#emc-devel | Logs for 2006-06-13

[00:17:48] <jepler> is OUTPUT_SCALE actually used for stepper systems, or can it safely be removed from the sample configurations?
[00:19:10] <SWPadnos> actually, I'd say that INPUT_SCALE should be remoevd, and use OUTPUT_SCALE for the steps/inch
[00:19:30] <SWPadnos> or even rename it - it's only used by HAL anyway, so the name can be anything
[00:19:33] <jepler> actually, INPUT_SCALE seems to be used by tkemc for somethin
[00:19:44] <SWPadnos> shouldn't be (bu tcould be)
[00:20:14] <jepler> er, xemc actually
[00:20:20] <jepler> ./src/emc/usr_intf/xemc.cc:// the size of the smallest increment to jog, = 1/INPUT_SCALE
[00:20:24] <jepler> ./src/emc/usr_intf/xemc.cc: if (NULL != (inistring = inifile.find("INPUT_SCALE", "AXIS_0"))) {
[00:20:43] <SWPadnos> ah
[00:20:44] <jepler> and I was just considering adding the same to axis, so that the lowest jog speed is about 1 step or encoder count per second
[00:21:07] <SWPadnos> well, just change that to OUTPUT_SCALE, and the problem is solved ;)
[00:21:17] <SWPadnos> it's not really an input scale, after all
[00:21:42] <SWPadnos> unless you look at from the viewpoint of the motor driver, I guess ;)
[00:40:04] <SWPadnos> umm - cant input_scale be used for things other than encoders? (like an analog scale)
[00:41:12] <jepler> no idea
[00:41:25] <jepler> I'll cross that bridge when the bug report is submitted
[00:41:29] <SWPadnos> heh
[00:41:40] <SWPadnos> that'll be Les or someone like him ;)
[00:45:32] <Roguish> SWPadnos: couple more questions after a long day of going nowhere very slowly?
[00:45:39] <SWPadnos> go ahead
[00:46:14] <Roguish> i have only gotten sporadic results with my simple drive. this is it: http://www.geocities.com/researchtriangle/lab/6584/XDDCMD_user_manual.pdfhttp://www.geocities.com/researchtriangle/lab/6584/XDDCMD_user_manual.pdf
[00:47:02] <Roguish> it should be simple. but recall i'm on the nuts and bolts side, not the elecrtronical side.
[00:47:10] <SWPadnos> ok
[00:47:46] <Roguish> dam, that link does not work. try www.xavien.com
[00:48:08] <SWPadnos> I have the manual from the first link - I just had to delete the extra copy of the location ;)
[00:48:25] <Roguish> oh. it's been a long day.......sorry
[00:48:42] <SWPadnos> heh
[00:49:01] <Roguish> page 5. signal logic table.
[00:49:23] <SWPadnos> ok
[00:49:57] <SWPadnos> you may need to invert the enable output, if it's connected to the brake pin
[00:50:08] <Roguish> i checked the 5i20 'enable' signal and it goes low when emc2 is 'on'
[00:50:28] <Roguish> maybe i'm backwards.
[00:50:52] <SWPadnos> hm. ok
[00:51:14] <Roguish> if i need to invert the signal, how?
[00:51:20] <Roguish> in hal?
[00:51:28] <SWPadnos> was that the actual pin, or the axis.0.amp-enable-out (or something like that)?
[00:51:53] <Roguish> the actual physical pin
[00:52:31] <Roguish> the encoder is working. turn the shaft and the display follows along
[00:52:36] <SWPadnos> ok. so what are you doing, and what actually happens?
[00:53:52] <Roguish> pwm goes to pwm, direction goes to dir, brake goes to enable. encoders to A and B.
[00:54:23] <Roguish> power in to the amp from an outside source, 12+ v, approx .5 amp.
[00:55:21] <Roguish> F1, F2 in emc to clear reset and enable.
[00:56:35] <Roguish> go to jog and click + or -. no motor on, but coordinate display changes momentarily, get following error, display returns to whatever.
[00:56:56] <Roguish> scope show pwm signal is working.
[00:57:08] <Roguish> real scope, not emc scope.
[00:57:25] <SWPadnos> ok
[00:57:33] <Roguish> what else?
