Back
[00:02:42] -!- Tecan has quit [Quit: Live Long And Phosphor!]
[00:09:16] -!- phantoneD has quit []
[00:20:20] -!- kjn has quit [Quit: Quit]
[00:26:24] -!- asdfasd has quit [Ping timeout: 256 seconds]
[00:27:28] -!- GammaX-Shop has quit [Ping timeout: 252 seconds]
[00:45:34] Bojangle1 is now known as Bojangles
[01:07:00] -!- tmcw has quit [Read error: Connection reset by peer]
[01:07:28] -!- andypugh has quit [Quit: andypugh]
[01:14:51] -!- ravenlock has quit [Ping timeout: 260 seconds]
[01:15:26] -!- hashfail has quit [Ping timeout: 240 seconds]
[01:36:35] -!- tmcw_ has quit [Remote host closed the connection]
[01:44:12] -!- stsydow has quit [Remote host closed the connection]
[01:46:23] -!- sumpfralle has quit [Read error: Operation timed out]
[02:25:37] -!- tjb1 has quit [Quit: tjb1]
[02:27:05] -!- tmcw has quit [Remote host closed the connection]
[02:29:44] -!- PetefromTn has quit [Client Quit]
[02:30:28] -!- BJfreeman has quit [Quit: had a good time]
[02:33:58] -!- fiesh has quit [Ping timeout: 252 seconds]
[02:41:47] -!- FinboySlick has quit [Quit: Leaving.]
[02:58:26] -!- ravenlock has quit [Ping timeout: 252 seconds]
[03:02:53] -!- Keknom has quit [Quit: Leaving.]
[03:05:37] -!- tmcw has quit [Ping timeout: 248 seconds]
[03:31:23] -!- ve7it has quit [Remote host closed the connection]
[03:37:37] -!- tmcw has quit [Ping timeout: 248 seconds]
[04:01:08] -!- Kyusuke has quit []
[04:05:41] -!- tjb1 has quit [Ping timeout: 240 seconds]
[04:30:57] -!- jerryitt has quit [Client Quit]
[04:37:57] -!- tmcw has quit [Ping timeout: 248 seconds]
[04:49:10] -!- Loetmichel has quit [Ping timeout: 276 seconds]
[05:01:53] -!- Fox_Muldr has quit [Ping timeout: 248 seconds]
[05:16:36] -!- mhaberler [mhaberler!~mhaberler@macbook.stiwoll.mah.priv.at] has joined #linuxcnc-devel
[05:19:37] -!- Tecan has quit [Changing host]
[05:38:09] -!- tmcw has quit [Ping timeout: 248 seconds]
[06:09:05] -!- tmcw has quit [Ping timeout: 248 seconds]
[06:18:13] -!- apexskier has quit [Ping timeout: 250 seconds]
[06:28:57] -!- jfire has quit [Quit: Leaving.]
[06:39:33] -!- tmcw has quit [Ping timeout: 248 seconds]
[06:44:20] -!- a-l-e has quit [Ping timeout: 260 seconds]
[06:47:48] Aero-Tec2 is now known as Aero-Tec
[07:14:57] -!- kmrhb has quit [Ping timeout: 240 seconds]
[07:21:05] -!- jpk has quit [Ping timeout: 248 seconds]
[07:27:06] -!- jst has quit [Read error: Operation timed out]
[07:27:20] -!- psha[work] [psha[work]!~psha@195.135.238.205] has joined #linuxcnc-devel
[07:29:05] -!- rob_h [rob_h!~rob_h@94.8.128.28] has joined #linuxcnc-devel
[07:39:39] -!- ler_hydra has quit [Remote host closed the connection]
[08:11:49] -!- tmcw has quit [Ping timeout: 248 seconds]
[08:42:45] -!- tmcw has quit [Ping timeout: 248 seconds]
[08:45:21] Cylly is now known as Loetmichel
[08:48:22] -!- toner has quit [Ping timeout: 276 seconds]
[08:50:16] whoelse is now known as Guest90608
[09:37:23] -!- oterral has quit [Read error: Connection reset by peer]
[09:53:58] -!- Guest90608 has quit [Quit: Page closed]
[09:58:25] -!- gammax-Laptop1 has quit [Ping timeout: 248 seconds]
[10:00:18] -!- rob_h has quit [Ping timeout: 252 seconds]
[10:16:13] -!- mhaberler has quit [Quit: mhaberler]
[10:30:25] -!- largecheesepuff has quit [Ping timeout: 248 seconds]
[10:41:43] -!- Simooon has quit [Quit: Leaving]
[10:42:55] -!- WrinkledCheese has quit [Read error: No route to host]
[10:46:15] -!- mrsun has quit [Quit: Leaving]
[10:48:30] -!- stsydow has quit [Remote host closed the connection]
[10:49:35] -!- nikola_ has quit [Quit: Page closed]
[10:52:10] -!- sumpfralle has quit [Read error: Connection reset by peer]
[11:16:21] -!- tmcw has quit [Ping timeout: 248 seconds]
[11:17:13] -!- mhaberler [mhaberler!~mhaberler@macbook.stiwoll.mah.priv.at] has joined #linuxcnc-devel
[11:22:55] -!- sumpfralle1 has quit [Ping timeout: 260 seconds]
[11:33:39] -!- dhoovie has quit [Read error: Connection reset by peer]
[11:34:20] -!- odogono has quit [Ping timeout: 252 seconds]
[11:43:57] -!- LeelooMinai has quit [Read error: Connection reset by peer]
[11:46:45] -!- tmcw has quit [Ping timeout: 248 seconds]
[11:56:59] -!- sumpfralle has quit [Ping timeout: 260 seconds]
[11:57:38] -!- phantoxe has quit [Remote host closed the connection]
[11:59:19] -!- fomox has quit [Ping timeout: 264 seconds]
