#linuxcnc Logs
Sep 10 2024
#linuxcnc Calendar
01:01 AM lcnc-relay: <meisterdippel> moin
04:05 AM JT-Cave: morning
04:39 AM Tom_L: morning
05:13 AM lcnc-relay: <Travis Farmer> Morning
06:10 AM JT-Cave: Tom_L, https://ibb.co/3vKBC4N
06:21 AM lcnc-relay: <skunkworks8841> https://youtube.com/shorts/lutj2HTfNyo?si=rMxr8Y8OpPauZGAZ
06:33 AM * JT-Cave brain was about to explode this morning but finally fingered it out
06:51 AM TurboTechnique: so I bought some parts, a 700x500mm gantry table, some nema23 stepper motors, an oversized spindle with VFD and a Mesa ethernet board to control the lot
06:52 AM TurboTechnique: What I'm wondering now, is there an optimum microstepping ratio?
06:53 AM rdtsc-wk: JT, the resources/import?
06:53 AM TurboTechnique: There's obviously no point in running with a much finer step granularity than the finesse of the machine allows
06:55 AM rdtsc-wk: Cool Turbo. Microstepping is mostly a trade-off: speed for precision. It also makes it slightly quieter. If machine must be fast, use little microstepping.
06:56 AM rdtsc-wk: Mesa Ethernet is fast enough to run full-speed with microstepping, but if you'd been using a parallel port break-out board, that has serious limitations with microstepping, so no worries there.
06:58 AM TurboTechnique: I remember when I converted my old machine that's been running linuxcnc since 2010 from parallel port to mesa ethernet, it changed the sound of the steppers significantly
06:58 AM TurboTechnique: only realised then how much jitter the parallel port really had
06:59 AM TurboTechnique: so I decided to build the new machine with familiar components
06:59 AM rdtsc-wk: I inherited KL3040 drivers... not impressed. Got some Gecko G251X drivers - night-and-day difference. Very excited about those.
07:01 AM JT-Cave: rdtsc-wk, yup
07:02 AM JT-Cave: one of the things I like about gecko drives is they morph from 10 micro steps to full steps as the speed goes up
07:03 AM rdtsc-wk: JT, so Tom was trying to use a "resource" inside a Qt .xml? Yeah that sounds like a brain-bender.
07:06 AM TurboTechnique: I don't need super high speed, 1200-1500mm/min rapids would be super
07:07 AM TurboTechnique: with ball screws
07:07 AM JT-Cave: I _think_ Tom was trying anything and everything yesterday
07:13 AM rdtsc-wk: Only thing tough about the G251X is the terminals are 3.0mm - very tiny, and too small for "normal" wire sizes. Think my motor wires are 12ga - had to peel off some strands to get them to fit for testing. Final version will use ferrules, but I think only the smallest (grey or white) would fit, requiring even thinner wire. The drive is the size of a matchbook, but can deliver 3.5A; crazy.
07:23 AM JT-Cave: I used the 203V drives on my plasma cutter
07:24 AM JT-Cave: time to ride
07:24 AM rdtsc-wk: Have a good one
07:42 AM TurboTechnique: I have always seen stepper drives as commodity items, never gotten a good education on why to choose one drive over another, because on the label they all do the same thing
07:42 AM TurboTechnique: But being an electrical engineer I know that more speed requires higher voltage to be able to maintain torque at speed
07:44 AM rdtsc-wk: Turbo, testing the KL-4030s had motors about to jump off the bench, also loud and hot; Geckos = very smooth, quiet and cool (with far more torque.) I'll stick to Geckos.
07:44 AM TurboTechnique: interesting
07:45 AM rdtsc-wk: Sliced my finger trying to stop the Gecko-driven motor, heh
07:46 AM rdtsc-wk: who knew a lovejoy sticker could be dangerous? ;)
07:48 AM * JT-Shop heads for wolf creek for a ride
07:52 AM mrec: how do you guys handle coolant? by yourself or through an external company?
07:53 AM -!- #linuxcnc mode set to +v by ChanServ
08:03 AM mrec: now in my setup I need some coolant which will be recycled so I'm a bit concerned about that.
