#linuxcnc Logs

Jan 18 2021

#linuxcnc Calendar

12:43 AM mrec: I think I'll do it another way ... store all the coordinates in the machine coordinates and add or decrease the detected tool offset
12:49 AM mrec: that should work best, I just want to have some setup for testing the accuracy that's not for cutting.
01:21 AM perry_j1987: o/
02:13 AM Deejay: moin
05:11 AM JT-Cave: morning
05:11 AM XXCoder: good morning
05:34 AM Tom_L: morning
05:36 AM XXCoder: good morning
06:46 AM Loetmichel: hehe, $me just threw the Samusng s7 of a coworker on the 3d-printer bed. 70°C for half an hour should weaken the glue enough to get the back off... (battery dead)... sometimes such a printer is useful for things you wouldn't think of :)
09:15 AM perry_j1987: morning
09:19 AM perry_j1987: that first sip of coffee where you chicken out cause its still too hot lol
10:22 AM unterhaus: I would hate to see Pi4 dropped. Got to be a solution.
10:52 AM CaptHindsight: not referring to Andy but the Rpi kernel devs are near worthless
10:54 AM CaptHindsight: unfortunately the dev here had to bow out but this was the way to do it for the Rpi4 https://github.com/sakaki-/gentoo-on-rpi-64bit
11:11 AM JT-Cave: only a million wires to go...
11:23 AM JT-Cave: lunch minute is over back to work...
11:36 AM unterhaus_: the rpi4 model b still works, right?
11:36 AM unterhaus_: or did they inline the usb hardware changes?
11:37 AM unterhaus_: I wonder if there is an exhale filter for my 3m mask
11:38 AM unterhaus_: I guess you just tape the exhale valve shut
11:51 AM JT-Shop: only 140 wires left to connect the buttons up
11:59 AM * roycroft has a glue-up to do, then shall begin dismantling the bandsaw
12:10 PM unterhaus_: thingiverse wins the price for light text on white background. I thought I did something wrong it was so light
12:10 PM unterhaus_: ^prize
12:11 PM unterhaus_: only on one of their account signup pages though
12:17 PM unterhaus_: okay, no point to having an account, it only saves one thing to collections
12:19 PM roycroft: i don't have light on light issues very much, but a huge number of websited don't do dark mode properly
12:19 PM roycroft: dark blue or dark grey text on a black background
12:20 PM roycroft: better to not support dark mode at all than to just change the background while leaving the text the same
12:34 PM unterhaus_: I just saw an article about how wfh is ruining everyone's eyesight and the trend is to make websites harder to read
12:34 PM unterhaus_: I don't know how you gray out some options when text is so close to white
12:36 PM unterhaus_: on that thingiverse page, even the button was close to white
12:55 PM norias: i did my website with black on light grey
12:55 PM norias: because white backgrounds are just so bright, it's annoying
01:41 PM unterhaus_: as long as there is decent contrast. Not sure when this craziness will end, but it's annoying for now
01:46 PM roycroft: dark mode is a good thing
01:47 PM roycroft: i've been running my xterms with a black background and various colors for text for decades
02:24 PM CaptHindsight: unterhaus_: sounds like different is better. The bad design lessons learned from long ago are the new new again.
02:35 PM unterhaus_: CaptHindsight what do you know about the mistakes of the past?
02:36 PM unterhaus_: not sure that's the right joke based on your nick, but it's close
02:37 PM CaptHindsight: I'll allow it
02:37 PM roycroft: iirc, capthindsight is an american, therefore he cannotknow of the mistakes of the past, as one of the fundamental rights of being an american is the right to be ignorant
02:38 PM roycroft: and all that we live for is to exercise our fundamental rights to their extremes
02:39 PM CaptHindsight: I have just as much to my ignorance and you are to your unignorance
02:39 PM roycroft: learning from the past sounds kind of socialist, doesn't it?
02:39 PM CaptHindsight: much right
02:40 PM rs[m]: soundtrack https://www.youtube.com/watch?v=cpbtU_vOUkY
02:40 PM roycroft: we learned from the brits - tally ho and forge ahead!
02:41 PM * gloops tips hat
02:41 PM CaptHindsight: I'm morning the loss of my yeast. Passed away last night.
02:42 PM rs[m]: that is a very strange pun
02:43 PM gloops: probably more than one yeast
02:43 PM gloops: trillions of yeastlets died
02:44 PM CaptHindsight: morning/mourning
02:44 PM rs[m]: morning at night
02:45 PM CaptHindsight: looks like a day I should stay away from anything expensive, sharp, words with letters etc
03:04 PM Juzzy: So I finally finished my cnc, I'm trying to move my Y forward, since I have 2 Ys syncronized, It doesnt let my Jog it until I home them. There's no way to Home the 2nd axis. (trying to home the normal Y does nothing?) However if I play the default "Linuxcnc" gcode, it errors out but then I can Jog around. hoever I cannot move it lower than when I started the gcode, it sets it as home, I need
03:04 PM Juzzy: to move it to like -800mm from where it's at.
03:04 PM Juzzy: is there a trick? or am I doing something vastly wrong?
03:04 PM gloops: hmm
03:05 PM gloops: are you using 2.8?
03:05 PM Juzzy: ya, latest
03:05 PM Juzzy: the mirror'd Ys are working, I just cant move them closer to the bottom
03:06 PM Juzzy: I'm trying to install Y switches but I cant get the axis down to test. Not even sure how to "find home" yet in lcnc
03:06 PM gloops: 2 Ys synchronised, you mean 2 motors on the Y axis?
03:06 PM gloops: or 2 Y axis
03:06 PM Juzzy: ya, 1 on left, 1 inverse on the right
03:07 PM gloops: ok you wont home without switches
03:08 PM gloops: can you see 'Home all'
03:08 PM Juzzy: ive seen it before, not sure its an option now, will have to go back to my shop and see
03:09 PM Juzzy: i dont have my Z home or Y installed yet
03:09 PM gloops: i did have a vid somewhere
03:09 PM Juzzy: b/c i cant move my
03:09 PM Juzzy: Y by hand without some help
03:10 PM gloops: there is a homing sequence, only one axis homes at a time
03:10 PM Juzzy: maybe i can tie a rope around each one and pull them together or something. seems like there should be a way to make them work tho
03:12 PM CaptHindsight: manual homing vs homing sequence
03:12 PM gloops: https://www.youtube.com/watch?v=iDTpnf7DIHM&feature=youtu.be
03:12 PM Juzzy: how do you manually home it?
03:12 PM gloops: apologies for the shocking video quality
03:13 PM Juzzy: when I try the manual commands it says it cannot jog until it's homed. X and Z work fine tho
03:13 PM gloops: you should be seeing something like that
03:13 PM CaptHindsight: http://linuxcnc.org/docs/2.8/html/config/ini-homing.html
03:13 PM Juzzy: ya mine is differenty
03:13 PM Juzzy: ya mine is different
03:13 PM gloops: different?
03:14 PM Juzzy: sec.
