#emc-devel | Logs for 2009-02-08

[00:08:15] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/ladder/images/watch_window.png: add image
[00:11:18] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/ladder/classic_ladder.lyx: add 2.3 image of watch window
[00:14:18] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: change papersides back to 1 ... didn't work out
[00:15:53] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/install/Latency_Test.lyx: change papersides back to 1 ... didn't work out
[00:18:05] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/hal/ (pyvcp.lyx rtcomps.lyx): minor edits
[00:21:33] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/ (main.lyx tool_compensation.lyx): getting started on tool comp
[00:22:30] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/images/ (comp01.dxf comp01.png): add file
[00:28:27] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/ (5 files): minor edit and format clean up
[00:32:24] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Getting_Started.lyx:
[00:32:24] <CIA-2> EMC: add usepackage{fancyhdr} to getting_started.lyx afaik it does not break
[00:32:24] <CIA-2> EMC: and a few minor edits
[00:32:27] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/config/ (emc2hal.lyx ini_config.lyx stepconf.lyx stepper.lyx):
[00:32:27] <CIA-2> EMC: add usepackage{fancyhdr} to getting_started.lyx afaik it does not break
[00:32:27] <CIA-2> EMC: and a few minor edits
[00:33:17] <BigJohnT> * BigJohnT is off to snuggle with a heating pad and a glass of wine to see which one makes my back feel better :)
[02:05:33] <cradek> yay, my primary machine is hardy now, and not much is broken
[02:08:08] <jmkasunich> you upgraded from dapper?
[02:08:12] <cradek> yeah
[02:08:21] <cradek> on my thinkpad
[02:08:28] <jmkasunich> just clicked on the button in the upgrade manager, or was it more complex?
[02:08:34] <cradek> replaced the hard disk with a compact flash
[02:08:43] <cradek> I did it from scratch
[02:09:00] <jmkasunich> so you didn't upgrade from dapper?
[02:09:12] <jmkasunich> you just installed hardy
[02:09:21] <cradek> yes I guess so
[02:09:30] <cradek> depends what you mean by "upgrade"
[02:09:33] <jmkasunich> don't confuse me ;-)
[02:09:44] <cradek> I don't think it was my doing
[02:10:14] <jmkasunich> upgrade from dapper means "take a disk that has dapper, and _upgrade_ it to hardy" not erasing it and installing hardy, or using a different disk
[02:10:32] <cradek> ok then
[02:10:40] <jmkasunich> the reason I asked is because this box (main dev machine, web server, compile farm,etc) is dapper, and I've thought about upgrading it
[02:10:59] <cradek> I have no experience with doing that
[02:11:38] <jmkasunich> right - you said that a while ago, and I thought that you might have just acquired said experience
[02:11:57] <cradek> no hard disk sound is disconcerting
[02:12:13] <jmkasunich> heh
[02:12:31] <jmkasunich> I've gotten used to this machine - the disk is quiet and the case is good
[02:12:34] <cradek> type :w, wait for disk to spin up, wait for file to save, alt-tab to AXIS, control-R
[02:12:35] <jmkasunich> I don't hear it
[02:13:02] <cradek> it will be strange for a while I think
[02:16:17] <jmkasunich> argh, I'm having a brain block
[02:19:22] <cradek> with what?
[02:20:00] <jmkasunich> my original algorithm for orientation is simple
[02:20:13] <jmkasunich> make a "panorama" from the camera image, and from the map
[02:20:20] <jmkasunich> panorama is wide and not very tall
[02:20:28] <jmkasunich> basically polar coords
[02:20:31] <jmkasunich> x = theta
[02:20:33] <jmkasunich> y = r
[02:20:46] <jmkasunich> bottom line of the image is a circle 9" out from the vehicle
[02:20:52] <jmkasunich> top line is a circle 22" out
[02:21:07] <cradek> ok, simple so far
[02:21:11] <jmkasunich> then do a correlation between the two panoramas, shifting along X till best match
[02:21:23] <jmkasunich> since x is theta, the shift at best match is the theta error
[02:21:36] <jmkasunich> works great, as long as there is no significant x or y error
[02:21:57] <jmkasunich> doesn't matter which way you are looking, if the error is in theta only, the image shift is the same
[02:22:08] <jmkasunich> but now suppose there is some x error (and no theta error)
[02:22:24] <jmkasunich> looking out the front, the image seems shifted to the left (if the vehicle is right of center)
[02:22:32] <jmkasunich> but looking out the back, the image seems shifted to the right'
[02:22:39] <cradek> so when you shift looking for best match, you wrap around in X
[02:22:48] <jmkasunich> yeah
[02:23:19] <jmkasunich> I make one of the panoramas cover 360 degrees, the other goes from -20 deg to +380 (or something like that)
[02:23:21] <cradek> what do you mean some x error?
[02:23:35] <cradek> do we have two Xes here?
