#linuxcnc Logs

Jan 14 2023

12:09 AM roycroft: i've been using hakko soldering irons for decades
12:19 AM roycroft: but i just use the small ones with the rheostat
12:19 AM roycroft: nothing fancy
12:19 AM roycroft: unfortunate that they're that way with their higher end stuff
12:19 AM Rhine_Labs: Yea, I have a bunch of the higher end stuff reliable affordable but no right to repair.
12:22 AM roycroft: that kind of reminds me of something we dealt with at the telco i worked for
12:22 AM roycroft: this was over 20 years ago, so i'm a little fuzzy on the details
12:22 AM roycroft: but basically, we had a bunch of phone switches, and they had some firmware burned onto an eprom
12:22 AM roycroft: that firmware was buggy
12:22 AM roycroft: and it gave us a lot of grief
12:23 AM roycroft: i was the manager for the data side, not the telecom side, so i wasn't directly involved in all of this
12:23 AM roycroft: but they bought a new version of the switch that had newer firmware, and it worked really well
12:23 AM roycroft: so the telco asked for new eproms for all the other switches
12:24 AM roycroft: the vendor refused to provide them
12:24 AM roycroft: and said they would have to send a field tech out, at the telco's expense, to reflash each switch
12:24 AM roycroft: my telco side counterpart was whining about it to me one day, and i asked to see the switch
12:25 AM roycroft: it turned out it was a common socked dip eprom
12:25 AM roycroft: socketed
12:25 AM roycroft: so i went home, got my eprom burner, and my magic uv eprom eraser, and went back and pulled the eprom out of the new switch
12:25 AM roycroft: read it, stored the file, and then flashed all the others in turn
12:26 AM roycroft: they then switched vendors for their phone switches after that episode
12:27 AM roycroft: i still have that eprom burner, but i haven't used it in many, many years
12:27 AM roycroft: iirc the machine that talks to it runs windows 95
12:28 AM XXCoder: can always vm it if needed ever
12:28 AM roycroft: it has an isa bus controller card
12:29 AM XXCoder: older pc with vm on it?
12:29 AM roycroft: possibly
12:29 AM roycroft: i bet the machine still works
12:29 AM roycroft: machines that old last forever
12:29 AM roycroft: it's newer machines that die quickly
12:29 AM XXCoder: unless it killed itself by capos
12:29 AM XXCoder: 90s is right era for poisoned caps
12:29 AM XXCoder: thankfully fixable
12:31 AM roycroft: i'll probably never use it again
12:31 AM XXCoder: dont sound like it
12:31 AM roycroft: when is the last time you saw a uv-erasable eprom?
12:31 AM XXCoder: infinity years ago
12:32 AM roycroft: they're hanging out in a box next to the vhs player and the palm pilot
12:34 AM roycroft: back when "never gonna give you up" was just a song by rick astley, and rick rolling was not even dreamt of
12:34 AM XXCoder: :) hmm though I do wonder how that started
12:55 AM roycroft: well, travis_farmer will be able to build his new shed, it appears
12:55 AM roycroft: a winning ticket for the $1billion+ lottery was sold in maine tonight
01:28 AM XXCoder: man would love to win billion usd lottery
01:35 AM Deejay: moin
02:52 AM travis_farmer[m]: roycroft: while i admit, i bought a ticket with high hopes, it was in fact a dud. so somebody else in Maine has my money. ;-)
02:52 AM travis_farmer[m]: G'Morning
02:52 AM XXCoder: lol drat
02:53 AM travis_farmer[m]: yeah, it would have been nice
04:14 AM travis_farmer[m]: the more i think about that Matrix tool changer magazine, the more i would love to build one... understand though, this is a dream, not actually what i will try an build. but it would be really neat to make. could be done on a small scale but it would still be too expensive to be practical. i mean, it would require it's own LCNC computer, just to be able to work independently of the main LCNC system for things like prefetch, and
04:14 AM travis_farmer[m]: storage of the previous tool holder. far beyond my finances... but the idea is still very cool :-)
04:47 AM GTKplusplus[m]: why would it require a dedicated PC?