03:14 PM CaptHindsight: Juzzy: post a link to a paste of your config file
03:14 PM gloops: you should have joint mode and global mode if you used 2.8 and synched 2 joints
03:14 PM unterhaus_: ot: anyone know what elongation 2" means as far as a mechanical property?
03:15 PM gloops: so you should see joint 0 joint 1 joint 2 etc, you wont get to the global/jogging screen until all the joints or homes
03:15 PM Roguish: test test.
03:16 PM Roguish: wow. i can post....
03:16 PM gloops: or you provide a signal to indicate they are homed, by pressing the switches
03:16 PM Juzzy: dang i need to network that machine, lol
03:16 PM CaptHindsight: unterhaus_: possibly elongation in % in 2"
03:16 PM Juzzy: k 1 sec on a couple pics...
03:16 PM gloops: Roguish the rest of linuxcnc is easy once you learned the posting bit
03:17 PM unterhaus_: Capt: yeah, it means something like that. No idea why it's a percentage
03:17 PM CaptHindsight: unterhaus_: https://pmpaspeakingofprecision.com/tag/elongation-in-2/
03:17 PM unterhaus_: so far all the answers have been garbled
03:17 PM Juzzy: does this work? https://cdn.discordapp.com/attachments/615653027952459865/800833309101260841/image1.jpg
03:18 PM Juzzy: errors: https://cdn.discordapp.com/attachments/615653027952459865/800833308316532736/image0.jpg
03:18 PM gloops: no
03:18 PM Juzzy: bleh
03:18 PM unterhaus_: Capt, that seems garbled.
03:18 PM Roguish: gloops, been here for years. just been having issues with freenode.....
03:18 PM unterhaus_: but I'm not getting anywhere near yield, so I guess I don't care much
03:19 PM gloops: Juzzy have you edited the INI and HAL files?
03:19 PM Juzzy: ya gloops.
03:19 PM gloops: any chance of a quick look at the ini
03:20 PM Juzzy: CaptHindsight asked too, need to get some networking on that box. or i can sneakernet it
03:20 PM gloops: https://forum.linuxcnc.org/49-basic-configuration/33079-how-to-2-or-more-motors-on-one-axis-gantry-linuxcnc-2-8-master
03:20 PM unterhaus_: I guess that's why I nearly flunked strength of materials
03:20 PM unterhaus_: best professor ever for that though
03:20 PM gloops: thats the basic procedure - written by a layman
03:21 PM gloops: well, written by a non layman, in laymans terms
03:23 PM CaptHindsight: unterhaus_: will be happy if I get my shoes properly tied today
03:24 PM unterhaus_: Capt: I am wearing crocs all day unless I change into some shoes I haven't tied for 20 years
03:25 PM unterhaus_: I might ride the trainer and those shoes have boa, so no tying them either
03:26 PM unterhaus_: can't remember the name of the laser cutting place someone recommended
03:26 PM unterhaus_: sendcutsend
03:26 PM Juzzy: homing: https://imgur.com/a/HBdjTyG
03:27 PM unterhaus_: you can't turn off the requirement to home?
03:27 PM Juzzy: error: https://imgur.com/a/1R4UHD1
03:27 PM Tom_L: Roguish, get your card all wired up?
03:28 PM gloops: not seeing anything Juzzy
03:28 PM Juzzy: ?
03:28 PM gloops: not loading
03:29 PM Juzzy: hrm
03:29 PM gloops: i never worked a way out to bypass the homing stage
03:29 PM Juzzy: reload: https://imgur.com/a/1R4UHD1
03:30 PM Juzzy: actually this one has both pics: https://imgur.com/a/HBdjTyG
03:31 PM unterhaus_: mesa cart is broken
03:31 PM gloops: might be my browser im not seeing anything
03:31 PM Roguish: Tom_L, yeah. pretty much. had to explicitly set up the IP addresses in the system network file.
03:32 PM Tom_L: sometimes you gotta do that but not always
03:32 PM Tom_L: depends on the linux flavor and moon phase
03:33 PM Roguish: the system kept getting confused with the wifi. hardcoded is better anyway.
03:33 PM Tom_L: yup
03:33 PM Juzzy: HAL: https://pastebin.com/ZGk3vfsr | INI: https://pastebin.com/UskHAsQv
03:34 PM Juzzy: Joint 1 and 3 are the Y's
03:35 PM Roguish: Tom_L, is your spindle set up with it's own encoder and PID in hal.
03:36 PM Juzzy: FWIW: net all-limit <= parport.0.pin-11-in = currently open bc I have not installed my limits yet
03:37 PM unterhaus_: yay, 7i48 in stock, ordered
03:37 PM Tom_L: yes but the vfd uses the hall from the motor
03:38 PM unterhaus_: gets rid of 4 relays in my system
03:38 PM gloops: 2 many axis, you havent got 2 Y axis
03:38 PM perry_j1987: so spindle speed is calculated by encoder or by the z index pulse
03:39 PM unterhaus_: CaptHindsight now that I have a 3d printer, I can finish my delta sigma a lot easier
03:42 PM gloops: thats a working INI for 2 motors on 1 axis https://pastebin.com/WZU4dJfG
03:43 PM W1N9Zr0: Juzzy, i don't think HOME_SEQUENCE can be -2, it's -1, 0, or positive numbers
03:43 PM Juzzy: ah, i got -2 off the forms as a way to do mirroring
03:44 PM gloops: and the hal https://pastebin.com/TEuXjGUJ
03:44 PM gloops: these pages will self destruct in 1 hour
03:44 PM Juzzy: so mine is COORDINATES = X Y Z Y and yours is just X Y Z
03:45 PM gloops: yes you only have 3 axis
03:45 PM Juzzy: also coppied from forums, i dont think that does much tho
03:45 PM gloops: you might be mixing pre 2.8 tutorials
03:45 PM Juzzy: very possible
03:47 PM Juzzy: I'll clean mine up some, and set it to 1 and 2 for X, remove Z like you have and see what happens
03:47 PM gloops: you do have a Z axis? youll need to keep that, remove one of the Ys
03:48 PM gloops: if i understand, you have an XYZ machine, with 2 joints on the Y
03:51 PM CaptHindsight: https://forum.linuxcnc.org/ are these just running slow for me?
03:53 PM Tom_L: loaded fine here
03:54 PM CaptHindsight: ok back up to speed here now as well
03:54 PM CaptHindsight: was looking at the Rpi4 issues
03:54 PM Tom_L: the new chip issue?
03:54 PM Tom_L: or an old one
03:55 PM Juzzy: gloops, that didnt work for me, -just- changing my HOME SEQ from -2 to 1, disconnected them and now jogging works but on 1 joint, not both,
03:55 PM W1N9Zr0: Juzzy i take it back, you can have -2 as the HOME_SEQUENCE, but you must number all your joints for home-all to work. the sequence must start at 0 or 1 (or -1 for synched pair of joints) and increment without skipping numbers. ie your sequence should be: X(0):0, Y(1):-1 Y(3):-1, Z(2):2 or similar
03:55 PM CaptHindsight: Tom_L: the issues that Andy posted just recently
03:55 PM Juzzy: W1N9Zr0: ah that makes sense.
