#emc | Logs for 2005-06-12

Back
[00:00:04] <anonimasu> it's tens at most
[00:00:11] <anonimasu> I
[00:00:28] <anonimasu> I'll still need drive electronics.. but the less complex it is the better
[00:00:42] <robin_sz> I made some flow sensors hot wires once .. like I said
[00:00:48] <robin_sz> (just for interest this ...)
[00:00:55] <anonimasu> * anonimasu nods
[00:00:59] <robin_sz> used a bridge,
[00:01:09] <robin_sz> resistors and lamp bulbs
[00:01:18] <robin_sz> broke the glass on one bulb
[00:01:23] <robin_sz> as the sensor
[00:01:43] <robin_sz> (miniture 12V bulbs)
[00:02:28] <robin_sz> and then used an op-amp so sense the bridge and drive it into balance, the more current, the hotter the wire, the greater the resistance
[00:02:53] <anonimasu> nice
[00:02:59] <robin_sz> in still air, they would detect if you just moved near them
[00:03:30] <robin_sz> after about 5kmh it was very linear
[00:03:36] <anonimasu> yeah :)
[00:03:44] <robin_sz> I calibrated them to about 100kmh
[00:03:52] <anonimasu> nice
[00:04:05] <robin_sz> used them for paragliding ...
[00:04:15] <anonimasu> I am going to have the prototype running at monday..
[00:04:16] <robin_sz> always handy to know the exact wind speed on the top of the hill
[00:04:49] <anonimasu> :)
[00:04:56] <robin_sz> especially when it is very close to the forward speed of you glider :)
[00:05:01] <anonimasu> yeah hehe
[00:05:04] <robin_sz> going backwards is not much fun
[00:06:09] <anonimasu> ideally I'd have a small tube and a \^ to generate a vortex for a piezo material..
[00:06:20] <anonimasu> but piezo material is hard to come by at sundays ;)
[00:06:25] <robin_sz> yeah
[00:06:32] <anonimasu> or well in sweden with <1day leadtime
[00:06:51] <robin_sz> I can imagine
[00:07:06] <anonimasu> they will plant seeds for 3 weeks more then the season is over
[00:07:08] <anonimasu> :)
[00:07:33] <robin_sz> so you feed it with low pressure air, and if the backpressure goes over a certain amount, you have an alarm?
[00:07:42] <anonimasu> yeah
[00:07:50] <robin_sz> * robin_sz nods
[00:08:01] <robin_sz> I used that to detect level of paint in a pot once
[00:08:23] <robin_sz> when the paint covered the end of the tube, the pot was full
[00:08:27] <dave-e> ray... hmmm your isp must be slow...still no files. :-(
[00:08:32] <weyland> weyland is now known as weyland_away
[00:08:46] <anonimasu> :)
[00:08:52] <robin_sz> it was pretty reliable ...
[00:09:04] <anonimasu> yeah
[00:09:05] <robin_sz> the paint was at 100KV
[00:09:12] <anonimasu> KV?
[00:09:17] <robin_sz> it was the only way 1000 volts
[00:09:25] <robin_sz> 100, 000 volts
[00:09:27] <anonimasu> powdercoat?
[00:09:30] <robin_sz> nope
[00:09:37] <robin_sz> water based
[00:09:49] <anonimasu> yeah but why KV?
[00:10:09] <robin_sz> paint gun
[00:10:14] <robin_sz> spray right?
[00:10:20] <anonimasu> yeah
[00:10:25] <robin_sz> put 100kv on the gun
[00:10:30] <robin_sz> pait becomes charged
[00:10:39] <robin_sz> sticks to the object you are painting
[00:10:39] <anonimasu> ah :)
[00:10:40] <anonimasu> very cool
[00:10:49] <robin_sz> heres a thing ...
[00:10:57] <robin_sz> the guy who invented it,
[00:11:13] <robin_sz> how much do you think he originally charged for the equipemtn to do this?
[00:11:21] <robin_sz> (he had a patent too)
[00:11:27] <anonimasu> 500$
[00:11:31] <robin_sz> answer: nothing
[00:11:46] <robin_sz> he simply said,
[00:11:53] <robin_sz> I will supply the equipment
[00:11:55] <robin_sz> for free
[00:12:07] <robin_sz> you pay me half of whatever it saves you in paint costs ...
[00:12:16] <anonimasu> lol
[00:12:24] <robin_sz> he made millions
[00:12:53] <anonimasu> * anonimasu does not like patents
[00:13:06] <robin_sz> they have plus and minus points
[00:13:07] <jacky^> :)
[00:13:21] <robin_sz> that to my mind was a good patent
[00:13:28] <jacky^> night ..