04:47 AM GTKplusplus[m]: just a dedicated mesa or other control board would work, with a PLC component in linuxcnc
04:52 AM sensille: and the next question: when i do an outside corner probe, it first moves <step_off_width> away from the workpiece so it is outside, but then moves Z down by "z clearance" + something, which is way too low and even touches the bed. imho it should only go down max z dist + something. what am i missing
04:52 AM sensille: ?
04:55 AM sensille: does it assume z to be touched off to something?
05:11 AM jpa-: it is apparently supposed to be that much above the workpiece to begin with
05:11 AM jpa-: i always found it pretty weird
05:11 AM sensille: the documentation states explicitly otherwise. what can you recommend for probing?
05:12 AM jpa-: not really anything, i ended up writing my own but it's not really well tested either
05:12 AM sensille: this stuff really makes me feel stupid :-/
05:12 AM jpa-: it seems the common solutions should work, but i never really understood how they are supposed to work
05:22 AM sensille: thanks for making me feel not totally alone :)
05:42 AM travis_farmer[m]: GTKplusplus: sorry, i drifted off to sleep. the reason i thought it needed a separate LCNC computer, was that i thought if it was operating independent of the running Gcode, to move steppers on the matrix magazine, i thought it had to be a different computer running it. if it can be done within a comp file, that would be interesting. i didn't find a component named PLC, so i assume you meant i had to write my own? how would it move
05:42 AM travis_farmer[m]: the steppers of the matrix magazine without affecting the running Gcode?
05:43 AM GTKplusplus[m]: classicladder
05:43 AM GTKplusplus[m]: that's the PLC component
05:43 AM travis_farmer[m]: ok, i haven't worked with classicladder...
05:46 AM JT-Cave: morning
06:40 AM Tom_L: morning
06:43 AM jpa-: travis_farmer[m]: unless you are well versed in ladder logic, i doubt classicladder will make it any easier - i found it easier to just write a custom HAL component in C or Python when more complex logic is needed
06:58 AM * JT-Cave finds ladder logic to be easy
07:10 AM * travis_farmer[m] finds JT to be bragging ;-)
07:20 AM GTKplusplus[m]: that would work too, just assuming he wanted to do something kinda built in
07:20 AM GTKplusplus[m]: I too would prefer a python component
07:38 AM JT-Woodshop: just the facts
07:45 AM JT-Shop: ladder logic is very simple... get inputs, solve logic, update outputs
08:39 AM * Tom_L wished orient was as simple
08:40 AM JT-Cave: who wrote orient?
08:40 AM Tom_L: harbler
08:41 AM Tom_L: i'll get it, i just haven't had enough time all in one spot
08:41 AM travis_farmer[m]: JT is telling me, someone who's mind is illogical, how easy logic is ;-)
08:41 AM Tom_L: JT-Cave, you see the gantry pics?
08:42 AM Tom_L: travis_farmer[m], just map it out on paper
08:43 AM travis_farmer[m]: so far i get by with C comps, so i will stick with that for now. ;-)
08:44 AM Tom_L: s/harbler/haberler
08:58 AM JT-Cave: no
08:59 AM JT-Cave: oh you mean the giant gantry, yes
09:30 AM markgonz: hello friends!
09:30 AM markgonz: Im planing a small granite frame cnc for watchmaking
09:31 AM markgonz: and in the end I would like to add linear encoder closed loop with linuxcnc
09:31 AM markgonz: my question is if going for servos or steppers make this easier in some way?
09:32 AM markgonz: I dont really need speed servos offer for micro machining I believe
09:32 AM markgonz: Im looking at leadshine closed loop steppers vs jmc servo with integrated driver
09:38 AM JT-Woodshop: I've used JVL servos with built in driver and plc
09:47 AM markgonz: what is plc?
