Back
[01:10:01] <minibnz> got a question.. have a mill that is run by linuxcnc and want to integrate the spindle control into EMC.. i have the hal file so i can get a PWM and a digital output out of a pin on my parport.. i want to get it to control a RC servo hooked to the speed knob..tried following the wiki but the servo wont move. i looked at the waveforms with my CRO and i can see the waveforms match my arduino test board.. anyone got any ideas?
[01:11:16] <minibnz> tried two different servo's a big one and a wee little one.. but both wont move under emc2.. happily work with arduino
[01:13:33] <minibnz> am i going to have to add a arduino to convert the PWM to RC PPM waveforms. seems like overkill
[01:49:14] <archivist> you could write that mode into the pwm driver
[02:01:09] <enleth> is there any open-source electrical design software optimized for relay logic and machine wiring?
[02:26:26] <Deejay> moin
[05:29:51] <jthornton> morning
[05:29:59] <XXCoder> no
[05:30:01] <XXCoder> heh
[05:30:05] <XXCoder> whats up
[05:31:41] <jthornton> first cup, setting up a J1900 on uspace for the BP
[05:49:37] <jthornton> trying to see if I can get linux mint to work with uspace
[05:49:53] <XXCoder> cool
[05:52:04] <jthornton> I have a 7i92 to experiment with
[05:52:38] <jthornton> I got it to work on debian 7
[06:37:16] <ehippiecncbuilde> hello!
[06:37:38] <ehippiecncbuilde> Can anyone suggest a logical signal that I can use for a spindle fault?
[06:38:03] <XXCoder> "launch all nukes" signal
[06:38:09] <XXCoder> or maybe batman signal?
[06:38:14] <ehippiecncbuilde> There're axis.?.amp-fault-in for drive faults, but none for the spindle
[06:38:17] <XXCoder> heh joking. dunno
[06:39:09] <jthornton-> I would think you want amp fault so the machine stops
[06:39:26] <ehippiecncbuilde> This is an issue I've wondered some time ago, and I can't but find it illogical that there is non.
[06:39:47] <ehippiecncbuilde> But in order to use amp-fault I have to set up an axis that doesn't exist.
[06:39:55] <jthornton> what would you want the spindle fault to do?
[06:40:04] <ehippiecncbuilde> stop the motion so nothing breaks up
[06:40:28] <archivist> OR it into your estop then
[06:40:29] <ehippiecncbuilde> E.g. a signal that linuxcnc would know that there is a problem with the spindle, so that it would stop all operations.
[06:40:35] <jthornton> or2
[06:40:51] <ehippiecncbuilde> That's one option, but I find it slightly illogical.
[06:41:02] <jthornton> reprogram the software
[06:41:09] <ehippiecncbuilde> I'm HW designer :D
[06:41:44] <XXCoder> sure you can solder in software! ;)
[06:42:53] <archivist> see this input motion.enable IN BIT
[06:42:53] <archivist> If this bit is driven FALSE, motion stops, the machine is placed in the "machine off" state, and a message is displayed for the operator. For normal motion, drive this bit TRUE.
[06:43:52] <archivist> also has motion.spindle-inhibit IN BIT
[06:43:52] <archivist> When TRUE, the spindle speed is set and held to 0.
[06:44:13] <XXCoder> I guess most spindle supports that
[06:44:14] <jthornton> and they all probably do the same thing internally
[06:44:27] <ehippiecncbuilde> that might work ok
[06:45:01] <archivist> motion.enable seems the right pin to me
[06:51:15] <ehippiecncbuilde> Do I understand correctly, that machine-off state turns off/disables drives?
[06:53:35] <archivist> to me a controlled stop is better than a drive off stop, it is a choice you have to make
[06:53:45] <ehippiecncbuilde> If drives are turned off and there's a heavy axis on motion, or heavy spindle on Z-axis (assuming vertical), the motion might continue or spindle might drop without control. Due to this, motion.feed-inhibit might be better.
[06:54:10] <ehippiecncbuilde> Uncontrolled stop in quite undesirable.
[06:54:14] <ehippiecncbuilde> is
[06:56:07] <jthornton> motion.feed-inhibit IN BIT
[06:56:08] <jthornton> When this pin is TRUE, machine motion is inhibited (this includes jogs, programmed feeds, and programmed rapids, aka traverse moves).
[06:56:08] <jthornton> If the machine is performing a spindle synchronized move when this pin goes TRUE, the spindle synchronized motion will finish, and any following moves will be inhibited (this is to prevent damage to the machine, the tool, or the work piece).
[06:56:08] <jthornton> If the machine is in the middle of a (non-spindle synchronized) move when this pin goes TRUE, the machine will decelerate to a stop at the maximum allowed acceleration rate.
[06:56:10] <jthornton> Motion resumes when this pin goes FALSE.
[06:56:30] <jthornton> http://linuxcnc.org/docs/2.7/html/man/man9/motion.9.html
[06:56:46] <jthornton> maybe you should read through them and select the one you need
[06:59:28] <ehippiecncbuilde> I did... from the pdf documentation, which doesn't say it all.
[06:59:50] <ehippiecncbuilde> Need to read the www-version it seems.
[07:00:16] <jthornton> the pdf unfortunately does not have the man pages
[07:00:35] <jthornton> well that is wrong they are there now
[07:00:59] <jthornton> CNC > Manual Pages
[07:01:29] <jthornton> unfortunately the pdf has no index
[07:02:40] <ehippiecncbuilde> Anyway, I think I got my answer. Thanks a lot!
[07:11:51] <Tom_itx> soggy day today
[07:12:04] <jthornton> I've not looked outside yet
[07:12:37] <Tom_itx> http://www.intellicast.com/National/Radar/Current.aspx?animate=true&location=USKS0523
[07:12:47] <Tom_itx> looks clear now but the ground is wet
[07:12:55] <jthornton> I doubt there will be any lumberjacking activities today
[07:13:12] <Tom_itx> paul bunyon on strike?