[00:13:28] <anonimasu> yeah
[00:13:37] <robin_sz> night jacky
[00:13:45] <jacky^> night
[00:13:47] <robin_sz> a real invention
[00:14:44] <anonimasu> yeah
[00:14:46] <anonimasu> agreed
[00:15:09] <anonimasu> good & bad
[00:15:14] <robin_sz> * robin_sz nods
[00:15:29] <robin_sz> too many obvious things get aptented
[00:15:40] <anonimasu> but unless you have enough money to fight people over your patent they suck.
[00:15:49] <robin_sz> right
[00:16:14] <anonimasu> ie you cant fight microsoft over a patent you've invented/owns
[00:16:23] <robin_sz> yeah
[00:16:27] <robin_sz> well, you can
[00:16:30] <anonimasu> s/owns/own
[00:16:34] <robin_sz> but you need lots of $$$ to do it
[00:16:42] <anonimasu> yeah, but your company will go out of buisness before you get a ruling
[00:16:48] <robin_sz> yeah
[00:16:50] <anonimasu> with 70% chance..
[00:17:08] <robin_sz> its funny watching M$
[00:17:17] <robin_sz> a few weeks ago they loved patents ...
[00:17:27] <anonimasu> well, they could just throw $ at you until you just nod..
[00:17:30] <robin_sz> now longhorn is stuffed byt them
[00:17:31] <anonimasu> everything is for sale.
[00:17:36] <robin_sz> and they now hate patents
[00:17:53] <anonimasu> :)
[00:17:55] <robin_sz> not *everyting*
[00:18:05] <robin_sz> ooh, joke
[00:18:08] <anonimasu> almost, it just depends on the ammount of money
[00:18:08] <robin_sz> right ...
[00:18:28] <robin_sz> this guy goes up to a very well dressed woman ..
[00:18:49] <robin_sz> "madam, for $1,000,000 .. would yo usleep with me?"
[00:19:01] <robin_sz> "why, yes .. for $1,000,000"
[00:19:11] <robin_sz> "ok, well, how about for $5?"
[00:19:24] <robin_sz> <slap> .. "what sort of woman do you think I am?"
[00:19:35] <anonimasu> * anonimasu nods
[00:19:46] <robin_sz> "well, we already established that, now we are just haggling on price" ;)
[00:20:04] <anonimasu> lol
[00:20:59] <anonimasu> hell if microsoft wants a patent they might just buy the company that owns it..
[00:21:05] <anonimasu> just like that..
[00:21:08] <robin_sz> well, yeah
[00:21:24] <robin_sz> makes you wonder if the FOSS community shold patent stuff
[00:21:28] <anonimasu> might be easier then negotiating a deal..
[00:21:31] <anonimasu> FOSS?
[00:21:37] <anonimasu> open source?
[00:21:42] <robin_sz> free and open source software
[00:21:47] <anonimasu> hm, yeah
[00:21:58] <anonimasu> the trouble is that it's costly to apply for patenst
[00:22:01] <anonimasu> patents
[00:22:02] <robin_sz> yep
[00:22:05] <fenn> good luck getting $15k for your little project
[00:22:18] <anonimasu> hm, that's not much
[00:22:19] <robin_sz> better to just keep going and let the code be free
[00:22:45] <fenn> make sure it's well distributed.. then it's "prior art"
[00:22:50] <anonimasu> yeah after you've released it it cant be patented though
[00:22:58] <robin_sz> well, deopends
[00:23:04] <robin_sz> in europe, no
[00:23:07] <robin_sz> in the US , yes
[00:23:18] <anonimasu> unless it's a "mod patent"
[00:23:25] <anonimasu> or well, a circumvented patent..
[00:23:31] <anonimasu> yeah
[00:23:33] <anonimasu> true
[00:23:45] <robin_sz> in the US, you can patent all sorts of stuff that is public kowledge
[00:23:56] <robin_sz> just so long as it has not been patented before
[00:24:07] <robin_sz> and not been published in some journals
[00:24:20] <robin_sz> (there is a list of what journals count somewhere_
[00:24:22] <robin_sz> )
[00:24:32] <anonimasu> yeah but you cant patent others stuff can you?=
[00:24:38] <robin_sz> others?
[00:24:44] <anonimasu> yeah
[00:24:47] <anonimasu> things other invented..