03:56 PM unterhaus_: Cap'n, any hope?
03:56 PM CaptHindsight: Tom_L: do you have a synopses of the issues?
03:56 PM Tom_L: i don't read the forums but i assume it's the new chip and loading from the image
03:56 PM Juzzy: I dont want to home Z automatically sint i dont have a touch plate
03:56 PM Juzzy: since*
03:57 PM Tom_L: iirc it was the usb chip?
03:57 PM CaptHindsight: I wasn't aware of any Rpi4 issues with Linuxcnc
03:57 PM unterhaus_: I thought you said you got up to speed
03:58 PM Tom_L: the latest boards have a different chip that requires loading from the rpi image file i believe
03:58 PM Tom_L: which chip exactly i'm not sure but i think it's the USB one
03:58 PM unterhaus_: like a binary blob?
03:58 PM unterhaus_: they disconnected the ethernet from the usb, as I understand it
03:58 PM Tom_L: some such thing
03:59 PM unterhaus_: isn't there a blob for the graphics already?
03:59 PM Tom_L: andy was trying to get a new board
04:00 PM CaptHindsight: unterhaus_: up to literal website access speed
04:00 PM CaptHindsight: pages were taking a minute to load
04:00 PM unterhaus_: okey
04:05 PM CaptHindsight: Ethernet left USB on the jump from Rpi3 to Rpi4, why Ethernet on the Rpi4 was usable for real time
04:06 PM unterhaus_: okay, can't keep track
04:07 PM unterhaus_: "You might recall that strongly I recommended extrusion for this application. There are a lot of reasons why, cost and effort being the big ones."
04:07 PM unterhaus_: welding aluminum angle iron instead tho
04:08 PM unterhaus_: I'm sure he has his reasons
04:22 PM CaptHindsight: Raspberry Pi 4 Model B vs Raspberry Pi 400
04:24 PM CaptHindsight: Rpi 400: the CPU is a newer revision of the same model and has a slight bump in clock speed
04:29 PM perry_j1987: hows it going guys
04:31 PM Tom_L: 05:10 PM andypugh: They dropped the eprom that supplied the USB chip firmware.
04:31 PM Tom_L: 05:10 PM andypugh: The USB chip is now programmed with a file from the SD card.
04:33 PM unterhaus_: well, the only reason the pi is so popular is that it's cheap
04:33 PM unterhaus_: someone was recommending an odrioid, which costs $200
04:34 PM CaptHindsight: andypugh: https://sourceforge.net/p/emc/mailman/message/37199548/ is there a list of changes to USB somewhere, I haven't found one
04:36 PM CaptHindsight: Tom_L: I can't find any details on this programming change "The USB chip is now programmed with a file from the SD card."
04:37 PM CaptHindsight: how does one know which model number (Rpi4 variant) uses this new programming method?
04:38 PM CaptHindsight: Rpi will never have u-boot since that would give users too much control over their hardware
04:39 PM Tom_L: no model number change afik
04:39 PM Tom_L: that's the kicker
04:39 PM CaptHindsight: wonderful
04:39 PM CaptHindsight: fix the Orange pi :)
04:39 PM Tom_L: 05:49 PM andypugh: I just ordered a new Pi4. I will be very annoyed if it’s one of the ones that works.
04:39 PM Tom_L: no way to tell
04:40 PM Tom_L: less a missing eeprom
04:40 PM CaptHindsight: fix the Orange Pi (Allwinner) ARISC driver
04:40 PM CaptHindsight: hell use the Orange Pi with mesa
04:41 PM unterhaus_: pis are subsidized?
04:42 PM CaptHindsight: China (Allwinner) vs Broadcom (Satan)
04:42 PM unterhaus_: there are benefits to running on the pi that can't really be equaled with other boards
04:42 PM andypugh: Tom_L: I think I am making progress.
04:43 PM andypugh: The pi doesn’t have much to recommend it as a controller compared to very many other things. But lots of people have them and are trying to find a use for them.
04:44 PM andypugh: Tom_L: Cross-compiled kernel + manual install direct to SD works fine on all three of my Pi’s.
04:44 PM andypugh: Now trying deb install onto a fresh stock SD to see what happens.
04:56 PM perry_j1987: im trying to figure out how to set scale on this spindle encoder
04:56 PM perry_j1987: motor is rated max 2000 acording to the site
04:56 PM perry_j1987: they said should have quadrature encoder 4000 lines or something
04:58 PM perry_j1987: i got A,B,Z index wired in
05:00 PM Juzzy: Are home switches NO or NC?
05:00 PM CaptHindsight: 2000 x 4000 lines is 8MHz pulses and 32MHz counts (quadrature)
05:01 PM CaptHindsight: Juzzy: up to you, or are you asking which works best?
05:01 PM Juzzy: well I am asking for the default in lcnc, i didnt define them as inverse
05:02 PM Juzzy: I'm wondering if maybe i have it backwards and its causing issues
05:02 PM Juzzy: still fighting with figuring this out
05:02 PM andypugh: Juzzy: HAL inputs to motion for liimits / home expect to go high when tripped.
05:03 PM Juzzy: ok so NO
05:03 PM andypugh: You can achieve what I described with NO or NC.
05:04 PM andypugh: Depends on external wiring, pull-uo / pull-down and what HAL pins you use in the HAL layer.
05:04 PM Juzzy: well if my default is +5v, and home = high, then that's NO
05:04 PM Juzzy: or am i missing something?
05:05 PM Juzzy: net home-x <= parport.0.pin-12-in
05:05 PM Juzzy: no invert
05:05 PM andypugh: I think that NC is probably safer. Then a broken wire stops the macine.
05:06 PM Juzzy: I would agree with that. so i would invert in pins and NC?
05:06 PM Juzzy: I would agree with that. so i would invert those input pins and NC?
05:06 PM andypugh: If you connect an NO switch between a parallel port pin and p-port gnd then it will go from high to low when the switch is closed.
05:07 PM Juzzy: I have one on pin 12, and one beside pin 12 +5.
05:08 PM andypugh: So I would use maybe use an NC switch, between P-port input and P-port Gnd (internal pull-ups) and use the in-not pin in HAL.
05:08 PM Juzzy: Are you saying 12 and gnd + NC ?
05:08 PM andypugh: But, then, are you actually wiring direct to parport, or via a BoB?
05:09 PM Juzzy: bob - C10
05:09 PM andypugh: OK, so just wire a switch across the terminals and see what it does, then adapt in HAL.
05:09 PM Juzzy: https://www.cnc4pc.com/pub/media/productattachments/files/ALLDriver.pdf
05:09 PM Juzzy: pretty close to that diagram
05:11 PM Juzzy: i'm using NC/NO switches, so I can wire them either way. I would rather NC
05:11 PM andypugh: Yes, so in that case wire an NC switch between 5V and 12. Then watch the HAL pins with halmeter or halshow.
05:11 PM andypugh: Choose 12-in or 12-in-not accordingly.
05:11 PM Juzzy: if I NC that means pin 12 will be HIGH all the time, so do I INVERT in hal or leave it default? :)
05:12 PM andypugh: The board has a jumper for pull-up or pull-down. You can invert the logic in at least three places.
