#linuxcnc-devel Logs

Mar 02 2017

08:27 AM stustev: cradek: If the MV: was changed to Dry Run: then the control would act/look like many previous controls.
08:29 AM stustev: I have looked (I think) everywhere to find how to change the MV: to Dry Run:. I even reinstalled Wheezy and cloned master and am still unable to get the label changed.
08:30 AM stustev: I grep MV: and change every glade file one at a time followed by a recompile every time. UGH! NOT CLEAR!! - still looking :)
09:02 AM cradek: stustev: did you try a sharpie?
09:03 AM cradek: src/emc/usr_intf/touchy/touchy.glade: <property name="label" translatable="yes">MV: 100</property>
09:03 AM cradek: src/emc/usr_intf/touchy/touchy.py: set_label(self.wTree.get_widget("mv").child, "MV: %d" % self.mv_val)
09:03 AM cradek: seems like you'd need to change these two and rebuild
09:06 AM stustev: ok - I changed the first. My search would not find the second - I will try that - thanks
09:09 AM stustev: HEH - just saw it - no, I did not try a sharpie. Does green work better?
09:10 AM stustev: changing those two files worked - thank you very much
09:41 AM cradek: I'm glad the fix for your user is simple, but is it ok with you if I find the whole affair a bit laughable?
09:42 AM cradek: do you have continuous jog switches? If so that value affects them too
09:44 AM stustev: yes it sure is - the whining has not stopped - I am adding a button for rapid override right now (more complicated) - they NEED that also as the Haas controls have that.
09:44 AM cradek: so many overrides
09:45 AM stustev: the feed rate of the continuous buttons?
09:45 AM cradek: I support adding a RO% button that works just like the FO% button - it would be great if you isolate that addition in a git commit so we can incorporate it
09:46 AM cradek: yes with touchy the continuous jog buttons move at the MV setting
09:46 AM cradek: you can even change MV during a continuous jog and it adjusts the current jog (so awesome!)
09:46 AM stustev: I have it working on the touchy screen. Looks good. I am now working to see how to make it function on the machine.
09:47 AM stustev: The MV(dry run) works the same on the other controls. Just the name is different.
09:47 AM cradek: making RO will be quite a few changes in touchy
09:49 AM stustev: I have made probably a dozen changes now in (I think) three place in the touchy.glade file. I have it pointed at self.linuxcnc.rapid_override(self.ro_val)
09:50 AM stustev: now I need to find just where that points or if I need to change the 'rapid' to something else
09:50 AM cradek: sounds like the right track to me
09:51 AM stustev: isolating a git commit is just another learning opportunity :)
09:51 AM cradek: do you think you'll become a developer or I'll become a machinist first?
09:52 AM stustev: I think I am just developing as a machinist
09:52 AM cradek: last night I fired up the jr and didn't have the valve on the air compressor open, but there was enough air pressure in the lines to come out of estop and get halfway through a tool change. yuck.
09:53 AM cradek: first time I'd made that mistake
09:55 AM stustev: was it difficult to recover?
09:57 AM cradek: after much hesitation, I just ran M6 again and it recovered, but then the tool table was wrong, so I used the manual unclamp button and a couple more M6 to get it back in the right pocket
09:57 AM cradek: it ended up with T6 and T0 (the no tool) swapped in their pockets after the failure
09:58 AM cradek: so I loaded T0 and popped it out, then loaded T6 and popped it back in
09:58 AM cradek: I must've been thinking about recovery when I wrote the ladder, because it worked fine - of course I don't remember
09:59 AM cradek: each step has a switch that shows when it's done (pot is down all the way, or up all the way, or the arm is back home) so it's pretty easy to get right
10:03 AM stustev: My tool change component for the enshu does the same thing. If the tool change doesn't complete for some reason you just restart the control. On restart all the switches are read and the tool change is positioned properly to restart. You then must make sure the tool are in the correct locations.
10:04 AM stustev: You are currently a much better machinist than I am a developer. :)
10:06 AM * cradek counts his fingers
10:06 AM cradek: yep guess so
10:06 AM cradek: any machining job you can walk away from was a good machining job
10:06 AM cradek: er wait that's landings
10:07 AM cradek: sorry, I had to come to work early to fix the internets, I'm a bit loopy
10:08 AM stustev: I am sure glad you can fix the internets. It was slow here until you got to work today.
10:12 AM cradek: stustev: I've been working on the bus again, getting ready for some trips this spring I hope
10:12 AM cradek: can't wait. got new tires so I better get started using them up.
