#emc | Logs for 2008-08-16

[00:20:21] <cradek> Open source licensing has become a widely used method of creative collaboration that serves to advance the arts and sciences in a manner and at a pace that few could have imagined just a few decades ago. (from the Jacobsen v. Katzer opinion)
[00:21:18] <archivist_ub> sure has :)
[00:22:13] <cradek> ok that encourages me to quit surfing and get back to my emc conversion project.
[00:22:34] <cradek> that decision is a very nice surprise.
[00:22:56] <archivist_ub> time for bed for me, whats next get that mazak from calif
[00:23:11] <cradek> not for me. I have a mill.
[00:23:21] <cradek> (and soon, a lathe)
[00:24:05] <jepler> with the RC filter and the leaded switch, I get great behavior when the switch clicks (quick rise to VCC, no chatter) and terrible when the switch opens (lots of chatter)
[00:24:39] <cradek> that sure is giving you trouble
[00:25:10] <jepler> it chatters for up to about 2ms
[00:25:17] <cradek> hey if you're bored, we could recut your part tonight
[00:25:53] <archivist_ub> either increase the time constant to kill the opening noise or test on close
[00:26:05] <jepler> I think the chatter on release isn't a problem as long as emc's digital debounce is long enough
[00:26:12] <jepler> I home by moving onto the switch
[00:27:47] <jepler> even though the chatter looks like up to 2ms on the scope I don't actually see it in halscope
[00:28:02] <tom1> skunkworks: thanks, that chip is interesting. only 60krpm max tho ;)
[00:28:03] <jepler> and the occasonal random values aren't popping up in halscope anymore either
[00:30:20] <jepler> so I'm very happy to think *that* item is fixed by better wiring
[00:30:47] <mozmck> you get the noise out of the switches?
[00:30:52] <jepler> mozmck: yes
[00:31:00] <mozmck> what did it take?
[00:31:21] <jepler> RC filter + better wires
[00:31:35] <jepler> + still using emc's digital debounce to improve switch release
[00:31:46] <mozmck> better wires? twisted pair?
[00:32:11] <tom1> yeah, on release sounds like twisted pair will help
[00:32:59] <jepler> the first time I had hand-soldered very small wires to the switch, then bent the wires pretty hard. I replaced it with a switch that had wires pre-soldered (and also the switch is sealed against dust), to exclude my bad soldering and wire treatment as a source of problem
[00:33:11] <jepler> these wires are larger gage too
[00:33:22] <mozmck> I see
[00:33:36] <mozmck> what is the pluto board you were mentioning?
[00:34:08] <mozmck> I haven't had a chance to test the idoubler code you gave me yet btw
[00:34:35] <jepler> mozmck: it is a board with an FPGA chip and parallel-port interface that I became aware of about 2 years ago and wrote some emc2 drivers for. but because of the low quality of the design and build I don't recommend it for anyone.
[00:35:24] <mozmck> I see.
[00:35:37] <jepler> on some PCs it simply doesn't work at all, and on others it's unreliable
[00:36:02] <mozmck> 3.3v inputs are not going to be very good overall for industrial use either I wouldn't think
[00:37:05] <jepler> yeah -- it's a trade-off that is tempting to justify for hobbyist-level machines but that's about it
[00:38:23] <mozmck> i'm curious about something.
[00:38:26] <jepler> (and my own machinery is strictly hobbyist-level stuff)
[00:38:50] <mozmck> how does mach3 get 100khz output
[00:38:53] <mozmck> ?
[00:39:22] <mozmck> looks like the best my 1.8ghz machine is going to do is about 35khz with emc2
[00:39:51] <jepler> I don't know the internals of mach3 so what I'm about to say is mostly trash talk
[00:40:40] <jepler> the thing that knocks down your step rate when you use stepconf is the value you enter for latency
[00:41:06] <jepler> stepconf is very careful to tune so that you are guaranteed not to violate the step timings you input, even when it means sacrificing speed
[00:41:14] <mozmck> yep. mine is around 11000
[00:41:27] <jepler> this is good, because stepper drivers will simply act flaky when you are not meeting the timings
[00:41:41] <jepler> if you want to see stepconf say you can get a higher step rate, then start by entering 0 for latency.
[00:42:21] <jepler> as far as I know, mach just lets you choose "XX Hz" and does not attempt to determine if that rate is doable in light of your system's latency -- so sometimes a step timing parameter (usually the time required between step pulses) is violated
[00:43:11] <jepler> if I choose PMDX-150 timings and enter 0 for latency test result, stepconf claims I'll get a step rate of 125000Hz
[00:43:32] <mozmck> I just wondered. I know the mach3 engine runs windows on top of it.
[00:44:23] <jepler> there are probably also technical differences in the way that mach3 and emc arrange for the CPU to be diverted from whatever it's doing to outputting the step pulses, but those are over my head in emc and I don't know the details on mach
[00:45:18] <mozmck> i see.
[00:48:02] <jepler> fwiw my machine is very sluggish running this configuration I just generated but it is going at 125000Hz step rate -- I don't have any motors that will step at that rate, of course
[00:48:31] <jepler> emc prints "unexpected realtime delay" messages, but if your interest is to step as fast as mach you could simply find those in the source code and comment them out
[00:48:34] <jepler> nobody likes bad news, after all
[00:57:34] <jmkasunich> we're going to find out how long a HF bandsaw takes to cut a 4x2 bar of cast iron.....
[00:57:42] <jmkasunich> (this is a timestamp, ignore)
[00:57:48] <cradek> haha
[00:58:07] <cradek> I cut 2x4 steel with my Jet, no problem
[00:58:21] <jmkasunich> I anticipate no problems, just time
[00:58:58] <jmkasunich> my timestamp is actually about a minute late, because I walked over here after starting the saw and read the conversation between moz and jeff before typing
[00:59:06] <cradek> I had a HF blade on it. it stopped cutting and it suspiciously bent when I took it out of the saw. I used pliers and could fould it flat against itself and it wouldn't break
[00:59:14] <jmkasunich> heh
[00:59:16] <cradek> could fold
[00:59:27] <jmkasunich> I'm pretty sure I replaced mine with a bimetal long ago
[01:00:04] <cradek> I folded it up into 6" segments to throw it away
[01:00:39] <jmkasunich> anybody want to do magnet experiments? I got the iron filings
[01:01:47] <jmkasunich> dang, I just realized now that the material (which I bought several months ago) is only 1-1/2, not 2"
[01:13:41] <jmkasunich> done!
[01:13:52] <jmkasunich> 17 minutes, more-or-less
[01:15:12] <jmkasunich> time to introduce freshly severed iron to Mr. Norman - bbl
[01:25:41] <jepler> http://www.alfonsomartone.itb.it/naxntb.html
[01:27:04] <tom1> the last line is like a threat "Ed il prossimo passo potrebbe essere la costruzione di una CNC-Mill..." ;)
[01:27:42] <jepler> yeah I saw that
[01:27:54] <jepler> (scroll to about the middle and you get the same in english)
[01:28:52] <tom1> oh!
[01:31:08] <jepler> http://tools.ietf.org/html/rfc5241
[01:35:47] <tom1> I waffled over it's reality, but finally scanned '1 April 2008'
[01:37:18] <tom1> goodnight all
[01:43:11] <Administrador> Hi everyone
[01:43:23] <Administrador> i'm a newbie of emc
[01:43:33] <Administrador> i have a few questions
[01:44:50] <Administrador> Administrador is now known as Angelo16m
[01:45:27] <Angelo16m> which is the hardware support of emc?
[01:48:22] <Angelo16m> somebody?
[01:48:44] <mozmck> I'm a newbie as well.
[01:48:59] <mozmck> have you looked over the emc website thoroughly?
[01:49:11] <Angelo16m> yes
[01:49:13] <fenn> Angelo16m: why are you PM'ing me?
[01:49:29] <Angelo16m> sorry
[01:50:05] <Angelo16m> appear information of caracterisitc of software
[01:50:09] <fenn> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_Supported_Hardware
[01:50:11] <jepler> Angelo16m: besides the parport, emc has drivers for a few specific I/O cards. most of them are listed in this document: http://linuxcnc.org/docs/html/hal_drivers.html
[01:51:04] <Angelo16m> but not the hardware supported
[01:51:15] <fenn> do you mean what computers it can run on?
[01:51:22] <cradek> can you be more specific?
[01:51:33] <cradek> hardware can mean many things. supported can mean many things.
[01:51:41] <mozmck> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Hardware_Requirements
[01:52:12] <Angelo16m> thanks all for help
[01:52:30] <Angelo16m> if found in this web page http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC_Supported_Hardware
[01:53:48] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Supported_Hardware
[01:54:16] <jepler> (slightly different URL, more recently updated)
[01:57:38] <Angelo16m> i have other question
[01:58:12] <Angelo16m> Vital Systems Motenc-Lite PCI Servo 4-Axis Card and I/O Boards (only support this type of machine or all virtual system?)
