#emc | Logs for 2008-01-17

[00:45:04] <ds2> another complain about DBUS and Bluetooth - http://gnumonks.org/~laforge/weblog/2007/
[00:45:20] <Gamma-X> is there a tutorial on hookin up the wires to a motion control card like the mesa. Im going to label the wires on where they are supposed to go on the mesa cards before i put them in.
[00:45:24] <ds2> blah
[00:45:27] <ds2> n/m
[01:01:08] <BigJohnT> Gamma-X this is a real good chart for machining and materials http://www.niagaracutter.com/techinfo/millhandbook/speedfeed/chart1.gif
[01:08:02] <jepler> and to help you convert those into F and S numbers: http://timeguy.com/cradek/01198949282
[01:12:20] <BigJohnT> jepler: love them charts, real simple to read
[01:45:26] <dmess> hi all anyone familiar with a mazak ajv32/405 machine.... cam2 control i believe?? im looking for a manual... anything with specs etc...
[01:46:52] <dmess> theres 1 in a warehouse... under power..but no one knows how to move it OFF it limit....
[01:48:56] <cradek> ha
[01:49:05] <fenn> ratchet
[01:49:27] <cradek> pipe wrench
[01:51:59] <toastydeath> anyone have a lathe with a c axis
[01:53:54] <toastydeath> or ever program one
[02:07:57] <dmess> yes c axis lathe....
[02:09:30] <toastydeath> like
[02:09:45] <toastydeath> how would you turn an oval
[02:10:13] <toastydeath> we have an application where we want to move the X axis based on the C axis position, with the C axis turning at a constant speed
[02:10:39] <toastydeath> and I don't even know where to start in terms of getting that programmed
[02:11:49] <skunkworks> lots of short line segments.. between c and x
[02:11:53] <skunkworks> ?
[02:12:19] <toastydeath> this is going to be done by hand, unfortunately
[02:12:25] <toastydeath> so i was hoping there was some elegant math solution
[02:12:32] <fenn> use HAL
[02:12:36] <toastydeath> not in emc
[02:12:42] <fenn> well, you're screwed then
[02:12:48] <skunkworks> basic - or python..
[02:12:54] <toastydeath> it's some control that can do it
[02:13:20] <skunkworks> spreadsheet?
[02:13:27] <toastydeath> hurr
[02:13:37] <toastydeath> see, the C axis is going to be spinning at 50-100 rpm
[02:13:49] <toastydeath> and is constant, so i don't know how that will go over as short line segments
[02:14:04] <toastydeath> but that is definately an idea
[02:14:17] <toastydeath> (the spreadsheet/scripting idea)
[02:14:20] <SkinnypuppY34> 50-100 errr that would prob excede xfeedrate
[02:14:29] <toastydeath> it's a 20" spindle nose
[02:14:54] <toastydeath> and the X axis doesn't have to move much
[02:15:04] <toastydeath> it's a very large radius we're trying to machine on a block
[02:15:36] <toastydeath> 4" ballscrews and big feed motors, so I'm hoping we can get at least a decent feed out of it
[02:15:50] <toastydeath> it's an ooollld lathe though
[02:16:03] <skunkworks> I don't concider that a c axis then.. but I am just a armchair cncer.. I could see doing it with spindle synced motion in emc..
[02:16:10] <skunkworks> maybe
[02:16:23] <toastydeath> yeah, that might be the right way to do it
[02:17:05] <SkinnypuppY34> I've done some flats on a haas live tool but not cam or oval shaped
[02:17:16] <toastydeath> this isn't live tool, that's the thing
[02:17:27] <fenn> toastydeath: you could use subroutines..?
[02:17:27] <toastydeath> it's a boring operation
[02:17:37] <toastydeath> how would it work with the synchronized motion
[02:18:07] <toastydeath> can you synchronize an in/out X pulse per revolution?
[02:18:20] <fenn> i dont think so
[02:18:25] <fenn> i think it's only linear movements
[02:18:28] <toastydeath> yeah
[02:18:35] <fenn> you could hack something together in hal with siggen though
[02:18:57] <fenn> um, wait not siggen
[02:19:39] <fenn> oh, i guess you could set frequency to 0 and input C to siggen.offset
[02:20:06] <toastydeath> i'm pretty sure this control has the ability to program asymmetrically
[02:20:10] <toastydeath> i just don't know how to do it math-wise
[02:20:20] <toastydeath> the cam programs for it are really expensive and are hard to find
[02:20:28] <fenn> isnt an ellipse just a sinusoidal offset in polar coords?