[07:14:12] <jthornton> mickey rooney bunyon doesn't like to lumberjack in the rain
[07:14:55] <Jymmm> What abut blu
[07:15:09] <jthornton> she passed away in July
[07:16:26] <Jymmm> Tom_itx: Poor Toto
[07:16:44] <jthornton> might not be too bad today, looks to be passing north of us
[07:16:44] <Tom_itx> toto is still spinning in the clouds
[07:18:43] * jthornton looks for the solder station
[07:18:55] <Tom_itx> still??
[07:21:17] <jthornton> found it
[07:22:25] <jthornton> it was either out in the shop/garage somewhere or done here in the beer cave
[07:22:55] <Jymmm> Crap, snow/showers today =(
[07:24:15] <Jymmm> but a HEAT WAVE of 63 next week ;)
[07:26:58] <jthornton> snow? I thought that was illegal in CA
[07:27:29] <Jymmm> No, just everything else is.
[07:28:55] <Jymmm> jthornton: We still have Tahoe and Mammoth snow skiing
[07:30:35] <Tom_itx> do they take the fin off their surfboard and use it as a snowboard?
[07:31:10] <Jymmm> Nah, that's silly... they use body boards for that ;)
[07:31:40] <Jymmm> snap in half like a popsicle... now two can snowboard
[07:32:47] <SpeedEvil> http://imgur.com/oDjyIOy - on that topic
[07:33:23] <Jymmm> YEAH!!!! Thats awesome
[09:53:37] <beikeland> so I'm making ever more progress with my mach2/emc conversion. Got motion and homing running. Lacking the speed. I keep getting "joint 2 following error"
[09:54:05] <beikeland> and i read its either too fast acceleration or not fast enough base thread; but that doesn't help me fixing it.
[09:54:27] <beikeland> Is it possible Mach3 only ran faster, at the cost of missed steps (hardware is the same)
[09:56:16] <beikeland> Currently i have:
[09:56:19] <beikeland> BASE_PERIOD = 55000 SERVO_PERIOD = 1000000 MAX_ACCELERATION = 25.0 STEPGEN_MAXACCEL = 25
[09:58:22] <skunkworks> stepgen_maxaccel needs to be atleast 25% higher than max_acceleration...
[09:58:38] <skunkworks> that is why you are getting following errors
[09:59:31] <skunkworks> MAX_ACCELERATION = 25.0 STEPGEN_MAXACCEL = 30
[09:59:40] <skunkworks> try that
[10:12:45] <beikeland> skunkworks: that seems to make it consistent at trigging following error when i do rapids but the rest is okay. so i reduce the velocity and it seems to be working. But its very much slower than it was; cannot speak to the accuracy mach3 had as it saw very little use
[10:13:25] <skunkworks> what is your input scale and such?
[10:14:37] <beikeland> scale for 2(z) is 800 and 0/1(x/y) is 400.
[10:15:36] <beikeland> are the acceleration parameters relative to the scale_
[10:15:37] <skunkworks> what max velocity are you trying for?
[10:16:09] <beikeland> It was running at 2500mm/min for xy and 1500 for z
[10:16:40] <beikeland> But i just want it as fast as it can go while maintaining accuracy
[10:17:12] <skunkworks> with linuxcnc those 2 things are not related..
[10:17:54] <skunkworks> you can tell linuxcnc how close you want it to follow the programmed path. (regardless of speed)
[10:18:14] <pcw_home> STEPGEN_MAXVEL also needs to be about 25% greater than MAXVEL
[10:18:45] <pcw_home> (or you will get following errors during rapids)
[10:19:30] <beikeland> it apears I don't have a STEPGEN_MAXVEL line in my .hal, i shall try adding
[10:19:56] <pcw_home> if its not there it should be OK
[10:20:13] <skunkworks> I think max vel was decided it didn't matter
[10:20:20] <pcw_home> (default is 0 which means no limit)
[10:21:22] <pcw_home> if you set stepgen maxvel it needs headroom just like maxaccel but 0= no limit is OK
[10:22:13] <beikeland> i have more pressing things than speed, so at least its usable albeit a tad slower now. Thanks!
[10:23:03] <beikeland> I'll revisit the speed when its been squared and it has milled itself an input board :)
[10:24:41] <beikeland> Actually that brings me to the switches. I presume I want N.C. and NPN type proximity switches?
[10:25:41] <pcw_home> for a parallel port, NPN since parallel port inputs are pulled up
[10:27:10] <pcw_home> if you are using a BreakOut Board (BOB), it may be different
[10:27:16] <beikeland> I'm using the output pins as well, so i need to add pullups (or pulldowns) but NPN removes the need to do level shifting. And normal closed is prefered to reveal cable issues?
[10:27:44] <beikeland> DIY on the breakout board. Its a breadboard at the moment
[10:28:15] <beikeland> I dind't find any reasonable BOBs with data lines as inputs
[10:28:52] <pcw_home> Yes NC (make sure you cant lose the ground connection on NPN proxes connected directly to a parallel port or you will fry the port)
[10:30:24] <beikeland> Gonna do opto isolators just to be on the safe side, cool, I'll go do some work. Thanks for the help so far :)
[11:12:12] * Loetmichel just lend himself the companys whore (2006 mercedes E-class diesel station wagon) because his car is at the Garage for spilling oil ... WHAT a pice of crap. Chec engine is continuously lit... the exhaust hits the bottom of the car at any slight bump (*CLANKCLANK*). the steering points slighty to the right when coasting, slightly more so when accelerating... and HARD LEFT when braking...
