#emc | Logs for 2009-06-28

[00:47:35] <jimbo655> where do I find the list of functions that are available in hal such as power on/off in axis?
[01:00:33] <cradek_> cradek_ is now known as cradek
[01:09:41] <jimbo655> I am trying to setup the power on of the main contactor for the machine and servos and the spindle control. I would like to use the power button in axis. I am most likly blind but I am unable to find where to add this pin.
[01:10:14] <cradek> you can put whatever you want in the chain of events leading to 'machine on'
[01:10:33] <jimbo655> Where do I find machine on?
[01:11:06] <jimbo655> I have looked in the hal files etc and I do not see it.
[01:11:19] <cradek> check out the demo_sim_cl config, which has some ladder in the estop chain
[01:11:42] <jimbo655> I will check ... thanks
[01:13:25] <cradek> the way I have my lathe set up is when I hit f1 to come out of estop, [if the hard estop buttons are out] there is a sequence of three contactors powering on various things, and if the last one succeeds, it has an auxiliary switch that is read back into ladder. only if they all succeed does the estop button in AXIS pop out. then I can turn 'machine on' which enables the servo amps.
[01:14:26] <cradek> this way if there is any fault [such as an estop button is pressed in] you see on the screen that it didn't make it of estop.
[01:15:04] <cradek> also my estop button has a red light in it which comes on when it comes out of estop.
[01:18:35] <Jymmm> Heh, I've debated when the light comes on
[01:19:43] <Jymmm> Like, should it come on when estop (hard or soft) is tripped as an indicator, or on so the button is easier to locate
[01:21:24] <Jymmm> but... I also have a green lit button, so could use that to indicate estop is ready to reset (no triggers)
[01:22:50] <cradek> Jymmm: I think of it as "danger - the machine is active" and also as a way to help locate the button
[01:23:52] <Jymmm> cradek: Yeah, I can understand that.
[01:25:07] <Jymmm> cradek: I also have a yellow lamp, so could use it to indicate that system is "live", really haven't drawn out all the controls and logic I want.
[01:25:33] <cradek> it's hard to decide how to do estop.
[01:26:17] <Jymmm> I like it lit when estop is triggered so when I go WTF?!??! I can look over and see "Oh, estop has been triggered"
[01:26:57] <Jymmm> but like I said I have a bunch of controls/lamps and just not sure what/why/how yet =)
[01:28:13] <Jymmm> I want one outlet for vac, for both auto and manual, and when in manual to bypass hard estop chain - still need to think that one out.
[01:28:58] <Jymmm> but, I also don't want to accidental plugin the spindle into the vac outlet without thinking
[02:16:35] <jimbo655> I found the demo_sim_cl files in the root file structure. I have looked at all of the files and I still see no referance to the power button. only the Estop.
[02:21:18] <jimbo655> I do find a referance to this button under halui called halui.machine.is-on. am I able to link this button to a pin in HAL
[02:30:20] <tomp> i see 'net machine-on pyvcp.machine.on halui.machine.on' in your-emc-root/configs/halui_pyvcp/pyvcp.hal
[02:30:52] <tomp> not sure what the effort is tho
[02:32:10] <Jymmm> cradek: I think we were talking about this a ways back (could have been jepler too)... http://www.linuxfordevices.com/c/a/News/Marvell-SheevaPlug-NAS/
[02:46:21] <tomp> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Halui as i read it, one signal is an action ( turn machine on ) other is a status ( is on or NOT is on )
[02:46:23] <tomp> jimbo655: ^^^
[02:49:38] <tomp> jimbo655: on same page "Since the machine on function is really dependent upon the state of estop, we should be able to implement the estop reset from this panel using the on/off pin. The mushroom head twist-to-release will be external motor power"
[02:50:08] <tomp> jimbo655: so it sounds as if there's no separate 'power button',
[03:05:36] <jimbo655> Pretty much that is it. I am not sure that I need one, except for the one in AXIS.