05:12 PM Juzzy: i guess that's the piece im missing. where I actually tell LCNC home = high/low
05:12 PM Juzzy: I guess I can configure it to do whatever.
05:12 PM andypugh: Wire NC, because that is better, and then look in HAL to see what happens.
05:13 PM Juzzy: k
05:13 PM Juzzy: I totally believe i can make the lights change in hal
05:13 PM andypugh: Choose a pin that goes high when you press the switch, and wire that to the motion input pins.
05:13 PM Juzzy: I just dont know if -shit- is causing my jog/home issue or if I'm making it worse
05:13 PM _unreal_: burrrrrrrrrrrrrrrrp
05:13 PM _unreal_: beer good
05:14 PM Juzzy: erm, -this-
05:14 PM _unreal_: how many beers should I consume before I start soldering the wires together on my laser etcher I'm bilding :)
05:14 PM _unreal_: building
05:14 PM Juzzy: depends on alc %
05:15 PM Juzzy: :)
05:15 PM CaptHindsight: perry_j1987: look at your pulse rates with high count encoders on a high speed spindle
05:16 PM _unreal_: pbr
05:17 PM _unreal_: also does anyone know the name of this connector? I'm trying to find it in KICAD
05:17 PM _unreal_: https://drive.google.com/file/d/1b2TqA5D0FLLsx77yy1YbBBcAqE9bwdmu/view?usp=sharing
05:23 PM JT-Shop: Tom_L, a few new photos of todays progress... https://gnipsel.com/images/customers/BobG/
05:24 PM JT-Shop: 5, 7, 10, 15, 20, 25, +1
05:25 PM JT-Shop: so you can put the target at any yard up to 25
05:25 PM norias: good enough for indoor
05:25 PM norias: does anyone actually shoot at 5 yards, i wonder?
05:26 PM CaptHindsight: grenades
05:26 PM JT-Shop: 7 yards is the qualifier for concealed carry
05:26 PM JT-Shop: 5 yards if you don't have a clue...
05:26 PM CaptHindsight: the target is the side of a barn
05:26 PM norias: what state?
05:26 PM JT-Shop: MO
05:26 PM norias: lol, grenades
05:27 PM norias: toss them fuckers as far as you can
05:27 PM JT-Shop: tannerite
05:27 PM norias: i'm in PA
05:27 PM JT-Shop: more fun
05:27 PM norias: the only qualification is be a resident and pass a background check
05:27 PM norias: which is the instant check system
05:27 PM norias: so, it takes like 15 minutes to get a license to carry
05:28 PM norias: mostly getting photographed and filing out the form
05:28 PM JT-Shop: open carry or concealed carry?
05:28 PM JT-Shop: anyone can open carry here
05:28 PM norias: no license for open
05:28 PM norias: but being in a car is considered defacto concealed
05:28 PM JT-Shop: here they want to know you can kill the invader at 7 yards
05:29 PM JT-Shop: even on the dash?
05:29 PM norias: yeah
05:29 PM perry_j1987: how about machine gun on a roof rack
05:29 PM norias: yeah, good question
05:29 PM perry_j1987: lol
05:29 PM norias: that and motorcycle has been debated
05:29 PM norias: but i don't think there's any caselaw on it
05:30 PM JT-Shop: we have a lot of those but it's hard to get the machine gun permit
05:30 PM CaptHindsight: norias: so it has to be locked in a box in the trunk or it's considered concealed?
05:30 PM JT-Shop: a local biker had a Jesse James themed bike once cops tried to take the guns off of it lol
05:31 PM JT-Shop: had/has
05:31 PM JT-Shop: chickens say it's time to retire inside...
05:31 PM XXCoder: lol later
05:31 PM unterhaus_: This discussion reminds me of the top guy guys trying to shoot
05:32 PM CaptHindsight: perry_j1987: that might be allowed if it's being held by a deer also strapped to the roof
05:32 PM perry_j1987: lol
05:32 PM perry_j1987: what about jackalopes
05:33 PM perry_j1987: ok thats odd
05:33 PM perry_j1987: im running some gcode from my other lathe on the new one
05:33 PM perry_j1987: same computer from the old lathe just new config
05:34 PM perry_j1987: its just hanging on an x move
05:34 PM perry_j1987: doublechecked the program didnt pause itself or something lol
05:34 PM perry_j1987: servo drives not faulted
05:34 PM perry_j1987: dro shows no movement just hanging on that line
05:34 PM perry_j1987: tried couple times same deal
05:39 PM perry_j1987: this computers on 2.7.15 or something like that
05:40 PM JT-Cave: is it a move that requires spindle sync?
05:40 PM perry_j1987: na just retractiong X
05:41 PM perry_j1987: unless there's something in there about constant speed in the gcode
05:41 PM perry_j1987: and since i dont have this spindle encoder setup properly yet?
05:43 PM JT-Cave: G0 or G1 move?
05:43 PM perry_j1987: thats odd there was no G0 on that line
05:43 PM perry_j1987: just X22
05:43 PM perry_j1987: put a G0 like the other lines in there and rerunning it
05:45 PM perry_j1987: there were a bunch of lines with no G0 and most had G0 for the x22 clearance moves odd
05:46 PM perry_j1987: nope that wasnt it
05:46 PM perry_j1987: still got held up on that line
05:47 PM JT-Cave: no errors?
05:47 PM perry_j1987: no popups
05:47 PM JT-Cave: what is the next line
05:48 PM perry_j1987: G1 X6. F25.
05:48 PM * JT-Cave has cracked the QProcess pipe code
05:48 PM JT-Cave: that is pretty odd then
05:48 PM perry_j1987: going to steal my parts now! :P
05:49 PM perry_j1987: i've given away all my gcodes lol
05:49 PM perry_j1987: there is not much different on this lathe than the last lathe, both had encoder though hmm
05:49 PM Juzzy: On Z axis homing, is that just taking a small thin piece of conductive metal + a clip on your bit and letting the 2 short each other out?
05:50 PM Juzzy: and setting an offset -<thicnkness> ?
05:50 PM perry_j1987: forming a switch basically
05:50 PM JT-Cave: usually it's a micro switch or similar on the axis... you seem to be talking about tool touch off not homing
05:50 PM Juzzy: oh. yea tool touchoff then
05:51 PM JT-Cave: I use a dowel
05:51 PM JT-Cave: on the material Z0
05:51 PM Juzzy: do you "home" to the part or to the base surface board?
05:52 PM JT-Cave: you home the machine then you set the material tool offset
05:52 PM JT-Cave: homing sets the machine limits...
