#linuxcnc Logs

Mar 31 2021

12:11 AM Juzzy: well fuck me.
12:18 AM Juzzy: the mirrored limit didnt work correctly
12:18 AM Juzzy: and it rammed through and broke all sorts of shit
12:18 AM Juzzy: before i could stop, the gantry is paartially derailed
12:18 AM Juzzy: what a fkn mess
01:19 AM Deejay: moin
01:40 AM CaptHindsight: Juzzy: sorry to hear, it happens, always design the limit switches to not be crushed by the end of travel
01:40 AM XXCoder: Juzzy: :( that sucks
01:41 AM XXCoder: hopefully parts is intact or easily fixable
01:41 AM CaptHindsight: always test at low speeds, keep your hand on the e-stop
01:41 AM XXCoder: indeed
01:42 AM CaptHindsight: the machine will just be better made now
01:44 AM CaptHindsight: the gantry should not be able to pull itself off the ballscrews
01:44 AM XXCoder: capt i was thinking on using ballscrew on big hypercube, i think .001mm would be enough precision for 3d printer
01:45 AM CaptHindsight: especially for a 0.02mm orifice
01:46 AM XXCoder: indeed
01:46 AM CaptHindsight: oh sorry, aren't gl;ue gun nozzles 0.2mm and higher?
01:47 AM XXCoder: theres .1 ones but yeah .4 is standard
01:47 AM CaptHindsight: 100x lower res is a bit of over kill
01:47 AM CaptHindsight: no need to step at that rate for a 0.1mm nozzle
01:48 AM XXCoder: i aim for .01 movement resolution
01:48 AM XXCoder: which means i want less than .001mm error
01:48 AM CaptHindsight: why a ballscrew?
01:48 AM CaptHindsight: you don't need the torque
01:48 AM XXCoder: it'll be 500mm^3 cube workspace
01:48 AM XXCoder: so it'll be quite heavy
01:49 AM CaptHindsight: I'd use a belt for non contact like that
01:49 AM XXCoder: both X and Y is belts
01:49 AM XXCoder: not z
01:50 AM CaptHindsight: is the build platform the Z as a load or the extruder?
01:50 AM enleth: Juzzy: consider installing separate home and limit switches, the former handled by linuxcnc, the latter wired into the e-stop circuit. linuxcnc supports home switches that are not at the very limit of an axis and you can install switches in a "run-over", not "run-into" position relative to the gantry
01:50 AM XXCoder: yeh load on z
01:51 AM XXCoder: enleth: yeah i remember that hard limit (switches) should be just behind "true home" of zeros
01:51 AM CaptHindsight: XXCoder: so your build platform moves?
01:51 AM XXCoder: only downwards. far better than bedflinger
04:25 AM Tom_L: morning
05:45 AM -!- #linuxcnc mode set to +v by ChanServ
05:50 AM JT-Cave: morning
07:46 AM drdoc: Morning, all
08:09 AM serp3 is now known as serp
08:40 AM drdoc: given the fact that I'll probably have to replace my 6i25 board, I'm looking at other options.
08:41 AM JT-Cave: 7i92M?
08:42 AM drdoc: I'm curious about the RPi4 and 7c81
08:43 AM drdoc: how's that shaking out, performance-wise?
08:45 AM drdoc: I'm running a shop-built gantry router, 3 axis, 4 motors, no encoders
08:46 AM drdoc: I'm thinking either the Pi+hat, or probably 7i92M with a j1900 thin client
08:48 AM drdoc: maybe a 7i96, but after blowing up the PCIe card I'm a little leary about eggs & baskets
08:50 AM drdoc: any input would be welcome.
09:06 AM mrec_: can anyone recommend a good parting tool for a mini lathe?
09:09 AM rs[m]: drdoc: maybe previews/backplots are a bit slower than on a PC and main memory is limited to 8gig max, but there should not be any issues regarding the machining
09:10 AM rs[m]: i'm not sure what the latest status of linuxcnc on the 8gig rpi4 is
09:12 AM drdoc: that's good to know. I'm also seeing a lot of references to using ssh -X for "remote" display to avoid running X on the linuxcnc machine
09:12 AM drdoc: does that help or hurt general performance?
09:20 AM drdoc: I s'pose I could dig out the Arduino and go back to grbl
09:20 AM drdoc: ;)
09:24 AM JT-Cave: I have a Rpi4 and a 7c81 but have not had the time to test it out
09:24 AM JT-Cave: I did finally get some idc40 connectors to make a cable lol
09:24 AM drdoc: hah!
09:25 AM JT-Cave: 7i96 is my go to board for stepper systems
09:25 AM * JT-Cave goes to finish his morning chores
09:26 AM drdoc: is it just a 7i92 with builtin breakout, ro does it offer more?