[03:06:37] <jimbo655> I am staring from scratch with the control I am trying to AXIS to control the machine and then be able to add hard buttons as I think they are needed. except for E-stop
[03:10:39] <cradek_> jimbo655: 'machine on' turns on axis.N.amp-enable-out
[03:21:55] <cradek_> what oil would you guys guess would go in my bus's differential?
[03:24:48] <skunkworks> something stinky I would guess.
[03:25:06] <cradek_> that's what came out...
[03:25:09] <skunkworks> * skunkworks hates the smell of gear oil
[03:25:18] <jimbo655> Not sure about a Bus but older cars use 90 weight. Not sure a bus dif would need anything different....
[03:25:50] <cradek_> I see talk about SAE80W90 GL-5 (GL-5 means for hypoid [sliding] gears)
[03:26:19] <jimbo655> My wife said my mill has dirty nappies so I had to clean it. now it smells fresh and clean.
[03:26:42] <cradek_> the stuff in there doesn't look much thicker than 10W40 - but it might be 18 years old
[03:28:25] <jimbo655> For a while they were trying lighter weight oils to reduce friction. it might be a lighter weight oil with additives that make it cling to the gears.
[03:28:42] <cradek_> this is a '92
[03:28:56] <cradek_> unfortunately I don't have much documentation for it.
[03:29:33] <jimbo655> Bummer
[03:56:08] <KimK> If I say "setp hm2_5i22.0.pwmgen.03.scale -20000", it works fine, but if I say "setp hm2_5i22.0.pwmgen.03.scale constant.0.out", I get "value 'constant.0.out' invalid for float". constant.0.out is reporting the correct value, according to hal configuration. Does this suggest anything as to where I should look for the trouble?
[03:56:51] <KimK> I forgot to say constant.0.out is reporting the correct *floating point* value.
[03:56:52] <SWPadnos> setp only accepts constants for the value to set, not variable or signal names
[03:57:55] <SWPadnos> and it doesn't do math either
[03:58:01] <KimK> OK, sorry, I also tried the other feature, constant.o.value, which is a parameter read write.
[03:58:31] <SWPadnos> oh, I should be more clear there, you can use an environment variable or ini variable using the various substitution symbols ([], ())
[03:59:17] <KimK> this has to be changable. An environment variable?
[03:59:24] <SWPadnos> well, if the scale is a parameter and not a pin, then you can only set it to constant values, you can't connect parameters to anything
[04:00:04] <SWPadnos> yes, you can set it to an env var, but it will not change when the variable changes, it will get the constant value of whatever the variable was at the time of the "setp" execution
[04:00:53] <SWPadnos> so fred=1 / setp foo.bar.param $fred / fred=2 will make foo.bar.param 1, even though fred changed later
[04:02:56] <KimK> Yes, that's what I want, an "occasionally changeable" constant. I thought "constant.n" was to make a pin out of a param? I'm not sure I follow the "$fred / fred=2" part of your example, what is that doing?
[04:03:32] <KimK> Oh, wait, I see
[04:03:46] <SWPadnos> constant is no longer useful, since you can now use setp to directly change pin states
[04:04:07] <SWPadnos> it used to be that the "p" in setp meant "param", but now it's either "pin" or "param" :)
[04:04:24] <SWPadnos> slashes represent carriage returns or separate statements :)
[04:07:00] <KimK> Isn't the advantage of using constant.n that I only have to enter a param once, and then can send it two places?
[04:07:11] <SWPadnos> no
[04:07:23] <SWPadnos> constant has a param as "input", and a pin as "output"
[04:07:34] <SWPadnos> that output can only be connected to pins, not to params
[04:07:40] <SWPadnos> you can't connect params to signals
[04:08:02] <SWPadnos> so you would only have the advantage of being able to change several pins with a single setp
[04:08:27] <KimK> how about constant.n'ds r/w param value, constant.n.value ?
[04:08:28] <SWPadnos> which sets does quite nicely, with no need to load the constant component, or to add its function(s) to a thread(s)
[04:08:47] <SWPadnos> what about it?
[04:08:52] <SWPadnos> that's the input parameter
[04:09:20] <KimK> isn't it r/w? man constant implies it's r/w?