[12:09:13] -!- Simooon has quit [Quit: Leaving]
[12:15:19] -!- andy1978 has quit [Ping timeout: 252 seconds]
[12:17:41] -!- tmcw has quit [Ping timeout: 248 seconds]
[12:28:14] -!- LeelooMinai has quit [Read error: Connection reset by peer]
[12:33:00] -!- DJ9DJ has quit [Quit: bye]
[12:34:34] -!- LeelooMinai has quit [Ping timeout: 252 seconds]
[12:43:13] -!- psha[work] has quit [Quit: Lost terminal]
[12:48:21] -!- mhaberler has quit [Quit: mhaberler]
[12:48:37] -!- tmcw has quit [Ping timeout: 248 seconds]
[13:00:39] -!- andreas_ has quit [Client Quit]
[13:00:44] -!- andreas has quit [Client Quit]
[13:12:42] -!- Kup has quit [Read error: Connection reset by peer]
[13:19:12] -!- tmcw has quit [Ping timeout: 256 seconds]
[13:30:09] -!- shurshur has quit [Ping timeout: 248 seconds]
[13:32:40] -!- LeelooMinai has quit [Ping timeout: 257 seconds]
[13:39:43] -!- skroon_ has quit [Remote host closed the connection]
[13:42:07] -!- stsydow has quit [Remote host closed the connection]
[13:54:35] -!- LeelooMinai_ has quit [Ping timeout: 252 seconds]
[13:54:45] -!- nikola_ has quit [Quit: Page closed]
[13:58:07] -!- tmcw has quit [Remote host closed the connection]
[13:58:18] -!- ravenlock has quit [Ping timeout: 256 seconds]
[14:06:31] -!- LeelooMinai has quit [Ping timeout: 264 seconds]
[14:10:41] -!- c-bob|| has quit [Ping timeout: 248 seconds]
[14:13:35] -!- c-bob|| has quit [Read error: Operation timed out]
[14:16:47] -!- stsydow has quit [Client Quit]
[14:18:06] -!- holst has quit [Quit: Leaving]
[14:21:20] -!- psha [psha!~psha@213.208.162.67] has joined #linuxcnc-devel
[14:25:50] -!- Tom_itx has quit [Ping timeout: 240 seconds]
[14:33:36] -!- syyl has quit [Quit: Leaving]
[14:37:10] _BJfreeman is now known as BJfreeman
[14:38:20] -!- LeelooMinai has quit [Ping timeout: 264 seconds]
[14:39:01] -!- krusty_ar has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
[14:41:57] -!- nikola_ has quit [Quit: Page closed]
[14:55:56] -!- LeelooMinai has quit [Ping timeout: 256 seconds]
[15:20:10] -!- dway has quit [Quit: NOOOOOOooooooooo……]
[15:29:19] -!- dway has quit [Quit: NOOOOOOooooooooo……]
[15:32:28] -!- ve7it [ve7it!~LawrenceG@S01060014d19d0b68.pk.shawcable.net] has joined #linuxcnc-devel
[15:37:27] -!- lando5446 has quit [Quit: Page closed]
[15:38:13] -!- Nick001-Shop has quit [Remote host closed the connection]
[15:38:15] -!- capricorn_1 has quit [Quit: Konversation terminated!]
[15:50:50] <cradek> seb_kuzminsky: if andy makes a new branch, I know the buildbot will build it - but are those packages accessible for people to easily use?
[15:52:27] <seb_kuzminsky> yes, they show up in the 'scratch' deb archive
[15:52:32] <seb_kuzminsky> hold on i'll find the docs for that
[15:53:08] -!- kwallace [kwallace!~kwallace@smb-96.sonnet.com] has joined #linuxcnc-devel
[15:54:05] <seb_kuzminsky> the buildbot docs are not very well organized, i'm no jthornton!
[15:54:08] <seb_kuzminsky> http://buildbot.linuxcnc.org/dev.html
[15:54:17] <seb_kuzminsky> look in the section called "Sorting of packages and other build results"
[15:54:59] <seb_kuzminsky> the user has to add a new apt-source, update, then select the version of the deb that corresponds to the new branch name
[15:55:31] <cradek> awesome
[15:59:19] -!- Bojangles has quit [Ping timeout: 276 seconds]
[16:05:14] -!- BJfreeman has quit [Ping timeout: 256 seconds]
[16:06:15] _BJfreeman is now known as BJfreeman
[16:14:21] -!- kostodo has quit [Quit: kostodo]
[16:23:38] -!- odogono has quit [Ping timeout: 252 seconds]
[16:25:30] -!- ktchk has quit [Read error: Connection reset by peer]
[16:45:50] -!- vladimirek [vladimirek!~vladimire@95.105.250.72] has joined #linuxcnc-devel
[16:50:21] -!- XiXora has quit [Ping timeout: 240 seconds]
[16:51:44] -!- morfic has quit [Ping timeout: 245 seconds]
[16:57:24] -!- exco has quit [Quit: Verlassend]
[17:05:08] -!- Bojangle1 has quit [Ping timeout: 256 seconds]
[17:11:26] <kwallace> I have an option to use TrueType-Tracer nearly as it is now, then prepend and append g-code for things like setting tool number, feed rate, x y z offsets and other parameters, or I can try to pass all these parameters to ttt.c, then add these features to ttt.c . Also, it would be nice to have a roughing and finish routine, and conventional or climb, and ...