09:41 AM * drdoc gets impatient, orders the damn thing
12:01 PM veegee: This is a bit silly, but what kind of jack is this: https://www.reddit.com/r/ProgrammerHumor/comments/mh2f67/answer_to_a_coding_interview_question_how_do_you/
12:02 PM veegee: It's a bottle jack, yeah, but has a big sturdy base
12:02 PM veegee: I guess I could just make a base and stabilizers for my own bottle jacks
12:05 PM gregcnc: seems to be a thing in the offroad world https://www.rv4x4.net.au/base-plate-big-boy-205mmx460mm
12:19 PM FinboySlick: gregcnc: Those wingnut bracket things don't seem very sturdy.
12:19 PM gregcnc: no idea
12:20 PM CaptHindsight: better than loose soil or mud
12:20 PM CaptHindsight: "hey go find me a flat rock to put the jack on"
12:21 PM gregcnc: i have to level a container so i've been looking around for something cheap
12:21 PM FinboySlick: The plate seems like a good idea, it's just that if you'Re stuck in mud or what not, I'd imagine you'd want the jack better secured to the plate.
12:22 PM CaptHindsight: gregcnc: put some "self levelling" concrete under it :)
12:22 PM gregcnc: we've had plenty of ideas from the other club members....
12:25 PM CaptHindsight: gregcnc: what do you have to move the container?
12:25 PM gregcnc: nothing
12:25 PM gregcnc: it's out in the middle of a field
12:27 PM CaptHindsight: 20 ft?
12:27 PM gregcnc: I guess they rented railroad jacks the last time they did it, but it was sketchy
12:27 PM gregcnc: 40
12:43 PM CaptHindsight: https://www.newegg.com/asrock-j4005b-itx-mini-itx/p/N82E16813157803 $77
12:50 PM FinboySlick: CaptHindsight: That's a pretty neat little board.
12:51 PM FinboySlick: All it's missing would be a 12v direct power connector so one can skip the PSU but that's nitpicking.
01:00 PM CaptHindsight: https://www.newegg.com/p/pl?d=pico+power+supply
01:17 PM FinboySlick: Wow, 450W for 40 bucks. I wonder how long that would last.
02:20 PM roycroft: so the sound-dampening mat arrived yesterday
02:21 PM roycroft: i just installed a decibel meter app on my ipad
02:21 PM roycroft: i'll get before and after measurements when i install it on my jointer
02:21 PM roycroft: the results might be interesting
02:22 PM roycroft: of course, when it's actually cutting, it's going to be really loud, and the dampening mat won't help with that at all
02:22 PM roycroft: but when it's not cutting there's a rather loud, low-pitched rumble and some vibration noises that i hope will go away
02:23 PM roycroft: my thicknesser has those same background noises - if it works well on the jointer i'll apply some to the thicknesser as well
02:34 PM enleth: do you know a fool-proof and convenient way to ensure people remember to close valves on welding gas bottles when they're done welding?
02:34 PM enleth: some sort of a flag on the bottle, like on a front porch mailbox?
02:37 PM gloops: no, you have to have baggings that dont leak
02:39 PM rs[m]: there is not. nature would immediately invent a better fool
02:52 PM gregcnc: enleth bill them for the gas
03:01 PM roycroft: enleth: put a big sign on the bottles saying if they are left open after use the user will be banned from the facility
03:01 PM roycroft: or do like i do - do not permit others to go near the welding gear
03:04 PM bjorkintosh: do all those cheap 3040s and < $300.00 pencil mills use linuxcnc?
03:06 PM enleth: gregcnc: problem is, *I* sometimes forget about it, too
03:07 PM enleth: roycroft: ^
03:09 PM -!- #linuxcnc mode set to +v by ChanServ
03:09 PM -!- #linuxcnc mode set to +v by ChanServ
03:10 PM enleth: I've seen someone mention somewhere that they've been getting gas from a place that installs ball valves, so the handle acts a very visible status flag
03:10 PM enleth: problem is, I've never seen a ball valve rated for gas cylinders
03:52 PM CaptHindsight: <FinboySlick> Wow, 450W for 40 bucks. I wonder how long that would last.
03:52 PM CaptHindsight: likely a long time with enough air flow and only drawing 250W max
03:52 PM XXCoder: 450w for 40 usd what
03:52 PM CaptHindsight: https://www.newegg.com/p/pl?d=pico+power+supply
03:53 PM XXCoder: ahh
03:53 PM CaptHindsight: https://www.newegg.com/p/2S7-01NV-00CD5
03:53 PM XXCoder: too bad, i thought it was power generation
03:56 PM CaptHindsight: 450W nuclear reactor will cost you much more until they get production volumes up
03:57 PM CaptHindsight: https://en.wikipedia.org/wiki/SNAP-10A
04:36 PM buzzmarshall: CaptHindsight: sorry i bugged out the other nite but i had fix the build server and get up for 5am
04:36 PM buzzmarshall: but i am looking at the github and routing for my rk info
05:06 PM roycroft: well, my mobile base is all pretty and painted
05:06 PM roycroft: but hammerite takes about 2 weeks to fully cure, so i might not be able to install it for a while
05:07 PM roycroft: i'm kind of bummed, because i wanted to do a dB reading of the jointer as it sits, a reading on the mobile base, a reading with the dampening mat as it sits, and a reading with the dampening mat on the mobile base
05:07 PM roycroft: and i don't want to wait 2 weeks to install the dampening mat
05:07 PM roycroft: but perhps i'll wait anyway
05:08 PM roycroft: i doubt putting it on the mobile base will make more than about 1.5 dB difference in the ambient noise level
05:11 PM roycroft: my gut feeling is that it will be a wee bit louder when it sits on the mobile base
05:27 PM _unreal_: FINALLY got my phone case
05:30 PM CaptHindsight: buzzmarshall: https://linux-sunxi.org/AR100
05:31 PM CaptHindsight: buzzmarshall: no problem, did you see the logs after you left?
