#emc | Logs for 2010-03-30

[00:25:45] <jimbo> OK feeling dumb now. When using mux4 to change to different values It uses a Binary input to control the value. How do you do the same thing with dedicated a switches?
[00:33:29] <tom3p> how to make a real switch act like mux4?
[00:35:47] <tom3p> connect the 'wiper' to the device that gets the selected signal. connect 4 signals to 4 positions that the wiper can connect to.
[00:43:53] <jimbo> mux4 has a single output with 4 variables IE .001, .01, .1 etc. based on the binary input it dumps the associated variable into the system. I see no way to assign a variable with a single switch closure. Any uses I have seen use a 74H148 to create the binary code to talk to mux4.
[00:45:32] <WalterN> hey mkillebrew!
[00:46:22] <WalterN> you stalker
[01:11:36] <Dave911> tom3p: So they were not doing shot position/velocity profiling etc on the Korean machines? That was a big deal back in the early 90's on high end machines. I wasn't sure how high end that kind of thing went.
[05:10:30] <Jymmm> This is awesome!!! http://www.antimonkeybutt.com/ You gotta buy it just to put on the shelf and for ppl to go WTF?!
[05:12:44] <pfred1> Jymmm no, whats awesome is you're posting a URL outside of graigslist's domain!
[05:12:56] <Jymmm> pfred1: STFU
[05:13:23] <Jymmm> pfred1: You just nag bitch and complain about everything, geeeze.
[05:14:59] <pfred1> Jymmm http://www.abymc.com/Shop/Ron_Monkey_Dust.jpg
[05:18:55] <pfred1> Jymmm want to see my latest board I designed?
[05:19:14] <Jymmm> sure
[05:19:23] <pfred1> Jymmm http://img94.imageshack.us/img94/9922/hsoptobrdc.png
[05:19:38] <pfred1> 88.6% complete!
[05:19:54] <Jymmm> what's it do?
[05:20:03] <pfred1> its a stepper motor driver
[05:20:35] <pfred1> with optical isolation
[05:20:59] <Jymmm> Not to be funny, but why? There are so many stepper driver boards out there.
[05:21:23] <pfred1> yeah and hmm how can I put this delicately?
[05:21:25] <pfred1> they suck
[05:21:49] <MrSunshine> how many amps/voltages?
[05:21:58] <pfred1> well some aren't so bad i guess but they are made from difficult to obtain parts
[05:22:16] <pfred1> MrSunshine well this is good to 44 V 3 amps per phase unipolar
[05:22:34] <Jymmm> Well sure. But gecko is only ~$150 with all the bells and whistles and compensation too. What's REALLY needed is a damn good breakout board.
[05:22:38] <MrSunshine> and its a unipolar board?!
[05:22:44] <pfred1> Jymmm only?
[05:23:01] <pfred1> Jymmm there's not $10 of parts on this board
[05:23:05] <MrSunshine> * MrSunshine has ordered 3x 80V 7.8A drives from the UK =)
[05:23:14] <Jymmm> pfred1: and the cost of the baord itself?
[05:23:25] <MrSunshine> about 150$ each :)
[05:23:27] <pfred1> Jymmm they're like $6
[05:23:36] <Jymmm> pfred1: from where?
[05:23:48] <pfred1> radio shack had copper clad board
[05:24:13] <pfred1> but i wire up on regular plain phenolic
[05:24:52] <Jymmm> http://axis.unpy.net/index.cgi/etchcnc
[05:25:17] <MattyMatt> http://www.goldmine-elec-products.com/prodinfo.asp?number=G2626
[05:25:27] <MattyMatt> 99c for 4x8.
[05:25:41] <Jymmm> $37.50 for shipping =)
[05:26:38] <MattyMatt> it's cheap to US, non-existent to EU, iirc
[05:26:48] <pfred1> I wish I could get 7062s
[05:27:01] <MrSunshine> is it hard to add encoder input to emc?
[05:27:03] <pfred1> or 7078s even
[05:27:16] <MrSunshine> i was thinking of adding these little magnetic encoders to my steppers to detect error conditions
[05:27:42] <MrSunshine> 1024 steps/rev on the encoder, atleast can detect if the motors has stalled
[05:27:50] <MattyMatt> if you can turn them into digital signals on parports, HAL can handle them
[05:28:31] <MrSunshine> would be nice with more then one parport then as i want limit switches on the mill also :)
[05:28:53] <pfred1> heck I bought a dual port card for $15
[05:29:05] <MrSunshine> http://www.rls.si/default.asp?prod=am4096
[05:29:08] <MrSunshine> oh 12 bit =)
[05:29:26] <MrSunshine> hellt hat has more bits then my stepper has steps atm :)
[05:30:19] <MattyMatt> an encoder usually uses 2 bits
[05:30:38] <MrSunshine> MattyMatt, humm ?
[05:30:39] <MattyMatt> maybe 3, 1 for an index
[05:30:50] <MattyMatt> 2 for quadrature
[05:31:57] <MrSunshine> MattyMatt, well thats ordenary encoders, these are for a rotating magnet? :)
[05:32:14] <Eric_K> you wouldn't need the index, but power is probably required
[05:32:29] <Eric_K> rotating magnet is some kind of hall effect device?
[05:33:04] <MrSunshine> dont know how they work realy
[05:33:14] <MrSunshine> but aparently people rebuild their servos for chips like that
[05:33:20] <MrSunshine> place a magnet inside of it and put the chip in
[05:33:50] <pfred1> what is the difference between an opto-coupler and an opto-isolator?
[05:34:01] <MattyMatt> none, afaik
[05:34:09] <pfred1> MattyMatt http://www.goldmine-elec-products.com/departments.asp?dept=1144
[05:34:24] <pfred1> MattyMatt apparently they warrant different departments!
[05:34:42] <Jymmm> http://en.wikipedia.org/wiki/Opto-isolator
[05:34:59] <MattyMatt> different warranties
[05:35:04] <Eric_K> pfred1: many, many different types
[05:35:09] <MrSunshine> but only 500hz at 4096 resolution
[05:35:24] <MrSunshine> oh well, hz is measured in seconds isnt it? :OP
[05:35:34] <MrSunshine> so 30k rpm :P
[05:36:19] <MrSunshine> aparently can be used as a tachometer also :)
[05:36:24] <MrSunshine> directly out of the chip :)
[05:42:10] <pfred1> oh damn i may have to get some of these!
[05:42:18] <pfred1> http://www.goldmine-elec-products.com/prodinfo.asp?number=A20538
[05:42:59] <MattyMatt> http://www.goldmine-elec-products.com/prodinfo.asp?number=G14197B
[05:43:27] <pfred1> MattyMatt they're floppy disc motors
[05:43:38] <MattyMatt> yeah, but 250 of them :)
[05:43:52] <MattyMatt> with nice connectors
[05:44:13] <pfred1> I tried to run some floppy drives here the other day all of mine are messed up!
[05:44:27] <pfred1> I couldn't make a boot disk on one machine and get another to boot off it
[05:44:28] <MattyMatt> that enough to make a whole team of robocup players
[05:45:22] <MattyMatt> or a whole wall of fake meters for "the lab"
[05:45:35] <elmo40> anyone looking for a 5-axis? http://toronto.en.craigslist.ca/tor/tls/1666146984.html
[05:45:40] <elmo40> HUGE machine!
[05:46:35] <MattyMatt> comes with 5t crane for work loading :)
[05:47:31] <elmo40> for $250k I sure hope so!
[05:47:42] <MattyMatt> when I get one that big, I'll build it into the fabric of the building
[05:48:04] <MattyMatt> supporting a roof is a trivial extra chore
[05:49:45] <MattyMatt> a stone cutting machine should go to the quarry and work the stone there :)
[05:49:56] <MattyMatt> for $250k it should anyway
[05:53:29] <MattyMatt> that would svae transport cost
[05:56:42] <elmo40> true
[05:56:45] <elmo40> http://www.teammachinetools.com/used/used-available5.php
[05:56:48] <elmo40> need parts?
