#emc-devel | Logs for 2008-04-14

[00:49:31] <jmkasunich> I'm gonna have to extract a few sections of this program and look at them a little closer - I'm seeing some jerkiness that might be a blending issue
[00:50:01] <jmkasunich> (or it might just be segments that are so short that one-segment lookahead isn't enough to allow smooth movement)
[01:31:42] <cradek> jmkasunich: that looks very nice
[01:36:18] <cradek> strange - I wonder what's wrong with the jogwheel on joint 8
[01:37:01] <cradek> well ... saying that made me figure it out
[01:37:28] <cradek> (he doesn't have a joint 8)
[01:37:53] <SWPLinux> wouldn't the quill be the joint for that?
[01:38:21] <SWPLinux> or is his kind not-quite-max-kins?
[01:38:25] <SWPLinux> s/kind/kins/
[01:38:31] <cradek> I think there is no quill
[01:38:51] <SWPLinux> hmmm. so he needs to be in teleop mode or something?
[01:39:09] <cradek> well he wants to jog W
[01:39:21] <cradek> other than teleop (which probably doesn't even work) there is no way to do that
[01:39:33] <cradek> that's part of what jmk has a plan for fixing
[01:39:45] <SWPLinux> unless you actually have a joint corresponding to W, which a quill could be
[01:39:51] <SWPLinux> (if you have one)
[01:39:53] <cradek> right
[01:40:21] <cradek> a quill would sure make TLO easy
[01:40:25] <SWPLinux> heh
[01:40:35] <cradek> but, it's totally redundant if the head is fully controlled
[01:40:40] <SWPLinux> I thought this was one of the machines with the pivoting head
[01:40:46] <cradek> yes
[01:41:29] <SWPLinux> I'm not sure it's redundant
[01:42:11] <SWPLinux> it certainly is if you ignore the possibility of hitting the workpiece
[01:42:14] <cradek> adding a quill would give you the same 5 DOF but a sixth joint
[01:42:26] <cradek> ??
[01:42:56] <SWPLinux> it's a question of clearance
[01:43:07] <cradek> oh I see what you mean
[01:43:11] <SWPLinux> the (massive) head clearing the workpiece
[01:43:24] <cradek> I meant redundancy in more of a mathematical sense
[01:43:28] <SWPLinux> heh
[01:43:33] <SWPLinux> but not a practical one ;)
[01:43:41] <SWPLinux> (theory :: practice)
[01:43:53] <cradek> I think the head on this machine only tilts 30 degrees or so
[01:44:43] <SWPLinux> yeah, 30-ish in one direction, a different amount the opposite way (same joint), and less on the other rotary IIRC
[01:55:13] <jmkasunich> +30 -20 IIRC
[01:55:29] <jmkasunich> and either +/- 20 or +/- 30 the other way
[01:59:36] <SWPLinux> right - side-to-side was supposed to be the same in either direction, but the head limited it in practice
[02:00:14] <jmkasunich> I like those machines - they're beasts
[02:00:19] <SWPLinux> heh
[02:00:29] <SWPLinux> beasts with little teeth when we saw them ;)
[02:00:55] <jmkasunich> true
[02:01:35] <jmkasunich> but I think the endmill that you took a photo of setting on Max fits the Cinci's
[02:02:24] <SWPLinux> heh. that and the 3-foot drills ;)
[02:06:09] <jmkasunich> well, I think I'm gonna do something really novel, and go to bed before 11pm
[02:06:21] <jmkasunich> 5 manifolds left, about an hour and a half - easy job tomorrow evening
[02:06:43] <jmkasunich> I'm thinking about making a video of one run
[02:07:08] <jmkasunich> but it might be pretty boring - most of the action is down in the port, and it gets buried in chips
[20:39:29] <SWPLinux> jepler: you around?
[20:40:06] <SWPLinux> I'm wondering if you remember our experimentation with USB input devices last year
[20:40:25] <jepler> only in the vaguest terms
[20:40:29] <SWPLinux> I recall you using xev to see what events the devices sent (like the volume knob)
[20:40:48] <jepler> I don't think it was xev
[20:40:59] <SWPLinux> do you know offhand how to tell X/the kernel to ignore a device so you can use it with hal_input
[20:41:23] <SWPLinux> oh - right, it was a program that used the python input library or something
[20:42:23] <jepler> there's also a C program called 'evtest' which is available on the internet
[20:42:30] <SWPLinux> ok, thanks
[20:43:15] <SWPLinux> I realized that I have a nice "display" to show some HAL stuff - the LEDs on an external keyboard
[20:43:58] <jepler> if that's big enough for your audience to see
[20:44:01] <SWPLinux> but I don't remember/know how to make sure it's available for hal_input
[20:44:18] <SWPLinux> it should be, the room is relatively small and it's only a few blinkenlights
[20:44:33] <jepler> keyboard LEDs should be supported by hal_input
[20:44:39] <SWPLinux> heh - firefox sees "blinken" as a misspelled word, but "blinkenlights" isn't ;)
[20:45:21] <jepler> Component Pins:
[20:45:22] <jepler> Owner Type Dir Value Name
[20:45:22] <jepler> 03 bit IN FALSE input.0.led-capsl
[20:45:22] <jepler> 03 bit IN FALSE input.0.led-numl
[20:45:22] <jepler> 03 bit IN FALSE input.0.led-scrolll
[20:45:37] <SWPLinux> yes, they're supported. I wasn't sure if there's a way to prevent core X events with a secondary keyboard (short of changing xorg.conf and restarting X)
[20:46:01] <jepler> If the first character of the inputspec is a "+", then hal_input
[20:46:04] <jepler> requests exclusive access to the device.