[00:58:04] <SWPadnos> you may have a simple problem of the encoder counting in the opposite direction than expected
[00:58:16] <SWPadnos> you can check that by swapping the A and B leads
[00:58:34] <SWPadnos> (physically change the A and B wires connecting to the mesa card)
[00:58:51] <Roguish> yeah, i've reveresed that a a few times. same with power from amp to motor. 4 combinations and only 1 works.
[00:59:33] <SWPadnos> ok. just so you can actually see what's going on, increase FERROR to something abominably large, like 1 inch
[01:00:07] <Roguish> yeah done that also. increases the time to fault
[01:00:30] <SWPadnos> did you start with the stock m5i20 config?
[01:00:36] <Roguish> yes.
[01:00:58] <SWPadnos> ok. have you changed any of the .hal files, or only the ini?
[01:01:07] <Roguish> only the ini
[01:01:16] <SWPadnos> ok. - that's good
[01:03:32] <SWPadnos> what's the speed of your motors if the PWM duty cycle is 100%
[01:03:54] <Roguish> ok, just to check that i'm not maybe backwards on the hi/low signal. i put the neg vom probe on the ground side of the breakout board, the pos probe on the desired terminal.
[01:04:25] <SWPadnos> it's not an inverting breakout board is it? ;)
[01:05:42] <Roguish> i do not think so. there are not electronics on it. just a header and term strips
[01:05:47] <SWPadnos> ok
[01:06:05] <SWPadnos> this could be anything, including bad PID tuning.
[01:06:29] <Roguish> vom reads 3.76 with emc machine 'off', .21 with emc machine 'on'
[01:06:30] <SWPadnos> it's best to try things out with a bare HAL setup, similar to the tutorial
[01:06:34] <SWPadnos> ok
[01:07:23] <SWPadnos> of course, with the motor(s) on the bench, not attached to a machine ;)
[01:07:31] <Roguish> all on the desk
[01:08:45] <Roguish> i know the amp works. if i disconnect the brake and direction leads and let them 'float', then touch the brake lead the motor will start and run.
[01:09:30] <SWPadnos> what are your P, I, D, FF0, and FF1 settings?
[01:09:31] <Roguish> thought i might have a grounding problem. but not sure
[01:09:43] <SWPadnos> or can you just put the ini on http://pastebin.com
[01:09:58] <SWPadnos> could be. the brake should be on by default - ie, if a wire breaks
[01:10:27] <Roguish> p i d ff0 ff1: 1, 0, 0, 0, 1
[01:12:04] <SWPadnos> hmmm
[01:12:19] <SWPadnos> I think P is normally much higher than that
[01:12:35] <SWPadnos> can you pastebin?
[01:15:33] <Roguish> not easily. i'm presently on my work computer (xp)
[01:15:33] <SWPadnos> ah. me too (win2k, but still)
[01:15:33] <SWPadnos> so I can't really search through the source code very well
[01:15:33] <Roguish> getting dizzy swing around from desk to desk.....
[01:15:41] <SWPadnos> heh
[01:16:04] <Roguish> well, this amp is so simple, i thought it would work. it should. only $37 to drive 2 motors. good for toys and testing.
[01:16:06] <SWPadnos> you should try just leaving the brake/enable disconnected, or connected so as to leave the motor enabled
[01:16:37] <SWPadnos> hmmm - try using MDI mode instead of a jog. I think jogs go at G0 speed
[01:16:37] <Roguish> that's what i've been trying to do.
[01:17:02] <SWPadnos> ok. you said jog + or - ...
[01:17:07] <Roguish> yeas
[01:17:09] <Roguish> yes
[01:18:02] <Roguish> the 'direction' from the 5i20 definitely goes up and down, based on the + or - direction.
[01:18:35] <Roguish> that elusive 'dir' that i was refering to this morning.
[01:19:01] <SWPadnos> ok. jog is different from MDI or auto ;)
[01:19:38] <Roguish> yeah, there should be a jog speed.....i don't know emc2 yet.
[01:19:59] <Roguish> should be set in the ini file?
[01:20:24] <SWPadnos> but you can test with MDI - ssue G1 X0.1 F10, and vary the F until you get errors
[01:20:30] <SWPadnos> slower speeds shouldn't error out
[01:20:47] <SWPadnos> increase the FERROR so you can see what's actually happening with the system
[01:20:56] <Roguish> good idea. i'll try that. i also increase the ferror..............