[04:09:42] <SWPadnos> the one thing about constant is that the changes are guaranteed to happen at a certain time in the thread execution, since the parameter is "forwarded" to the pin only when the function executes
[04:09:44] <cradek> I didn't know we still had constant blocks - that is a holdover from when you couldn't setp a pin
[04:09:57] <SWPadnos> or a signal
[04:10:00] <SWPadnos> (maybe)
[04:10:08] <cradek> yeah in the very early days I think you couldn't sets a signal either
[04:10:12] <SWPadnos> KimK, that means that you can read it back
[04:10:32] <SWPadnos> all parameters can be read (makes no sense to prevent it), but only some can be written
[04:10:39] <SWPadnos> so you can do this:
[04:10:45] <SWPadnos> setp come.parameter 22
[04:10:48] <SWPadnos> err
[04:10:55] <SWPadnos> halcmd setp some.parameter 22
[04:11:04] <SWPadnos> halcmd show param some.parameter
[04:11:17] <SWPadnos> and you can see that the number printed is what you set it to
[04:12:05] <SWPadnos> for params that might get limited by some internal calculation (like a stepgen max_vel, dependent on its internally calculated max frequency), you can read back the actual value after the component applies any limiting to it
[04:12:07] <KimK> where would the output go if it was just show param some.parameter in the hal file?
[04:12:18] <SWPadnos> it would likely get printed to the terminal
[04:12:27] <SWPadnos> there is no HAL connection between parameters. none.
[04:12:54] <SWPadnos> they are only visible using halcmd (or another utility like it, if someone writes one)
[04:13:32] <KimK> OK, well, maybe I'll have to setp or sets twice then. I'll look at it some more. Thanks for the help, sorry to be dense.
[04:13:45] <SWPadnos> welcome
[04:14:02] <SWPadnos> don't worry, we're all dense at times - it just depends on what we're talking about :)
[05:06:11] <Jymmm> SWPadnos: what do you mean 'at times'
[06:08:50] <jimbo655> xminlim and xmaxlim if they are the same limit do I just comment the xminlim out?
[06:33:06] <jimbo655> Do you have to use 2 pins for each limit 1 for Xminlim and 1 for Xmaxlim?
[06:34:25] <Jymmm> You can wire them all up to one if you wish
[06:34:35] <Jymmm> xyz min max etc
[06:36:38] <jimbo655> My problem is if i assign pin 00 to Xmaxlim and Xminlim I get a lik error on startup. do you know how to get around this/
[06:36:49] <jimbo655> Link error
[06:37:34] <Jymmm> *I* don't, maybe someone else might
[06:44:17] <micges> jimbo655: can you pastebin.ca line you're trying to run?
[06:47:15] <jimbo655> Are you looking for the error on startup?
[06:47:38] <micges> too
[06:50:45] <jimbo655> the error is a HAL error pin 5i20.0.in-00 is linked to Xminlim and cannot be linked to Xmaxlim. How do I get arround this my Xminlim and Xmaxlim are the same switch.
[06:52:18] <micges> what are Xminlim and Xmaxlim? signals names?
[06:54:14] <micges> it will be faster if you show me part of that hal file you working
[06:54:24] <jimbo655> they are both linksp <= 5i20.0.in-00-not
[06:54:57] <jimbo655> I am on two diferent machines with no communication between them.
[06:55:25] <jimbo655> And one is windows
[06:55:45] <micges> ok. first. what you are trying to do?
[06:57:10] <jimbo655> I have a single limit switch for the limits on all axis. I am trying to rename the pins to match my wiring.
[06:57:49] <jimbo655> I have home switches also but do not think this will be a problem
[06:58:11] <jimbo655> That is a single for each axis
[07:01:30] <micges> I would write this: net limit-xy 5i20.0.in-00 => axis.0.neg-lim-sw-in axis.0.pos-lim-sw-in axis.1.neg-lim-sw-in axis.1.pos-lim-sw-in
[07:02:57] <micges> one signal name is connected to all hard switches
[07:03:47] <jimbo655> do I include axis 2 for the Z
[07:04:09] <micges> yes you can
[07:04:33] <jimbo655> Thank you I will try it......