[05:57:31] <MattyMatt> not in america
[05:57:57] <elmo40> it was a question for the room ;)
[05:58:00] <elmo40> and Canadians ;)
[05:58:12] <MattyMatt> it's a different market here in UK. a whole new ecosystem of old UK machines, although there is a bridgeport on ebay atm
[06:01:45] <MattyMatt> probably came over in WW2
[06:02:48] <MattyMatt> yet today, it's cheaper to get one transported from china than made here or north america :)
[06:04:23] <MattyMatt> NAFTA & EU need a reciprocal agreement to drop import duties etc
[06:04:42] <MattyMatt> we're all getting carved up by the new powers
[06:05:39] <MattyMatt> I'm moving to India while I can still afford to :)
[06:06:11] <MattyMatt> Bangladesh, now there's somewhere that needs stonecutting robots
[06:06:50] <MattyMatt> mountains in the north, floods in the south. the solution seems obvious to me
[06:13:05] <mkillebrew> have killbots wipe out India so we don't have to worry about flooding?
[06:13:14] <mkillebrew> or am I on the wrong track
[06:13:57] <MattyMatt> they would win a killbot race, if they tooled up to retaliiate
[06:45:38] <MattyMatt> does anyone else think BiSS interfaces are a good idea?
[06:46:15] <MattyMatt> machine control has to go serial sometime
[06:46:57] <Eric_K> seems to me that serial interface implies off-board intelligence
[06:48:16] <MattyMatt> yep. one cpu per motor/sensor
[06:48:24] <MattyMatt> cheap as chips
[06:51:32] <Eric_K> it could work if the interface looked like a dac/encoder :)
[06:52:25] <MattyMatt> nope, the controller software would work with the serial protocol
[06:52:45] <MattyMatt> no legacy quadrature stuff anywhere
[06:53:33] <MattyMatt> dunno how you'd sync it all, I'd have a separate system wide timebase clock pulse
[06:54:01] <MattyMatt> rather like the PS/2 keyb interface
[06:55:25] <Eric_K> the idea that the off-board processors would be cheaper than a PC seems unlikely to me
[06:56:48] <MattyMatt> look at the price of usb adaptors. they cost barely more than the connectors, yet have a 50¢ cpu with a protocol converter
[06:57:03] <Eric_K> wrong model
[06:58:34] <MattyMatt> PC doesn't scale well. a biped robot could use more steppers than emc2 could handle
[06:59:09] <MattyMatt> a serial system could delegate more, like ethernet routers
[06:59:16] <Eric_K> step/dir interface is a bit of a kluge
[07:00:03] <Eric_K> if you could get an off board processor that looked like a dac/encoder, emc could easily do a large number of steppers
[07:00:49] <MattyMatt> for curve cutting, the advantages of having a central cpu doing the stepping is worth it
[07:01:12] <MattyMatt> otherwise you'd need a global timebase or sth to coordinate it
[07:02:32] <archivist> there are a couple of timebase systems out there
[07:03:03] <archivist> one is toy stepper system and the other is high end servo
[07:04:07] <MattyMatt> guess which one I'm likely to be interested in :) clue http://www.goldmine-elec-products.com/prodinfo.asp?number=G14197B
[07:05:47] <archivist> you need all 250 to get sensible torque
[07:06:39] <MattyMatt> 250 rubber wheels under the bridgeport
[07:07:29] <Eric_K> looks like my motor collection, but I've got significantly more power output potential
[07:07:30] <elmo40> MattyMatt: the Yaskawa Motoman robots have one CPU per motor. with another CPU keeping them all in line.
[07:07:37] <elmo40> I think that is a great way to do it
[07:07:53] <elmo40> cpu's are cheap. not like you need a quad-core or anything :-P
[07:08:29] <Eric_K> otoh, those drives cost 20x my PC motherboard
[07:09:17] <MattyMatt> I found my PC at the dump
[07:09:36] <Eric_K> so infinity x your motherboard
[07:09:44] <MattyMatt> but it needed a new PSU, so that's $5 worth
[07:11:08] <MattyMatt> nice machine. 1.2 athlon, 512MB, dvd burner, usbs on front
[07:12:14] <MattyMatt> I suppose moores law will keep PC mobos cheap for a few more years
[07:12:56] <MattyMatt> but they aren't really compact, compared to an embedded cpu
[07:13:43] <Eric_K> there is no reason for commercially available embedded processors to get too cheap
[07:13:58] <MattyMatt> there is. competition
[07:14:27] <Eric_K> unless you're looking at something an arduino can do
[07:14:40] <MattyMatt> that's why mips are resurging. they are 2¢ cheaper than arm
[07:16:08] <MattyMatt> it doesn't take much of a cpu to manage 1 motor and a few sensors
[07:16:12] <Eric_K> but that price doesn't translate to you, it's going to the people using a million chips
[07:16:41] <Eric_K> it doesn't take much, but then why isn't the grex a raging success?
[07:18:13] <MattyMatt> the airbrush & nailgun people?
[07:19:22] <MattyMatt> BiSS doesn't seem to be catching on much either, that's supposed to be the industry wide open standard
[07:20:45] <Eric_K> grex is a gecko product
[07:21:10] <Eric_K> truth is, it's a lot of work to take all this stuff off-board
[07:22:24] <MattyMatt> at the design stage, but in making a motor it's just one small chip on the driver board
[07:22:39] <Eric_K> and since nobody will ever make the off-board processor look like a dac/encoder, the modifications to emc would be non-trivial as well
[07:23:11] <Eric_K> the chip is nothing, consider it done
[07:23:27] <Eric_K> then you still have more work than you can handle to do the rest
[07:25:52] <MattyMatt> yeah I accept it's non-trivial, but I do think it's the inevitable future
[07:26:23] <MattyMatt> like airliners and battleships are fly-by-wire now, just to save all the power cables running everywhere
[07:26:55] <MattyMatt> model railways are now, even
[07:30:39] <Eric_K> much easier though, dcc doesn't keep you from crashing 2 trains together
[07:31:21] <MattyMatt> software could :)
[07:31:49] <Eric_K> and you might get a magic pony tonight too
[07:36:09] <MattyMatt> http://www.youtube.com/watch?v=dyMXYE_50Ts&NR=1
[07:37:53] <herron_> herron_ is now known as herron
[07:41:06] <MattyMatt> hmm bridgeport still 94 quid. that'll go less than 500 I bet
[08:04:30] <MattyMatt> does the tidemark in the rustline imply it's been used for production?
[08:04:39] <MattyMatt> http://cgi.ebay.co.uk/Bridgeport-Milling-Machine-with-DRO_W0QQitemZ330417292947
[08:07:13] <MattyMatt> 3 genuine bidders
[08:19:27] <MattyMatt> DRO implies it hasn't been used for repetetive production doesn't it? for that you'd use stops
[08:20:26] <MattyMatt> well I've not got room, or a flatbed to pick it up
[08:26:24] <MattyMatt> Invisible lathe -> http://cgi.ebay.co.uk/myford-lathe_W0QQitemZ320504799044
[09:28:07] <piasdom> g'mornin all
[12:04:29] <JT-Dev> anyone run the sample g76.ngc on the lathe sim?
[12:11:43] <herron> yer, it assumes too much
[12:12:07] <herron> needs hacking to suit
[12:12:40] <JT-Dev> it needs fixing :)
[12:12:42] <herron> and took too much per pass on my little toy
[12:13:02] <JT-Dev> It won't run on the sim at all
[12:18:34] <herron> I started writing a modded version to feed back but never got enough round tuits in stock
[12:19:09] <herron> JT-Dev, btw /me is archivist at a clients :)
[12:19:32] <JT-Dev> ok, I was wondering who it was :)
[12:24:32] <JT-Dev> good to hear your at a client
[12:27:26] <herron> just a small bit of web site work
[13:03:36] <elmo40> morning people. 1125oz-in is good for a typical bridgeport-type kneemill?