05:52 PM Juzzy: ah ok got it. so I can use a switch on Z
05:52 PM JT-Cave: https://www.youtube.com/watch?v=VmRUJk0ptN4
05:53 PM Juzzy: got it. thanks
05:54 PM JT-Cave: in that video I'm setting the tool Z to a fixture ie the spindle face then when I put a piece of material in the chuck I set ONE tool to the material Z0
05:54 PM CaptHindsight: https://imgur.com/gallery/aSY6Yqm
05:56 PM Juzzy: dang i just knew if was gunna fall back in that hole and drench him
05:56 PM JT-Cave: lol
05:56 PM JT-Cave: time to switch on the magnetassswitch on the couch
05:56 PM perry_j1987: gha this is so weird
05:57 PM perry_j1987: i've never experienced linuxcnc just "stopping" like this
05:58 PM Tom_L: JT-Cave, good progress
06:01 PM perry_j1987: took out all the S M3 commands and lets see if it completes without spindle/vfd running
06:01 PM Tom_L: andypugh, was the rpi you got the one you'd hoped to get?
06:02 PM Tom_L: perry_j1987, S won't do anything without a directional command
06:02 PM perry_j1987: yep its passed that line
06:02 PM perry_j1987: with the vfd / spindle not running
06:02 PM Tom_L: of course you set S to S 15000000 for the first test right?
06:05 PM perry_j1987: lol
06:05 PM perry_j1987: was suppose to be 500rpm
06:07 PM roycroft: my mbp finally shipped today!
06:07 PM * roycroft should have it on wednesday
06:09 PM andypugh: Tom_L: Well, it has the missing USB EEPROM, so yes.
06:12 PM CaptHindsight: andypugh: is there no way to tell which version has the EEPROM or not when you place the order?
06:13 PM unterhaus: I guess a bare rp4 can't be bought from amazon
06:13 PM andypugh: Not that I know of.
06:14 PM andypugh: unterhaus: https://www.amazon.co.uk/Raspberry-Pi-ARM-Cortex-A72-Bluetooth-Micro-HDMI/dp/B07TC2BK1X
06:14 PM andypugh: CaptHindsight: You can tell once it arrives…
06:14 PM CaptHindsight: handy
06:15 PM andypugh: The one in the photo has the eeprom.
06:15 PM andypugh: Whether the one that arrives has…
06:15 PM unterhaus: andy, is that what you bought?
06:15 PM andypugh: No, I bought mine from the Pi Hut
06:16 PM Tom_L: closest to the source would likely have the latest revs
06:17 PM andypugh: What I did get from Amazon was one of these cases. In purple: https://www.amazon.co.uk/Aluminium-Heatsink-Case-Raspberry-Rhodolite/dp/B082Y2LVGB
06:17 PM andypugh: I like that case, it makes the Pi a lot better to handle apart from anything else.
06:17 PM Tom_L: sexy
06:17 PM Tom_L: i do too, mine is black
06:17 PM andypugh: Yes, my old Pi4 is black, I chose a different colour to distinfuish them.
06:18 PM Tom_L: just wish they could make it so you could plug a header into the board
06:18 PM Tom_L: no room though
06:19 PM andypugh: I have seen some with a header extender
06:19 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Rpi4/Case_top_idc.jpg
06:20 PM CaptHindsight: andypugh: the Rpi4 stored the bootloader in a serial EEPROM, is this still the case?
06:21 PM andypugh: I am not sure about the bootloader.
06:21 PM andypugh: The eeprom I am talking about is the USB one, I think.
06:21 PM unterhaus: couldn't find a purple one, but the case i found came with header extenders
06:22 PM CaptHindsight: https://heise.cloudimg.io/width/2000/q75.png-lossy-75.webp-lossy-75.foil1/_www-heise-de_/ct/imgs/04/2/7/4/3/7/8/1/RPi4-Block-e4359fdb4be588cf.png
06:22 PM perry_j1987: ok whats goin on with this... last night i played with scale to get the x axis to move dead on what i commanded it to
06:22 PM perry_j1987: come back today and its moving little less than half what i command it to
06:23 PM Tom_L: G90 vs G91?
06:23 PM Tom_L: that can bite ya
06:24 PM perry_j1987: was just in MDI giving g0 xblahblah
06:24 PM andypugh: Did you change microsteps on the driver, but not power-cycle it?
06:26 PM Tom_L: perry_j1987, what mode is it currently in?
06:26 PM Tom_L: it shows a list of active codes
06:27 PM perry_j1987: nope havnt touched drivers
06:27 PM Tom_L: unless you forgot to save the config or such
06:27 PM perry_j1987: i see g90 i just powercycled the whole box though
06:27 PM Tom_L: that's default iirc
06:28 PM Tom_L: so try G0 G91 X1 Fxx and see what happens
06:28 PM Tom_L: or Z in your case... forgot it was a lathe
06:29 PM Tom_L: or better use G01
06:29 PM perry_j1987: ya g91 worked heh
06:31 PM CaptHindsight: I'm not even sure of what the problem is with the Rpi4 and LCNC
06:32 PM Tom_L: building the image for the new rev i think
06:32 PM CaptHindsight: how is a USB EEPROM involved in this?
06:32 PM Tom_L: or at least testing it
06:32 PM Tom_L: no idea
06:32 PM Tom_L: maybe he just wants to test it
06:33 PM CaptHindsight: the non native English on the forums is difficult to follow
06:34 PM asdfasd: on the website it says current release 2.8.1 but cant find download, am I blind?
06:34 PM Tom_L: yes
06:35 PM Tom_L: asdfasd, what version are you currently running?
06:36 PM Tom_L: http://buildbot.linuxcnc.org/
06:36 PM asdfasd: 2.7.13 but the PC is very old and upgrading the whole PC now
06:36 PM asdfasd: now on new PC need new version :)
06:36 PM Tom_L: point your source list to the new version
06:36 PM Tom_L: or did you want the iso?
06:36 PM Tom_L: you will have to upgrade from the iso to get the latest
06:37 PM asdfasd: iso
06:37 PM asdfasd: so the latest iso is 2.8.0?
06:37 PM Tom_L: http://linuxcnc.org/downloads/
06:37 PM Tom_L: afik yes
06:37 PM CaptHindsight: asdfasd: so you were right, there is no link to a 2.8.1 ISO
06:37 PM asdfasd: I see
06:37 PM Tom_L: http://buildbot.linuxcnc.org/dists/buster/
06:37 PM Tom_L: then upgrade to the latest
06:38 PM CaptHindsight: http://www.linuxcnc.org/iso/
06:38 PM asdfasd: yep, I seen the last link with iso folder
06:39 PM asdfasd: and downloading iso now
06:39 PM CaptHindsight: I think what asdfasd meant was that are no clear instructions on how to get 2.8.1
06:39 PM Tom_L: then set your sources and do the upgrade
06:39 PM Tom_L: i agree
06:39 PM asdfasd: Im reading something about the kernel
06:39 PM Tom_L: no iso though
06:40 PM asdfasd: its not rtai now, should I change it to rtai for LPT port ?
06:40 PM asdfasd: is it better?
06:40 PM Tom_L: there is no rtai for buster yet
06:40 PM Tom_L: wheezy would be the last one i believe
06:40 PM asdfasd: there is some instructions on the download page
06:41 PM CaptHindsight: asdfasd: test the latest using preempt_rt, see what your latency results are
06:41 PM asdfasd: LinuxCNC 2.8.0 Debian 10 Buster RTAI
06:41 PM asdfasd: The more adventurous can install the Buster ISO and then install the experimental RTAI kernel as described in 2.8 documents
06:41 PM Tom_L: ahh i see... you're one of those :)
06:41 PM Tom_L: welcome
06:41 PM asdfasd: those?