[02:23:39] <jmkasunich> imaging you are sitting in a car
[02:23:51] <jmkasunich> the car is sitting still in the middle of the road
[02:24:06] <jmkasunich> if you turn the car (as on a turntable), it doesn't matter which way you are looking
[02:24:18] <jmkasunich> if the car turns CCW, what you see will appear to move to the right
[02:24:31] <jmkasunich> whether you are looking frontwards, backwards, left, right
[02:24:47] <jmkasunich> when I say "to the right", I mean "your right", regardless of which way you are facing
[02:24:55] <jmkasunich> follow so far?
[02:25:09] <cradek> I think so
[02:25:12] <jmkasunich> ok
[02:25:24] <jmkasunich> now imagine the car is doing 60 mph down the road
[02:25:41] <jmkasunich> if you look forward, you see little or no lateral movement of "things"
[02:25:58] <jmkasunich> if you look out the left window, things are moving from your right to your left (front to back of the car)
[02:26:21] <jmkasunich> if you look out the right window, things are moving from your left to your right
[02:27:11] <jmkasunich> also, when you look out the side windows, things nearby (telephone poles) are moving much faster than thing far away (mountains on horizon)
[02:27:22] <cradek> sure
[02:27:35] <cradek> in front, things are getting bigger too (back, smaller)
[02:27:42] <jmkasunich> those differences should be enough to let me distinguish between "turning" and "moving"
[02:28:00] <jmkasunich> yeah, but bigger/smaller is a less obvious effect
[02:28:13] <jmkasunich> and invisible when I do a 1-D correlation
[02:28:19] <jmkasunich> 2-D is much more expensive
[02:28:43] <cradek> what's the goal here? know how you moved since the last photo?
[02:28:48] <jmkasunich> yes
[02:28:57] <jmkasunich> actually, not quite
[02:29:03] <jmkasunich> I'm not comparing photo to photo
[02:29:08] <jmkasunich> I'm comparing photo to map
[02:29:38] <jmkasunich> I think I went 12" and turned 3 degrees, so I generate a map panorama of what I'd see in that case
[02:29:55] <jmkasunich> then I compare it to the camera panorama and realise that I went 13" and turned 4 degrees
[02:30:59] <jmkasunich> I'm pretty sure I want to split the panoramas into 4 quadrants, front, left, right, back (or maybe north, east, south, west)
[02:31:08] <jmkasunich> then I want to correlate in each quadrant individually
[02:31:18] <jmkasunich> and use the result to get both linear and angular displacements
[02:31:29] <cradek> does it mostly move forward and turn in place?
[02:31:40] <cradek> or will it be moving along curves etc
[02:31:50] <jmkasunich> yes, but the most critical measurement is lateral movement
[02:32:07] <jmkasunich> more than about 1-3/4" off centerline and a wheel will fall off the edge of the track
[02:32:31] <cradek> seems like you could do a correlation in X of the original picture, and a correlation in "new X" of the polar converted one
[02:33:01] <cradek> you don't have to do both operations on the same image - isn't lateral displacement easier before it's converted to panorama?
[02:33:30] <jmkasunich> yeah, sort of
[02:33:37] <jmkasunich> "before" is a misnomer
[02:33:53] <jmkasunich> the map starts out flat, so for that, panorama is a processing step
[02:34:02] <jmkasunich> the camera image is distorted by the fisheye mirror
[02:34:22] <jmkasunich> I make the panorama directly from the fisheye image, and then later (after I know the angular error) I can make a flat image
[02:34:54] <jmkasunich> my original plan (and maybe still) was to get orientation only from the panorama, then generate a flat image and do X and Y correlations (independent 1D , not 2D) for linear displacement
[02:35:19] <jmkasunich> the problem is that linear displacements can muck up the orientation measurement if I ignore them
[02:35:42] <cradek> hm, that's true
[02:36:19] <jmkasunich> if I'm driving down a straight "road", it isn't that hard to deal with
[02:36:40] <jmkasunich> if I'm turned 5 degrees but centered, the front and back quadrants will both be shifted 5 deg in the same dir
[02:36:58] <jmkasunich> if I'm off center (but not turned), the front and back views will shift in opposite directions
[02:37:20] <jmkasunich> so if I just average them, the off-center component should cancel out, and the angular component add
[02:37:32] <jmkasunich> problems happen when I am at a corner for example
[02:37:41] <jmkasunich> the back view might be featureless plain
[02:38:11] <jmkasunich> what I'm trying to wrap my head around is assessing not only the shift from the correlation,but also the "quality" of the correlation
[02:38:34] <jmkasunich> if there are few or no features in a quadrant, I can't trust the correlation much
[02:40:22] <jmkasunich> I've also been thinking about breaking the panorama into rings as well as quadrants
[02:40:51] <jmkasunich> take advantage of the "telephone pole vs mountain" effect, so even if the back view is featureless, the front can tell me the difference between rotate and shift
[02:40:57] <SWPadnos> it may be easier to look for certain shapes in the fisheye view
[02:41:15] <jmkasunich> why?
