#emc-devel | Logs for 2010-07-23

[10:32:06] <SteveStallings> SteveStallings is now known as steves_logging
[12:49:09] <IchGuckLive> why are there 2 channels?
[12:52:08] <SWPadnos> this one is for development discussions (why certain code does what, changes to be made, etc)
[12:52:33] <SWPadnos> the other is for user questions, like how to set up EMC2, how to get it to do something, etc.
[18:46:54] <robh__> i am having a problem in that you can not change from G99 to G98 (and G98 to G99) in a cycle as it seems to ignore the change in G99 and G98 - i have a test program
[18:54:05] <robh__> http://pastebin.com/sNfkj8gL
[18:56:45] <jepler> on which line(s) is the motion different than you expect?
[18:59:24] <robh__> line 12
[18:59:30] <robh__> it does it right
[18:59:42] <robh__> but on line 15 it does not do a full retract as exspected
[19:02:04] <robh__> i would exspect as detailed in the ( ) i ran it in Sim to check it on what i saw on the machine
[19:02:42] <jepler> so if I delete everything after line 15 up to %, and then delete 13 and 14, I'll still see the problem?
[19:03:45] <robh__> if u did 17 and on correct, as 15 should be done still with retract of Z2 then lift to Z20 to go to line 16
[19:04:48] <robh__> i watched EMC status window and noticed it does not change the G98 to G99 when reading the lines
[19:06:19] <jepler> the modal codes on the mdi tab represent the readahead state not the current motion line state
[19:06:35] <robh__> i ment if u do EMC status from menu in Axis
[19:06:47] <jepler> that's the same source of information, same limitation
[19:06:57] <robh__> aah i see sorry
[19:07:26] <robh__> is there away one can one see the current read states
[19:08:41] <cradek> a G4 in the program will force the current states to match
[19:08:44] <jepler> on line 15 the sticky Z should be -2.7, the sticky R should be 2., the old Z should be what it was at the end of line 14, and G98 should be active
[19:09:47] <jepler> the last motion on line 14 is to Z2., so that's the old Z on line 15
[19:10:16] <jepler> G98 is active so you get option (2) retract to old Z which is 2.
[19:10:17] <robh__> yes but at end of the motion on L15 should return to G99 Z
[19:10:47] <robh__> so it can pass onto L16 at the cleance of Z20
[19:12:15] <jepler> at the start of line 15, the Z axis is at 2 not 20
[19:16:18] <robh__> correct, it should be there untill it has completed motion then retact back to inital Z as asked then process next line
[19:17:17] <robh__> i took this program from the fanuc machine which does as i have described then found EMC did not process the quick plane change in the cycle as exspected
[19:17:36] <jepler> the initial Z (the Z position at the end of line 14 / beginning of line 15) is 2, and that's where it retracts to.
[19:18:54] <jepler> if you change that line to
[19:18:54] <jepler> G98 X51. R1 (drill co-ord in G99, on finish G98 - retract to celarnce Z20)
[19:19:08] <jepler> you'll see that it still retracts to Z2, not the specified R
[19:21:00] <jepler> you expect "the original Z" to mean a different than than emc does. Maybe your definition is something like: the Z position back on line 12 when G82 was first programmed
[19:21:09] <robh__> its staying in cleanernce plane, my inital is alwasy 20 as Z has never changed only the R planes have. so it should rember that G98 is inital plane and not clearnce like G99
[19:21:54] <jepler> but emc's definition is "the Z position when starting to interpret this line of gcode"
[19:23:44] <robh__> one way to to chancel cycle, and redefine all the time, but its very wastfull time of programing etc where a quick change on the plane type and R and Zs and you can soon jump over clamps or boss's
[19:27:47] <robh__> i hope i am making sence what i am trying to explane
[19:28:37] <jepler> I think I understand what you expect. I've tried to explain why what emc does is consistent with what its documentation says. I don't have an opinion about which one is better.
[19:31:48] <robh__> maybe this helps, http://imagebin.ca/view/OI11Ho3M.html
[19:32:36] <robh__> i understand i just through it might a improvment to make EMC more power full and user friendly so what some of us cnc guys are used to, this is why i brought it up
[19:35:31] <cradek> robh__: what control are you using that has the behavior in the top picture?