[13:04:53] <elmo40> Nema 34 sized servo motors.
[13:19:50] <SWPadnos> elmo40, continuous or peak?
[13:48:14] <elmo40> peak
[13:48:25] <elmo40> don't know the continuous
[13:48:27] <elmo40> will look that up
[13:48:55] <SWPadnos> ok. continuous is what you design for, and it'll be 20-25% of peak (usually)
[13:49:03] <SWPadnos> those are a little small, but usable
[13:49:22] <elmo40> wow. 'constant' torque is 1/3!
[13:49:32] <elmo40> 226
[13:49:39] <SWPadnos> the motors I have, which are from an Anilam conversion, are ~500 oz-in continuous, ~1800 peak
[13:49:40] <elmo40> ok, less then 1/3
[13:49:49] <elmo40> ok
[13:49:55] <SWPadnos> around 1/4.5 :)
[13:51:09] <elmo40> would a 1125oz-in stepper be like that as well?
[13:51:22] <SWPadnos> no, the stepper would suck more
[13:52:28] <SWPadnos> from what I've heard, you need something closer to an 1800-2000 oz-in stepper to get close in performance
[13:52:42] <SWPadnos> the 1125 will work, but at lower speed and with lower accel
[13:55:54] <elmo40> for such a size I am surprised they are so weak
[13:56:27] <elmo40> what can the Nema34 frame handle? 2000? 4000?
[13:56:32] <SWPadnos> dunno
[13:57:49] <Jymmm> 4 gazillion metric tones
[13:58:14] <Jymmm> ...and one small diet coke
[14:01:16] <elmo40> ok, I think I found my spindle :) http://www.kelinginc.net/CNCSpindleandController.html#e50
[14:01:40] <elmo40> diet? all I have is caffeine free
[14:05:33] <sealive> the tourch of nema34 is 3Nm
[14:05:43] <sealive> at 4,2A
[14:24:13] <elmo40> a tourch? did he mean Torque?
[14:24:25] <elmo40> anyways... anyone use USB? http://www.warp9td.com/
[14:25:50] <elmo40> would that add latency to the system compared to parallel?
[14:27:32] <elmo40> so many people make Mach3 stuff (that USB thing has a mach3 'plugin' for something)
[14:27:42] <elmo40> why wouldn't people make open source stuff?
[14:28:15] <elmo40> ya, I know, long winded conversations ensue over that one... I guess people need to learn the Linux world first, eh
[14:29:06] <cradek> elmo40: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Emc2HardwareDesign
[14:29:39] <elmo40> However, no one has ever submitted a driver that uses rt-usb
[14:29:40] <elmo40> :P
[14:29:57] <elmo40> my point exactly... can't take that driver from the site and fenagle it?
[14:30:09] <elmo40> but there is a timing issue.
[14:30:42] <elmo40> on a production machine, such as a Haas or Mazak, do they use parallel communication from their computers? Is that just the 'norm' ?
[14:31:37] <celeron55> in a professional setup you're supposed to use some realtime i/o cards i guess
[14:31:39] <elmo40> I know the latest Mazaks use fibre optics to their motors
[14:31:40] <cradek> it depends what you mean by parallel
[14:31:52] <elmo40> cradek: instead of serial
[14:32:23] <cradek> do you mean a pc-style serial or parallel port? because they surely use neither
[14:32:24] <Jymmm> Maximum Step Frequency is 4 MHz
[14:33:09] <cradek> they probably use something much more like a mesa pci card to give lots of simultaneous fast realtime IO
[14:35:05] <Jymmm> MPG == Miles Per Gallon
[14:35:49] <Jymmm> ?
[14:36:07] <cradek> Jymmm: are you off your meds again?
[14:36:40] <JT-Work> lol
[14:36:45] <Jymmm> cradek: http://en.wiktionary.org/wiki/MPG
[14:38:17] <Jymmm> cradek: http://en.wikipedia.org/wiki/MPG_%28gene%29
[14:39:29] <cradek> are you seriously asking a question, and if so, what is it?
[14:40:02] <cradek> "my car gets 26 mpg" => miles per gallon
[14:40:22] <cradek> "you need a mpg to run touchy" => manual pulse generator
[14:41:12] <Jymmm> What's that? Like a heartbeat of sorts?
[14:41:47] <cradek> http://www.motioncontrol.com/media/articlesProducts/2008/Micronor112308_1.jpg
[14:42:07] <Jymmm> rotary encoder?
[14:42:31] <cradek> yes it's an encoder that you turn by hand. it has 100 marks on the wheel, with a detent for each one
[14:43:03] <Jymmm> TY =)
[14:43:20] <elmo40> ooohhh, this would be awsome to have! http://pdscolombo.com/prod_compact.php
[14:43:29] <Jymmm> Heh, it looks a lot like a safe dial
[14:43:33] <cradek> welcome - I had no idea that is what you were asking :-)
[14:43:46] <elmo40> cradek: never heard of 'mesa pci card'.
[14:44:27] <elmo40> does EMC2 support them?
[14:44:45] <cradek> yes
[14:44:49] <Jymmm> elmo40: http://webcache.googleusercontent.com/search?q=cache:2P99HiDEoIUJ:www.mesanet.com/motioncardinfo.html+mesa+5i20&cd=1&hl=en&ct=clnk&gl=us
[14:44:57] <elmo40> I am seriously designing and *going* to make a large scale 4-axis machine.
[14:45:34] <Jymmm> elmo40: Heh, and I'd like to make a super small scale one =) (3 axis)
[14:46:50] <elmo40> super small? how small? 2" cube small?
[14:46:56] <elmo40> just uber-accurate, right? ;)
[14:47:28] <Jymmm> http://sfbay.craigslist.org/sby/tls/1668225613.html
[14:47:50] <elmo40> 40-pitch thread? 3000 oz-in motors? cat-40 tool holder? all for 2" cube table :-P
[14:47:54] <Jymmm> Are BP still being made?
[14:49:18] <elmo40> I hate those ads. they never give any real information! is it an are-8 or cat-30? single or 3-phase? 'brand new' means nothing to me...
[14:49:28] <elmo40> Jymmm: prob in china...
[14:50:03] <cradek> that's going to be 3 phase with R8 like pretty much all of them are
[14:50:44] <elmo40> like all the 'german' Hofmann tire changers... all made in China
[14:51:10] <elmo40> cradek: there, my point exactly. 'pretty much all of them' != all of them
[14:51:22] <elmo40> our shop has the ones with cat-30
[14:52:06] <elmo40> annoying sounding things, though. they don't use a tang on the end, just the regular threaded rod into the tool :P Silly, actually.
[15:14:18] <elmo40> http://www.axyz.com/index.html My future competitors ;)
[15:18:08] <JT-Work> do you use CSS when parting off
[15:21:10] <elmo40> cascading style sheet? :-/
[15:21:23] <JT-Work> nope
[15:21:24] <elmo40> constant spindle speed?
[15:21:25] <elmo40> ;)
[15:21:28] <JT-Work> yep
[15:21:32] <JT-Work> nope
[15:21:40] <JT-Work> constant surface speed
[15:21:44] <elmo40> you should... you want to keep the same chip load all the time, for any tool
[15:21:52] <elmo40> ok, surface speed... sorry
[15:22:02] <elmo40> that is what I was thinking, just didn't type it.