[02:41:16] <SWPadnos> lines and corners change shape as you pass by them in the fisheye
[02:41:32] <SWPadnos> if you're rotating, they don't
[02:42:21] <SWPadnos> you may also want to look up some object tracking code (I know there's some out there)
[02:42:37] <SWPadnos> that would help to find "features" in the image and track how they move
[02:42:43] <SWPadnos> even if they change shape a bit
[02:42:47] <jmkasunich> so I'd have to generate multiple maps (transform from map space to fisheye space, for several possible camera positions), then compare those maps to the camera image
[02:43:08] <SWPadnos> I don't think so
[02:43:15] <jmkasunich> then how?
[02:43:30] <SWPadnos> if you use object tracking of some kind, that will tell you how far each object has moved from one frame to the next
[02:43:42] <jmkasunich> btw, I can't really track objects or features from one frame to the next
[02:43:46] <SWPadnos> this would be comparing successive frames from the camera, not involving the map
[02:43:53] <jmkasunich> I'll be moving too fast for that
[02:44:03] <jmkasunich> target speed, 12-18 inches/second
[02:44:05] <SWPadnos> oh, how long between frames?
[02:44:33] <jmkasunich> frame rate, hard to say - no better than 10 fps, probably more like 3-5, worst case 1-2
[02:44:42] <jmkasunich> depends on how complex my code gets
[02:45:04] <SWPadnos> hmm
[02:45:10] <SWPadnos> on, I'll go back to lurking now :)
[02:45:11] <jmkasunich> an intersection, bend, etc (a node in the map) is 12" square
[02:45:42] <jmkasunich> so I'll only have a couple frames to recognise one when I approach it
[02:46:23] <jmkasunich> if I'm traveling down a long straight road (long is anything over 44", my view radius is 22"), then I won't have any camera knowledge about progress, only lateral and angular drift
[02:46:42] <jmkasunich> progress will be based on encoders only, until I see the next node
[02:47:27] <jmkasunich> I've been told that nodes are on a 2' grid, so when I see one, I will "round" it to the nearest grid point to correct my encoder measurement of progress down the road
[02:48:17] <jmkasunich> I think what I need to do is extend my correlation routine to give me some measure of the quality/strength of the correlation
[02:48:34] <jmkasunich> then break the panorama into quadrants and rings
[02:48:44] <jmkasunich> and look at the data to decide what to do with it
[05:09:54] <The_Ball> how many pulses per revolution on the spindle is recomended when converted a lathe so that thread cutting works well?
[05:10:04] <jmkasunich> that depends
[05:10:42] <jmkasunich> what is the maximum speed at which you want to thread (which in turn depends on the axis speed and accel limits and the range of thread pitches you want to cut)
[05:11:02] <jmkasunich> also, what is the maximum rate at which you can count - that depends on whether you are using software or hardware to count
[05:15:59] <The_Ball> I'll be using a pluto or a mesa card for hardware counting
[05:16:22] <jmkasunich> that makes it simpler, because they can count quite fast
[05:16:45] <jmkasunich> you'll need to read up on the card(s) to see exactly how fast (I think the mesa is faster)
[05:17:06] <jmkasunich> then choose an encoder ppr such that max spindle speed doesn't exceed the limit of the card
[05:17:38] <jmkasunich> for example, 1800 RPM = 30 revs/second, * 2500 ppr * 4 counts/cycle = 300,000 counts per second
[05:17:47] <jmkasunich> I think pluto can do that, and I'm sure mesa can
[10:44:42] <The_Ball> jmkasunich, i think the mesa does 2 or 20mhz, but what i was wondering was if there is a point where more ppr doesn't improve performance. i think i read that emc could do rigid tapping with only a single ppr?
[10:46:26] <alex_joni> that's somehow possible, but by all means not good
[10:46:34] <alex_joni> the more ppr the better
[10:46:42] <alex_joni> at some point price might be an issue
[10:49:57] <The_Ball> for the lath spindle i was hoping to make a disk which i drill holes in and use two ir light gates as the encoder, but i would only get 50 holes or so i would guess, so 200ppr
[10:50:39] <alex_joni> hmm.. the only problem with that approach is spacing the signals
[10:50:54] <alex_joni> although the mesa will probably be pretty tolerant
[10:51:31] <The_Ball> i'll use a scope to make adjustments to the second light gate so that it's at a 90 degree phase
[10:52:09] <The_Ball> might put a index pulse in there as well
[10:53:44] <The_Ball> oh sorry didn't see i was in #emc-devel and not #emc
[10:54:12] <alex_joni> The_Ball: it's ok ;)
[10:54:18] <alex_joni> although #emc is better
[12:48:03] <BigJohnT> morning alex_joni
[14:04:19] <CIA-2> EMC: 03tissf 07TRUNK * 10emc2/src/po/ (fr_axis.po fr_rs274_err.po): french translation update
[14:27:58] <cradek> it is impossible to do rigid tapping with a single ppr
[14:29:25] <jepler> because typically you'll be a half revolution off at guessing where the true reversal took place
[14:54:05] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/images/ (comp02.dxf comp02.png): add files
[15:05:19] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/tool_compensation.lyx: a bit more on comp... I'm getting there
[15:18:21] <cradek> yay
[15:19:30] <cradek> hm, I just realized that I can't edit the docs anymore. I didn't think of that consequence.
