#emc | Logs for 2010-11-28

Back
[00:04:18] <theorb> theorb is now known as theorbtwo
[00:56:54] <AeroTec> so if one was willing to spend a little more money, what would be the best way to connect the computer to the CNC, Mesa 7i43 was recomended but is there a better way to go or is this getting close to the top of the line?
[00:57:30] <AeroTec> I when to there web site and wow, they have a ton of card one can buy
[00:57:42] <AeroTec> over welming
[00:58:10] <andypugh> 7i43 is OK, and plugs conveniently into the parport.
[00:58:11] <AeroTec> went not when
[00:59:31] <AeroTec> can one run more then one of their Mesa 7i43 cards in a computer?
[01:00:09] <andypugh> There are equivalent PCI cards with more pins (5i20, 5i23). Top of the line is probably the 3X20 which plugs into cabled PCIe and has 144 IO pins
[01:00:41] <andypugh> More than one 7i43 is possible, but probably silly.
[01:00:58] <andypugh> Better to go with one of the internal cards.
[01:02:06] <andypugh> The 7i43 is around $80, the 5i23 is $200.
[01:02:56] <AeroTec> I cut for cash so spending some $$ is not a problem as long as I am not wasting $$
[01:03:18] <AeroTec> want to get a good system not cheap out to much
[01:03:33] <AeroTec> but not waste $$ ether
[01:04:23] <AeroTec> sound like the Mesa 7i43 is a good starting level for me
[01:04:43] <AeroTec> but may want to upgrade soon
[01:07:23] <tom3p> i dont 'know' git (haha)... where can i read about how to use 'git' for the newest release (already used git for a rip install)
[01:07:34] <andypugh> 7i43 makes for easy cabling, one cable from PC to control box, then all the 48 little wires are in one place.
[01:10:12] <tom3p> hotay panky... http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Git
[01:11:22] <AeroTec> I would love to play around with robot arms and other robo mechs
[01:11:45] <AeroTec> I guess there is no easy movement programming tool
[01:11:46] <andypugh> Then EMC2 is a good package to get to know
[01:13:02] <andypugh> As far aw we know, EMC2 with genserkins is about as good as it gets, certainly in the free and Free arena
[01:13:05] <AeroTec> what shouyld I read ot get or what ever to learn more about doing robot arms and such in/with EMC2?
[01:14:02] <andypugh> Probably the integrators manual. http://www.linuxcnc.org/docs/EMC2_Integrator_Manual.pdf
[01:14:30] <andypugh> Especially Kinematics
[01:14:33] <tom3p> AeroTec, ? as in Aerotech.com?
[01:16:07] <AeroTec> no
[01:16:35] <AeroTec> I did a google search for genserkins
[01:16:36] <tom3p> np ( i got an A3200 4 axis setup )
[01:16:50] <AeroTec> what is it and how do I get or use it?
[01:17:39] <tom3p> the docs are online or in the live cd... afik its a set of stacked up joints ( thus serial )
[01:18:24] <AeroTec> is it part of EMC?
[01:18:49] <AeroTec> I am as green as they come to EMC, but not CNC
[01:19:30] <andypugh> A normal config files has "loadrt trivkins" at the top. To use genserkins (General serial kinematics) you just change that word. Then you spend ages getting all the DH parameters right.
[01:20:16] <andypugh> There is a nice video about DH parameters here: http://en.wikipedia.org/wiki/Denavit-Hartenberg_Parameters
[01:20:51] <tom3p> yes its part of emc2, it defines the realtions between the swivel-ly bits of the robot arm
[01:21:08] <tom3p> or relations :|
[01:21:32] <jay_> jay_ is now known as fjay
[01:24:26] <tom3p> are you sure kinematics is where you want to start with in EMC2? or are you just interested? np, but its mathamagical
[01:24:43] <andypugh> EMC2 has a number of kinematics modules, trikins maps one motor to each cartesian joint, then there is gantrykins which is still cartesian but maps one or more motors to each joint. Then it gets complicated with hexapods, serial robots, parallel robots, 4-wire winches, 2-wire winches....
[01:26:28] <tom3p> re: git "git checkout v2.4.6_branch" doesnt work, what is the correct 'tag'?
[01:26:34] <andypugh> But yes, kinematics is advanced stuff. Be aware that that video earlier of a robot arm moving in cartesian space is something the commercial robot companies don't do. (I hesitate to say that they can't, but they apparently look blank when you ask)
[01:27:46] <andypugh> tom3p: I would say ask on -devel but I think that they are all here
[01:28:56] <AeroTec> will start with lathe and mill
[01:29:18] <AeroTec> but would like to start playing with arms soon
[01:30:49] <AeroTec> you say the commercial guys do not use genserkins, do they just use Gcode and play with the Gcode till it moves the way they want?
[01:36:01] <tom3p> look up 'kinematics' and get that idea first. gcode is not directly connected to the things that move, gcode is directly connected to the way you measure stuff.
[01:36:42] <tom3p> in a robot you dont directly have 'X', but you can move along it according to a set of relations called the kinematics
[01:38:27] <andypugh> Most industrial robots are just taught end points. The programmer moves the robot manually to a postion and the joint angles are stored, then the next position, and so on. Then the robot repeats that sequence. Nothing inside the controller actually knows where in space the end of the arm is (and, in most cases, it doesn't matter)
[01:39:49] <AeroTec> would it be easy to do that in EMC2?
[01:40:01] <AeroTec> just record end points
[01:40:31] <andypugh> Actally, I suspect not.
[01:41:14] <andypugh> I have seen some teach-in stuff somewhere in the project, I don't know how live it is.
[01:43:19] <AeroTec> when ones uses a prob and scans some 3D object, does not the axis get stored?
[01:43:34] <AeroTec> x,y,z,a?
[01:44:07] <andypugh> Yes, but that is the cartesian coordinates, not the joint angles
[01:44:31] <AeroTec> could one not manualy trigger the probe pin and store the data?
[01:44:59] <AeroTec> a axis would be a angle
[01:45:22] <AeroTec> can one not program all axis to be angles?
[01:46:04] <AeroTec> then one stores all the angles at one point in time with the trigger for scanning
[01:46:31] <AeroTec> so move, store, move, store
[01:46:51] <andypugh> Yes, no reason that G0 X0 can't set an angle, you can send that command value to any motor
[01:47:21] <AeroTec> my point, yes
[01:48:04] <AeroTec> and can x to be rotary like the A axis
[01:48:06] <andypugh> With a probe-input switch and all the motors unpowered but with encoders running you might even be able to use the arm as a manually-moved digitiser.
[01:48:32] <AeroTec> that would be so cool
[01:48:41] <AeroTec> I could use one of them
[01:48:53] <AeroTec> that is a great idea
[01:49:08] <andypugh> We have one at work (well, it has not motors, just encoders in the joints).
[01:49:31] <tom3p> look at www.romer.com portable cmm's arms that are used to digitize
[01:49:41] <andypugh> That does do the inverse kinematics though, and measures in cartesian space.
[01:49:46] <AeroTec> yea that was what I was thinking, a arm with just encoders
[01:50:33] <AeroTec> for a 3D scanner
[01:50:35] <andypugh> I think ours is an old Romer. They are frighteningly expensive.
[01:51:14] <AeroTec> but for the arm it would have moters and whould hit a manual switch to store the joint angles
[01:51:48] <AeroTec> so can all axis have a 360 rollover?
[01:52:01] <AeroTec> so goes to 360 and then 0?
[01:52:15] <tom3p> very expensive, how about emc2/xbox/OpenKinect to sense end end effector position :)
[01:52:19] <andypugh> I am not sure.
[01:53:45] <andypugh> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?WrappedRotaryAxes
[01:54:18] <andypugh> Suggests that it should only be used on ABC axes, but it is possible that you can define X as angular and wrap it.
[01:58:08] <AeroTec> one could make a small encoder arm and have it control the robot arm and them teach it that way by hitting the scan trigger
[01:58:30] <AeroTec> getting ones end points
[01:59:42] <pcw_home> Short long enough to blow up the 8I20 (we had 900 uF of filter at 360V)
[02:00:24] <pcw_home> Oops wrong window
[02:24:42] <AeroTec> when one hit the scanner trigger pin, does that store all active axis locations in the file?
[02:25:45] <AeroTec> so if one had 6 axis for the arm, EMC would store all 6 data points with one hit of the scanner trigger
[02:26:02] <AeroTec> would that be right? or is there more to it then that?
[02:28:02] <tom3p> yes, but substitute 'joints' for 'axis' in your question :) it would store 6 data for 5 joints, and kinematic could translate that data into XYZABC position ( or some polar format )
[02:28:25] <tom3p> make that ^^^ 6 data for 6 joints
[02:30:15] <AeroTec> cool
[02:30:17] <AeroTec> thanks
[02:31:20] <andypugh> If the feedback was encoders and a kinematics module was loaded then the probe would store the cartesian coordinates. If you were using a simple mapping with X redifined as rotary then you would get the same effect as using any other kineamtics that does not suit the application.
[02:32:15] <andypugh> What would get confusing in the "trivkins robot" is that the preview window in Axis would be horribly wrong.
[02:37:14] <AeroTec> this all sounds like fun stuff
[02:37:24] <AeroTec> will have to give it a try
[02:39:50] <Wesbaker> Is there a way to change the default g code file name from .ngc to something else?
[02:43:31] <andypugh> For opening or saving?
[02:46:38] <andypugh> Wesbaker: for loading you can just add the required extension with PROGRAM_EXTENSION = in the [FILTER] section of the ini
[03:06:47] <kb8wmc> does the axis simulator allow for setting different x and y zero? if so, how?
[03:10:06] <Portly> hello
[03:10:07] <the_wench> hello Portly, you have a question?
[03:10:40] <Portly> yeah. :) Have you ever had a problem with reading the parallel port?
[03:12:04] <andypugh> kb8wmc: touch-off?
[03:12:11] <Portly> I have the 8 LED schematic made and hooked up to my parallel port. The first 5 LED's are on, but faint. The last 3 are not. And nothing I do seemes to change this. I ran linux cnc and the lights did not flicker at all.
[03:12:30] <kb8wmc> rgr that Andy
[03:12:35] <kb8wmc> and thanks
[03:14:05] <Portly> this is the schematic I made. http://www.epanorama.net/circuits/lptcircuit2.jpg
[03:14:21] <andypugh> Portly: You might find this software equivalent useful http://wiki.linuxcnc.org/emcinfo.pl?Parallel_Port_Tester
[03:14:52] <Portly> ok, will have a look.
[03:16:24] <Portly> Will be right back. Have to boot into my other os...