[02:20:34] <toastydeath> i believe so
[02:21:14] <toastydeath> the one vain hope i have is that this machine apparently reads DXF files directly
[02:21:19] <fenn> so x = (major-minor)*sin(c) + minor
[02:21:25] <toastydeath> ooh.
[02:22:10] <toastydeath> oh wait, that works for an ellipse, but what we are doing is a bigger bore
[02:22:12] <toastydeath> than the lathe can handle
[02:22:19] <toastydeath> 50"-100" radius
[02:22:27] <toastydeath> in one side of a block
[02:22:41] <toastydeath> there's got to be an adaptation of that
[02:23:06] <fenn> http://mysite.du.edu/~jcalvert/math/ellipse.htm
[02:23:39] <toastydeath> hurr
[02:23:40] <skunkworks> with emc - you can do a bunch of short movements synced to the spindle.. Which I assume is something others can do.. so you would have a bunch of movements that blend togather to make the shape you want. (threading on steroids) It would be something to ask cradek.. I think he has a really good understanding of it as he wrote it. :)
[02:23:51] <toastydeath> the issue is that C is not programmable
[02:24:17] <fenn> if c isn't programmable then it's just a spindle, not an axis
[02:24:18] <toastydeath> i'm just going to have to wait and see
[02:24:48] <fenn> but basically, if you're asking hacky questions about a control nobody's ever used...
[02:24:50] <toastydeath> for whatever reason, asymmetric programming treats it as a c axis
[02:25:24] <toastydeath> and i was hoping someone who had used a lathe with asymmetric capability could shed some light on it because i've never looked at that kind of program before
[02:25:35] <fenn> skunkworks: hmm can you do arcs?
[02:27:42] <skunkworks> The way I understand it.. If you don't go faster than the axis can move.
[02:28:02] <skunkworks> (at least in my head)
[02:28:37] <skunkworks> but like I say - cradek is the one to ask
[02:28:53] <fenn> i think emc tries to go as fast as the spindle moves (ignoring joint constraints) (maybe)
[02:29:51] <fenn> uh, no that wouldnt' work because hal stuff has accel/vel constraints too
[02:30:29] <fenn> i'm not sure which is worse - violate joint constraints or break the part (and tool)
[02:30:40] <skunkworks> http://www.linuxcnc.org/docs/2.2/html/gcode_main.html#sec:G33,-G33.1:-Spindle-Synchronized
[02:31:37] <fenn> it's harder to think about stuff where your synchronized axis is a short linear move for less than one rotation
[02:31:58] <skunkworks> It is an error if:
[02:32:09] <skunkworks> the requested linear motion exceeds machine velocity limits due to the spindle speed
[02:32:27] <fenn> yes but that's in the interp
[02:32:33] <fenn> the spindle may or may not be going the requested speed
[02:33:13] <skunkworks> hmm - maybe I don't understand what you mean
[02:34:47] <fenn> your motors can do 10ipm: S1000 G33.1 Z-1 K0.1
[02:34:58] <skunkworks> I would assume you would get a joint following error if the axis can't keep up
[02:35:27] <fenn> uh, yes, but that's not what it means
[02:35:38] <fenn> you will get the error before the program runs
[02:36:31] <fenn> my example would request a 100ipm move right?
[02:36:52] <skunkworks> ah - I see what your saying - I think. If you say s1000 - doesn't mean that your spindle is going to go that speed. so the first check (interp) may allow it to run - but the spindle may actaully be going faster.
[02:37:32] <fenn> well, the interp would error in this example because the motors cant spin 100ipm
[02:38:14] <skunkworks> right
[02:38:19] <fenn> if you change it to S90 then interp lets it through, but the spindle hasn't slowed down yet and so you get a joint following error
[02:38:38] <fenn> (maybe)
[02:38:41] <skunkworks> I see
[02:44:37] <fenn> unfortunately the man page is pretending to be some sort of language spec, and doesnt explain how emc2 actually does stuff
[02:46:52] <fenn> i would like to see an elliptical thread though :)
[02:47:03] <fenn> DIY forming tap
[02:48:24] <toastydeath> somebody needs to put emc on a tool and cutter grinder
[02:55:47] <dmess> gimme a grinder...
[02:57:26] <dmess> eliptical thread FORM?? or thread on an elipse??
[03:10:29] <fenn> thread on an ellipse
[03:26:50] <Gamma-X> anyone with mesanet cards do u use the power coming from the 5i20 or an external power connecter?