[20:46:19] <jepler> so in that case you'd figure out how to specify the external keyboard, and put "+" at the front
[20:46:33] <jepler> that'll take away the whole external keyboard from linux / X though, so you can't type on it at that point
[20:46:50] <SWPLinux> ok. hmmm. this all works on sim?
[20:46:53] <jepler> yes
[20:47:00] <jepler> hal_input is userspace
[20:47:03] <SWPLinux> well then I guess I can try it now :)
[20:47:05] <SWPLinux> thanks
[20:48:53] <jepler> my recollection is that if you don't do anything in X that would change the LED state, that it won't change the state that hal_input sets. I'm not sure, though.
[20:49:06] <jepler> test first is my advice
[20:49:16] <SWPLinux> yep - doing that now
[20:53:54] <SWPLinux> hmmm. I need to be root to load hal_input, yet strangely sudo halcmd loadusr ... doesn't work (halcmd: command not found)
[20:54:31] <jepler> if using run-in-place, sudo may blow away the environment set by . emc-environment
[20:54:44] <SWPLinux> it seems that way
[20:54:53] <jepler> but see the man hal_input for how to set devices so that users may open them with hal_input
[20:56:32] <SWPLinux> a-ha. the correct device number is 8, not 9 :)
[20:57:44] <SWPLinux> heh - interesting names:
[20:58:00] <SWPLinux> ...
[20:58:02] <SWPLinux> 21044 bit OUT FALSE input.0.key-z
[20:58:04] <SWPLinux> 21044 bit OUT TRUE input.0.key-z-not
[20:58:06] <SWPLinux> 21044 bit OUT FALSE input.0.key-zenkakuhankaku
[20:58:08] <SWPLinux> 21044 bit OUT TRUE input.0.key-zenkakuhankaku-not
[20:58:09] <SWPLinux> 21044 bit IN FALSE input.0.led-capsl
[20:58:11] <SWPLinux> 21044 bit IN FALSE input.0.led-compose
[20:58:13] <SWPLinux> 21044 bit IN FALSE input.0.led-kana
[20:58:15] <SWPLinux> 21044 bit IN FALSE input.0.led-numl
[20:58:17] <SWPLinux> 21044 bit IN FALSE input.0.led-scrolll
[20:58:18] <SWPLinux> ...
[20:58:22] <SWPLinux> and there are only 3 visible lights
[20:58:40] <cradek> scrolll!
[20:58:51] <SWPLinux> yep, that's what it says
[20:59:12] <SWPLinux> I ws wondering where the zenkakuhankaku key is
[20:59:15] <SWPLinux> wsa
[20:59:16] <SWPLinux> was
[20:59:38] <jepler> apparently your keyboard reports that those keys exist ...
[20:59:48] <jepler> Specify -L to only create pins for LEDs
[20:59:51] <SWPLinux> heh
[20:59:51] <jepler> hal_input -L +keyboard
[20:59:56] <SWPLinux> yep, saw that
[21:00:05] <SWPLinux> there's a yen key too
[21:00:39] <SWPLinux> I guess it's time to make a vcp panel with all those keys in it
[21:00:57] <SWPLinux> since I'm now curious as to what a zenyaddayadda key is
[21:02:29] <cradek> I think zenkakuhankaku is the key that switches between half and full width characters on a japanese keyboard
[21:03:38] <cradek> http://www.suse.de/~mfabian/suse-cjk/kinput2-canna-hankata-zenro.html
[21:03:44] <cradek> ... or something
[21:03:59] <SWPLinux> ok, this is weird
[21:04:20] <SWPLinux> the normal keys aren't passed to X. "special" keys are, like volume control, email, web ...
[21:04:25] <SWPLinux> but not suspend
[21:04:28] <SWPLinux> (luckily)
[21:05:40] <jepler> huh
[21:07:06] <SWPLinux> I may have keyboard suspend disabled, so that could be getting sent
[21:07:28] <jepler> perhaps there are two separate /dev/input items for the keyboard?
[21:07:50] <SWPLinux> hmmm. actually that could be. it is listed for input 9
[21:08:12] <SWPLinux> yep. that's it
[21:08:36] <SWPLinux> I was wondering why I got such a short list before, which is why I switched to 8
[21:09:16] <SWPLinux> what's really cool is that I finally plugged in my ATI Remote Wonder, which actually works pretty well
[21:09:27] <SWPLinux> (USB remote control with lots of buttons)
[21:19:34] <jepler> neat
[23:32:13] <jmkasunich> SWPLinux: when is the big event?