08:03 AM mrec: the current coolant in the machine is around 4 years old and smells wonderful.
08:03 AM Tom_dev: JT-Auto-Shop, i started fresh and went thru the docs and get this: Could not create pixmap from ./images/inX-Y+.png
08:03 AM mrec: (well not used for 4 years, not sure how long it was inside before that)
08:03 AM Tom_dev: it isn't reading the resources.py file
08:03 AM Tom_dev: it's still reading them from the images folder
08:03 AM Tom_dev: i know because i renamed the images to images1
08:03 AM Tom_dev: lemme double check
08:06 AM lcnc-relay: <captainhindsight_.> Turbo https://www.forum.linuxcnc.org/24-hal-components/1492-tool-changer-m06-control-for-carousel-system
08:07 AM lcnc-relay: <captainhindsight_.> https://forum.linuxcnc.org/10-advanced-configuration/41504-m6-remap-for-atc-with-carousel
08:08 AM lcnc-relay: <captainhindsight_.> good night zzzzzzzzzzzzz..........
08:11 AM TurboTechnique: thanks
08:13 AM Tom_L: JT-Auto-Shop, in the docs where you show <file>inX+.png</file>, one is that way then just below that is: <file>images/inX+.png</file>
08:13 AM Tom_L: which is correct?
08:24 AM Tom_L: it depends if you create the qrc file in the working directory or images directory
08:25 AM Tom_L: i still think it's reading from the images folder and not the resources.py file but i'll check again when i have more time
08:26 AM Tom_L: try renaming your images folder and see if they go away
09:49 AM lcnc-relay: <meisterdippel> https://jauriarts.org/_matrix/media/v3/download/jauriarts.org/cjGaJRNUwjxFDiOLXcGhiZNQ/IMG_20240910_164124.jpg
09:49 AM lcnc-relay: <meisterdippel> new toy 🙂 lets play
10:14 AM JT-Cave: Tom_L, I think you found something, I didn't even import resources.py and it finds the images
10:16 AM JT-Cave: don't even need the resources.qrc just put the images in a sub-directory and use that path in the qss file...
10:44 AM rdtsc-wk: Nice, meister. Have fun :)
11:07 AM lcnc-relay: <meisterdippel> thanks 🙂
11:34 AM JT-Cave: so the mityvac brake bleeder just sucks air when you open the bleeder valve so my thought is to get a bleeder and drill a small hole through the center and tighten it up so it can't suck air
11:36 AM JT-Cave: or I could make a sorta pressure bleeder\
11:39 AM lcnc-relay: <Travis Farmer> teflon tape?
11:40 AM JT-Cave: I think that would still leak as the threads are straight... maybe sprayed on sealant?
11:41 AM lcnc-relay: <Travis Farmer> gravity bleeding works, if you have the time
11:42 AM lcnc-relay: <Travis Farmer> i have pressure bleed before, too
11:43 AM JT-Cave: I made a pressure bleeder for my chevy truck with a screw on cap
11:43 AM roycroft: i have a brake bleeder that is like a lawn sprayer - you fill the reservoir partially with brake fluid, put the top on, pump it up to pressurize it, then attach it to the master cylinder (it has a large assortment of gaskets/clamps/lids to fit almost anything)
11:43 AM roycroft: it works really great
11:44 AM JT-Cave: I read a bunch of 1 star reviews on the motive products power brake bleeder
11:44 AM * roycroft has no idea where the thing is now, though - it's in one of dozens of boxes in one of his storage units
11:45 AM JT-Cave: of course they could be written by people that shouldn't attempt to bleed brakes too
11:45 AM JT-Cave: never can tell
11:45 AM lcnc-relay: <Travis Farmer> lol
11:46 AM JT-Cave: they have a "speed bleeder" that's a bleed valve with a check valve in it
11:46 AM JT-Cave: but the same thing could be done with a bleeder with a hole through and the hose in a container with brake fluid I suppose
11:47 AM Tom_L: JT-Cave, the subdirectory path is what i was using prior to this. i was hoping to get away with that with the resources.py file but it doesn't seem to be working yet
11:48 AM Tom_L: the resources.py i think is supposed to imbed them with the application
11:49 AM Tom_L: so you don't need the sub files
11:50 AM Tom_L: with the subdirectory files you can either link them with the qss file or the stylesheet per button widget
11:50 AM Tom_L: that's my findings so far anyway
11:54 AM Tom_L: it's certainly easier to just use the subdirectory method
11:54 AM rdtsc-wk: There may be a way to "in-line" a file as data bytes within Python, but that is only valid for the running Python code. It's not going to help for an .xml or .qss file, as Python doesn't execute those.