[15:22:04] <elmo40> :P
[15:22:39] <elmo40> * elmo40 goes to take a nap... brain fart occurring too often now
[15:43:10] <elmo40> does this still work? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?GcodeGenerator
[15:43:34] <elmo40> I tried installing it, put it in the directory as listed but when I restarted Blender I couldn't find it
[15:48:45] <elmo40> my little home mill has 80 oz-in motors with a HobbyCNC board. You think a 230 oz-in motor would still be drivable by that board? Says it handle 3A per phase. No clue what the 230 motors specs are.
[15:49:56] <elmo40> the hobbyCNC site offers 400+ oz in motors.
[15:50:09] <elmo40> would they sell motors their boards can't handle?
[15:53:21] <Jymmm> elmo40: caveat emptor
[16:02:45] <elmo40> wow, firefox is consuming 995Mb :/
[16:03:37] <elmo40> glad I have 4Gb :P
[16:08:15] <JT-Work> any ideas one where to start for speeds and feeds for parting off 6061?
[16:10:24] <SWPadnos> four thousand quatloos
[16:10:53] <JT-Work> ok
[16:11:20] <SWPadnos> JT-Work, same as milling I'd imagine - a couple hundred SFM for HSS, plus more if you have coolant, coatings, or a different tool material
[16:13:07] <JT-Work> on my old manual lathe I would usually go real slow to keep it from chattering... so I don't know where to start with the Hardinge
[16:13:41] <cradek> for parting try 80
[16:13:49] <cradek> use fpr+css and you'll be surprised how well it works
[16:14:00] <JT-Work> ok, thanks
[16:14:15] <Jymmm> FPR == Feet Per Revolution?
[16:14:21] <JT-Work> I have a new parting tool due here tomorrow :)
[16:15:23] <Jymmm> * Jymmm hands JT-Work a rusty old hacksaw blade
[16:15:52] <elmo40> JT-Work: your hardinge will be MUCH more stable then the manual lathe. I bet you could easily double it.
[16:19:34] <JT-Work> what depth of cut works well? for G95?
[16:19:49] <sealive> Hi do i loose always the ini if i change things at stepconf?
[16:20:06] <JT-Work> yes for ini
[16:20:14] <elmo40> loose? no. you can make a new one or save over the old one
[16:20:19] <sealive> if i press edit nothiing i usr mousepad and in the ini is gedit
[16:25:09] <sealive> the new free heekscad cam G-code generation is very good
[16:25:41] <elmo40> it is default to mm, drives me nuts :P
[16:26:48] <elmo40> ok, price check... 3x 230oz-in NEMA-23 steppers with a HobbyCNC board. Any ideas?
[16:27:17] <sealive> 259Eur just bouth one
[16:27:42] <elmo40> 3 motors and board?
[16:27:44] <sealive> with PC-CNC optokoppler
[16:27:48] <elmo40> 260 E ?
[16:27:50] <sealive> jes
[16:27:50] <elmo40> expensive!!
[16:28:02] <elmo40> oh, with pc and opto coupler?
[16:28:06] <elmo40> what is an optocoupler?
[16:28:12] <sealive> tourch is 1,2NM at 2A 24V
[16:29:11] <sealive> HPCL2630
[16:29:26] <sealive> both direktions
[16:30:54] <sealive> 8Databits and 4Switches and all inputs iof Sub-D25 are in full galvanic devided
[16:31:15] <elmo40> any pictures?
[16:32:04] <sealive> http://www.mechapro.de/catalog/images/hsopto_komplett_frei.jpg
[16:33:16] <elmo40> well, this guy is asking $140 for 3x steppers and one hobycnc board.
[16:33:19] <sealive> Driver->http://www.mechapro.de/catalog/images/step3d_komplett_27.jpg
[16:33:36] <elmo40> board is like $60. steppers, depending on where, range from $30 to $80 each
[16:34:02] <elmo40> where do you plug the motors?
[16:35:50] <sealive> right hand to the cooler it fits the screwdriver even to zero
[16:36:09] <sealive> Full and Half step
[16:36:20] <sealive> as the L297/298 does
[16:36:22] <elmo40> next to the blue?
[16:36:41] <elmo40> what are the LED's for. trouble shooting?
[16:36:50] <sealive> jes there are 15 plugs
[16:37:22] <sealive> power enable the Axis limit home
[16:40:33] <sealive> also the 4th axis is on the separate plug
[16:42:39] <elmo40> I made an offer of $120 for the kit.
[16:42:51] <elmo40> he was going to CNC a Sieg X3
[16:42:54] <elmo40> but stopped
[16:45:29] <sealive> the layout boards are available without parts for 35Dollars
[16:46:38] <elmo40> your boards?
[16:46:45] <elmo40> the hobbyCNC Pro board is $80
[16:47:00] <sealive> wher in china
[16:47:13] <elmo40> hobbycnc.com
[16:47:36] <elmo40> http://www.hobbycnc.com/products/hobbycnc-pro-chopper-driver-board-kits/
[16:48:01] <elmo40> 3-axis, though.
[16:48:07] <elmo40> $100 for 4-axis
[16:49:22] <sealive> is this finisht or to buld kit
[16:49:35] <Jymmm> Probably asking too much... http://sfbay.craigslist.org/sby/tls/1667950794.html
[16:51:51] <elmo40> I think you have to solder everything together.
[16:53:25] <Jymmm> kit
[16:53:37] <Jymmm> "Kits requires PCB soldering skills"
[16:53:55] <Jymmm> Unipolar, not Bipolar too.
[16:54:24] <sealive> ok the shipping to germany will sertenly be not available
[16:55:40] <elmo40> Jymmm: I don't know. looks about right to me.
[16:58:28] <Jymmm> Please, take it from someone who bought a xylotex to save the money from getting geckos... BUY THE GECKOS http://www.geckodrive.com/product.aspx?c=3&i=14469 It already comes with opto isolated breakout board built in, as well as VFD, chargepump, Mid-band resonance compensation, 50VDC, and a hell of a idiot warranty!
[16:58:55] <elmo40> http://www.k2cnc.com/2009/09_Software_CAD_Cam_Rhino_MasterCam.asp EMC is on the list :)
[16:59:17] <elmo40> though they recommend mach3. I find mach3 to be very 'junior' level. controls are NOT what I have on a real CNC machine
[16:59:41] <sealive> the cooler V6716Z is missing
[17:00:19] <elmo40> Jymmm: wow, $300 for that board.
[17:00:25] <elmo40> looks very professional, though!
[17:00:59] <Jymmm> elmo40: Marriss is very smart and friendly guy to talk to
[17:01:22] <Jymmm> elmo40: I waasn't kidding about the idiot warranty either.
[17:02:32] <Jymmm> elmo40: If you "fess up" that you screwed up and miswired or something along those lines and blew up your gecko, He will one-time replace it free of charge. but you have to be honest about it.
[17:03:02] <elmo40> nice guy! I like that.
[17:03:21] <Jymmm> elmo40: Plus the Mid-band resonance compensation alone will be a big frustration saver
[17:03:57] <Jymmm> It helps prevent the steppers from stalling == ruined work.
[17:04:04] <elmo40> I wonder why they mounted the ball screws in this fashion http://www.k2cnc.com/shop/proddetail.asp?prod=KG-1414-G&cat=26
[17:04:07] <elmo40> side ways
[17:04:14] <elmo40> puts all the load on the top set.
[17:04:23] <elmo40> err, linear bearings...
[17:04:27] <elmo40> not ball screws, sorry
[17:04:47] <elmo40> for the gantry.
[17:05:01] <Jymmm> elmo40: There is an underplate that you can't see
[17:05:09] <elmo40> I understand for the Y+Z.
[17:05:11] <Jymmm> I have one.
[17:05:13] <elmo40> underplate? for what
[17:05:20] <elmo40> oh, I see that now
[17:05:22] <Jymmm> And I've had issues with it for years.
[17:05:39] <elmo40> but that still puts most of the load on the top set of rollers
[17:07:09] <Jymmm> Keep swarf off?