06:41 PM Tom_L: adventurous types
06:42 PM asdfasd: haha, not actually
06:42 PM asdfasd: that is why Im asking now
06:42 PM asdfasd: is it worth or not
06:42 PM CaptHindsight: test preempt_rt first
06:42 PM asdfasd: I need stable machine, and because its on LPT port the timing is important
06:42 PM Tom_L: ^^
06:42 PM CaptHindsight: building RTAI is easier said than done
06:43 PM CaptHindsight: I'd say that if you have to ask how to build it then you probably shouldn't
06:44 PM asdfasd: well, there is instructions, Im not a guru, but I can follow them if that will make it better
06:46 PM asdfasd: I was stuck with that wheezy for a long time, with network problems, hard to share files with windows 10
06:46 PM asdfasd: some protocol issues, so decided to upgrade now everything
06:46 PM Tom_L: i'm still running wheezy on my mill
06:47 PM asdfasd: me too, but cant open shared folders from windows 10 machine
06:48 PM Tom_L: i use 7 for my cad work and once in a while it burps trying to access it from the mill
06:48 PM asdfasd: win 7 is fine
06:48 PM CaptHindsight: asdfasd: there is no support to do so, some people don't follow the RTAI howto and then post their lunch menu, bowling scores and random typos their cat makes....
06:49 PM Tom_L: i can't say much good about 10 personally
06:49 PM asdfasd: but in 10 they decided to disable the old protocols as not secure
06:49 PM Tom_L: if that's the only issue, i'd resort to the tennis shoe net when required
06:50 PM Tom_L: i keep a thumb drive in my pocket
06:51 PM asdfasd: yeah, but some time need to make a lots of quick changes
06:51 PM Tom_L: incentive to get it right!
06:51 PM asdfasd: most of the time I dont use network for a year
06:52 PM asdfasd: but 2 days ago decided to mill some PCBs
06:52 PM asdfasd: and it was an adventure
06:52 PM Tom_L: some have success doing that, i never did
06:53 PM Tom_L: i etched most of what i needed until i got lazy and ordered prototypes
06:53 PM asdfasd: haha
06:54 PM asdfasd: my problem is that I do it once in year
06:54 PM asdfasd: and until next time I forget what and how to make it
06:54 PM asdfasd: haha
06:55 PM Tom_L: i was a whiz at eagle but i'd have to start over now it's been so long
06:56 PM andypugh: CaptHindsight: One of the problems is that the current LinuxCNC realtme pi kernel boots the Pi OK, but with no USB it’s not as useful as one might wish.
06:57 PM andypugh: CaptHindsight: But, yes, after a bit over a week of chasing my tail, compileng packages and SD images only to get “No, that doesn’t work either” I feel inclined to give up.
06:58 PM asdfasd: ok, now running live iso, how to test the latency
06:58 PM andypugh: It doesnt help that others are saying “Bit you can just download this, and that, them that, and compile from source”. Yes, of course you can do that. That’s what I mean by ginving up, limiting Pi use to those who can figure out all that stuff.
06:58 PM andypugh: latency-test
06:59 PM andypugh: asdfasd: CNC menu, latency test. Or command line, latency-test
06:59 PM andypugh: Or, if you prefer, latancy-histomgram
06:59 PM andypugh: latency-histogram, that is
07:00 PM * JavaBean would offer andy a b33r for his work on the raspi version... notes the mail services in the us don't like mailing single b33rs
07:02 PM asdfasd: found it
07:03 PM asdfasd: it show low values until I try to resize the window
07:03 PM asdfasd: then values go big
07:04 PM asdfasd: it was approx 20000 ns and after resizing it is 248000
07:04 PM asdfasd: is that normal?
07:05 PM Tom_L: not very good for a parallel port system
07:06 PM Tom_L: you need to exercise the system or the test isn't 'real'
07:06 PM roycroft: javabean: a few years ago some sports fans in my brew club made a bet with a brew club in georgia on the outcome of a football game
07:07 PM roycroft: the georgie folks won
07:07 PM Tom_L: asdfasd, you say you got a new MB to upgrade to?
07:07 PM roycroft: at the time, homebrew was still illegal in georgia, but my brew club had to pay up by sending them some homebrew
07:07 PM asdfasd: well... new mean better than the old one, but still old haha
07:07 PM roycroft: what they did was send some "yeast samples" that were suspened in a liquid medium to keep them alive
07:07 PM Tom_L: seems you may be better off with what you currently have
07:08 PM asdfasd: the one I currently have is very, very old
07:08 PM asdfasd: I build it 2007
07:09 PM Tom_L: and it still works
07:09 PM asdfasd: yep
07:09 PM Tom_L: don't mess with a good thing...
07:09 PM asdfasd: I will keep it as back up for a while
07:10 PM Tom_L: i certainly would
07:10 PM asdfasd: is there any graphics drivers for linux?
07:10 PM asdfasd: do I need to update anything?
07:18 PM asdfasd: disabled some cpu features in bios but still the same
07:19 PM Tom_L: probably not a good candidate for preempt-rt
07:21 PM asdfasd: :(
07:21 PM asdfasd: its much much faster than the old one
07:22 PM Tom_L: onboard graphics?
07:22 PM Tom_L: might fiddle with that and see if it improves
07:24 PM asdfasd: yes, and there is no option to disable it, it was with external graphics until now, just removed it and will try with internal
07:24 PM asdfasd: much better now
07:25 PM perry_j1987: heh first chips with the new lathe made
07:25 PM asdfasd: so external graphics make troubles
07:33 PM asdfasd: tried different external graphics
07:33 PM asdfasd: now is better
07:52 PM perry_j1987: still trying to figure out why linuxcnc is hanging on this gcode line without progressing
07:52 PM perry_j1987: i was able to get the spindle encoder reading the commanded speed
08:14 PM Centurion-Dan2 is now known as Centurion_Dan
08:20 PM Juzzy: So I have X and Z homing. but Y is being stubborn still. I have the switch in hand since I cant get it to move any closer to me. INI: https://pastebin.com/2sWiD3AH | HAL: https://pastebin.com/kBHj6dpi
08:21 PM Juzzy: So Y is still saying I must home it before I can JOG it. but when I click Home, it works on X and Z, but Y doesnt do anything.
08:21 PM Juzzy: I've tried HOME_LATCH_VEL = 1.5 and HOME_LATCH_VEL = -1.5 in Y.
08:22 PM Juzzy: I keep them both the same, but should I remove the home_latch and search from the mirrored joint?
08:38 PM W1N9Zr0: Juzzy - your joint.3.home-sw-in is not connected
08:44 PM CaptHindsight: W1N9Zr0: good catch
08:51 PM perry_j1987: is there any special way to coming up with steps per mm for these g320x driven servos
08:51 PM perry_j1987: or just new_e_steps = old_e_steps * (100 / distance_actually_moved) till i dial it in
08:54 PM W1N9Zr0: perry_j1987 - leadscrew or belts? use the appropriate calculator from here https://blog.prusaprinters.org/calculator_3416/
08:55 PM perry_j1987: ballscrews
08:56 PM perry_j1987: not exactly sure their specs
08:56 PM CaptHindsight: perry_j1987: what pitch screws?