05:38 PM _unreal_: NoGodDamnIdea, you live?
05:39 PM NoGodDamnIdea: sadly yes
05:39 PM NoGodDamnIdea: waddap
05:39 PM _unreal_: so how goes the project?
05:39 PM NoGodDamnIdea: well I made some metal supports today for the endstops
05:39 PM NoGodDamnIdea: but some issues appeared at work and didnt get to install them
05:39 PM NoGodDamnIdea: because fuck me right
05:39 PM _unreal_: with luck you made it so they have some adjustment ya? so you can TUNE the point of limit
05:40 PM NoGodDamnIdea: well its pretty much done
05:40 PM _unreal_: do you have a spindle yet?
05:40 PM NoGodDamnIdea: just these supports needs to be made so the Z endstop limits the Z axis travel
05:40 PM NoGodDamnIdea: yes I have had a spindle since forever but I do not have a VFD for it
05:41 PM _unreal_: then you dont have a spindle for it LOL
05:41 PM NoGodDamnIdea: first thing im gonna do with it is cut aluminium foil anyways
05:41 PM NoGodDamnIdea: yeah I guess I dont
05:41 PM _unreal_: foil?
05:41 PM _unreal_: aliexpress is a good source for VFD's
05:41 PM _unreal_: just have to wait a month
05:41 PM NoGodDamnIdea: yes foil
05:41 PM _unreal_: unless you can $ afford shipping
05:41 PM NoGodDamnIdea: I am trying to make planar drivers
05:41 PM _unreal_: like tin foil?
05:42 PM NoGodDamnIdea: pretty much
05:42 PM NoGodDamnIdea: just thinner
05:42 PM _unreal_: for what ? speakers?
05:45 PM NoGodDamnIdea: that too
05:45 PM NoGodDamnIdea: earbuds :D
05:45 PM NoGodDamnIdea: there is this niche I think has potential
05:45 PM NoGodDamnIdea: so I am trying that
05:45 PM NoGodDamnIdea: if that wont work, I can try for speakers
05:46 PM _unreal_: NoGodDamnIdea, https://photos.google.com/share/AF1QipPp1f0dUHlqqTwZCQ32JntR_n59omq5yjLi0sI1AFxK8ghJ-Fgv4MRm1h6_kwMyBw/photo/AF1QipPhxMC9W7kfJG5kVfu8KEEhbzj1GtsFuqBgo3lx?key=b196bk50Mk9GMFd4ZFBSY0RkclVCVW9QYm0tbWFB
05:46 PM _unreal_: I designed that
05:46 PM _unreal_: https://photos.google.com/share/AF1QipPp1f0dUHlqqTwZCQ32JntR_n59omq5yjLi0sI1AFxK8ghJ-Fgv4MRm1h6_kwMyBw/photo/AF1QipNGF4oC8nAbKTu3KckPmJmX4Q_L3NklQcya4oxX?key=b196bk50Mk9GMFd4ZFBSY0RkclVCVW9QYm0tbWFB
05:46 PM _unreal_: its a pen holder
05:47 PM _unreal_: pen holder with "" probe "" function
05:47 PM NoGodDamnIdea: lol im using the exact same switches
05:47 PM NoGodDamnIdea: pretty cool
05:49 PM _unreal_: https://drive.google.com/file/d/1RdHmmcXoBX_gtgXFcT8jmDTXr2A8LFLe/view?usp=sharing
05:49 PM _unreal_: that is my daughters cnc machine
05:49 PM _unreal_: she "" helped "" build it
05:57 PM buzzmarshall: CaptHindsight: ya ive read the ar100 link but haven't gone and looked at the logs yet
05:58 PM buzzmarshall: later i will make some time to find the full docs on the H6 as well
05:59 PM buzzmarshall: the nda'd material will be sitting on someones server someplace just need to find it
05:59 PM buzzmarshall: specially the register manual
06:01 PM buzzmarshall: what would be decent sbc style of board with the H6 on it?
06:02 PM _unreal_: ve7it, have you looked at my laser etcher project?
06:04 PM ve7it: yes... looks like fun... wish I had a dedicated laser engraver... I swap the laser on and off my 3d printer, works, but extra time involved.
06:05 PM _unreal_: those BLDC motors I gave you would be PERFECT for a laser
06:05 PM _unreal_: good encoder setup
06:05 PM _unreal_: with a FAST controller
06:05 PM _unreal_: ve7it, speaking of old. boy.... what have you done with those motor/controllers since?
