#linuxcnc Logs

Jun 02 2025

12:02 AM xxcoder: https://youtu.be/dXIkUkwzTQU
12:16 AM xxcoder: forgot url lol
12:54 AM lcnc-relay: <meisterdippel@> moin
03:43 AM lcnc-relay: <travis_farmer@> Morning... just removed myself from all discord servers, but this one... apparently i am a annoying worthless individual that nobody wants to be around... 😦
04:27 AM Tom_L: morning
04:35 AM lcnc-relay: <meisterdippel@> travis_farmer@: that's the wrong attitude, i don't know you but you're not hurting anyone - everyone in the chat is an enrichment, i think i annoy people more with my videos and RIO stuff 🙂
04:35 AM lcnc-relay: <z3rni3@> Morning👋
04:36 AM lcnc-relay: <z3rni3@> travis_farmer@: I doubt thats true! Just be yourself and the right people will appreciate you!
04:53 AM Deejay: hi
07:05 AM rdtsc-w: morning Tom, Deejay, meister, all - life is too short to waste on feeling negative about anything... the only good negative thing is rake angle
09:45 AM roycroft: hi folks
09:45 AM * roycroft woke up this morning for the first time without worrying about finishing his move
09:46 AM bjorkintosh: pleasure to welcome you to the rest of your life.
09:49 AM roycroft: well, to the next phase, at least
09:50 AM roycroft: i'm still two years behind on other stuff - plenty to worry about still
09:50 AM roycroft: but at least not that any more
09:52 AM roycroft: my car is finally going in the shop today to fix an annoying sensor problem
09:52 AM roycroft: that will be another huge relief - the sensor has been flaky since last fall
09:56 AM roycroft: honestly, the shaper community collectively have no experience with technical design tools
09:56 AM roycroft: there was just a "major" update to their shaper studio design product
09:57 AM roycroft: "One of the unique capabilities of Draw is the ability to specify both length and angle of any of the segments or sides of your shape."
09:57 AM roycroft: folks are talking about that as though it's a revolutionary, game-changing feature
09:58 AM roycroft: not as shaper studio finally catching up to what cad software has been doing for decades
11:03 AM Unterhaus_ is now known as Unterhausen
01:31 PM roycroft: so the country fair takes place in a month
01:31 PM roycroft: there are over 20,000 volunteers whose information is needed for them to get their camping permits and other inventory
01:32 PM roycroft: i am responsible for managing the server on which the inventory database runs, but i have nothing to do with the inventory database
01:32 PM bjorkintosh: that's really large.
01:32 PM roycroft: for 2 years i've been asking about how that database is being backed up, and offering to manage the backups as part of my machine management tasks
01:33 PM roycroft: i've been refused access of any sort to the database
01:33 PM roycroft: on may 29 one of the database people nuked a large portion of the database
01:33 PM roycroft: early this morning i was asked to restore it
01:33 PM roycroft: i restored the system backups that include the directory where the database team's backups are stored
01:34 PM roycroft: but i have never backed up the live database because it would be completely useless, as to back up a database the tables need to be locked during the backup to ensure integrity of the backup
01:34 PM roycroft: and they knew this, and assured me that their own backups are adequate
01:34 PM roycroft: so now they have them
01:35 PM bjorkintosh: is it a mysql db?
01:35 PM roycroft: and they are demanding that i restore backups of the live database that they know do not exist and would be useless
01:35 PM roycroft: because their own backups are not full backkups - they are partial backups
01:35 PM Tom_L: sounds like they're desperate
01:35 PM roycroft: no, and let's not even get into that
01:35 PM roycroft: it's mongodb
01:35 PM bjorkintosh: oh.
01:35 PM roycroft: they are
01:36 PM bjorkintosh: someone did a hit and run on them.
01:36 PM roycroft: but instead of fixing their problem they're looking for a scapegoat
01:36 PM roycroft: and trying to make that me
01:36 PM Tom_L: seems like they have a big problem
01:36 PM bjorkintosh: I bet whoever decided it had to be mongo is long gone.
01:37 PM roycroft: i already produced the email trail where they very explicitly told me NOT to do my own backups of the database, and assured me and the executive director that their own backups were complete and tested regularly
01:37 PM roycroft: they are, however, still trying to find a way to blame me
01:37 PM roycroft: the person who decided on mongodb is still there and doubling down on the decision
01:37 PM roycroft: and is the main person blaming me
01:38 PM bjorkintosh: aha!
