#linuxcnc Logs
Oct 26 2022
#linuxcnc Calendar
02:13 AM Deejay: moin
03:33 AM travis_farmer[m]: G'Morning, 62F with a high of... 62F (odd) and foggy as heck right now.
04:08 AM pere: hi
04:29 AM Tom_L: morning
04:29 AM Tom_L: 42F
04:44 AM KristjanK[m] is now known as krj2680[m]
04:46 AM JT-Cave__ is now known as JT-Cave
04:46 AM -!- #linuxcnc mode set to +v by ChanServ
04:47 AM JT-Cave: morning
06:00 AM -!- #linuxcnc mode set to +v by ChanServ
06:44 AM * c101horse8332[m] uploaded a video: (18922KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/FLMEmnnjqWLSgSVJhFjALmXh/WhatsApp_Video_2022-10-26_at_13.24.45.mp4 >
06:44 AM c101horse8332[m]: anyone that can explain this? im isolated from the machine, holding only the positive leg of the led, all the terminals im touching are for the Y axis, step dir and enable pins
06:49 AM CloudEvil: you have a large AC voltage on you (>>10V) due to mains pickup in some manner and are trying to kill the pins you're connecting to
06:51 AM CloudEvil: (Or the machine has a large voltage on it)
06:56 AM c101horse8332[m]: i was trying to see if my mesa is even sending out pulses to the drivers, i dont have a scope at the ready here thought i'd use a led across the pulse pins, no pulsing, only a flash on startup
06:57 AM CloudEvil: Attach LED with a 1-10K or so resistor to ground, and you can then use it as a probe.
06:58 AM c101horse8332[m]: i have a resistor on it
07:14 AM ZincBoy[CAON][m]: That LED lighting up shows a serious grounding issue. As CloudEvil mentioned there is a voltage present to ground (you). This should not happen and is a safety issue.
07:14 AM CloudEvil: Holding it freehand, with a large voltage across you may damage the inputs
07:14 AM CloudEvil: Also you, but...
07:15 AM ZincBoy[CAON][m]: The other issue with using a LED is that the step pulse is quite short and may not have enough duty cycle to be visible.
07:37 AM c101horse8332[m]: im still wondering how the led can even light up, im not in contact with anything live
07:39 AM c101horse8332[m]: in my mind it should then only light up when im touching the ground pins not all the pins
07:40 AM ZincBoy[CAON][m]: You are ground. A LED only needs a few uA to light up and you are the leakage path to ground. This means there is significant voltage present on that pin. This is a safety issue. You need to figure out why there is significant potential on that node because there should not be any.
07:40 AM c101horse8332[m]: so i should be measuring that pin to ground?
07:42 AM ZincBoy[CAON][m]: Yes. Your machine should be tied to the main ground but I would check that to be sure.
07:43 AM c101horse8332[m]: the controll box is grounded to the mains
07:44 AM ZincBoy[CAON][m]: The machine frame should be grounded as well.
07:44 AM c101horse8332[m]: when measuring from that pin to the mains ground i get 11v ac
07:45 AM ZincBoy[CAON][m]: What about control ground?
07:45 AM Tom_L: google charlieplexing and that might help make sense of it
07:45 AM Tom_L: or not...
07:49 AM c101horse8332[m]: you mean between two pins?
07:49 AM Tom_L: that's what the example is about but consider yourself a pin
07:50 AM ZincBoy[CAON][m]: No, between the mains ground and the ground for the drive control pins.
07:50 AM Tom_L: ok, i didn't read the whole log
07:50 AM Tom_L: so is he using star ground?
07:51 AM Tom_L: there is leakage somewhere
07:52 AM Tom_L: you can get just as inconclusive results from a bad breadboard circuit
07:52 AM Tom_L: breadboards can cause havock too
07:53 AM ZincBoy[CAON][m]: The fact that there is 11V AC on the control pin and it lights up a LED means there is leakage. Could just be capacitive or inductive coupling but still an issue that needs to be resolved.
07:53 AM ZincBoy[CAON][m]: @Tom_L, it is in a machine control cabinet.