[11:12:12] * Loetmichel and to top it off it reeks like someone smoked a full box of havannas ... last week. and didnt clean the asthray... but its still better than walking the 12km one way to the comany ;)
[11:28:51] <_methods> hahah just scored a cincinnati #2 horizontal for $100
[11:30:13] <_methods> heh another one just went for $100
[11:30:17] <_methods> cat50 varnamo
[11:30:30] <CaptHindsight> oh sure now you tell us
[11:31:13] <_methods> heh i don't want anyone outbidding me lol
[11:31:39] <_methods> there's another cinci model 520 coming up
[11:31:45] <_methods> that thing is huge
[11:32:12] <CaptHindsight> I sold my last Cincinnnattitti (sp?) a few years ago
[11:32:57] <_methods> i'm tryin to find a small horizontal like a harding um-tm
[11:33:03] <CaptHindsight> those horizontals are often on ebay
[11:33:04] <_methods> one day i'll find one
[11:33:15] <_methods> yeah they want too much for them
[11:33:24] <_methods> i just snatched a #2 for $100
[11:33:34] <_methods> they want $3k for a umtm on ebay
[11:33:52] <CaptHindsight> sitting here waiting for snow but it's raining
[11:33:55] <gregcnc> anything that fits easily in a garage is expensive
[11:34:04] <_methods> yeah
[11:34:21] <CaptHindsight> you need bigger garages
[11:34:25] <gregcnc> and can be moved by one or two people
[11:34:28] <gregcnc> yes I do
[11:34:29] <_methods> i got my eyes on a 00-8 coming up
[11:35:14] <_methods> 0-8 that is
[11:36:11] <_methods> king 42" vtl comin up
[11:36:22] <_methods> be interesting to see what that goes for
[11:36:54] <gregcnc> I'm looking for one of these there is still room in the garage.
http://www.machineseeker.com/ma2/bilderanzeigen-A2329808-1-english.html
[11:37:25] <_methods> ah that's pretty spiffy
[11:37:57] <gregcnc> not too many around
[11:38:01] <_methods> you'll have to wait till some school gets rid of their's
[11:38:12] <gregcnc> they are starting
[11:38:23] <_methods> not many shops would buy something like that so they'll all be at schools
[11:38:52] <gregcnc> Haas has a great deal so none of the schools are replacing the tiny Emco with newer Emco.
[11:39:03] <_methods> yeah
[11:39:13] <_methods> haas basically gives them to schools
[11:39:14] <Sync> well, it is the same with solidworks
[11:39:24] <Sync> they push it into the schools so people prefer it in their jobs
[11:40:05] <gregcnc> Fusion is trying hard to do the with startups
[11:41:21] <gregcnc> I know of two of these machines.one has little tooling the other is expensive
[11:50:49] <CaptHindsight> gregcnc: a bit over priced
http://chicago.craigslist.org/nwc/tls/5428180002.html
[11:52:55] <gregcnc> I contacted him a year ago? he's in LITH irc, but over priced.
[11:53:48] <gregcnc> There was a very dirty 320 near Milwaukee for 3900
[11:54:18] <CaptHindsight> HWACHEON CNC-LATHE Model#Hi-ECO31A - $35 I think that this is not the actual price
[11:54:25] <CaptHindsight> http://milwaukee.craigslist.org/tls/5429490053.html
[11:58:57] <Sync> I'd take it for $35
[11:59:30] <CaptHindsight> I'd go as high as $40 :)
[11:59:54] <gregcnc> $35k
[12:00:43] <Sync> nah
[12:00:45] <gregcnc> http://www.exapro.com/used-hwacheon-hi-eco-31a-2-axis-cnc-lathe-with-fanuc-otc-control-10-chuck-p40123114/
[12:01:48] <gregcnc> what it's actually worth i don't know
[12:26:36] <Cromaglious> bbiab
[12:26:47] <Cromaglious> rebooting this POS winbloz machine
[12:40:07] <DaPeace> damned that fusion 360 is great for being free :-D
[12:48:33] <CaptHindsight> anyone else having trouble getting packages to the UK that involved Royal Mail?
[12:48:54] <CaptHindsight> they are starting to screw up every one
[12:54:57] <CaptHindsight> 400w china spindle showed up and the bearings don't spin smoothly...whoodah thought it?
[12:55:16] <CaptHindsight> yah think i got taken for $49.95? :)
[12:56:27] <gregcnc> brush spindle? The listing didn't say "kit"?
[12:59:01] <Sync> let it run in and see if it improves CaptHindsight :P
[12:59:20] <CaptHindsight> it's a DC router motor
[12:59:36] <CaptHindsight> you have to break in the bumpy grease
[13:00:13] <DaPeace> i used one of that 400w-spindles too but after some month i made a jump to 1500w and i will never miss the old one :-D
[13:02:09] <CaptHindsight> the 1.5KW also arrived, the bearings are much better
[13:02:25] <DaPeace> yeah the newer one are not that bad.
[13:02:36] <CaptHindsight> not using either for cutting anyway
[13:02:52] <DaPeace> so you are collecting them? :-D
[13:03:08] <CaptHindsight> well maybe the 400w for films
[13:03:18] <CaptHindsight> just spinning stuff
[13:03:40] <CaptHindsight> coatings, inks, etc
[13:04:25] <DaPeace> ah ok.. a little other task than milling :-D
[13:06:01] <CaptHindsight> DaPeace: which version 1.5KW spindle did you get?
[13:06:21] <CaptHindsight> did it mention the 4 Germany bearings?
[13:06:32] <CaptHindsight> air or water colled?
[13:06:58] <gregcnc> the Germany bearing scam is the same as the Swiss skate bearing scam.
[13:07:02] <DaPeace> i think i have one with 3 bearings but they mentioned the germany-bearings (which are form a company named "germany" as far as i know) and i have a air cooled
[13:07:40] <DaPeace> yeah i think the bearings cost more than the whole spindle when you take german ones..