06:06 PM _unreal_: I'm guessing NoGodDamnIdea passed out LOL
06:07 PM NoGodDamnIdea: nah im just sad
06:07 PM NoGodDamnIdea: and searching for some books that I cant find a pdf of
06:07 PM _unreal_: THAT i showed you the "coolest" toy ever for a 12 yo and my daugher is like ya thats nice. OH LOOK prity thing on my phone
06:08 PM NoGodDamnIdea: pretty solid
06:08 PM buzzmarshall: CaptHindsight: i was thinking about the pine h64 for that
06:08 PM ve7it: not much... been resurecting old S100 bus machines and relearning z80 assembly language programming. Its fun to work on hardware where you can just about see the bits moving about!
06:08 PM _unreal_: one bit at a time even
06:10 PM ve7it: current exercise is reverse engineering a cpu board that I cant find a manual for. I need to know how to set the eprom address, power on jump address and the i/o addresses.
06:11 PM _unreal_: ve7it, have you tried searching archive.org
06:12 PM _unreal_: ve7it, better yet is there/WASS..... there a web site for the board?
06:12 PM ve7it: yes.... seems the board was only made for a couple of years in the early 80's... I found byte magazine ads for it, but that is all.
06:12 PM _unreal_: OH 80's
06:12 PM _unreal_: :S
06:12 PM ve7it: way before google!
06:13 PM _unreal_: I'm about to order this https://www.amazon.com/gp/product/B07C7LDKXZ/ref=ox_sc_act_title_1?smid=A2DGP45YASMA2M&psc=1
06:15 PM _unreal_: HAY any suggestions from anyone on a good fan to order from amazon? 120mm range for dealing with smoke from my laser etcher enclosure?
06:16 PM ve7it: yup... I have mine by a window and use a bank of 3 big cooling fas as an exhaust system
06:21 PM _unreal_: think I like this one https://www.amazon.com/ASHATA-Wind-Force-250-3CFM-Dissipation-Antminer/dp/B07K8Q7NBY/ref=sr_1_25_sspa?dchild=1&keywords=PFC1212DE&qid=1617232003&refinements=p_85%3A2470955011&rnid=2470954011&rps=1&s=electronics&sr=1-25-spons&psc=1&smid=A4IWNATC64MMF&spLa=ZW5jcnlwdGVkUXVhbGlmaWVyPUExVTY0UVhGSkw3OTI0JmVuY3J5cHRlZElkPUEwOTYxMzI5UjZIUUVRSE85UlVRJmVuY3J5cHRlZEFkSWQ9QTA0MDc1ODIxVFVPMzhaMEhaRFBMJndpZGdldE5hbWU
06:21 PM _unreal_: 9c3BfYnRmJmFjdGlvbj1jbGlja1JlZGlyZWN0JmRvTm90TG9nQ2xpY2s9dHJ1ZQ==
06:21 PM _unreal_: think I like this one https://www.amazon.com/ASHATA-Wind-Force-250-3CFM-Dissipation-Antminer/dp/B07K8Q7NBY/
06:22 PM CaptHindsight: buzzmarshall: https://wiki.pine64.org/wiki/PINE_H64
06:22 PM CaptHindsight: buzzmarshall: RAM Memory Variants: 1GB, 2GB, and 3GB LPDDR3
06:22 PM _unreal_: ok ordered that CONE and fan
06:22 PM _unreal_: should be good
06:23 PM _unreal_: I'm going to have to setup ether a PIC or atmel chip to control that fan and the pump RPM's
06:23 PM CaptHindsight: buzzmarshall: LCNC will run on the 1GB version, the issue with RAM is running out when you run a big program, the backplots will eat it all up
06:25 PM _unreal_: buzzmarshall, you want 4gb minimum really 2gb is the least you should ever run lcnc with
06:25 PM buzzmarshall: CaptHindsight: whatever i buy i will make sure to buy with the most ram available which is usually what i do
06:25 PM _unreal_: buzzmarshall, I ran a tinkerboard first gen with 7i96 just fine
06:26 PM roycroft: you don't need 128GB of ram for linuxcnc machine
06:26 PM roycroft: 8GB is fine for almost anything you would be doing with it
06:26 PM _unreal_: roycroft, who said 128gb is ideal?
06:26 PM buzzmarshall: most of my current boards are all 4g as i like to run a few custom tools under the linux kernel that lets me do somethings
06:26 PM roycroft: buzzmarshall said 'with the most ram available'
06:26 PM _unreal_: took him at his word I see ;)
06:27 PM roycroft: i have several servers with 64GB of ram that can take 128GB or 256GB
06:27 PM roycroft: i was just suggesting a realistic number
06:27 PM buzzmarshall: sorry i meant embedded linux boards
06:27 PM _unreal_: 128 is realistic, realisticly HUGE lol
06:27 PM roycroft: i do not run linuxcnc on any of those
06:27 PM buzzmarshall: my build servers here all have at least 128g in them
06:28 PM CaptHindsight: buzzmarshall: have you looked deeply into the machinekit stepping code for the PRU's?