[03:41:07] <fenn> you need both
[03:41:39] <fenn> power from the 5i20 drives the opto's on one side of your 7i37, and power from elsewhere drives the rest
[03:45:46] <Gamma-X> fenn how would u jumper it properly then?
[03:46:26] <Gamma-X> with the jumper its either or.
[03:47:15] <fenn> i have nfc
[03:48:23] <jepler> the jumper selects the power on the non-isolated side of the 7i37 board -- from pin 49 (or is it 50?) of the 50-pin connector, or from the 4-pin connector.
[03:50:15] <jepler> the isolated side doesn't require a "power supply" as such -- current on the inputs turns on the input optoisolators, and the outputs switch whatever loads you attach.
[03:57:05] <jepler> 'night all
[04:04:23] <Gamma-X> jepler are u using the mesa cards?
[04:06:09] <fenn> he's not, and he's not here
[05:50:41] <juanjo_> juanjo_ is now known as juanjo
[06:45:38] <tranzient> tranzient is now known as notranc
[10:36:42] <micges> hello all
[10:37:09] <micges> alex_joni: have moment ?
[10:43:44] <alex_joni> micges: yeah
[11:02:17] <alex_joni> micges: what's up?
[11:30:38] <micges> my translations doesnt work :(
[11:31:12] <alex_joni> how did you try it?
[11:31:42] <micges> my .mo files which I copy to LC_MESSAGES doesnt work
[11:31:52] <micges> ok
[11:32:01] <micges> I run make emc
[11:32:15] <alex_joni> can you send me the file to see if it works here?
[11:32:24] <alex_joni> do you run emc installed or run-in-place?
[11:33:27] <micges> file is rs274_err.mo which has translated interpreter messages
[11:34:28] <micges> ok from begining
[11:34:33] <alex_joni> the .mo is the translated file
[11:34:44] <alex_joni> but you need to have the language designation in the name too
[11:35:20] <micges> moment
[11:36:11] <micges> I must restart
[11:44:01] <micges> Im back
[14:19:38] <skunkworks_> logger_emc: bookmark
[14:19:38] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-01-17.txt
[14:27:26] <alex_joni> bbl
[14:43:57] <skunkworks_> http://www.cnczone.com/forums/attachment.php?attachmentid=48745&d=1197947544
[14:44:16] <skunkworks_> http://www.cnczone.com/forums/showpost.php?p=380884&postcount=8
[14:54:57] <xemet> hello
[14:55:19] <skunkworks_> Hi
[14:55:52] <micges> bye
[14:56:19] <cradek> skunkworks_: emc2 only requires a tiny bit of gnome. you certainly don't have to run the gnome desktop environment
[14:56:55] <skunkworks_> did you see the picture? pretty techie looking.
[14:57:19] <cradek> yeah I've seen that desktop cube thingy
[14:57:28] <xemet> I had a problem with gdepth, some one could help me?
[14:58:41] <cradek> what is gdepth?
[14:59:11] <xemet> http://axis.unpythonic.net/01169521961
[14:59:12] <skunkworks_> Hi cradek. http://axis.unpythonic.net/01169521961
[14:59:15] <skunkworks_> heh
[15:02:17] <cradek> xemet: I think that was a fun project jepler did, but he might not be interested in keeping it up to date - it might be up to you :-)
[15:03:10] <xemet> yes, but I need some indication....
[15:04:30] <xemet> it works with the old 2.1.7, but not in the 2.2.2
[15:04:49] <xemet> I get an error: spiral.ngc:0: Bad character 'g' used
[15:05:22] <xemet> the error is generated by a line I don't rememeber exactly in this moment, but is is something like gcode.parse(........)
[15:05:39] <xemet> so I think this gcode.parse has changed in the 2.2.2
[15:08:20] <skunkworks_> Its snowing - again :)
[15:15:03] <xemet> here only rain http://www.albaria.com/mondellofoto.jpg
[15:16:28] <skunkworks_> that is pretty
[15:21:27] <jepler> There's no documentation on what has changed in the 'gcode' interface from emc 2.1 to emc 2.2. However, by noting which functions have been changed or added here, you can get some idea of what differences there are: http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/emc/rs274ngc/gcodemodule.cc.diff?r1=1.12.2%3Av2_1_branch&tr1=1.1&r2=1.24.2%3Av2_2_branch&tr2=
[15:21:37] <jepler> I haven't used or updated gdepth in months, and I have no plans to.