[13:08:52] <CaptHindsight> came with English using of the manual
[13:09:36] <CaptHindsight> or a double sided photocopy of a list of motors
[13:10:41] <CaptHindsight> this ones from near Shanghai
[13:10:59] <CaptHindsight> I'll have to stop by their factory
[13:11:34] <CaptHindsight> have to find that Germany Bearing in China some day
[13:12:54] <CaptHindsight> one student could proofread a few hundred Engrish data sheets in a day
[13:13:08] <CaptHindsight> but they just won't do that
[13:15:40] <DaPeace> i got one without any information.. tried to auto-measure it from the inverter and i get hot as hell..
[13:19:41] <CaptHindsight> who makes a 5 or 10V out digital motor speed controller without feedback that has a digital display of rpm and controllable ramps?
[13:20:20] <CaptHindsight> so similar to VFD display and control
[13:20:40] <gregcnc> how will you get rpm with no feedback?
[13:20:48] <gregcnc> with a DC motor?
[13:21:17] <CaptHindsight> sorry meant with feedback
[13:21:31] <CaptHindsight> dyslexic typing today
[13:26:15] <CaptHindsight> looks either programming a PLC or a *duino
[13:27:18] <CaptHindsight> even the
http://www.omega.com/pptst/OMDC-MD.html doesn't let you set the ramps
[13:28:59] <CaptHindsight> well it does but only one setting for accel rate and another for decel
[13:33:45] <gregcnc> regular servo amp with encoder feedback won't work with Soooo commands?
[13:34:32] <CaptHindsight> need UI, LCD and some buttons, somewhat user friendly
[14:06:52] <aventtini6> hello guys
[14:15:37] <CaptHindsight> heh order SSR boards from China get *duinos
[14:16:09] <CaptHindsight> and the box smells like diesel
[14:48:37] <enleth> finally
[14:48:41] <enleth> the mesa card is there
[15:04:21] <aventtini6> whereeeeeeeeee
[15:04:46] <aventtini6> ??
[15:06:34] <enleth> right here, in my workshop
[15:07:35] <aventtini6> :)))
[15:07:38] <aventtini6> good one
[15:07:48] <aventtini6> i was thinking on the website
[15:07:50] <aventtini6> :))))
[15:08:40] <jtektool> Can anyone out there recommend a good panel-mount jog wheel, my setup is mesa 5i25/7i77
[15:16:30] <CaptHindsight> jtektool: some have used this
http://www.automationtechnologiesinc.com/products-page/mpgs/rotary-maunal-pulse-generator
[15:16:55] <yasnak> nothing beats the smell of brand new moblemet 426. mmmmm
[15:19:32] <CaptHindsight> jtektool: for higher quality (and price)
https://www.euchner-usa.com/Category/45_1/HANDWHEELS.aspx
[15:21:00] <jtektool> definitely looking at more sturdy this machine is shaping together but i had a couple snaffus, such as i installed a touchscreen and got it working only to find out that gmoccapy has too many lines of resolution for my monitor. lol
[15:23:09] <jtektool> currently having issues with touchy too, as in it wouldn't load at all when i put it in the config. but right now im just working on getting the estop and my panel physically together. I suppose some rotary switches to select axis might be nice as well. if anyone else has any ideas what would be nice on a panel feel free to chime in
[15:24:35] <CaptHindsight> jtektool: post your configs if you need help
[15:25:13] <cradek> what do you mean by wouldn't load at all? do you get an error?
[15:25:42] <jtektool> brb going to go out to the machine....
[15:35:42] <jtektool> looks like something in my xorg.conf maybe; it says randr extension missing in Xlib.
[15:39:24] <jtektool> which is weird since gmoccapy works
[15:52:38] <jtektooltoo> so yeah it needs a touchy.hal and doesnt have one... does anyone have some sample files for a 5i25/7i77 combo?
[15:54:54] <cradek> http://linuxcnc.org/docs/html/gui/touchy.html#_hal_connections
[16:15:15] <enleth> does anyone use separate boxes for linuxcnc backend and GUI processes?
[16:15:23] <enleth> with X forwarding over network or something?
[16:15:49] <cradek> I'm sure some people do that - it's a normal feature of X
[16:19:23] <enleth> just wondering if there's a tangible benefit
[16:19:36] <enleth> I guess graphics cards are the worst offenders when it comes to hard RT
[16:20:01] <CaptHindsight> HMI at the beach, machine in the shop
[16:20:14] <enleth> so it would make sense to set up a box that just doesn't have one (with a BIOS that boots without a VGA adapter)
[16:20:24] <CaptHindsight> smartphone is too small to see Axis :)
[16:20:36] <FinboySlick> My attempts to do that weren't very successful actually.
[16:20:57] <FinboySlick> Though it was a while ago.
[16:21:07] <enleth> FinboySlick: separating the GUI over network?
[16:21:12] <CaptHindsight> FinboySlick: headless or at the beach?
[16:21:29] <enleth> there's also the possibility to use Xvnc
[16:21:38] <FinboySlick> CaptHindsight: I must have neglected the beach aspect of things ;)
[16:22:01] <FinboySlick> enleth: running axis as a remote X application.