[17:07:16] <elmo40> they could have put them in the proper direction and still have that bar underneath
[17:07:24] <elmo40> nope, swarf will still stick.
[17:08:45] <elmo40> Jymmm: don't think $120 for those two steppers and hobbycnc board is a good idea?
[17:10:26] <elmo40> * 3 steppers
[17:11:45] <sealive> it will simply work
[17:11:58] <Jymmm> elmo40: I know that seems like a "good deal" and that $300 seems like a LOT of money, but you'll save yourself *SO* much headaches... Be it wiring because you don't need a breakout board too, everything is self contained just add PS, higher voltages/amps, and you can use it years down the road on other projects.
[17:12:41] <celeron55> the hobbycnc board supposedly doesn't need a break-out board either
[17:12:50] <celeron55> at least that's what they say on the page
[17:13:25] <Jymmm> no mid band compensation
[17:13:29] <celeron55> it does look a bit cheap though
[17:13:37] <elmo40> celeron55: ya, it does look cheap
[17:13:43] <elmo40> I have the 'econo' version
[17:13:55] <elmo40> doesn't look very much different then their pro version
[17:14:00] <celeron55> if was building a serious mill i'd get the gecko
[17:14:08] <Jymmm> it's also UNIPOLAR, not BiPolar
[17:14:16] <celeron55> +I
[17:14:26] <elmo40> the only difference is the PRO has idle current reduction, whatever that is
[17:14:52] <sealive> no step no power
[17:14:54] <elmo40> who uses bipolar?
[17:14:56] <Jymmm> elmo40: When your motors arent spinning, it reduces the holding current
[17:15:12] <elmo40> why would I want to reduce the holding power?
[17:15:18] <celeron55> less heat
[17:15:20] <Jymmm> heat
[17:15:26] <elmo40> or am I looking at this a little differently
[17:15:46] <Jymmm> elmo40: Do you rev your car at 4000RPM when parked?
[17:15:53] <elmo40> will it know that if I travel in X the Y also gets affected by tool load?
[17:16:33] <elmo40> I don't understand. will the holding power be reduced as well?
[17:16:50] <Jymmm> yes, but you won't care.
[17:16:51] <sealive> no only if all are quiet
[17:17:02] <elmo40> >_<
[17:17:16] <Jymmm> elmo40: you won't care
[17:17:24] <elmo40> I know the current adjusters are silly one-turn styles.
[17:17:26] <JT-Work> holding power is only the measure of how much work the stepper is not doing
[17:17:28] <elmo40> Jymmm: why not?
[17:17:45] <sealive> as i do no care the min power is still more then it needs
[17:17:52] <Jymmm> elmo40: holding isn't to prevent the machine from moving, it's to maintin current position. not a big deal
[17:18:22] <sealive> the pitch is also doing its best
[17:18:43] <elmo40> I guess the screws hold 99% of the position, in nature of screw threads.
[17:18:55] <sealive> never lost a single step by reducing the power at quiet axis!
[17:19:25] <elmo40> 1/10 ACME threads right now. going to ball screws for the next table-top machine.
[17:19:28] <LawrenceG> current foldback is evil..... if you are cutting a circle slowly, when you go through each of the quadrants, the un moving axis gets pushed out of position...
[17:19:48] <elmo40> LawrenceG: and you have seen this?
[17:20:01] <LawrenceG> mostly a problem when using ball screws (easy to back drive)
[17:20:35] <JT-Work> isn't there some kind of delay before this kicks in like 5 minutes or something
[17:20:37] <LawrenceG> yes... it happened to me.... took me a while to figure out why I could not cut a circle
[17:21:32] <elmo40> very odd
[17:21:35] <elmo40> and lame
[17:21:58] <elmo40> why would that be put in there? steppers run HOT, that is their design.
[17:22:11] <LawrenceG> my foldback was only like 1 second.... if it was in the minute range it would probably be ok... but steppers are designed for continuous current and they dont use much power when stopped (using a pwm driver)
[17:22:32] <archivist> a cold stepper is not driven properly
[17:22:53] <elmo40> cold stepper is turned off :(
[17:22:57] <Jymmm> elmo40: Also, gecko says "18VDC to 50VDC supply voltage" excluding BEMF, HobbyCNC says "42VDC maximum input voltage" including BEMF. (If I'm reading them correctly)
[17:23:13] <elmo40> bemf?
[17:23:40] <archivist> elmo40, a machine tool does not turn off the steppers
[17:23:42] <Jymmm> elmo40: BEMF = Back voltage produced from motors when they slow down
[17:24:04] <elmo40> Jymmm: and how do you calculate that back voltage?
[17:24:23] <elmo40> archivist: not supposed to. but this sounds like it does (momentarily)
[17:24:30] <Jymmm> elmo40: http://www.geckodrive.com/upload/Step_motor_basics.pdf
[17:24:42] <elmo40> ah, read a little more ;) Idle Current Reduction to 50% when idle for 10 seconds
[17:25:01] <elmo40> so, if no activity in 10 seconds it idles down
[17:25:49] <sealive> thats nice for correcting the programm on emc2
[17:25:51] <JT-Work> you would have to be cutting a circle REAL SLOW for that to kick in
[17:26:35] <sealive> ej by the way can i start a programm from wherever i'am in the programm on EMC
[17:26:58] <sealive> as i mean at N450 for example
[17:27:21] <Jymmm> elmo40: Go download the manual from HobbyCNC
[17:27:41] <archivist> its dumb for X to go into idle just because you had a long cut on Y
[17:28:23] <celeron55> maybe it goes to idle only when none of the axes are moving?
[17:28:50] <celeron55> that would be reasonable
[17:29:05] <Jymmm> Go download the manual and see what it says
[17:29:21] <sealive> i got to go
[17:29:25] <sealive> BB and G8
[17:29:45] <archivist> celeron55, the driver has no clue what the other axes are doing
[17:34:56] <elmo40> Jymmm: was that a joke? go download their manual... I don't see it anywhere
[17:35:42] <Jymmm> elmo40: Well, the joke is that they don't offer a manual to download.
[17:36:09] <Jymmm> elmo40: How the hell are you know what you're buying if you can't even read about it.
[17:36:43] <JT-Work> they have a manual to modify a pc power supply
[17:37:13] <elmo40> lame
[17:37:15] <cal> cal is now known as Guest98852
[17:37:59] <Jymmm> elmo40: Can't tell from the low-res photos, but if it uses Allegro chips, they are prone to thermal issues.
[17:38:29] <elmo40> they edited out the numbers in the pics
[17:38:36] <elmo40> it is a shifty site
[17:39:17] <JT-Work> http://www.geckodrive.com/product.aspx?c=3&i=14471
[17:39:19] <Jymmm> "The board is based around the SLA7062M Stepper Driver Chip from Allegro Microsystems."
[17:39:26] <Jymmm> http://www.mibot.com/hobbycnc/driver.php
[17:40:23] <Jymmm> Nice... NOT! Product support is provided via the HobbyCNC Yahoo Groups discussion forum. This forum is only open to people who have purchased HobbyCNC products.
[17:40:59] <JT-Work> Jymmm: if that ain't a big red flag I don't know what one is
[17:42:03] <elmo40> wow. what a sham indeed
[17:42:08] <elmo40> well, he IS from Ohio...
[17:42:09] <elmo40> :P
[17:42:13] <Jymmm> JT-Work: no doubt.
[17:42:50] <Jymmm> elmo40: you'll NEVER see cra like that from Mariss. Hell you could call and ask him a question right now and he'll be happy to answer.
[17:43:09] <Jymmm> elmo40: But he'll talk your ear off too =)
[17:43:10] <elmo40> americans... think what they slap together is the cream of the crop and they 'own' it. such typical mentality
[17:43:22] <JT-Work> yep take your first $69 and buy a gecko
[17:43:59] <elmo40> I have to say, though, my current machine only works in X and Y under EMC2.
