#linuxcnc Logs
Sep 26 2018
#linuxcnc Calendar
12:16 AM CaptHindsight: peep
12:20 AM skunkworks: peep
12:24 AM CaptHindsight: no logs yet for today, thought I'd make a peep
12:24 AM CaptHindsight: up late machining
12:32 AM veek: what kind of wire do they use in making nails
12:43 AM veek: nm http://www.nailmakingmachine.net/news/html/453.html
12:44 AM veek: SAE 1008/1006) have #45 ,#65 steel wire for making concrete nails and C1022 steel wire for making screw nails. 380-400 per MT.
12:44 AM veek: steel wire coil diameter is 6.5 mm or 5.5 mm ,2 mt per coil.
12:45 AM veek: hmm how do they reduce the wire down to 1/2 for a nail
12:47 AM veek: http://www.nailmakingmachine.net/news/html/428.html *purrs*
01:19 AM veek: anyone have a life size wire gauge chart
01:58 AM Deejay: moin
03:08 AM diverdude: Anyone knows how accelstepper handles the situation where i ask it to move X steps at an acceleration which is so slow that there is not enough steps to go to full speed? How much will it accel and deaccelerate in this case?
04:57 AM jthornton: morning
05:00 AM XXCoder: hey
05:19 AM Tom_L: morning
05:20 AM Tom_L: 53°F Hi 69
05:24 AM jthornton: that arrives here tomorrow we are still 70°F with a high of 71°F
07:00 AM jthornton: damn that code drove me nuts till I figured out what stupid mistake I made
07:01 AM XXCoder: what was wrong
07:01 AM jthornton: or should have been and
07:02 AM XXCoder: oh thats fun
07:02 AM jthornton: yea, it's like WTF is wrong then it hits you
07:02 AM XXCoder: my worse so far is LACK of one line
07:02 AM rmu: diverdude: it will accelerate half time and then decelerate, will not reach top speed
07:02 AM XXCoder: something->pointerthing = null;
07:03 AM jthornton: rmu: that depends on how the trajectory planner is setup
07:03 AM jthornton: http://linuxcnc.org/docs/2.7/html/user/user-concepts.html#_trajectory_control
07:03 AM rmu: ok, assumption was initial speed = final speed = 0
07:03 AM XXCoder: jthornton: that bug was hell to debug. In fact I didnt discover the bug till I tested script that repeated a small loop
07:04 AM XXCoder: second loop would run infinitely. really confused me
07:04 AM jthornton: I hate when that happens lol
07:04 AM jthornton: anyway on to make a i2c buss and get the oled working
07:04 AM XXCoder: well as it turns out without null assignment, it had dirty data and it pointed back to old data that then ran into that one again
07:04 AM jthornton: ouch that is bad
07:05 AM rmu: jthornton: don't know what you mean
07:05 AM XXCoder: yeah. it was so evil that it worked perfectly unless there is tiny chance
07:05 AM XXCoder: then it freezes
07:05 AM XXCoder: took me 8 hours to figure out
07:06 AM jthornton: rmu: G61 is different than G64 which is different from G64Pn
07:07 AM rmu: jthornton: i was referring to diverdude's question from 4 hours ago
07:07 AM rmu: should have looked at timestamp
07:07 AM jthornton: XXCoder: I thought I would go for a ride on the mountain bike till I walked outside to let the chickens out
07:08 AM jthornton: yea the answer to his question depends on how the trajectory planner is programmed
07:09 AM rmu: the trajectory planner is only involved if there is more than one move
07:09 AM XXCoder: gloomy weather?
07:10 AM jthornton: rained last night so the trails will be muddy
07:10 AM jthornton: rmu: yea that is another unknown for his question, so the answer is not enough information to answer
07:14 AM rmu: whatever the planner decides to do, for linear moves, each axis is accelerated/decelerated with const acceleration, depending on start/exit velocity and length on segment, optionally a part with constant velocity happens in between. very short moves can directly accelerate from start to exit velocity if any of those are nonzero.
07:15 AM rmu: relevant for diverdude: acceleration limits will not be exceeded, if they are, it is a bug ;-)
07:15 AM jthornton: I wonder if that is what he was asking?