09:47 AM Tom_L: programmable logic control
09:49 AM jpa-: markgonz: either will work, or you can even just have DC motors & make them into servos using the feedback
09:49 AM JT-Woodshop: Tom_L, is that gantry at the new place your boy works at?
09:50 AM Tom_L: yes
09:50 AM Tom_L: it was put in years ago by the previous owner
09:50 AM Tom_L: there are... i think 4 or 5 in that building
09:50 AM Tom_L: at least 2 of the dual gantries like you see
09:51 AM Tom_L: they were setting new rails on one of them
09:51 AM JT-Woodshop: wow
09:51 AM Tom_L: that plate in 1 took 2 weeks to skin cut
09:52 AM Tom_L: ~180' bed
09:52 AM Tom_L: notice the middle spindle is missing.. out for repair
09:53 AM Tom_L: actually it was sitting on the bench
09:53 AM Tom_L: the cmm is brand new
09:54 AM Tom_L: that's one of the ones he runs
09:57 AM Tom_L: you can see it's twin in pic 2
09:58 AM JT-Woodshop: I had to look hard to even find the spindles lol
09:58 AM Tom_L: you can't miss em.. they're huge
09:59 AM Tom_L: pic doesn't give the feel of the size
09:59 AM * JT-Woodshop is going to Chaumette Vineyards for lunch yummy
09:59 AM Tom_L: nice.. i'm going to a 4yr old's party :)
09:59 AM Tom_L: i'd rather be going with you
10:00 AM JT-Woodshop: you would have to ride in the back of the Corvette :)
10:00 AM Tom_L: still...
10:00 AM JT-Woodshop: for an hour and a half...
10:00 AM Tom_L: meh.. i _am_ getting older..
10:00 AM JT-Woodshop: I couldn't do it
10:01 AM Tom_L: i doubt i could either
10:01 AM JT-Woodshop: I don't like ladders any more
10:01 AM JT-Woodshop: I rented a lift to install the starlink dish
10:01 AM Tom_L: heh
10:01 AM JT-Woodshop: clear sky here today
10:02 AM Tom_L: last time i painted, i borrowed some scaffold
10:02 AM Tom_L: yeah same here but chilly
10:02 AM JT-Woodshop: I have two sections of scaffold and 3 scaffold planks
10:03 AM Tom_L: that's about what i borrowed but it's been long enough it could use another coat
10:05 AM JT-Woodshop: I used mine when I built the machine shop but not since
10:07 AM markgonz: so if buying closed loop servos is posible or makes sense to disable the motor rotary encoder and just use linear encoder on the axis?
10:07 AM markgonz: closed loop steppers i meant
10:07 AM Tom_L: use both
10:08 AM Tom_L: don't ask me how they do it, i've never done that
10:08 AM Tom_L: you may want the encoders for the closed loop
10:09 AM Tom_L: they were built as a unit
10:09 AM Tom_L: then use the rail as position for lcnc
10:10 AM markgonz: ok thank you
10:15 AM skunkworks[m]: I didn't - link?
10:16 AM Tom_L: skunkworks[m], i'm sure your dad will get a kick out of it
10:16 AM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/cincinnati/
10:17 AM skunkworks[m]: Wow - huge.. Wonder what kind of speeds they can get..
10:17 AM skunkworks[m]: Do you know what they were making? aerospace?
10:18 AM Tom_L: the heavy metal ones rapid around 200ipm, aluminum around 4-600
10:18 AM Tom_L: yes airplane/chopper parts
10:18 AM Tom_L: he had some blackhawk floors sitting on the floor when i was there
10:20 AM Tom_L: the 3 spindle setup is interesting.. you set one then the other 2 are offsets
10:22 AM Tom_L: i didn't get any pics of the new starrag ecospeed they just put in
10:22 AM Tom_L: 6 axis
10:36 AM travis_farmer[m]: (regarding a Mesa THCAD-10) when an encoder input is in counter mode, and i am reading the velocity HAL pin, what number can i expect? trying to convert the value to spindle RPM, where 0v = 0rpm and 10v = 24000rpm.