[03:20:36] <Connor> Hmm.. okay.. so I have a few options now for Speed control on my Bosch. I can combine all 3 leads on the TRIAC and use a external speed controller that I can buy, or build.. or, possible remove the POT from the existing speed controller, and replace with some sort of variable resistor that's controlled via voltage.. (I think a FET can do this as well..)
[03:28:15] <Porty> in order to use that command. halrun. do I need to install the cnc version of linux?
[03:29:10] <Porty> halrun -I -f ptest.hal
[03:29:24] <Porty> No command 'halrun' found,
[03:29:49] <tom3p> halrun is part of emc2 suite
[03:30:54] <Porty> Ok, I was running it from the CD. Can I run the cnc version with my linux I have right now if I install it?
[03:31:20] <Porty> Or even overwrite this linux is fine too
[03:32:07] <tom3p> "the CD"... does that mean the emc2 live cd? i think halrun should be available on the livecd
[03:32:21] <Porty> yeah, the live.
[03:32:35] <Porty> I just can't get on this chat with the live cd.
[03:33:19] <Porty> and not sure if I can save files.
[03:33:27] <Porty> to test the port
[03:34:48] <tom3p> again, i thought halrun should be available, yet you say it said 'No command 'halrun' found" i dont have the livecd to test now
[03:36:19] <tom3p> and i gotta restart from some recent updates, bbl
[03:54:43] <Wesbaker> Andypugh: Thanks for the info on the file extension directive in the ini file. I'll check that out.
[04:16:45] <Guest524> hello
[04:16:45] <the_wench> hello Guest524, you have a question?
[04:17:39] <Guest524> Yes, just found this chat and am not familiar with it. Looking for help.
[04:21:16] <Guest524> Are theer rules to posting questions?
[04:22:04] <Valen> just to ask it in its full form
[04:22:11] <Valen> and don't expect a reply quicky lol
[04:22:29] <Guest524> tnx valen.
[04:23:44] <Guest524> I've been playing with EMC2 and have built a DIY CNC. I've gotten it working and running. However, having trouble configuring it.
[04:25:23] <Guest524> I have a cheapo 4-axis stepper board with Toshiba TB6560 driver chips and can't seem to get the configuration of the stepper conf to be accurate. Anyone have any experience with the Toshiba chips?
[04:26:09] <Portly> hello again
[04:26:48] <Portly> ok, i tried the Parallel_Port_Tester and I am getting no change in my LEDs
[04:27:12] <Portly> I installed linuxCNC on my computer
[04:27:38] <Guest524> I've found some usefull info on the EMC2 forums, but nothing relating to the stepper configuration. The datasheets for the Toshiba TB 6560 is not much help.
[04:28:45] <Portly> I made this schematic here. http://www.epanorama.net/circuits/lptleds.gif
[04:29:14] <Portly> except the ground pin I only haved hooke up to pin 25
[04:30:30] <Portly> any ideas why I am getting pin 2,3,4,5,6 with dim lit leds and 7,8,9 off. (This is how it is right from turing on my computer)?
[04:30:45] <Portly> when I try to do the port tester nothing happens.
[04:31:00] <tom3p> is this what you tried? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Parallel_Port_Tester
[04:33:32] <Portly> yes
[04:33:47] <tom3p> trying it now...
[04:34:14] <Portly> in my bios I have the port set to ecp
[04:34:42] <tom3p> does the window popup?
[04:34:46] <Portly> i don't think the bios setting make much difference though
[04:34:54] <Portly> yeah the window pops up
[04:35:03] <Portly> I can click the buttons.
[04:35:11] <tom3p> do the colors change when you click the buttons?
[04:35:14] <Portly> and they turn green in the program
[04:35:19] <Portly> yes
[04:35:32] <Portly> but my led's do not from my parallel port
[04:36:32] <tom3p> ok, the change means the software outputs the command, the lack of change on the port says the command doesnt make the change you want
[04:37:12] <tom3p> is there 1 input that is red and others all green?
[04:37:44] <Portly> 10 - 15 are all green
[04:38:59] <tom3p> ignore last q, it may be my hardware
[04:39:52] <Portly> I really want to make this CNC machine. But this port thing is driving me batty
[04:40:41] <Portly> is anyone else having problems with the port not getting the messages sent?
[04:40:59] <tom3p> try point 14 on http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Testing_Parallel_Port_Outputs
[04:41:37] <Valen> Portly: what mbo you using btw?
[04:44:06] <tom3p> btw use ^D to exit the pport tester after closing the window
[04:46:25] <tom3p> and the reference to "Hal COnfiguration Screen" means you should be running EMC2 with Axis display already
[04:46:26] <Portly> will check
[04:48:38] <Portly> my computer is this one HP_Pavilion_Media_Center_a1310n_Desktop_PC
[04:50:28] <tom3p> Valen were you asking because some mobos are 3.3V and others 5V?
[04:50:38] <tom3p> coffee shop closing bye bye
[04:51:11] <Portly> my mobo is Motherboard Name: A8AE-LE
[04:51:40] <Portly> my computer = http://h10025.www1.hp.com/ewfrf/wc/document?docname=c00571922&tmp_track_link=ot_faqs/top_issues/en_us/c00571922/loc:3&lc=en&dlc=en&cc=us&product=1815823
[04:53:59] <Portly> motherboard specifications = http://h10025.www1.hp.com/ewfrf/wc/document?docname=c00496280&tmp_track_link=ot_recdoc/c00571922/en_us/c00496280/loc:1&lc=en&dlc=env&cc=us&product=1815823
[05:03:10] <Portly> ok, i tried point 14 on http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Testing_Parallel_Port_Outputs
[05:03:21] <Portly> no change to my LED's when doing this
[05:07:58] <Portly> anyone got any ideas why my parallel port is not responding? I have to be missing something.
[05:08:28] <omb> How is it not responding?
[05:09:21] <Portly> ok, I made the 8 led circuit, hooked it up to my parallel port.
[05:09:37] <omb> WHat are you using to light the LED's?
[05:10:19] <Portly> pin 2-6 are lit dimmly right from turning on the computer. 7-9 are not. and not matter what i do with prot tessing, etc. there is no change in the leds'
[05:10:38] <Portly> http://www.epanorama.net/circuits/lptleds.gif
[05:10:46] <Portly> this is the circuit
[05:11:14] <Portly> <tom3p> try point 14 on http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting#Testing_Parallel_Port_Outputs
[05:11:20] <Portly> this is the port tester
[05:11:38] <omb> Sorry, I'm new to the chat. Can you explain "prot tessing"?
[05:12:20] <Portly> ok. one sec.
[05:12:52] <Portly> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Parallel_Port_Tester
[05:13:02] <Portly> this is the tester, it explains
[05:13:32] <Portly> click a button and it turns on the 5v from that pin.
[05:13:43] <Portly> but it does nothing for me
[05:14:01] <omb> Cool, what OS are you using?
[05:14:10] <Portly> I installed the linux cnc.
[05:14:21] <omb> on what OS?
[05:14:28] <Portly> ubuntu
[05:15:18] <Portly> my computer = http://h10025.www1.hp.com/ewfrf/wc/document?docname=c00571922&tmp_track_link=ot_faqs/top_issues/en_us/c00571922/loc:3&lc=en&dlc=en&cc=us&product=1815823
[05:15:51] <Portly> I also tried accessing the port from windows, same thing
[05:19:26] <omb> Not sure if the following will be helpfull, but has some interesting info on P Ports
[05:19:28] <omb> http://www.beyondlogic.org/spp/parallel.htm
[05:20:40] <Portly> ok, will check it out
[05:23:07] <omb> Are you using the port on the motherboard?
[05:23:32] <Portly> would you happen to know why when I hook up a voltometer I get more voltage from hooking up the ground 25 pin than the other pins?
[05:24:45] <omb> Do you mean you get a higher voltage when testing between pin 25 and ground?
[05:25:42] <Portly> pin 18-25 are grounds
[05:25:54] <Portly> so from pin 2 to 25 = more voltage
[05:26:04] <Portly> from pin 2 to 24 about half
[05:26:21] <omb> What are the voltages?
[05:28:04] <Portly> ok, between 2-25 when my voltometer is set to 1000 DC Voltage I get a reading of 5
[05:28:54] <Portly> actually all the other ground pins read 0
[05:30:40] <omb> According to http://www.nullmodem.com/DB-25.htm#parallel you should only see 5v between (2 and 24) and (2 and 25). Pin 2 is (DATA) and pins (18-25) are ground.
[05:31:35] <Portly> yeah, so why am i getting different readings when using different ground pins?
[05:31:50] <omb> You should see a reading of 5v from pin 2 to (18-25) assuming pin 2 is high (i.e 5v).
[05:32:34] <Portly> yeah so it should not matter what ground pin right?
[05:32:47] <Portly> using pin 24 or 25 should be the same
[05:37:47] <omb> Different info on DB25 pinouts located at http://zytrax.com/tech/layer_1/cables/tech_rs232.htm#pins
[05:38:25] <Portly> ok, checking it out
[05:38:31] <omb> what volateg to you get measureing between pin 2 and 7?
[05:41:21] <Portly> checking
[05:42:09] <Portly> 0v
[05:42:52] <omb> according to http://www.lammertbies.nl/comm/cable/parallel.html#pins pins (18-25) are ground. So you should see 5v when measuring pin 2 to any of (18-25), assuming pin 2 is high (i.e. 5v).
[05:43:31] <omb> If you are getting anything other than 5v...
[05:43:39] <omb> 1. Something is fried.
[05:44:11] <omb> 2. Check the bios and make sure you have the parrallel port turned on
[05:45:01] <Portly> yeah it is enabled
[05:45:01] <omb> 3. make sure what ever you are using to turn on the port is properly configured
[05:46:12] <Portly> i checked the bios and tried the different modes. it is lpt 1 0x378 all the same as what linuc cnc uses.
[05:46:24] <omb> do you have another PC to try with?
[05:47:12] <omb> that is....try your tester circuit with
[05:47:44] <Portly> it is a really old 486 comp but it has a parallel port. :)
[05:55:16] <Portly> hmm, I will give that a shot. thanks for the help
[05:55:35] <omb> have a look at 9 and 10 of the following. Hope it helps! http://wiki.linuxcnc.org/emcinfo.pl?TroubleShooting ..... Good luck!
[10:01:44] <theorb> theorb is now known as theorbtwo
[10:04:10] <archivist_emc> archivist_emc is now known as archivist
[10:51:09] <_Fox_Muldr> _Fox_Muldr is now known as Fox_Muldr
[11:00:23] <psha> mhaberler_: good day
[11:00:44] <mhaberler_> psha: good morning!