[15:22:41] <BigJohnT> hey cradek is it an error to program a G0/1 without an X or Y after a G41/42?
[15:23:26] <cradek> do you mean something like g41 g1 z-1?
[15:23:34] <BigJohnT> yep
[15:23:51] <BigJohnT> at least that's what I remember from yesterday :)
[15:24:43] <cradek> I don't think it's an error, but I'm trying to decide whether I think it's a bad idea
[15:24:49] <cradek> did you try it?
[15:25:02] <BigJohnT> yep
[15:25:11] <cradek> does it error?
[15:25:26] <BigJohnT> I think I did a G42 one sec I'll find it
[15:26:26] <BigJohnT> Is it by design that I get an error if I program a G41 G1 on the same line without any X or Y?
[15:26:29] <BigJohnT> 12:41:05 <BigJohnT> The error is "Lenght of cutter compensation entry move is not greater than the tool radius"
[15:26:36] <BigJohnT> from yesterday
[15:26:43] <cradek> ah
[15:27:09] <cradek> if you tried it and got an error, yes, it is an error :-)
[15:27:23] <BigJohnT> LOL ok then I'll document it
[15:27:25] <cradek> do you think it shouldn't be?
[15:27:31] <cradek> I can't decide
[15:28:01] <BigJohnT> if your not moving in X or Y it doesn't seem like your doing an entry move
[15:28:14] <BigJohnT> It suprised me anyhow
[15:28:57] <cradek> well you can exit in Z only. the move becomes diagonal and takes you (in XY) from the comp to the nominal path
[15:29:18] <cradek> it seems like you could enter that way too, but I think you would always gouge the part on the way down
[15:29:30] <alex_joni> cradek: I have a dapper VM still around for working on the docs
[15:29:44] <cradek> alex_joni: I guess I have other machines too
[15:29:57] <alex_joni> it should be fixable for 2.4.x
[15:30:38] <alex_joni> jepler reported the lyx guys fixed the conversion package, so we can probably use hardy in the future
[15:30:41] <BigJohnT> it does follow the rule that the first non rapid move must be at least the cutter radius...
[15:31:06] <cradek> non-rapid isn't part of the rule - rapids can be entries
[15:31:39] <BigJohnT> yep... the last rapid or first non rapid :)
[15:31:51] <cradek> oh about your bug report
[15:32:03] <BigJohnT> is it a feature bug?
[15:32:10] <cradek> suppressing the initial rapids is a "feature" that I wrote on purpose
[15:32:14] <cradek> here's why
[15:32:14] <alex_joni> a buggy feature?
[15:32:31] <cradek> say I have a lathe part that fits in a 1" cube
[15:32:41] <cradek> the turret moves a foot to the right to change tools
[15:33:02] <cradek> without suppressing those initial rapids, the preview is over a foot wide, with the extents showing 13"
[15:33:21] <cradek> the actual part of the program I'm interested in seeing is a fly speck when zoomed to show everything
[15:33:36] <BigJohnT> I can see where that is a problem
[15:33:41] <cradek> so, that was the "fix"
[15:34:08] <cradek> I like the preview results a lot better, but I also would like to see the entry rapids sometimes. I don't know how to solve both problems.
[15:34:14] <BigJohnT> doesn't show/hide also do that?
[15:34:25] <alex_joni> this sounds like a need for configuration..
[15:34:28] <BigJohnT> or hide/show
[15:34:37] <alex_joni> show rapid extents on/off
[15:35:18] <alex_joni> otoh, too many configuration options only lead to confusion
[15:35:47] <BigJohnT> what happens when you use (AXIS,hide) and (AXIS,show) to wrap the tool change code?
[15:35:50] <cradek> with a tool change position and large tool offsets, there are rapids scribbled *everywhere*
[15:36:29] <cradek> BigJohnT: that could work, but I don't want to leave the burden of getting a good preview on every gcode programmer
[15:37:15] <BigJohnT> it does kinda confuse someone on a mill that might have a couple of rapids to go around a clamp or something
[15:38:39] <cradek> yeah, a mill will have long entries too, not just a lathe thing
[15:40:24] <BigJohnT> is it the same behavior outside of a comp move?
[15:43:36] <cradek> is what?
[15:46:27] <jepler> alex_joni: yes, but the "fix" involves substantial effort that I don't want to do for 2.3.