07:18 AM rmu: how much it will accelerate and decelerate in case full speed can not be reached because move is too short
07:19 AM rmu: if initial speed = final speed = 0 then it will accelerate to some speed for half the move and decelerate for the other half
07:20 AM rmu: whatever. diverdude can ask again. EOT.
07:21 AM sensille: he was talking of some android library
07:23 AM jthornton: awesome this 2.1 oled works on 3.3v
07:24 AM rmu: mea culpa. my fault then, should have left it alone in linuxcnc channel
11:18 AM fragalot: Hey
11:22 AM Tom_L: hi
11:26 AM fragalot: even with direct instructions to delivery guys, they still hand off packages to my 104 year old neighbour >.>
11:28 AM Tom_L: ppl don't listen much anymore
11:28 AM Tom_L: goes in one ear and right out the other
11:28 AM fragalot: not sure if it makes it that far :P
11:29 AM Tom_L: it's hollow inside
11:31 AM pcw_home: I think Amazon delivery persons are the worst I've seen (dumping stuff in the bushes if you don't answer the door in 10 seconds)
11:40 AM Loetmichel: pcw_home: in germany "Hermes" delivery guys easily top the Amazin guys.
11:40 AM Loetmichel: -i+o
11:41 AM Loetmichel: ( if its not the same guy, Amazon uses all kinds of delivery services here in germany)
11:41 AM Loetmichel: (whatever is cheapest at the time i assume)
12:26 PM jd3 is now known as Guest83508
12:54 PM * JT-Shop gets all the way to town and the clamp that holds the seat on the road bike breaks... was a fun ride back
12:55 PM Tom_L: don't sit down
12:56 PM JT-Shop: I managed to jam the seat down on the post and gently sit on it but it was then very low so lots of extra work to pedal lol
12:57 PM JT-Shop: crap the part that broke is not a simple part...
12:57 PM Tom_L: i snapped a pedal off once going up a hill
12:57 PM Tom_L: that was fun
12:57 PM JT-Shop: ouch
12:58 PM Tom_L: you can machine a new one...
12:59 PM Tom_L: have you ever used a thread mill?
01:00 PM Tom_L: i'm wondering how to touch off
01:00 PM JT-Shop: I have one but never used it
01:01 PM * JT-Shop wanders to town to see if the local bicycle shop has one first
01:15 PM gregcnc: Tom_L I touch off the bottom, as surely you don't want to crash it in the bottom of a bore by mistake. For cutting threads I account for the distance from bottom to tip in the program
01:15 PM Tom_L: that's what i was thinking
01:16 PM Tom_L: i considered adding half a thread but depending on the pitch it could be more or less than that
01:17 PM gregcnc: check the catalog for the distance (or measure/eyeball) I also add a small amount for clearance
01:17 PM Rab: 15% off eBay right now (US/Canada): https://pages.ebay.com/promo/2018/0927/69157.html
01:17 PM Tom_L: nice thing is you can cut inside, outside, right or left hand threads with one tool
01:18 PM gregcnc: yes versatile if slow
01:20 PM gregcnc: ooh 15% off, I still have 10% ebay bucks through today
01:21 PM Tom_L: not good here
01:21 PM Tom_L: valid on Sept 27
01:22 PM gregcnc: oh start tmr
01:23 PM Rab: Egad, my mistake
01:23 PM Tom_L: better than having it for yesterday
01:23 PM * Tom_L makes a note to order something tomorrow
01:25 PM gregcnc: sometimes tooling is so tough to track down
01:26 PM beachbumpete1: Tom_L whatcha workin' on?
01:26 PM Tom_L: nothin really. i wanna try out thread milling since my spindle motor is too small for rigid tappnig
01:27 PM Tom_L: if it works good i'll have an alternative way to make threads
01:27 PM gregcnc: http://www.iscar.com/eCatalog/Family.aspx?fnum=2448&mapp=IS&app=267&GFSTYP=M
01:27 PM beachbumpete1: ah
01:27 PM gregcnc: these aren't that common but it seems every make has a slightly different ISO code
01:28 PM Tom_L: maybe that's for cam tool description codes for the industry
01:28 PM beachbumpete1: https://i.imgur.com/03jXWn7.jpg
01:29 PM beachbumpete1: https://i.imgur.com/imzgUAc.jpg
01:29 PM beachbumpete1: https://i.imgur.com/WRbF7RW.jpg
01:29 PM Tom_L: 3d scanner?