11:54 AM rdtsc-wk: Prolly easiest to put them into a folder (such as /images) and link to them from the .qss. Will relative links work, such as "../images/X+Y-.png"?
11:55 AM Tom_L: as long as the path is valid to the working directory
11:55 AM rdtsc-wk: that may be the easiest method :)
11:56 AM Tom_L: that's what i've been doing up til now but wanted to see if the resources.py method would work
11:56 AM Tom_L: lunch, then maybe i'll mess with it some
11:56 AM rdtsc-wk: I don't think it can jump-the-gap to .xml or .qss - what do you think JT?
12:04 PM Tom_L: rcc converts it to a py file
12:04 PM Tom_L: but i think we're still missing a step somewhere
01:00 PM he1 is now known as hel
01:08 PM Tom_L: JT-Cave, busy?
01:10 PM -!- #linuxcnc mode set to +v by ChanServ
01:11 PM Tom_dev: an alternate to adding the images to the qss file: http://paste.debian.net/1328996/
01:11 PM Tom_L: both methods work
01:18 PM JT-Shop: fixing to head to town
01:18 PM JT-Shop: ran out of brake fluid...
01:22 PM JT-Shop: and gotta build a pressure bleeder as there's not enough difference in height from the mc to caliper to gravity drain
01:32 PM Tom_L: fun project ehh
01:32 PM Tom_L: jack the front of the car up :)
01:33 PM bjorkintosh: ramp it up?
01:33 PM Tom_L: we always had 2, one to push the pedal down and one to open/close the bleeder and keep the reservoir full
01:35 PM Tom_L: nap.
01:49 PM NetPipe: https://www.amazon.ca/dp/B07KDFD3WP deal on this stuff right now
02:04 PM rdtsc-wk: hope the lines aren't gunked up
02:19 PM rdtsc-wk: Once got an old CB360 running with a frozen front brake cylinder, and took it around the block for a quick spin. No drag on the front wheel so figured it would be ok. Almost back, it was bogging down and I couldn't figure out why - thought it was a the carb - then the front wheel totally slid and down we went. Luckily, was going slow. Lesson learned!
03:16 PM JT-Shop: I put new pads in my old chevy truck and one caliper was somewhat frozen made it to town and back but had to stop often to let the rotor cool off
03:43 PM Tom_L: you get all 4 calipers changed ?
03:43 PM lcnc-relay: <skunkworks8841> I know we have had 2 situations where the flexible brake line plugs up and causes the calipers to stick
03:47 PM JT-Shop: I had a brake hose on the truck close up and made it pull to one side
03:48 PM JT-Shop: Tom_L, just working on the back two and the master cylinder
03:51 PM rdtsc-wk: It's gonna have great brakes when he's done with it
03:54 PM xxcoder: hopefully not too great lol (as in locked in closed)
04:01 PM rdtsc-wk: With all the effort in rebuilding them, assuredly they'll be fully functional. :) Can't mess around with brakes - need to work.
04:08 PM Tom_L: JT-Shop, when you get a min, rcc -o resources.py resources.qrc then from the working folder: pyuic6 yourapp.py yourapp.ui and tell me if that's anything useful
04:12 PM Tom_L: err, add the rcc -g python -o resources.py......
