some info from the motor label, but some of it is hidden. It is a " Mitsubishi SJ 9 A"
03:43 AM hansu: I quick search gives me the info that it is an induction motor. That makes it much easier to find an inverter :-)
03:45 AM hansu: I have an inverter home in Germany, which should at least be powerful enough to test the spindle motor. But unfortunately not here :-/
03:48 AM pere: how powerful is powerful enough? I got two vfd boxes, but do not know how to get them working with this motor.
03:52 AM hansu: If I have it correct in mind it's a 7.5 kW 15 A (18 A) VFD. So should be enough even for normal operating conditions. A very nice VFD which also supports encoder feedback.
03:52 AM hansu: What VFDs you have?
03:53 AM pere: aha. Mine are 4.5kW max.
03:53 AM pere: one 4.5 kw and one 1.5kW.
03:54 AM hansu: For testing without load should work as well
03:58 AM pere: I agree, but I could not figure out how to configure the 4.5kW VFD, as it has a horrible 4 letter display and I lack documentation.
03:59 AM pere: the 1.5kW VFD is currently in the wrong location and I have not had time to test it since the summer holiday
03:59 AM hansu: Hmm without documentation it's hard to parametrize the VFD, I agree :(
04:03 AM pere: it has some auto-tuning mechanism I managed to invoke, but it is very unhappy with the motor I connect. :)
04:14 AM hansu: Yeah it's easier not to use auto tuning. In normal U/f mode it's quite uncomplicated to set up.
04:15 AM pere: I could not find a way to start it (lacking doc), and the only way I could get something happening was by enabling tuning, but it show an over current error. :(
04:43 AM hansu: Yeah even for auto tuning you have to set the motor parameters.
12:48 PM CaptHindsight[m]: pere: why sometimes paying the extra 10-30% more for a name brand VFD is worth it just for the docs and predictable behavior
02:29 PM -!- #linuxcnc-devel mode set to +v by ChanServ
04:09 PM CaptHindsight[m]: how LCNC gets LPT address and type info from the kernel isn't working properly
04:10 PM CaptHindsight[m]: on some LPT controllers I can only get inputs but not outputs working
04:10 PM CaptHindsight[m]: some LPT input signals make it into LCNC HAL, but don't make it to the GUI
04:38 PM linuxcnc-build: build #9178 of 0000.checkin is complete: Failure [4failed fetch branch to local git repo] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/9178 blamelist: Sebastian Kuzminsky <seb@highlab.com>
04:49 PM andypugh: CaptHindsight[m]: I don’t see how LPT signals (but not other signals) could get as far as HAL and then get lost.
04:49 PM andypugh: Once in HAL the source is unknown,
06:46 PM seb_kuzminsky: hansu: that commit failed the tests, so the buildbot chose not to package it
06:47 PM seb_kuzminsky: http://buildbot.linuxcnc.org/buildbot/builders/1401.rip-wheezy-rtai-i386/builds/6975
06:47 PM seb_kuzminsky: http://buildbot.linuxcnc.org/buildbot/builders/1400.rip-wheezy-i386/builds/7299
06:47 PM seb_kuzminsky: http://buildbot.linuxcnc.org/buildbot/builders/1640.rip-buster-rtpreempt-amd64/builds/2768
06:47 PM seb_kuzminsky: http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/6824
06:49 PM seb_kuzminsky: i don't think your gmoccapy change could have caused those test failures, so it's probably an infrastructure problem :-(