06:28 PM buzzmarshall: i am only interested in LinuxCNC on embedded boards
06:28 PM _unreal_: Well now I feel like an underdog. I think the most ram I have in any machine is like 16gb
06:28 PM buzzmarshall: CaptHindsight: not very deeply... i messed a bit with mk when i got the original bbb
06:28 PM buzzmarshall: but didn't really care for it
06:29 PM buzzmarshall: it seems more like something for more platforms then i am interested in
06:29 PM CaptHindsight: buzzmarshall: wondering if that code is easily adapted to the ARISC and the Cortex M0 in the allwinners and rk3399
06:30 PM buzzmarshall: and found most of the embedded stuff i am interested in not there so i thought i would just go to the source and look at linuxcnc as thats really my interest
06:30 PM _unreal_: CaptHindsight, I published a buildon that
06:30 PM _unreal_: based on the tinkerboard its on the armbian forums
06:30 PM buzzmarshall: CaptHindsight: i am sure its adaptable but how easy i am not sure
06:31 PM _unreal_: buzzmarshall, I had to build lcnc for tinkerobard from source
06:31 PM buzzmarshall: either way tho its good to have examples to learn from
06:31 PM _unreal_: after shutting down a lot of usless services. the latency test was great
06:31 PM CaptHindsight: buzzmarshall: wondering if the machinekit PVR should be ported vs starting over for either ARISC or the M0
06:31 PM buzzmarshall: thats cool
06:31 PM _unreal_: 20,000-30,000 range
06:32 PM _unreal_: with 4 glxgears a you tube video and I forget what else
06:32 PM CaptHindsight: Orangecnc wrote their own but it doesn't work properly on the ARISC
06:33 PM _unreal_: buzzmarshall, again that was tied with a 7i96
06:33 PM CaptHindsight: yet another crop of firmware to go through
06:33 PM buzzmarshall: for me currently the quickest would be to take on the rk as i have the full register manual and a partial soc model built for IDA Pro that lets me work at assembly level
06:33 PM buzzmarshall: also helps when reversing their code
06:33 PM -!- #linuxcnc mode set to +v by ChanServ
06:33 PM buzzmarshall: so having jtag access to the board lets me see whats actually going on
06:34 PM _unreal_: buzzmarshall, I wanted to use the GPIO on the tinkerboard but the coding was byond me and the person who was going to help ended up disapearing
06:34 PM _unreal_: gpio as STEP/DIR signals
06:34 PM CaptHindsight: buzzmarshall: the rk3399 manual did not have a diagram of how the Cortex M0 is connected to the rest of the SOC
06:34 PM buzzmarshall: the H6 i dont enough info on yet to work in the way i normally do
06:35 PM Juzzy: do you have to set a max (pos) limit when setting a min (neg) limit or is there a trick to assigning all of the max travels to a same pin?
06:35 PM buzzmarshall: CaptHindsight: ya the nda'd material has all of that
06:35 PM buzzmarshall: as well as the undocumented registers or at least most of them
06:36 PM buzzmarshall: rk like the amlogic stuff typically never gives out all the real important stuff so one gets bogged down trying to figure things out the hard way
06:36 PM buzzmarshall: i am assuming allwinner will be similar
06:37 PM buzzmarshall: rpi's are as well but i do have a good amount of broadcom docs
06:37 PM buzzmarshall: just not a fan of the rpi's
06:37 PM buzzmarshall: good board just not my thing
06:37 PM CaptHindsight: me neither, rather broadcom in general
06:38 PM buzzmarshall: to be blunt i hate their crap
06:38 PM buzzmarshall: lol
06:38 PM buzzmarshall: most of the old sat tuners i used to hack were broadcom crap
06:38 PM buzzmarshall: i preferred stm's stuff as the tools were easier to find
06:39 PM CaptHindsight: 15-20 years ago I talked to them about an ARM SOC, it was all vapor
06:39 PM buzzmarshall: what i would like to do is adapt a tool i built a few years ago to sit under linux on the embedded boards
06:39 PM buzzmarshall: and use it in a similar manner as rtai
06:40 PM CaptHindsight: they just poked around looking into what potential customers might be up to
06:41 PM buzzmarshall: ya there one of those old style companies that are really hard core in maintaining control and give out very little thats not necessary to any of the downstream devs
06:41 PM buzzmarshall: broadcom that is
06:41 PM CaptHindsight: they had a program called Beachead
06:42 PM CaptHindsight: how they got inside very cable box and sat TV tuner
06:42 PM CaptHindsight: every/very
06:43 PM CaptHindsight: anyway I just ignore anything they make
06:43 PM CaptHindsight: useless due to the control issues, like Intel
06:43 PM buzzmarshall: ya i agree
06:44 PM buzzmarshall: when i was messing the rk stuff i liked there stuff but was into some other things
06:45 PM buzzmarshall: the H6 tho and the ar100 is interesting
06:45 PM buzzmarshall: i really only looked at the bbb's because the idea of the pru seems interesting
06:46 PM buzzmarshall: never seen any of the other socs i played with have anything like that
06:46 PM CaptHindsight: https://github.com/LinuxCNC/linuxcnc/tree/andypugh/bb_pru/src/hal/drivers/hal_pru_generic
06:46 PM buzzmarshall: but for motion control i could see it eliminating the need for a asic/fpga/cpld
06:46 PM CaptHindsight: I didn't know that the PRU's were also used for reading the encoders for closed loop
06:47 PM buzzmarshall: ya i got that last night from you and started to look thru it
06:47 PM buzzmarshall: actually if your into assembly code you can do quite a bit with them
06:47 PM drdoc: ve7it: do you know "glitch", of GlitchWorls?
