#linuxcnc Logs

Sep 26 2022

#linuxcnc Calendar

12:05 AM mrec: linuxcnc is a great piece to learn ...
12:14 AM ZincBoy[CAON][m]: Yes, but simple_tp is only for single axis moves. If you want more than one axis you need to use the normal TP. That is the one that is hard to modify or switch. The thing is, if you implement this as a kinematics module, your virtual axis will use the simple_tp as it is a single axis even though it is moving two joints. There are some implications when configuring the speed and acceleration of the virtual axis so as to not
12:14 AM ZincBoy[CAON][m]: violate the limits on the other axis.
12:24 AM randy: morning
12:54 AM mrec: ZincBoy[CAON][m]: I think about injecting the jog requests into the coordinated motion pipeline - which already exists as you know
12:57 AM mrec: what you wrote is also interesting, I will have a look at that later too
12:58 AM jpa-: kinematics way looks much easier, new kinematic modules are easy to code up
12:59 AM jpa-: but on the other hand it would be cool to have a way to jog with G10 L2 rotation applied
12:59 AM mrec: well no need to G10 L2 for that
01:00 AM mrec: getting the angle in is the smallest problem :-)
04:05 AM Deejay: moin
04:42 AM Tom_L: morning
04:53 AM JT-Cave: morning
05:44 AM mrec: I more and more get the feeling that injecting the handwheel data into the coordinated motion / command queue might work
05:45 AM mrec: does anyone see any disadvantage doing so?
07:07 AM mrec: /** G / M mode flags for simple states like inch / mm, feedhold enable, etc.
07:07 AM mrec: * This stores packed bits in one field (since we can't use a bitset in a
07:07 AM mrec: * pure C struct).
07:07 AM mrec: */
07:07 AM mrec: I wonder where that comes from
07:07 AM mrec: unsigned long int packed_flags;
07:07 AM mrec: probably pre-gcc times
07:30 AM mrec: FIXME - this struct is broken into two parts... at the top are
07:30 AM mrec: structure members that I understand, and that are needed for emc2.
07:30 AM mrec: Other structure members follow. All the later ones need to be
07:30 AM mrec: evaluated - either they move up, or they go away.
07:30 AM mrec: haha
11:21 AM JT-Shop: Tom_L, I think I have it fingered out, test more when I get back from Revere
11:22 AM daight4206460[m]: Anyone have a clue why my part is inclemental in y lmao
11:23 AM * daight4206460[m] uploaded an image: (3241KiB) < https://libera.ems.host/_matrix/media/r0/download/jauriarts.org/EmWXUdyukYfHtgwJdezkPVgX/20220926_092014.jpg >
11:23 AM * daight4206460[m] uploaded an image: (2926KiB) < https://libera.ems.host/_matrix/media/r0/download/jauriarts.org/mngxnTKkMAQNNRBJDfUXvIta/20220926_092009.jpg >
11:23 AM daight4206460[m]: Thinking g91.1
11:30 AM daight4206460[m]: Using hsm works linux cnc post
12:44 PM ZincBoy[CAON][m]: Stepper based system? Missing steps on the y axis?
12:45 PM ZincBoy[CAON][m]: If it was g91.1, you would see it in the preview.
01:06 PM perry_j1987: this cell booster seems to be having a positive effect out here at the shop!
01:07 PM Tom_L: JT-Shop, great
01:11 PM Tom_L: it would likely be just as useful in the hal file
02:12 PM perry_j1987: why do i keep buying tooling for this lathe im selling
02:12 PM perry_j1987: lol
02:34 PM roycroft: hopefully the tooling you're buying will work with another lathe that you have or will have
02:38 PM perry_j1987: i just grabbed a traveling steady for the techno isel lathe off ebay
02:43 PM roycroft: oh, that won't transfer to another machine easily
02:43 PM roycroft: but it will add value to your current one, so you should be able to recover at least some of the cost when you sell it
02:43 PM perry_j1987: nope im still "reserving" the right to keep this lathe but i got it listed just incase someone nabs it heh
02:43 PM thunder-gunner[m: I really need to get a lathe, but I'm debating should I just build a long bed 4 axis cnc to go alongside the smaller one. All I'll really be doing with the lathe is threading and profiling barrels and reaming chambers, with the occasional other odd turning job
02:43 PM perry_j1987: i scored the travel rest for only 40 bucks
02:47 PM roycroft: i would buy a lathe and get some fixtures that would allow me to do milling operations on the lathe before i would buy a milling machine and attempt to do lathe operations on it
02:51 PM perry_j1987: been seeing a few listings recently of tomrachs with rapidturn lathe heads mounted on the bed
03:33 PM CaptHindsight[m]: the problem with milling on a lathe is limited travels
03:34 PM CaptHindsight[m]: I mill way more than I turn
03:34 PM CaptHindsight[m]: i turn things that spin and mill things that don't
03:35 PM CaptHindsight[m]: but live tooling and extra axis on a lathe is great
03:35 PM perry_j1987: there we go just installed the breaking resistor on this techno isel lathe
03:36 PM perry_j1987: pretty nice stuff! being able to stop the lathe quickly
03:39 PM CaptHindsight[m]: perry_j1987: DC spindle motor?
