#emc-devel | Logs for 2009-12-11

[00:50:38] <CIA-13> EMC: 03micges 07tlo_all_axes * rf2c74845369c 10/ (4 files in 4 dirs): Update python code and Axis to use new tool table format
[09:38:17] <micges_work> alex_joni: what do you think about adding halui.home-all command pin?
[09:54:15] <micges_work> it was requested few times so I think we can add it to master finally
[11:15:03] <alex_joni> micges_work: fine by me
[11:15:17] <alex_joni> wonder if we should make it smart (like AXIS)
[11:15:33] <alex_joni> so it first checks if HOMING_SEQUENCE is defined, and only export it then
[11:25:23] <micges_work> I think halui needs some rethink, it has many duplicating pins with motion and io,and it locking gui on many commands through hal
[11:26:00] <micges_work> but yes I'll make home-all pin sequence smart
[16:20:45] <alex_joni> http://wiki.python.org/moin/PythonDebuggers might come in handy
[19:05:28] <cradek> I wonder what john wants to do with more than 9 fixture offsets - kim left before I could ask
[19:12:03] <robh_> maybe alot of parts from a plate, 100+ parts from a 6 foot sheet if plasma etc, but then thats why you have subs and G10 L20 for unlimited data shift
[19:12:37] <cradek> he has a bp series 2, so not that huge a table
[19:12:42] <robh_> cradek, id like to help add some Gcode cycles, some lathe functions like grooving, area clearnce etc how can i help to add these
[19:13:41] <cradek> robh_: currently, the only way to add new cycles is to put them in the interpreter. I'd be happy to review any of your submissions.
[19:14:42] <robh_> im worls best programing of C that is 1 of my major problems here i was looking over how the interpreter works and how files go together other day to see how things work
[19:15:27] <cradek> sorry what is worls best?
[19:15:41] <robh_> not the worlds best programmer
[19:15:47] <cradek> ah
[19:15:56] <cradek> I am sure I have never met the world's best programmer
[19:16:07] <robh_> maybe i should start by typing up the cycles and laying out what is needed etc
[19:16:11] <cradek> if you have questions just ask, maybe someone can help
[19:16:34] <cradek> for area clearance it seems like there's a lot of geometry work involved
[19:17:01] <robh_> yea, when you want to detect under cuts in the geometry layout of part
[19:17:03] <cradek> you have to cut up the area into slices, and calculate the slice boundaries by cutting up the original moves accordingly
[19:17:18] <cradek> yes for undercuts it's even worse - a slice can break apart into several pieces
[19:17:23] <robh_> i have a area cleance layed out in varible programming
[19:17:58] <cradek> I don't think I see how that can be done all in gcode
[19:18:16] <cradek> I'm picturing the kind of cycle where you program the outline of the area (made of lines and arcs) and the cycle clears it
[19:18:31] <robh_> its not true Gcode varible programming on fanuc etc, its abit like basic with goto etc and basic math operations
[19:18:36] <cradek> I don't think you can do that in gcode without a huge huge mess
[19:18:48] <cradek> oh I didn't understand what you meant then
[19:19:03] <cradek> I thought you had something like emc subroutines that did all the geometry work to cut up an area into slices
[19:20:33] <robh_> stick to easy ones first then take on the more complex cycles is wise
[19:21:00] <cradek> for a clearing cycle, I'd use the new queue that the interpreter has for cutter compensation. it lets you queue up a set of moves and operate on their geometry
[19:21:13] <cradek> I see that as a very advanced project
[19:22:15] <cradek> maybe you could start with a round pocket (helical mill hole) cycle instead. I've seen a lot of people ask for that.
[19:22:34] <cradek> ramp in + spiral out + finish pass, or something like that
[19:23:01] <robh_> im not sure how you would layit out yet, wether you ask for a lot of points in cycle line make messy 1 line cycle code , or use ask for a Nblock
[19:23:08] <cradek> that would be quite a bit easier to accomplish, but would require some of the same kinds of additions
[19:23:10] <robh_> as you end up with quite afew points defind
[19:24:10] <cradek> yeah for lathe roughing, I've seen the system where you say everything from HERE to N12345 is the outline to be roughed in (expressed in regular gcode lines and arcs)
[19:24:37] <cradek> again that's a major project - we don't do anything like that kind of thing with N words currently.
[19:25:55] <robh_> i think first step to do, is think out cycles, what varibles to use with them, and what motions of the cycle realy do do, then start to develope them
[19:27:32] <robh_> thx for advice and all
[19:27:41] <cradek> welcome