[15:46:43] <alex_joni> jepler: that's why I said for 2.4
[15:46:49] <jepler> alex_joni: ok
[15:46:54] <jepler> besides, we won't support dapper for 2.4 anyway
[15:47:04] <alex_joni> I'll try to help with the substantial effort :)
[15:48:33] <BigJohnT> my question didn't make any sense :)
[15:48:53] <BigJohnT> had to try and find Mom in law's keys
[15:49:05] <cradek> now that's multitasking
[15:49:14] <BigJohnT> yea
[15:49:26] <alex_joni> were they behind an email?
[15:50:42] <BigJohnT> anyway I "expected" to see the rapids and when they disapeared with 2 or more I was more confused than normal
[15:51:07] <BigJohnT> sounds like she found the keys but I don't speak spanish
[15:51:11] <ehj> Alex: Is there any doc on how to make binary packages for EMC (to go with rtaismp of course)?
[15:51:33] <alex_joni> ehj: debian/configure -r
[15:51:36] <alex_joni> debuild
[15:51:47] <cradek> BigJohnT: yeah the rule is all rapids between a tool change (or TLO change) and the first feed are supressed
[15:52:00] <ehj> k, thanx
[15:52:08] <BigJohnT> ahh, ok
[15:52:36] <cradek> BigJohnT: might be nice if the docs show feed moves instead of rapids for entry... but up to you.
[15:53:13] <BigJohnT> that's what I'm working on now and yes to keep the confusion down I'm using feeds for everything :)
[15:54:01] <BigJohnT> cradek: is it all rapids if more than one on the tool change?
[15:55:29] <ehj> Alex: debian/configure -r goves error "your kernel '2.6.24-16-rtaismp' is not known. There might be needed dependencies which won't get set automatically.
[15:55:37] <ehj> *gives
[15:56:30] <jepler> this mostly affects whether the Build-Depends line will be correct in your source package
[15:56:37] <cradek> BigJohnT: you mean like the quill up and reference point move stuff? yes those are all rapids
[15:56:41] <jepler> edit debian/configure to add lines for your kernel if this is important to you.
[15:57:00] <BigJohnT> ok, thanks
[15:58:44] <cradek> "If several rapid moves are issued after a G41/42 only the last one will move the tool to the compensated position." this is not true - the first move after g41/g42 does the full entry
[15:59:56] <BigJohnT> if you do rapids in the middle like from part to part you get that I think... must have got confused
[16:00:41] <cradek> that's confusing. the endpoints of the rapids are still on the compensated path, but it may go through corners to get there.
[16:01:00] <BigJohnT> yes
[16:03:48] <cradek> bbl
[16:04:00] <ehj> Alex: Configure worked, where is 'debuild'?
[16:04:00] <BigJohnT> ok
[16:05:05] <ehj> Alex: Install devscripts?
[16:11:29] <alex_joni> apt-cache search debuild
[16:11:57] <ehj> alex: I think I got it, there is just a ton of things to install (325MB).
[16:14:21] <ehj> alex: BTW, I started a wiki on this, here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_With_Custom_Kernel
[16:15:01] <alex_joni> ehj: ok, nothing wrong with that
[16:15:11] <alex_joni> it'll only be outdated in 6 months though :D
[16:15:21] <ehj> Yea, I figured.
[16:15:24] <alex_joni> I think we have 10? pages on building a rtai kernel already
[16:15:31] <alex_joni> doesn't hurt to have another one :P
[16:15:47] <BigJohnT_> BigJohnT_ is now known as BJT-Hardy
[16:16:39] <BJT-Hardy> anyone understand the gouging error with this short g code? http://pastebin.ca/1330802
[16:21:01] <ehj> Alex: I may have missed it, but I did not see any wikis showing how to build the kernel from the linuxcnc repositories, particularly the part about editing files under debian/binary-custom.d/...
[16:21:37] <alex_joni> ehj: well, that's new
[16:21:53] <ehj> Yea, makes things a lot easier.
[16:22:22] <alex_joni> doesn't work on anything besides ubuntu sources from git
[16:23:27] <ehj> Yea, I understand that, but then I expect that would be the way many would want to do it. At least for taking advantage of features particular to specific processors.
[16:37:03] <ehj> Alex: A tar.gz and dsc file was created, but also got a fatal error after getting an error "no convert, documentation cannot be built.
[16:37:46] <ehj> Do I need to install LyX?
[16:38:38] <ehj> No, LyX is already installed.
[16:39:17] <SWPadnos> imagemagick
[16:39:38] <ehj> k
[16:39:44] <SWPadnos> and you'll probably see another dependency, maybe dvips, after you install imagemagick
[16:39:56] <BigJohnT> and dvipng and groff I think...
[16:40:09] <BigJohnT> according to my notes
[16:47:19] <ehj> Got past that, now missing epstopdf. I was not able to find it through synaptic.
[16:52:49] <jepler> http://packages.ubuntu.com/search?searchon=contents&keywords=epstopdf&mode=exactfilename&suite=hardy&arch=any
[16:59:57] <ehj> Looks like it got past the dependencies related to documents. Thanx.