[00:24:47] <robin_sz> yes
[00:24:51] <anonimasu> others
[00:25:26] <robin_sz> it seems in many cases, that whoever patents it first is the "owner"
[00:25:34] <robin_sz> not who wrote about it first
[00:25:43] <robin_sz> (this is the US system)
[00:25:46] <anonimasu> I guess I should stop ranting about stuff/prototypes/ideas :)
[00:26:02] <anonimasu> *grins*
[00:26:16] <robin_sz> not really
[00:26:20] <anonimasu> not that I care for any patents..
[00:26:22] <fenn> you ain't gonna patent sheeit
[00:26:28] <robin_sz> most products have a 2yr lifel cycle
[00:26:53] <robin_sz> better jsut getting it to market quick and riding the wave
[00:27:11] <anonimasu> fenn: not anything I am speaking about here.. for sure :)
[00:27:26] <robin_sz> right, bedtime
[00:27:37] <anonimasu> robin_sz: most cases yeah
[00:27:45] <robin_sz> remind me to make a picture of the sign after I mount it on the wall tomorow
[00:27:54] <anonimasu> yeah do that :)
[00:28:04] <anonimasu> I'll be heading to bed also
[00:28:21] <robin_sz> right,
[00:28:23] <robin_sz> later
[00:28:40] <anonimasu> fenn: why not?
[00:29:16] <fenn> once you patent something it becomes a huge hassle
[00:29:32] <fenn> you're no longer doing what you want to do.. instead you've become a lawyer
[00:30:01] <fenn> it takes a lot of time.. and you don't have that much time
[00:30:14] <anonimasu> fenn: that's why you have lawyers. :)
[00:30:28] <fenn> even with a lawyer though
[00:30:35] <anonimasu> yeah
[00:30:42] <fenn> dont take my word for it.. ask les
[00:31:12] <anonimasu> * anonimasu nods
[00:31:34] <anonimasu> the trouble with the buisness I am in is that there are loads of copycats ;)
[00:32:09] <fenn> really? forest equipment controls?
[00:32:22] <anonimasu> hm, the whole buisness
[00:32:29] <anonimasu> forest/machine buisness
[00:33:29] <anonimasu> the competitor came and videotaped the prototype when we ran it.
[00:33:40] <fenn> that's pretty low
[00:33:50] <anonimasu> that's how bad it is..
[00:33:58] <anonimasu> they wanted to know more about the huydralics ;)
[00:34:02] <anonimasu> and about the control.
[00:34:14] <anonimasu> the guy running it wouldnt tell them.
[00:34:26] <fenn> should've made up some fake info
[00:34:37] <anonimasu> oh, there's nothing special with the plc..
[00:34:59] <anonimasu> just the code inside of it :)
[00:35:47] <anonimasu> I love the valves we use for the huydralics..
[00:36:31] <anonimasu> some italian valves, flow compensated and proportional
[00:36:52] <fenn> what's "flow compensated" maen
[00:37:52] <anonimasu> do you have any idea how a constant pressure system works?
[00:38:05] <anonimasu> you have a LS line going back to the pump, changing the angle of it..
[00:38:48] <anonimasu> to be able to set the pressure of one function, without affecting the other ones you need a flow compensated valve..
[00:39:07] <fenn> like a pressure regulator?
[00:39:39] <fenn> hmm... i'll figure this out some other time i think
[00:39:57] <fenn> brain's got a clog in it today
[00:40:12] <anonimasu> very clever valve's :9
[00:40:13] <anonimasu> :)
[00:40:37] <anonimasu> I cant remember how it's done.. but it's all hardware
[00:41:18] <anonimasu> it also allows you to have shock valves that's way lower then the system pressure
[00:41:51] <anonimasu> :)
[00:42:36] <anonimasu> * anonimasu yawns
[00:42:38] <anonimasu> good night
[00:42:41] <fenn> night
[03:28:29] <LawrenceG> anybody on tonight that has used gplot? I'm using bdi-4.2 and gplot seems to be looking for librcs.so which doesnt seem to exist on the bdi stuff
[03:30:45] <fenn> gplot as in Axis frontend?
[03:32:14] <fenn> sorry, i guess not
[03:32:19] <LawrenceG> yes... it looks like it is supposed to be a standalone gcode viewer... I am working on some dxf to gcode coverter utilities and I would like to find a way to plot the gcode without having to run the realtime kernel
[03:32:58] <fenn> most of the axis stuff uses "ldd" to figure out what linked libraries to use
[03:33:32] <LawrenceG> lg@speedy3:~/Desktop $ gplot
[03:33:33] <LawrenceG> Traceback (most recent call last):
[03:33:33] <LawrenceG> File "/usr/bin/gplot", line 42, in ?