[01:20:59] <SWPadnos> hard to tell when the motor only spins for 1/10 second
[01:21:07] <Roguish> right.
[01:21:35] <SWPadnos> and fiddle with halscope - you can see what the motion controller thinks is happening
[01:21:44] <Roguish> at least i'm getting used to using my scope (Tek 2246)
[01:21:49] <SWPadnos> plot things like following error and that kind of thing
[01:22:01] <Roguish> 4 channel, way tooooo many buttons.
[01:22:01] <SWPadnos> cool - love to have new equipment
[01:22:35] <SWPadnos> I just got an Agilent MSO6104 - 4 analog + 16 digital channels, 1 GHz bandwidth ...
[01:22:59] <Roguish> blazzzzin
[01:23:16] <SWPadnos> heh - yep. 1024x768 color LCD display too - very snazzy
[01:23:45] <Roguish> what are 'input scale' and 'output scale' in the ini file under each axis?
[01:24:15] <SWPadnos> they're used to set scaling constants for the output (motor driver) and input (encoder / linear scales)
[01:24:27] <SWPadnos> I think the meaning depends on the system being used
[01:25:01] <SWPadnos> for the m5i20, I believe the OUTPUT_SCALE is used to scale user units to the motor velocity
[01:25:26] <SWPadnos> so when the motion controller feeds a 1 into the 5i20, the table moves 1 inch/second
[01:25:45] <SWPadnos> INPUT_SCALE is the encoder counts per user unit (inches, I'm assuming)
[01:25:59] <SWPadnos> those need to be right, or nothing works very well
[01:26:11] <Roguish> with or without the quadrature?
[01:26:24] <SWPadnos> with
[01:26:30] <SWPadnos> that's how the 5i20 counts
[01:27:20] <Roguish> this is a simple HPheds 5500 encoder.
[01:27:55] <SWPadnos> that's fine. x1 or x4 is a function of the encoder reader, not the encoder itself (in general)
[01:28:01] <Roguish> it's 500 cpr
[01:28:14] <Roguish> so, input scale = 2000 ?
[01:28:21] <SWPadnos> ok, so probably 2000 counts/rev.
[01:28:23] <SWPadnos> not quite
[01:28:51] <SWPadnos> you expect to put this on a machine, right? (or is this for testing only?)
[01:29:10] <Roguish> for testing only rght now.
[01:29:14] <SWPadnos> ok
[01:29:41] <SWPadnos> normally, you'd take the encoder counts/rev (2000) * screw TPI * drive ratio
[01:29:55] <Roguish> not really enough counts for a good machine with direct drive. i usually run gearboxes of 10 or 20 to 1
[01:30:01] <SWPadnos> so 2000 * 5TPI screw * 2:1 ratio = 20000 counts/inch, for example
[01:30:13] <Roguish> right.
[01:30:31] <SWPadnos> but for now, 2000 is fine - then a move of 1 in emc should be one rev
[01:30:53] <SWPadnos> do you have index on the encoders?
[01:31:30] <Roguish> unfortunately not on these. i have a different test motor with index, i can try it.
[01:32:01] <SWPadnos> ok. no matter - it could just help in checking that emc is moving the motor the right amount
[01:32:59] <SWPadnos> have you done much fiddling with the INPUT_SCALEand OUTPUT_SCALE?
[01:33:27] <Roguish> a little.
[01:33:46] <SWPadnos> ok. then it's entirely possible that the problem lies there
[01:34:10] <Roguish> ?
[01:34:59] <SWPadnos> well, if emc doesn't know how hard to push the motor, or how to interpret the encoder feedback, it's pretty hard to control the motor position
[01:36:51] <Roguish> just found (again) the INI documentation.
[01:36:56] <SWPadnos> heh
[01:53:44] <SWPadnos> hiya pete - you just missed a farly long discussion regarding the 5i20 ;)
[01:53:55] <petev> hmm
[01:54:02] <SWPadnos> though Roguish is still here, so there may be more questions
[01:54:22] <petev> do u know if the new CVS server carried along the old passwds?
[01:54:26] <SWPadnos> how are you doing?
[01:54:32] <SWPadnos> no - it didn't.