03:40 PM perry_j1987: the techno isel has that bauldor motor that i got the durapulse gs10 vfd on now
03:40 PM CaptHindsight[m]: battery charger vs braking resistors
03:42 PM perry_j1987: ?
03:56 PM perry_j1987: what do yoou mean
03:58 PM roycroft: think of a prius
03:59 PM perry_j1987: oh regenerative breaking
03:59 PM roycroft: unless you need the heat, it's better to recover the energy then to just radiate it
04:02 PM perry_j1987: prob would have to run looooong while before the resistor heats up much
04:02 PM perry_j1987: depends on the gcode as well
04:13 PM JT-Shop: Tom_L, I create a new main hal and io hal file each time now
04:14 PM CaptHindsight[m]: where did the configs go for steppers with encoders on an LPT port?
04:15 PM CaptHindsight[m]: Step/Dir + incremental encoders on LPT
04:16 PM CaptHindsight[m]: I thought Pncconf was able to as well but it does not with LPT only
04:17 PM CaptHindsight[m]: should this feature be added to Stepconf, Pncconf or some new tool? Or is it just too obscure an app to be concerned about it?
04:33 PM perry_j1987: hmm whats a good way to measure the spindle centerline height on this lathe
04:33 PM perry_j1987: i need to make a riser block for the gang tooling block to fit on there
04:37 PM Tom_L: center a known diameter dowel pin and measure to the bottom
04:37 PM perry_j1987: ah good idea
04:38 PM perry_j1987: other optioin i was just eyeballing this other chunk of aluminum and thinking about making up a new gang block so its directly in the spindle line for this lathe
04:40 PM JT-Shop2: 26% rh
04:53 PM XXCoder: wow. recyclable turbine blades
04:53 PM XXCoder: a big step
04:54 PM perry_j1987: hmm odd. unexpected realtime delay on task 0 with period 1000000. hm2/hm2_7i96.0: error finishing read! iter=8219 joint 0 following error
04:56 PM JT-Shop2: you running a 5.x kernel?
04:56 PM JT-Shop2: with a mesa ethernet card?
04:56 PM perry_j1987: uname -r says 4.19.0-21-rt-amd64
04:59 PM JT-Shop2: hmm thats odd
04:59 PM JT-Shop2: intel nic?
05:00 PM perry_j1987: ya
05:00 PM JT-Shop2: sec
05:00 PM perry_j1987: ga350n motherboard
05:00 PM JT-Shop2: https://gnipsel.com/linuxcnc/debian-10-eth.html
05:00 PM perry_j1987: with a pci intel nic
05:01 PM JT-Shop2: middle of that page tells you how to fix intel nic
05:05 PM * JT-Shop2 goes to make some chicken crack
05:08 PM JT-Cave: Tom_L, https://pastebin.com/RCTEw7zw
05:08 PM * Tom_L puts his glasses on
05:09 PM Tom_L: another test?
05:10 PM perry_j1987: rebooting now lets see if that does the trick
05:10 PM Tom_L: it will do _some_ trick
05:13 PM perry_j1987: hmmm no dice
05:14 PM Tom_L: did you edit the intel chip settings?
05:15 PM perry_j1987: ya added that line
05:19 PM perry_j1987: let me try the realtek onboard nic
05:24 PM perry_j1987: hmm how is this having issues all of asudden lol
05:32 PM JT-Cave: Tom_L, if you want to
05:32 PM Tom_L: i am
05:32 PM perry_j1987: switched ehternet to the realtek onboard and configred its ip
05:32 PM perry_j1987: launch axis and it opens
05:33 PM Tom_L: all the constants = Changed
05:33 PM perry_j1987: if i try to jog an axis then it gives the joint following error
05:34 PM Tom_L: one thing..
05:34 PM Tom_L: it gets confused on one of them
05:35 PM JT-Shop2: how so?