08:56 PM CaptHindsight: ah hah
08:56 PM perry_j1987: http://www.techno-isel.com/Education1/Ed_MetalLathe.htm all i know lol
08:57 PM CaptHindsight: how many steps per rev on the axis?
08:57 PM CaptHindsight: rotate the screw for one complete turn, measure the travel
08:58 PM CaptHindsight: might have been easier when the motors were not yet on the screws
08:58 PM perry_j1987: it seems to move acording to calipers almost dead on 10mm when commanded to move 10 with 1579 steps per mm
08:58 PM CaptHindsight: are there any belts or pulleys involved?
08:58 PM perry_j1987: none
08:59 PM CaptHindsight: it should be dead on vs almost
08:59 PM perry_j1987: im not quite to dead on yet
09:00 PM CaptHindsight: just about to walk out the door here
09:06 PM W1N9Zr0: what's your encoder lines/rev or count/rev?
09:08 PM W1N9Zr0: that page says Ball screw pitch: 2.5mm
09:08 PM W1N9Zr0: if your encoder is 160 lines/rev (640 counts/rev), 1600 steps/mm would be the correct number
09:08 PM W1N9Zr0: 2.5 * 640 = 1600
09:09 PM perry_j1987: they said 1000 4000
09:09 PM perry_j1987: 4000 per rev someone said
09:10 PM W1N9Zr0: oh oops, i have it backwards. it's counts/rev / leadscrew pitch
09:10 PM W1N9Zr0: so 4000 / 2.5 = 1600
09:11 PM W1N9Zr0: (counts / rev) / (mm/count) = (counts / mm)
09:11 PM perry_j1987: and on the drivers g320x pulse multiplier dips is their microstepping ?
09:12 PM W1N9Zr0: oops, i can't type anymore. (counts / rev) / (mm / rev) = (counts / mm)
09:12 PM W1N9Zr0: yes, pulse multiplier is microstepping
09:13 PM perry_j1987: ok so mine are set to none so in stepconf i put motor steps as 4000 for the encoder
09:13 PM perry_j1987: change driver microstepping to 1 and the leadscrew pitch to 2.5 and shows 1600 at the bottom
09:13 PM W1N9Zr0: yup
09:14 PM perry_j1987: ok
09:14 PM perry_j1987: testing
09:15 PM perry_j1987: what should max velocity and acceleration be for servos
09:16 PM Juzzy: W1N9Zr0: sorry for the delay was trying stuff. in RE: joint.3.home-sw-in - how can i mirror that input to the same as the other Y?
09:17 PM Juzzy: I already have a "net home-y => joint.1.home-sw-in"
09:17 PM W1N9Zr0: just add another line "net home-y => joint.3.home-sw-in"
09:18 PM Juzzy: ah ok.
09:18 PM Juzzy: but I dont have another input port available :(
09:18 PM W1N9Zr0: but the whole point of setting up separate joints for a gantry is to home them to separate switches
09:19 PM W1N9Zr0: if you're just going to have 1 switch, you might as well just use 1 joint, and wire two step/dir pins to the same stepgen
09:19 PM Juzzy: ah crap i just noticed this also
09:19 PM Juzzy: i have "net aenable joint.3.amp-enable-out => stepgen.3.enable"
09:19 PM Juzzy: net a*
09:19 PM Juzzy: does that need to be y* ?
09:20 PM Juzzy: so, I did the wizard with XYZA, because I thought it was how you did it. then went back and swapped it to XYYZ
09:22 PM Juzzy: W1N9Zr0: there isnt a way to move them independantly though. like, maybe 3-4 steps and then it binds up. I want to avoid that. having two switches slight off will be a problem
09:22 PM W1N9Zr0: aenable in this case is just a signal name, so it doesn't matter
09:24 PM Tom_L: gantrys seem like a pita
09:24 PM Juzzy: So question then, if I setup a home-pin in .hal is there a way i can use the same paraport pin w/o causing issues?
09:25 PM Tom_L: not if you're homing both YY separately
09:25 PM Juzzy: i will never do that
09:25 PM Tom_L: they each need to know when they've reached their home point
09:25 PM Juzzy: they are working the same arm, 5' apart
09:26 PM Tom_L: if you want to tie the switches together use an and gate to the input pin then
09:26 PM Juzzy: this is ok to do then? net home-y <= parport.0.pin-13-in ; net home-yy <= parport.0.pin-13-in
09:27 PM Tom_L: the <= does nothing but it technically is backward
09:27 PM Juzzy: if I had another available input I could put them in parallel
09:27 PM W1N9Zr0: why do you need separate signal nets for this?
09:27 PM Juzzy: Tom_L: that's a copy/paste from my hal file
09:28 PM Tom_L: he doesn't
09:28 PM Juzzy: correct. I just want to home both joints off 1 switch
09:28 PM Tom_L: i don't see that you can if you're running the steppers independently
09:29 PM Juzzy: ok so if I have two joints on an axis, is there a way to manually home them?
09:29 PM W1N9Zr0: perry_j1987 - http://linuxcnc.org/docs/html/config/stepconf.html#sub:finding-maximum-velocity
09:30 PM Juzzy: that woulds be absolutly fine, I can manually do the other axis np
09:30 PM Juzzy: would*
09:30 PM Tom_L: you're just running out of pins
09:31 PM Tom_L: are all the enable lines separate or tied to one enable?
09:31 PM Juzzy: ya unless i repurpose my limits
09:31 PM Juzzy: I could repurpose my Z
09:31 PM Juzzy: and then link the 2 Y inputs
09:32 PM Tom_L: do what you need to but you need 2 inputs for Y YY
09:32 PM Juzzy: and just manually do Z since that'll never change really
09:32 PM W1N9Zr0: if your motors are moving together, and they will not misalign relative to each-other, you should just use 1 joint and output its stepgen on to sets of step/dir pins
09:32 PM W1N9Zr0: *two sets
09:32 PM W1N9Zr0: or even output it to 1 set of pins, and tie those pins to two motor drivers
09:32 PM W1N9Zr0: (unless you need to invert direction on one of them)
09:33 PM Tom_L: very likely
09:33 PM Juzzy: W1N9Zr0: I could do that, not sure if just sending the signal from the same pin and just flipping DIR is safe?
09:34 PM Juzzy: i dont know how DIR works, is it sent -5 vs +5 or +5 vs 0v ?
09:34 PM Juzzy: or is 0V meaning the same direction as the last puse?
09:34 PM Juzzy: pulse
09:34 PM Tom_L: in the [AXIS] or [JOINT] section of the ini, just set scale negative on one and that will invert it
09:35 PM W1N9Zr0: you can wire the same stepgen dir signal to two pins
09:35 PM W1N9Zr0: and then set the pin one of the pins to be inverted using parport.<portnum>.pin-<pinnum>-out-invert
09:36 PM Juzzy: but then i'll have to have 2 joints, right?