10:37 AM travis_farmer[m]: does the encoder become a frequency counter, and therefore the output is the actual frequency?
10:39 AM jpa-: travis_farmer[m]: AFAIK yes, but of course makes sense to test with a known voltage
10:40 AM jpa-: (and assuming encoder.scale is left at the default 1)
10:40 AM travis_farmer[m]: yeah, i will. i just was wondering before i head back out to the shop after lunch. :-)
10:41 AM travis_farmer[m]: i can dead-short the input of the THCAD-10, and the frequency should be the marked 0v frequency on the board. should be a good test, i suppose
10:56 AM * JT-Shop closes up shop for a while
10:57 AM JT-Shop: Tom_L, have fun today
11:26 AM _unreal_: got some new updates done. https://github.com/berin-aquaquad/orange-pi-5
12:04 PM travis_f[SHOP]: regarding my THCAD-10, i am getting VERY erratic output. encoder velocity reads 0 when the spindle is not turning, yet at running RPM, it jumps all over. maybe i don't have the THCAD-10 connected right? +5 to 5v supply, gnd to star ground, and FO+ to the encoder input. is there something else i need to connect? the input is connected to the VFD
12:06 PM travis_f[SHOP]: the LED on the THCAD-10 is blinking at about a 1 second interval, and speeds up a bit when the VFD starts
12:07 PM Tom_L: sounds like pid needs tuning
12:08 PM travis_f[SHOP]: what PID?
12:08 PM Tom_L: or lack of one :)
12:09 PM travis_f[SHOP]: the VFD outputs a 0v - 10v in reference to the rpm. the THCAD-10 reads in the volts and converts it to a frequency, read by the encoder input
12:11 PM travis_f[SHOP]: the VFD and spindle runs rock steady, and i watched the volts output, and it too is steady. so i don't know why it is so erratic...
12:14 PM pcw--home: Maybe a noise issue, you should consider using differential signaling if the encoder input supports it
12:16 PM travis_f[SHOP]: hmmm, i will try and check
12:20 PM travis_f[SHOP]: maybe the noise is introduced in my analog wire from the VFD... i will hunt down some shielded wire and try that...
12:21 PM jpa-: do you have a scope?
12:21 PM travis_f[SHOP]: it is inside the house...
12:21 PM travis_f[SHOP]: going to grab it
12:34 PM travis_f[SHOP]: ok, i have one probe on the 0v - 10v incomming, and the other probe on the frequency out, with the trigger on the frequency out channel. seems to be a little noise on the analog, at 0v, 0rpm. and when the noise happens, the freq output varies.
12:34 PM travis_f[SHOP]: will try with the VFD running...
12:36 PM travis_f[SHOP]: same slight noise on input, very nice slow rise as the spindle revs up
12:36 PM travis_f[SHOP]: freq output varies a lot more
12:38 PM travis_f[SHOP]: a little less erratic at 24000 rpm, 10v output
12:39 PM travis_f[SHOP]: think a small filter cap at the input of the THCAD-10 would help?
12:48 PM travis_f[SHOP]: i am thinking shielded cable for the analog signal may help. the noise is not huge, but apparently enough to disrupt things. can't remember, twisted pair wire ok for a analog signal, or not?
12:56 PM travis_f[SHOP]: on the THCAD-10, should i connect FR- and FR+ to anything?
01:00 PM travis_f[SHOP]: this would have been so much easier if mb2hal worked...
01:06 PM * travis_f[SHOP] thinks his help vanished, or is otherwise busy...