[17:11:35] <kwallace> My main problem now is passing a large number of parameters to ttt.c using "-? xxx" for each. Anyone have an opinion, for an easy way to pass parameters to ttt.c, or using ttt.c as a basic path generator?
[17:12:37] -!- shurshur has quit [Remote host closed the connection]
[17:16:21] <seb_kuzminsky> kwallace: i think ttt has a 'subroutine' option, where it emits a sub instead of a whole gcode program
[17:16:34] <seb_kuzminsky> then you could have your own wrapper that sets up those things and calls the sub
[17:17:08] <cradek> and if it doesn't, that's a fine idea
[17:19:28] <kwallace> I didn't find that option, but it would be easy to add.
[17:19:30] -!- bpuk [bpuk!~ben@boopotter.plus.com] has joined #linuxcnc-devel
[17:20:19] <kwallace> But I'm running out of -? letters.
[17:21:46] <cradek> you can always use --long-words to work around that problem, but I agree with how you are questioning whether that's what you really want anyway
[17:23:14] <cradek> truetype scaling is a mystery to me -- other than guess-and-check I'm not sure how to get output exactly 1" tall for instance
[17:23:41] -!- andypugh [andypugh!~andy2@cpc16-basl9-2-0-cust685.20-1.cable.virginmedia.com] has joined #linuxcnc-devel
[17:24:20] <cradek> hey andypugh
[17:24:31] <cradek> thought you might be showing up...
[17:24:31] <andypugh> Hi
[17:24:35] <seb_kuzminsky> hi!
[17:24:44] <kwallace> Right now I have a note next to the UI scale entry saying "Adjust to suit".
[17:24:54] <cradek> haha yep
[17:25:00] <andypugh> Just got home from work.
[17:25:28] -!- shurshur has quit [Remote host closed the connection]
[17:25:36] -!- phantoxe has quit []
[17:26:43] <cradek> the thought of using spaghetti makes me smile
[17:27:54] <andypugh> Cheap, readily available and frangible. Otherwise it is really hard to tell if the shafts are out of phase.
[17:28:15] <kwallace> If it sticks, it's done.
[17:28:44] <cradek> kwallace: as a simulation of a very fragile gantry
[17:29:48] <andypugh> http://www.youtube.com/watch?v=_pejHjPJlcE
[17:33:12] <cradek> that's awesome
[17:33:15] <kwallace> That's pretty cool.
[17:35:38] <andypugh> Talking releases, have you seen mungkie's ready-to-go RPi image?
[17:37:18] <andypugh> http://soundproofingforum.co.uk/rpi_linuxcnc/raspberrypilinuxcnc.htm
[17:38:04] <andypugh> DMA-based stepgen included.
[17:38:35] <cradek> nope I hadn't seen that
[17:39:33] <cradek> do these systems use apt, or are whole disk images the standard way to distribute things?
[17:39:59] <cradek> (I know it's often wrong to think of them like PCs...)
[17:40:30] <andypugh> I have only ever seen whole-disk images (and, curses! I forgot to buy a micro-SD for the Beaglebone on the way home again).
[17:40:57] <seb_kuzminsky> there is a fork or branch or something of debian forrpi:
http://www.raspbian.org/
[17:41:19] -!- jfire has quit [Quit: Leaving.]
[17:41:59] <andypugh> referring to your email, the rapid-to-home is a real worry with gantry homing. I think that as it stands you would need to have the home position extremely close to the home-switch release point.
[17:42:38] <cradek> yes
[17:42:38] <andypugh> What you really want is each slaved axis to wait for the others to finish step N before moving to step N+1
[17:42:51] <cradek> also yes (that's what my mail was about!)
[17:43:10] <cradek> any other solution is only approximate
[17:43:27] <cradek> there are lots of approximate solutions that are possible to implement without touching motion
[17:44:24] <cradek> to me, the problem of initially moving in opposite directions seems quite possible too
[17:44:29] -!- Tecan has quit [Ping timeout: 252 seconds]
[17:44:30] <andypugh> I don't think that the already-on-the-switch side needs to back off. All it has to do is nothing, until the other axes are on their switches.
[17:45:05] <cradek> if you don't back off you aren't sure you'll get the correct index (the correct one is the first one after the switch trips)
[17:45:48] <andypugh> Ah, yes.
[17:45:51] <cradek> I think you must back off because it's the switch's rising edge that's important
[17:46:08] <cradek> and I guess that's true whether or not you use index
[17:46:31] <andypugh> This breaks my "panic if states don't change almost simultaneously" monitor, though.
[17:47:02] <cradek> yeah I am not sure how you could successfully watch for things going wrong
[17:47:25] <cradek> if it were my hardware, I'd want a mechanical strain estop of some kind
[17:47:40] <cradek> servo systems sure like to runaway in various failure modes
[17:48:35] <pcw_home> Torque mode limit during homing?