01:39 PM xxcoder: probably guy who used bad command too
01:39 PM roycroft: it would have taken me seconds to write a backup script to back up their entire database, and maybe a couple hours to set up an environment where i could restore backups and ensure data integrity
01:39 PM roycroft: but they very explicitly told me that it was none of my business and that i was not to do so under any circumstances
01:39 PM xxcoder: thats why records and papertrail is amazing
01:39 PM roycroft: they made their bed
01:40 PM roycroft: and when they made it they put a big hole in the middle of it
01:40 PM roycroft: so when they had to lay on it they fell through the hole
01:42 PM rdtsc-w: reminds me of a time I was asked to look at an "analog" spark plug coil tester for automotive QC testing... built by a PhD student (whom quickly moved on) - brilliant design, but natively unstable and highly finnicky... they were not happy when I suggested a $2 microcontroller for much more accuracy...
02:26 PM Tom_L: pcw---home, i adjusted the FERROR to .005 and MIN_FERROR to .0005 and ran thru all my tests without a hitch
02:27 PM Tom_L: i may lower that until i get an error so i know what's going on
02:27 PM pcw---home: Wonder what was different?
02:28 PM pcw---home: (plotting the error is a good way to see whats going on)
02:28 PM Tom_L: i'd lowered all my accel rates yesterday but didn't change feror
02:28 PM Tom_L: prior to that i think my accel was just too high
02:28 PM Tom_L: i did change that hal setting to 200
02:29 PM Tom_L: i don't think i've ever set up the plot
02:29 PM rdtsc-w: the whole accel-per-axis, and per-joint, combined with two following errors, makes this all super confusing - I'd had to bump up initial accel to avoid following errors...
02:29 PM pcw---home: Just look at the joint following error in halscope
02:30 PM Tom_L: is FERROR = .001 unreasonable?
02:30 PM pcw---home: No that fine if there are not other issues
02:30 PM Tom_L: i'm not going to hope for much better than that on this mill
02:30 PM Tom_L: what does MIN_FERROR do?
02:31 PM Tom_L: seems kinda redundant to me
02:31 PM pcw---home: note that those are only limits of when to fault, they have no effect on the accuracy
02:31 PM Tom_L: that could change under load too
02:32 PM Tom_L: i realize that
02:34 PM pcw---home: Ferror is at maximum speed, min_ferror is at 0 speed. Actual trip point calculated as somewhere between min_ferror and ferror depending on joint velocity
02:35 PM Tom_L: i associate that to be similar in cad where a line tolerance can deviate that amount
02:35 PM Tom_L: when generating cam
02:38 PM rdtsc-w: Tom, try a G64 P0.0005 before some real (big and complex preferred) gcode and see how it handles it - each position should be within a thou; could setup a DTI and verify
02:38 PM pcw---home: The DPLL setting may have helped as too low a value can cause enough real sampling jitter to trigger bad behavior in the built-in position mode
02:39 PM Tom_L: would you try going lower than 200?
02:39 PM pcw---home: (since you are using position rather than velocity mode in your configuration)
02:39 PM pcw---home: There is no advantage in going lower
02:39 PM Tom_L: ok
02:40 PM Tom_L: rdtsc-w, that's exact path mode right?
02:40 PM pcw---home: what -200 means is sample the stepgen position 200 usec before the nominal read time
02:40 PM rdtsc-w: kinda, that's "you're allowed to be off by 0.0005" mode :)
02:42 PM Tom_L: rdtsc-w, i updated the hole ops in flex recently if you're using any of em
02:42 PM Tom_L: and a couple probe routines
02:42 PM pcw---home: if this DPLL number is too small you can get in a situation where sample does not happen before the read ( due to base line wander ) to you get 1 ms old position data and the control loop goes crazy
02:44 PM pcw---home: so you get
02:44 PM Tom_L: is that like waiting for propagation delay to settle?
02:53 PM rdtsc-w: cool.. I'm still plugging away at the articulating arm in spare time... someday it'll be done, someday...
02:53 PM lcnc-relay: <skunkworks8841@> I feel like it kinda makes the latency consistant..
02:53 PM lcnc-relay: <skunkworks8841@> or - not effecting the loop
02:55 PM pcw---home: The DPLL reduces sampling jitter by sampling things like stepgen or encoder positions at a relatively fixed time rather than at the time LinuxCNC reads them
02:57 PM pcw---home: that is, say an Ethernet device might have 200 usec of sampling jitter. This causes a major position sampling error (noise basically) in the position feedback
02:58 PM lcnc-relay: <skunkworks8841@> it is crazy how well it works.
02:58 PM pcw---home: The DPLL reduces that jitter to less then 1 microsecond
02:59 PM pcw---home: less than
03:00 PM bjorkintosh: so the DPLL acts as a buffer?