05:35 PM -!- #linuxcnc mode set to +v by ChanServ
05:35 PM JT-Shop2: all the values get changed to Changed
05:35 PM JT-Shop2: it's just test code
05:35 PM Tom_L: i know
05:36 PM Tom_dev: [KINS]
05:36 PM Tom_dev: KINEMATICS = trivkins coordinates=XYZA
05:36 PM Tom_dev: JOINTS = 4
05:36 PM Tom_dev: [TESTKINS1]
05:36 PM Tom_dev: code = 0
05:36 PM Tom_dev: files = 6
05:36 PM Tom_dev: # comment5
05:36 PM Tom_dev: wait this doesn't show blank lines
05:36 PM Tom_dev: i'll paste it
05:37 PM JT-Shop2: ok
05:39 PM Tom_dev: i screwed up the test.. hang on
05:40 PM JT-Shop: lol ok
05:41 PM Tom_dev: http://paste.debian.net/1255100/
05:41 PM Tom_dev: it gets scrambled there
05:42 PM Tom_dev: i tried various line spacing between
05:42 PM Tom_dev: the rest look ok
05:42 PM Tom_dev: i'll post the whole test section now with the result
05:43 PM JT-Shop: before and after if you can
05:43 PM Tom_dev: 2 files or one ok?
05:43 PM JT-Shop: 2 would be better
05:43 PM Tom_dev: i'll make 2 separate files, a before and after
05:43 PM JT-Shop: ok
05:47 PM Tom_dev: http://paste.debian.net/1255101/
05:47 PM Tom_dev: http://paste.debian.net/1255102/
05:48 PM Tom_dev: comment 4 lost a line above it
05:48 PM Tom_dev: comment 5 is out of order
05:49 PM JT-Shop: that will screw up the order so more work to be done
05:49 PM Tom_dev: it _could_ happen that way is why i tested it
05:49 PM JT-Shop: your test helps make it good
05:49 PM JT-Shop: extra space after comment 1 too
05:50 PM JT-Shop: close but no cigar...
05:52 PM Tom_dev: i think the extra line on 4 is what starts the chain reaction
05:53 PM JT-Shop: hmm I better go make a copy of the posts they expire in 24 hours I think
05:53 PM Tom_dev: i can reproduce them if needed
05:53 PM Tom_dev: i saved the before file
05:54 PM JT-Shop: looks like it got cornfused on line 9
05:56 PM Tom_dev: let me run it one more time to make sure i did it right
05:56 PM JT-Shop: ok
05:58 PM Tom_dev: http://paste.debian.net/1255103/
05:58 PM Tom_dev: original file
05:58 PM Tom_dev: http://paste.debian.net/1255104/
05:59 PM Tom_dev: after test
05:59 PM JT-Shop: more work to be done for sure but getting close
06:00 PM Tom_dev: yeah line 9 lost a blank line
06:00 PM JT-Shop: 9 gained a blank line...
06:00 PM Tom_dev: right..
06:01 PM Tom_dev: sry
06:01 PM Tom_dev: i was looking at the original :)
06:01 PM JT-Shop: I got your back...
06:01 PM JT-Shop: time to wander inside and relax
06:01 PM Tom_dev: later..
06:02 PM JT-Shop: 4 hours diagnosing a leak checker to finger out a wire is loose or broke somewhere... tomorrow to find and fix it
06:02 PM Tom_L: revere?
06:02 PM JT-Shop: yep a Briggs gas tank leak tester
06:03 PM JT-Shop: old machine that's been upgraded a couple of times
06:05 PM JT-Shop: see you in the morning
06:06 PM Tom_L: looking at the code, i never see it print 'blank'
06:06 PM Tom_L: oh wait.. nevermind
06:06 PM -!- #linuxcnc mode set to +v by ChanServ
06:06 PM Tom_dev: tom@bookworm:~/linuxcnc/configs/test_ini$ python3 parse.py
06:06 PM Tom_dev: blank
06:06 PM Tom_dev: blank
06:06 PM Tom_dev: blank
06:07 PM Tom_dev: blank
06:07 PM Tom_dev: blank
06:07 PM Tom_dev: blank
06:07 PM Tom_dev: blank
06:07 PM Tom_dev: blank
06:07 PM Tom_dev: 7 #comment1
06:07 PM Tom_dev: 8
06:07 PM Tom_dev: 13 #comment2
06:07 PM mrec: actually injecting the jog positions into the command loop for coordinated motion seems to work I'm getting a perfect 45° angle
06:07 PM mrec: but I still have lots of bugs in my experimental code
06:07 PM mrec: the feedback loop & co I did not take care about yet
06:07 PM Tom_L: JT-Shop, do the commands show as blank too
06:08 PM mrec: so simulating the compound on a lathe seems to be reasonable
06:08 PM mrec: and it doesn't seem to be so complex actually
06:09 PM Tom_L: mm they must
06:40 PM CaptHindsight[m]: where is the latest 2.9 Deb package?