[17:44:03] <elmo40> Z doesn't move
[17:44:05] <JT-Work> I killed one of my 203v drives and Mariss replaced it no problem
[17:44:14] <elmo40> I tried editing the setup, but no go
[17:44:21] <elmo40> under Mach3 everything runs.
[17:44:26] <Jymmm> JT-Work: did you "fess up" ?
[17:44:51] <JT-Work> I told him I didn't know what killed it and I don't really know
[17:45:04] <JT-Work> elmo40: is your step and direction pins mixed up?
[17:45:22] <Jymmm> JT-Work: Yep, that's Mariss for ya... dont bullshit him and he'll bend over backwards for ya.
[17:48:17] <elmo40> nice kit. thick mounts and good looking ball screws http://www.cncfusion.com/images/X3mounts/ABBA%20ballscrew%20kit/3-14-08/IMG_1102.JPG
[17:48:32] <elmo40> JT-Work: I tried everything
[17:48:46] <elmo40> the motor seems like it is 'on'
[17:49:03] <elmo40> maybe it is a current issue. I could try playing with the 1-turn current pot
[17:49:15] <archivist> trace your step signal
[18:02:00] <JT-Work> elmo40: swap drives and or motors one at a time
[18:02:33] <archivist> if the drive works from mach3 then its a configuration error
[18:05:51] <elmo40> I don't know what else to configure
[18:06:10] <elmo40> I can't find the setup from mach3...
[18:12:25] <JT-Work> will your Z drive move the X or Y axis?
[18:12:52] <JT-Work> I have had a wire break before and other odd things like a drive that quit
[18:13:55] <elmo40> JT-Work: I have to head to work soon. will test when I get home. sorry, no time atm.
[18:14:02] <JT-Work> ok
[18:14:31] <elmo40> damn work, getting in the way of life. Rather annoying how that is
[18:16:11] <L84Supper> has anyone seen an off the shelf non-contact solution to power a few mA circuit in a jaw of a chuck that will spin at up to 8K rpm?
[18:16:46] <archivist> yes but a long time ago
[18:18:07] <celeron55> getting a few mA's with induction isn't too much of a problem i guess... but off-the-self is 8)
[18:18:34] <L84Supper> yeah, I don't see anything I can just order
[18:18:50] <archivist> iirc there was a chip/design idea in one of the trade mags like new electronics
[18:19:05] <elmo40> why induce anything? put a magnet on the end, have a coil on the chuck pass the magnet. (epoxy the coil to keep it there). it will make a few mA no problem. add magnets for more mA.
[18:19:46] <archivist> but but, the magnet does the induction :)
[18:19:47] <L84Supper> yes, the magnet and coil solution or lamp + solar cell will be cheap and simple
[18:19:48] <elmo40> use a diode and such to clean it up
[18:20:48] <celeron55> elmo40: isn't what you just described exactly what induction is?
[18:21:50] <celeron55> of course to get more current with a little space you could use a high-frequency coil to transmit the power
[18:22:04] <celeron55> but that isn't really needed i guess
[18:22:07] <elmo40> how about RFID technology?
[18:22:19] <elmo40> send a radio wave to it
[18:22:21] <celeron55> i guess they use more like a few nA's
[18:22:33] <celeron55> or uA's
[18:23:07] <L84Supper> we're reading a strain gauge to measure the force of the chuck while it's spinning
[18:24:38] <L84Supper> I guess they are seeing too many flying objects per year. millions of parts per year with a few ppm that get loose
[18:25:23] <archivist> use a pendulum chuck that increases the grip with rpm
[18:26:03] <L84Supper> that would seem obvious but they want to have a strain guage :)
[18:30:14] <archivist> L84Supper, iirc analog devices had a chip for the return signal via non contact from strain gauge
[18:30:38] <archivist> I bet the app note will show power supply too
[18:36:15] <LawrenceG> Jymmm, http://imagebin.ca/view/R0I0E13.html talk about a bad day at the office
[18:37:53] <Jymmm> LawrenceG: No, THIS is a bad day at the office... http://boldt.us/8548-3/capsized_cargo_ship
[18:38:52] <Jymmm> It looks like some of the containers crushed like tin cans in your pic
[18:39:12] <Jymmm> Can you say "GVWR"
[18:42:23] <LawrenceG> yea.. the dock cranes would be pretty useless.. the lifting rigs assume the container is close to horizontal.... each container would need to be manually rigged.. nice job climbing over those piles
[18:44:59] <L84Supper> If only they had used an enormous bungee wrapped on that last stack for safety before they left port
[18:51:08] <Administrator_> wow yea looks like they lost a few on the right
[18:51:30] <Administrator_> hope they had insurance!
[18:52:10] <Administrator_> there has to be some serious weight there to crush those containers like that
[18:52:27] <Administrator_> Administrator_ is now known as aa-danimal-shop
[18:52:49] <aa-danimal-shop> dammit thats getting annoying
[18:56:09] <aa-danimal-shop> why would i be getting unexpected realtime delay errors?
[18:56:28] <pcw_home> L84Supper a rotary transformer might be a good solution, Ive seen it done with a pot core...
[18:58:56] <aa-danimal-shop> hmm maybe it's the video card, it had a big black bar across the screen
[19:00:42] <aa-danimal-shop> so noone knows what can cause an unexpected realtime delay error?
[19:06:16] <micges> archivist: hello
[19:06:46] <pcw_home> video is a one of the usual suspects...
[19:06:56] <micges> hello
[19:07:11] <pcw_home> Hi micges
[19:08:05] <micges> someone asked me if writing component that will be drive hi current to stepper driver when moving and driving lo current when it stops will be working?
[19:08:11] <frallzor> me ordered a usb>rs485 today
[19:08:18] <aa-danimal-shop> pcw_home, thanks, i got another card, i can try swapping it out. i did notice it was a bit loose in the agp slot though.
[19:08:48] <aa-danimal-shop> frallzor, for the modbus thing you were asking about yesterday?
[19:08:51] <micges> iirc this will caouse some steps loosing at begining of move
[19:08:52] <frallzor> si
[19:08:57] <frallzor> to try it at my main pc
[19:09:05] <frallzor> same shit just different workplaces to sit at =)
[19:10:33] <aa-danimal-shop> frallzor, let me know how it works. I'm assuming that would save some outputs on my mesa cards
[19:12:36] <frallzor> seems pretty straight forward
[19:12:46] <frallzor> just need to figure out the greek language
[19:13:46] <aa-danimal-shop> ehh, i cant figure out anything but portuguese
[19:14:54] <frallzor> did you see my nice little Z-test video?!
[19:15:24] <aa-danimal-shop> no i dont think so
[19:15:30] <aa-danimal-shop> linkie?
[19:15:45] <frallzor> 3933.026-1
[19:15:48] <frallzor> baah
[19:15:49] <frallzor> I hate mirc
[19:15:58] <frallzor> http://vimeo.com/10530991 thar we go
[19:16:06] <frallzor> rattle = not being tightened =)
[19:18:38] <aa-danimal-shop> nice!
[19:19:48] <pcw_home> miges. I think its possible to make this work, but how do you tell when EMC is idle?
[19:20:07] <frallzor> aa-danimal-shop did a y test too today
[19:20:18] <frallzor> 16000mm/min seems pretty doable
[19:20:26] <frallzor> at least dry running it
[19:21:22] <aa-danimal-shop> nice!
[19:21:27] <aa-danimal-shop> steppers or servos?