06:48 PM CaptHindsight: orangecnc put the firmware back up for the ARISC https://gitlab.com/orangecnc
06:48 PM buzzmarshall: i get why the motor/drive companies want to use fpga to protect their product but having a soc with the ability to do all of that seems more interesting
06:49 PM CaptHindsight: buzzmarshall: I think orangecnc does not keep track of Step timing well enough, it drifts over time
06:50 PM buzzmarshall: at some point i need to grab a orange product as they seem to be catching on... just something ive not yet bought
06:51 PM CaptHindsight: "the lost steps are in the arisc stepgen and the arisc step gen doesnt respect the preconditions of the dir_setup and doesnt respect the postcondition of dir_hold."
06:51 PM buzzmarshall: as someone thats been at this cnc stuff much longer then me... what can you see as the max'd out clock/pulse chain speed?
06:52 PM CaptHindsight: the integrated microcontrollers have been tested at up to ~400KHz
06:52 PM CaptHindsight: I think the Mesa cards step to ~1MHz
06:52 PM buzzmarshall: something with enough speed for a bit of future expansion
06:53 PM drdoc: _unreal_: you around?
06:53 PM buzzmarshall: k... ya i was thinking in around 1mhz
06:53 PM drdoc: CaptHindsight: the website claims 10MHz
06:54 PM buzzmarshall: but if the codes done in assembly i am sure faster should be doable as the socs these days really crank things out
06:54 PM CaptHindsight: buzzmarshall: it depends on the machine, but say you have 10um per step, how many meters per second do you want to jog?
06:54 PM CaptHindsight: drdoc: thanks, I always forget
06:54 PM drdoc: no problem
06:55 PM drdoc: I reviewed them all this morning
06:55 PM drdoc: :(
06:55 PM CaptHindsight: ballscrews only move so quick
06:55 PM drdoc: Yes. I'm still butthurt about smoking that 6i25
06:55 PM CaptHindsight: and if you are using linear servos you likely won't care about a ARM SOC vs PC for control
06:56 PM buzzmarshall: to be honest as a newbie at this i am not sure what to expect that would be reasonable
06:56 PM CaptHindsight: the highest speed open loop steppers would be for glue gun printers, inkjets and similar non-contact applications
06:56 PM buzzmarshall: and honestly i really only went down this road as i got sick of marlin and all the other old software overusing the atmels that made me start looking at going my own way
06:57 PM Juzzy: considered odroids?
06:57 PM CaptHindsight: moving a CNC at 2m/s with 10um per step = 200,000 steps/s
06:57 PM buzzmarshall: my 4-color printer uses open loop but the newer machine i am on now ive changed over to close-loop to learn more about better control
06:58 PM * CaptHindsight back in 5 minutes
06:58 PM buzzmarshall: np
06:58 PM buzzmarshall: actually im gonna go make a jd
07:06 PM buzzmarshall: juzzy... i have a couple of n2's from hk and got pee'd off over their poor quality so i stopped buying their products and never got around to any of the other boards
07:06 PM Juzzy: ah. I have 5x xu4's no complaints here
07:07 PM Juzzy: one does RT audio streaming over wifi
07:07 PM buzzmarshall: they used to make a decent product but i don't know what happened as their reflow's seem to get worse
07:07 PM Juzzy: 4 in a k8s cluster
07:08 PM CaptHindsight: Jepler got LCNC working on an odroid 5-7 years ago
07:09 PM CaptHindsight: connected to a Mesa card though over SPI
07:10 PM buzzmarshall: i wanted to get a h2 but after 2 bad n2's out of 2 and then the whole paying CE fees thing abunch of use just boycoted hk
07:10 PM Juzzy: so i cant figure out why my mirrored axis plows through my home switch
07:10 PM Juzzy: ive tested the switch and it's NC and working at the controller
07:11 PM Juzzy: the other 2 home switches are working fine, Y and X
07:11 PM CaptHindsight: Juzzy: I thought that Y was the gantry with the 2 home switches
07:12 PM Juzzy: it does, 1 stops the other just keeps on going
07:13 PM Juzzy: it's configured "correctly"
07:13 PM Juzzy: on pin 13...