07:53 AM Tom_L: i dare say it's poor grounding
07:53 AM ZincBoy[CAON][m]: I suspect that as well.
07:54 AM Tom_L: and lack of shielded wires where necessary
07:55 AM * c101horse8332[m] uploaded an image: (200KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/JSojTQskYltlHAIRlvitDbPT/20221026_145252.jpg >
07:55 AM c101horse8332[m]: The whole back plate is grounded an has worked like this for a year
07:55 AM c101horse8332[m]: Only now that im on a new machine has it stoppen working
07:55 AM Tom_L: maybe you've been lucky for a year
07:55 AM c101horse8332[m]: After it has worked for a day or two
07:56 AM ZincBoy[CAON][m]: Yes, the back plate is grounded. But what about the control logic ground?
07:56 AM Tom_L: fix the grounds as a start
07:56 AM Tom_L: a good pannel will have a ground lug
07:56 AM c101horse8332[m]: its grounded to the mains by the ground terminals at the bottom din rail
07:57 AM ZincBoy[CAON][m]: What voltage do you measure at the control ground pin on the drive?
07:57 AM Tom_L: i'm not saying you may not have a problem but cover all your bases first
07:58 AM Tom_L: where is the braided wire grounded?
07:58 AM Tom_L: not on the ends i see...
07:59 AM ZincBoy[CAON][m]: Star connect all power and ground connections.
07:59 AM Tom_L: ^^
07:59 AM Tom_L: do not make a loop wire and ground ---.---.---.---.
07:59 AM ZincBoy[CAON][m]: @Tom_L, you referring to the shielding wire on the white cable?
08:00 AM Tom_L: yes
08:00 AM Tom_L: or any other signal wire
08:00 AM Tom_L: then route them away from the high voltage areas
08:03 AM c101horse8332[m]: currently the shields are not connected to anything, havent gotten to that as it ran fine without
08:03 AM c101horse8332[m]: :smollface:
08:04 AM Tom_L: i can show you a really shitty looking cabinet that works well
08:04 AM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/rue/Control2.jpg
08:04 AM Tom_L: :)
08:04 AM Tom_L: but everything is grounded
08:05 AM c101horse8332[m]: i trusted my brother on doing a good job on the panel as he does industrial control panels for a living
08:05 AM ZincBoy[CAON][m]: The shielding (or lack thereof) is not the issue causing the 11Vac on the control line. It is a problem with the grounding.
08:05 AM Tom_L: i had problems on mine when i started building that on the sherline
08:05 AM Tom_L: i agree with zinc
08:05 AM c101horse8332[m]: ill have him gander at it again
08:05 AM c101horse8332[m]: as im really out of my depth here
08:05 AM ZincBoy[CAON][m]: c101horse🇳🇱 (@_discord_181695718124290048:jauriarts.org) your panel is very neat but that does not mean the grounding is done properly. Do you have a schematic?
08:07 AM c101horse8332[m]: 🤣 no
08:07 AM c101horse8332[m]: that part is my brothers doing
09:21 AM pere: new linuxcnc was just accepted into debian. :)
09:27 AM CloudEvil: :)
09:27 AM CloudEvil: Hwo new?
09:29 AM Tom_L: yeah, does 'new' come with a version number?
09:30 AM Tom_L: https://tracker.debian.org/pkg/linuxcnc
09:31 AM Tom_L: probably a newer update somewhere..
09:32 AM JT-Cave: a guy from luxomburg ordered a card and used finland as the country to calc price...
09:32 AM Tom_L: ship it to finland
09:32 AM Tom_L: tell him to pick it up there
09:33 AM JT-Cave: I didn't have finland setup right so he selected express flat rate and I don't do that
09:36 AM Tom_L: only a 1600 mi trip
09:56 AM pere: CloudEvil: very new. :)
10:54 AM mrec: does anyone know do those inside gear hobber have a special name?
10:55 AM mrec: https://www.youtube.com/watch?v=EefFxEGVbWo
10:59 AM c101horse8332[m]: welp, seems the Y axis on my mesa is dead, all others pulse nicely
11:07 AM c101horse8332[m]: i might've had something plugged in wrong when troubleshooting something else. is there a way to have it diagosed and maybe repaired?