[17:50:26] <pcw_home> Which reminds me it would be nice to be able to do the home to stop trick
[17:50:28] <pcw_home> but I think that mean you need to disable ferror or change the ferror limits dynamically
[17:50:39] <andypugh> cradek: I guess the test is for "panic if states are divergent for more than a cerain length of time"
[17:51:00] <pcw_home> not distance?
[17:51:14] <cradek> I dunno, that assumes a lot
[17:52:08] <cradek> as far as I can tell, the ONLY thing you know is that the original divergence, whatever it is, didn't break the machine
[17:52:34] <cradek> during homing the divergence will change, but you can't know whether you are improving it or making it worse
[17:53:07] <cradek> I don't think you can sanity-check much of anything
[17:53:14] <pcw_home> right, without absolute encoders you dont know until both indexes have occured
[17:54:24] <andypugh> I meant if the homing states are not converging. I don''t know the numerical values, but if you have one in state 3 and one in state 5, then it is OK if the 3 becomes a 4 or the 4 becomes a 5, but 2 and 6 are right-out.
[17:55:49] <andypugh> (Please insert numbers that make sense in the abov sentence)
[17:57:26] <cradek> but if they're in lockstep (all joints must be satisfied at the end of state N in order to move to state N+1) you get this automatically
[17:59:05] <andypugh> I was trying to cover for a broken home switch, but I guess in that situation they both carry on merrily on to the hard stop.
[17:59:18] <pcw_home> this is like the simple trick of anding the home switches
[17:59:33] <cradek> well one would stop and the other would keep going forever
[17:59:46] <cradek> same with a missing index
[17:59:48] <andypugh> "Yes, I have seen the home switch, but I am going to continue the search move until you see yours...."
[18:00:12] <cradek> no, the one that finds it has to stop
[18:00:16] <pcw_home> well if you get one you could limit the search for all
[18:00:33] <cradek> stopping will make the alignment better
[18:00:48] <cradek> and if it keeps going, it might skip over the desired index
[18:01:36] <andypugh> Ah, so, you exit "searching" but don't enter "latching"? That does make more sense
[18:02:28] <pcw_home> That does look like the best (and works with just switches as well)
[18:02:39] <cradek> there are so many states - often many are skipped - we have to figure out how to handle each transition. I don't claim to have thought it all through.
[18:03:45] <cradek> there are ~25 states
[18:05:19] <andypugh> I thought you had 52? :-)
[18:05:56] <andypugh> (checks Wikipedia) Ah, only 50.
[18:07:24] <cradek> sure but about half of them are crappy
[18:17:24] -!- tanepiper has quit [Remote host closed the connection]
[18:18:42] Guest21469 is now known as davec_
[18:20:01] * KimK heard one of our leaders say we have 57 states
[18:20:39] <pcw_home> The United States of Heinz?
[18:21:06] <KimK> It must be
[18:24:53] <cradek> eh snopes has the actual clip
[18:26:51] <bpuk> Evenin' all
[18:27:10] <bpuk> Daft question time: Is there any time a Type 1 G71 is more desirable than a Type 2?
[18:27:34] <cradek> I have no idea what you're talking about
[18:27:43] <bpuk> Sorry, lathe roughig cycles
[18:28:16] <cradek> we don't have a G71 roughing cycle
[18:29:39] <bpuk> My brain is apparently fried: Let me start over - Some time ago I started work on implementing one - which had some issues. A remapped version was suggested as a reasonable compromise. I finally got time to look at it again
[18:30:05] <cradek> ah ok.
[18:31:17] <bpuk> When looking at it this time, something clicked and I realised Type 2 is actually not that hard. A couple of issues with cutter compensation, but nothing too major. At which point I started wondering if it was worth having type 1 at all if type 2 is available
[18:31:41] <bpuk> Type 1 - No pockets, retracts at 45 degrees. Type 2 - Pockets allowed (but no undercuts), follows profile out of cut before moving to next cut - so no stairstep
[18:32:32] <cradek> aha, now I have enough information to answer - and my answer is I have no idea :-)
[18:33:30] <cradek> does pocket mean X decreases during a cut?
[18:34:12] <bpuk> Yes -
http://www.cnccookbook.com/img/GWizardE/GCodeCourse/G71Pockets.jpg is an example (not mine)
[18:34:38] <cradek> ooh that's complex
[18:35:01] <cradek> you need to use the tool's backangle to know how to not gouge
[18:36:03] <bpuk> yes. The other major part is knowing how deep it's possible to cut. That part is so machine dependant it's difficult to come up with a one size solution
[18:36:14] <cradek> I'd love to have these cycles and I'd dread having to implement them...
[18:37:04] <cradek> it's not clear to me that you can do them in remap land
[18:37:20] <bpuk> Well, I had type 1 working fairly well in C space - it's mostly ported over now. I'm concentrating on getting the profile support solid at the moment.
[18:37:23] <cradek> you have to specify a series of lines that don't do the usual thing but instead define the outline
[18:37:58] <cradek> do you key off N number ranges or something to denote those special lines that belong to the cycle?