[11:01:39] <mhaberler_> ok, will give them widgets a spin shortly
[11:02:26] <psha> i've added zero point to bar
[11:02:45] <psha> so if you have +- spindle speed it will grow somewhere from center and not from min side
[11:04:33] <psha> color zones are still calculated from min side and not from zero point
[11:04:56] <mhaberler_> I've started repainting my pyvcp panel in glade, yes
[11:04:58] <mhaberler_> the interesting part will be doing probing.. now that mdi ocalls and probing work that could be an approach; my current one is pathetic with hal: http://mah.priv.at/cgi-bin/viewvc.cgi/cvs/emc-fraese-hm2/pymdi.py?revision=1.4&view=markup
[11:06:05] <mhaberler_> I'll look into not only hooking hal pins to widgets but also passing a user.py file to gladevcp fro some normal callback handling
[11:07:39] <psha> mhaberler_: maybe some sort of 'EMC script' button would be sensible
[11:08:21] <psha> something that runs you function on button click with initialized emc.command object
[11:08:58] <psha> wow
[11:09:14] <mhaberler_> at this point, probably just a plain python callback; so it would be standard button linked to py method as usual
[11:09:21] <psha> i thought that nobody is still using cvs in 2010's :)
[11:09:35] <mhaberler_> I'm phasing out... yes yes
[11:10:27] <psha> git-cvs-import will help you ;-)
[11:11:13] <psha> python scripts have some disadvantages over G-code
[11:11:34] <psha> there is no reliable way to know that command is finished
[11:34:27] <mhaberler> psha: looks great! http://imagebin.org/125151
[11:35:37] <psha> i'll add 'force width' 'force height' properties
[11:36:16] <psha> without them you have to check off 'fill' and 'expand' packing properties
[11:36:46] <psha> btw there have to be labels for values (min,cur,max)
[11:37:01] <psha> and they are missing on your screenshot...
[11:37:21] <mhaberler> I was looking for those.. hold on
[11:38:27] <mhaberler> well? which properties are these?
[11:39:41] <psha> choose 'packing' and uncheck 'fill' and 'expand'
[11:39:51] <psha> forced width/height are not commited yet ;)
[11:41:09] <mhaberler> I was referring to the labels..dont see a prop for that
[11:41:40] <psha> i've found where are labels :)
[11:41:51] <mhaberler> share that secret!
[11:41:56] <psha> i've commited code with commented lables to make icons :)
[11:43:45] <psha> it will be back in ~10 minutes with forced width/height props
[12:02:08] <psha> mhaberler: pushed
[12:02:20] <mhaberler> ok
[12:02:43] <psha> labels + forced height/width not depending on widget size request
[12:04:15] <psha> so bars won't bloat on resize
[12:06:34] <mhaberler> they are perfect: http://imagebin.org/125153
[12:07:55] <psha> try setting force width to 30 for VBar and force height to 20 for HBar
[12:08:01] <psha> they'll look nicer
[12:09:06] <psha> also when they are longer linear gradient looks better :)
[12:11:19] <mhaberler> good to go! http://imagebin.org/125154
[12:12:57] <psha> working :) last thing that have to be added is adjustable font and it will be done
[12:16:37] <psha> be back later
[12:17:03] <mhaberler> psha:cu; working on usermodule import into gladevcp
[12:17:27] <jthornton> you guys are having too much fun
[12:17:54] <mhaberler> warp speed!
[14:09:23] <kb8wmc> G'morning all
[14:10:34] <andypugh> Good Afternnon.
[14:30:33] <kb8wmc> hey there Andy
[14:48:22] <psha> mhaberler: custom handlers looks fine
[14:49:15] <psha> but i've strong feeling that something is wrong... still don't understand what ;)
[14:54:01] <SteveStallings> SteveStallings is now known as steves_logging
[15:10:39] <mhaberler> psha: hi - back from breakfeast ;-) what's fishy with the handlers?
[15:12:08] <JT-Hardinge> one down (cycle timer) one to go (ngcgui)
[15:26:24] <psha> mhaberler: i believe that this is better to do via 'wrapper' script
[15:26:59] <psha> something like you'r foo.py but with last line 'import gladevcp; gladevcp.doitforme(handlers)'
[15:27:18] <mhaberler> psha: ah, upcall. sure-
[15:27:33] <psha> your solution is more generic since it allows combinations of different files
[15:27:37] <psha> foo, bar, ...
[15:27:39] <psha> it's good
[15:28:17] <mhaberler> so in that case gladevcp.py would be a module just exposing the doitforme method, or the whole thing?
[15:29:37] <psha> 90% of gladevcp is option parsing
[15:29:56] <mhaberler> fair enough. will doit.
[15:30:41] <psha> give it some time to lay untouched :)
[15:30:59] <mhaberler> will start to smell
[15:31:03] <psha> maybe today better solution will come to us :)
[15:31:45] <mhaberler> I was about to make handlers a method which will be called by gladevcp
[15:34:05] <psha> wrappers are ugly...
[15:34:49] <psha> you solution is not
[15:35:01] <mhaberler> maybe there's a way through an init method
[15:35:02] <psha> but i think that 'handlers' dict is not needed
[15:35:15] <mhaberler> its gone already
[15:35:24] <psha> you may enumerate module contents with 'dir'
[15:35:29] <psha> and leave only callables
[15:36:04] <psha> i'll be back soon
[15:42:25] <mhaberler> psha: I know, but then I potentially would import garbage handlers - maybe with a pattern match like on_*
[15:45:53] <psha> mhaberler: no harm since only handlers which are set in glade will count
[15:46:10] <psha> nothing bad in 'i', 'MAX_INT', 'f' methods
[15:46:19] <mhaberler> good point
[15:46:24] <psha> if there are no such signal handlers defined
[15:49:08] <mhaberler> I get the dir(mod) thing.. otoh: how do you select functions only from dir?
[15:50:30] <psha> for f in dir(m): if not hasattr(m, '__call__'): continue
[15:50:35] <psha> oops
[15:51:01] <psha> for n in dir(m): f = getattr(m, n); if not hasattr(f, '__call__'): continue
[15:51:10] <psha> one more stage was needed
[15:51:17] <mhaberler> thanks
[15:51:39] <psha> otherwise store handlers[n] = f
[15:52:09] <psha> non obvious point is what to do with equal names?
[15:52:22] <psha> last overrides first? maybe it's better to call all of them?
[15:52:55] <mhaberler> shit happens; or distingush by mod name and construct names like foo_on_event..., bar_on_event
[15:53:25] <mhaberler> gets py dependency into glade though
[15:54:00] <psha> py dependency is in glade already :) via gladevcp
[15:54:30] <psha> no, problem is not in 'foo_on_event', 'bar_on_event'
[15:54:40] <psha> consider 'quit' event
[15:54:50] <psha> connected to detroy signal of main window
[15:55:00] <mhaberler> yeah, but the event handler names shouldnt care about which source file they are defined in
[15:55:02] <mhaberler> yes, I saw that connect
[15:55:23] <psha> both foo and bar wants to handle quit
[15:55:25] <psha> for some cleanup
[15:55:37] <psha> is it possible to connect both of them to 'quit'?
[15:55:53] <mhaberler> will think about it
[15:56:39] <mhaberler> would need a trampoline function I guess
[15:58:30] <atmega> anyone speak/read russian?
[15:59:34] <atmega> if so, does http://services.achilies.biz/ look legit?
[16:04:27] <mhaberler> psha: autodetect is in my repo; no double func calls yet
[16:04:37] <mhaberler> more elegant! thanks
[16:05:39] <psha> atmega: text on the page is sensible but i don't know if it's worth to trust them :)
[16:06:53] <psha> they state that they are working till 2006 but it's hard to tell if it's long or short :)
[16:11:22] <atmega> psha: that guy bought a radar antenna from me on eBay
[16:14:15] <psha> via this service?
[16:14:58] <atmega> no, he just had that URL in his message... normal ebay transaction
[16:16:32] <poppabear> Greetings One and All.......
[16:18:14] <mhaberler> psha: multiple ident handler names are called in import order. up in repo.
[16:18:53] <mhaberler> psha: call as gladevcp -d -u foo -u bar -u baz -H gladevcp-test.hal gladevcp-test.glade
[16:20:01] <psha> poppabear: hi
[16:20:03] <psha> mhaberler: good
[16:21:31] <poppabear> hey psha, you live on here? hehehhehe
[16:21:45] <mhaberler> psha: hold on.. does work despite glaring bug..
[16:21:58] <poppabear> I see you hear every time I come on........ mabey its me..... :)
[16:23:19] <psha> :) my active time is in the evernings :) so you just choose proper moment :)
[16:23:38] <psha> mhaberler: are all callbacks done with one param?
[16:23:47] <psha> for button clicks it's true
[16:24:02] <psha> but for other? value-changed signal for example?
[16:24:02] <mhaberler> psha: I think gtk does multiple registers for us...
[16:24:04] <mhaberler> re: numargs: good question
[16:24:14] <mhaberler> maybe add named default params
[16:24:23] <psha> with gtk you may connect several signal handlers
[16:24:35] <psha> question was if signal_connect supports this or not
[16:24:46] <skunkworks> poppabear: there are a lot of people logged in 24/7 - the just might not be 'here'
[16:24:46] <psha> if you do it by hand it's ok
[16:24:53] <skunkworks> *they
[16:25:08] <mhaberler> it obviously does despite my bug
[16:25:31] <poppabear> skunkworks....... got it
[16:26:04] <mhaberler> psha: then I failed to understand the problem you were trying to solve
[16:26:06] <poppabear> so is the best choice for larger machine installs (analog servos), the MESA stuff?
[16:26:49] <skunkworks> I am using mesa stuff. Love it!
[16:27:48] <psha> mhaberler: it's possible to connect multiple handlers to one signal with 'widget.connect("signame", cb)'
[16:27:51] <poppabear> ok
[16:28:21] <psha> but i don't know if it's possible to pass multiple callbacks (for one signal) into builder.connect_signals() call
[16:28:28] <psha> if it's possible - than trampoline is not needed
[16:28:39] <mhaberler> good. so my stuff isnt needed in the first place. will test with libglade and glade funcs
[16:28:46] <andypugh> I think that the Pico stuff probably works well on big machines too, and is rather more of a system than a kit of parts.
[16:28:56] <skunkworks> My current project is using 2 5i20 boards (72 i/o each) with 2 servo daughter boards (one 6 axis and one 4 axis) so it gives me 96i/o and 10 encoder counters/ +/-10v
[16:29:19] <poppabear> I am slapping together a 5 axis machine ( 1kw AC servos ), arm type ATC, and pallet changer, my IO count looks around 2-300 est. (usually use a PLC for that kinda of stuff), but it appears I can do most of that by expanding the number or Pins and Analog IO that EMC can do, correct?