11:18 AM pere: <URL: https://i18n.debian.org/l10n-pkg-status/l/linuxcnc.html > show the need for more work. :)
11:26 AM jpa-: c101horse8332[m]: you could use a different pin instead, if you only fried one of them
11:28 AM c101horse8332[m]: guess ill have to do some reading and figure out how to tell the software side joint1 is somewhere else now
01:06 PM Tom_L: c101horse8332[m], talk to pcw-home
01:10 PM CaptHindsight[m]: pere: is there a decoder ring for that chart somewhere?
01:11 PM Tom_L: you didn't get one?
01:13 PM CaptHindsight[m]: Tom_L: pretty sure it's an obvious secret
01:14 PM Tom_L: i bet red is bad, yellow meh, green is good
01:14 PM pere: CaptHindsight[m]: the language codes, you mean? Sure. <URL: https://en.wikipedia.org/wiki/List_of_ISO_639-1_codes >. <URL: https://hosted.weblate.org/projects/linuxcnc/ > is a more useful representation, though. :)
01:14 PM CaptHindsight[m]: po vs src/po
01:14 PM jpa-: there is mouseover text
01:15 PM CaptHindsight[m]: Language po
01:15 PM CaptHindsight[m]: docs/po src/po src/po/gmoccapy
01:15 PM pere: and yeah, red is missing, yellow is "there is a suggestion someone should look at" and green is OK.
01:17 PM CaptHindsight[m]: all in black text
01:21 PM CaptHindsight[m]: Directory docs/po src/po src/po/gmoccapy
01:22 PM CaptHindsight[m]: pere: https://hosted.weblate.org/projects/linuxcnc/ far more :)
01:23 PM pere: but I like the compactness of the debian link. :)
01:24 PM pere: with one glance it is obvious which languages are missing which component, and the level of translations across all languages. :)
01:24 PM CaptHindsight[m]: obvious once you know
01:25 PM CaptHindsight[m]: thanks
01:26 PM c101horse8332[m]: Who's that?
01:27 PM CaptHindsight[m]: pere: how many people are working on translations?
01:27 PM pere: CaptHindsight[m]: let me try to count...
01:28 PM CaptHindsight[m]: does debian have a pool of translators?
01:29 PM pere: using the git source, "git log docs/po src/po/|grep Author:|sed 's/^ A/A/'| sort -u|wc" tell me it is 100 different author lines. I got three, so the real count is less.
01:29 PM pere: sorry, I got four.
01:29 PM pere: CaptHindsight[m]: not really a pool, but some people interested in translations.
01:30 PM pere: CaptHindsight[m]: note, the stats I show here are not for debian initiated translations, but translations done by the people involved in linuxcnc.
01:31 PM pere: and this is counting everyone since 2004. at the moment I suspect less than 5 people are active translating.
01:34 PM CaptHindsight[m]: pere: does Weblate actually perform any language translation or is that performed by the human translator?
01:35 PM pere: CaptHindsight[m]: humans.
01:36 PM pere: it provide assistance, but humans do the work. there is an 'automatic translation' option for users to check out and perhaps use.
01:39 PM CaptHindsight[m]: pere: google translate with an editor and version control
01:40 PM pere: CaptHindsight[m]: heh, kind of. :)
01:40 PM pere: CaptHindsight[m]: instead of guessing, you can give it a go and provide translations for a language you know. :)
01:41 PM pere: it is a lot easier to text than to understand the mechanism from me explaining it.
01:41 PM pere: s/text/test/
01:41 PM CaptHindsight[m]: I was wondering how much translating it actually does vs humans
01:42 PM CaptHindsight[m]: but I get how it's a framework, editor, version control etc
01:42 PM pere: my personal experience with google translate for en->nb is that it is not really good enough, but can be helpful.