[18:39:15] <bpuk> that was the 'issues' last time.
http://permalink.gmane.org/gmane.linux.distributions.emc.user/39158 was the post clarifying how to do it
[18:39:38] <bpuk> can't key off N ranges - instead define a remapped G-code with a P word to define a profile ID
[18:40:12] <cradek> ah I see
[18:40:39] <cradek> you don't care about being compatible with something - you'd rather be able to do it in terms of remap
[18:40:42] <bpuk> the advantage of doing it this way is it gives arbitrary profile support - which I would *love* to get working in AX on the lathe (andy's eccentric turning stuff)
[18:41:16] <cradek> my hat is off to you
[18:41:26] <bpuk> The original was almost compatible with fanuc controls - two word letters changed - but it broke the execution abstraction too badly (apparently)
[18:44:14] <cradek> if you do it in remap land it smells like you will have to reimplement cutter compensation
[18:44:29] -!- ler_hydra has quit [Remote host closed the connection]
[18:44:37] -!- IchGuckLive has quit [Quit: ChatZilla 0.9.87 [Firefox 20.0/20130329043827]]
[18:45:00] -!- L84Supper2 [L84Supper2!~TheLarch@211.155.168.68] has joined #linuxcnc-devel
[18:45:15] <cradek> I am not sure if your remapped move-generator can issue moves that then get compensated
[18:45:22] <bpuk> quite possibly - that's the one part I haven't looked at yet. In theory it can be broken down into a series of g0/g1/g2/g3 commands - which *should* have cutter comp enabled, but it's something to check
[18:45:28] <cradek> I don't know enough about this to picture how the whole thing would work
[18:45:30] -!- zlog has quit [Remote host closed the connection]
[18:45:36] -!- L84Supper2 has quit [Changing host]
[18:45:36] -!- L84Supper2 [L84Supper2!~TheLarch@unaffiliated/l84supper] has joined #linuxcnc-devel
[18:45:43] -!- Tom_itx has quit []
[18:46:20] <bpuk> The original mailing list thread is at:
http://comments.gmane.org/gmane.linux.distributions.emc.user/39104 - there were quite a variety of viewpoints there.
[18:47:07] -!- Willburrrr2003 has quit [Quit: Do fish get thirsty?]
[18:47:11] <bpuk> I suspect I'll get it almost complete - then ask mhaberler if he can point me in the right direction for the last few bits ;)
[18:48:44] -!- exco has quit [Quit: Leaving]
[18:49:39] -!- mhaberler [mhaberler!~mhaberler@macbook.stiwoll.mah.priv.at] has joined #linuxcnc-devel
[18:49:53] <andypugh> Cutter comp on lathes actually rarely makes a difference on profiles that matter.
[18:50:17] <cradek> andypugh: I'm, um, surprised to hear you say that
[18:50:42] <andypugh> Tapers end up a bit fat, and radii are a bit wrong.
[18:50:55] <cradek> I see mhaberler answered my question - I asked the same thing in that thread - he thinks it would be ok
[18:51:06] <andypugh> But the mechanically important diameters and shouldes come out right regardless.
[18:51:41] <andypugh> What I am saying is that if the G71 can't have cutter-comp then it is still well worth having.
[18:52:02] <cradek> I was wondering if that was your point...
[18:52:38] <cradek> seems like without ccomp you often cut not quite enough - in that case you could do a separate finish pass with comp to "fix" it -- sort of
[18:52:55] <cradek> but surely it would be better to choose a design where it can actually work right
[18:53:31] <bpuk> also, for roughing, where you're leaving an allowance, it matters less than usual - since the finish pass should clean it up correctly
[18:53:40] <bpuk> but if doable, yes, it should have it
[18:53:50] <cradek> well obviously it's doable
[18:54:03] <cradek> you just have to implement it in a way that allows it
[18:54:12] <andypugh> If the tool can reach the programmed profile then the finish path is perfectly normal and cutter comp should work.
[18:54:52] <andypugh> And I find it hard to imagine the roughing pass leaving too much on because of a lack of cutter comp.
[18:55:32] <andypugh> (Of course, my imagination has its limits)
[18:55:45] <cradek> if the finish pass should cut the same amount everywhere, you must compensate the roughing
[18:55:58] <cradek> if the finish pass DOESN'T do that it's not much of a finish pass
[18:56:21] <bpuk> pre-finish and finish? (kidding!)
[18:56:21] <cradek> for instance if it rubs on the straights and cuts on the arcs and tapers, it's going to be pretty inadequate
[18:56:54] <mhaberler> bpuk: the fallout of this thread was the example here:
http://git.mah.priv.at/gitweb?p=emc2-dev.git;a=shortlog;h=refs/heads/list-recorder-by-remap
[18:57:05] <mhaberler> I probably should add this to master
[18:57:16] <mhaberler> basically it's a gcode recording-and-replay scheme
[18:57:28] <bpuk> indeed, I used it as the base for the current code
[18:57:30] <cradek> can you replay with comp?
[18:57:43] <andypugh> I think you need to make a profile cut rather than a roughing cut to guarantee consistent finish cut thickness anyway?
[18:57:46] <mhaberler> you and your questions.. ;)
[18:58:33] <mhaberler> I would think that boils down to recording state and replaying properly too, but I dont see that as a major issue
[18:58:49] <cradek> andypugh: I'd think the last cut of the cycle would be the finish cut, and the cut before that would leave a constant allowance for it
[18:59:06] <mhaberler> basically it records a list of actions in a python list, and it should be straightforward to add a state observer/setter there
[18:59:09] <bpuk> This does bring back my earlier question though - a type 1 profile leaves a stairstep pattern making the finish inconsistent - which is why some implementations of G71 allow a fre-finish pass as part of the cycle
[18:59:13] <cradek> I suppose you'd want a different tool for the very last finish cut
[18:59:39] <cradek> ok we have a term for it: pre-finish (leave the part consistently oversized everywhere)
[18:59:56] <cradek> then you'd change tool and run the same profile once more for the real finish cut
[19:00:22] <bpuk> but with a type 2 you never get that stairstep - the profile is correct, but with a consistent allowance
[19:00:38] <bpuk> so only one finish pass is needed, no seperate pre finish
[19:01:00] <cradek> then I suspect controls have/had type1 because it was easier to implement, and later because the gcode using type1 existed
[19:02:02] <cradek> I bet type2 is useful in all cases where type1 was useful, plus more
[19:02:18] <andypugh> But Type 1 may be able to remove more material, faster.