07:13 PM CaptHindsight: Juzzy: post your issues on the forums with your current INI and HAL files
07:13 PM Juzzy: aight
07:14 PM CaptHindsight: you sound really close, someone might see the problem right away
07:14 PM CaptHindsight: I'm in the middle of 10 other things
07:14 PM Juzzy: aye.... I was hoping it was just some little thing im missing
07:14 PM Juzzy: i am close, cant believe that dang driver dipswitch issue, omg
07:15 PM CaptHindsight: it happens
07:16 PM -!- #linuxcnc mode set to +v by ChanServ
07:17 PM Tom_L: Juzzy, http://linuxcnc.org/docs/2.8/html/config/ini-config.html#_joint__lt_num_gt_section ( HOMING_IS_SHARED )
07:18 PM CaptHindsight: buzzmarshall: some people microstep way way down to under their machines mechanical precision
07:18 PM Tom_L: http://linuxcnc.org/docs/2.8/html/config/ini-homing.html#sec:homing-section
07:18 PM CaptHindsight: so worst case might be 1um steps at 2m/s = 2,000,000 steps/s
07:18 PM Tom_L: Juzzy, set the home sequence negative for those joint switches
07:18 PM Juzzy: yea it's -2
07:18 PM CaptHindsight: lots of steps for nothing
07:20 PM CaptHindsight: Juzzy: I think I asked before but forgot the answer, are your 2 home switches in parallel with each other or in series?
07:20 PM Juzzy: parallel
07:20 PM CaptHindsight: so you either need 2 inputs, one for each switch or...
07:21 PM Juzzy: i have 2 inputs.
07:21 PM CaptHindsight: ok
07:21 PM Juzzy: whats the hal gui that shows the pins?
07:21 PM CaptHindsight: then it sounds like a HAL or wiring issue
07:21 PM Juzzy: red/green debug tool
07:22 PM CaptHindsight: do you see a state change when either Home switch is activated?
07:22 PM Juzzy: its a cli command, not a menu one
07:22 PM Juzzy: yea i was going to test it, make sure it isnt a bad solder joint on the board or something
07:23 PM Juzzy: also, can i possible manually home a gantry? seems like i wouldnt be able to if I have to use the sequence = -2 to bind them
07:24 PM Juzzy: right now i have a stepconf up with "x" and "y" confiugred for RY and LY so i can move them back off the switches and repeat
07:27 PM CaptHindsight: http://linuxcnc.org/docs/html/hal/tools.html#_halshow
07:28 PM CaptHindsight: http://linuxcnc.org/docs/html/hal/tools.html#_halreport
07:28 PM Juzzy: ahhh ya there ya go thanks
07:32 PM Juzzy: ah found the old test hal files for halrun
07:32 PM Juzzy: so the pins are working correctly and mapped correclty
07:33 PM Juzzy: so gotta be something wrong somewhere else
08:00 PM drdoc: it sounds like ARM support has changed a *lot* since I was last active
08:23 PM -!- #linuxcnc mode set to +v by ChanServ
08:59 PM Juzzy: ffs im about to give up on this shit
09:01 PM Juzzy: so just focusing on X... I did option 3: touch the switch, then back off.
09:03 PM Juzzy: https://youtu.be/kUHJc1rySCo
09:03 PM Juzzy: touches the switch and just keeps going until i abort
09:04 PM Juzzy: makes no sense that it doesnt reverse direction. I have -0.5 but thats what it says to use, a negative number
09:07 PM drdoc: Juzzy: single motor?
09:08 PM W1N9Zr0: if your switch is on the left (X-), your HOME_SEARCH_VEL should be negative and HOME_LATCH_VEL positive
09:10 PM drdoc: http://linuxcnc.org/docs/html/config/images/emc2-motion-homing-diag.png
09:11 PM drdoc: that diagram helped me a lot
09:16 PM W1N9Zr0: diagram shows homeing switch in axis+ direction, so for axis- switches the signs for both velocities must be swapped
09:25 PM Juzzy: wait, whaaat?!
09:26 PM Juzzy: it searches in the correct direction... I have it inverted/not inverted correctly
09:26 PM Juzzy: I swapped it back to option 2 though seems to work on X just not Y
09:30 PM drdoc: silly question - are all your switches wired the same? All NO or all NC?
09:34 PM W1N9Zr0: did you use Machine -> HAL Meter to check the status of the joint.N.home-sw-in and confirm it's not inverted?
09:37 PM W1N9Zr0: https://linuxcnc.org/docs/devel/html/config/ini-homing.html#_home_search_vel read the description, the sign of your SEARCH_VEL must be in the direction of your home switch. if your machine is moving X- with a positive SEARCH_VEL, your switch is either not wired in properly, or inverted
09:38 PM W1N9Zr0: you should also turn HOME_USE_INDEX = FALSE to simplify debugging, even if you have your encoder configred correctly already
09:42 PM -!- #linuxcnc mode set to +v by ChanServ
09:43 PM XXCoder: heyy jt can talk now!!