[16:22:15] <enleth> it's still more reasonable to expect the pure software implementation of Xvnc and the ethernet adapter to maintain RT than it would be with regard to a graphics card
[16:22:32] <Jymmm> JT-Shop: cradek
http://www.bonappetit.com/recipe/simplest-asian-dressing
[16:29:21] <CaptHindsight> X and graphics drivers also talk direct to hardware if the developer was lazy and hoped nobody would care or notice
[16:30:12] <CaptHindsight> they stopped caring about low latency and real time, just getting their end to work
[16:34:04] <enleth> CaptHindsight: thus Xvnc
[16:34:13] <enleth> all software, fully preemptible
[16:34:57] <CaptHindsight> X without a GPU would be all software
[16:36:48] <enleth> I think that's what I just said
[16:38:48] <_methods> that
[16:38:53] <_methods> is why his name is CaptHindsight
[16:38:59] <malcom2073> lol
[16:39:07] <Deejay> gn8
[16:39:12] <_methods> it's his super power
[16:46:44] <Jymmm> It feels cheap, can lift/carry with one finger, but damn does it work well, I'm VERY impressed for what it is...
http://www.amazon.com/Eureka-Bagless-Upright-Vacuum-AS2113A/dp/B00WGEW9CA
[16:47:14] <Jymmm> The 'ASS ONE' vacuum =)
[16:51:02] <CaptHindsight> you gotta wake up pretty late in the afternoon to fool me :)
[16:56:35] <CaptHindsight> maybe remote USB display
[16:57:06] <CaptHindsight> USB GPU on Ethernet to USB bridge
[16:57:55] <CaptHindsight> last time I tried USB graphics on Linux it was pretty awful to setup
[16:58:39] <Jymmm> CaptHindsight: Try these graphics...
https://hackaday.com/2016/01/31/tv-transmitter-uses-esp8266/
[17:00:57] <CaptHindsight> Jymmm: so video over wireless! What will they think of next?
[17:01:40] <CaptHindsight> reminds me of that Latin phrase tele vision
[17:43:08] <Loetmichel> *meh* half second power net failure. Note to self: have to get new batteries for the UPS.. completely dead. didnt even sufficient for half a second... :-(
[17:54:12] <Tecan> whats the easiest way to get the linuxcnc 2.7.3 iso ?
[17:56:44] <JT-Shop> internet
[17:56:59] <JT-Shop> http://linuxcnc.org/
[17:58:06] * JT-Shop needs to clean his desk
[18:03:42] <__rob> anyone know if tormachs source code for path pilot is available ?
[18:03:55] <__rob> not on their website, but presumably they have to have put it out if lcnc is gpl ?
[18:04:54] <JT-Shop> you have to pay for a cd or go to jeffs web site, look on the forum for clues
[18:04:59] <malcom2073> They don't have to make it available to anyone except those who buy their software
[18:05:10] <__rob> well I have one of their machines
[18:05:17] <__rob> guess I can email them
[18:05:17] <malcom2073> Assuming their code is linked in such a way that it requires GPL compliance
[18:05:33] <malcom2073> Yeah I'd email them
[18:06:12] <__rob> also, their machine just boots straight into path pilot, is it possible (at least in linux cnc) to get a terminal up ?
[18:06:21] <__rob> I really want the machine to be on the network
[18:06:47] <JT-Shop> doubt it, why path pilot?
[18:07:00] <__rob> thats what they shipped with the mill
[18:07:20] <JT-Shop> ah ok that makes sense then
[18:07:25] <__rob> their mod of linuxcnc, so supports everything out of the box
[18:07:42] <__rob> but I dont want to be messing around with usb keys to transfer across toolpaths
[18:08:57] <spooq> pathpilot looks pretty good
[18:09:22] <JT-Shop> looks pretty limiting to me
[18:09:28] <__rob> the layout is pretty decent
[18:09:35] <__rob> far nicer than mach3 ever was
[18:09:49] <spooq> I watched their video, it solves a bunch of crap with mach3
[18:10:02] * JT-Shop never could get mark3 to work
[18:10:04] <spooq> doesnt really matter if its limiting if it does what oyu want
[18:10:07] <spooq> I use it
[18:10:12] <__rob> think thats why they switch.. alof of their old "how to" pdfs for mach3 have warnings all over the place
[18:10:13] <spooq> its... meh
[18:10:40] * JT-Shop goes inside to start some chow
[18:10:53] <__rob> only thing I dont see is an easy way to add my own scripts for setup etc..
[18:10:56] <spooq> I use mach3, I mean
[18:11:10] <__rob> they have pre done ones for the digital probe for centre finding and boss centre
[18:11:25] <__rob> but I was assuming as it was linuxcnc it would be more easily customisable
[18:12:29] <PCW> I dont think they want you to customize, from a support standpoint
[18:12:43] <__rob> yea :/
[18:13:03] <spooq> they are generally ok with customising, just not, as you say, combined with using their support to look after whatever you do
[18:13:39] <spooq> fwiw, if you google for it, you can find the src
[18:13:42] <__rob> what I was hoping was to be able to have a script for orientation finding, so if you dont have a part in a squared up vice, and its not square on the table, you can just probe around it
[18:14:13] <cradek> you can do that all in gcode. just run the gcode program.
[18:14:16] <__rob> and find out how much its rotated by and have the toolpath rotated by that
[18:14:29] <__rob> yes, but that means i have to remake the toolpath ?
[18:14:32] <cradek> probing, math, rotating is all done in gcode
[18:14:39] <cradek> no, just rotate the coordinate system
[18:14:47] <__rob> yea, but then I have to manually probe and do all that
[18:14:48] <__rob> is what I mean
[18:14:57] <cradek> no, probing is done with gcode
[18:14:58] <__rob> I need to generate the rotation value
[18:15:05] <cradek> yes I understand
[18:15:21] <cradek> gcode and rotation the coordinate system are both done by running gcodes
[18:15:23] <__rob> so if its some random part, i would need to run an optimisation solve for the orientation
[18:15:25] <cradek> err
[18:15:33] <cradek> probing and rotating the coordinate system are both done by running gcodes
[18:15:35] <__rob> find the best fit of the tool path to the probe points
[18:15:45] <__rob> so I could litterally slap the part down, and click probe, then go
[18:16:04] <cradek> we're obviously talking past each other
[18:16:17] <malcom2073> How do you make the part square in the first place if your vise isn't square?
[18:16:19] <Jymmm> CaptHindsight: Actually no, it's the I2S (not I2C) bus modulating that NTSC signal, nothing at all todo with wifi actually.