10:17 AM stustev: tires, batteries and fuel all have a limited life and must be used before they are thrown away
10:18 AM cradek: yep, got new batteries too. had to jump start it from a prius. the last ones lasted 10 years!
10:19 AM stustev: 10 years is a very good life for batteries - do you have a maintaining charger on them?
10:19 AM cradek: yes, the solar panels keep them full
10:20 AM stustev: that is very necessary for max battery life
10:20 AM cradek: the prius could (barely) start the generator, the generator and prius together could (barely) start the main engine
10:20 AM cradek: nice to know it can be jumped
10:21 AM stustev: GOOD jumper cables and charging time are the key if the main batteries have any life left
10:22 AM cradek: yeah mine are only 5' long and super thick wire and hard springs on the clamps
10:22 AM cradek: I plan to be buried with them
10:23 AM cradek: nobody has ever said "jeez I wish I had bought $20-cheaper jumper cables"
10:23 AM stustev: I have a couple good sets. I keep a close eye on them.
10:24 AM cradek: yep, gotta watch that kind of stuff
10:25 AM cradek: also the set of wrenches where they're all still there
10:26 AM stustev: wrenches are much more difficult to watch - my solution is to get much more wrenches than you need
10:27 AM stustev: the 'lost' wrenches seem to return but when they are lost I don't want to search for them
10:28 AM cradek: oh look the 1/2 is missing from this set too
10:29 AM cradek: stab stab
11:00 AM stustev: there are a few wrenches that are missing more often - I have piles of wrenches and lots of unmatched sets to try to combat that
11:14 AM seb_kuzminsky: thanks pcw_mesa
11:27 AM cradek: I love that format of I did this, I expected this to happen, I see this instead
11:27 AM cradek: it's so useful
01:10 PM CaptHindsight: https://forum.linuxcnc.org/18-computer/32170-latency-about-7000-with-latency-test-raises-to-above-100-000-with-latency-plot#85923 I noticed this post and had to do some recent testing myself
01:11 PM CaptHindsight: I get very similar results with latency-plot and -histogram, latency-test gives me and occasional jump (sometimes 10x the jitter) over the course of several hours
01:12 PM CaptHindsight: and/an
01:12 PM CaptHindsight: I haven't looked at the source of either
01:14 PM CaptHindsight: have only tested this on one board for 2 days using the latest RTAI kernel in 2.7.8
01:16 PM CaptHindsight: so I get the opposite results from what they posted
01:17 PM CaptHindsight: their latency-test had lower results than latency-plot
05:34 PM andypugh: I feel slightly soiled by this solution. Is there a better way? https://forum.linuxcnc.org/10-advanced-configuration/32412-bug-in-current-pocket#88904
05:40 PM andypugh: pcw_home: Purely out of curiosity, what does W3 do on the 8i20? It isn’t mentioned in the manual
05:44 PM pcw_mesa: I dont think it does anything
05:45 PM pcw_mesa: its just an option on the low side power supply PIC processor but its ignored
06:19 PM andypugh: Right. I just noticed it as I hooked up exernal logic power to my drive box. It took all might, mainly because i was determined to use some of my collection of second-hand Lemo connectors.
11:39 PM cradek: andypugh: not sure what possible answer there is except that fixing the actual bug would be the better way
11:39 PM seb_kuzminsky: i'm looking at that bug right now
11:40 PM cradek: sweet
11:40 PM seb_kuzminsky: it's tricky
11:41 PM seb_kuzminsky: the pocket number from the tool table file is not the pocket number that interp uses to think about the tool
11:41 PM cradek: except sometimes (random)?
11:41 PM seb_kuzminsky: not even then
11:41 PM seb_kuzminsky: interp compactifies the tool table, to make all the pockets sequential starting from 1
11:42 PM cradek: ah, so it's not even the problem that's everywhere else
11:42 PM seb_kuzminsky: :-/
11:42 PM cradek: oh good grief
11:42 PM cradek: I'm sorry I made that all even worse with random
11:43 PM cradek: but also I'm happy my machine works
11:43 PM cradek: sigh
11:43 PM seb_kuzminsky: yeah it's nice that it works :-)
11:44 PM cradek: sleeeep
11:44 PM cradek: goodnight!
11:44 PM seb_kuzminsky: how was music night?
11:45 PM cradek: it was the very best ever
11:45 PM seb_kuzminsky: yay!
11:45 PM cradek: joyous
11:45 PM seb_kuzminsky: sleep well :-)
11:45 PM cradek: 'night
11:46 PM seb_kuzminsky: yeah, io knows the real pocket number, but interp does not