01:29 PM beachbumpete1: Some point clouds from scans we took of a house we are siding
01:30 PM beachbumpete1: yeah
01:30 PM beachbumpete1: I thought it looked pretty cool
01:30 PM Tom_L: yep
01:30 PM beachbumpete1: now we are trying to wrap head around using the data inside 3D Autocad to create planes on the faces
01:31 PM beachbumpete1: then we can measure and show the shimming and firring necessary to make our way around the house without too much drama.
01:32 PM gregcnc: what's the accuracy of a scan like that?
01:32 PM beachbumpete1: thats a damn good question
01:32 PM beachbumpete1: :)
01:34 PM gregcnc: sorry i ask the tough questions
01:34 PM beachbumpete1: the accuracy of the scanner is listed at 10mm at 10 meters but the volume of the data overlaps considerably allowing for presumably more accuracy. In our actual scans measurements compared to take measured by hand measurements it has not been off more than an 1/8" or so at most
01:34 PM gregcnc: ok so there are multiple scans to make that
01:34 PM beachbumpete1: most of the time it is fairly spot on to actual dimensions
01:35 PM beachbumpete1: oh hell yeah
01:35 PM beachbumpete1: that is actually like 25 scans from numerous locations atop scaffolding etc....
01:35 PM gregcnc: wow ok
01:35 PM beachbumpete1: the scanner was normally no more than a few feet from the walls we are trying to get data on.
01:36 PM beachbumpete1: those images are also cropped considerably it used to have all the trees and cars and whatnot around the property.
01:36 PM gregcnc: are datums required or does software find overlaps
01:37 PM beachbumpete1: This is my first big scan and we are moving slowly to learn how to use this tool and what data we can use as well as how accurate it is...
01:37 PM beachbumpete1: software finds overlaps as long as you give it half a chance. Otherwise you can manually register the scans using a series of common points.
01:37 PM gregcnc: sounds like fun
01:38 PM beachbumpete1: it is interesting indeed
01:38 PM gregcnc: new stuff always is
01:38 PM beachbumpete1: Autocad pro is impressive to say the least.
01:38 PM beachbumpete1: you have a good bit of manipulation available to you in regard to the cloud points.
01:39 PM Tom_L: nice big files too
01:40 PM gregcnc: isn't that why they make bigger and faster computer hardware?
01:40 PM Tom_L: they do that stuff at the vo-tech here. we got to go to that department once and see all their toys
01:40 PM Tom_L: it still takes time
01:41 PM gregcnc: surely much faster than the alternative
01:43 PM FinboySlick: Tom_L: I'm stuck with thread milling too. Works pretty darn good mind you.
01:43 PM FinboySlick: https://www.youtube.com/watch?v=ZX6bvPdfElE
01:44 PM Tom_L: near as i can tell i can do #10 and M5 M6 all with one cutter
01:44 PM Tom_L: you start from the bottom?
01:45 PM FinboySlick: Yes, so the chips don't all mash up at the bottom.
01:45 PM gregcnc: I always climb cut
01:45 PM FinboySlick: In this case it isn't a blind hole so it doesn't matter as much.
01:46 PM Tom_L: gregcnc i found it works quite a bit better on my mill too
01:47 PM gregcnc: in regard to thread milling climb cut internal right hand means you start at the bottom.
01:50 PM Tom_L: do you program the major diameter as given in the data sheets?
01:51 PM FinboySlick: Tom_L: I had to tweak a bit from whatever HSMWorks and Solidworks picked for me, but it's a weird thread size.
01:51 PM gregcnc: usually I have a mating part to match which has been made/ measured
01:52 PM gregcnc: if it was important I'd use a gage
01:52 PM FinboySlick: Was too tight 'as modeled' but I don't know where the error lies.