[19:02:41] <cradek> yeah I suppose so
[19:02:57] <cradek> I suppose always following the profile out means a lot of rubbing
[19:03:23] <bpuk> and a lot slower than a retract - which in production obviously matters
[19:04:02] <andypugh> I can imagine that there are tools that chatter unless they are cutting a big bite too.
[19:04:16] <bpuk> RCMT?
[19:04:41] <andypugh> But I would say that folk woulf be grateful for anything you care to offer :-)
[19:05:24] <cradek> yeah for RCMT I bet you'd want type 1 only
[19:05:43] <bpuk> perhaps, but if there is a use case for type 1 - theres no reason not to add it. if there isn't then it makes writing gcode less confusing as you don't need to consider whether the profile needs to be type 1 or 2
[19:05:44] <cradek> you probably want a certain load, a straight cut, and nothing else
[19:06:37] <bpuk> I find RCMT generally works well for roughing (3-4mm DOC), but I always get chatter when finishing with one.
[19:06:49] <cradek> yeah I believe it
[19:07:05] <bpuk> I'm sure theres a feedrate that wouldn't chatter, but I just switch tools
[19:07:33] <cradek> you don't ever want to finish with the roughing tool anyway do you?
[19:08:17] <bpuk> depends on what the part is. My toolchanger is in need of replacement - best case toolchange is ~30 seconds. So yeah, I sometimes finish with the roughing tool
[19:09:41] <andypugh> I have never even dared look at using RCMT on my lathe.
[19:10:54] <andypugh> 0.4 DOC and 0.15 feed per rev is about right. Luckily it is at least CNC so I can play with computers or watch TV until it is done.
[19:11:15] <cradek> oh, mm :-)
[19:11:41] <seb_kuzminsky> mmm, mm
[19:14:28] <bpuk> mhaberler: I'm assuming numpy isn't available from remap on a default install?
[19:15:01] <mhaberler> it is available if you install numpy and import it where you need it
[19:15:36] <mhaberler> standard python
[19:15:36] <mhaberler> you plan to use it in embedded python, just import it in toplevel.py
[19:15:36] <bpuk> I don't want to add that as a requirement: so no :D I'll just do the few bits of math I need longhand
[19:16:08] <mhaberler> toplevel.py is specific for your config so it isnt a build or install prerequisite
[19:16:17] <mhaberler> you give a decent error message and be done with it
[19:16:25] -!- holgi has quit [Quit: ChatZilla 0.9.90 [Firefox 20.0/20130329043827]]
[19:16:25] <mhaberler> try:
[19:16:31] <mhaberler> import numpy
[19:16:37] <mhaberler> except ImportError:
[19:16:50] <mhaberler> inter.error(bitch as needed)
[19:17:08] <mhaberler> (need to look at the details for interp.error(), I just made that up)
[19:17:35] <bpuk> heh
[19:18:02] <bpuk> It's still an added requirement for anyone else to install the same code - only need about three functions from it - so easy enough to reimplement
[19:18:14] <mhaberler> all standard python rules apply; the only difference is: you dont fire it up by running 'python mycode.py'
[19:18:25] -!- KimK has quit [Ping timeout: 248 seconds]
[19:18:41] <mhaberler> actually I'm not sure if numpy isnt pulled in anyway
[19:18:46] <mhaberler> let me see
[19:19:02] <bpuk> and is there any way to debug except adding "self.set_errormsg(err)" all over the place :P
[19:21:32] <mhaberler> yes, that is outlined in the manual, including screenshots - a very plush way actuall
[19:21:48] <mhaberler> you just won in the package prerequisite lottery;
http://git.mah.priv.at/gitweb?p=emc2-dev.git;a=blob;f=debian/control.in;h=fa428e4235e89bad1be5da0532ec0443b2a6c46c;hb=a1baaafebe638f2d37a69590d279a6c1012bab5b#l48
[19:21:52] <bpuk> oooh, *reads manual*
[19:22:06] <bpuk> and yay!
[19:22:41] <mhaberler> note you can always gdb -p the process, gdb understands python
[19:22:59] <mhaberler> it isnt totally intuitive but possible
[19:23:38] <bpuk> if I can get that working this could start to go a bit faster :D
[19:24:37] <mhaberler> my suggestion would be to start with a config using python and the rs274 standalon interp, this is a bit more straighforward than when the full monty of linuxcnc processes, ui etc runs
[19:24:55] -!- jthornton has quit [Read error: Connection reset by peer]
[19:24:55] -!- JT-Shop has quit [Read error: Connection reset by peer]
[19:25:20] -!- JT-Shop [JT-Shop!~John@99.198.67.11] has joined #linuxcnc-devel
[19:25:22] -!- jthornton [jthornton!~john@99.198.67.11] has joined #linuxcnc-devel
[19:25:41] -!- jfire has quit [Quit: Leaving.]