03:01 PM pcw---home: I think of it as a flywheel that's synced to LinuxCNCs read cycle
03:02 PM pcw---home: so things like latching the stepgen position count happen at a fixed flywheel angle
03:03 PM pcw---home: (with minimal jitter, typically in the 150 ns region)
03:05 PM bjorkintosh: ah. quite clever.
03:05 PM pcw---home: PLLs are a pretty old idea
03:05 PM bjorkintosh: a flywheel.
03:06 PM pcw---home: Yep flywheels are old as well...
03:06 PM bjorkintosh: digital pll
03:08 PM lcnc-relay: <skunkworks8841@> lol - that is kinda how I think of it. Having a fixed known time for things to happen.
03:55 PM rdtsc-w: if you think of the period of a clock signal, say it goes high every second: a PLL will phase-lock onto this; phase meaning at some angle (0-360deg), which creates a time-altered version of the clock, which can then be used to cause something to happen before (or after) the clock normally would. PLLs can also divide and/or multiply the clock if needed. DPLLs are modern digitial versions of the old analog standard PLL, not to be confused with https://en.
03:55 PM rdtsc-w: wikipedia.org/wiki/DPLL_algorithm which is a whole other can of worms :)
03:56 PM xxcoder: thanks for warning, for me, the can stays firmly sealed lol
04:09 PM Tom_L: so i ran FERROR to .001 and MIN_FERROR to .00005 and all the tests i ran went great
04:10 PM Tom_L: the flywheel seems to be spinning in the right direction
04:21 PM roycroft: i started watching a video yesterday about a wood thicknesser rebuild
04:21 PM roycroft: i noticed early on that all the screws/bolts they were removing were left-hand thread
04:21 PM Tom_L: rdtsc-w, the 68332 pwm used a phase shift pwm using a primary pulse it was easily able to produce 16 channel pwm
04:21 PM roycroft: i thought that odd, and watched more closely, and found that the video was a mirror image of reality - the text was backwards
04:22 PM roycroft: why would someone do that?
04:22 PM roycroft: it was too irritating to watch any further
04:22 PM Tom_L: for those down uner
04:22 PM Tom_L: under
04:22 PM roycroft: i'll cancel my plans to emigrate to australia
04:22 PM Tom_L: just be glad you're on the top side of the globe
04:23 PM Tom_L: i'm pretty happy with this config tuning so far
04:23 PM xxcoder: thats odd.
04:23 PM xxcoder: wonder if its copied video and they flipped it to avoid being found out
04:24 PM xxcoder: only reason I can think of
04:24 PM roycroft: i'm sittin' on top of the world
04:24 PM roycroft: i'm rolling along, yes rolling along
04:24 PM roycroft: it's irritating as heck
04:24 PM xxcoder: stop the world, I want to get off
04:25 PM roycroft: my e-stop button arrived a few minutes ago
04:25 PM roycroft: it's really nice, but it's comically large for the application
04:25 PM roycroft: i'm not sure i can use it
04:26 PM roycroft: it's about as big in diameter as the height of the controler chassis
04:26 PM xxcoder: how large is it?
04:26 PM roycroft: probably 80mm in diameter, the mushroom part
04:27 PM roycroft: i expected it to be about half that size for some reason
04:27 PM roycroft: if i don't use it on this project, i'll find a good use for it eventually
04:27 PM roycroft: maybe on the bigger router
04:28 PM roycroft: but that would leave me with finding something for the mini router
04:28 PM roycroft: and i really don't want to use one of those red plastic buttons
04:28 PM xxcoder: time to wood it?
04:29 PM roycroft: perhaps
04:30 PM Tom_L: at least you won't miss it if you need to hit it quick
04:32 PM Tom_L: i may have gotten mine from delta electric locally.. i don't honestly remember now
04:34 PM roycroft: i was too tired to stay up late last night to see the aurora borealis
04:35 PM roycroft: and tonight's aurora is predicted to be just out of view
04:35 PM roycroft: visible down to a bit south of portland, but not down to california like last night
04:35 PM roycroft: xxcoder should be able to see it
04:35 PM roycroft: except tacoma might be too bright
04:35 PM xxcoder: ohh thanks for tip
04:35 PM xxcoder: last time it happened, I used phone to be able to see it
04:35 PM roycroft: it's worth a go, though
04:35 PM xxcoder: it was very faintly viewable otherwise
04:36 PM roycroft: yeah, phone cameras can sometimes pick it up when it's not visible to the naked eye
04:36 PM xxcoder: yep
04:45 PM roycroft: my fans did not arrive today as scheduled
04:46 PM roycroft: they are supposed to be identical to one that i already have, but i want to have them in hand before i route the controller cabinet to fit them
04:46 PM roycroft: i'm funny that way
04:48 PM rdtsc-w: lol https://www.youtube.com/watch?v=dXIkUkwzTQU
05:55 PM JT-Shop: 87°F
05:57 PM Tom_L: haha same here
05:57 PM roycroft: it's going to get hotter than that here by week's end
05:57 PM Tom_L: and breezy
05:57 PM roycroft: 36c by sunday
05:57 PM xxcoder: oh boy. 80s f starting sat
05:57 PM JT-Shop: gotta start my ride a lot earlier now that summer is here
05:58 PM Tom_L: supposed to rain tomorrow
05:59 PM roycroft: my car is fixed and i am happy that i did not have to deal with it myself
05:59 PM JT-Shop: what was wrong with it?