04:12 PM Tom_L: typo
04:13 PM Tom_L: but not on the pyuic6 line
04:14 PM Tom_L: try once more: rcc -g python -o resources_rc.py resources.qrc
04:14 PM Tom_L: notice the output file
04:29 PM rdtsc-wk: Trying out GitKraken, seems pretty nice (was in my repo): https://www.gitkraken.com
04:30 PM * JT-Shop takes 5 with Brubeck
04:31 PM lcnc-relay: <Travis Farmer> gitkraken is fine if you don't need to connect to a self-hosted gitlab repo
04:32 PM JT-Shop: Tom_L, pyuic converts the designer xml to a python file
04:34 PM JT-Shop: I've seen resources_rc.py in pyqt4 iirc
04:39 PM * JT-Shop prefers git gui over gitkraken
04:43 PM lcnc-relay: <JT> https://jauriarts.org/_matrix/media/v3/download/matrix.org/dBCLvgCpUabgtEpOZGgzFcFH/1000002329.jpg
04:47 PM xxcoder: air powered car ;)
04:53 PM Tom_L: JT-Shop, so not that useful
04:54 PM Tom_L: unless you were packaging an exe
04:54 PM Tom_L: don't drip fluid on the paint...
05:07 PM rdtsc-wk: or in your eye!
05:07 PM Tom_L: is there a limit on halui MDI_COMMANDS?
05:07 PM Tom_L: rather, what is the limit?
05:09 PM Tom_L: i see a post from '15 that says 63
05:14 PM Tom_L: JT-Shop, is it possible to call a subroutine from a 'function mdi' pin?
05:17 PM JT-Auto-Shop: yup
05:18 PM Tom_L: i'm wondering which is better mdi_command or using function mdi?
05:19 PM Tom_L: function mdi would be shorter
05:19 PM Tom_L: and not require postgui
05:21 PM JT-Auto-Shop: ?
05:21 PM JT-Auto-Shop: mdi button vs ?
05:22 PM Tom_L: the mdi function you can put it in the code. using mdi_command you need to go thru postgui don't you?
05:22 PM Tom_L: my sub is going to blink a hal show pin to start since i have no probe
05:25 PM JT-Auto-Shop: link?
05:26 PM JT-Auto-Shop: https://gnipsel.com/linuxcnc/flexgui/mdi.html#mdi-buttons
05:26 PM JT-Auto-Shop: mdi buttons
05:26 PM Tom_L: right
05:26 PM Tom_L: command could be the call to the sub
05:26 PM Tom_L: right?
05:27 PM JT-Auto-Shop: and what's the other one you speak of?
05:27 PM JT-Auto-Shop: oh yeah
05:27 PM JT-Auto-Shop: I see a typo
05:27 PM Tom_L: using postgui to call mdi_command from the ini
05:27 PM JT-Auto-Shop: oh no it's not a typo
05:27 PM Tom_L: mdi buttons would be the short way to get there
05:27 PM JT-Auto-Shop: ah the emc way
05:40 PM Tom_L: yeah that works great
05:52 PM JT-Shop: well I _think_ the back brakes are done
05:54 PM Tom_L: cool
05:54 PM Tom_L: 2 reservoirs so you don't have to re bleed the rear for the front?
05:54 PM JT-Shop: it has a nice firm peddle now
05:54 PM JT-Shop: right
05:55 PM Tom_L: i've always wondered how many PSI a brake system generates
05:56 PM JT-Shop: depends on how scared you are I assume
05:56 PM Tom_L: well yeah
06:00 PM Tom_L: skunkworks, when you use cradek's probe routines, where does it put the result?
06:05 PM JT-Shop: iirc they save the result to a parameter
06:06 PM Tom_L: JT-Shop, on this screen: http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/Probe_screen6.png
06:06 PM Tom_L: in the first 2 boxes
06:06 PM Tom_L: the X+ Y+ X- Y- would repeat there wouldn't they?
06:06 PM Tom_L: just in different positions
06:07 PM JT-Shop: yup looks the same to me x+ is x+
06:07 PM Tom_L: you would just locate the probe a bit differently
06:08 PM JT-Shop: how's that
06:08 PM Tom_L: well maybe not :)
06:08 PM Tom_L: so that's 9 probe calls including the ID circle
06:08 PM lcnc-relay: <roguish> Tom_L: you pink points are good starting placews
06:09 PM Tom_L: and i think i have one for material and maybe the tool touch
06:09 PM Tom_L: roguish are you willing to test?