09:43 PM XXCoder: ;)
09:43 PM XXCoder: hows things jt
09:52 PM veegee: https://app.aws.org/forum/topic_show.pl?tid=17839 lol forks really aren't that expensive
09:52 PM veegee: why anyone would even consider welding a broken one is beyond me
09:54 PM drdoc: ugh, especially at the bend
09:55 PM drdoc: when I was doing shop maintenance for a red iron company, the site crew sent me a design for a boom extension for their forklift
09:56 PM XXCoder: " Thanks! guys for putting my head back on my shoulders, this is one job I will be leaving alone. I would rather have a mad customer than a hurt customer, because no matter how well I repair the fork sooner or later he will over stress the fork again and it will break again."
09:56 PM XXCoder: glad
09:56 PM drdoc: Essentially an 8' boom on tubular rails, adjustable angle, with a fixed chain & hook
09:57 PM drdoc: they wanted it rated at 5000 pounds
09:58 PM drdoc: when I asked if they were serious, they insisted that the forklift is rated to 12,000 pounds
10:00 PM * drdoc pointed out that that rating was 18" from the throat of the fork, and please do that math, guys
10:00 PM drdoc: they shut up and went away
10:02 PM CaptHindsight: just weld an extra 5 tons of weight to the back end
10:03 PM drdoc: CaptHindsight: these guys were setting 40 & 55 foot trusses
10:03 PM drdoc: using a forklift instead of a crane was unsane to begin with
10:04 PM drdoc: I was like the welders in that thread - I Do Not Want this on my dime.
10:06 PM drdoc: are cheap ball screws better than good lead screws?
10:07 PM Tom_L: i think so
10:07 PM XXCoder: go for minium C level you want
10:08 PM XXCoder: c7 is .001mm along 200mm for example
10:08 PM Tom_L: those become not cheap
10:08 PM XXCoder: mystery ones can be whole mms off
10:08 PM XXCoder: tom c7 is $40 something for 500mm
10:09 PM drdoc: it seems that the ball *screw* isn't where the expense is
10:14 PM drdoc: hah!
10:14 PM drdoc: "C7 ... Country of Origin: Unknown"
10:15 PM Tom_L: https://tech.thk.com/en/products/pdf/en_a15_011.pdf
10:15 PM drdoc: thank you sir
10:17 PM CaptHindsight: drdoc: also depends on the application
10:18 PM Tom_L: http://www.cbmind.com/linear/thk/pdf/Ballscrew%20Product%20Specifications.pdf
10:19 PM drdoc: shop built router, mostly doing PCBs, epoxy laminates & polymers
10:19 PM drdoc: maybe a little aluminum & brass
10:19 PM Tom_L: lots of grit
10:20 PM drdoc: I'll be using a mister for anything the dust collector can't manage, but yeah
10:20 PM drdoc: lightweight grit at that
10:20 PM Tom_L: FR4 is rather abrasive
10:20 PM drdoc: yup
10:21 PM drdoc: I'm thinking about making a separate machine for that
10:21 PM CaptHindsight: https://www.hennigworldwide.com/telescopic-springs
10:22 PM drdoc: with flood cooling
10:23 PM drdoc: for one thing, a PCB mill needs to be faster than strong, and the stresses are way different from 2.5D milling
10:24 PM CaptHindsight: use linear servos
10:24 PM drdoc: instead of steppers?
10:25 PM CaptHindsight: the price is much higher for new retail parts
10:27 PM drdoc: but?
10:27 PM CaptHindsight: bargins on ebay
10:28 PM drdoc: what's the advantage of servos? I know nothing about them except as fancy solenoids
10:28 PM CaptHindsight: https://www.youtube.com/watch?v=ICN2iO3nbiQ
10:29 PM drdoc: ummm
10:29 PM drdoc: you have anything I don't need to hear?
10:29 PM drdoc: <--- deef
10:39 PM CaptHindsight: https://www.youtube.com/watch?v=TPCgbfWZ6IQ
10:47 PM roycroft: servos are fast
11:09 PM Juzzy: i may have found my problem with my ganry mirror issue...
11:10 PM Juzzy: I noticed when i did the stepconf i could move in both directions, but i hit PLAY on the defaul job and cancelled it asap and i could move the gantry w/o homing it, both FRD and Back both moved in the same direction
11:10 PM Juzzy: so i looked and saw I had both step and dir inverse
11:11 PM Juzzy: however, I removed step inverse and just left direction.... and now it moves the opposite way and the other Y doesnt move at all while i see joint errors
11:11 PM Juzzy: getting closer though
11:15 PM Juzzy: I have a feeling i need to parport.0.pin-XX-out-reset, but the docs are not super clear when to use reset and when not to
11:16 PM Juzzy: from: http://linuxcnc.org/docs/2.4/html/hal_parallel_port.html
11:23 PM W1N9Zr0: reset is only used on step pin to double the step rate (aka doublefreq or doublestep), otherwise disable it. look at newer doc pages, 2.7 or 2.8, or whatever you're running
11:23 PM W1N9Zr0: https://linuxcnc.org/docs/2.7/html/hal/parallel-port.html