01:43 PM CaptHindsight[m]: a feature rich computer aided tool for translators
01:43 PM pere: or as I think of it, a fancy PO file format editor. :)
04:22 PM -!- #linuxcnc mode set to +v by ChanServ
04:28 PM JT-Cave: Tom_L, I'm rewriting the loadini.py so allow both the halfile and mdi multiple times and get rid of the source nonsense
05:20 PM Tom_L: not clear on what you mean but.. ok
05:23 PM JT[m]: Loadini.py uses configparser so to have multiple duplicate keys I have to change the way I read the ini file
05:24 PM Tom_L: yeah i remember that about configparser
05:24 PM Tom_L: you rolled your own
05:24 PM JT[m]: I'll use a similar way to the updateini.py
05:25 PM Tom_L: i'm betting the end results will be better than anything else out there
05:25 PM JT[m]: For sure
05:25 PM Tom_L: how hard _would_ it be to add parport support?
05:26 PM Tom_L: for generic configs
05:26 PM Tom_L: treat it just like another card maybe?
05:26 PM JT[m]: Probably not too hard
05:26 PM JT[m]: Possibly
05:26 PM Tom_L: that would sure throw it over the top
05:27 PM JT[m]: I just found some bronze shift fork pads that will be better than the OEM nylon ones
05:29 PM xxcodery: interesting. shortage is so bad theres now rpi zero 2 to rpi3 adaptors
05:29 PM JT[m]: I suppose the parallel port could be like another card
05:29 PM Tom_L: was gonna ask how hard it would be to make some brass ones...
05:30 PM JT[m]: They seem to be cast
05:30 PM Tom_L: worth a shot probably
05:31 PM JT[m]: https://www.hawksmotorsports.com/t56-reverse-bronze-shift-fork-pads-set-upgrade-tremec/
05:31 PM xxcodery: adoptors if curious https://www.cnx-software.com/2022/10/25/adapters-convert-raspberry-pi-zero-2-w-into-raspberry-pi-3-or-4/
05:31 PM Tom_L: those are obviously cnc'd
05:32 PM Tom_L: ran the chamfer tool rpm a little slow
05:33 PM Tom_L: unless it's just the pic angle
05:33 PM JT[m]: I see that now, but for $20 it's not worth the R&D
05:33 PM Tom_L: right
05:33 PM Tom_L: would be a no brainer to make though
05:34 PM JT[m]: Yup
05:34 PM JT[m]: If you had the dimensions
05:35 PM Tom_L: i did a similar part for a guy that wanted a stronger lock recptacle on a desk
05:35 PM Tom_L: i forget how many maybe 200 i dunno
05:36 PM Tom_L: just a block with a slot and 2 recessed screw holes
05:37 PM Tom_L: i'm surprised somebody hasn't tried some sort of composite
05:37 PM JT[m]: I ordered a insulated 1/4" thick tunnel cover and my debit card was hacked
05:37 PM Tom_L: damn
05:38 PM Tom_L: i had somebody get my cc a while back for $800
05:38 PM Tom_L: but i got it resolved
05:39 PM Tom_L: also that fedex overnite i paid for and didn't get... they refused a refund but the cc took it from them anyway
05:39 PM JT[m]: The bank caught it and reversed the charges
05:39 PM Tom_L: that was like $130 shipping
05:39 PM Tom_L: good
05:40 PM JT[m]: About $1500
05:40 PM Tom_L: during covid ppl in Ga couldn't get meds they needed so we shipped some
05:40 PM Tom_L: i think the overnite wound up being 3-4 days
05:47 PM JT[m]: The bank put my debit card on restricted and refused another $1500
05:48 PM xxcodery: debit definitely has less protections
05:48 PM xxcodery: I should get cc or something but meh
05:49 PM xxcodery: so many credit card stuff has fees up woozoo
05:49 PM Tom_L: on the cc they just issued new cards
05:49 PM roycroft: i only use my debit card at a few places, and mostly at atms
05:49 PM JT[m]: Yep, my bank lady says she uses PayPal
05:49 PM roycroft: not only is there better protection when using credit cards, all the credit cards i use have cash rewards programs
05:50 PM Tom_L: as do mine
05:50 PM roycroft: the least rewards i get with any card is 2%, and my amex blue card has 6% rewards on groceries
05:50 PM JT[m]: Send me a CC
05:50 PM roycroft: which takes some of the sting out of inflation
05:50 PM roycroft: but i am the consumer that the cc companies hate
05:50 PM xxcodery: I guess you avoid most fees by pay in full every month
05:51 PM xxcodery: is there other fees
05:51 PM roycroft: i use their cards for almost everything, reap the rewards, and pay in full every month
05:51 PM roycroft: the amex has a $90/year annual fee
05:51 PM roycroft: that's the only card i have with an annual fee
05:51 PM roycroft: but i alwasy get over $500/year back in rewards, so it works out
05:51 PM xxcodery: not too bad then.
