#emc-devel | Logs for 2007-07-08

[00:01:11] <jmkasunich> I think I need food
[00:02:52] <cradek> yay ohmage
[00:03:04] <jmkasunich> ?
[00:03:21] <cradek> motor
[00:03:45] <cradek> has a better chance of turning if both brushes are touching something
[00:03:51] <jmkasunich> true
[00:05:12] <jmkasunich> even better if they're touching something that spins
[00:05:43] <cradek> I wonder why jon took the filtering off index. that's the most important place for it.
[00:05:59] <jmkasunich> that is one of those questions that I just don't ask
[00:08:07] <cradek> yay it spins
[00:08:13] <jmkasunich> cool
[00:08:18] <jmkasunich> what are you spinning anyway, the spindle?
[00:08:35] <cradek> the spindle motor
[00:09:34] <jmkasunich> bbwf
[00:09:41] <jmkasunich> (when fed)
[00:17:44] <jepler> cradek: what are you doing to your spindle motor?
[00:20:01] <cradek> trying to get it to work again
[00:20:08] <cradek> lost to time is the reason I took it apart in the first place
[00:21:40] <cradek> (it is working again now)
[00:24:37] <jepler> that's good to hear
[00:24:42] <jepler> your lathe, or your mill?
[00:25:14] <cradek> mill
[00:31:03] <cradek> ssh-add
[00:31:05] <cradek> oops
[01:11:10] <jmkasunich> back
[01:11:24] <jmkasunich> amazingly I was able to walk the dog around the block with no insanity
[01:50:03] <jmkasunich> If I was still running my old PC, I'd be getting annoyed at jepler about now
[01:51:47] <jepler> jmkasunich: haha
[01:52:07] <jepler> not so bad with the new one?
[01:52:38] <jmkasunich> not at all
[01:52:59] <jmkasunich> incremental builds take about 30 seconds each, and have no noticable effect on the responsiveness of the system
[01:53:38] <jmkasunich> If I wasn't seeing the CIA entries I wouldn't even know you were committing
[01:53:39] <jmkasunich> far cry from before
[01:55:27] <jepler> it's so nice to have the compile farm
[01:55:49] <jmkasunich> most of the time lately we haven't had the RT dapper slot
[01:56:04] <jmkasunich> because that was (and is) running native on my old machine, not on a VM
[01:56:11] <jmkasunich> I don't run the old machine 24/7 anymore
[01:56:15] <jepler> what's your desktop machine?
[01:56:17] <jmkasunich> (its running today)
[01:56:20] <jepler> not dapper?
[01:56:30] <jmkasunich> dapper with stock (non-RT) SMP kernel
[01:57:01] <jepler> it should be possible to test the build of the RT version without running the RT kernel
[01:57:12] <jmkasunich> bah
[01:57:17] <jepler> just have the necessary packages installed
[01:57:51] <jmkasunich> I don't even have the RT kernel installed on this box
[02:21:14] <jmkasunich> jepler: the tabbed widget that lets you choose between MDI and manual in Axis - what's that called?
[02:21:23] <jmkasunich> is it tkinter, or bwidget, or ?
[02:25:31] <jmkasunich> ah, its a bwidget notebook
[02:36:02] <jepler> yes
[02:36:15] <jepler> and it's a PITA to use from python, if I recall correctly..
[02:36:23] <jmkasunich> oh joy
[02:36:36] <jepler> similar to the scrolled whatsis, there's a trick to putting the stuff in the tabs
[02:37:14] <jepler> page = notebook.insert(Tkinter.END, i, text=os.path.basename(f))
[02:37:18] <jepler> maybe it's not too tricky
[02:37:26] <jepler> http://mail.python.org/pipermail/tkinter-discuss/2004-November/000251.html
[02:38:02] <jmkasunich> thanks
[02:38:11] <jmkasunich> * jmkasunich shamelessly copies
[02:48:41] <jmkasunich> interesting - when first created, no content shows for the notebook pages - only appears once you click a tab
[02:48:59] <jmkasunich> do I have to fake a click on a tab to make it paint the page?
[02:53:22] <jepler> ummmmm
[02:55:07] <jepler> I dunno
[02:55:14] <jepler> did my example *ahem* actually work for you?
[02:55:24] <jepler> I get an error when I run it
[02:55:39] <jmkasunich> I didn't try the example itself, I copied pieces of it
[02:56:23] <jmkasunich> the tk/bwidget notebook has a method (I think its a method) called "raise" that I believe makes a page active
[02:56:36] <jmkasunich> trying to figure out how to invoke it from py
[02:57:49] <jepler> notebook.raise_page(0)
[02:58:17] <jepler> and yes, if I don't do that, no page is displayed by default
[02:58:34] <jmkasunich> that did the trick
[02:58:46] <jmkasunich> I guess "raise" by itself is a reserved word in python
[02:58:53] <jepler> yes
[02:59:00] <jepler> it has to do with exceptions
[03:27:12] <jmkasunich> jepler: thanks for the help - the notebook is working perfectly now
[03:27:44] <jmkasunich> of course, I'm still flailing around trying to figure out how I want to manage the data, but thats my problem
[03:27:50] <jmkasunich> gonna sleep on it
[15:28:06] <jepler> oops, looks like I porked something up in pluto stepgen before I checked it in
[15:28:21] <jmkasunich> murphy
[15:29:01] <jepler> I have another copy that works, so it's just a matter of finding the mistake
[15:29:17] <jepler> I decided to change the number of whole and fractional bits in the step generator, and I assume that's the cause
[15:49:38] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/something-wrong.png
[15:50:09] <jepler> ideally the "tuned feedback loop" wouldn't be slamming the velocity between +-30ips every ms
[16:04:11] <jepler> aha, found it
[16:08:36] <jepler> hi skunkworks
[16:08:40] <skunkworks> jepler: if I did it right - the watchdog works.