[03:33:33] <LawrenceG> import emc
[03:33:33] <LawrenceG> ImportError: librcs.so: cannot open shared object file: No such file or directory
[03:33:33] <LawrenceG> lg@speedy3:~/Desktop $ locate librcs.so
[03:33:57] <LawrenceG> nobody named librcs.so on my box...
[03:34:15] <fenn> try using the emc2 patched version
[03:34:15] <LawrenceG> I think that gets built with the emc1 stufff
[03:35:41] <fenn> i still haven't gotten axis to compile right.. it fails looking for openGL
[03:35:46] <LawrenceG> I may give that a try.... download a fresh copy of axis, patch and attempt to build.... I dont have emc2 sources onboard... only the bdi stuff which seems to be half way between emc 1 and 2..
[03:36:57] <fenn> although.. if you don't have the emc2 sources it will probably fail when it tries to "import emc2"
[03:37:12] <LawrenceG> have you a copy of librcs.so on a pentium box?
[03:37:24] <fenn> no i haven't compiled emc1
[03:37:54] <LawrenceG> anybody?
[03:38:13] <LawrenceG> kind of quiet tonight
[03:38:24] <fenn> it comes and goes
[03:39:05] <fenn> there are probably other gcode viewers out there too btw
[03:39:58] <LawrenceG> I tried a few windows ones running under wine and did not have any success... the ones that survived an install, died when trying to run. any suggestions?
[03:41:12] <LawrenceG> I have emc bdi running on a laptop.... I may just beam the g code files over there and look at them with the old viewer... the laptop is too braindead to run the axis stuff.
[03:42:05] <fenn> i found it annoying to have to actually run a program just to look at the output
[03:42:25] <LawrenceG> me to
[03:43:27] <fenn> i would say.. get the emc2 cvs, and axis-1.0, and the patch to make them work together
[03:43:35] <fenn> then you're in the same boat as me :)
[03:44:48] <LawrenceG> .. I have tried to build the axis stuff several times without success... It would really be helpful if there was an rpm or deb that had all the pieces.
[03:45:38] <fenn> i wrote a wiki page about compiling axis... it shows some of the packages you might need: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CompilingEMC2
[03:46:31] <fenn> ugh usemod always mangles the formatting of everything i write
[03:48:39] <LawrenceG> how close to compiling axis have you come?
[03:48:54] <fenn> dunno.. it's not like it gives you a progress meter or anything
[03:49:02] <fenn> fix one error, another pops up
[03:49:17] <LawrenceG> right.... been there
[03:49:26] <LawrenceG> do you have skype?
[03:49:32] <fenn> no.. what's that?
[03:49:50] <fenn> a VoIP
[03:49:54] <LawrenceG> its a voice over ip prog that works pretty well for windows or linux
[03:50:11] <LawrenceG> I have run it on both os's
[03:50:35] <LawrenceG> good for chats when working on stuff as its full duplex
[03:51:05] <LawrenceG> http://www.skype.com/
[03:51:38] <LawrenceG> it works best with a headset with boom mic
[03:51:43] <fenn> hmm.. talking out loud.. sounds difficult
[03:51:52] <fenn> i'd have to rewire my brain to make my thoughts come out of my mouth
[03:52:20] <LawrenceG> there is a certain amount of heavy breathing that gets sent around the world...
[03:53:02] <LawrenceG> where are you? I am off the west coast of Canada
[03:53:10] <fenn> indiana.. midwest usa
[03:53:47] <LawrenceG> it has been a little wet out that way... any flodding in your area
[03:54:35] <fenn> only in my bedroom
[03:54:44] <LawrenceG> k
[03:56:09] <LawrenceG> have you used qcad? it works with dxf files in and out.... 2 1/2D only
[03:57:08] <LawrenceG> I am trying to get from drawings to gcode to chips all on my linux box....
[03:58:05] <LawrenceG> http://www.ribbonsoft.com/qcad.html
[04:01:53] <fenn_> well, that's interesting
[04:02:03] <fenn_> skype locks up hard.. no ssh or anything
[04:02:33] <LawrenceG> strange.... what os you using?