[16:30:17] <andypugh> I try not to prefer one over the other when answering forum/email queries, but as I have a lot more involvement with the Mesa stuff than Pico stuff its hard to be equal.
[16:31:55] <andypugh> You can put 768 48V 2.5A output drivers and 768 5-48V input drivers on one 5i20....
[16:32:02] <mhaberler> psha: aha, breaks with gtkbuilder format.. ok with libglade
[16:32:24] <skunkworks> poppabear: http://www.youtube.com/watch?v=KplU8hkI0AQ
[16:33:49] <andypugh> (that would be using a full set of 7i64 cardson a 4-header mesa card)
[16:34:02] <skunkworks> I like that the mesa is directly pluggable with opto22 cards (24 port 50 pin scsi pin)
[16:37:40] <poppabear> skunkworks: I had to make a "comp" file to be able to do movement code inside of a running M6, ( steve at stuarts emc fest showed me how ), I see your moving your Z axis during the m6, did you do a work around as well?
[16:38:35] <andypugh> It might be moving to the pre-defined (M30?) position
[16:40:59] <andypugh> Section 2.2.10 here: http://linuxcnc.org/docs/2.4/html/config_ini_config.html
[16:41:31] <poppabear> for my machine, I need to move the Z axis during the tool change, so I had to "Fool" emc with the comp, so it thought no movement was going on then at the end of the m6 bring the Z back to where it was, and "reconnect" EMC2 to the Z and keep on rolling
[16:42:43] <andypugh> I will have to do something rather similar with my machine.
[16:43:35] <skunkworks> poppabear: I have TOOL_CHANGE_AT_G30 = 1 in the ini
[16:43:50] <poppabear> I understand
[16:43:50] <poppabear> but
[16:44:34] <poppabear> after I reache tool change position, I have to move the Z head up/down to grab/put holders in/out of holder claw
[16:44:48] <skunkworks> poppabear: ah
[16:44:54] <poppabear> so I have to have movement of the Z during the M6, no other way on that machine
[16:45:21] <skunkworks> I did do a comp for the gear shift (16 gear transmission with orient)
[16:45:46] <skunkworks> poppabear: did it work well?
[16:46:05] <psha> mhaberler: libglade support lists in handlers dict and gtkbuilder - not?
[16:46:06] <andypugh> The docs say you can combine tool changes. That means you could use TOOL_CHANGE_POSITION & TOOL_CHNGE_QUILL_UP & TOOL_CHANGE_AT_G30 to get a three move cycle. (And, hopefully, they happen in a consistent order :-)
[16:46:36] <mhaberler> psha: yes; gtkbuilder connect_signals - doesnt work for me yet
[16:47:15] <poppabear> skunkworks: Yes
[16:47:39] <poppabear> andypugh: My CAM post out only M6, for changes, so, I did it all in the M6
[16:48:18] <andypugh> Those are M6 options. I am not saying it works as-documented or is useful, but it might be.
[16:48:40] <andypugh> That is three different INI file options, which allegedly can be used together.
[16:49:06] <andypugh> The docs definitely say that you can do quill up then move.
[16:49:42] <mhaberler> psha: its in repo - connect_signals doc is pretty meagre
[16:49:47] <andypugh> So I think you could have quill-up, move, then down onto a tool. But then you would be stuck.
[16:50:21] <poppabear> true, thats why I did the "pet tricks" in the comp for my m6 routine
[16:50:31] <mshaver> andypugh: Want to continue the 8i20 stuff, or do you want a few more hours to restore your sanity?
[16:50:41] <psha> mhaberler: side note: suffix for gtkbuilder files is .ui
[16:50:52] <mhaberler> ah
[16:50:54] <poppabear> I was VERY encouraged to hear that you could do Axis motion within a custom Mcode as well
[16:51:33] <psha> poppabear: you may do everything, even emc2 recompilation in custom M codes
[16:51:40] <andypugh> I might have it fixed. My config with a motor connected is giving me a null pointer problem. I am unsure why.
[16:51:46] <poppabear> what do you folks use? Bash?, a compiled something??
[16:51:53] <mshaver> poppabear: did you commit your M6 .comp to the git repository?
[16:52:16] <poppabear> no
[16:52:33] <poppabear> it is for a specific/custom machine set up
[16:52:52] <mshaver> do you have an account on the emc git repo?
[16:52:56] <andypugh> How do you move in a custom M?
[16:53:15] <mshaver> i'm thinking it would be a good "example" component
[16:53:40] <poppabear> redirecting the output commands to the input feed back
[16:53:44] <psha> andypugh: import emc; c = emc.command(); c.jog?
[16:53:45] <psha> ?
[16:53:50] <poppabear> so EMC doesnt see things move
[16:54:38] <psha> andypugh: it's not hard i think. have not tried but see no possible issues
[16:55:32] <andypugh> You can link directly to functions inside loaded modules from the command line? I suppose it is obvious that it should be possible, I had just never considered it.
[16:55:56] <psha> andypugh: no. but you may load new module and work inside it
[16:57:01] <poppabear> then pull the motor stats, and do the 3rd order derivitive in the comps for handleing accel/velocity of the motors
[16:57:09] <poppabear> http://video.google.com/videoplay?docid=8421475437071753563#
[16:57:43] <poppabear> that is a Tormach that I did a Mach3 ATC m6 on, so I adapted its logic to do the same thing, but run EMC
[16:57:52] <mhaberler> psha: from web: It used to be the case (at least in GTK+ 2.12, not sure if it's still the same) that you could call connect_signals only once, any signals which are not connected during the first invocation will never be connected. This was different in glade, so be careful if you relied on that feature before.
[16:58:02] <psha> mhaberler: you have too 'connect_signals' calls
[16:58:05] <psha> second one is ignored
[16:58:09] <psha> same :)
[16:58:14] <psha> s/too/two/
[16:58:18] <mhaberler> psha: just found it.
[16:58:43] <psha> btw i don't undestand purpose of this line
[16:58:51] <psha> builder.connect_signals(builder)
[16:59:16] <mhaberler> i dont understand it either
[16:59:17] <mhaberler> psha: in that case, I dont see how multiple handlers can be connected in gtkbuilder
[17:01:11] <skunkworks> poppabear: you need to mount a encoder on the spindle and do some rigid tapping with emc. :) then post on cnczone :)
[17:01:40] <poppabear> :)
[17:02:01] <poppabear> the currently machine I am designning will have a 5 axis trunion
[17:02:20] <skunkworks> are you going to do kins for the machine?
[17:02:29] <poppabear> so, REALLY like the idea of TRUE working kienematics!! Definatly going to retro a Fanuc Robot also
[17:02:40] <poppabear> yes
[17:02:46] <poppabear> kins is fun
[17:02:57] <skunkworks> neat!
[17:02:58] <poppabear> got a handle on it at the emc fest at stuarts
[17:03:00] <poppabear> just math
[17:03:31] <poppabear> the ability for the machine to do True 5 Axis simultaneous milling is AWESOME!!!
[17:04:06] <poppabear> lately, I have been doing 4 axis live, and 5 positional since the CNC control (in this case M3), does not do axis traslations/kienematics
[17:05:09] <psha> mhaberler: i've tried to pass multiple handlers in one dict and it's not working
[17:05:20] <psha> seem that it won't work in that way at all...
[17:05:38] <mhaberler> the trampoline method is still an option
[17:05:43] <mshaver> is this for a retrofit? (5 axis)
[17:05:46] <psha> so either to leave 'trampoline' or ignore first/last handles
[17:06:13] <mhaberler> I think the trampoline makes sense, I'll get it back in
[17:06:17] <poppabear> mshaver: a ground up, design/build
[17:06:22] <mshaver> andypugh: got the comp file - need a minute to compile, etc.
[17:06:35] <andypugh> No worries.
[17:06:44] <poppabear> thinking of using Polymer concreate for the machine base/sturcture with a steel "skeletion"
[17:06:44] <mshaver> whole machine, castings & all?
[17:06:55] <mshaver> ah
[17:07:13] <mshaver> for you, or as a product?
[17:07:17] <poppabear> yeap, most of my buiz, is retrofit or custom build
[17:07:22] <poppabear> both
[17:08:10] <skunkworks> poppabear: does mach3 do blending on more then 3 axis?
[17:08:14] <mshaver> wish I could get the customers - or at least customers who waned to pay a price that would make the work required pay better than minimum wage :)
[17:08:17] <andypugh> Thinking about it, magnetically homing an 8-pole spindle is simple and probably fine. It will only ever twitch by 45 degrees, and will not be under load at the time.
[17:08:21] <poppabear> Arm type random ATC, pallet changer, spindle chiller, swarf convayor, lights, bells, whistles... Hooter Girls...
[17:08:31] <poppabear> skunk: yes
[17:09:13] <andypugh> I think I would prefer to be independently wealthy and _be_ the customer.
[17:09:17] <skunkworks> poppabear: http://www.youtube.com/watch?v=9xDPqFXo_5w&feature=mfu_in_order&list=UL
[17:09:39] <poppabear> mshaver: mostly I go after OEMs, and Industrial clients, there is no money in hobbie as an SI
[17:12:11] <mshaver> I do work for a couple OEMs, but they really aren't interested in paying much - industrial customers with specific needs tend to be better, but the work is transitory - project by project
[17:13:13] <poppabear> mshaver: my findings as well
[17:13:18] <poppabear> were are you located?
[17:13:46] <andypugh> I actually used to work for a company that, if I was still there, EMC2 would have been great for.
[17:13:50] <mshaver> maryland - 20 miles west of baltimore
[17:14:03] <mhaberler> psha: any clue how we can merge the sig handlers from builder and from py files?
[17:14:07] <mshaver> you TN right?
[17:14:14] <poppabear> yes
[17:14:44] <poppabear> mostly hit the 2cd/3rd tier automotive clients here, sadly alot of that work has dwindled
[17:15:00] <andypugh> Brilliant little company, 2 programmers, 2 electrical engineers, 2 mechanical designers, a machinist, 2 wiremen and a fitter. All in one office. Specialised in making bits of kit work together.
[17:15:11] <andypugh> Went bust..
[17:15:16] <poppabear> probably that this point gonna do some turn key control systems to sell, both Mach3 and EMC
[17:15:35] <mshaver> surprised they don't require 100% Allen-Bradley
[17:15:44] <poppabear> most do
[17:16:44] <poppabear> usually I put together a machine cell with AB PLC/HMI, Fanuc Robot, then use Mach3 (or now gonna use EMC) to handle syncronized motion, they only see a "Black Box", I just tell them it is an integrated MOtion cotrol computer
[17:16:48] <mshaver> I've got some plans for turn-key controls as well - kinda been dragging my feet as I don't know how well they would sell
[17:16:59] <poppabear> they nod knowingly with a blank stare
[17:17:41] <andypugh> Hide a computer well enough and folk never feel the urge to upgrade it either.