[01:59:44] <jepler> I believe that kind of card is only for servo motors with +-10V analog inputs. It is not for step&direction motors.
[02:00:40] <Angelo16m> jeples
[02:00:49] <Angelo16m> de page that yuo send
[02:01:22] <Angelo16m> is all card or machines (i'm relly a newbie) supported for emc?
[02:01:48] <jmkasunich> "machine" can mean many things - some people mean "computer", others mean "machine tool"
[02:02:09] <jmkasunich> that page is mostly a list of interface cards
[02:03:07] <Angelo16m> but i no refer the computer like machine , i refer mahine the bject controlated by computer
[02:03:33] <cradek> Angelo16m: do you want to use emc2 to control a lathe or mill?
[02:03:45] <jmkasunich> EMC is designed to control milling machines and lathes. routers and plasma cutters can also be controlled - they are similar to milling machines
[02:05:33] <Angelo16m> ok
[02:05:52] <Angelo16m> i don't now yet
[02:06:52] <JymmmEMC> Angelo16m: what are you wanting emc to do?
[02:07:27] <Angelo16m> i watching the theme of emc
[02:07:38] <Angelo16m> for choose the correct machine
[02:07:52] <jmkasunich> usually you choose the machine based on what you want to do
[02:07:59] <jmkasunich> if you want to make round things, you use a lathe
[02:08:06] <JymmmEMC> Angelo16m: What do you want to do with emc?
[02:08:06] <jmkasunich> if you want to make square things, you use a mill
[02:08:21] <jmkasunich> if you want to cut sheets of metal, you use a plasma cutter
[02:08:39] <JymmmEMC> jmkasunich: Hey now... the first thing I ever made on a mill was round =)
[02:08:59] <jmkasunich> you always have been a little "different", now haven't you ;-)
[02:09:10] <JymmmEMC> jmkasunich: And always will be =)
[02:09:24] <Angelo16m> i working with a friend
[02:09:45] <Angelo16m> he said lathe
[02:09:51] <JymmmEMC> jmkasunich: Now I wonder if the first time I touch a lathe if I'll make something square =)
[02:10:09] <jmkasunich> that will be a bit harder
[02:10:29] <JymmmEMC> jmkasunich: an 8 sided square =)
[02:10:56] <jmkasunich> Angelo16m: lathes need less I/O (usually)
[02:11:04] <jmkasunich> two axes of motion, plus a spindle encoder
[02:11:16] <jmkasunich> mills usually have three axes of motion
[02:11:33] <Angelo16m> ok
[02:12:36] <Angelo16m> how can i know how much resist weight a machine
[02:13:34] <jmkasunich> do you mean "how strong do the motors need to be?"
[02:13:55] <Angelo16m> yes
[02:14:16] <jmkasunich> best way is to measure on the actual machine
[02:14:45] <Angelo16m> maximun weight of plates
[02:15:08] <jmkasunich> don't understand question
[02:15:10] <Angelo16m> plates wich cut with de lathe
[02:15:56] <jmkasunich> lathes do not cut plates
[02:16:20] <jmkasunich> http://en.wikipedia.org/wiki/Lathe_(tool)
[02:17:08] <Angelo16m> with machines cutes plates of metal?
[02:17:30] <jmkasunich> describe the plates you want to cut?
[02:17:31] <Angelo16m> which machines cutes plates of metal?
[02:17:34] <jmkasunich> how thick? how wide?
[02:18:18] <jmkasunich> for large thin plates, plasma cutter : http://www.youtube.com/watch?v=KvhvQu5Xxtk
[02:21:24] <Angelo16m> the wide is 20 x 10 feet
[02:21:25] <Angelo16m> the thick are
[02:21:27] <Angelo16m> 3 milimeters
[02:21:58] <jmkasunich> 20 x 10 feet = 6 x 3 meters, is that correct?
[02:22:04] <Angelo16m> yes
[02:22:11] <jmkasunich> you probably want a plasma cutter for that
[02:22:18] <jmkasunich> did you see the youtube video I posted?
[02:22:23] <Angelo16m> yes
[02:22:33] <jmkasunich> is that the kind of cutting you want to do?
[02:23:12] <Angelo16m> Is it too expensive?
[02:23:19] <Angelo16m> plasma cooter
[02:23:29] <jmkasunich> anything that will cut 6m x 3m is going to be "expensive"
[02:24:09] <mozmck> If you build the table yourself it can be cheaper...
[02:24:28] <jmkasunich> Angelo16m: where are you?
[02:24:43] <Angelo16m> why?
[02:24:50] <Angelo16m> for my english?
[02:25:09] <jepler> I'm guessing Chile
[02:25:38] <Angelo16m> yes
[02:25:42] <jmkasunich> http://www.youtube.com/watch?v=Bxr8UWczofU&feature=related
[02:25:49] <Angelo16m> how you kow?
[02:26:00] <jmkasunich> spanish I think: http://www.youtube.com/watch?v=Bxr8UWczofU&feature=related
[02:26:32] <Angelo16m> jepler
[02:26:41] <Angelo16m> for my ip?
[02:26:44] <jepler> Angelo16m: yes, from your IP
[02:26:44] <jmkasunich> he probably looked at your IP address
[02:27:42] <jmkasunich> http://es.wikipedia.org/wiki/Corte_por_plasma
[02:27:48] <jmkasunich> the english article has more detail
[02:27:57] <jmkasunich> http://en.wikipedia.org/wiki/Plasma_cutter
[02:28:02] <Angelo16m> my firend say need the machine to make Figures 2d and 3d
[02:28:30] <Angelo16m> is posible with thw plasma cutter?
[02:28:51] <jmkasunich> plasma cutters cut thin flat sheets of metal
[02:28:59] <jmkasunich> that is 2D only
[02:29:16] <jmkasunich> you can bend them after cutting, or weld pieces together, to make 3D
[02:30:25] <jepler> plasma always cuts all the way through the metal sheet (when operating correctly)
[02:33:03] <Angelo16m> i need
[02:33:04] <Angelo16m> a machine can make
[02:33:10] <Angelo16m> a steel profiles
[02:40:48] <jepler> 'night all
[02:41:13] <jmkasunich> I think I need a de-magentizer: http://jmkasunich.com/pics/fuzzy-drill-2562.jpg
[02:50:18] <Angelo16m> i need do this
[02:56:08] <Angelo16m> http://i33.tinypic.com/210gj9g.jpg
[02:56:29] <jmkasunich> you don't make those, you buy them
[02:56:40] <Angelo16m> why?
[02:56:45] <jmkasunich> they are made using very large rolling machines - rolled red hot
[02:57:12] <Angelo16m> like this
[02:57:13] <Angelo16m> http://www.voortman.net/steel-machinery/products/beams/vb600.html
[02:57:44] <jmkasunich> that saw cuts beams to the right length
[02:57:54] <jmkasunich> it does not make the shape of the beam
[02:59:07] <jmkasunich> this is how beams are made: http://www.youtube.com/watch?v=TulekDb1zBc&feature=related
[03:04:20] <jmkasunich> slick: http://www.youtube.com/watch?v=TulekDb1zBc&feature=related
[03:11:33] <Angelo16m> i have a confusion
[03:12:02] <Angelo16m> i need a machine can write letters and numbers
[03:21:36] <JymmmEMC> an engraver?
[03:26:07] <Angelo16m> si
[03:26:11] <Angelo16m> yes
[03:34:11] <Angelo16m> plasma cutter?
[03:34:29] <jmkasunich> can you see youtube videos?
[03:36:39] <Angelo16m> yes
[03:36:40] <Angelo16m> no i wrong
[03:36:42] <Angelo16m> i need a steel
[03:36:43] <Angelo16m> cutting steel
[03:38:23] <jmkasunich> this is a plasma cutter - it can cut letters and numbers in steel: http://www.youtube.com/watch?v=KvhvQu5Xxtk
[03:39:30] <jmkasunich> this is a milling machine - it can engrave letters and numbers:
[03:39:36] <jmkasunich> http://www.youtube.com/watch?v=Nvq4f9AHOTA
[03:44:47] <Angelo16m> buy milling machine is very slow
[03:45:05] <Angelo16m> and i don't thik so milling cut steel
[03:45:30] <jmkasunich> that video is milling aluminum
[03:45:38] <jmkasunich> you can mill steel, but it is even slower
[03:46:18] <jmkasunich> bigger and stronger milling machines can go faster, but are very expensive
[03:46:32] <jmkasunich> even very big milling machines are slower than plasmacutter
[03:50:28] <pminmo> jmlasunich, should I post this error as a bug? http://pastebin.com/d7f361cbf
[03:51:59] <jmkasunich> no idea
[03:52:17] <pminmo> I get it when I exit emc, everything runs fine
[03:52:35] <jmkasunich> can it be replicated, does it happen with our standard packages on standard ubuntu?