01:53 PM gregcnc: and use cutter comp to adjust final pass
01:53 PM Tom_L: it's a necessity but i'm not fond of using cutter comp
01:54 PM gregcnc: cutter comp is wonderful and long as you don't screw it up
01:54 PM Tom_L: yeah
01:55 PM gregcnc: figure out what works and put it into a spreadsheet of macro
01:56 PM gregcnc: or macro
01:56 PM Tom_L: yeah, once i do some i'll have the cad data to get them right
01:58 PM Tom_L: i'm not gonna go out and buy thread gages though :)
02:43 PM diverdude is now known as Guest37106
02:44 PM diverdude: Hi, so I have motors setup like this: https://imgur.com/a/6OwONIu Is it possible to buy external encoder to put on or do motors come with encoders internally incorporated, so I have to buy new motors with encoders builtin?
02:45 PM fragalot: encoders can be added externally if you need them
02:45 PM diverdude: fragalot: oh cool - can you maybe point me to some really good ones :)
02:45 PM fragalot: phoenix, sick, ...
02:45 PM fragalot: what resolution do you need?
02:46 PM diverdude: fragalot: is it possible to get submillimeter?
02:46 PM gregcnc: you can get 25bits per rev...
02:46 PM fragalot: diverdude: sure. either you can get a linear encoder and mount it elsewhere, or you can get a rotary encoder
02:46 PM diverdude: oh nice
02:46 PM fragalot: depends on what your need is
02:46 PM fragalot: or what you want to accomplish
02:47 PM gloops: put an encoder on the shaft
02:47 PM diverdude: fragalot: hmm ok...i just want to know where im at even if i am slipping :D
02:47 PM gloops: (screw shaft)
02:47 PM fragalot: diverdude: how is the shaft that you turn moving the axis?
02:47 PM diverdude: gloops: sorry...im not native english...which one is the screw shaft again?
02:48 PM fragalot: diverdude: if you have a shaft that is positively coupled, you can use an incremental rotary encoder on there
02:49 PM Wolf__: why not ask what the machine is first ;P
02:49 PM fragalot: if you do not, or don't trust it, you can put an optical or magnetic scale completely separately
02:49 PM fragalot: Wolf__: NEVER! :D
02:49 PM fragalot: always assume
02:49 PM Wolf__: lol
02:50 PM diverdude: fragalot: hmm im not really sure if its positively coupled or not
02:50 PM fragalot: diverdude: do you have an overview picture
02:50 PM diverdude: yeah, hang on
02:50 PM gregcnc: you could squeeze a though hole encoder right next to the bearing spports on the shaft
02:54 PM diverdude: fragalot: mmm i dont know if this gives you more insight: https://imgur.com/a/yWjszB4
02:54 PM diverdude: fragalot: there are 2 motors....its the same motor in both pictures
02:54 PM diverdude: fragalot: but the other one works in same way
02:54 PM gloops: sorry ive got sausages in the pan
02:55 PM fragalot: if you trust the belts to not slip, you can put a rotary encoder on the shaft
02:55 PM Wolf__: looks like co2 laser setup
02:55 PM fragalot: if you don't, you can still install linear encoders fully separate from the belt drive system
02:55 PM diverdude: fragalot: hmm im not sure if the belts will not slip...they may actually
02:55 PM fragalot: then install either an optical or magnetic linear encoder scale :)
02:55 PM gregcnc: why should they slip?
02:56 PM diverdude: gregcnc: well...im just a skeptic kind of guy...i dont trust those belts :D
02:56 PM fragalot: https://www.machine-dro.co.uk/digital-readouts-display-and-encoders/linear-encoders/easson-linear-encoder.html <== cheapie but alrighty
02:56 PM diverdude: gregcnc: better play it safe hehe
02:56 PM Wolf__: whats the cutting tool load?
02:57 PM diverdude: fragalot: optical or magnetic encoder sounds cool
02:57 PM diverdude: Wolf__: me?