[19:35:40] <robh__> Fanuc
[19:35:42] <cradek> (by the way your picture is awesome and helpful)
[19:35:45] <cradek> yes which one?
[19:35:49] <robh__> from Fanuc 6 to 31I does it
[19:36:04] <robh__> i can do a video if u like
[19:36:14] <jepler> I don't think that's necessary
[19:36:31] <jepler> what if line 15 said explicitly G82 ... ? would the fanuc still look like your 'expected' picture?
[19:36:33] <cradek> I believe you -- I was wondering if I could find a matching manual online that describes it
[19:36:38] <jepler> and what if it said G83 instead?
[19:36:53] <robh__> we use it alot if we have a grid of holes with boss's clamps or islands in the way, so u can put in all the co-ord then just skip up and over anything and no redefines etc
[19:37:17] <robh__> fanuc manual details it as i have talked about,
[19:37:33] <robh__> i dont know of any online ones sorry
[19:38:00] <robh__> never tryed it, i belive it would ignore it as it would see its still in G82 there
[19:41:19] <robh__> EMC seemed to start the program out ok, which made me thing maybe theres a bug or something, as not many people try and use it in this way i have done here
[19:43:17] <jepler> if you do the R1 on line 15 you'll see what emc is thinking -- that the old Z is 2, not that it's not really getting into G98 for that line..
[19:43:26] <jepler> I do see your side of things
[19:43:50] <robh__> cradek, we have a spare fanuc 10 manual, if you would be intrested?
[19:44:57] <cradek> robh__: does that have all the details about how the gcodes work? I'd be super interested if so.
[19:45:04] <robh__> yes
[19:45:09] <robh__> its programmers manual
[19:45:13] <robh__> big thick book
[19:45:30] <robh__> its Mill i dont think we have a spare Turning one we might
[19:46:14] <cradek> robh__: I have a fanuc lathe manual but I don't remember the model. it is a really old one with no roughing cycles
[19:46:25] <robh__> from your HNC?
[19:46:39] <cradek> no, that was a GE
[19:46:55] <cradek> this is one some other helpful person sent me
[19:47:06] <robh__> i see
[19:47:41] <robh__> well if you PM me your address or some where to send it, i will post it out no problem all free of charge
[19:47:59] <cradek> that would be great, thank you
[19:49:13] <cradek> (wish fanuc would put them all online...)
[19:49:27] <robh__> i keep telling them that, when we see the reps
[19:49:34] <robh__> as every once else does now days
[19:49:54] <robh__> but fanuc service people still have to pay for online access to them, which is quite costly
[19:50:13] <cradek> that's sure stupid
[19:50:31] <robh__> i found they will send the older stuff out now but you have to twist there arm
[19:51:16] <cradek> maybe they can't afford a scanner, since everyone's using EMC instead of buying from them now :-)
[19:52:14] <robh__> haha
[19:52:26] <jepler> that's a good one :-/
[19:53:03] <robh__> like when i do my fanuc mill, i will probly have to buy the servo manuals and spindle to set them up and get the pinouts as service manual does not tell you too much there
[19:54:29] <robh__> every time a machine rep comes into ours and see the EMC machine at work, they cant realy belive what they see on speed and price on a old machine :)
[20:10:51] <skunkworks> robh__: that is funny :)
[20:11:07] <robh__> funny?
[20:11:13] <skunkworks> machine repgs
[20:11:17] <skunkworks> reps
[20:11:55] <robh__> aah soz, yel they dont laugh much when u tell them how much you can do it for tho
[20:12:15] <jepler> can anybody think of a way that changing emc to work the way robh details would change an existing emc2 ngc file to retract less far than it does today?
[20:12:35] <jepler> personally, I can't..
[20:13:04] <jepler> (the change seems safe if you never retract lower than before, but unsafe if you could ever retract lower / less far than now_
[20:13:07] <jepler> )
[20:13:45] <jepler> .. not that I've figured out precisely what one would change to get robh's behavior..
[20:48:09] <cradek> jepler: not me
[22:01:06] <ries_> ries_ is now known as ries