[03:53:10] <pminmo> yes, and it only started today when I did a package update
[03:53:32] <jmkasunich> what config?
[03:54:01] <pminmo> not sure I understand the question?
[03:54:27] <pminmo> ubuntu 8.04
[03:54:34] <jmkasunich> what machine configuration? one of our examples? or yours? if yours, is it stepper, servo, etc
[03:54:59] <jmkasunich> if somebody like jepler can't reproduce the problem, then there isn't much that can be done
[03:55:12] <jmkasunich> to reproduce it, we need to know exactly what you are doing
[03:55:35] <pminmo> i.e stepper, my ini and hal happens when I exit emc after using it with no issues
[03:56:11] <pminmo> everything is fine loading running, but when I exit I get an error dump
[03:58:03] <pminmo> I mean it doesn't hurt my usage, I can start emc back up and it starts fine and runs fine, just the error dump on ecit
[03:58:12] <pminmo> exit
[04:02:24] <jmkasunich> it seems to be a segfault in axis, so that would be something for jepler to look at
[04:02:34] <jmkasunich> he'll probably see this discussion tomorrow, and your pastebin
[04:03:35] <pminmo> ok thanks
[04:05:40] <cradek> looks like we now have a pci step generator
[04:05:42] <cradek> that's pretty neat
[04:07:09] <pminmo> ?
[04:07:17] <jmkasunich> the 5i20?
[04:07:27] <cradek> yes
[04:07:43] <cradek> seb says it works now
[04:15:22] <jmkasunich> goodnight
[04:24:51] <JymmmEMC> cradek: I thought that's been around for a while, just "beta"
[04:25:13] <JymmmEMC> so to speak
[07:31:20] <mumu> jepler: I'am still trying to control emc with python script, but still no luck, i tested this ( http://pastebin.ca/1174058 ) code, and it returns: " c = emc.command() TypeError: expected string or Unicode object, NoneType found". I don't know what am i doing wrong.
[08:26:49] <Vq^> thats weird
[08:28:42] <Vq^> do you run the script from a different directory than the ini-file?
[08:29:07] <Vq^> i see it's done like this in axis.py:
[08:29:10] <Vq^> emc.nmlfile = os.path.join(os.path.dirname(sys.argv[2]), inifile.find("EMC", "NML_FILE"))
[08:43:45] <mumu> yes, i run the script from different directory, can this be a problem?
[08:44:20] <Vq^> it can be if the nmlfile path is relative in the ini file
[08:44:41] <Vq^> but use something like that row i showed you and you'll be alright
[08:57:02] <mumu> i tied that, now i receive this error: "emc.nmlfile = os.path.join(os.path.dirname(sys.argv[2]), inifile.find("EMC", "NML_FILE")) NameError: name 'inifile' is not defined"
[10:33:42] <robin_sz> evening ...
[10:35:41] <Vq^> g'day
[10:35:57] <Vq^> mumu: i hope you solved that yourself :)
[10:39:13] <mumu> not yet, now i switched to mdi example...this one works for me
[10:40:05] <Vq^> mumu: you called it ini, not inifile
[10:43:27] <mumu> now i received: emc.nmlfile = os.path.join(os.path.dirname(sys.argv[2]), ini.find("EMC", "NML_FILE"))
[10:43:27] <mumu> File "/usr/lib/python2.5/posixpath.py", line 60, in join
[10:43:27] <mumu> if b.startswith('/'):
[10:43:27] <mumu> AttributeError: 'NoneType' object has no attribute 'startswith'
[10:46:46] <Vq^> sounds like it doesn't find the NML-path in the ini-file
[10:47:52] <mumu> i solved it now, i simply add this line: emc.nmlfile = "/home/mumu/emc2/configs/stepper-xyza/emc.nml"
[10:47:56] <mumu> :)
[10:48:31] <Vq^> yeah, hardcoding solves the problem temporarily :)
[11:12:40] <OoBIGeye> hardcoding solves everything... forever :O)
[11:14:57] <Vq^> :/
[11:26:27] <alex_joni> usually you start from the same folder where the ini is
[11:26:33] <alex_joni> and the nml is in the same folder
[11:37:05] <BigJohnT> afternoon alex_joni
[11:48:29] <alex_joni> hi BigJohnT
[11:48:51] <alex_joni> how goes it?
[11:49:38] <BigJohnT> not having fun... had to take dads car away
[11:49:59] <BigJohnT> he is 82
[11:51:00] <alex_joni> oh
[11:51:21] <alex_joni> close to my age.. only inverted
[11:51:34] <BigJohnT> LOL then you have a ways to go
[11:51:48] <alex_joni> I would hope so
[11:52:31] <BigJohnT> it is like constant acceleration... the older you get the faster it gets here
[12:05:55] <OoBIGeye> BigJohnT: nice way of puttiong it
[12:06:01] <OoBIGeye> *putting
[12:06:43] <BigJohnT> hi OoBIGeye
[12:07:19] <BigJohnT> when I was 17 it took forever to get to 18... now the years just fly by
[12:08:36] <OoBIGeye> when your plans only went as far as the next weekend
[12:08:47] <BigJohnT> :)
[12:08:51] <OoBIGeye> now i'm a year ahead...
[12:09:04] <BigJohnT> I just quit planning
[12:15:58] <BigJohnT> well I better go fetch Dad his breakfast see you guys later
[12:32:25] <ra3vat> ra3vat is now known as dimas_
[12:48:49] <alex_joni> anonimasu: around?
[13:09:41] <robin_sz> * robin_sz yawns
[13:10:01] <robin_sz> wonderful weather this week ...
[13:10:12] <robin_sz> except monday when it rained ...
[13:10:16] <robin_sz> and tuesday ..
[13:11:14] <robin_sz> foreach $day (@daysOfWeek){ print "it rained on $day\n"; }
[13:11:44] <robin_sz> which is good for staying inside and messing with emc i guess :)
[13:22:36] <alex_joni> heh
[13:22:50] <alex_joni> it was darn sunny here all week
[13:22:57] <alex_joni> 39C yesterday
[13:23:42] <alex_joni> 102.2F
[13:26:08] <DanielFalck> It was just a little higher here yesterday 103F
[13:27:13] <alex_joni> only 90 today :D
[13:27:30] <DanielFalck> supposed to be only 98 here today
[13:27:57] <DanielFalck> "Plentiful sunshine" : )
[13:28:11] <alex_joni> Chance of Rain
[13:28:11] <alex_joni> 20% chance of precipitation
[13:31:31] <alex_joni> http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=ITIMISUN2&day=15&year=2008&month=8&graphspan=year
[13:52:40] <mozmck> down in texas here, and it 68 degrees - i'm about to freeze!
[14:53:56] <Angelo16m_> somebody?
[14:55:55] <DanielFalck> Angelo16m_: what's your question?
[14:56:26] <Angelo16m_> i can buy a plasma cutter cnc for less than 7000 dolarrs
[14:56:29] <Angelo16m_> ?
[15:01:46] <alex_joni> new?
[15:21:12] <JymmmEMC> Well, I guess the answer is yes..... http://cgi.ebay.com/3D-CNC-ROUTER-MILLING-MACHINE-Plasma-Cutter-TOP_W0QQitemZ370076578471QQihZ024QQcategoryZ57122QQssPageNameZWDVWQQrdZ1QQcmdZViewItem and http://cgi.ebay.com/THERMAL-DYNAMICS-CUTMASTER-TRUE-39-PLASMA-CUTTER_W0QQitemZ360078714433QQihZ023QQcategoryZ113743QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[15:21:39] <alex_joni> bbl
[15:21:47] <JymmmEMC> alex_joni:
[15:24:10] <anonimasu> JymmmEMC: that's a shitty machine for plasma
[15:24:34] <anonimasu> JymmmEMC: exposed screws and plasma dust.
[15:24:35] <anonimasu> yum
[15:24:38] <JymmmEMC> anonimasu: Never said it wasn't. Just under $7k
[15:24:54] <robin_z> expensive too
[15:25:46] <BigJohnT> plasma?
[15:26:16] <robin_z> theres a few US made plasmas that are just about OK for around 5k
[15:26:34] <robin_z> i think plasmacam is 4K when I last looked
[15:26:39] <anonimasu> I actually think he meant the cutter.. not the table..
[15:26:47] <robin_z> oh, then 7K is a lot!
[15:27:11] <robin_z> it should say "hypertherm" on the side ;)
[15:27:17] <BigJohnT> :)
[15:29:08] <robin_z> ive messed with a few plasmas in my time, at the "big boys" end of the market, there are several players worth lookign at, at the low/mid end, hypertherm have it sewn up ...
[15:29:30] <robin_z> you get a lot of neat technology for your dollar
[15:30:51] <robin_z> i got the PID sorted out on that axis now .. quite sweet it is too
[15:32:02] <cradek> robin_z: you got it tuned finally? what made the difference?