[19:28:10] -!- shurshur has quit [Ping timeout: 252 seconds]
[19:30:44] <mhaberler> there might be easier ways to debug embedded python nowadays; at the time the Eclipse pydev plugin was the only game in town
[19:31:46] <bpuk> Compared to what I have been doing, Eclipse+pydev is a huge improvement
[19:32:03] -!- KimK [KimK!~Kim__@wsip-184-176-200-171.ks.ks.cox.net] has joined #linuxcnc-devel
[19:33:47] -!- Simooon has quit [Quit: Leaving]
[19:34:12] -!- fjay has quit [Remote host closed the connection]
[19:38:42] <mhaberler> yes, its by far the best Python debugging environment I found; I just find Eclipse a bit too clunky for everyday use
[19:45:20] -!- LeelooMinai has quit [Ping timeout: 240 seconds]
[19:49:57] -!- i_tarzan has quit [Ping timeout: 248 seconds]
[19:56:42] <andypugh> What do you prefer? (I am using Eclipse now, after XCode stopped really working for non-Mac apps
[19:59:23] <mhaberler> I've never tried Xcode except the hello-world iphone app so I cant tell; I use emacs
[20:03:56] -!- gammax-Laptop1 has quit [Ping timeout: 256 seconds]
[20:46:46] -!- vegeetz has quit [Quit: Leaving.]
[20:48:58] -!- tmcw has quit [Remote host closed the connection]
[20:55:10] -!- FinboySlick has quit [Remote host closed the connection]
[20:56:42] -!- toastywork has quit [Quit: Page closed]
[20:58:22] -!- nikola_ has quit [Quit: Page closed]
[21:00:53] s1dev is now known as s1dev|away
[21:01:58] <andypugh> cradek: Is there any penalty in all axes that share a homing sequence number waiting for each other, all the time, by default? I am not seeing one.
[21:02:27] <cradek> very interesting question
[21:02:44] <cradek> you do have to wait for homing to complete on the slowest joint
[21:03:14] <cradek> I bet there are some cases where it would be slower. consider one joint has a very long initial search, and another has a very long final search
[21:03:22] s1dev|away is now known as s1dev
[21:03:37] <cradek> your change would not let them happen simultaneously
[21:06:05] <andypugh> I guess a rotary with a full half-circle cam would be an eample of the slow final-search, if it started on the cam. Or would you configure that to have a fast second-search and slow final latch?
[21:06:12] <cradek> andypugh: are you thinking that change plus ja3+gentrivkins would give everything that slaved joints would? it seems like it might, and if so, that's gotta be the easy way
[21:07:16] <andypugh> <puts on Tyrolean hat> Ah, yes, but is it the _right_ way?
[21:07:27] -!- DJ9DJ has quit [Quit: bye]
[21:08:01] -!- tjb1 has quit [Ping timeout: 240 seconds]
[21:08:28] <cradek> ha, "right"
[21:08:40] <andypugh> I think it might be, gantrykins isn't horrible (the problems are non-triv kins related). Gentrivkins apparently works well. All that is left is a nagging doubt about unfotunate homing sequences.
[21:08:56] -!- bpuk has quit []
[21:09:07] <cradek> I think you're on to something
[21:09:19] <cradek> that might be a relatively minimal change
[21:09:45] -!- bpuk [bpuk!~ben@216.108.232.205] has joined #linuxcnc-devel
[21:11:56] -!- theorbtwo has quit [Ping timeout: 256 seconds]
[21:12:06] theorb is now known as theorbtwo
[21:12:22] <andypugh> The problem might be if simultaneous axes had different homing sequences. I guess in that case you would assume that they weren't intended to be slaves, and that there is no need to wait.
[21:12:49] <andypugh> You wouldn't have one side of a gantry with 3-step homing and the other with 2-step.
[21:13:13] <andypugh> (well, you would hope not).
[21:13:19] <cradek> not on purpose you wouldn't
[21:13:26] <mhaberler> this is a Styrian hat;
http://www.kleinezeitung.at/steiermark/voitsberg/516827/index.do# , and _this_ is a Tyrolean hat:
[21:13:30] <mhaberler> this is a Styrian hat;
http://www.kleinezeitung.at/steiermark/voitsberg/516827/index.do# , and _this_ is a Tyrolean hat:
[21:13:31] <mhaberler> this is a Styrian hat;
http://www.kleinezeitung.at/steiermark/voitsberg/516827/index.do# , and _this_ is a Tyrolean hat:
[21:13:50] <mhaberler> this is a Styrian hat;
http://www.kleinezeitung.at/steiermark/voitsberg/516827/index.do#
[21:13:53] <cradek> it may turn out that we want the user to still explicitly state when a joint is a slave of another. there's no need to guess
[21:13:59] <cradek> mhaberler: have you been drinking?
[21:14:22] <mhaberler> this is a tyrolean hat:
https://www.google.at/search?q=tirolerhut&client=firefox-a&hs=84f&rls=org.mozilla:en-US:official&tbm=isch&tbo=u&source=univ&sa=X&ei=1tefUezZNueI7AaIpYGICA&ved=0CEwQsAQ&biw=1493&bih=748#facrc=_&imgrc=ZnuvONfqBT4alM%3A%3B_tXmIHGefTAVWM%3Bhttp%253A%252F%252Fmembers.aon.at%252Ffomalhaut%252Fkeinsoein.jpg%3Bhttp%253A%252F%252Fmembers.aon.at%252Ffomalhaut%252Ftirolerhut.html%3B1220%3B880
[21:14:27] <mhaberler> no, andy bungled it
[21:15:02] <andypugh> The Styrian hat is rather less camp.