[01:54:37] <Roguish> yes, i'm here. trying MDI
[01:54:47] <petev> so what happened to all of the old dev passwds?
[01:54:50] <petev> they are no good?
[01:54:50] <SWPadnos> you need to send cradek a public pgp key
[01:54:57] <petev> ok
[01:55:13] <SWPadnos> the server is actually sitting in jepler's basement
[01:55:18] <SWPadnos> way more reliable than SF lately
[01:55:34] <petev> yeah, SF has gotten bad
[01:55:47] <SWPadnos> it was mostly unusable for the last 2-3 months
[01:56:01] <SWPadnos> at least for projects with certain starting letters (like E)
[01:57:26] <cradek> petev: just send me your .ssh/id_dsa.pub
[01:57:47] <cradek> or id_rsa.pub, don't think it matters
[01:57:55] <petev> cradek: I'll generate a key now
[01:58:01] <cradek> ssh-keygen -t dsa
[01:58:11] <petev> what format do you want?
[01:58:14] <petev> ok, dsa
[02:02:58] <petev> ok, sent
[02:03:46] <cradek> hmm, that doesn't look like the format I want
[02:04:27] <cradek> are you using something other than openssh?
[02:04:46] <petev> that was from my doze box
[02:04:53] <cradek> hmm
[02:04:57] <petev> never had a problem using it on linux before
[02:04:59] <cradek> is that what you'll be using to do cvs?
[02:05:06] <petev> what format do you want?
[02:05:53] <cradek> one line, something like: ssh-dss AAA..........== user@host
[02:05:53] <jepler> The line should look something like this: ssh-rsa AAAAB3...(about 4 lines at 80 cols).....Q== petev@example.com
[02:07:41] <petev> how many bits in the keys you are using?
[02:07:58] <cradek> I think people use various numbers
[02:08:09] <petev> I sent you a 1024 bit dsa key
[02:08:38] <cradek> I'm sure that's true (it says so) but it's not in a format I can put in your ~/.ssh/authorized_keys
[02:09:04] <petev> let me boot my notebook and try open-ssh
[02:09:12] <SWPadnos> has it been mime encoded or something silly like that?
[02:09:18] <SWPadnos> by the mailer
[02:09:25] <cradek> it's "armored" differently
[02:09:28] <SWPadnos> ok
[02:09:32] <cradek> ---- BEGIN SSH2 PUBLIC KEY ----
[02:09:36] <cradek> Comment: "dsa-key-20060612"
[02:09:38] <cradek> ...
[02:10:24] <petev> I think u can just delete those lines and add the user name for open-ssh
[02:11:53] <cradek> These others all end with ==, whatever that means (yours doesn't even if I concatenate the lines), I can try it if you want
[02:12:32] <petev> give it a shot, I have to change network config on my notebook so it will be a few minutes
[02:12:47] <cradek> ok
[02:14:50] <petev> does the server know the key yet?
[02:15:17] <cradek> done
[02:15:22] <petev> ok, hang on
[02:15:24] <cradek> login petev
[02:15:46] <petev> got an auth error
[02:15:55] <petev> let me try open-ssh keygen
[02:16:04] <cradek> it should not ask for a password
[02:18:14] <petev> key looks the same except for the key type at beginning and user name at end
[02:18:55] <cradek> I agree they look similar
[02:25:20] <Roguish> SWPadnos: IT LIVES ! ! ! got commanded motion. grounded EVERYTHING. raised P, sorted out scales, redid encoder and motor power several times. now, + goes forward and counts up, - goes reverse and counts down. Thannks.
[02:25:34] <SWPadnos> great! glad to help
[02:26:40] <Roguish> time for a beer..........................................
[02:27:01] <SWPadnos> heh. it's still early enough for the House of Prime Rib ;)
[02:27:49] <cradek> petev: ready
[02:28:45] <petev> ok, let me try the open-ssh key on the doze box
[02:29:16] <cradek> try it on unix just to make sure I didn't mess it up
[02:30:08] <petev> I only have svn on the notebook
[02:30:16] <petev> not near my other linux box with cvs
[02:30:22] <cradek> ssh petev@cvs.linuxcnc.org cvs server
[02:30:31] <petev> the private key on doze looks quite different
[02:30:57] <petev> will the server support multiple public keys for the same user?