[19:21:59] <aa-danimal-shop> gotta be servos
[19:27:06] <frallzor> steppers
[19:27:13] <frallzor> and rack and pinion
[19:27:24] <frallzor> its like 7-8mm/rev
[19:27:48] <frallzor> X might be another deal, it got the heavy loads
[19:28:26] <aa-danimal-shop> pretty fast for steppers
[19:33:15] <frallzor> http://vimeo.com/10560773 weeey
[19:35:26] <frallzor> wish I didnt test until all done, tht rattle is bugging me :P
[19:41:30] <Eric_K> I can't hear a rattle
[19:41:54] <frallzor> when the car is close to the camera
[19:41:56] <frallzor> that rattle =)
[19:42:27] <frallzor> but there is shitloads of stuff that can cause it, I bet its the 10000000 untigthened screws though :P
[19:43:25] <frallzor> might be the rack and pinion too, first run together
[19:44:40] <Eric_K> yeah, could be that
[19:44:57] <frallzor> but all in all, its all goood
[19:45:14] <Eric_K> nice
[19:45:16] <tom3p> idle = a sort of charge pump run off the sum of all 'steps' or non-zero analog commands, if integration of this pump drops lowe enuf, ... thats idle
[19:45:18] <Eric_K> is that z axis done?
[19:47:17] <frallzor> pretty much, not really assembled as it should e
[19:47:18] <frallzor> be
[20:02:55] <archivist> micges, yes powering down can miss steps, when the power comes back on there is no guarantee it settles in the same place
[20:03:48] <micges> archivist: that I was think too
[20:04:46] <archivist> micges, and microsteps complicate matters too
[20:04:57] <pcw_home> If you just lower current when idle (no mechanical load) I think it will work
[20:08:13] <pcw_home> But if you have any static load, when you reduce current you will get unwanted motion
[20:08:40] <archivist> the motor will just balance its magnetic forces at the lower current, I would expect a move with no load
[20:10:07] <pcw_home> Well maybe a tiny move based on the ration of cogging torque to current generated torque
[20:10:17] <pcw_home> (ratio)
[20:12:50] <micges> thanks, I'll tell him all that cases
[20:24:52] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[20:35:48] <MarkusBec> MarkusBec is now known as MarkusBec_away
[20:49:09] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[20:49:36] <frallzor> btw, do I have to configure something to use mini?
[20:49:50] <frallzor> tried the onscreen jogging without result
[21:07:58] <MattyMatt> "matt is not in the sudoers file. This incident will be reported." debian ain't ubuntu :)
[22:14:52] <MarkusBec> MarkusBec is now known as MarkusBec_away
[22:36:35] <skunkworks> random picture... http://www.electronicsam.com/images/KandT/conversion/morewiremess.jpg
[22:47:29] <andypugh> Nice retrofit target: http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=260576164044
[22:58:14] <morfic> i see a lot of manual to cnc conversions using EMC2, or old nc to cnc conversion, it appears most use drives/servos to do the job, is anyone using EMC2 on machines formerly operated by Fanuc (18-T, 18-M, 0-T) controllers, driving the fanuc servos?
[23:02:12] <andypugh> I think a Mazak machining centre was done that way. Near the bottom of http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Case_Studies
[23:02:51] <skunkworks> what kind of servos are those? brushed? brushless?
[23:03:03] <skunkworks> * skunkworks doesn't know his fanuc models
[23:03:48] <tlab> anyone use pcb-gcode?
[23:11:04] <morfic> skunkworks: i admit, none of the companies with their windows based solutions asked that, so i did not check, i can always find out
[23:12:35] <morfic> skunkworks: the reason i look at emc2 is that my boss wanted an open solution (while he likes open source, he more meant, buy a package and do not come back paying for options we need later), running on std pc hardware
[23:14:51] <morfic> the further we get with the quoting process the more we buy hardware from them, the more we pay for options to do stuff we want to do, the further basically from simple pc hardware that makes us independent from one source, and paying for option over option is not much different than paying fanuc
[23:15:30] <morfic> linux geek/machinist i am, i would like to figure out if emc2 is worth a consideration
[23:16:22] <skunkworks> it sure is :)
[23:17:21] <skunkworks> there are a bunch of inexpensive interface cards that allow you to run closed loop with emc2 closing the pid loop.
[23:18:08] <andypugh> If EMC doesn't do what you want, you can make it do what you want. Either by writing the code you want, or getting someone else to write it. But it seems to do most stuff already.
[23:19:27] <andypugh> Moving the PID into the software gives you a lot more control, and you can watch the values and tune them easily. And you can see what is going on rather more than you can with the PID running in opaque firmware.
[23:21:26] <morfic> which is kind of what i thought my boss wanted, use what's there, make what is not, but i guess our design guy who does the quoting is more comfortable with a windows based solution, and before a decision is made, i don't want to ignore EMC2, and i was told on the way out i should get in contact with one of the windows solutions, which means i just gave myself permission to add 2, not just one if i do it on my time
[23:23:00] <pfred1> morfic Windows is great its a path to endless support and bugs!
[23:25:44] <andypugh> The windows solutions clearly work. I have no idea what they are like to work with, but I have found EMC2 to be fairly straightforward even as a total linux noob.
[23:25:48] <morfic> i guess i made mistake to stop looking at EMC2, when i told him about EMC2 early on and he was not going for it, "i am not the one gathering the info, so i won't" and now it's "oh can you contact them, i have too much on my plate", so i screwed myself out of a lot of time. cause i am lacking details to go and ask you guys the right questions, not annoying "dude that is there on the wiki" stuff
[23:26:50] <skunkworks> heh
[23:26:51] <pfred1> morfic documentation is the achillies heal of FOSS
[23:26:54] <morfic> andypugh: i had a fadal with windows embedded that we ended up drilling a large hole into, from the front cover through the way cover.....so we could et a socket on the ballscrew to unjam it, because it went haywire in the middle of a proven old program that ran fine on the old fadal it replaced......
[23:26:56] <pfred1> heel even
[23:27:33] <morfic> pfred1: which makes presenting it as an option a little problematic, i search forums, google, and love freenode, seems weird to my old boss maybe though
[23:27:43] <L84Supper> I met with a large automation and tool company earlier today, 100+ years in business, $16B in sales last year
[23:28:12] <L84Supper> they had never heard of EMC, they were using Windows until about 4 hours ago
[23:28:13] <pfred1> morfic EMC2 is better than most but still not extremely polished all around
[23:28:34] <morfic> our main thing is, he wants to do more automation, so we need more I/O than before for part changers, logic to do things based on a probe measuring the part
[23:29:14] <morfic> pfred1: it looks good enough (site) to make me think it's worth a shit, the fact the channel is active is a very good sign to me
[23:29:17] <morfic> shot*
[23:29:43] <morfic> L84Supper: and now they use what?
[23:30:01] <pfred1> man I need google for my bookmarks!
[23:30:01] <andypugh> <digression> Any tips for getting little bits of metal out of your fingers?
[23:30:35] <morfic> caliper
[23:30:37] <L84Supper> I introduced them to EMC2 and they jumped on it
[23:31:04] <skunkworks> morfic: emc has an embedded ladder logic program called 'classic ladder' also - a mesa pci card will give you 72i/o (and you can run more than one)
[23:31:10] <andypugh> My calipers are all blunt from being used to scribe marks :-) (heck, they are cheap enough nowadays)
[23:31:26] <L84Supper> Windows + machine control is just asking for trouble
[23:31:35] <skunkworks> andypugh: I do the exact same thing...