06:42 PM Tom_L: buildbot is dated 09/25
06:43 PM Tom_L: http://buildbot.linuxcnc.org/dists/buster/master-rtpreempt/binary-amd64/
06:44 PM CaptHindsight[m]: thanks, daily builds
06:44 PM Tom_L: looks like 3 that day
06:46 PM CaptHindsight[m]: no Bullseye or Bookworm 2.9 packages?
06:47 PM CaptHindsight[m]: https://www.youtube.com/watch?v=cVWF3u-y-Zg for more better remote machining
06:48 PM Tom_L: bookworm will install from package manager i think
06:48 PM Tom_L: but not bullseye
06:48 PM CaptHindsight[m]: 2.8 for bookworm?
06:48 PM Tom_L: 2.9
06:48 PM Tom_L: pre
06:49 PM Tom_L: https://www.gnipsel.com/linuxcnc/debian-12-emc.html
06:49 PM CaptHindsight[m]: oh is 2.9 pre on Debi servers?
06:49 PM Tom_L: i think so
06:49 PM Tom_L: just not for bullseye
06:49 PM Tom_L: that i know of unless it changed
06:49 PM CaptHindsight[m]: how recent is that 2.9 pre?
06:50 PM Tom_L: if you want the very latest i'd build it but on their site, i dunno
06:50 PM CaptHindsight[m]: yeah I've been building, but i was wondering about a package
06:50 PM Tom_L: i'd guess the buildbot to be the most recent
06:50 PM CaptHindsight[m]: ok, thanks
06:50 PM Tom_L: i dunno if debian does daily builds or not
06:51 PM Tom_L: pere would probably know
06:52 PM CaptHindsight[m]: I needed the latest 2.9 with the encoder fix
06:52 PM Tom_L: for mesaflash, you have to build to get the latest
06:52 PM CaptHindsight[m]: I'm encodering away now on the LPT port
06:52 PM Tom_L: for 7i96s etc
06:52 PM Tom_L: nice
06:54 PM Tom_L: someone had a deb for bullseye but i don't have the link
06:56 PM Tom_L: http://linuxcnc.org/2022/03/03/LinuxCNC-in-Debian/
07:01 PM CaptHindsight[m]: qtpyvcp had some
07:01 PM CaptHindsight[m]: but now they link to the buildbot
07:05 PM mrec: Tom_L: do you know when the patches for 7i96s were added?
07:06 PM mrec: my 2.9 installation is a bit "old" already
07:14 PM CaptHindsight[m]: current max speed is 18m/minute with 1K pulse encoder
07:15 PM CaptHindsight[m]: on LPT port
07:16 PM CaptHindsight[m]: 8:1 ratio of step travel to encoder count length
07:46 PM daight4206460[m]: Servo machine
08:53 PM Alicrow-phone: I hooked up the old rotary encoder for the y axis, and it seems I get a quicker response from the motors on that vs the linear encoder I have set up. About 10ms to get a positional change from the linear scale, vs about 5ms for the rotary encoder.
08:58 PM Alicrow-phone: And the rotary encoder is actually slightly less sensitive than the linear scale. (5080 counts per inch for linear scale, vs 4000 counts per inch for the rotary encoder)
09:22 PM srk- is now known as srk
10:20 PM CaptHindsight[m]: sounds like magic
10:30 PM ZincBoy[CAON][m]: What type of encoders? optical or magnetic?
10:57 PM CaptHindsight[m]: Alicrow-phone: how are you measuring the latency between encoder movement and a pulse change?
11:17 PM mrec: ZincBoy[CAON][m]: do you see any problem with injecting the motion requests via tpAddLine and setting the motion state to EMCMOT_MOTION_COORD?
11:17 PM mrec: I still have problems with syncing the position eg. syncing between keyboard and handwheel jogs (the motion itself looks good)
11:40 PM boozeman[m]: CaptainHindsight (@_discord_702231000050040894:jauriarts.org) Bookworm has linuxcnc-uspace (2.9.0~pre0+git20220906.02744cdef6-1. So it is rather fresh 🙂