[02:31:01] <cradek> yes
[02:31:01] <SWPadnos> what cvs software do you use in Windows?
[02:31:17] <petev> ok, let me modify the doze public key and send it to u
[02:31:37] <cradek> ok but I did that already...? maybe I messed it up.
[02:31:46] <petev> leave the linux one and I'll xfer it to my other linux box later
[02:31:59] <cradek> sure
[02:42:40] <cradek> petev: ready
[02:42:44] <petev> ok
[02:43:31] <cradek> yay
[02:43:53] <petev> it authed, but looks like I have something else messed up
[02:44:11] <cradek> Jun 12 21:48:09 cvs sshd[37739]: Accepted publickey for petev from ...
[02:44:28] <cradek> looks fine from here too
[02:44:49] <petev> yeah, I think my root file or repository is messed up
[02:45:01] <petev> I thought I fixed it all, but I must have missed something
[02:45:08] <SWPadnos> did you do a new checkout?
[02:45:11] <cradek> you're not trying to update an old checkout?
[02:45:23] <petev> yes, I was
[02:45:32] <cradek> down that path lies madness
[02:45:57] <petev> it's not that bad, a few find commands, greps, awks... ok, maybe you're right
[02:46:35] <cradek> chris@buster2:~/emc2.head$ cat CVS/Root
[02:46:35] <cradek> :ext:cradek@cvs.linuxcnc.org:/cvs
[02:47:20] <petev> what's your repository name, still emc2?
[02:47:33] <cradek> yes
[02:49:24] <cradek> be sure you up -dP if you mess with the files manually (some dirs have been removed and added)
[02:49:51] <petev> seems like my cvs client is getting old data, maybe it cached stuff
[02:49:56] <jmkasunich> just get a fresh one
[02:49:58] <petev> I'll get it sorted
[02:50:13] <petev> yeah, I had some additional files in the tree
[02:50:16] <petev> shouldn't be this hard
[02:50:38] <cradek> are you on dialup or something?
[02:50:45] <petev> no, why?
[02:50:57] <cradek> just wondered - it would be done by now
[02:51:24] <cradek> sorry, I'll leave you be, looks like my part is done
[02:51:27] <petev> true, but I would have to find all the new and modified files and put them back
[02:51:48] <SWPadnos> do a new checklout, and diff ;)
[02:51:52] <petev> plus it's annoying me that it's not working
[02:52:18] <SWPadnos> ah - the real problem ;)
[02:57:03] <petev> now we're talking
[02:57:34] <cradek> yay
[04:00:19] <SWPadnos> SWPadnos is now known as SWP_Away
[11:52:23] <jepler_> jepler_ is now known as jepler
[13:48:43] <SWP_Away> SWP_Away is now known as SWPadnos
[14:48:48] <jepler> I notice that sourceforge has taken to removing the "CVS is down" messages once it's no longer down
[14:48:59] <jepler> as opposed to leaving them, and adding a new note that it's back up
[14:49:10] <SWPadnos> heh
[14:49:22] <SWPadnos> "leave no trace" ...
[14:49:56] <jepler> cradek: we've had two users note that it's clumsy to use a jog wheel when the gui is axis, because axis waits to switch to manual mode.
[14:50:31] <jepler> cradek: should we do something about that in axis? I am not sure. When the jog wheel was going through halui, I believed halui should do the same thing axis does to switch modes as needed.
[14:50:41] <jepler> but that's much harder with the jog wheel being all in realtime
[14:50:52] <SWPadnos> does it hurt to switch modes when people hit F3/F5?
[14:51:06] <cradek> only in that there are three modes and two tabs
[14:51:44] <SWPadnos> perhaps change to the mode where the jogwheel works when the "manual" tab is raised
[14:51:54] <cradek> but with the new jogwheel, do these gui modes even matter? I think the motion controller modes are completely separate
[14:52:04] <jepler> the problem I anticipate is the need to detect that "the program is done (or aborted or ...)" and switch back to manual mode
[14:52:29] <jepler> cradek: jmk was the second person who noted it, so I assume it does affect the realtime jogging
[14:52:32] <cradek> right.