[21:15:58] <mhaberler> yes,Tyrol is a bit of an island if you will, so I'm not surprised about your preferences
[21:16:12] -!- psha has quit [Quit: Lost terminal]
[21:17:24] <andypugh> I am rather fond of the Tirol, it is lumpy and snowy in a very entertaining way.
[21:19:10] -!- chillly has quit [Quit: Leaving]
[21:19:41] <andypugh> Next thing you will be telling me that you won't be wearing lederhosen in Wichita.
[21:20:52] <mhaberler> nope, I am an asphalt creature, just living in the pampa
[21:21:34] -!- tjb1 has quit [Ping timeout: 256 seconds]
[21:21:34] tjb1_ is now known as tjb1
[21:22:32] <andypugh> I guess it is rather like folk assuming that I will be wearing a kilt.
[21:22:52] <cradek> currently, kilts and fedoras are the style in the US
[21:22:55] <mhaberler> we'd be delighted. No haggies please.
[21:22:57] <andypugh> (Same country, wrong region)
[21:23:25] <mhaberler> Arent those forbidden by the Geneva convention?
[21:23:39] <cradek> the US doesn't pay any attention to that kind of thing
[21:23:46] <andypugh> cradek: Have you considered that you are getting that impression soley because you hang out with the Linuxistas?
[21:24:35] <mhaberler> I honestly hope the DOA beagles will do their job when Andy immigrates
[21:24:51] <cradek> are you implying that my avoidance of popular culture means I'm out of touch with the current fashion trends?
[21:24:58] <andypugh> As for Haggis, it is supposedly illegal to make it in the US, but it may be imported as a regional delicacy. Lungs are not considered meat in the US, so may not be added to food.
[21:25:36] <cradek> other things that strangely aren't considered meat: dogs, cats, people
[21:25:39] <mhaberler> we're doomed
[21:25:46] <cradek> horses are right on the edge
[21:26:19] -!- odogono has quit [Quit: odogono]
[21:26:36] <andypugh> A lot if British folk started eating horse meat recently. They didn't acually intend to.
[21:26:37] -!- tjb1 has quit [Quit: tjb1]
[21:26:45] -!- vladimirek has quit [Remote host closed the connection]
[21:27:08] <bpuk> not exactly recently
[21:27:31] <cradek> meat is meat. people only have strong feelings and revulsion about it because they're made of it
[21:27:44] -!- mrsun has quit [Quit: Leaving]
[21:28:47] <andypugh> bpuk: Current affairs is not my strong point, but it was this year, I thought?
[21:29:07] <mhaberler> regional, ok, but how can it pass as a 'delicacy'? beats me.
[21:29:16] <andypugh> It's lovely.
[21:29:33] <bpuk> it was discovered this year, been going on a lot longer than that
[21:29:39] <andypugh> (Though the only times I have ahd it, it was served to me by an American)
[21:29:41] <mhaberler> and what if we bring you in by teleconference?
[21:30:12] <bpuk> nothing wrong with horse, so long as you know it's horse (and hasn't been pumped full of fat-soluble painkillers)
[21:30:41] <andypugh> I read that as "rat-soluble". Which I guess is also true.
[21:31:05] <andypugh> And yes, the horse I ate was very tasty.
[21:36:34] -!- vegeetz has quit [Quit: Leaving.]
[21:51:14] <PCW> the whole thing?
[21:53:28] <mhaberler> he is from the UK
[21:53:43] <PCW> Ahh
[21:53:57] <mhaberler> hippophagues
[21:54:57] <mhaberler> the fifth column of the Royal Raspberry Society
[21:59:19] -!- tjb1 has quit [Ping timeout: 264 seconds]
[22:01:13] <bpuk> a full horse in one sitting might be a bit much
[22:01:19] <bpuk> but only just
[22:18:24] <mhaberler> doggie trucl needed
[22:18:28] <mhaberler> truck
[22:19:07] -!- mhaberler has quit [Quit: mhaberler]
[22:24:18] -!- vegeetz has quit [Read error: Connection reset by peer]
[22:33:05] -!- jef79m has quit [Excess Flood]
[22:44:28] -!- bpuk has quit []
[22:45:58] -!- ravenlock has quit [Ping timeout: 246 seconds]
[22:49:24] -!- LeelooMinai has quit [Ping timeout: 260 seconds]
[22:53:42] -!- LeelooMinai has quit [Client Quit]
[22:57:56] -!- CBJamo has quit [Quit: Page closed]
[23:14:27] -!- vegeetz1 has quit [Quit: Leaving.]
[23:28:15] -!- vegeetz has quit [Client Quit]
[23:40:57] -!- asdfasd has quit [Disconnected by services]
[23:40:59] asdfasd1 is now known as asdfasd
[23:51:06] -!- vegeetz has quit [Client Quit]
[23:52:07] -!- fomox_ has quit [Ping timeout: 264 seconds]
[23:53:16] -!- mozmck has quit [Ping timeout: 252 seconds]
[23:53:49] -!- Nick001-Shop has quit [Ping timeout: 276 seconds]