05:51 PM roycroft: i do carry a balance sometimes on large purposes
05:52 PM roycroft: but only when the cc company offers me 0% apr
05:52 PM JT[m]: PITA to make monthly payments to pay off a CC
05:52 PM roycroft: i do not believe in paying credit card interest or late fees
05:52 PM JT[m]: Just another thing to remember
05:52 PM roycroft: i set up autopay
05:53 PM ZincBoy[CAON][m]: Autopay FTW!
05:53 PM JT[m]: Hmm that would work
05:53 PM roycroft: and my bookkeeping software sends me email reminders when i need to pay bills
05:53 PM Tom_L: you have to pay them back??
05:53 PM roycroft: i also arrange for the payment due date on all my cards to be a week after i get my paycheck
05:53 PM JT[m]: Sometimes
05:54 PM roycroft: so on payday i deposit the check, and then i have a few days to pay bills
05:54 PM ZincBoy[CAON][m]: Now in Canada the merchants can charge a surcharge if you pay by credit card. Driving more of my purchases onto debit.
05:54 PM roycroft: that's always been the case here
05:54 PM roycroft: but a few years ago we passed a federal law requiring merchants to post any cc surcharges on the entry door to the business
05:55 PM roycroft: they used to just sneak them in
05:55 PM * JT[m] wanders inside for the night
05:55 PM roycroft: but most merchants don't charge the surcharge, as the majority of customers pay with credit cards
05:55 PM roycroft: they just build that cost into the price of the products
05:55 PM roycroft: so if you're paying cash or with a debit card you're subsidizing the cc payers
05:55 PM roycroft: because you don't get the cashback rewards or mileage points that cc users get
05:56 PM roycroft: i know my approach doesn't work for a lot of people who don't have the financial discipline to pay their cards off every month
05:57 PM roycroft: or even know how much money they have
05:57 PM roycroft: but if you keep track of things it works fine
05:57 PM ZincBoy[CAON][m]: Yes, that is why I use the credit card for pretty much everything.
05:58 PM roycroft: and as recently witnessed by my experiences with gst3g, paying with a cc gives one protections that paying with cash or debit does not
05:58 PM roycroft: speaking of which, would anyone like some pla+ that is utterly useless for 3d printing?
05:58 PM roycroft: i have 10kg of it
05:58 PM roycroft: well, more like 9kg now
05:58 PM Tom_L: use it for weed eater cord
05:58 PM roycroft: since i spent a lot of time trying to make it useful
05:58 PM ZincBoy[CAON][m]: Burn it?
05:59 PM Tom_L: it will go fast that way
05:59 PM roycroft: if someone went fargo on the stuff and then fed it to a hopper extruder it might work
05:59 PM roycroft: i think the plastic is probably ok
06:00 PM roycroft: it's just that the extrusion is so uneven it won't work
06:00 PM ZincBoy[CAON][m]: Trying to think if you could create a rocket engine with lengths of filament surrounded by oxidizer...
06:00 PM roycroft: so the guy that runs the local makerspace is a jerk
06:00 PM roycroft: i've known him for almost 30 years, and he's always been a jerk
06:00 PM roycroft: and he screwed me over recently
06:01 PM roycroft: as in costing me dozens of hours of extra time working on our election at a time where i could not spare any
06:01 PM Tom_L: why go if you don't like the guy?