[14:52:38] <cradek> hmm
[14:52:43] <SWPadnos> well - this exposes a sort of dichotomy in the "NIST principle"
[14:53:06] <cradek> I thought the motion controller modes were homed/in position/free/coordinated (the real machine modes)
[14:53:26] <cradek> and jogging is allowed when in free mode
[14:53:28] <cradek> I'd have to look at it
[14:53:49] <SWPadnos> in position? do you recall what that was for?
[14:53:59] <cradek> SWPadnos: can you elaborate about what you mean?
[14:54:08] <SWPadnos> re: the NIST principle?
[14:54:15] <cradek> I think in position means we're stopped with no pending motion
[14:54:17] <jepler> I guess my position is AXIS's current behavior isn't a bug, and when worded as a feature request it's not something to be done for 1.4.
[14:54:19] <SWPadnos> ok
[14:54:34] <cradek> yes nist
[14:54:57] <cradek> jepler: before changing axis, I'd want to understand the problem, if there is one, regarding the new jogwheel implementation
[14:55:27] <SWPadnos> ok. they say that (a) ther eshould be only one locus of control and (b) that all user interfaces must use "stateless" controls, so that there are no surprises when changing the locus of control
[14:55:48] <cradek> ok
[14:56:00] <cradek> so the gui modes violate nistness
[14:56:04] <skunkworks> locus - great - another word I have to look up
[14:56:07] <cradek> (very true)
[14:56:13] <SWPadnos> not really a dichotomy, I guess, but here, we're trying to make multiple locations coordinate together, and have multiple simultaneous or near-simultaneous controllers
[14:57:05] <SWPadnos> that's the problem - it's just as I was saying to etla - it makes no sense to use axis to change the behavior of a physical button on a panel
[14:57:25] <cradek> I agree that's a bit bogus
[14:58:01] <cradek> but imagine (not saying this is the case) the 3 gui modes represent something real on the machine
[14:58:02] <SWPadnos> this is the same thing, only the RT jogwheel implementation requires help from some other UI
[14:58:25] <cradek> just like you don't have to have a Flood button on each gui, you don't have to have 3 mode buttons on each gui.
[14:58:45] <SWPadnos> only on those GUIs where you want to control flood or machine mode, respectively ;)
[14:59:03] <cradek> it's that these modes are an accidental implementation detail, not a real feature of the machine, that is making everything clash
[14:59:38] <SWPadnos> perhaps the "real" problem is that the RT jogwheel has added another control point to the motion controller. ther used to be only one, NML messages
[15:00:09] <SWPadnos> now you have two, and they can conflict, so the mode becomes truly important
[15:00:30] <cradek> I maintain that the gui modes aren't important
[15:00:41] <SWPadnos> I'm thinking of machine modes at the moment
[15:00:52] <cradek> ok I agree machine modes are important
[15:01:04] <SWPadnos> they apparently exist, since jmk had trouble with the RT jogwheel ...
[15:01:06] <cradek> for instance "it's already moving" is important to a jogwheel
[15:01:25] <SWPadnos> in some cases - ie, already moving under program control
[15:01:34] <cradek> I should test/try this before I say much more
[15:01:39] <SWPadnos> already moving as a result of a previous jog isn't as important
[15:02:09] <cradek> that's right
[15:02:20] <cradek> it's coordinated mode (a program or MDI is running) that we have to worry about
[15:02:54] <SWPadnos> right. (though it would be nice for jogs to be coordinated as well, so that you get straight lines if you move diagonally)
[15:03:05] <cradek> that's a can of worms
[15:03:10] <SWPadnos> heh
[15:03:19] <cradek> that's no longer free mode jogging
[15:03:23] <SWPadnos> I've got a worm-seeking can opener, I guess
[15:04:28] <cradek> I should put the motion controller (real) modes in the stat buffer
[15:04:57] <cradek> then we could lose the accidental gui modes and the guis could work right
[15:05:18] <cradek> for instance, disable the jog buttons while an mdi is still running
[15:05:37] <SWPadnos> hmmm - something's got to be ther already, or e.g. tkemc wouldn't change the estop / auto / MDI indicator when you cahnge it in halui
[15:05:43] <cradek> I bet today if you start a long-running MDI, switch to manual mode and jog, you get an error
[15:06:08] <cradek> no you're confusing the MC (real) modes with the gui modes
[15:06:29] <cradek> sure you can change the gui modes with halui etc
[15:06:49] <cradek> but you can do that whenever you want, and it just gives you errors when you try to use them, like the example I just gave
[15:07:01] <SWPadnos> could be. but the gui modes would need to be in the stat buffer for them to get transferred to all attached UIs, no?