[15:32:19] <robin_z> using much lower P values :)
[15:32:36] <robin_z> because its metric, the "standard" starting points are way off
[15:32:45] <cradek> ahh
[15:33:01] <cradek> scale and units and stuff, I see
[15:33:16] <robin_z> and an error in the m5i20 sample setup didnt help
[15:33:38] <cradek> what's that? is it fixed now?
[15:33:45] <robin_z> it is in mine ;)
[15:34:05] <robin_z> therea MAX_OUTPUT = 1 line in there
[15:34:18] <robin_z> so the range is -10 to +1V
[15:34:35] <cradek> hm, that doesn't seem right
[15:34:39] <robin_z> indeed not
[15:34:41] <cradek> seems like that should limit it to -1..+1
[15:34:48] <robin_z> nope
[15:34:55] <robin_z> -10 to +10 is normal
[15:35:16] <cradek> I know, but MAX_OUTPUT should let you limit that - for instance my amps say to use 9 max
[15:35:28] <cradek> but if it only limits the top, it's useless
[15:35:32] <robin_z> and -9 minimum?
[15:35:38] <cradek> yes I assume so
[15:35:54] <robin_z> you neesd MIN_OUTPUT=-9 as well then
[15:35:59] <cradek> are you using hostmot2 or the old driver?
[15:36:07] <robin_z> * robin_z shrugs
[15:36:08] <cradek> oh is there a MIN?
[15:36:12] <robin_z> yep
[15:36:18] <cradek> oh then forget what I said
[15:36:41] <robin_z> but the sample setup only comes with a MAX line .. and thats at 1V ...
[15:37:25] <robin_z> good options would be neither line, or both lines ...
[15:38:49] <robin_z> just been having fun winding up the speeds and accels now to see where it tops out ;)
[15:39:04] <robin_z> the screw is 20mm pitch, so its fairly fast
[15:39:38] <cradek> either you are mistaken or there is a bug. there is no min output for pid, and the manpage says maxoutput is absolute value like I expected at first
[15:39:57] <cradek> you are talking about the m5i20 sample config, right?
[15:40:02] <robin_z> * robin_z nods
[15:40:18] <robin_z> well, then I read wrong documentations someplace
[15:40:39] <cradek> the source agrees that maxoutput limits the absolute value
[15:40:53] <robin_z> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Inch_Freq
[15:41:16] <robin_z> INPUT_SCALE = 16000 0
[15:41:16] <robin_z> OUTPUT_SCALE = 16000.000 0.000
[15:41:16] <robin_z> MIN_LIMIT = -10.0
[15:41:16] <robin_z> MAX_LIMIT = 10.0
[15:41:16] <robin_z> MIN_OUTPUT = -10
[15:41:18] <robin_z> MAX_OUTPUT = 10
[15:41:20] <robin_z> for example ...
[15:41:36] <robin_z> but its in the integrators manual as well I think
[15:42:04] <cradek> that wiki page is an emc1 config. I will just remove it.
[15:42:20] <robin_z> k
[15:42:33] <robin_z> im pretty sure I got it in the current integrators manual ...
[15:42:42] <robin_z> could be wrong though
[15:42:43] <cradek> I bet BigJohnT would be happy to hear about doc bugs. please let him know what was wrong or misleading.
[15:42:56] <robin_z> kk
[15:45:18] <cradek> I can't figure out how to delete a wiki page, so I blanked it out
[15:47:04] <robin_z> so, the current codebase has no use for MIN_OUTPUT?
[15:48:33] <cradek> well there are some remnants - it is read in, but I don't think it's used
[15:48:42] <robin_z> ok
[15:48:49] <cradek> that limiting happens in the pid hal module now
[15:49:09] <robin_z> using the handy google search on the linuxcnc.org site ...
[15:49:10] <robin_z> http://www.google.com/custom?domains=linuxcnc.org&q=MIN_OUTPUT&sitesearch=linuxcnc.org&sa=Search&client=pub-noaccount&forid=1&channel=pub-nochannel&ie=ISO-8859-1&oe=ISO-8859-1&cof=GALT%3A%23008000%3BGL%3A1%3BDIV%3A%23336699%3BVLC%3A663399%3BAH%3Acenter%3BBGC%3AFFFFFF%3BLBGC%3A336699%3BALC%3A0000FF%3BLC%3A0000FF%3BT%3A000000%3BGFNT%3A0000FF%3BGIMP%3A0000FF%3BFORID%3A1%3B&hl=en
[15:49:19] <robin_z> about 40 pages on there refer to it
[15:50:21] <cradek> well it is/was used in emc1, which still exists and there's a lot of stuff about it on that site
[15:50:23] <robin_z> if its read but not used I gues its not doing any harm though
[15:50:49] <robin_z> i guess thats part of the confusion
[15:50:55] <cradek> yeah
[15:51:08] <robin_z> emc 1 and 2 are so different in places now
[15:51:15] <cradek> it's really hard to tell what information is current.
[15:51:18] <robin_z> yep
[15:51:40] <robin_z> at least the manuals that ship with emc2 should be right for emc2 ...
[15:51:53] <cradek> yes if they aren't, that's something to fix
[15:53:56] <robin_z> yeahm its wrong in the current integrators manual too, and even says MIN_OUTPUT (hal)
[15:54:23] <robin_z> chapter4, page 29
[15:54:35] <cradek> good find, let's let BJT know when he gets back
[15:54:42] <robin_z> will do
[15:54:50] <robin_z> hey, you have a cat?
[15:54:51] <JymmmEMC> cradek: potentionally could move stuff to an emc1 archive instead of deleting it.
[15:55:30] <JymmmEMC> just for historical purposes
[15:55:47] <cradek> yeah, I had second thoughts about doing that. it was not a useful page though. on the wiki you can always visit old revisions of pages which is nice for history
[15:55:51] <robin_z> mines bugging the hell out of me ... want another? I'llpay shipping
[15:56:13] <cradek> I have around a half dozen cats depending on how you count
[15:56:25] <robin_z> one more would be no problem then
[15:56:29] <cradek> so I probably do not need any more
[15:56:38] <robin_z> hummph
[15:56:52] <cradek> it can be trouble if they outnumber the people in the house.
[15:59:35] <robin_z> i might modify it
[16:00:13] <robin_z> might make a nice pair of slippers
[16:00:32] <cradek> and six tacos
[16:01:16] <robin_z> hmmm ...
[16:01:50] <robin_z> so,
[16:02:04] <robin_z> just so I get my head around this ...
[16:02:54] <robin_z> im planning on a nice big tube plasma ...
[16:03:53] <robin_z> basic motion is X and A
[16:04:04] <robin_z> and a bit of Z for torch height
[16:04:37] <mozroute> the troubleshooting page on the wiki says you don't want parport_pc installed
[16:05:03] <mozroute> but what about lp parport and ppdev?
[16:05:29] <cradek> I think parport_pc is the only one you need to blacklist
[16:05:45] <mozroute> also, I see the ipv6 is installed, will it help if I get rid of it?
[16:05:49] <cradek> without it (with those others) you can still print to usb printers etc.
[16:05:56] <cradek> help what?
[16:06:10] <mozroute> latency
[16:06:31] <cradek> latency generally cannot be affected by removing or installing software
[16:06:43] <cradek> it is simply a matter of the hardware
[16:06:43] <mozroute> mine seems to be running around 17000 to 18000
[16:06:56] <cradek> that's pretty acceptable
[16:07:33] <mozroute> I would have thought so, but why can it make a difference which video driver you use?
[16:07:57] <cradek> because various drivers trigger the video card to do things
[16:08:15] <mozroute> on a Mach64 card I got quite a bit lower latency with the vesa driver than the standard one
[16:08:30] <JymmmEMC> mozroute: Also, in your BIOS, turn off things you don't need. serial ports, IDE controllers, etc.
[16:08:56] <mozroute> I need my harddrive!
[16:09:13] <mozroute> I turned off all the power stuff
[16:09:14] <cradek> lots of 3d cards have latency killer features. using the vesa driver avoids tickling them
[16:09:38] <JymmmEMC> mozroute: If you have hdd and cdrom on primary controller, turn off the secondary controller.
[16:10:05] <cradek> I swear by matrox cards for realtime (I use millenium II and G450 cards in all my realtime machines)
[16:10:16] <cradek> I just seem to have a lot of them around
[16:10:29] <mozroute> JymmmEMC: I think they're on two different channels. I may put them together and try what you say.
[16:10:39] <mozroute> I'm using a g400 right now.
[16:11:05] <mozroute> the Mach64 seemed to get just about as good of results though and I have two of them
[16:11:25] <cradek> whatever works!
[16:11:46] <cradek> are you having a problem where you think lower latency numbers would help? I don't think 18000 is a problem.