01:27 PM * travis_farmer[m] is now back in the house :-)
01:54 PM travis_farmer[m]: yeah, i wish this spindle had the tach output it was advertised as having... then i could input that into an encoder channel, and directly read the RPM, rather than cobbling together odd stuff to make it work. for that matter, if the ModBus worked, that would help too, but regardless of what i do, mb2hal reads in nothing from the VFD. but i guess i can make it work with analog.
01:58 PM travis_farmer[m]: i wonder how an optical tach sensor would work? i would just have to paint half the exposed part of the spindle output black. though dust may cause errors... say nothing about paint chipping off at 24000rpm ;-)
02:14 PM sensille: travis_farmer[m]: why does mb2hal not work?
02:15 PM travis_farmer[m]: it neither controls nor reads status from the VFD, so i gave up on it
02:20 PM sensille: i checked with pymodbus.console first
02:27 PM XXCoder: https://www.youtube.com/watch?v=eh3pASckWGE
02:27 PM XXCoder: that guy is interesting. that video is first of 3 video series
02:50 PM XXCoder: watching other video of his, on upgrades of his hf mill. "I made brass knob of course, because plastic one was ugly, it offended me". gonna love artists
03:03 PM roycroft: i can appreciate being offended by plastic knobs
03:03 PM XXCoder: you might like that channel
03:08 PM roycroft: i'll check it out, thanks
03:09 PM roycroft: i'm actually caught up on my work for the fair right now, for the first time in weeks
03:09 PM XXCoder: guys interesting. more artistic than many channels
03:09 PM roycroft: and i'm thinking that this afternoon i'll head out to the shop and start working on the brass knob i just got
03:09 PM XXCoder: I subbed to channel for now
03:09 PM XXCoder: nice :)
03:10 PM roycroft: i think you know that i like parts that have aesthetic considerations
03:10 PM roycroft: which is why i like folks like clickspring and uri tuchman
03:10 PM XXCoder: yeah his ongoing series is interesting. he found a thing with large angle ball thing, curved
03:10 PM XXCoder: so hes making a large clock using that as time mark
03:11 PM roycroft: i'm making a circle jig for my disc sander
03:11 PM roycroft: i want to be able to adjust the diameter of the circles with a granularity of 0.001"
03:11 PM roycroft: and this brass knob is my adjuster knob
03:11 PM roycroft: so i need to grave some index marks on it
03:11 PM XXCoder: nice
03:12 PM roycroft: it will be a fun job with my rotary table on the mill
03:14 PM roycroft: i have to figure out how to make a drag graver for the mill again - i haven't done that in years
03:48 PM -!- #linuxcnc mode set to +v by ChanServ
05:00 PM -!- #linuxcnc mode set to +v by ChanServ
06:02 PM XXCoder: blondihack is awesome
06:40 PM XXCoder: hmm is cork fine for lining mic and calipers enclosure?
06:42 PM Bleepshop: XXCoder: Should be. It'll take a light oil and help control rust.
06:42 PM XXCoder: cool :)
06:42 PM XXCoder: what kind oil?
06:47 PM roycroft: i use cork drawer liners in my machinst tool chests
06:47 PM XXCoder: anything special on prepping those cork?
06:47 PM roycroft: i'd like to use felt, but it is pretty hard to find real wool felt, and not synthetic felt
06:47 PM roycroft: synthetic felt will actually be counterproductive
06:47 PM roycroft: it traps moisture and will make things rust faster
06:48 PM roycroft: wool felt will absorb oil and help keep things not rusty
06:48 PM roycroft: i just buy cork drawer liner in rolls from jerry's
06:48 PM roycroft: the rolls are about 4mm thick
06:49 PM roycroft: i cut to fit the drawer, spray 3m adhesive on the drawer bottom and on the cork, and press it in place
07:47 PM skunkworks: Tom_L: How old is that cincinati? We have a cincinati malicron lathe from the 70's
07:51 PM Tom_L: skunkworks, i know it was put in by the previous owners but i'm not sure exactly when.