05:59 PM roycroft: and just a short drive after i picked it up shows that i'm getting at least 4mpg better fuel economy than before
05:59 PM roycroft: crankshaft position sensor was bad
06:00 PM roycroft: it would work when the engine was cold, but as soon as the engine started warming up it would fail, and the ecu would go into derate mode
06:00 PM roycroft: which severely limited power
06:00 PM roycroft: and really hurt fuel economy
06:01 PM roycroft: i don't have a vehicle shop or a lift
06:01 PM roycroft: i did not want to jack the car up on the driveway and crawl under it to fix it
06:02 PM JT-Shop: I got tired of doing that and put in a 4 post lift
06:02 PM roycroft: it cost me $170 in labor to have someone else do it, and i'm quite happy with that
06:02 PM JT-Shop: that's not bad
06:02 PM roycroft: yeah, if i had a big shop i'd consider doing that myself
06:02 PM roycroft: but i don't so i won't
06:04 PM JT-Shop: transitioning from a machine builder to a hobbyist is a lot of work
06:05 PM Tom_L: retirement in itself is alot of work
06:05 PM roycroft: let's face it
06:05 PM roycroft: life is a lot of work
06:06 PM JT-Shop: I don't want to retire fully but want to play more with things I like working on/with
06:06 PM JT-Shop: and most important is to stay busy and keep my brain thinking
06:07 PM Tom_L: exactly
06:15 PM * JT-Shop calls it a night
06:43 PM roycroft: my retirement will mean doing the things that i really love to do instead of doing what i need to do to make a living
06:43 PM roycroft: i'll still need to make money, but not a lot, and if i can do that doing something i love then it's a win
06:44 PM roycroft: i am almost wanting at this point to find a way to fit the mill/drill into my current shop setup, and figure out how to move it and set it up by myself
06:45 PM roycroft: it would make so many things so much easier
06:45 PM roycroft: but then i would say the same thing about my welding gear
06:45 PM roycroft: and my other wood bandsaw
06:45 PM roycroft: and everything else that is in storage
06:45 PM roycroft: which will not fit in the current shop
11:34 PM roycroft: i partially unloaded my pickup today, but then i stopped when i felt like stopping
11:34 PM roycroft: it was not so feel no pressure to get it done
11:34 PM xxcoder: thats good not to push too hard
11:34 PM roycroft: yup
11:35 PM roycroft: but then i went to find a piece of aluminium that was 10mm or a bit more thick, for the limit switch indexes, and could not find anything quite that thick
11:36 PM roycroft: 3/8" was as thick as i had, and that's just a wee bit too thin
11:36 PM roycroft: so maybe, even after having moved all my metal stock here, i might have to buy a small piece for this job
11:37 PM roycroft: i have plenty of steel that is thick enough, but i don't want to manually make the indices out of steel
11:37 PM xxcoder: stacking not an option?
11:37 PM roycroft: i'd rather not
11:39 PM roycroft: i just got rid of an aluminium cylinder head for a toyota r22 engine
11:40 PM roycroft: i bet i could have cut pieces out of that that would have worked
11:40 PM xxcoder: yeah. you dont do any casting?
11:40 PM roycroft: no
11:40 PM roycroft: although i'd like to
11:40 PM roycroft: but if/when i do, i'll focus on bronze and precious metals
11:40 PM roycroft: and maybe copper
11:40 PM roycroft: but not aluminium or steel
11:41 PM roycroft: i was moving
11:41 PM roycroft: it was big enough and useless enough (it was cracked and warped)
11:41 PM roycroft: so it went away
11:41 PM xxcoder: I do want to try casting evenually, but with lower temp metals
11:41 PM roycroft: i had kept it thinking i might be able to harvest small bits of it occasionally
11:46 PM roycroft: then i got rid if it right before i needed it for my first harvest
11:47 PM xxcoder: its impossible to predict
11:47 PM roycroft: yeah
11:47 PM roycroft: no regrets
11:47 PM roycroft: i did what i needed to do
11:50 PM xxcoder: yeah