[15:23:13] <xemet> ok, thank you very much
[17:17:56] <SWPLinux> hey JymmmEMC, you up?
[17:18:48] <JymmmEMC> SWPLinux: No, leave a msg at the beeeeeep.... how was your flight?
[17:22:47] <JymmmEMC> SWPLinux: Yes, I'm here.
[17:23:02] <alex_joni> JymmmEMC: too late.. he left a message
[17:23:14] <JymmmEMC> =)
[17:23:48] <JymmmEMC> This looks good... http://www.cachecreek.com/dining/harvest_buffet/menu.aspx
[17:24:55] <alex_joni> I just had some really great pilaf http://en.wikipedia.org/wiki/Pilaf
[17:25:33] <SWPLinux> it was interesting - the hotel had sushi-like stuff at breakfast
[17:25:47] <JymmmEMC> SWPLinux: Sushi-like?
[17:25:53] <SWPLinux> I got there about 10 minutes before closing, but there was still rice and miso soup
[17:26:02] <JymmmEMC> What, is that ... like turkey sushi???
[17:26:10] <alex_joni> JymmmEMC: raw unidentifiable thingies
[17:26:16] <JymmmEMC> heh
[17:26:19] <SWPLinux> I assume there was other stuff (there were condiments - wasabe and ginger and stuff)
[17:26:25] <JymmmEMC> ah
[17:26:30] <SWPLinux> too bad the coffee is cap
[17:26:32] <SWPLinux> crap
[17:26:53] <SWPLinux> I'm going to head up to SFO for MacWorld, what's your schedule later today?
[17:27:11] <alex_joni> hmm.. just started listening to some really old CDs of mine
[17:27:22] <SWPLinux> click click pop
[17:27:25] <JymmmEMC> SWPLinux: Just the optomitrist at 1pm - then I'm all yours baby!
[17:27:43] <SWPLinux> (my old CDs have holes in the foil from some manufacturing problem :) )
[17:27:47] <SWPLinux> JymmmEMC: ok
[17:28:09] <SWPLinux> I'll call you when I'm on the way back. when do you think it would be smart to avoid the roads southbound?
[17:28:11] <JymmmEMC> SWPLinux: I figured you wouldn't be back this way till around 4pm or so, but I could be wrong.
[17:28:18] <SWPLinux> probably at least that
[17:28:21] <JymmmEMC> SWPLinux: you driving up?
[17:28:33] <SWPLinux> I think so, but I may do a park/ride thing on CalTrain
[17:29:20] <SWPLinux> or BART
[17:29:38] <JymmmEMC> looking up map now...
[17:29:48] <SWPLinux> millbrae is the southernmost BART station
[17:30:13] <JymmmEMC> there also calTrain
[17:30:20] <SWPLinux> yep
[17:30:21] <JymmmEMC> the station is in Sunnyvale
[17:30:31] <SWPLinux> I'm in Santa Clara, whatever that means
[17:30:37] <JymmmEMC> hang on
[17:30:44] <JymmmEMC> you type too damn fast
[17:31:43] <SWPLinux> s o r r y
[17:32:30] <JymmmEMC> http://www.caltrain.org/news_2008_01_08_macworld_expo.html
[17:34:02] <JymmmEMC> SWPLinux: Now, I'm closest to the Blossom Hill station, so it's either you backtrack now, or drive thru traffic on the return trip.
[17:34:04] <SWPLinux> hmmm. $15.00 round trip + $2 parking (or more, since there's a conference here)
[17:34:43] <JymmmEMC> SWPLinux: And traffic isn't THAT bad, but it's not pretty either.
[17:35:05] <SWPLinux> hmmm, and then there's the cost of parking in San Fran, which is not cheap
[17:35:11] <anonimasu> macworld sucked -_-
[17:35:15] <anonimasu> nothing revolutionary released
[17:35:28] <anonimasu> just a ultraportable that weighs more then my ultraportable pc..
[17:35:41] <SWPLinux> I'm more there to look at the peripheral stuff - storage servers, render farms, that kind of thing
[17:35:47] <SWPLinux> I don't give a shit about Macs in general
[17:36:08] <skunkworks_> heh.. but it is unix based!
[17:36:19] <anonimasu> I'd like a mac.