[07:15:04] <jimbo655> micges: life is good thanks much.......
[07:15:22] <micges> welcome
[16:43:00] <eric_unterhausen> from now on, I'm sticking with stainless steel
[17:14:39] <awallin> eric_unterhausen: ouch, that's difficult to machine and hard on the tools!
[17:22:44] <noel> noel is now known as Guest29584
[17:22:48] <Guest29584> Guest29584 is now known as Roguish
[17:24:20] <eric_unterhausen> awallin: but it looks so nice :)
[17:30:24] <Roguish> good day all. is git down?
[17:56:45] <Jymmm> Jeees, that's 4 this week... http://www.nydailynews.com/news/2009/06/28/2009-06-28_tv_pitchman_billy_mays_found_dead_in_florida_home.html
[18:00:11] <eric_unterhausen> Jymmm: I'm sure there have been more deaths than 4
[18:00:42] <Jymmm> celebs
[18:01:39] <eric_unterhausen> two 50 year old guys, I'm watching my back
[18:01:52] <eric_unterhausen> MJ was born 7 days after I was
[18:03:07] <Jymmm> And ironically their initials spell.... BEMF! You better watch out eric_unterhausen!
[18:03:38] <eric_unterhausen> whoosh, sound of that one going over my head
[18:04:32] <Jymmm> BEMF = Billy, Ed, Michael, Farrah == Back EMF, as in what happens when one stepper motor produces when it slows/stops.
[18:05:14] <eric_unterhausen> you're right, I'm gonna die
[18:05:28] <Jymmm> eric_unterhausen: Are you a carnivor?
[18:05:35] <Jymmm> Like a good steak?
[18:05:49] <eric_unterhausen> I don't eat many, I knew someone who died of mad cow
[18:06:15] <Jymmm> eric_unterhausen: Well, you'll love this.... add your initial and it's: M Beef!
[18:06:38] <Jymmm> Which I take as: Mmmmmmmm Beef!
[18:07:02] <Jymmm> eric_unterhausen: So, no beef for you for at least a week! LOL
[18:07:10] <Jymmm> or the month of june
[18:07:58] <eric_unterhausen> ever hear of the story of death in Tehran?
[18:08:12] <Jymmm> nope
[18:09:08] <eric_unterhausen> basically says you can't avoid death by skipping meat for a week :)
[18:09:25] <Jymmm> eric_unterhausen: RIP.... We'll Miss you!
[18:32:18] <kakeman> does anything working with mach3 work with emc?
[18:33:01] <cradek> kakeman: ask a more specific question to get a good answer
[18:33:15] <kakeman> ok
[18:33:16] <cradek> kakeman: your question is nearly impossible to answer.
[18:34:20] <kakeman> yes, very hard to
[18:34:42] <cradek> that's why nobody has answered :-)
[18:36:20] <kakeman> maybe i have to read littlebit more and ask less
[18:37:35] <cradek> Roguish: git is up and down. like SWPadnos said on the list, the ISP is having trouble with a router bug but is working on it.
[18:38:07] <cradek> bbl
[19:22:34] <gefink> hello
[19:32:53] <micges> hi
[22:43:59] <skunkworks> http://rc.cis.si/projects/cnc_motion/
[23:36:56] <skunkworks> did this weekend fly by again>?
[23:37:54] <Roguish> ciao, all......
[23:44:08] <DaViruz> what kind of maximum step freqency is reasonable from the parallell port of a reasonably recent computer?
[23:46:31] <DaViruz> my steppers seems to misbehave beyond 20kHz (1200mm/min) on both X, Y and Z which is odd considering Z requires alot more torque when traversing upwards..
[23:49:16] <parag0n> has anyone here tried CNC converting a Seig c0 lathe to use with emc2?
[23:56:13] <pcw> DaViruz: have you run the latency test?
[23:56:15] <pcw> That will tell you how fast you maximum step rate will be