[17:23:32] <ehj> Got to now signing changes. gpg: skipped ... secret key not available, clearsign failed, fatal error, running debsign failed.
[17:24:52] <cradek> BJT-Hardy: nope, I don't understand it
[17:24:59] <cradek> it might be a bug
[17:27:38] <cradek> noooooo they changed xemacs to use fuzzyfonts
[17:28:52] <jepler> cradek: oh no!
[17:30:24] <jepler> cradek: did you get emacs22-gtk or emacs22? emacs22 doesn't depend on xft, so maybe it's not fuzzyfonts
[17:30:49] <cradek> xemacs21
[17:31:37] <cradek> also, http://timeguy.com/cradek-files/emc/lovely-dired.png
[17:32:20] <jepler> odd
[17:32:29] <jepler> try not scrolling down that far
[17:32:39] <cradek> **sigh**
[17:33:35] <alex_joni> ehj: signing shouldn't be an issue
[17:35:13] <ehj> k, it looked like it built the debs. I see that it built debs for both binaries and source. I assume I will be able to just install the binaries.
[17:45:24] <jepler> cradek: I get bitmappy fonts in xemacs21 .. http://emergent.unpy.net/index.cgi-files/sandbox/emacsen.png
[17:45:31] <jepler> nothing in .emacs.d but auto-save-list
[17:46:09] <jepler> I do get the crap at the end of dired
[17:47:40] <jepler> cradek: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=399483
[17:48:47] <jepler> the elisp in there fixed that little problem for me
[18:22:09] <cradek> jepler: aha, my real problem was a font path issue...
[19:41:02] <jepler> I suppose there's no strtod in rtapi
[21:05:25] <BigJohnT_> BigJohnT_ is now known as BigJohnT
[21:16:02] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/images/comp02.dxf: update base drawing
[21:17:18] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/tool_compensation.lyx: remove incorrect statement on rapid moves
[21:19:20] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Getting_Started.lyx: *** empty log message ***
[21:19:45] <BigJohnT> dang log messages
[21:24:55] <alex_joni> who ate it?
[21:26:41] <cradek> BigJohnT: I think the gouge error when you use rapids like that is a bug - thanks for the program that reproduces it so nicely
[21:26:49] <cradek> I have not found it yet though.
[21:27:48] <BigJohnT> ok cool
[21:28:34] <BigJohnT> dunno alex_joni but it automagically disappeared
[21:30:04] <BigJohnT> I wonder if the cat walking across the keyboard while I commited had anything to do with it :)
[21:41:41] <SWPadnos> hey, anyone need opto-22 /O boards?
[21:42:10] <jmkasunich> I thought you did?
[21:42:25] <SWPadnos> well, I can get a lot of 23 of them, and I don't need that many
[21:42:31] <jmkasunich> ah
[21:42:44] <SWPadnos> the I/O modules are in the $1-2 range on eBay, and the boards are ~$25
[21:42:55] <SWPadnos> so it's actually less expensive than a 7i37T + terminals
[21:43:08] <SWPadnos> and you get completely flexible I/O configuration
[21:43:14] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Master_User.lyx: nice header/footer IMHO
[21:43:15] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/ (axis.lyx image-to-gcode.lyx tkemc.lyx): nice header/footer IMHO
[21:43:15] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/ (Glossary.lyx user_intro.lyx): nice header/footer IMHO
[21:43:16] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/main.lyx: nice header/footer IMHO
[21:44:42] <SWPadnos> I'd just get them and hope to sell a few, but with the CNC workshop up in the air ...
[21:51:58] <jepler> I don't need any
[21:52:12] <jepler> cradek might for his mill conversion, but I wouldn't want to speak for him
[21:52:28] <jmkasunich> sell them on ebay
[21:52:42] <SWPadnos> the guy is selling them on ebay already :)
[21:52:50] <jmkasunich> I guess $25 x 23 is a non-trivial investment
[21:53:02] <SWPadnos> $575
[21:53:09] <jmkasunich> he's only selling them as a lot of 23, right?
[21:53:22] <SWPadnos> no, I can probably get fewer, and maybe at a lower cost too
[21:53:23] <jmkasunich> I bet most people are like you, don't need that many
[21:53:25] <SWPadnos> yeah
[21:53:32] <jmkasunich> oh, then just get what you need
[21:53:54] <SWPadnos> unless I decide to trigger the cameras with them. in that case I can make a killing reselling them :)
[21:54:01] <jmkasunich> if he was only selling them as a lot, then you could turn around and sell in smaller quantities (at a somewhat higher price)
[21:54:36] <SWPadnos> yep. I need to email him back, I just wanted to see if anyone here thought they might want some so I could get a better estimate of quantiry
[22:02:18] <ehj> swp: There is a complete set of debs on my ftp site for smp, including packages for installing EMC and separate packages for firmware (Mesa).
[22:02:52] <ehj> Let me know if you don't remember the URL, user name or password.