[17:36:29] <tomp2> caltrain python yesterdays freshmeat http://freshmeat.net/projects/caltrainpy/?branch_id=72972&release_id=269822
[17:37:06] <SWPLinux> they're not too bad as computers go, but they're not that great, and there are still all sorts of "hidden functions", even though the user interface is supposed to be perfect
[17:37:47] <SWPLinux> (for instance, I minimized a Firefox or Safari window, and couldn't figure out how to get it back. I'm sure there's a keystroke that will do that though)
[17:37:54] <anonimasu> im sick of windows...
[17:38:11] <anonimasu> and well, the lack of great linux graphics drivers pisses me off,
[17:38:22] <SWPLinux> buy a commercial X server then
[17:38:22] <cradek> Results 1 - 10 of about 153,000 for mac minimized how to get it back
[17:38:39] <SWPLinux> well, I suppose I could have done a web search
[17:38:40] <anonimasu> SWPLinux: accelerated x didnt support ati last time i looked.
[17:38:45] <SWPLinux> but I can do that on Linux too :)
[17:38:47] <skunkworks_> but you can't search if it is minimized..
[17:38:54] <SWPLinux> you can open another window
[17:39:06] <anonimasu> not last time i looked
[17:39:24] <cradek> click-and-hold on the hidden application’s dock icon. When the contextual menu appears, you’ll see all of that program’s windows listed at the top of the menu. Minimized windows have a gray diamond next to their name
[17:39:26] <SWPLinux> that was my point though - it's no more obvious on the Mac than it is on Linux or Windows
[17:40:06] <JymmmEMC> SWPLinux: FYI http://www.caltrain.org/news_2008_01_08_macworld_expo.html
[17:40:09] <SWPLinux> file -> new window still works, even if you can't find the minimized window
[17:40:19] <SWPLinux> saw that, thanks
[17:40:49] <JymmmEMC> SWPLinux: Sorry, wasn't sure what link I gave previously
[17:41:24] <SWPLinux> yay - BART is $8, with free parking at Millbrae :)
[17:41:37] <JymmmEMC> SWPLinux: Why BART?
[17:42:24] <SWPLinux> it's faster (fewer stops), cheaper, and parking is free :)
[17:42:37] <JymmmEMC> CalTrian parking is free too.
[17:43:05] <SWPLinux> no, it's $2 at most stations, more on conference days at the San Jose <whatever> station
[17:43:34] <SWPLinux> at least that's what the website says
[18:03:54] <xemet> hello
[18:04:09] <xemet> please if possible help me with the pluto board
[18:05:31] <renesis> wtf is pluto board
[18:06:05] <xemet> http://www.knjn.com/board_pluto-P.html
[18:13:22] <skunkworks_> xemet: ask away
[18:21:25] <xemet> skunk
[18:21:31] <xemet> I've got the new pluto
[18:21:38] <xemet> the board works, I'm sure.
[18:21:40] <xemet> BUT
[18:21:53] <xemet> I cannot succeed in getting my servo turn in both directions
[18:22:05] <xemet> it is like the pwm works only in one direction
[18:22:49] <bill2or3> buh
[18:23:23] <skunkworks_> odd
[18:23:55] <skunkworks_> can you swap the pwm pins and see if it is the pluto or the amp?
[18:24:08] <skunkworks_> physical pins
[18:32:08] <xemet> sorry, girlfriend at telephone...
[18:32:17] <xemet> I'll be back in a minute
[18:38:58] <xemet> ok
[18:39:08] <xemet> let's see if I can work now
[18:39:35] <xemet> skunk, the amp works, because if I use it with the parallel board everything works fine
[18:40:52] <xemet> have you used it in pwm+direction mode?
[18:43:32] <skunkworks_> No. I use pwm+pwm
[18:44:07] <fenn> do you have a scope? or maybe even a multimeter would work
[18:44:58] <fenn> a servo amp is not a multimeter..
[18:45:51] <xemet> I've a multimeter
[18:45:59] <xemet> and I've seen
[18:46:19] <xemet> that when the pwm value is positive, the pwm pin has a +3.33 V
[18:46:44] <xemet> when the pwm vale became negative, the led on the board is on, but on the pin I've 0V
[18:47:06] <xemet> this explain why the motor does not turn...I think
[18:47:37] <xemet> I use pwm+direction
[18:48:33] <skunkworks_> seems like you have it set to pwm+pwm
[18:48:45] <bill2or3> what environment do you use to program that pluto board?
[18:48:45] <xemet> but I've set the parameter pwmdir to true
[18:49:04] <xemet> bill2or3, what?
[18:49:15] <xemet> I do not program it...
[18:49:26] <bill2or3> like, what language, and IDE, etc.