[18:16:36] <__rob> malcom2073, I mean if you dont have a vise
[18:16:44] <__rob> if you had a vaccum table
[18:16:45] <cradek> malcom2073: imagine finishing some castings that have a flat surface you mount to the table
[18:16:57] <cradek> malcom2073: of course it's stupid to NOT square a vise, if your parts go in a vise
[18:16:58] <malcom2073> Ah, so I guess you can plop a piece down, square it up, then when you remove/replace, garuntee it's stright
[18:17:00] <__rob> ohh, different question
[18:17:18] <spooq> you don't really have to square a vice...
[18:17:26] <spooq> not if you are willing to probe it
[18:17:28] <cradek> sure but it's stupid not to
[18:17:30] <__rob> cradek, so is this possible, I understand the gcodes are there to set the work offsets
[18:17:47] <__rob> but to get that rotation I would need to be able to run aa python script or something
[18:17:58] <spooq> well its not a lot of effort, and comes with some benefits, but I have run programs perfectly well with the vice at like 10 degrees to square
[18:17:59] <cradek> no, you can do it in gcode
[18:18:03] <Jymmm> JT-Shop: cradek: That dressings... might add a lil honey or sugar. server over finely sliced cabbage, half a bunch of chopped cilantro, roasted peanuts, could add leftover chopped chicken.
[18:18:04] <spooq> just to see if I could, really
[18:18:42] <malcom2073> cradek: You can have an angle offset in gcode?
[18:19:04] <spooq> g68 I think
[18:19:16] <cradek> you can rotate the various work coordinate systems around Z
[18:19:23] <cradek> each one can have its own rotation
[18:19:31] <malcom2073> Hmm, that's useful
[18:19:44] <cradek> yes then you can mount several castings and align a coordinate system to each one
[18:20:44] <__rob> cradek, but the problem I mean is to get the actual value of the rotation and translation offset for the Gcode
[18:21:14] <cradek> do you mean you want to take the results of your probing and do some trig on them?
[18:21:18] <__rob> that requires solving an optimisation problem
[18:21:34] <__rob> well I would want to fit the outside edge of the toolpath to the probe points, yea
[18:21:36] <cradek> heh, I think it only requires very simple trig
[18:21:56] <cradek> like probe two points that should be aligned (or at a certain angle) and then rotate so they are
[18:21:59] <__rob> well, only if you have already matching sets of points
[18:22:01] <__rob> but you dont
[18:22:08] <__rob> you have random locations on a part in the machine
[18:22:18] <spooq> I dont think you want to change the gcode to match your probe points
[18:22:18] <cradek> then I don't understand your task
[18:22:27] <spooq> like, after the probe
[18:22:41] <cradek> no, you want to align the coordinate system to the part and then run the unmodified gcode
[18:22:48] <__rob> the task is flip a part on a vaccum table to mill the other side. dont have to worry to much about alignment or jigs
[18:23:01] <__rob> have the probe probe the part to find its orientation, and any translation offset
[18:23:05] <spooq> so you dont have to align any features on either side?
[18:23:13] <__rob> then set the work offset (including rotation)
[18:23:17] <spooq> are there any features?
[18:23:18] <__rob> start milling other side
[18:23:21] <cradek> the part has to be nearly in the right place to probe it, right? otherwise how does the probe find it?
[18:23:45] <__rob> well, thats it, if you come in from each side, it wouldn't
[18:24:00] <__rob> but yea, I would flip and palce roughly the same place
[18:24:22] <cradek> yeah, if it's aligned against a linear feature, you can sure find it automatically by probing parallel to the feature
[18:24:34] <__rob> yup, but I would want it genertic
[18:24:48] <cradek> there's no generic solution :-)
[18:24:50] <__rob> obviously if you have 2 edges then its some trig
[18:24:57] <__rob> well there is
[18:25:33] <cradek> how you can usefully probe depends on the shape of the part and how it's (approximately) on the table
[18:25:52] <cradek> how you can go from those probe results to offset/rotation depends on what the points are
[18:26:01] <enleth> cradek: you could hook up a kinect or something...
[18:26:03] <spooq> you only need to probe 1 edge to get rotation, if its parallel to the x
[18:26:18] <__rob> enleth, I thought something like that would be nice
[18:26:24] <__rob> but I dont know ifthe resolution would be good enough
[18:26:25] <cradek> the answer is "it depends" but I guarantee you can do it in gcode :-)
[18:26:26] <enleth> like, determine that there's *something* on the table roughly *there*
[18:26:27] <__rob> to get decent alignment
[18:26:33] <enleth> and start probing in that direction
[18:26:39] <spooq> whats good enough in this case?
[18:26:41] <__rob> the digital probe is pretty accurate
[18:27:01] <cradek> my renishaw repeats extremely well - a few microns
[18:27:03] <__rob> so with enough probe points there would (depending on the part ) be a unique solution
[18:27:15] <cradek> the probe is better than the machine I bet
[18:27:18] <enleth> __rob: it's good enough to locate an object within a couple mm
[18:27:36] <enleth> assuming a distance of ~30cm
[18:27:39] <__rob> yea, guess that would be good to feed to the probe
[18:27:54] <__rob> so it probes in the right place
[18:28:00] <cradek> schmutz on the part will affect your results more than your probe's accuracy will
[18:28:14] <__rob> so are all the plugin scripts for linux cnc written in C or python or something ?
[18:28:14] <enleth> good point
[18:28:28] <__rob> like the probe routines from tormach for their digital probe
[18:28:55] <__rob> I mean, they talk to the ATC over a usb cable, so it must be written in a real language
[18:29:00] <cradek> I bet they're written in gcode
[18:29:11] <cradek> (probing routines)
[18:29:21] <cradek> the tool changer must have a hal driver
[18:29:26] <cradek> hal drivers are written in C
[18:29:43] <spooq> mach3 scripts are vbscript I believe
[18:29:45] <cradek> well realtime hal drivers are written in C. userland hal components can be just about anything.