07:53 PM -!- #linuxcnc mode set to +v by ChanServ
07:54 PM XXCoder: ugh
07:54 PM XXCoder: having pc reboots problem
07:56 PM Tom_L: skunkworks, i was told they have 10 in various conditions.. some for parts now i'm sure
07:57 PM Tom_L: the day i was in, they had 2 controls out sitting on the bench :)
07:58 PM Tom_L: 70's sounds about right though judging from when i was last in that shop under the old personel
08:05 PM _unreal_: so.... I'm cutting steel for the first time on my cnc.....
08:05 PM _unreal_: ????
08:06 PM _unreal_: its going REALLLLLLLY well and quiet?
08:06 PM _unreal_: I'm dumbfounded its so quiet?????? I'm cutting 1/16th sheet metal steel, with a 4 flute 1/8th bit
08:06 PM _unreal_: 15krpm
08:08 PM _unreal_: I would have thought a big flat sheet would be SCREAMING. but no???
08:08 PM Tom_L: you're just cutting the edge of it
08:08 PM _unreal_: its ... pleasant
08:08 PM _unreal_: no
08:08 PM Tom_L: and have it clamped down i would guess
08:08 PM _unreal_: I'm cutting out the brackets to mount a heat sink on an OPI5
08:09 PM _unreal_: latest photos https://github.com/berin-aquaquad/orange-pi-5
08:11 PM Tom_L: skunkworks.. btw, it's not that far from where MPM is
08:15 PM Tom_L: according to google it became 'milacron' in 1970 so it would surely be some time after that
08:19 PM skunkworks: Tom_L: neat!
08:19 PM Tom_L: one of the workers i spoke with was there when the previous owners had it and he remembers them going in. he's been there 13 years
08:19 PM Tom_L: so do the math :)
08:20 PM Tom_L: so that would put them newer than 70's if that is correct
08:21 PM Tom_L: definitely a shop to visit..
08:22 PM Tom_L: the yellow grate you see in the pics between the 2 gantrys is the coolant tank for it
08:25 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/cincinnati/Cincinnati_4.jpg
08:25 PM Tom_L: there
08:26 PM Tom_L: we'd walked about halfway down the rail when i took that one
08:34 PM _unreal_: just modified the gcode and trying again. had some depth issues
08:53 PM _unreal_: And there is that scream I was looking for. Finally broke through the steel sheet metal
08:54 PM Tom_L: more likely to break a bit now
09:04 PM _unreal_: likely
09:04 PM _unreal_: has not yet
09:04 PM _unreal_: very slow though
09:04 PM _unreal_: only going 20mm / min
09:05 PM _unreal_: I smeared the surface with super lube to help
09:05 PM _unreal_: clearly helping
09:05 PM XXCoder: making dust?
09:05 PM _unreal_: no
09:05 PM _unreal_: regardless its got the dust shoe on and extraction running
09:06 PM XXCoder: so its making visible chips? thats good
09:06 PM _unreal_: I've almost got one part cut
09:06 PM _unreal_: lot of tabs though :(
09:06 PM _unreal_: then again its SHIT metal so I dont have a choice the way I'm cutting it. WISH I had a fiber or co2 laser.
09:07 PM _unreal_: could have been done in the blink of an eye
09:07 PM _unreal_: :/
09:10 PM Tom_L: or blind in even less
09:26 PM Tom_L: ... inconsistencies... some comps have inputs that start with '0' and others start with '1' :/
11:07 PM xxcoder8265[m]: quack
11:08 PM XXCoder: nice its fixed
11:08 PM TurBoss8676[m]: 🦆
11:15 PM Tom_L: for how long?
11:15 PM TurBoss: ... 😄
11:15 PM XXCoder: what happened reason is now known so if it happens again, it should be easy fix :)
11:17 PM Tom_L: i'll hold you to it :)
11:17 PM XXCoder: lol