[23:31:55] <pfred1> morfic getting started and the integrator manual off this page are nice: http://linuxcnc.org/docs/2.3/
[23:32:39] <morfic> skunkworks: it's what i am getting at, the companies that quote us have fancy operator panels, and "yes it's all std parts you find in any PC", sorry to say, none of my PCs has ever had IO cards in it, so i am not at all familiar with how easy it is to source io boards, drives for the servos and all that jazz
[23:32:51] <pfred1> L84Supper I had to look at some easm files today so ran a viewer in wine even that left me with 2 zombies on my system I felt just like bad old times here
[23:33:02] <L84Supper> they were all trying out the latency test on the embedded PC's in the lab
[23:33:53] <pfred1> about all I'd do with Windows is play Solitare or Minesweeper
[23:34:26] <skunkworks> morfic: then the wiki is your friend.. ;) like http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Supported_Hardware
[23:34:29] <morfic> latencies is what has me wondering the most, how well the ubuntu lts with rt kernel does compared to windows with whatever proprietary realtime extension they wrote, and in the end, it's all hardware driven
[23:34:30] <L84Supper> we only use Win7 for gaming and we rebuild that box every week
[23:34:54] <morfic> skunkworks: i saw the list, i asked the designer to ask camsoft about those
[23:35:17] <pfred1> pfft I've run Linux for over 14 years now through machine meltdowns and everything only time I ever upgrade is when some distros just stopped updates
[23:36:31] <andypugh> morfic: Have a look at http://www.mesanet.com for good (and inexpensive) IO boards and http://pico-systems.com/motion.html for some more specific problem-solving interfaces (including resolver to quadrature and Fanuc encoder convertor amongst others
[23:36:33] <pfred1> last Linux distro I had going was Suse 8.2 and that lasted me about 6 years or so
[23:36:38] <skunkworks> morfic: that is the fundimental difference between something like camsoft and emc. emc uses dumb/inexpensive interface cards that really just do encoder counting and pwm/+/-10v output.. while camsoft has an embedded motion control in hardware.
[23:37:04] <skunkworks> emc wants to do the motion control.
[23:37:41] <morfic> L84Supper: i used vista just fine, until i had issues with a firewire audio card on it, and pretty much told me to go eff myself when i tried to install drivers newer from intel, that day vista died and i came back to my senses, unfortunaly, ubuntu has died since, they decided to no longer support armv5, so i lose the "oh yeah i run ubuntu at home" bit ;P
[23:37:58] <morfic> andypugh: thanks!
[23:38:48] <pfred1> morfic hardware support is not really Linux's responsibility
[23:39:04] <morfic> skunkworks: when my boss asks if emc is capable to do the same work w/o the hardware, how do i say "sure" in more words?
[23:39:13] <morfic> pfred1: how do you mean?
[23:39:21] <L84Supper> ubuntu is moving to all the new ARM cortex stuff
[23:39:29] <pfred1> morfic you think Microsoft is writing hardware drivers?
[23:40:18] <L84Supper> I designing a dual ARM core Marvell board currently, quad core ARM's are due later this year
[23:40:23] <morfic> L84Supper: i would not be so mad if distupgrade had a tiny check for /proc/cpuinfo and not just go do its mery thing and unpack armv6 bins on my armv5 box ;P
[23:41:01] <pfred1> morfic https://help.ubuntu.com/community/PinningHowto
[23:41:05] <L84Supper> well to me Ubuntu has always been a broken version of debian
[23:41:36] <pfred1> L84Supper as if Debian isn't broken all on its own
[23:41:43] <L84Supper> :)
[23:42:03] <morfic> pfred1: no, they don't write them, i guess i am wondering in reply to what i said you made the comment, i was commenting on ubuntu dropping support for my cpu's instruction set, was it that? :)
[23:42:13] <pfred1> last time i tried Debian I ran a stable and because of its package manager my sound didn't work it installed conflicting packages
[23:43:14] <aa-danimal-shop> is there any way to have a program in que so that when it finishes one program, it goes straight to the next?
[23:43:20] <pfred1> morfic I don't know all about it but imagine it might just have been bad juju to keep it going
[23:43:29] <L84Supper> yeah, the Debian installer could use some work, last time we downloaded it they left out support for our network adapters
[23:43:38] <pfred1> aa-danimal-shop &&
[23:43:53] <aa-danimal-shop> ?
[23:44:07] <pfred1> well in shell that works or is that at the same time I forget
[23:44:23] <aa-danimal-shop> hmmi'll look into it
[23:44:31] <morfic> pfred1: widely used cpu, dropped, was bad juju to distupgrade, that's sure ;P
[23:44:32] <pfred1> its either that or just a single ampersand
[23:45:02] <pfred1> try this date && ls in your terminal
[23:45:35] <pfred1> or just date & ls see which you like better
[23:46:21] <pfred1> then there's the whole &2>&1 deal
[23:48:38] <pfred1> you can also chain commands with a semicolon if memory serves date;ls
[23:48:58] <jackc> or && if you want it conditional on the succes of the previoud
[23:49:11] <pfred1> oh is that what the two ampersands mean?
[23:49:12] <jackc> oh, sry
[23:49:14] <jackc> yes
[23:49:17] <pfred1> heh
[23:49:22] <jackc> if the previous fails, the second doesnt fo
[23:49:23] <jackc> &go
[23:49:27] <jackc> whereas ; means "do both"
[23:49:34] <pfred1> just one?
[23:49:44] <jackc> like chmod 777 file && echo hi >> file
[23:49:47] <pfred1> well there is a single ampersand in there too
[23:49:50] <morfic> pfred1: pinning wouldn't work if all the packages come in for a cpu instruction set newer than the one the hardware supports, thanks though maybe deb source the newer stuff would have worked, at which point i'm rather on a system which is meant to be compiled, but i am majorly getting side tracked here :)
[23:50:06] <jackc> pfred1: iirc, the single runs both and returns the & of both exit values to the shell
[23:50:16] <andypugh> Time to sleep. Good luck with your enquiries morfic. I think it is safe to say that the EMC2 approach will be cheaper and more flexible than the alternatives, at the expense of a steeper learning curve.
[23:50:18] <pfred1> ah so a single is a fork
[23:50:23] <morfic> did i mention i was told to gather info on Soft Servo today, and the next meeting is tomorrow, w/o access to company email from home :P
[23:50:48] <pfred1> and a double is a conditional
[23:50:48] <morfic> andypugh: thanks a lot, yes the curve is higher, i still think it may just be worth it
[23:51:14] <andypugh> I reckon so. But then we have all already made that choice, and are bound to think it was the right one.
[23:51:18] <pfred1> and a semicolon is just do it who cares
[23:51:25] <morfic> && needs previous to finish make && make modules_install, you can use ; if you do not care if success or not of previous commant
[23:51:37] <morfic> * morfic loves &&
[23:51:57] <morfic> finish successfully*
[23:52:07] <pfred1> morfic even if Windows was absolutely perfect I still much prefer the Linux environment
[23:52:08] <morfic> i hate when i think the word, but fingers don't type it :/
[23:52:27] <morfic> pfred1: i am sold, i am looking at strong points to pimp to my boss now
[23:52:52] <pfred1> the strongest point is if yo ureally really want to do something there is nothing in Linux stopping you from doing it
[23:53:07] <pfred1> sit there go through the code and make it happen!
[23:53:29] <pfred1> and that ain't never gonna be on Windows
[23:54:34] <pfred1> billy G ain't never going to invite you over to lunch then hand you a disk with the windows sourcecode on it on your way out
[23:54:41] <morfic> it's waht i am scared of, emc2 may not be perfect, we might something noone else was bothered by, but there is a chance it's fixable and we can fix it, if the windows systems do not work (cause they all "sell the best") as we want, much more limited choice of solutions
[23:55:16] <morfic> we might hit something noone else was bothered by*
[23:55:47] <morfic> see what i mean, totally "nottyped" the word "hit"
[23:55:49] <pfred1> well then you're doing something awefully strange and i doubt it works on any platform
[23:57:50] <pfred1> to me Microsoft's dominance does not extend past the desktop
[23:57:50] <MattyCNC> bleh heekscad.exe runs fine in windows
[23:58:15] <MattyCNC> yeah it's a workstation os
[23:59:17] <pfred1> this is my favorite: http://www.top500.org/stats/list/34/os
[23:59:45] <morfic> we do have a linux server, boss attempted a linux desktop, use could not adjust to gnome and paths though, so the idea died out again, so EMC2 may have a chance