09:36 PM Tom_L: don't think it works that way
09:37 PM Juzzy: I could send PUL to 2 drivers (i assume) and then reverse one of the DIRs between the driver and the motor. i just don't know if that works or not
09:37 PM W1N9Zr0: i think that will work, yes
09:37 PM Juzzy: i dont think it would.
09:37 PM Juzzy: ok maybe so then
09:37 PM Tom_L: you have a JOINT section for each of the motors right?
09:38 PM Juzzy: ya tom.
09:38 PM Tom_L: just invert scale
09:38 PM Tom_L: done.
09:38 PM Juzzy: INI: https://pastebin.com/2sWiD3AH | HAL: https://pastebin.com/kBHj6dpi
09:38 PM Juzzy: that doesnt help with homing
09:38 PM Juzzy: I already invert then pins to do this
09:39 PM Tom_L: if scale is inverted, the motor will spin the opposite direction
09:39 PM Juzzy: ok so jogging actually works
09:39 PM Juzzy: (sorta)
09:40 PM Tom_L: so one of them would be SCALE = -533.333333333
09:40 PM Juzzy: like it throws errors initially, but if I run the linuxcnc demo job. it errors out, after that I can actually jog + postition, but not -
09:41 PM Juzzy: where the gantry is at, it sets that as home. but that's the only way I can home it (probably like W1N9Zr0 pointed out the missing pin line)
09:41 PM Tom_L: add another pp card :)
09:41 PM Juzzy: I dont think changing the scale will solve the home problem, yanno?
09:41 PM Juzzy: ya.. or an ESS or something
09:41 PM Tom_L: try it i think it might
09:42 PM W1N9Zr0: i'm sure Tom_L's 2-joint scale solution can work too, but if you only have 1 home switch, and your motors naturally stay aligned together when powered down, i would use 1 joint and 1 stepgen to drive both motors for simplicity sake
09:42 PM Tom_L: ^^
09:42 PM W1N9Zr0: if you're going to invert a joint with scale, i think you have to invert all the joint limits and home offsets and such too
09:42 PM Tom_L: you would probably
09:42 PM Juzzy: so i might just steal the Z input, and split the Y pin with the z pin, that way they'll both home at the same time on the same switch, just will lose my z
09:42 PM Tom_L: for that joint
09:43 PM Tom_L: or step up to mesa
09:43 PM Juzzy: well first ima try to set both home inputs on the same pin. if that works, i'm done.
09:44 PM Tom_L: i'm running a 7i90 on my parallel port
09:44 PM Tom_L: 72 io
09:44 PM Juzzy: i'll do the -scale trick last, sounds like a pita.
09:44 PM Tom_L: not really
09:44 PM Tom_L: it's all software
09:44 PM Tom_L: it's actually the easiest
09:44 PM Juzzy: how does that fix the 2nd home tho?
09:44 PM Juzzy: does that -scale joint not need a home?
09:45 PM W1N9Zr0: you can also use one switch (or two switches in parallel on 1 pin, etc) for homing separate axis in sequence http://www.linuxcnc.org/docs/devel/html/config/ini-homing.html#_home_is_shared
09:45 PM Tom_L: if you tie step - direction pins to both drivers you'll have some pins to use for switches
09:45 PM Tom_L: or run them in series
09:46 PM Juzzy: W1N9Zr0: i've read this doc like 4 times. I thought that meant something completly different
09:46 PM Tom_L: my home / limit switches are shared
09:46 PM Juzzy: hell maybe i can just set HOME_IS_SHARED = 1 and be done
09:47 PM Tom_L: there are plenty of options
09:48 PM W1N9Zr0: you can share X and Z home for example, the idea is that X homes and gets off the switch so Z can use it next
09:48 PM W1N9Zr0: but you can't share home switches for joints that are part of the same gantry, that doesn't make sense
09:49 PM W1N9Zr0: the whole idea of multiple joints on a gantry is to home them separately to align them
09:52 PM Tom_L: so what you're saying is in the home sequence, home Z first then use it's switch input and the Y switch to home Y & YY
09:52 PM Tom_L: i've never and never will build a gantry so i haven't experimented with that
09:55 PM W1N9Zr0: yup, that's one way to get around your limited number of inputs
09:57 PM Juzzy: so it looks like home-y and x|z are built internally
09:57 PM Juzzy: like there's a net home-y
09:57 PM Juzzy: if I make a net home-yy
09:58 PM Juzzy: what ties home-yy to home-y ?
09:59 PM Juzzy: net home-y => joint.1.home-sw-in
09:59 PM Tom_L: nothing. they are 2 ends of a stick
09:59 PM Juzzy: net home-yy => joint.3.home-sw-in
09:59 PM Juzzy: ?
09:59 PM Juzzy: how do you define home-sw-in ?
10:00 PM Tom_L: to the physical pin
10:00 PM W1N9Zr0: home-y is the name of a "wire" connecting pins. when you do net home-y => joint.1.home-sw-in you tie the output end of the wire to a pin of a component
10:00 PM Tom_L: net switch3 parport.0.xxx => joint.3.home-switch-in
10:01 PM W1N9Zr0: earlier you had net home-y <= parport.whatever you tied the input end of the wire to another component pin
10:03 PM Tom_L: 3 components to a net
10:03 PM Tom_L: net name, signal and wire
10:03 PM Tom_L: or some such nomenclature
10:04 PM Juzzy: ya im reading hal -> net naming
10:04 PM Juzzy: http://www.linuxcnc.org/docs/2.5/html/hal/basic_hal.html
10:04 PM Tom_L: in my example, switch3 is the net name
10:04 PM Tom_L: parport.0.xxx is the wire
10:04 PM W1N9Zr0: you don't need to create a home-yy net, you can just reuse home-y, nets can tie to multiple pins. net home-y => joint.3.home-sw-in
10:04 PM Tom_L: joint.3.xxx is the signal
10:05 PM Juzzy: ahhhhh
10:05 PM Juzzy: "nets can tie to multiple pins"
10:06 PM Tom_L: multiple signals
10:06 PM Juzzy: got it.
10:06 PM Tom_L: err wires
10:06 PM Juzzy: damn i wish there was a "best practices" for this
10:07 PM Juzzy: gets hair quick. could only imaging doing an A B C axis
10:07 PM Tom_L: not really any different
10:07 PM Tom_L: they just spin
10:08 PM Juzzy: aight so, i like the idea of 2 switches on both sides of the gantry, that way it'll ensure squareness
10:08 PM Juzzy: i'll steal the Z home pin and use that I guess
10:09 PM Juzzy: so no matter what hackery I do, nothing will fix that issue
10:13 PM W1N9Zr0: shared home switches are still an option, you can put X and Z on one pin, or any other combination except Y1 and Y2
11:05 PM CaptHindsight: the yeasts live!
11:05 PM CaptHindsight: was merely resting
11:45 PM Juzzy: beer? wine?
11:49 PM Juzzy: always keep some ec-1118 on hand just in case