[18:29:45] <__rob> so they will have used custom gcodes ?
[18:29:49] <__rob> to talk to it
[18:29:53] <__rob> via their driver ?
[18:29:57] <cradek> to talk to what?
[18:30:01] <__rob> the ATC for instance
[18:30:11] <__rob> tell it to shuffle next
[18:30:22] <__rob> the actual board is hooked up over USB to the control machine
[18:30:26] <__rob> the ATC board that is
[18:30:33] <cradek> the tool interface of linuxcnc is a few hal pins for 'load this tool' and 'prep this tool'
[18:30:43] <spooq> its normally parallel
[18:30:45] <cradek> tool changes are standard gcodes
[18:30:56] <spooq> Im really surprised they went with usb for anything
[18:31:05] <__rob> well its only for ATC
[18:31:13] <__rob> doesn't need to be realtime
[18:31:36] <__rob> I assumed they would have just used a com usb driver and a python sript or similar to talk to iy
[18:31:38] <__rob> it
[18:31:43] <__rob> bit hard to say without hte source code
[18:31:43] <cradek> yeah, if it's usb it's not realtime, but for slow stuff like talking to humans and tool changers, that's fine
[18:32:23] <spooq> its probably serial over usb
[18:32:29] <__rob> yea
[18:32:59] <__rob> cant see how to get a terminal up to look around..
[18:33:52] <__rob> http://www.tormach.com/blog/wp-content/uploads/2015/03/atc_screen.png
[18:33:58] <spooq> lsusb, check the /dev/tty files, etc
[18:33:59] <__rob> this is what you get for the atc
[18:34:12] <__rob> I would if I could get a terminal :D
[18:34:30] <__rob> its similar under "probe" for probe routings
[18:34:35] <__rob> but theres only like 3 pre done ones
[18:34:45] <__rob> this is what I am interested in finding out how its done so I can add my own
[18:35:12] <spooq> you tried port scanning it?
[18:35:21] <cradek> if you're asking about adding gui stuff to the pathpilot gui, nobody here can help you - probably nobody but tormach will answer those questions
[18:35:27] <spooq> maybe they left sshd open or telnetd?
[18:35:37] <__rob> yea, thats a good point
[18:35:43] <__rob> I'll take a look
[18:36:11] <__rob> ohh - assumed the gui was just their own layout of standard linux cnc with a few scripts for their tools
[18:36:16] <enleth> ah, the irony of Erickson QC tooling - there are no automated tool changers that support EQC, as far as I can see
[18:36:17] <__rob> maybe they've done more then I thought
[18:36:18] <spooq> failing that, pull the drive, mount it somewhere else, see whats on it
[18:37:00] <cradek> __rob: they did a lot, including discarding the packaging work and making their own system of updates, and they locked in to a certain OS release and a certain monitor size
[18:37:26] <cradek> so ... it's not a whole lot like linuxcnc anymore, although deep down the brains are the same
[18:37:39] <spooq> apparently shift-alt on boot goes to gnome
[18:37:54] <__rob> right, thats good to know. To be honest most of my problems can be solved by just re processing the toolpath
[18:38:07] <__rob> if I can get the probe values exported then that would be an easier route
[18:38:47] <__rob> or just make a jig :)
[18:45:22] <spooq> try ctrl-alt-x also, gives you a prompt
[18:51:51] <__rob> cool, will give that a go
[19:24:50] <PetefromTn_> hi folks
[19:40:27] <malcom2073> Hi
[20:18:45] <trentster> howdy all
[20:19:53] <trentster> Anyone have a good collection of LCNC code routines to share e.g. center finding and edge finding probing?
[20:24:59] <Jymmm> trentster: Look at the wiki
http://wiki.linuxcnc.org/cgi-bin/wiki.pl, also here are some gcode generators as example
http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Simple_LinuxCNC_G-Code_Generators
[20:25:36] <cradek> trentster:
http://timeguy.com/cradek/01262579508
[20:27:25] <Jymmm> cradek: See how you are, keeping it all to yourself ;)
[20:41:31] <trentster> cradek: thanks a ton
[20:41:43] <trentster> Jymmm: danke her
[20:58:29] <Jymmm> trentster: bete
[21:01:09] <trentster> Jymmm: I think you mean "Bitte"
[21:01:16] <Jymmm> yeh, that ;)
[21:01:31] <trentster> :-)
[23:09:15] <SpeedEvil> http://cen.acs.org/articles/94/web/2016/01/Strong-Stretchable-Carbon-Nanotube-Films.html - holy fuck
[23:09:21] <SpeedEvil> that sounds 'easy' to make.
[23:09:24] <SpeedEvil> CVD
[23:09:45] <SpeedEvil> Using a stream of nitrogen gas, they injected ethanol, with a small amount of ferrocene and thiophene added as catalysts, into a 50-mm-wide horizontal tube placed in furnace at 1,150–1,130 °C. A hollow cylinder with walls made of aligned carbon nanotubes forms in the furnace and emerges from the other end of the tube, driven by the nitrogen. As the tube emerges, the researchers wind the floating carbon nanotube cylinder onto a rotating drum.
[23:09:53] <SpeedEvil> Modulo 1150C of course
[23:13:38] <Jymmm> eh, blow it out your nitrogen tube!
[23:14:24] <Ralith> how long are the resultant nanotubes?
[23:15:58] <Jymmm> 9.6 gigapascal = 200500168.638 pound-force/foot²
[23:16:55] <SpeedEvil> Ralith: dunno - but 9.6GPa is fucking excellent for what seems like an initial cut of the technique.