[17:17:54] <mshaver> do you like the electronics end? or more the SI part?
[17:18:05] <poppabear> I use the Robots for MIG Spot, weld and resitive, and the motion control for loading unloading parts, carrage motion as well
[17:18:32] <andypugh> There were hundreds of microvaxes running x-ray analysers attached to electron microscopes decades after they were obsolete.
[17:18:33] <poppabear> SI hands down
[17:18:58] <poppabear> I just buy the electronics that do what I want
[17:19:17] <mshaver> I've never had a customer with a budget > $10,000, most less than that
[17:19:33] <mshaver> it's hard to justify
[17:19:37] <poppabear> AB is CRAZY assed expensive, so, I have been getting some "Omeron", Mitusbishi and ADC acceptance on smaller jobs
[17:20:10] <poppabear> the last big cell I had was 875,000.00
[17:20:43] <poppabear> for a machine that would clip, punch, weld, a bumper system for a Honda Odysee
[17:21:13] <mshaver> you know, if you have an outlet for this stuff, I'd love to provide you with ready-to-use control hardware - EMC2/Mesa based assemblies
[17:21:30] <poppabear> but they run that kinda machine 24/7 for 4-5 years, then 1-2 years after that for "supply parts" then scrap the machine
[17:21:55] <poppabear> working that angle my self
[17:21:56] <mshaver> they don't reconfigure for a new job?
[17:22:02] <poppabear> the control sys
[17:22:08] <poppabear> block idea
[17:22:27] <mshaver> what form factor for the blocks?
[17:22:33] <poppabear> hang it, configure it, debug it, prove it, get paid, leave......
[17:22:45] <poppabear> standard Operators Panel
[17:23:38] <psha> mhaberler: i'll be back in an hour
[17:23:42] <mshaver> there's a standard size?
[17:24:03] <mshaver> other than 19"
[17:24:06] <poppabear> I downloaded the Xlinx IDE, so I am going to start learning the FPGA programming stuff for the Mesa boards
[17:24:12] <mhaberler> ok. think its starting to work - the first connec_signals isnt needed...
[17:24:23] <poppabear> I HATE the idea of another learning curve on top of all this other stuff
[17:24:45] <poppabear> for industrial touch 17"
[17:25:21] <poppabear> for EMC2, since wide scxreen is so common there are cheaper than 17, or 19 in touch panels, I use the ELO 22.5"
[17:25:24] <poppabear> back mount
[17:25:47] <poppabear> I am hoping to be up to speed on this EMC in about 3-6 months or so
[17:26:13] <mshaver> Mach3 no longer enough for your apps?
[17:26:18] <poppabear> building a Table top sized, proof of concept, 5 axis trunion machine to play on
[17:26:26] <poppabear> not for what I need to do
[17:26:59] <mshaver> understood
[17:27:25] <poppabear> Honestly Mach is much easier to set up, use, modify..... but I will be using EMC for the stuff that HAS to run with out down time
[17:27:37] <poppabear> Windows in the end it the REAL killer of Uptime
[17:27:59] <poppabear> I have run Mach on embedded Win CE, and it still (windows), crashes
[17:28:11] <poppabear> service cost me to much
[17:28:11] <skunkworks> <sarcastic> I thought mach4 was going to take care of all of this</sarcastic>
[17:28:19] <poppabear> if under warrenty
[17:28:25] <mshaver> I have a friend down in TN who I work with on CNC/machining (Lebanon) - next time (if there is one) I'm down that way I should stop by!
[17:28:32] <poppabear> unknown if it will at this time
[17:29:01] <poppabear> Rev 4, has been promised for over a year now.........
[17:30:06] <mshaver> they did hire a new programming guy
[17:30:41] <poppabear> also, if you do high end work, and you have to have working Modbus serial and Ethernet, Brains and other higher end functions of mach, when you ask for them to be fixed (the bugs.....), you are pushed to the back burner since M3 caters to the bulk, simple machine crowd....... I have waited on fixes for several years now, and no joy...
[17:31:20] <poppabear> no just Brian, and one single new "support guy".
[17:32:01] <poppabear> mshaver: if you come down, hollar, BTW, if you want to do some JOINT project work, at your end or mine, give me a hollar
[17:32:29] <poppabear> send me your rates, requirements, liability coverage amount.. etc.... to
[17:32:32] <poppabear> :
[17:32:37] <mshaver> well, that's kind of where I was headed
[17:32:43] <poppabear> Poppabear<at>hughes.net
[17:33:29] <skunkworks> how has your satellite internet been running? (I had it for a few years a few years ago)
[17:33:40] <poppabear> it SUCKS!!!!!
[17:33:49] <poppabear> I HATE HUGHES NET!!!!!!
[17:33:57] <poppabear> ........sorry.........SORE nearve there
[17:34:15] <skunkworks> heh
[17:34:21] <skunkworks> I remember that :)
[17:34:23] <poppabear> they are suppossed to get Cable in my area soon, and when they do....... I AM SO GONE
[17:34:26] <skunkworks> I was hoping it got better
[17:34:59] <mshaver> Rate: $75/hour ($85 if traveling), but I could accept more Requirements: Checks must not bounce :) Liability Coverage: I don't have any
[17:35:00] <skunkworks> do they still have the 'bucket' download throttling?
[17:35:05] <poppabear> they limit you to a max of 200mb download/24 hrs and if you exceed, they PUNISH you with less than dial up for the following 25 hours
[17:35:27] <poppabear> you willhave to have liability
[17:35:30] <skunkworks> yep - that is what I remember. I think it was 100mb when I had the service.
[17:35:38] <poppabear> all shops in my area require it
[17:36:03] <poppabear> if you can do 75/hr, and 45 when traveling that may be doable
[17:36:18] <poppabear> industry wont pay any more than that
[17:36:41] <poppabear> been there done that, UNLESS you are a Fanuc Robot, or Fanuc Control tech.........
[17:36:43] <poppabear> hehheehe
[17:36:52] <mshaver> I could get it, but it would have to be figured into the cost. My problem is that all of my work has been on such a low dollar level that there's no way you could support that
[17:37:06] <mshaver> why less $ on the road?
[17:37:26] <poppabear> if you want to run with the big dogs, you gonna have to put up, or shut up
[17:37:36] <mshaver> I don't mean travel time - I mean if I work away from home
[17:37:42] <poppabear> Liability and sometimes Bonding is required
[17:37:49] <mshaver> I don't really charge for travel time itself
[17:38:04] <poppabear> they only pay an hourly rate period
[17:38:18] <mshaver> $75 would be fine if it was for more than say 2 days
[17:38:32] <poppabear> sample is 75/hr under 10 hour day, overtime after that, traveling to/from 45/hr.
[17:39:02] <mshaver> that would be fine!
[17:39:19] <poppabear> I suggest you start bidding some OEM/Industrial jobs in your area, and get your name out.
[17:39:45] <IchGuckLive> flyers are a good idea
[17:39:57] <IchGuckLive> or its time for X-mas cards
[17:40:03] <mshaver> Maryland dude - there is no industry!
[17:40:03] <poppabear> HIGHLY recommend that you set your self up as an LLC, otherwize taxes will totaly KILL you
[17:40:20] <poppabear> have car? will tavel?
[17:40:28] <poppabear> look on the internet for work
[17:40:39] <mshaver> I've been a corp since 1987, so that end I have covered
[17:40:41] <poppabear> I travel all over the US, and world, go where the work it
[17:40:49] <poppabear> NO LIFE when traveling
[17:41:18] <IchGuckLive> poppabear: agree here SIEMENS World Wide
[17:41:26] <IchGuckLive> no live at all
[17:41:38] <poppabear> yeap, you HAVE to be willing to travel
[17:42:13] <IchGuckLive> travel means Airport Job-Hotel-Job-Airport
[17:42:24] <mshaver> do you do extensive planning prior to the on site work?
[17:42:51] <IchGuckLive> for me NO
[17:42:56] <poppabear> and if you go to places like Venezualla, Dubia, and other rough 3rd world, be ready and willing to purchase and use a firearm from a street vendor, otherwize you can/and will be pumked. Yeap, been there, it is cheaper to shoot first, and pay off the cops...... hehehehe
[17:43:14] <IchGuckLive> China as well
[17:44:30] <andypugh> Not wanting to lower the tone too far, but Siemens have an office to the west of London. Some poor secretary has the job of answering the phone with "Hello, Siemens Staines"
[17:44:39] <mshaver> I'm not greedy - I'll let someone else have the jobs in those places :)
[17:45:04] <mshaver> ok, going to the other computer to compile that .comp...
[17:45:12] <poppabear> went to China ONCE, and learned that you get ALL agreements IN WRITTING UP FRONT before you step foot in that country, since after the job is over, they will try and NOT pay, and then they can and WILL hold you and stop your way out of that country until the dispute is resolved!!
[17:45:50] <IchGuckLive> Siemens is searching for 6000 Engineering people around the world
[17:46:35] <poppabear> IchGuckLive: mabey so, but they always want some guy with a "Degree", not self taught guys, who actually do the work....
[17:46:35] <poppabear> hehehehe
[17:46:56] <andypugh> I can lend you one of mine, I am not using it.
[17:47:02] <poppabear> hehhehee
[17:47:09] <micges> poppabear: I agree
[17:47:16] <poppabear> sure, fax it to me, I am sure they dont do background checks.......
[17:47:31] <IchGuckLive> poppabear: i lost a costemised notebook at the customs service :D they found it 1 week later all the stuff wars 2month later on the internet
[17:48:22] <IchGuckLive> poppabear: i now that Papers are all you need ask your russen friend he can give you all you need
[17:48:46] <poppabear> IchGuckLive: I will work for Siemans IF they provide full time SI work, and good pay, and dont require a degree.......
[17:49:16] <poppabear> Played with Sinumeric and Step7 stuff about 4 years ago, havent touched it since then...
[17:49:18] <andypugh> ... Don't forget the Dancing Girls too
[17:49:21] <IchGuckLive> thats inposibal on there thinking
[17:49:24] <poppabear> yeap
[17:49:31] <poppabear> and Spider Monkeys!!