[22:02:57] <SWPadnos> ok, thanks
[22:04:54] <jmkasunich> ehj: these are new since I downloaded them friday?
[22:05:13] <jmkasunich> you mentioned something about rebuilding...
[22:05:17] <alex_joni> hi guys
[22:05:18] <ehj> Yes, ijncluding turning off the switch I mentioned earlier.
[22:06:13] <jmkasunich> I got my DC input supply yesterday, just hooked it up
[22:06:56] <jmkasunich> 2.3A at 12V (including hard disk)
[22:06:58] <jmkasunich> not bad
[22:07:58] <ehj> Did you get the 4 Pin 12V adapter (obviously you did if you have it running)?
[22:08:03] <jmkasunich> yes
[22:08:13] <jmkasunich> the ebay guy I got my PS from throws that in free
[22:09:10] <ehj> Cool. I ordered both from a different supplier just because they had that adapter. I was too cheap to pay shipping twice. :)
[22:09:19] <jmkasunich> yeah
[22:09:33] <jmkasunich> the PS had free shipping ;-)
[22:09:47] <ehj> That's not bad.
[22:10:09] <jmkasunich> M2-ATX, with all cables for $72 delivered
[22:11:06] <ehj> I will have to check what I paid for shipping, mine was like $62 + $2 for the adapter + shipping.
[22:11:34] <jmkasunich> wow
[22:11:47] <jmkasunich> the going rate seemed to be mid to upper 70's when I was looking
[22:11:57] <ehj> That was the 12-32 VDC PS.
[22:12:16] <SWPadnos> ok, after finding the second password, all the files are transferred :)
[22:12:17] <jmkasunich> I think the 6-24 was about the same
[22:12:18] <SWPadnos> thanks
[22:12:26] <SWPadnos> the new debs are in /experimental/smp/
[22:12:42] <ehj> Yea, they were within a dollar or two of the same price.
[22:12:46] <fenn> what happens if you exceed the 150W rating?
[22:13:00] <jmkasunich> fenn: I'm not worried
[22:13:07] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Master_Integrator.lyx: nice headers and footers :)
[22:13:07] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/integrator_intro.lyx: nice headers and footers :)
[22:13:08] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/config/ (ini_config.lyx stepper.lyx): nice headers and footers :)
[22:13:09] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/drivers/opto22.lyx: nice headers and footers :)
[22:13:09] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gui/halui.lyx: nice headers and footers :)
[22:13:11] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/hal/ (drivers.lyx pyvcp.lyx rtcomps.lyx): nice headers and footers :)
[22:13:14] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/ladder/ (classic_ladder.lyx ladder_examples.lyx ladder_intro.lyx): nice headers and footers :)
[22:13:18] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/install/installing_emc2.lyx: nice headers and footers :)
[22:13:19] <fenn> jmkasunich: i want to run a desktop motherboard on one of those
[22:13:22] <jmkasunich> fenn: I'm drawing 27.6 watts right now
[22:13:38] <ehj> Likewise, only powering an IDE flash drive and either 7i43 or 5i20 controller.
[22:14:10] <jmkasunich> if you overload it, either something on the PS will release its magic smoke, or the supply will shut down on overload
[22:14:13] <ehj> swp: Great.
[22:14:14] <jmkasunich> no telling which one
[22:15:17] <jmkasunich> wires are gonna be a pain on this robot
[22:17:11] <ehj> Yea, one of the reasons I wanted to use the 7i43, I put everything in a travla panel mount enclosure. I have already closed the wires up on the CPU fan and had it shut down on thermal overload.
[22:20:11] <jmkasunich> my problem is that the entire robot isn't very big
[22:20:18] <jmkasunich> the mobo is the widest part
[22:20:43] <jmkasunich> we're planning on mounting it on an angle at the back, so the camera can see over it
[22:20:58] <jmkasunich> the wires are gonna stick into the camera view more than any other part of the vehicle
[22:23:17] <ehj> Those PS's are high $50 on down on mini-box, see: http://www.mini-box.com/DC-DC
[22:24:44] <jmkasunich> M2-ATX is $79.50 on mini-box.com
[22:24:45] <SWPadnos> jmkasunich, stick it on top of the mirror
[22:25:11] <jmkasunich> SWPadnos: the whole thing has to fit thru a 12" high opening at the start and finish
[22:25:20] <jmkasunich> if it could be higher, I'd move the mirror up
[22:25:21] <SWPadnos> oh right. forgot about the height
[22:25:41] <ehj> ah, I know you said it above, but thought you were using one more like mine. Mine on mini-box is $59.50.
[22:26:08] <jmkasunich> one of the "pico" ones?
[22:27:34] <ehj> Yes, this one: http://www.mini-box.com/PicoPSU-80-WI-32V?sc=8&category=981
[22:27:50] <jmkasunich> was just looking at that
[22:28:18] <jmkasunich> since I'm using a single 12V battery (motorcycle batt), I need to run at less than 12V
[22:28:38] <jmkasunich> the pico ones either don't do that, or are very marginal on 12V output
[22:28:47] <ehj> Yea, most of them are made for 12 Batteries, to run in automobiles.