[16:11:59] <mozroute> no, i'm just trying to squeeze as much out of this setup as I can
[16:12:26] <mozroute> it will be a stepper system driven from the parport
[16:12:55] <mozroute> i'll be using gecko drives on about a 3' x 3' router table
[16:13:39] <cradek> geckos are nice. I like the ones with the step multiplier so the pc's step rate can be lower
[16:14:33] <mozroute> These don't have that :(
[16:14:43] <mozroute> I think they are 201s
[16:15:47] <mozroute> I asked last night how mach3 can get up to 100khz step rate and emc2 can't? jepler didn't know, maybe you have an idea?
[16:15:59] <mozroute> is it something in rtai?
[16:16:15] <cradek> jepler gave you an answer that is better than anything I could give
[16:16:20] <mozroute> I'm sure it depends on the computer for mach too...
[16:16:32] <mozroute> ok.
[16:16:41] <cradek> he said he got 135kHz if he told stepgen to not worry about how badly the pulses were spaced
[16:16:44] <mozroute> wish I knew more and had more time
[16:16:48] <mozroute> :-)
[16:16:49] <cradek> or something like that
[16:17:04] <mozroute> I don't know what mach does on the spacing
[16:17:39] <mozroute> at those speeds that is
[16:17:42] <cradek> microstepping drives can probably tolerate very crappy step timings
[16:17:59] <cradek> mozroute: if you have a scope, you could run some tests and see for yourself
[16:18:20] <mozroute> like the geckos - hmmm, hadn't thought of that
[16:18:28] <cradek> but it sounds like you are prematurely optimizing - programmers say this is the root of all evil.
[16:18:42] <mozroute> i have a scope, might do some checking
[16:19:02] <cradek> do you know your design will need a high step rate? if so maybe you could consider step generating hardware. or, use different leadscrews, or who knows what
[16:19:18] <mozroute> :-) I've done that too much sometimes, but I do believing in doing it early
[16:19:53] <cradek> wood routers do need to move fast. you are right to think about it in the design stage.
[16:20:04] <mozroute> I'm not sure exactly what the specs are on the router. it's already built
[16:20:15] <cradek> but since you bought geckos already, some good options are out of the picture already
[16:20:41] <mozroute> I've written quite a bit of software, and I spend more time in the code design sometimes than I do write the code!
[16:20:49] <cradek> ah, seems like you should find that out and go from there.
[16:21:26] <cradek> if it has 5tpi ballscrews, or some pulley gearing, it's a very different story from 20tpi all-thread
[16:21:42] <mozroute> the whole router is done. I think the motors are 200 steps/rev, and I think the leadscrews are 10 pitch ACME
[16:22:02] <mozroute> the pulley ratio is 1:1
[16:22:18] <robin_z> and 10 microstep drives?
[16:22:40] <mozroute> uh, I guess :-) I forget...
[16:23:20] <robin_z> i suspect you re going to need some sort of stepgen hardware
[16:24:23] <cradek> if my fiddling with stepconf is right, it looks like you can get about 100ipm with your numbers
[16:24:45] <mozroute> I hope not... I'm getting on some trading from Tom at CandCNC who I work for.
[16:25:06] <mozroute> we are running it there with Mach3 with no problems out of the parport
[16:25:20] <cradek> that is probably fast enough for cutting wood but moving across the table will take a bit.
[16:25:41] <robin_z> mach3 does seem to get decent step rates out of it ...
[16:25:57] <cradek> at what speed do you cut?
[16:26:06] <mozroute> I'll try and get the specs on it today. It is not just real fast. I think it runs about 150ipm right now
[16:28:06] <cradek> bbl
[16:40:13] <Angelo16m_> hi again
[16:40:43] <Angelo16m_> the plasma cutter which show me
[16:40:52] <Angelo16m_> resiste a steel profile?
[16:47:06] <Angelo16m_> the link that you send are a millie
[16:59:06] <mozroute> I presume "invert" in Stepconf means active low?
[17:01:28] <SWPadnos> yes
[17:03:57] <mozroute> I'm getting a RTAPI realtime error when I run axis
[17:04:15] <mozroute> [ 5197.790386] In recent history there were
[17:04:15] <mozroute> [ 5197.790388] 1787408, 1790978, 1785412, 1788106, and 1787445
[17:04:15] <mozroute> [ 5197.790389] elapsed clocks between calls to the motion controller.
[17:04:15] <mozroute> [ 5197.790394] This time, there were 3011472 which is so anomalously
[17:04:15] <mozroute> [ 5197.790395] large that it probably signifies a problem with your
[17:04:15] <mozroute> [ 5197.790397] realtime configuration. For the rest of this run of
[17:04:17] <mozroute> [ 5197.790399] EMC, this message will be suppressed.
[17:04:29] <mozroute> what does this mean?
[17:04:59] <mozroute> my config is generated by stepconf
[17:06:40] <SWPadnos> that error means that yourt computer may not be suitable for use as a machine controller
[17:06:48] <SWPadnos> there are troiubleshooting tips onthe wiki
[17:07:55] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TroubleShooting
[17:11:53] <mozroute> what a pain...
[17:12:28] <SWPadnos> yes, it can be. not all PCs are suitable for realtime control
[17:12:54] <SWPadnos> if it's an Intel chipset, the SMI fix may help. It's hard to know though
[17:13:07] <SWPadnos> (easy to find out, but hard to predict)
[17:13:58] <mozroute> it's a via chipset
[17:14:03] <SWPadnos> bummer
[17:14:20] <mozroute> yah
[17:14:51] <mozroute> I'm going to look through the bios and see if I missed something in a minute
[17:14:59] <mozroute> how do I use halscope?
[17:16:05] <SWPadnos> err - that's a big question :)
[17:16:10] <SWPadnos> do you know how to use a real scope?
[17:16:12] <mozroute> :-)
[17:16:20] <mozroute> yes
[17:16:30] <SWPadnos> ok, then it should be easy
[17:16:43] <SWPadnos> click on a channel number, select the HAL "thing" to look at, and go from there
[17:17:25] <mozroute> I set channel 2 on halscope to parport.0.pin-02-out
[17:17:49] <SWPadnos> if you want to look at step/dir signals, you need to be sure the sampler is running in the base thread
[17:18:03] <mozroute> then I jog the x axis in Axis which has pin 2 for the step pin
[17:18:11] <SWPadnos> (click the "N samples at X Hz" button th cahnge that)
[17:18:13] <mozroute> uh, how do I do that?
[17:18:20] <SWPadnos> s/th/to/
[17:19:32] <mozroute> I set it to base thread when it started
[17:19:45] <SWPadnos> ok
[17:19:52] <SWPadnos> did you set the trigger?
[17:20:33] <mozroute> no, it's on normal
[17:20:37] <SWPadnos> you can trigger on any channel (you can then disable the channel if you want, it only has to be enabled when you set the trigger)
[17:20:52] <SWPadnos> well, you need to select the channel to trigger on, and the level (like a normal scope)
[17:21:30] <SWPadnos> and the direction, of course (rising/falling)
[17:22:26] <SWPadnos> if you're trying to measure jitter or something, it won't work. since the scope sampler runs in a HAL thread, it doesn't know when there's an error in the sample time
[17:22:34] <mozroute> I see nothing when I jog. I set the trigger up a little.
[17:23:06] <mozroute> on a real scope it will show something with the trigger anywhere.
[17:23:15] <SWPadnos> only in auto mode
[17:23:23] <mozroute> I don't see anything
[17:23:24] <SWPadnos> in normal mode, you won't get anything unless there's a trigger
[17:23:31] <SWPadnos> is it in run mode or stop?
[17:23:54] <mozroute> uh, stop :-)
[17:23:59] <mozroute> i'm a dummy
[17:23:59] <SWPadnos> heh
[17:24:17] <SWPadnos> as they say in polite society - "shite happens" :)
[17:24:50] <mozroute> I'm seeing step pulses now
[17:24:55] <SWPadnos> yay!
[17:25:14] <mozroute> what I'm not seeing is a charge pump signal
[17:25:38] <mozroute> I set pin 17 to charge pump out in stepconf
[17:25:49] <mozroute> do I need to do something else to make it run?
[17:45:09] <SWPadnos> you need "machine on" to get the charge pump output
[17:45:22] <SWPadnos> gotta run - going for a bike ride with the wife. good luck
[17:54:59] <mozroute> i did that, no luck
[18:32:32] <robin_z> * robin_z blinks
[18:33:06] <robin_z> * robin_z loves halsope ...
[18:33:24] <robin_z> i love the pre-trigger facility ...
[18:33:31] <robin_z> that is so neat
[18:33:42] <robin_z> "heres what happened just before ..."
[18:34:06] <robin_z> i guess modern digital scopes do that too
[18:37:24] <mozroute> I like a DSO
[18:37:34] <JymmmEMC> SWPadnos: !!!
[18:37:52] <JymmmEMC> * JymmmEMC waits for SWPadnos to disconnect...
[18:38:05] <JymmmEMC> 5... 4... 3.... 2... 1....