[17:49:56] <IchGuckLive> andypugh: 2500 < 25Years in Goungdong at the fectory
[17:49:56] <poppabear> if no one is slinging POO at you, it just doesnt feel right
[17:50:35] <IchGuckLive> they work 24/7 for 6month then they ged payd and firerd at the same time
[17:50:55] <IchGuckLive> 12H shifts 40 in 1 bedroom
[17:51:22] <IchGuckLive> not allowd to leave the factory for the hole time
[17:51:53] <andypugh> That's how you can buy a USB charger for <50p delivered from china.
[17:52:33] <IchGuckLive> Annual salary 6month 565USDollars - criminal rewarding
[17:53:56] <IchGuckLive> why buying a forklift to load a 25tonns truck when you can get 40 chinese to that in half the time and 1/5th of the money
[17:54:14] <poppabear> well it has been Tons of Fun, but, gotta get back to the grind, Ol' Lady just HAS to go to freaking walmart..... I figure when I die, to be burried there, then she would visit me.....
[17:54:28] <poppabear> later gents
[17:54:37] <IchGuckLive> Hat jop is he aplying for
[17:54:47] <poppabear> and thanks for the info
[17:55:03] <poppabear> gonna go the Mesa route it appears
[17:55:07] <IchGuckLive> what job he wants to run for
[17:56:00] <IchGuckLive> Technical Toolmaker Mastercheffe
[17:56:16] <IchGuckLive> or simple CNC G-code developer on Heekscad ?
[17:57:40] <mshaver> must reboot...
[17:58:34] <IchGuckLive> how far is linux spreed to the machining business in the USA
[18:00:05] <psha> mhaberler: i'm back now
[18:00:40] <mhaberler> hi
[18:00:40] <the_wench> hello mhaberler, you have a question?
[18:00:59] <psha> :)
[18:01:12] <mhaberler> more than one
[18:01:26] <psha> so i think it's better to use trampoline (or something similar)
[18:01:55] <psha> i personaly prefer functinal wrappers to classes but it's up to you
[18:01:57] <IchGuckLive> merry X-mas CNC greetings from me -> http://euroschall.de/postkarte.jpg
[18:03:52] <psha> not only since libglade is slowly considering obsolete
[18:04:45] <psha> so ping me when you are satisfied with current state and i merge it
[18:05:22] <IchGuckLive> Questioon can i drive 1 Servomotor with a paralellport ?
[18:06:52] <mhaberler> psha: its in repo. it works great with libglade, but signal connecting is severely broken with gtkbuilder files
[18:06:53] <mhaberler> I dont see how to extend the list of handlers done with the first connect_signals
[18:07:14] <andypugh> IchGuckLive: Not without extra hardware
[18:07:52] <mhaberler> or a 800 pound parport ;-)
[18:07:58] <andypugh> But with a speed/direction drive , yes.
[18:08:51] <mhaberler> psha: with gtkbuilder, it seems even the scrollbar handlers dont connect correctly
[18:08:53] <mhaberler> I dont get it
[18:09:22] <andypugh> In theory you could do it with nothing more than an IGBT and a software three-phase PWM generator (I have considered writing one)
[18:10:18] <psha> mhaberler: what's purpose of connecting builder handlers?
[18:10:42] <mhaberler> make standard events work?
[18:11:35] <IchGuckLive> i try to search the datrasheed i have from Motor and controler
[18:12:07] <psha> mhaberler: no
[18:12:33] <psha> standard events are not connected at all. gladevcp events are connected in GladePanel
[18:12:46] <psha> i think it's safe to kill it
[18:13:36] <mhaberler> psha: if it dont call the frst connect_signals it seems to work..
[18:14:13] <mhaberler> psha: except the sliders dont work ? did that ever work with gtkbuilder files?
[18:14:29] <psha> you have to add GtkAdjustment to sliders
[18:15:01] <psha> in general settings for slider select adjustment and click 'create new'
[18:15:12] <psha> it's not obvious at first but then it seem very reasonable
[18:15:20] <psha> since you may define one adjustment for many widgets
[18:15:31] <mhaberler> why does it work in libglade output?
[18:16:40] <psha> libglade does not handle out-of-widgets adjustments,models,etc
[18:16:55] <psha> so when loading UI definition with glade it automaticaly creates stub ones for you
[18:17:13] <psha> gtk builder works with out-of-widgets definitions
[18:17:18] <mhaberler> so its not a bug I introduced..
[18:17:22] <psha> yes
[18:17:31] <psha> you've just forgot to add adjustment
[18:17:43] <psha> i agree that it's not obvious at first
[18:17:49] <IchGuckLive> there is the pdf -> http://tinyurl.com/3af29ws some one can help me with this i never setup a Servo machine
[18:18:11] <mhaberler> mucho better!
[18:18:11] <psha> to create combobox you have to go for a long way :)
[18:18:29] <IchGuckLive> i own a R7M-A10030 S1-D motor
[18:18:30] <mhaberler> ok, but then that is out of band..
[18:18:45] <IchGuckLive> 100w 230V
[18:19:32] <andypugh> Do you have the matching drive?
[18:19:47] <IchGuckLive> and a XD-01MN01 XtraDrive
[18:19:57] <IchGuckLive> yes
[18:20:12] <IchGuckLive> they fit together
[18:20:18] <andypugh> I think the SmartStep drive would have been easier to interface with EMC2
[18:20:26] <mshaver> andypugh: I tried the new bldc.comp, it does do more during the init process. Two problems are: 1. after init, the motor runs away (with the same pid.0.Pgain and bldc encoder polarities as before), and 2. I can't remove the component by an "unload". It complains that bldc is not in /proc/modules and I have to reboot to clear memory.
[18:21:10] <andypugh> 1) is because the Pgain sign it wrong, I reckon.
[18:21:12] <IchGuckLive> andy it wars 1 kit for 50euro
[18:21:57] <andypugh> mshaver: Or, alternatively, try -90 as the drive-offset parameter
[18:22:41] <IchGuckLive> i only whant to test the kit if it works but there where no Cables with it supplyd what a miss
[18:22:55] <psha> mhaberler: i need to merge all your gladevcp fixes in one commit :( there is too much noise on userfuncs branch
[18:23:28] <mhaberler> wait a bit... not yet done. Do as you like.
[18:23:40] <mshaver> thoughts on not being able to unload? that's a killer problem
[18:23:57] <andypugh> open another window and type "dmesg"
[18:23:59] <psha> mhaberler: i'll push it in minutes, i'll do second merge if you add something
[18:24:06] <mshaver> doing...
[18:24:40] <mhaberler> commit itch?
[18:26:32] <andypugh> IchGuckLive: It looks like you should be able to connect that drive direct to a stepgen via the P-Port
[18:26:44] <mhaberler> ok, give it a try now, I guess we're good with both ui and glade file, multiple handlers
[18:27:58] <psha> mhaberler: side note: use 'continue' in loops, that's save you lot indenting
[18:28:26] <mhaberler> line?
[18:29:35] <psha> for example ImportError handling
[18:29:48] <psha> try:... except ImportError:... else: ...
[18:30:01] <andypugh> IchGuckLive: You can drive that drive pretty much any way you like. Either as a smart-positioning drive with step/direction or as a speed/direction or torque/direction drive with PWM from emc2 (converted to voltage by a simple filter)
[18:30:04] <psha> else is not needed if you place 'continue' in except test
[18:30:16] <psha> i'll push it in a short
[18:30:23] <mhaberler> good point
[18:30:37] <IchGuckLive> thanks andypugh
[18:30:57] <IchGuckLive> i will try to find some fitting cables
[18:31:15] <psha> also there is useless 'if handlers:'
[18:31:26] <andypugh> The cables will cost twice what you already paid. Each.
[18:31:34] <psha> if there are no handlers no harm to call 'connect_signals({})'
[18:31:39] <psha> it'll just connect nothing :)
[18:31:43] <mhaberler> I#m not even pretending I'm a programmer
[18:32:00] <psha> i've never told that code is bad :)
[18:32:03] <psha> but it may be better :)
[18:32:30] <andypugh> Hey guys, stop being reasonable, we were looking forward to a fight.
[18:32:53] <mhaberler> I am ot suffering from insanity. I enjoy every minute!
[18:33:04] <psha> andypugh: i think pluto boards are better then mesa!
[18:33:06] <psha> andypugh: :-P
[18:33:17] <psha> * psha have not even seen any of them
[18:33:30] <andypugh> On a simple price comparison, they win hands-down.
[18:33:31] <mhaberler> psha: you didnt miss much.
[18:33:39] <mshaver> ok, the "can't remove" problem has disappeared.
[18:34:00] <mshaver> I now get bi-directional motion as well!
[18:34:05] <andypugh> Yes, I was puzzled by that, it worked for me.
[18:34:21] <mshaver> the speed tops out at the same values in each dir
[18:34:21] <andypugh> Should be faster and smoother too?
[18:34:43] <mshaver> yes, except at 0 speed it oscilllates
[18:35:00] <andypugh> Driven by PID or direct?
[18:35:07] <mshaver> pid
[18:35:22] <andypugh> Add a couple of encoder counts of deadband
[18:35:34] <mshaver> doing...
[18:35:35] <psha> mhaberler: also, builder prints warnings about missing handlers so no need in doing it again (i'm about commented code)
[18:35:53] <pcw_home> (or some D term)
[18:36:00] <andypugh> Or less P.
[18:36:20] <andypugh> Basically you are into PID tuning now, my work is done :-)
[18:36:30] <pcw_home> especially if you use the new PID comp with input for velocity
[18:38:42] <IchGuckLive> andypugh: so shoudt i cut the plugs and do my own connection ?
[18:40:17] <mshaver> ok, quiet now at 0
[18:40:27] <andypugh> I found some hard-drive power connector internal crimps that fitted the motor pins, and that is how I am experimenting with my servos. Basically individual wires to the pins as the proper connectors are £16 each.
[18:40:41] <mshaver> I think you are off the hook for now! :)
[18:40:55] <andypugh> mshaver: What's your max current at the moment?
[18:41:31] <mshaver> Here's the next part: 1. maybe try the hall sensors - see if they are worth the trouble
[18:41:44] <andypugh> For a spindle, possibly not/
[18:42:04] <mshaver> hold on (the current commanded is "1.0" at top speed)
[18:42:08] <andypugh> Though it is a better way of doing it.
[18:42:36] <pcw_home> Andy does your "thump" start routine change the angle (to guarantee start if at TDC)
[18:42:59] <andypugh> Yes, 1 at top speed is right, but that only means == 8i20.max_current. Last time i saw your hal file that was set to 2
[18:43:04] <mhaberler> psha: cleaned up, all examples working including rules, spinbox, got rid of some warning etc
[18:43:33] <mshaver> 7.5A
[18:43:41] <andypugh> pcw_home: It does a slow ramp to -90, then ramps to +90 and then to zero, holds a little while, then measures.