[23:17:21] <Ralith> that is pretty impressive, yeah
[23:17:31] <SpeedEvil> Double that, and you're almost at the point you can get a space elevator. (triple is better)
[23:17:43] <Ralith> sounds like it might actually be commercially useful
[23:18:37] <Jymmm> Kinda funny, I was reading last week about the lady who invented Kevlar =)
[23:18:42] <Ralith> wonder how much it weighs
[23:18:53] <Ralith> would be interesting to explore applications to e.g. light aircraft skin
[23:19:36] <SpeedEvil> Ralith: The fundamental theoretical strength is 230GPa, and about 1300kg/m^2.
[23:19:55] <SpeedEvil> Ralith: but - that's for closely packed ideal single-wall nanotubes
[23:20:14] <SpeedEvil> The density will probably be a bit lower as it's badly packed.
[23:20:28] <Ralith> did you mean kg/m^3?
[23:20:32] <SpeedEvil> yes
[23:20:32] <Ralith> or, no, it's a film
[23:20:35] <Jymmm> Kevlar is 50 years old
[23:20:37] <Ralith> oh
[23:20:41] <Ralith> well then I wonder how thick it is!
[23:21:04] <SpeedEvil> Strength is of course one thing.
[23:21:06] <Jymmm> Ralith: the video shows a filament
[23:21:11] <Jymmm> https://en.wikipedia.org/wiki/Stephanie_Kwolek
[23:21:28] <SpeedEvil> I ahven't checked out the paper - but it may be lots stiffer than the polymer 'strings'
[23:21:41] <SpeedEvil> the polymers are great, but they have an extension of several percent
[23:22:04] <Jymmm> looks like goo
http://cen.acs.org/content/cen/articles/94/web/2016/01/Strong-Stretchable-Carbon-Nanotube-Films/_jcr_content/articlebody/subpar/articlemedia_0.img.jpg/1454095819163.jpg
[23:24:06] * Jymmm waits for the nanotube printing fad to begin
[23:24:54] <SpeedEvil> It looks almost like it's felt-like
[23:25:16] <SpeedEvil> I do wonder what happens if you roll two layers together at right angles
[23:25:18] <Jymmm> carbon paper (if you rememebr that)
[23:25:23] <SpeedEvil> yeah
[23:25:40] <Jymmm> or nori (on the left)
[23:26:13] <Jymmm> SpeedEvil: tyvek
[23:26:29] <SpeedEvil> yeah - I mean what's the interlayer bond strength like
[23:28:15] <Jymmm> how you looked at tyvek close up and in a bright light?
[23:28:18] <Jymmm> have*
[23:28:36] <SpeedEvil> yes, I know what it's like
[23:28:49] <SpeedEvil> I mean without binders
[23:28:52] <Jymmm> it's barly a fabric (mesh)
[23:28:57] <SpeedEvil> Felt doesn't have binders
[23:29:05] <Jymmm> here's a microscopic view
http://img.edilportale.com/products/prodotti-7832-reld32e5d1f8ccc4797a529b9327700f9cf.jpg
[23:29:27] <SpeedEvil> tyvek is melted partially though
[23:30:02] <Jymmm> It's really HDPE
[23:30:36] <Jymmm> I'm sure there is some carbon glue of sorts, graphite?
[23:30:59] <SpeedEvil> yes, you can do that, but if you could make it work like felt, that would be enormously better in some ways
[23:31:10] <Jymmm> and at 2000F, they could make some sort of fabric, or even eae one
[23:31:22] <Jymmm> weave
[23:31:23] <SpeedEvil> Felt is with a largely mechanical process interweaving the fibres
[23:31:36] <SpeedEvil> it has no binder, just wool
[23:32:38] <Jymmm> oh that's pretty for kevlar fabric
http://g03.s.alicdn.com/kf/HTB1ER7hHXXXXXcxaFXXq6xXFXXXT/325g-Jacquard-Black-Basalt-Colorize-Kevlar-fabrics.jpg
[23:33:03] <Jymmm> I saw just make it into threads and weave it into a fabric
[23:33:06] <Jymmm> say*
[23:33:18] <SpeedEvil> yeah - that's the other option
[23:33:46] <SpeedEvil> weaves have very different properties from a hypothetical 2d 'felt' though
[23:34:09] <SpeedEvil> The other fun issue is inhaled nanotubes
[23:34:30] <Jymmm> cant be much different than fiberglass
[23:35:02] <SpeedEvil> No
[23:35:39] <Jymmm> http://www.aliexpress.com/store/product/SUPER-STRONG-Carbon-3K-Aramid-Fiber-Hybrid-Fabric-Carbon-Kevlar-1500d-W-pattern-Cloth-190g-m2/324101_1224275045.html
[23:36:53] <Jymmm> I guess the stuff is *SO* available, they are just making fabric out of it for car bodies/custom helmets
[23:37:56] <Jymmm> pick a color, any color...
http://img.photobucket.com/albums/v687/wolfness-xi/100_3983.jpg
[23:38:16] <SpeedEvil> Being very careful as sometimes when you see 'carbon' it's actually dyed kevlar
[23:39:03] <Jymmm> But still, kevlar fabric that cheap for what it is?
[23:40:01] <SpeedEvil> you typically need several layers and stuff to make a finished product.
[23:40:35] <Jymmm> Hell, if it wasn't for the planned obsoelsnse of everythign these days, they could make a car bdy out of it, and nobody would die in an accident
[23:41:14] <Jymmm> "Yeah, just hit it with the hair dryer and it'll be good as new" lol
[23:41:54] <SpeedEvil> https://en.wikipedia.org/wiki/Ruppert_Archaeopteryx
[23:42:13] <SpeedEvil> This is a CF glider, which costs as much as silver.
[23:42:18] <SpeedEvil> $1/g or so
[23:43:21] <Jymmm> eeeesh
[23:43:23] <SpeedEvil> Once you put kevlar into a matrix - it snaps in crashes
[23:43:45] <Jymmm> nice