[18:38:15] <mozroute> that's all I use now - although I still don't know everything about them
[18:38:16] <tom1> he's riding his bike
[18:38:27] <JymmmEMC> tom1: (I know =)
[18:38:59] <mozroute> anyone have an idea why I'm not getting a charge pump signal?
[18:39:19] <JymmmEMC> mozroute: did you pay the electric bill?
[18:39:45] <tom1> no, is it connected to a thread?
[18:40:06] <tom1> deos the signal show up in hal-scope?
[18:40:08] <mozroute> :-)
[18:40:09] <tom1> does
[18:40:19] <mozroute> no signal in hal-scope
[18:40:31] <tom1> no signal change or no signal name?
[18:40:36] <mozroute> I used stepconf to set up everything
[18:41:06] <mozroute> I'm looking at parport.0.pin-17-out
[18:41:29] <mozroute> which is what I set to be charge pump in stepconf
[18:42:17] <JymmmEMC> mozroute: http://www.youtube.com/watch?v=jmBMcHK479M
[18:43:06] <mozroute> I don't have sound on this computer
[18:45:26] <tom1> out of my experience, but I'd use halscope, look at parport.0.pin-17-out, see that it ...pin17 is 'connected' to the signal 'charge pump'. I dont know the names but if pin17 is not connected to anything, you wont get the heartbeat you expect.
[18:45:29] <JymmmEMC> mozroute: Wait till you have, it's worth it =)
[18:46:17] <JymmmEMC> mozroute: do you have a cp signal at the parapport ?
[18:46:32] <mozroute> no
[18:46:44] <JymmmEMC> at the cp pcb?
[18:47:31] <mozroute> I haven't put a real scope on the pin because my scope is set up in the other room
[18:47:59] <mozroute> but hal-scope does show stuff when I jog the axes on the correct pins
[18:48:04] <JymmmEMC> mozroute: Well, comfirm that you have a REAL signal, then can assist in getting emc to see it
[18:48:22] <tom1> yeah, this cp originates on hardware ( not generated by emc )
[18:48:32] <mozroute> I looked at the charge-pump signal in hal-scope and see nothing.
[18:48:54] <JymmmEMC> mozroute: agian, confirm that you see the signal on the paraport cable first.
[18:48:55] <mozroute> ? I thought charge pump out would be generated by emc?
[18:49:04] <mozroute> that's what I need
[18:49:22] <JymmmEMC> mozroute: No, CP singal is generated by ext hardware, not emc.
[18:49:32] <JymmmEMC> well read by ext hw that is
[18:49:38] <tom1> i read that the pmdx board generates it ( only wiki reference i found, similar to your setup, not exactly like yours)
[18:50:09] <JymmmEMC> You can setup a CP to be full loop (that's what I thoutght you were tlaking about at first).
[18:50:40] <JymmmEMC> the PC generates a pulse, but it's read by a external hw device to make sure the PC is still alive and responding.
[18:50:40] <mozroute> well, mach3 puts out a charge pump signal at about 12 khz that we use to enable the outputs on our breakout boards
[18:51:29] <tom1> you can see if pin17 'works' by pulling it up with a resistor or down to gnd DONT TRY THIS AT HOME ( esp with mobo parports)
[18:51:47] <JymmmEMC> or toss on an led + resistor
[18:52:01] <JymmmEMC> two pin bi-colored LED's aere the best for this
[18:52:02] <mozroute> we have a jumper to bypass cp on most of our boards, but this one needs it for input switching as well so we don't have a way to bypass it
[18:52:25] <JymmmEMC> why would you bypass a safety device/feature?
[18:52:36] <mozroute> actually, this board has an LED to indicate charge pump. but it has to have a signal from the computer
[18:52:48] <tom1> this is not mach, dont expect same thing. and dont expect much mach knowledge here.
[18:54:01] <mozroute> JymmmEMC: well, i guess for a system that doesn't have a charge pump signal like EMC?
[18:54:24] <mozroute> or we use it for testing as well
[18:55:13] <mozroute> tom1: understood, I just figured charge pump was charge pump.
[18:55:41] <JymmmEMC> http://www.linuxcnc.org/docview/html//man/man9/charge_pump.9.html
[18:55:49] <mozroute> JymmmEMC says the PC generates a pulse - is that a one time pulse?
[18:56:13] <JymmmEMC> just found that in a search
[18:56:42] <tom1> Square wave if ’enable’ is TRUE or unconnected, low if ’enable’ is FALSE
[18:56:53] <mozroute> ok that sounds like what I expected.
[18:57:19] <tom1> it gnr8s a stream of pulses, you can connect that to p17
[18:57:42] <mozroute> setp parport.0.pin-17-out-invert 1
[18:57:42] <mozroute> net charge-pump => parport.0.pin-17-out
[18:57:42] <mozroute> here's the lines from the .hal file
[18:57:57] <mozroute> loadrt charge_pump
[18:57:57] <mozroute> net estop-out charge-pump.enable iocontrol.0.user-enable-out
[18:57:57] <mozroute> net charge-pump <= charge-pump.out
[18:58:27] <mozroute> the last three lines are toward the top of the file, the first two are toward the bottom
[18:58:45] <tom1> i dont see it connected to a thread
[18:58:57] <mozroute> addf charge-pump base-thread
[18:59:03] <tom1> k
[19:01:59] <tom1> does iocontrol.0.user-enable-out go TRUE ?
[19:05:34] <mozroute> net estop-out charge-pump.enable iocontrol.0.user-enable-out
[19:05:42] <mozroute> is that what you're talking about?
[19:06:00] <tom1> yes, looking at the enable's source
[19:06:04] <mozroute> I'm still quite new. I'm reading the HAL manual right now
[19:06:16] <tom1> np, still new to me too
[19:06:17] <mozroute> is that line correct?
[19:06:32] <robin> oh dear ...
[19:06:42] <robin> you'll not get on in here at all ..
[19:06:45] <robin> reading the manual!
[19:06:50] <robin> waht next! ;)
[19:06:51] <tom1> its ok, but i get a bit worried that the signal that enables the pump is called estop
[19:07:07] <mozroute> hmmm
[19:07:09] <tom1> ok/valid
[19:07:43] <tom1> its just words, but i am very literal
[19:07:48] <mozroute> net estop-out <= iocontrol.0.user-enable-out
[19:08:03] <mozroute> that line is nearly to the end of the file
[19:08:08] <tom1> out as in off ... haha ok
[19:08:22] <tom1> very visual
[19:08:38] <mozroute> bad?
[19:08:51] <tom1> still you should get the enable before you get the heartbeat
[19:08:55] <tom1> not bad
[19:09:02] <tom1> got enable?
[19:09:22] <mozroute> what does that mean?
[19:09:53] <tom1> does the enable signal go TRUE in halscope/halmeter?
[19:10:22] <tom1> as long as its LOW the comp will output LOW
[19:11:11] <tom1> sorry ... heartbeat/keepalive/chargepumpsquarewave same things
[19:11:40] <mozroute> halmeter shows a charge-pump signal as FALSE
[19:11:59] <robin> hmm im sure I had the 7i33 pinout someplace ...
[19:12:26] <tom1> right, thats the output of the comp, one of the inputs is ENABLE, it must be TRUE before the output starts to oscillate
[19:12:42] <mozroute> ok, i'm lost there
[19:13:51] <mozroute> you mean a port input is HIGH?
[19:15:14] <tom1> the 'comp' called 'charge-pump' has an input 'charge-pump.enable'. this input must be TRUE if the 'comp' will output pulses
[19:15:18] <robin> ok how do I figure out which pin of my m5120 is hooked up to which limit switch signal?
[19:16:07] <tom1> robin: physical pins on 5i20 had a big discussion months ago, can you search for it?
[19:16:35] <mozroute> so I need a line somewhere like charge-pump.enable = TRUE
[19:18:06] <robin> mmm, not really, im assuming its possible to work it out from the HAL stuff
[19:18:08] <tom1> mozroute: no, it should react the way you wired it, we're looking to see the state right now. not saying something is wrong, just want to know what it is right now. (it defaults to TRUE, i suspect something is making it FALSE)
[19:18:50] <robin> i just need + and - limits, plus home switch for each channel
[19:19:08] <tom1> robin: HAL talks about halnames, not really physical pins, thus my confusion (i was in the discussion too and ended up beginning a map)
[19:20:06] <mozroute> in stepconf I have an input (parport pin 10) set to ESTOP In
[19:21:23] <robin> well, color me stupid, but what users need is a "this pin does this" section for the sample configs
[19:21:24] <mozroute> aha!
[19:22:09] <mozroute> I set the pin 10 to unused in stepconf
[19:22:27] <mozroute> now I have a charge pump signal out
[19:22:35] <mozroute> the LED lights on the board
[19:24:09] <robin> sigh ..
[19:24:18] <tom1> mozroute: great! i dont see what pin10 had to do with it, but progres is great
[19:24:33] <robin> it really should not be this hard to figure out which pin is the limit nad home switch ...