[18:43:50] <mshaver> I should see what raising this does
[18:44:11] <Guest796> hello
[18:44:11] <the_wench> hello Guest796, you have a question?
[18:44:27] <andypugh> She's a chatty one, this wench
[18:45:03] <Guest796> rtapi.conf and SMI
[18:45:11] <andypugh> Ah, yes, that
[18:45:13] <Guest796> do we still need to edit this file?
[18:45:25] <andypugh> I believe so.
[18:45:49] <andypugh> I haven't tried 10.04 on my SMI-infected machine.
[18:45:56] <Guest796> so where is the best place to find the info?
[18:45:59] <pcw_home> OK good, we had some 7I39 users that didn't do it and had intermittent startup troubles (start at ~180 degrees fails , because of 0 torque)
[18:46:43] <andypugh> 7i39 + bldc.comp works a treat in all modes.
[18:47:00] <mshaver> Also: 2. This is a 1kW motor we have lots of - we rigged it up with an encoder to see if this idea had any hope of working (which it certainly seems to!) - Now I need to get an encoder mounted on the 2.2kW motor we really want to use, heatsinking on the 8i20, and the whole mess on a machine to try
[18:47:39] <Guest796> is there a test to see if you computer has the SMI problem and the file needs to be edited?
[18:47:46] <andypugh> If you have Hall sensors you could skip the encoders. bldc.comp can drive an 8i20 on hall-sensors alone
[18:48:10] <pcw_home> Neat, that means that you could make a bigger 7I39 with just an IGBT module and isolation
[18:48:17] <andypugh> You will be trapezoidally comutated, but that might not matter all that much.
[18:48:37] <mshaver> 3. Also, it would be nice to have a bldc.0.brake pin (or an 8i20.0.brake pin) that we could use to hold the spindle still
[18:48:59] <Guest796> does one just use the latency test to see if there is a problem?
[18:49:02] <mshaver> 4. rigid tapping
[18:49:05] <mhaberler> psha: gee, I guess I start to like this - this was a glaring omission from pyvcp and now we have it!
[18:49:15] <pcw_home> I dont suggest coarse commutation with the 8I20 (steps into the PID are not desirable)
[18:49:32] <mshaver> plus, don't I need an encoder for smooth low speed rotation?
[18:49:33] <andypugh> Guest796: Have you seen http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?FixingSMIIssues
[18:51:12] <mshaver> * mshaver looks to see what these 7i39s and 7i64s are...
[18:51:13] <andypugh> pcw_home: Would it be OK in direct mode?
[18:51:17] <psha> mhaberler: http://psha.org.ru/cgit/psha/emc2.git/commit/?h=gladevcp-userfuncs2
[18:51:48] <psha> list of cleanups: killed 'paths' dict since you may check paths in sys.path directly
[18:51:56] <psha> load trampolines late, not early
[18:51:58] <Guest796> they talk about 9 and 10 I am using 8.04
[18:52:06] <psha> killed connect_builder flag (useless as i may see)
[18:52:18] <Guest796> is there a older version of this doc?
[18:52:30] <Guest796> or should I upgrade to version 10
[18:52:50] <psha> Guest796: first try running from livecd
[18:52:55] <psha> don't upgrade blindly!
[18:53:03] <andypugh> Guest796: Which EMC2 version?
[18:53:04] <mhaberler> psha: I'll pull and give it a try
[18:53:30] <psha> it won't merge cleanly on your branch since there are many cleanups and all merged into one commit
[18:53:53] <psha> that's why i've pushed it as 'userfuncs2' branch not userfuncs :)
[18:54:06] <psha> oops
[18:54:09] <andypugh> Ubuntu 8.04 + EMC2 2.3.0 also already has the smi.ko object, so the Section4 instructions apply
[18:54:09] <psha> don't pull
[18:54:18] <psha> i'll amend it with small patch
[18:55:07] <Guest796> axis 2.4.6
[18:55:18] <Guest796> so use section 4
[18:55:22] <Guest796> right?
[18:55:30] <andypugh> Oh, you definitely should have a precompiled smi.ko with that.
[18:55:35] <psha> http://psha.org.ru/cgit/psha/emc2.git/commit/?h=gladevcp-userfuncs3
[18:55:35] <andypugh> Yes.
[18:55:54] <andypugh> But only if you definitley have a latency problem.
[18:56:01] <mhaberler> we forgot one thing: the usermods need to get at the widget tree so they can set props in widgets
[18:56:06] <psha> mhaberler: forget userfuncs2, userfuncs3 is morecorrect
[18:56:09] <Guest796> so should one just blindly do this mod?
[18:56:21] <Guest796> or should one do a test first to see if you need it?
[18:56:22] <psha> mhaberler: we've not forget it :)
[18:56:31] <psha> mhaberler: it's different task, not handler related
[18:56:45] <Guest796> oops
[18:57:05] <psha> btw how do you want to set props?
[18:57:11] <Guest796> so how does one know if one has a latancy problem?
[18:57:19] <mhaberler> we somehow need to pass/expose builder to py files
[18:58:01] <psha> mhaberler: for what purpose? you may connect 'realize' signal to widgets you want to modify and then fix all you need in the handler
[18:58:04] <mhaberler> psha: AND the fact wether its libglade or gtkbuilder (get_widget vs get_object)
[18:58:31] <andypugh> Guest796: from the EMC2 mene, select "latency test"
[18:58:50] <Guest796> all ready have it running
[18:58:56] <andypugh> ie, rather than start emc2, start the latency test from the same applications dropdown.
[18:58:56] <mhaberler> psha: assume I have button press handler and want to set image in a different widget . all you get is the button object ref
[18:59:30] <andypugh> If you see a regular huge spike every minute or so, then SMI is a likely cause
[18:59:52] <Guest796> whats the best torture to test latency?
[18:59:59] <mhaberler> psha: need to get at the widget with builder.get_(object/widget)("pimagewidget") from buttonpress handler
[19:00:08] <andypugh> glxgears is normally reccomended
[19:00:35] <psha> Guest796: and something that eats CPU
[19:00:41] <andypugh> But SMI issues will show up in an idle system
[19:00:53] <Guest796> what is glxgears?
[19:00:59] <psha> i'm running some glxgears and streaming compression of /dev/urandom to /dev/null
[19:01:11] <psha> Guest796: it's simple graphical program
[19:01:17] <psha> run 'Alt-F2' and type 'glxgears'
[19:01:23] <andypugh> First question, the latency test has been running a few minutes now, what does it say?
[19:01:25] <psha> s/run/press/
[19:01:30] <mhaberler> psha: we could pass it as userdata to the handlers
[19:01:40] <psha> mhaberler: it's not wise i think...
[19:01:48] <Guest796> what one is the one to use?
[19:01:52] <Guest796> it has 2
[19:01:58] <andypugh> Give us both
[19:02:11] <mhaberler> psha: I'm all ears for a wise param passing
[19:02:33] <Guest796> 6359 6340
[19:02:43] <psha> why i think it's not good - in most of handlers you don't need knowledge about other widgets so you don't need to know where they are
[19:02:54] <psha> but somebody is pushing another userdata for you!
[19:02:55] <Guest796> but I think this computer has the mod
[19:03:09] <andypugh> Yes, that's extra-good latency
[19:03:10] <pcw_home> andypugh: big angle steps are not ideal for the twin PI loops and may cause U,V,W current faults if running near 30A
[19:03:32] <Guest796> but my other computer that is the same modle is running around 20,000
[19:03:33] <psha> so we have to allow strange things along with easy ways for easy handlers
[19:03:58] <andypugh> 20,000 is fine too. SMI gives you 300,000 or so, or maybe a million
[19:04:01] <psha> Guest796: wait for a while
[19:04:41] <psha> mhaberler: let me describe you example with 'realize' callbacks
[19:04:51] <andypugh> mshaver: OK, pcw says you need encoders :-)
[19:04:59] <psha> i'll make example for you
[19:05:09] <mhaberler> thanks
[19:07:04] <mshaver> So, having increased max current to 20.0 Amps, I still can't go faster than 25-27 RPS. Why?
[19:07:42] <mshaver> Is commutation speed limited by thread speed?
[19:07:48] <andypugh> Yes.
[19:07:58] <mshaver> Oh.
[19:08:00] <andypugh> I am trying to work out if you are at that limit
[19:08:12] <pcw_home> You cannot increase the current without setting the EEPROM variable
[19:08:22] <psha> mhaberler: http://psha.org.ru/tmp/realize
[19:08:31] <psha> look into bar.py
[19:08:52] <psha> it's a bit ugly with 'global' vars but you may use object and store values inside of it
[19:08:53] <mshaver> So, just increasing that var "on the fly" doesn't actually do it?
[19:08:57] <pcw_home> also your voltage will limit the max speed as well (and the commutation rate)
[19:09:29] <pcw_home> no the VAR does not do it (you need to write the EEPROM value and power cycle the 8I20)
[19:09:52] <andypugh> mshaver: You probably should have got an error message that you can't have more than 7.5A in the current eeprom setup?
[19:09:56] <mhaberler> psha: I see. Ok, tat's a way too. Fine with me.
[19:10:06] <mshaver> I've got about 150V, but it's (I think) a 300V motor (usually runs off 220Vac)
[19:10:06] <Guest796> do we look at max jitter or max intervals?
[19:10:27] <mshaver> in the kernel log? or...
[19:10:41] <Guest796> ran glxgears and the max jitter is now 9900 10477
[19:10:43] <pcw_home> Current is not limiting your speed, voltage is
[19:10:46] <psha> Guest796: max jitter
[19:11:08] <Guest796> cool
[19:11:15] <pcw_home> (you may want to up your servo thread to 2 or 4 KHz) to improve commutation
[19:11:21] <andypugh> I reckon you should be able to get to 7000rpm with your current servo thread perid
[19:11:24] <Guest796> so what good is the glxgears program?
[19:11:37] <andypugh> Guest796: It's just a graphics test
[19:11:57] <pcw_home> Not without big angle steps which will cause problems (until we add interpolation)
[19:12:05] <pcw_home> bbl
[19:12:06] <psha> Guest796: it's loading your system so you may see how it behaves in bad situation
[19:12:08] <andypugh> mshaver: But you might get a bit lumpy at the top end
[19:12:25] <mshaver> I'll speed it up just to see if that affects it...
[19:12:27] <pcw_home> Lumpy may mean faults
[19:12:33] <Guest796> so do I need to punish my computer any to test it further?
[19:12:56] <andypugh> I would say not.
[19:13:09] <psha> andypugh: is not it too short period?
[19:13:14] <Guest796> if so what would be the best thing to do for the testing?