06:01 PM roycroft: i'm tempted to donate it to him :)
06:01 PM roycroft: i actually nothing to do with the makerspace
06:01 PM roycroft: it's just coincidental that he runs it
06:02 PM roycroft: anyway, he deserves this filament
06:02 PM roycroft: but the users of the makerspace don't
06:02 PM roycroft: so i'll not donate it
06:02 PM roycroft: and i won't donate it to the goodwill either
06:02 PM roycroft: they would sell it for $5.99, and someone would buy it, go home, and get really frustrated
06:03 PM roycroft: i like the gst3d boxes better than the atomic filament boxes, so i'll probably keep the boxes
06:03 PM roycroft: and send the filament to the landfill
06:03 PM roycroft: anyway, if i had paid for that stuff with a debit card i'd be out $250
06:03 PM roycroft: but since i used a cc, i got all my money back
06:12 PM xxcodery: I just had a thought
06:12 PM xxcodery: filament feeder torch lol
06:13 PM xxcodery: of course it burns plastic so its not great but silly idea
06:13 PM xxcodery: roy wonder if something like giant bore nozzle .8mm for example wouldnt care about bad filament for prints you wouldnt care on quality on
06:18 PM roycroft: it jams before it gets to the nozzle
06:20 PM xxcodery: oh yeah it was that bad. forgot
06:20 PM roycroft: tomorrow is trash day
06:20 PM roycroft: so some of it will probably go out in the trash then
06:20 PM xxcodery: many uses I could think, pla is too brittle for
06:21 PM roycroft: because i can't think of anything useful that i can do with the stuff
06:21 PM xxcodery: like making mesh for shades or something
06:22 PM roycroft: with a heat gun, it may be possible to sculpt things out of it, kind of like coil built pottery
06:24 PM xxcodery: yeah
06:25 PM roycroft: maybe before i toss all the stuff i'll post a free ad on cl
06:25 PM xxcodery: i bought really hot heat gun from amazon. it wasnt bad, though it does have flaws
06:25 PM roycroft: letting people know that it's not suitable to use as 3d printer filament
06:26 PM xxcodery: I wonder how well it work work for 3d pen
06:27 PM xxcodery: i bought one for $5. got very lucky lol. its kinda fun to play with a bit. well worth $5 but morr than that? eh dunno. theres few very good 3d pen yt channels lol
06:29 PM xxcodery: roy if youre curious on heat gun i can link you to it, though amazon dont load well on rpi4
06:33 PM roycroft: i have a good heat gun
06:37 PM xxcodery: thats good then
06:37 PM xxcodery: what I got isnt amazing but its good enough for what I needed
06:39 PM roycroft: https://www.masterappliance.com/master-mite-heat-gun/
06:39 PM roycroft: that's the one i have
06:40 PM roycroft: i've had it for about 35 years, and it still works like it's new
06:40 PM roycroft: i dropped it on a concrete floor a year or so ago, and cracked the outer case
06:40 PM roycroft: i was able to buy a replacement from master appliance that fits and works perfectly
06:42 PM xxcodery: nice
06:49 PM xxcodery: one I have is pretty small. basically very thick wand
06:49 PM xxcodery: can do 200c and 420c
06:49 PM xxcodery: honestly id derate it by half but still plenty for my needs
06:50 PM roycroft: i have a butane one as well, which i like to use for heat shrink tubing on wiring harnasses
06:50 PM roycroft: it's actually a butane soldering iron that has a heat gun feature
06:50 PM xxcodery: nice
06:51 PM roycroft: https://www.amazon.com/gp/product/B0892S4MDM
06:51 PM roycroft: if you can load that
06:52 PM roycroft: i had a really old one that i bought at radio shack in the '80s, but it finally died a year or so ago
06:53 PM xxcodery: pretty strightforward
06:53 PM roycroft: yeah, but you have rpi issues, so i wasn't sure
06:57 PM xxcodery: I meant the butane thing is pretty strightforward
06:57 PM xxcodery: yeah site load wasnt too bad this time
10:35 PM solarwind: ZincBoy[CAON][m] about to pull the trigger on some DMM DYN4 drives. They have good specs and those guys are responsive. 100µs sample time. Only concern is the ADC, it's 12 bit from -10 to +10V
10:35 PM solarwind: ZincBoy[CAON][m] you think 12 bit is too little for velocity and torque control applications?
10:35 PM solarwind: Not a CNC machine, something like a robotic arm