[18:49:33] <bill2or3> ahh, ok.
[18:49:48] <fenn> bill2or3: the firmware is written with altera quartus, but you dont need to mess with it to use the chip
[18:49:54] <SWPLinux> bill2or3: you would download the (windows-only) Altera Quartus development system
[18:50:03] <bill2or3> thanks.
[18:50:19] <bill2or3> windows makes me sadfaced
[18:50:33] <fenn> its altera's fault
[18:51:03] <xemet> anyone has use the pluto in pwm+direction mode?
[18:51:39] <skunkworks_> what do you get on the other pwm pin?
[18:53:02] <xemet> ah...the direction pin...I've renounced to use it
[18:53:41] <xemet> My board needs two pin for direction. The classic Dir pin and the inverted
[18:53:51] <xemet> like this: http://emergent.unpythonic.net/projects/01142347802
[18:54:46] <xemet> I'had a lot of difficult in using the dir pin of the pluto, so for the moment I used two output pin and send them a dir signal generated with the comp component
[18:56:44] <xemet> comp has two inputs: I've set one to 0 and the other is linked with the pwm value
[18:56:56] <xemet> so the output of comp is the dir signal
[18:57:07] <xemet> jepler suggested this way to me
[18:57:45] <xemet> I link this signal to a output pin of the pluto and to another inverted
[18:57:59] <xemet> so I get the dir and the not-dir signal
[18:58:13] <xemet> and this works because I've measured it with the multimeter
[18:58:28] <xemet> the thing that does not work is the PWM signal
[18:58:43] <xemet> one direction is ok, other is always 0V
[18:59:32] <SWPLinux> xemet: do you have a MIN_OUTPUT set in the ini file?
[18:59:38] <SWPLinux> or MIN_DC or something like that
[18:59:48] <xemet> I don't think so
[19:00:03] <SWPLinux> well, look :)
[19:00:16] <xemet> wait a moment
[19:02:09] <skunkworks_> could you pastbin.ca your hal file?
[19:02:54] <xemet> uhm, I've it on the other PC, wait a moment, I transfer it here
[19:06:10] <xemet> http://www.pastebin.ca/858826
[19:06:32] <xemet> consider the X axis only
[19:06:54] <xemet> I do not have the Y at the moment
[19:14:43] <xemet> any suggestion?
[19:15:23] <skunkworks_> not off the top of my head.. Looks correct for pwm+dir. (as much as I can tell)
[19:16:43] <fenn> you could do the absolute value of the pwm value in hal (hacky)
[19:16:51] <xemet> well, I've to go now, I will try to ask again tomorrow, I've to find someone that used the pwm+dir
[19:16:52] <skunkworks_> could you set it to pwm+pwm and see if one direction outputs on one pin - and the other direction outputs on the other?
[19:16:59] <xemet> fenn, yes, I've thought it
[19:17:28] <xemet> skunk, I will try...I will try everything...
[19:18:00] <xemet> I'm getting crazy with this pluto, it seems I'm the only one that cannot get it working :)
[19:18:13] <fenn> what's comp.0 supposed to do?
[19:18:33] <xemet> comp creates the dir signal, explained above
[19:18:36] <SWPLinux> it gives a sign for the PWM command
[19:18:52] <SWPLinux> err - from the command going to the PWM, that is
[19:18:59] <xemet> I need the dir signal and its inverted
[19:19:32] <xemet> so I used comp first to generate the inverted signal because I thought I could use the pluto dir signal
[19:19:47] <xemet> now I've used comp and not to generate a dir signal and its inverted,
[19:19:51] <fenn> ok
[19:20:27] <xemet> with this configuration, If, like you suggested, I get the absolute value of pwm, it should work
[19:20:57] <xemet> so I will try this hack tomorrow. If it works, well, after I would like to succeed in getting it work in the right way...
[19:22:49] <xemet> ok, now I go, I will try again tomorrow, thank you all
[19:23:17] <xemet> bye
[19:28:47] <tomp2> ? halscope to see the signal does change sign when it should? ( does the command to the hdwr actually change )
[19:29:29] <SWPLinux> seems like it must if the comp output changes
[19:31:12] <skunkworks_> the main issue is that the pwm pin only has pwm on it when the signal is positive.