[19:24:39] <tom1> robin: i been looking, and am same color as you ;)
[19:25:44] <robin> im going to resort to monitoring axis0.home on halscope and dabbing volts around the likely input pins ...
[19:26:56] <tom1> robin: sorry, the m5i20 box is not here, so notes are'nt either. but i did figger it from the hal-pin-show and the mesa docs. hard to reconstruct right now. no need to jeopardize your expensive hardware.
[19:27:13] <tom1> well be careful anyway
[19:27:47] <jmkasunich> robin: the home switch is on whatever pin you put it on
[19:28:22] <jmkasunich> if you are using a sample config, there will be a "net" or "linksp" line in the hal file that connects things
[19:28:55] <robin> ah ha .. and that says someting like "P2, pin 6"
[19:29:05] <jmkasunich> well, probably not
[19:29:27] <jmkasunich> it probably says something like m5i20.0.io-12-out
[19:29:42] <jmkasunich> mapping that "12" to a physical pin is the confusion?
[19:29:51] <robin> uh,
[19:29:57] <robin> let me read the hal file ...
[19:30:00] <jmkasunich> (I must admit I'm not sure myself what that mapping is, I'd have to go doc-digging
[19:30:02] <tom1> i think robin is asking... when the stock config file says linksp Xhome <= m5i20.0.in-02, then what physical pin is in-02
[19:30:34] <tom1> iirc, it is not 2nd pin
[19:30:38] <jmkasunich> tom1: the answer to that question is specific to the board you are using - 5i20 in this case
[19:30:48] <tom1> m5i20, yes
[19:31:00] <robin> so its m5120 in 0, in 1 and in 2 for neg, pos and home ...
[19:31:19] <jmkasunich> I agree it could be documented better - the answer is in a mix of mesa docs (physical pinouts) and driver docs (mapping to hal pins)
[19:31:38] <robin> nah,m i must have that wrong
[19:32:21] <robin> no, could be ..
[19:32:31] <robin> I was confused there for a moment ...
[19:33:07] <tom1> imo you can read the ...emcdir/configs/m5i20/m5i20_io.hal for the netlists, and must refer to the mesa docs for the physical side
[19:33:15] <robin> part of the confsion is the integrators manual ..
[19:33:24] <robin> itfor example .. port P3
[19:33:34] <jmkasunich> when we started the mazak at the CNC workshop we had the same issues - physical pinouts can be odd, and in our case we also had two different I/O boards, and they were driving opto-22 interface boards that added further confusion
[19:33:51] <robin> for the secodnary hal function it gives tis use, axis 4,ENC A
[19:33:55] <jmkasunich> so I wrote lots of notes in the hal file: http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/configs/demo_mazak/demo_mazak.hal?rev=1.20
[19:34:10] <robin> for the primary function it just says "in 00"
[19:35:20] <robin> maybe adding some descriptive text to the manual there would help? obviously it would only be correct for the sample config ..
[19:36:07] <jmkasunich> manuals can almost always be improved
[19:36:30] <robin> anyway, i know how to figure it out now
[19:36:38] <jmkasunich> in the case of 5i20, the issue is further complicated by the fact that a new family of drivers for it is in the latest rev, and I don't know which driver you are using
[19:36:45] <tom1> haha, i looked up at the wall behind the monitor... very long lists of logical-physical for another control ( a dozen 50 pin connectors , looks like a clothes line full of sheets :)
[19:36:47] <jmkasunich> the new one has (or will have) better docs
[19:38:06] <jmkasunich> well, I gotta go unload the truck
[19:38:14] <tom1> may be of help (a pre-test) http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?M5i20_Halvcp_Test_Panel
[19:38:14] <jmkasunich> as soon as I figure out where to put everything
[19:38:31] <tom1> moving John? be careful!
[19:38:37] <jmkasunich> no
[19:38:46] <jmkasunich> just bought some shelving from a store that was going out of business
[19:38:53] <tom1> cool
[19:39:09] <jmkasunich> 24 pcs of 22" x 48" nice formica covered particleboard, plus 4 bays of 6' high framing
[19:39:39] <tom1> store racks? (gondola's) good storage stuff
[19:41:59] <jmkasunich> they were back room storage, not fancy
[19:42:04] <jmkasunich> but still quite nice
[19:48:23] <robin> hmmm
[19:48:26] <robin> so ..
[19:49:37] <robin> linksp Xhome <= m5i20.0.in-02
[19:49:37] <robin> linksp Xhome => axis.0.home-sw-in
[19:49:54] <robin> thats the M5i20 input 02 right?
[19:50:31] <robin> which, is port P3, pin 5 right?
[19:51:26] <robin> * robin puts his real world multimeter on pin 5 and cofirms it flips 0 to 5V
[19:52:08] <robin> ah wait,
[19:52:12] <robin> * robin slaps his head
[19:52:21] <robin> opti isolated
[19:52:22] <tom1> i thin... P2pin5 is IO2
[19:52:34] <robin> it is ...
[19:52:39] <robin> ?
[19:53:07] <tom1> connector2 not connector3 ( reading mesa manual )
[19:53:10] <robin> no, P2 is all the analogu stuff, encoders and DACs
[19:53:17] <robin> ah yes
[19:53:25] <robin> ok connector 3, port 2 ...
[19:53:37] <tom1> P2 CONNECTOR PINOUT
[19:53:37] <tom1> 1 IO0 2 GND 3 IO1 4 GND
[19:53:37] <tom1> 5 IO2 6 GND 7 IO3 8 GND
[19:53:58] <robin> then theres the 7i33 card ...
[19:54:05] <mozroute> ok, I'm looking at the idoubler.comp file jepler made for me last week, but I'm not sure how to use it.
[19:54:30] <tom1> oh, you want to make this easy ;)
[19:55:11] <robin> 7i37 I meant
[19:56:11] <tom1> sorry, i got a panel to layout, but it seems you are on the way. IF you doc anything, please post it, cuz i know i've re-constructed the map too many times
[20:07:04] <robin_z> ah john t
[20:07:14] <robin_z> cradek suggested I drop you a note ...
[20:07:27] <BigJohnT> ok
[20:07:40] <robin_z> integrators manual ...
[20:07:52] <BigJohnT> * BigJohnT listens
[20:08:02] <robin_z> references to MIN_OUPUT = blah ah incorrect, apparently
[20:08:18] <BigJohnT> what section?
[20:08:23] <robin_z> hal only uses the MAX_OUTPUT and uses it symmetrically ..
[20:08:33] <robin_z> so crade tells me from the code anyway
[20:08:45] <robin_z> um section 4, page 29 was one instance
[20:08:49] <robin_z> but there are others
[20:09:15] <robin_z> apparently MIN_OUTPUT is not used anymore ...
[20:09:20] <robin_z> so can be zapped
[20:09:21] <BigJohnT> give me a key word or three my section numbers might not be the same
[20:09:48] <robin_z> cant you just search for MIN_OUTPUT?
[20:10:01] <BigJohnT> yep :)
[20:10:23] <BigJohnT> just got up from a nap... best excuse I could come up with
[20:10:25] <robin_z> its on an diffeent machine in a diffeent room on a different floor, or id go look
[20:10:39] <BigJohnT> np
[20:11:14] <BigJohnT> MIN_OUTPUT = -10 (HAL) The minimum value for the output of the PID compensation that is writ-
[20:11:16] <BigJohnT> ten to the motor amplifier, in volts. The computed output value is clamped to this limit. The
[20:11:18] <BigJohnT> limit is applied before scaling to raw output units.
[20:11:20] <BigJohnT> this one?
[20:19:08] <BigJohnT> found it!
[20:28:34] <BigJohnT> robin_z: it has been corrected and commited, thanks for helping
[20:28:50] <robin_z> thats the one
[20:44:11] <jymm> jymm is now known as MrAsshole
[20:44:24] <MrAsshole> MrAsshole is now known as Jymm
[20:49:36] <robin_z> * robin_z plays http://fantasticcontraption.com/
[20:50:51] <alex_joni> robin_z: we finished it in here a couple times :P
[20:51:20] <robin_z> level 12
[20:51:57] <alex_joni> those are still pretty simple :P
[20:52:22] <alex_joni> but after I finished them I still spent a lot of time looking at other ways of solving them
[20:52:26] <alex_joni> you should try too
[20:52:27] <alex_joni> :D
[21:06:23] <jepler> ooh my new spindle came
[21:09:01] <BigJohnT> :)
[21:09:14] <BigJohnT> a high speed one?
[21:09:34] <robin_z> defined as 100K rpm +?
[21:10:30] <robin_z> theres a nice youtube thing of a 120k rpm spindle somewhere
[21:15:10] <jepler> no, only about 20k
[21:29:14] <mozroute> do I want apic mode enabled in the bios?
[21:32:15] <mozroute> also, i noticed my parport was set to ECP with DMA 3. do I want it to be EPP or does it make a difference?