[22:29:05] <jmkasunich> the pico's really _aren't_ made for auto/battery
[22:29:27] <jmkasunich> things like starting the engine would droop the battery enough to make the supply drop out
[22:29:31] <SWPadnos> they have some that don't regulate 12V, and others that do
[22:29:56] <jmkasunich> SWPadnos: the ones that do, either expect at least 13V or so input, or have very limited 12V output
[22:29:58] <SWPadnos> the computers aren't meant to run when the ignition is off
[22:30:05] <jmkasunich> IOW, they buck the 12V only, not boost it
[22:30:25] <SWPadnos> there are "smart power controllers" that sense when the ignition is on, and once power has been stable for some time they turn on
[22:30:27] <SWPadnos> sure
[22:30:33] <jmkasunich> right - when the engine is running, the battery will do a nice 13+ V
[22:30:49] <jmkasunich> but I don't have the luxury of an alternator ;-)
[22:30:53] <SWPadnos> heh
[22:30:59] <jmkasunich> the M2-ATX is one of those smart ones
[22:31:09] <jmkasunich> 6-24V, plus sequencing options
[22:31:11] <SWPadnos> nor the, uh, problem of a starter
[22:31:51] <jmkasunich> yeah, but I do have a couple modest motors, and its a small battery that might already be a bit low by the time the run starts
[22:32:15] <jmkasunich> an extra $20 to ensure that it will work down to 6V seemed prudent
[22:32:23] <SWPadnos> oh, definitely
[22:33:14] <jmkasunich> I bet I could use this board as my webserver, and the UPS time would go from 15 mins to several hours
[22:33:54] <SWPadnos> heh. days (since there would be nothing connecting to it during a power outage)
[22:34:19] <SWPadnos> especially if you configure it to power down the HD when idle
[22:34:30] <jmkasunich> well, one thing about this board - the power consumption is not very load dependent
[22:34:44] <jmkasunich> the chipset draws more (much more) than the CPU, and that doesn't vary much
[22:34:54] <SWPadnos> not running the RT kernel
[22:35:00] <SWPadnos> ah, true
[22:35:11] <jmkasunich> I think the CPU is ~3W, and the chipset is close to 20
[22:35:28] <SWPadnos> power would be reduced if you ran it headless (or at least don't load X)
[22:35:30] <jmkasunich> right now I'm using a laptop drive, thats probably only about 3W
[22:35:42] <jmkasunich> I don't think so
[22:35:56] <jmkasunich> I've been watching the current needle on my bench supply
[22:36:06] <alex_joni> SWPadnos: don't think there's an option to disable accel stuff form the BIOS for graphics
[22:36:12] <alex_joni> s/form/from/
[22:36:15] <jmkasunich> it is virtually constant, even during BIOS and text mode boot time
[22:36:20] <SWPadnos> if you don't use it, it's close enough to disabled
[22:36:28] <SWPadnos> interesting
[22:36:54] <SWPadnos> uh-oh. the wife's home, time to go clean some stuff. bbiab
[22:36:56] <SWPadnos> :)
[22:39:36] <BigJohnT> she back from fishing?
[22:41:47] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/ (4 files): update axis image
[22:50:41] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Master_HAL.lyx: remove duplicate chapters
[22:50:42] <CIA-2> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/hal/intro.lyx: remove duplicate chapters
[22:58:46] <CIA-2> EMC: 03cmorley 07TRUNK * 10emc2/src/hal/classicladder/calc.c: Add define so Cl compiles with 2.6.27
[23:08:47] <tissf> hi all
[23:09:24] <tissf> John, one for you:
[23:09:35] <tissf> francis@aldi2:~/emc2.TRUNK/src$ make
[23:09:37] <tissf> ../docs/src/Submakefile:389: *** Required image file /home/jet/emc2-trunk/docs/src/gui/images/axis_2.3.png does not exist. ArrĂȘt.
[23:10:15] <tissf> ? ;)
[23:20:11] <tissf> this is in the V1.11 2009/02/08 22:41:33 of Getting_EMC.lyx
[23:27:46] <tissf> and elsewhere !
[23:29:01] <tissf> in integrator_intro.lyx
[23:33:37] <tissf> in user_intro.lyx
[23:51:07] <CIA-2> EMC: 03tissf 07TRUNK * 10emc2/docs/src/common/ (Getting_EMC.lyx integrator_intro.lyx user_intro.lyx): fix axis_2.3.png absolute link
[23:51:09] <CIA-2> EMC: 03tissf 07TRUNK * 10emc2/docs/src/common/images/axis_2.3.png: fix axis_2.3.png absolute link
[23:51:10] <CIA-2> EMC: 03tissf 07TRUNK * 10emc2/docs/src/gcode/main_fr.lyx: fix axis_2.3.png absolute link