[19:31:14] <skunkworks_> iirc
[19:31:43] <SWPLinux> tomp2's question is whether the input is actually going negative, or stopping at 0
[19:31:44] <jepler> ah looking at the pluto_servo.comp I think I've spotted a bug that prevents pwmdir from working properly
[19:32:08] <SWPLinux> oh, that's helpful :)
[19:32:48] <jepler> I can't test this change right now, but it may help xemet: http://www.pastebin.ca/858870
[19:32:59] <jepler> if he comes back later and I'm not here, it'd be nice if somebody could point him at it
[19:33:18] <fenn> man gtk file chooser dialog sucks
[19:34:23] <tomp2> is xemet == manfredi?
[19:34:28] <skunkworks_> yes
[19:34:43] <fenn> typing ~ always goes to a nonexistent path
[19:35:41] <fenn> and actually typing out the filename doesn't open it
[19:36:14] <tomp2> i dont have manfredi's email of today, but someone could just send him a quik note
[19:36:58] <tomp2> (just flushed the email device)
[19:37:11] <tomp2> bwooosh
[19:38:51] <jepler> ah I found the address .. I'll e-mail him
[19:39:03] <tomp2> just did, he just sent a new one
[19:39:27] <tomp2> 2 emails is ok too (too emails is ok 2 )
[19:40:42] <SWPLinux> to send two emails is OK too?
[19:40:54] <tomp2> email email is ok ok
[19:40:58] <jepler> I have no idea what any of you are saying
[19:41:04] <jepler> please to be in english?
[19:41:04] <SWPLinux> that's a good sign
[19:41:12] <tomp2> suspenders and a belt is ok
[19:41:20] <SWPLinux> yoda-speak not is better, think you?
[19:41:33] <tomp2> yoda spoke forth
[19:41:38] <SWPLinux> indeed
[19:42:17] <fenn> yoda spoke backwarth
[19:59:49] <SWPLinux> ok, time to hit the road. BBL
[21:22:35] <alex_joni> good night all
[21:22:46] <tomp2> gnite
[23:13:41] <GNieport> Hi guys, anyone on tonight familiar with the 5i20 card?
[23:14:51] <seb_kuzminsky> not really, but what's your question? :-)
[23:17:38] <GNieport> I am just getting started with EMC. I am using the 5i20 an analog servo output card, 7i33. Anyways, the 7i33 is supposed to output a 0-10volt signal to command the servo. I have not been able to get any voltage on the output, yet I can see that the "dac-value" is changing in Hal
[23:18:01] <GNieport> the 5i20 digital IO does work correctly
[23:18:19] <GNieport> so I know the card is at least adressed correctly
[23:18:23] <seb_kuzminsky> right
[23:18:50] <seb_kuzminsky> the 5i20 is a PCI card, right? how does it get 10V to output? or is that done on the 7i33?
[23:19:02] <GNieport> PCI, yes.
[23:19:18] <GNieport> there is a buss connection (think daughtercard)
[23:19:47] <seb_kuzminsky> oh i see, the 5i20 gives pwm&dir signals to the 7i33, and the 7i33 converts it to a +/- 10V voltage
[23:19:48] <GNieport> the 10v is generated internally
[23:19:55] <GNieport> exactly.
[23:21:34] <GNieport> I've been toying around in hal for about a week, and belive I have all the bits set right. EMC will come out of ESTOP, and the Machine power will turn on
[23:21:56] <seb_kuzminsky> so the 7i33 makes the 10V from the 5V you feed it
[23:22:12] <GNieport> I know the servo amps are enabled, because I was able to command motion with a small battery on the analog inout
[23:22:22] <seb_kuzminsky> are you providing external 5V power to the 7i33 on the P1 connector
[23:22:24] <GNieport> i suppose so, the docs are a little sparse
[23:22:41] <GNieport> no, just power over the 50 pin
[23:23:49] <seb_kuzminsky> since the 5i20 gives the signal to the 7i33 via pwm, are you sure you're executing the pwm function, and is the pwm enable on?
[23:23:51] <GNieport> I thought my self-terminated output ribbon was suspect, but i checked the pins directly on the logic card
[23:24:10] <seb_kuzminsky> ah you see the pwm signal at the end of the 50-pin cable from the 5i20?
[23:24:14] <GNieport> and am not seeing more than 0.035V or so
[23:24:30] <GNieport> I haven't checked that side yet, I need to dig out my scope
[23:25:11] <seb_kuzminsky> can you pastebin 'halcmd show all'?