[21:34:41] <jepler> mozroute: unless the signals don't get correctly output, there's no need to change the setting in the BIOS
[21:34:55] <jepler> for step & direction emc uses the oldest (SPP) style of interface
[21:35:21] <mozroute> ok. I had seen people mention EPP and didn't know if it made a performance difference
[21:35:43] <mozroute> how do I go about using the idoubler.comp file you made last week?
[21:36:19] <jepler> well -- step 1 is "understand HAL" and step 2 is "understand idoubler.comp".
[21:36:44] <mozroute> ok :-)
[21:37:18] <mozroute> I'm reading the HAL stuff. I've got to leave for a while right now though.
[21:37:32] <jepler> basically, you create 5 connections between idoubler and the parallel port in your hal files (net or newsig/linkXX), make the idoubler code be executed (addf) and then it will have 8 values available in hal that should represent the values read from the multiplexed pins
[21:37:59] <mozroute> one, thing though. we use the charge pump signal right now to switch the input banks.
[21:38:39] <mozroute> do we just turn off the charge pump and use the signal in idoubler?
[21:38:50] <jepler> that's one way to do it
[21:38:52] <mozroute> generate the signal i mean...
[21:39:07] <jepler> you could also alter idoubler so that it takes a signal which indicates which bank to read, instead of generating it
[21:39:16] <mozroute> ok. I'll look at it some more this evening.
[21:39:40] <mozroute> oh, like looking at the cp signal?
[21:39:44] <mozroute> bbl
[21:39:45] <jepler> yeah
[21:40:53] <jepler> loadrt idoubler
[21:40:53] <jepler> net mult0 parport.0.pin-90-in => idoubler.0.in-0 # and similar for 1, 2, and 3
[21:40:53] <jepler> net msel idoubler.0.select => parport.0.pin-94-out
[21:40:53] <jepler> addf idoubler.0 base-thread
[21:40:59] <jepler> the hal file fragment might look a bit like this
[21:41:26] <jepler> then you hook the idoubler.0.out-# pins to whatever you want, like axis.0.neg-lim-sw-in: net idoubler.0.out-0 => axis.0.neg-lim-sw-in
[21:50:44] <jepler> pro tip: don't start the spindle while the tool is pressing against the tool length sensor
[21:52:01] <BigJohnT> actually HAAS does that when you have an insert type tool and runs it slowly backward to find each insert
[21:56:28] <BigJohnT> jepler: do you have pictures?
[21:57:21] <jepler> BigJohnT: not yet
[21:57:31] <BigJohnT> :)
[21:59:28] <jepler> it's this spindle http://i20.ebayimg.com/06/i/000/8b/67/547d_1.JPG with motor http://i18.ebayimg.com/03/i/000/a2/29/85e9_1.JPG in a mount like this http://i8.ebayimg.com/04/i/000/8d/22/f462_1.JPG
[22:00:40] <BigJohnT> * BigJohnT wishes he didn't have to use a string and two tin cans to connect to the internet
[22:01:13] <BigJohnT> nice spindle
[22:02:25] <BigJohnT> * BigJohnT goes back out to the shop to finish the new drive shaft for the Y axis gantry
[22:12:01] <jepler> first test looks good .. it can cut just the tiniest little trace instead of the .010" or wider from the dremel
[22:17:23] <tom1> how tiny? sounds impressive
[22:18:36] <jepler> well .. I don't have a very good tool for measuring
[22:19:32] <jepler> I performed a test where the spindle is running and you touch the tool to a piece of copper clad, and then try to measure the size of the copper it mills away
[22:19:50] <jepler> but the best thing I have to do this with is a set of calipers and my eyes
[22:20:08] <jepler> with the dremel the dot was bigger than the gap between the calipers when I set it to a reading of .015 inches
[22:20:20] <tom1> lets see, lay a hair in it and deep mike onto it and find the height of the chord.... :)
[22:20:57] <jepler> with the new spindle it no longer has visible roundness, it's just a dot
[22:21:45] <tom1> nano nano (metric mork)
[22:25:43] <jepler> (the quoted runout for the spindle is .0004" and for the tool .0002")
[22:28:51] <robin> so
[22:29:05] <robin> to home the X axis .. i enter what on the MDI?
[22:29:42] <robin> hmm, it homes in the wrong direction
[22:29:49] <robin> the home sw is at the +ve end ...
[22:31:39] <pjm> invert the X direction pin in stepconf just to test it
[22:31:49] <robin> que?
[22:31:50] <jepler> robin: if the first step of homing goes the opposite direction from what you want, negate HOME_SEARCH_VEL
[22:31:59] <robin> * robin nods
[22:32:05] <jepler> inverting the pin will have a similar effect, as long as it's not a shared home and limit switch
[22:32:06] <robin> just done that .. restartying
[22:32:20] <jepler> pjm: (I think robin has a servo interface board, so he's not using stepconf)
[22:32:24] <robin> this is servos and mesa
[22:32:46] <pjm> jepler ahh ok thanks
[22:33:01] <pjm> i've only ever used the parport drivers
[22:35:32] <pjm> i ran a program to engrave a picture today with EMC with 84974 lines, and it ran flawlessly
[22:37:25] <jepler> great
[22:39:15] <pjm> yes its an extreamly impressive piece of software, the developers deserve lots of praise for sure
[22:46:28] <jepler> hmph, I've still got backlash pushing my holes off center
[22:46:37] <jepler> I thought I'd mostly overcome that with software backlash correction
[22:47:30] <robin> humoh
[22:47:35] <robin> doesnt seem right ...
[22:47:46] <robin> aproachs home ...
[22:47:51] <robin> finds it at speed
[22:47:56] <robin> backs off ...
[22:48:02] <robin> aproaches agains slowly ...
[22:48:14] <robin> activates switch ...
[22:48:18] <robin> keps going ...
[22:48:33] <robin> passes over it ....
[22:48:37] <robin> hmmm
[22:50:21] <robin> maybe latch velocity needs to be negative too ...
[22:52:21] <jepler> in my config I my SEARCH_VEL and LATCH_VEL are both negative. both times (coarse and fine steps of homing) it works by moving onto the switch from right to left.
[23:03:05] <robin> so, tell me ...
[23:03:27] <robin> yeah, making both -ve worked a treat, thanks
[23:03:30] <robin> so ..
[23:03:33] <robin> limit switches
[23:03:51] <robin> I have +ve and -ve switches for the axis ...
[23:04:00] <robin> if ive run into the +ve limit ...
[23:04:19] <robin> it wont let the machine come on to jog it off ...
[23:06:04] <robin> some machines allow you to jog -ve off a +ve limit switch .. and vice versa
[23:06:46] <robin> apart from tyhat the homing and softlimits seem to work a treat :)
[23:07:51] <robin> and its picking up the index pulse too
[23:08:49] <jepler> in emc you have to use "override limits" even if you are only on one hard limit switch
[23:09:03] <jepler> but if the limit switches are independent, emc will only let you jog off of the switch that's tripped
[23:23:44] <robin> hmmm
[23:23:51] <robin> im indpent
[23:24:28] <jepler> spindle blogged, photographed: http://axis.unpy.net/index.cgi/01218927215
[23:27:17] <BigJohnT> nice setup jepler
[23:27:50] <robin> but as soon as I try to tunr it back on I get "joint 0 limit switch error" .. so no opportunity to jog off
[23:29:22] <BigJohnT> BBQ time
[23:33:46] <tom1> jepler: looks very nice ( and the traces are way better ), where did the spindle come from? with motor?
[23:34:02] <mozroute> I don't find comp
[23:34:22] <mozroute> do I need the emc source to compile a comp file?
[23:34:48] <tom1> docs at 'man comp'
[23:35:08] <tom1> no need for emc src to make new comps
[23:36:00] <mozroute> ok, man comp pulls up the manpage, but 'comp' gives me a file not found
[23:36:01] <jepler> mozroute: no, you need the emc2-dev package and the build-essential package.
[23:36:04] <tom1> src is its just a text file you write and then follow the docs to 'build' the comp
[23:36:20] <tom1> oh, what jepler said
[23:36:35] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ContributedComponents#How_to_compile_and_install_a_component
[23:37:04] <mozroute> ah. i'll get it. this is to be my router controller so I didn't install all the dev stuff on it. I may have to though at some point
[23:37:28] <jepler> I don't think it fits on the CD or we'd do it by default
[23:37:45] <jepler> (at least enough to build .comp files, probably not enough to rebuild emc2 completely from scratch)
[23:38:02] <mozroute> I installed from the web with the script.
[23:38:13] <mozroute> I have all the dev stuff on my main machine
[23:40:01] <mozroute> did you get that spindle off ebay? it looks nice
[23:41:38] <jepler> yeah - they are listed on and off by this guy http://stores.ebay.com/Wolfgang-Engineering
[23:59:20] <tom1> yes, wolfgang, plz jury it after some use. they look great