[19:13:35] <Guest796> ok
[19:13:36] <Guest796> thanks
[19:13:55] <andypugh> I guess you could leave it 24 hours. Some computers do wierd things once a day
[19:14:32] <psha> mine increased jitter from ~12-13us to 15-18 after ~10 hours :)
[19:14:43] <andypugh> But the 6k latency computer looks very good, and the 20k one is fine. I wouldn't bother looking for more from either.
[19:15:07] <psha> mhaberler: have you any suggestions for Bar widget?
[19:15:14] <andypugh> SMI gives you spikes much bigger than the normal.
[19:15:19] <mhaberler> no more!
[19:15:30] <psha> mhaberler: if not i'll create one commit for it and merge into main gladevcp branch
[19:15:48] <psha> and i think it's good time today for a merge request :)
[19:15:57] <mhaberler> yess!
[19:16:30] <andypugh> mshaver: I guess I need to get to work on the nvram setting code then?
[19:17:18] <mhaberler> psha: if you get it in, I'll do the wiki footwork :-/
[19:18:28] <mshaver> andypugh: that would be great!
[19:18:52] <mshaver> cutting the thread time in half doesn't increase the speed
[19:18:56] <mhaberler> my current last wish is getting a gladevcp window in the pyvcp place in axis.
[19:19:43] <mshaver> in fact, at servo_period_nsec=300000 or shorter, the motor runs away
[19:20:37] <andypugh> Interesting!
[19:20:52] <mshaver> I did notice that at servo_period_nsec=500000 I could get 30RPS vs. 27RPS at servo_period_nsec=1000000, but that may be just the result of a better ability to display the encoder velocity
[19:21:32] <andypugh> Do you know the rated voltage and speed of the motors?
[19:23:24] <mshaver> I think it's 300V and 3500RPM
[19:23:40] <mshaver> so at 150V, half that is about what I'm getting
[19:24:52] <andypugh> There is a linear relationship between speed and voltage, I am almost sure that you are currently voltage-limited.
[19:25:03] <mshaver> I think so too.
[19:25:13] <andypugh> Can you half your voltage (assuming doubling it is hard)
[19:25:17] <psha> mhaberler: i'm intersted in it also :)
[19:25:39] <mhaberler> psha: please get out line 37 from gladevcp.py !!
[19:25:47] <mshaver> So, for now, we're where we can go. Tomorrow I'll ask the boss what she wants to do next!
[19:25:55] <andypugh> Great.
[19:26:33] <mshaver> I'm basically rectifying 110V now, I could with a transformer
[19:26:38] <Guest796> can one drag and drop copy? all I get is move
[19:26:44] <andypugh> You know, you might not actually want or need more than 7.5A
[19:27:24] <andypugh> Guest796: In what software/os/hardware?
[19:27:26] <psha> Guest796: try right-button dnd (may be not correct)
[19:27:44] <psha> andypugh: nautilus i think )
[19:27:52] <Guest796> what is the fastest way to copy files? all I found was copy paste
[19:27:59] <Guest796> 8.04
[19:28:12] <Guest796> the live cd install for EMC2
[19:28:16] <psha> mhaberler: oops
[19:28:18] <mshaver> very true at 340Vdc (240Vac full wave rect), 2.2Kw is 6.47A
[19:28:32] <mhaberler> its the commit itch ;-)
[19:28:38] <SWPadnos> drag with the middle button and you'll get a menu that asks you to copy/move/link
[19:28:49] <andypugh> ctrl-drag
[19:28:51] <psha> mhaberler: btw all three lines are very bad
[19:28:51] <Guest796> cool
[19:28:53] <Guest796> thanks
[19:28:56] <SWPadnos> you can also hold shift/alt/ctrl to get different options
[19:29:11] <SWPadnos> (the + sign means copy, I don't recall what the other ones are)
[19:29:12] <mhaberler> fix as you see fit
[19:31:03] <mhaberler> psha: wrt wiki: are we staying at the psha gladevcp-userfuncs3 branch for now?
[19:33:51] <psha> mhaberler: branch 'gladevcp'
[19:34:12] <mhaberler> you mean on your repo?
[19:34:15] <psha> yes
[19:34:22] <psha> it's HAL Bar and Userfuncs
[19:34:31] <psha> and it'll go into master today i hope
[19:34:32] <mhaberler> ok
[19:34:46] <psha> and i think i'll keep this branch as merge one
[19:34:50] <psha> instead of gladevcp-modules
[19:35:31] <psha> and halfile
[19:36:04] <andypugh> I wonder if psha/mharbeler should move this stuff to #emc-devel so that they don't frighten the newbies who appear?
[19:36:24] <psha> andypugh: :) sure master ;)
[19:36:27] <mhaberler> this is the coding boot camp !
[19:36:45] <andypugh> (I speak as one of the newbies)
[19:53:56] <Guest796> is there a dwell in mil sec? G4 in second seem a tad long
[19:54:54] <SWPadnos> try using something like 1.234 seconds
[19:55:41] <Guest796> so one could do G4 0.2 and that would work?
[19:55:42] <Guest796> cool
[19:55:47] <SWPadnos> dunno, try it :)
[19:56:41] <Guest796> can one run EMC and the latency test at the same time?
[19:56:49] <jthornton> no
[19:56:54] <Guest796> did not want to shut down the set right now
[19:57:09] <jthornton> you can do G4P0.000001 if you like
[19:57:16] <SWPadnos> you can check latency while emc is running, but it's not easy
[19:57:45] <SWPadnos> the EMC latency test uses a HAL component, which can be added to a running emc
[19:57:50] <SWPadnos> but you have to do it by hand, which is a pain
[19:58:07] <mhaberler> psha: I've updated the http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?GladeVcpSetup wiki page and added user callbacks.
[19:58:09] <mhaberler> I guess I need a good MDI example to inspire the folks about its usefulness..
[19:58:16] <SWPadnos> and then display the results somehow, which isn't trivial (though I think halmeters could do it)
[19:58:32] <Guest796> will let the test run for 24+ hours and see where it goes
[19:59:07] <Guest796> at 11396 and 13839
[19:59:34] <Guest796> so what number does one use in the setup wisard?
[20:00:07] <Guest796> servo thread or base thread
[20:00:21] <Guest796> or the largest of the 2
[20:00:54] <SWPadnos> largest is safer, else use the one from the base thread, since that's the one that matters for software step generation
[20:01:49] <psha> mhaberler: thanks
[20:21:31] <psha> bb all
[20:33:27] <juri__> juri__ is now known as juri_
[23:14:26] <poppabear> Hey Can EMC2 do a front and Rear tool post for lathes?
[23:16:04] <JT-Hardinge> yes
[23:16:21] <poppabear> know of any docs, or how to's?
[23:16:42] <JT-Hardinge> you have both or one or the other
[23:17:36] <Guest666> Any ideas on this one? using emc on a knee mill setting up my first axis all is well except when the axis is at standstill. The motor will turn very slowly and the dro on emc will not register any change. however my DRO attached to my sclaes moves about a .001" evey two or three sec.
[23:17:49] <JT-Hardinge> front or rear just depends on your setup for axis directions
[23:20:10] <poppabear> JT any docs or how tos you are aware of?
[23:20:16] <JT-Hardinge> Guest666: servos?
[23:20:47] <JT-Hardinge> poppabear: there is no difference in programming either one
[23:21:07] <Guest666> JT Yes
[23:21:38] <JT-Hardinge> Guest666: are your encoders connected to EMC?
[23:22:12] <poppabear> JT: What I am hoping for in Turn is I can designate tools in the table as either front or rear and then when that tool is called the X axis will move accordingly
[23:22:40] <poppabear> [6:22:26 PM] hood: means I can have all positive values for X and dont have to piss about with editing the code or doinga G51X-2 and remembering to do a G50 at the right point
[23:22:52] <Guest666> JT yes thats the weird part i can see the counts on the drive changing however if i look at the 7i43 encoder counts in hal meter nothing is changing
[23:25:13] <Guest666> The encodes are connected i can issue MDI commands and jog no problem
[23:27:12] <Guest666> It's just the odd time one the axis is standing still it will start to drift. Lets say .030" per min. when it drifts the DRO on EMC still thinks the axis is still at the position it last stopped at
[23:28:04] <JT-Hardinge> Guest666: sounds like something is not connected correctly
[23:28:34] <JT-Hardinge> poppabear: I'm not sure how that works from that stand point
[23:28:45] <toastydeath> poppabear, i don't recall emc having that feature
[23:29:05] <toastydeath> which is usually called rear/reverse/mirror turning
[23:29:18] <JT-Hardinge> seems to me if you touch off front or back then the X should be even
[23:29:32] <toastydeath> negative
[23:29:51] <toastydeath> x is negative for rear tool posts, i.e. facing to x-.05 or something
[23:30:12] <Guest666> JT When i jog or issue mdi commands the proper HAL pins update. Just not when sitting still.
[23:30:50] <toastydeath> because even though it is a lathe it's still a three dimensional coordianate system
[23:31:18] <toastydeath> once you pass x0 from the positive direction, it goes negative even though it's positive from the absolute magnitude/radius perspective
[23:32:11] <JT-Hardinge> yea and tools on the back turret are aways negative right?
[23:32:18] <toastydeath> right
[23:32:38] <JT-Hardinge> so it is just a matter of proper programming?
[23:32:40] <toastydeath> more importantly than absolute position, they're reversed - making them more negative backs the tool out
[23:32:54] <toastydeath> sort of?
[23:33:13] <toastydeath> it kind of removes the convenience of having a rear tool post if you have to program it explicitly for one tool position
[23:33:24] <Portly> hey, I just wanted to let a few of you know, the ones on last night trying to help be debug my parallel port, that I got it fixed.
[23:33:28] <toastydeath> i.e. the paths are not universal and you've boned yourself
[23:34:08] <JT-Hardinge> Guest666: somehow EMC does not know of the drift or it would correct it as if you have it disconnected when your at idle
[23:34:36] <Portly> If anyone else comes with the same problem, it was the chord I was using from my parallel port. It was one used to transfer data from one computer to another. With those chords the pins at one end do not match the other.
[23:34:41] <poppabear> darn satelite.........
[23:35:49] <JT-Hardinge> Portly: and some don't have all the connections between all the pins
[23:36:07] <JT-Hardinge> * JT-Hardinge wanders inside to be with the wifey
[23:36:20] <JT-Hardinge> say goodnight Gracie
[23:36:35] <Portly> yeah, this took me days of wiring, re wiring, reconfiguting my os, my bios, etc..
[23:44:01] <andypugh> I wish I had seen Portly at the time, those Laplink cables really are a menace. (especially as some of the cheap chinese stepper drives seem to ship with one included, despite the fact it is totally wrong for the job)