06:09 PM Tom_L: i haven't finished it just yet but probably this evening i'll have at least those
06:09 PM lcnc-relay: <roguish> Tom_L: sometime, yes.
06:09 PM Tom_L: i just have no way to test
06:10 PM lcnc-relay: <roguish> can they run in sim?
06:10 PM Tom_L: and we know cradek's routines are solid
06:10 PM Tom_L: i dunno
06:10 PM Tom_L: maybe
06:10 PM JT-Shop: if you have a sim way to trip the probe they could
06:10 PM lcnc-relay: <roguish> ya make an independent button to 'click' in and trip the probe
06:11 PM JT-Shop: you have to be super fast on the button release...
06:11 PM Tom_L: run it thru a debounce
06:11 PM JT-Shop: wonder how I know?
06:11 PM Tom_L: heh
06:12 PM lcnc-relay: <roguish> one thing I really like, is to have a view going on during the probe.
06:12 PM JT-Shop: you can have the plot now...
06:14 PM lcnc-relay: <roguish> just got a pack of x-acto blades so I can pick the nickel slivers out of my fingers....
06:14 PM Tom_L: let's hope that's all you do to them
06:15 PM xxcoder: I have tweezer with lens and light attachment
06:15 PM xxcoder: it have pulled so many metal chips from my fingers
06:15 PM lcnc-relay: <roguish> Tom_L: so, are you going to add outside and inside widths ?
06:17 PM Tom_L: we
06:17 PM * JT-Shop calls it a day
06:17 PM Tom_L: we'll see
06:17 PM Tom_L: i want to test a few first
06:17 PM Tom_L: i'd be writing them blind
06:18 PM lcnc-relay: <roguish> Tom_L: I need to do some machining this week. if you let me know the easiest way to install, I can probably test...
06:18 PM lcnc-relay: <roguish> I have not loaded Flex on my mill machine
06:18 PM Tom_L: just install the examples in the linuxcnc/configs folder and make an icon to rpi_mill_3
06:19 PM Tom_L: i haven't uploaded anything yet
06:19 PM lcnc-relay: <roguish> can't today, maybe tomorrow. or wed.
06:19 PM Tom_L: i may not be ready that soon anyway
06:20 PM lcnc-relay: <roguish> no hurry here....
06:20 PM Tom_L: this must be a sim since i can run it on the test pc
06:20 PM Tom_L: i can post the files somewhere ahead of putting them on git
06:21 PM lcnc-relay: <roguish> well, I
06:22 PM lcnc-relay: <roguish> well, I'll be at my desk in the next couple monnings....
06:23 PM lcnc-relay: <roguish> nice day here today. topped out at 73.
06:24 PM lcnc-relay: <roguish> 88 yesterday, 95 sunday
06:29 PM lcnc-relay: <roguish> ok, rum and tonic time on the deck..........be back tomorrow.......caio
06:53 PM rdtsc_away is now known as rdtsc
06:55 PM rdtsc: Leinenkeugel's Lakeside Cherry here
06:56 PM rdtsc: Re: Tom_L, car brake line pressure: 556PSI https://www.brakeandfrontend.com/brake-math-psi-clamping-forces-what-happens-in-the-real-world/
08:12 PM Tom_L: JT-Cave, if increments isn't in the ini, it generates an error
08:13 PM Tom_L: err wait.. it's the ',' comma error i ran into again
08:20 PM Tom_L: roguish one is working
08:54 PM Tom_L: ok, X+- Y+- id hole working
09:01 PM Tom_L: workpiece top working
09:19 PM rdtsc is now known as rdtsc_away
09:38 PM Tom_L: JT-Cave, PR #77 for minimal probe sim
09:40 PM Tom_L: rdtsc, roguish, http://tom-itx.no-ip.biz:443/~webpage/cnc/JT-SHOP/flexgui/touch_probe_sim/
10:24 PM Tom_L: JT-Shop, may need some help on the tool probe sensor code
10:25 PM Tom_L: looks like it uses a couple spinboxes for probe level and search vel
10:25 PM Tom_L: and an ini section [TOOLSENSOR] with a few settings there