[15:07:17] <cradek> yes I think so
[15:07:22] <cradek> brb, work
[15:07:25] <SWPadnos> or is there another path that I'll remember after I've had more coffee? ;)
[15:07:27] <SWPadnos> ok
[19:27:45] <alex_joni> hello
[19:28:31] <jepler> hi alex
[19:28:49] <alex_joni> * alex_joni is up in the mountains :)
[19:28:59] <alex_joni> crappy net connection
[19:35:54] <skunkworks> we will just type slow
[19:36:34] <alex_joni> nah.. the box running screen is OK
[21:03:30] <SWPadnos> hi Ray - get any sleep?
[21:04:37] <kerry_> Yes I did
[21:04:48] <SWPadnos> good deal
[21:04:53] <kerry_> Sun is just creaping up
[21:05:07] <SWPadnos> yeah - 5AM
[21:05:08] <alex_joni> hi ray
[21:05:15] <kerry_> Hi alex
[21:05:20] <alex_joni> 00:10 < SWPadnos> yeah - 5AM ?
[21:05:27] <SWPadnos> in Beijing ;)
[21:05:47] <alex_joni> oh, ray switches sides :P
[21:05:53] <SWPadnos> heh
[21:06:15] <kerry_> We met a guy from the tool show on the elevator last night.
[21:06:32] <kerry_> Works for a Spanish cnc maker
[21:08:34] <SWPadnos> cool
[21:09:06] <kerry_> Any pressing issues with the software?
[21:09:34] <SWPadnos> I don't think so. jmk added RT jogwheel support to motion recently
[21:09:48] <SWPadnos> there's been a fair amount of discussion on blending and cutter comp
[21:10:43] <jepler> some day soon we'll have to decide when to branch for "2.1"
[21:10:45] <kerry_> I saw that jeppler had done some work and was asking for testing
[21:11:10] <SWPadnos> ah - for the segment concatenation addition
[21:11:18] <jepler> kerry_: cradek found one small bug that would only affect programs that have a G1 motion as the last motion .. besides that I haven't heard any complaints
[21:11:28] <jepler> and it does improve contouring, as shown my this halscope graph: http://emergent.unpy.net/files/sandbox/speed.png
[21:12:04] <kerry_> * kerry_ goes looking
[21:12:25] <SWPadnos> is that vector velocity? (with the hypot)
[21:16:30] <kerry_> That plot as good as I've seen.
[21:22:39] <cradek> kerry_: have you had a chance to play with AXIS?
[21:25:47] <kerry_> Not since the addition of the scope and such.
[21:27:25] <kerry_> I think that I've got the latest installed on this laptop so will plan some time to do that.
[21:27:56] <kerry_> Kerry is impressed that usb works properly, including his digital cam
[21:28:21] <cradek> nice
[21:28:23] <SWPadnos> even firewire / a video camera worked for me
[21:30:59] <kerry_> Really. I've got a board but nothing to connect.
[21:31:07] <kerry_> Will try that when I get back home.
[22:04:32] <Roguish> SWPadnos: is it possible (read:easy) to have the 'testing' along with the 'released' versions on the same computer?
[22:04:46] <SWPadnos> yes
[22:05:09] <SWPadnos> when you get the CVS files, just use ./configure --with-run-in-place
[22:05:21] <SWPadnos> or was that --enable-run-in-place
[22:05:32] <Roguish> i saw in some of the docs to do a cvs checkout. then the run in place.
[22:05:46] <SWPadnos> anyway - do that, and the CVS version will be run from the directory that you put the source
[22:06:08] <SWPadnos> right. you also need to 'sudo make setuid' after building emc2
[22:06:30] <Roguish> sounds good. think i'll ttry it. is a lot of that new stuff i've been seeing here in the testing?
[22:08:28] <SWPadnos> I don't know exactly, but it seems likely. I think 2.0.1 was released several weeks ago
[22:09:05] <Roguish> yeah, that's what i have now.
[22:10:12] <Roguish> i would like to see the work on halscope
[22:10:18] <Roguish> and axis
[22:10:37] <SWPadnos> yes -the scope colors should be quite nice