[23:25:39] <GNieport> brb
[23:29:01] <seb_kuzminsky> the gnieports are multiplying
[23:32:30] <GNieport1> yes, I had to go to the garage
[23:32:33] <fenn> GNieports: you have to pull enable low
[23:32:35] <GNieport1> http://pastebin.ca/859186
[23:32:51] <GNieport1> on the 7i33, yes, I am
[23:33:07] <GNieport1> to one of the GND connections on the same 50 pin connector
[23:34:38] <seb_kuzminsky> all your axis enables are false, maybe the signal is stopping in the 5i20?
[23:34:59] <GNieport> well, the control cabinet is off just now.
[23:35:35] <GNieport> pin axis enable?
[23:36:25] <seb_kuzminsky> yeah, axis.X.amp-enable-out
[23:36:45] <seb_kuzminsky> oh but wait, you're using testbittwo for the dac-00-enable, nm
[23:37:02] <GNieport> right, I was hoping to get a few ides for the next evening I can work on it. I have the baby on my lap tonight :)
[23:37:30] <seb_kuzminsky> heh yes hard to type with babies crawling on you
[23:37:31] <GNieport> I loaded EMC, then changed a few things to let me inject a sine wave
[23:37:49] <GNieport> well, also the garage is a not-s-friendly 30 degrees, lol
[23:37:50] <fenn> is dac-enable the same as turning on enable for the 7i33?
[23:38:00] <fenn> (i dont think it is)
[23:38:00] <GNieport> I believe not
[23:38:11] <GNieport> the enable line on the 7i33 is hardware only
[23:38:25] <GNieport> i.e. force the dac output to zero
[23:39:01] <seb_kuzminsky> do you have a breakout board for the 50-pin ribbon cable coming out of the 5i20?
[23:39:20] <GNieport> I built an opto coupler interface
[23:39:24] <GNieport> so, yes :)
[23:39:27] <seb_kuzminsky> if so put a multimeter on the dac-00 pin and get the 7i33 out of the loop
[23:39:37] <seb_kuzminsky> to start narrowing the debugging down
[23:39:42] <GNieport> Okay
[23:40:21] <seb_kuzminsky> you should be able to measure the pwm signal with a multimeter, it should be a voltage between 0 and 5, proportional to the value you write to the dac-00-value
[23:41:06] <GNieport> The 5i20 has many hardware goodies, I was not sure if dac-00,etc were the same signals as what I would expect on the analog signal output
[23:41:17] <GNieport> I will check the 5i20 output
[23:41:54] <seb_kuzminsky> yeah, there's kind of a disconnect between HAL stuff and real hardware stuff...
[23:43:45] <seb_kuzminsky> it's strange to me that the 5i20's pwmgens are exported as dacs... they're only dacs if there's a 7i33 (or similar) plugged into them
[23:44:14] <GNieport> i find it fascinating, especially when I strted to remember the Hal commands; i can't stand the hurry up and wait of my situation :)
[23:44:47] <GNieport> I was wondering the same thing, perhaps the analog interface is how most people use the card
[23:44:57] <GNieport> step and direction may or may not be working yet
[23:45:09] <GNieport> i haven;t heard, it has been discussed for a while
[23:45:58] <GNieport> I know that 2.2.2 included a new 5i20 driver
[23:46:12] <GNieport> anyhow, I'll have to check the pwm output next.
[23:46:33] <seb_kuzminsky> the pastebin info looks right to me for dac-00
[23:46:43] <seb_kuzminsky> oh, you're not on 2.2.2? what are you running?
[23:46:56] <GNieport> Yes, 2.2.2
[23:47:05] <GNieport> Installed from the LiveCD
[23:47:46] <GNieport> I had 2.1.7, and updated all packeges, which hosed the realtime kernel somehow
[23:47:47] <GNieport> lol
[23:47:51] <seb_kuzminsky> heh
[23:47:59] <GNieport> so I nuked the drive and started fresh
[23:48:37] <seb_kuzminsky> take off an nuke the site from orbit, it's the only way to be sure
[23:49:20] <GNieport> Exactly
[23:50:24] <seb_kuzminsky> yea, the hal setup looks ok to me, dac-00 seems set up right, looks like it should be outputting +10V (you should see 100% duty cycle on the PWM, so +5V on pin, uh, whatever pin it is)
[23:51:18] <seb_kuzminsky> ping 15 maybe? that's pwm0 on the 7i33
[23:51:54] <GNieport> Most likely, the chanel numbering seems to be consistent
[23:52:00] <seb_kuzminsky> oh good :-)
[23:52:35] <seb_kuzminsky> ok I've got to go put *my* baby on my lap now, good luck and see you later
[23:52:41] <GNieport> thnaks, bye