02:57 PM gregcnc: I forget, do we have a example of steppers with encoders working
02:57 PM diverdude: Wolf__: its not a cutting tool...its carrying an aluminium frame which is a couple of kilos
02:58 PM Wolf__: yeah, application would determine if I would worry about slipping or not
02:58 PM diverdude: Wolf__: its quite light load
02:58 PM fragalot: gregcnc: my build is a hybrid setup like that
02:58 PM fragalot: except the drive takes care of it :P
02:59 PM gregcnc: tha'ts different
02:59 PM Wolf__: I’m thinking about using clear path servos on my plasma cnc for that reason
02:59 PM diverdude: fragalot: wow they seem quite big
03:00 PM fragalot: diverdude: magnetic ones are tiny
03:00 PM fragalot: it's like a 1x10mm strip of magnetic material, and a tiny read head
03:00 PM diverdude: fragalot: nice - do you have a link to some good magnetic encoder?
03:01 PM XXCoder: behold: the sphere. https://www.amazon.com/dp/B00YV3B2SG/
03:01 PM fragalot: I used the oens from M-DRO on the website I linked
03:01 PM diverdude: XXCoder: wooow....that looks dope :D
03:01 PM fragalot: https://photos.app.goo.gl/bmMh3jHoxC9KCHHRA
03:02 PM fragalot: https://photos.app.goo.gl/5Ns33ZvVdsbKJJ3K6
03:02 PM diverdude: fragalot: yeah they look cool
03:02 PM fragalot: etc
03:03 PM gloops: i forget the really cheap hack on ebay for a reasonable encoder
03:03 PM gloops: bicycle milometer or something
03:03 PM fragalot: low resolution, but sure
03:04 PM diverdude: fragalot: ahh the magnetic ones how low resolution?
03:04 PM fragalot: diverdude: 0.005mm for the standard, or 0.001mm if you pay more and hate yourself
03:05 PM diverdude: fragalot: i hate myself but i like 0.001mm :D
03:05 PM fragalot: you really won't :D
03:05 PM fragalot: sneeze nearby and the number changes
03:05 PM diverdude: fragalot: oh...show me da stuff
03:05 PM diverdude: ohhh
03:05 PM diverdude: thats bad
03:06 PM diverdude: i guess the rattling from the moving machine will induce erros in those encoders then if they are so sensitive
03:06 PM fragalot: yup
03:06 PM fragalot: https://www.machine-dro.co.uk/digital-readouts-display-and-encoders/linear-encoders/m-dro-magnetic.html <==
03:06 PM fragalot: they're not really errors
03:06 PM fragalot: your machine really IS moving that much
03:07 PM diverdude: heh yeah ok
03:07 PM diverdude: oh, so many...and they got a nifty little price tag on em too :D
03:07 PM fragalot: magnetics cost money :)
03:08 PM gloops: https://www.ebay.co.uk/itm/4-Digital-LED-Tachometer-RPM-Speed-Meter-Hall-Proximity-Switch-Sensor-Bracket/163108794005?
03:08 PM gloops: these dont
03:08 PM fragalot: gloops: that's because that is not an encoder
03:08 PM fragalot: it's just a switch
03:08 PM gloops: yeah thats not what i was looking for
03:09 PM diverdude: fragalot: which one has less error? optical or magnetic?
03:09 PM fragalot: magnetic puts less strain on your machine
03:10 PM fragalot: optical drags because of the seal
03:10 PM fragalot: but optical is _cheap_ :)
03:10 PM diverdude: sounds like magnetic is what i want...but they all seem to have a serial port connection
03:10 PM diverdude: :/
03:11 PM fragalot: that's just a connector, it's a standard 5V encoder signal
03:11 PM fragalot: I should have the manual for them lying around in my shed
03:11 PM fragalot: I can send you a copy tomorrow if you want
03:11 PM diverdude: fragalot: oh yeah that'd be awesome.
03:11 PM fragalot: alright - remind me tomorrow :)
03:11 PM diverdude: fragalot: would be cool to have usb interface instead
03:12 PM fragalot: usb is far too slow for encoders
03:12 PM diverdude: oh really
03:12 PM fragalot: if you want them to be useful, yeah
03:12 PM diverdude: i thought usb was way faster than serial
03:12 PM fragalot: it is faster at continuous throughput
03:12 PM fragalot: but it's not realtime
03:12 PM diverdude: aha
03:12 PM Rab: Faster, maybe less deterministic?
03:59 PM Deejay: gn8