[16:08:47] <skunkworks> Hi
[16:08:57] <jepler> skunkworks: great, thanks for testing
[16:09:19] <skunkworks> I have the index hooked up physically - I just need to muddle thru the hookups.
[16:12:49] <skunkworks> jepler: did you notice <Fritz> was wondering about the pluto_step?
[16:13:20] <skunkworks> he is not on right now.
[16:13:26] <jepler> yeah I saw that
[16:14:03] <jepler> I wonder if he reads the commits, or the axis blog
[16:17:08] <skunkworks> jepler: I wonder if 3 stepgen + encoder and pwm would be better?
[16:17:17] <skunkworks> if it fit
[16:17:37] <alex_joni> jepler: is it possible to use 2 plutos?
[16:17:59] <alex_joni> using 2 parports I mean
[16:18:19] <skunkworks> oh - that would be one hell of a machine controller...
[16:18:24] <skunkworks> Hi alex
[16:18:28] <alex_joni> hi samco
[16:19:40] <skunkworks> did you see my mess? http://www.electronicsam.com/images/KandT/servostart/ampmess.JPG
[16:20:08] <alex_joni> nice
[16:20:36] <skunkworks> poor little thing takes 10amps ;)
[16:21:21] <skunkworks> make for neat micro/mini mill servos.
[16:22:25] <jepler> alex_joni: you can't have two pluto_servo or two pluto_step, but you could probably mix
[16:22:37] <jepler> it would be a modest amount of work to support 2 boards of the same kind on separate ports
[16:22:42] <jepler> (and a pin renaming)
[16:22:59] <jepler> nothing subtle about it, just a bit tedious
[16:24:12] <jepler> 10A? that little thing?
[16:24:16] <skunkworks> stalled
[16:24:23] <skunkworks> gets warm ;)
[16:24:50] <skunkworks> I have no real specs on them - came off of some old line printers
[16:25:55] <skunkworks> 2 brush.
[16:27:12] <skunkworks> probably from the late 70s early 80s
[16:27:13] <alex_joni> nice
[16:28:01] <skunkworks> I am just thinking there are more people with 3 axis machines that would like spindle control..
[16:54:17] <skunkworks> or even 4 axis + pwm (no encoder)
[16:58:08] <skunkworks> or all of the above ;)
[17:16:14] <Skullworks-PGAB> you can only shoehorn so many hippies into a VW bug
[17:16:39] <Skullworks-PGAB> errr logic units into an FPGA
[17:36:36] <jepler> skunkworks: yeah, I think 3 stepgen + 1 PWM + 1 encoder would be a nice compromise for a lot of people
[17:36:46] <jepler> 4+1+1 is too big for the device
[17:43:18] <skunkworks> jepler: how hard is it to have a few different firmwares? one 4 stepper, 3 + encoder + pwm, 4 encoder + pwm?
[17:44:09] <skunkworks> sorry - meant 4 stepper + pwm (if it fits)
[17:57:01] <jepler> similar to what I said about more than one board: more tedious than subtle
[17:57:23] <jepler> more variations means more bugs, though
[20:28:47] <jmkasunich> need an outside opinion
[20:28:55] <jmkasunich> the 5i20 has 3 connectors with 24 I/O pins each
[20:29:18] <jmkasunich> should the hal numbering be 0-71, or 000-023, 100-123, and 200-223?
[20:29:42] <jmkasunich> the latter makes it easy to figure out what is the 17th signal on the 2nd connector
[20:29:58] <jmkasunich> but I dunno if that matters much to anybody
[20:31:09] <jepler> strangely enough they're labeled "P2, P3, P4" on the blard
[20:31:11] <jepler> board
[20:31:25] <jmkasunich> I've already decided that I'm not using those numbers
[20:31:27] <jepler> I think 200-223 is confusing
[20:31:45] <jepler> A00..A23, B00..B23, C00..C23?
[20:31:52] <jmkasunich> hmm
[20:31:59] <jmkasunich> that works
[20:32:34] <jepler> on the third hand, the silkscreen has the 0..71 numbering printed
[20:32:50] <jmkasunich> I like the 00-23 and repeat, because the LEDs on the 7i31 boards are numbered 00-23
[20:32:54] <jepler> just do something .. it'll be fine
[20:33:49] <jmkasunich> I think I'm gonna do the A00 thing
[20:35:44] <jmkasunich> decisions decisions
[20:36:02] <jmkasunich> do I name them blah.in-A00, or blah.A00-in?
[20:36:13] <jepler> just do something .. it'll be fine
[20:36:14] <jmkasunich> if the former, all inputs will sort together
[20:36:42] <jmkasunich> you're a lot of help
[20:37:31] <jepler> keep in mind that you can use wildcards in 'halcmd show xxx' so in that respect '*-in' or 'in-*' is just about as easy
[20:38:22] <jmkasunich> true, its better than before when you could only specify a prefix
[20:38:51] <jmkasunich> starting with the number will make it easy to see the "pinout"