[04:02:39] <fenn_> maybe it didn't like the RTAI kernel
[04:02:50] <fenn_> fedora core 3
[04:03:04] <LawrenceG> that could be... I have tried it with the rt kerenl running
[04:03:34] <LawrenceG> I am using ubuntu, with kernel 2.6.10 something
[04:04:43] <fenn_> yes, lots of people seem to be using ubuntu these days
[04:05:31] <LawrenceG> It works great, but you really do need the box to have high speed net access as it is gear to do most updates/installs off the net
[04:11:40] <dave-e> anyone using floppies under bdi-4.2?
[04:12:21] <fenn_> floppies? what's that?
[04:12:30] <dave-e> no joke
[04:12:44] <fenn_> mount /dev/fd0 /floppy ?
[04:12:57] <dave-e> nope...it complains
[04:13:01] <fenn_> what's it say
[04:13:20] <dave-e> not a device or something like that
[04:13:26] <fenn_> hrm
[04:13:45] <jmk_dinner> I think I used a floppy once on 4.2... lemme try a couple things
[04:14:03] <fenn_> ls -l /dev/fd0 see what it says
[04:15:05] <fenn_> nick fenn
[04:15:07] <fenn_> erk
[04:15:10] <fenn_> fenn_ is now known as fenn
[04:15:26] <jmk_dinner> jmk_dinner is now known as jmkasunich
[04:16:16] <dave-e> brw-rw-... root fd0
[04:16:24] <jmkasunich> "mount: /dev/fd0 is not a valid block device"
[04:16:44] <dave-e> so what does it expect?
[04:16:58] <jmkasunich> dunno... it's in /etc/fstab
[04:17:06] <dave-e> no I checked
[04:17:38] <dave-e> I may have to go back and rewrite to the cd
[04:17:49] <jmkasunich> from /etc/fstab:
[04:18:05] <jmkasunich> /dev/fd0 /floppy msdos,vfat noauto,users 0 0
[04:18:24] <dave-e> hmmmm don't think mine says that
[04:18:26] <dave-e> brb
[04:18:41] <jmkasunich> even tho it is in fstab, the mount still fails
[04:18:57] <fenn> well.. mount /dev/null will fail too
[04:19:01] <fenn> even if it's in fstab
[04:19:20] <jmkasunich> wiseguy
[04:19:30] <jmkasunich> the fact that it is in fstab means that "somebody
[04:19:33] <jmkasunich> oops
[04:19:43] <jmkasunich> "somebody" thought it should work
[04:20:23] <dave-e> my fstab has hdd as cdrom but no fd0
[04:21:08] <jmkasunich> mine has fdc as cdrom... I think that was figured out at install time to suit the actual hardware
[04:21:13] <dave-e> how is the flashdrive usb support ?
[04:21:19] <jmkasunich> sorry. hdc, not fdc
[04:21:39] <jmkasunich> no idea, never tried any usb stuff
[04:22:02] <dave-e> I keep looking at the memory sticks as a way to transfer files.
[04:22:21] <jmkasunich> yeah... floppies on steriods
[04:22:33] <fenn> i never had any luck with floppies even back in the eighties
[04:22:39] <dave-e> btw...I have emc2 up on the machine I'm bringing to fest for a servo test machine
[04:22:47] <LawrenceG> hey... I use my digital camera to transfer files.... 128mb card
[04:23:09] <fenn> i use a PCMCIA hard disk :)
[04:23:10] <jmkasunich> dave_e - ray was telling me about that... good to hear it
[04:23:30] <dave-e> Hi Lawrence...and they are cheap too ...even the Gb ones are not too bad
[04:23:36] <jmkasunich> dave_e - you already know I'm gonna be there, right?
[04:23:48] <jmkasunich> looking forward to working on the mazak
[04:23:50] <dave-e> Yep...that is really good news
[04:24:19] <dave-e> It is different than mine ... 6m rather than 7m controller
[04:25:22] <dave-e> cpu cabinet is much different
[04:25:35] <LawrenceG> jmk - do you have librcs.so from a pentium box you could send me? I am trying the get the axis gplot to run, but I dont have emc1 compiled for my box.
[04:25:50] <jmkasunich> I don't have emc1 compiled either
[04:26:04] <LawrenceG> rats
[04:26:24] <dave-e> the best laid plans of mice and men.....
[04:27:22] <fenn> how do i know what opengl libraries my system is using?
[04:28:07] <fenn> erf.. nevermind
[04:28:07] <dave-e> try doing a locate on opengl and see if you get a string of file names that make sense
[04:28:17] <fenn> xorg-x11-Mesa
[04:28:50] <dave-e> anyone have a command line for cdrecord that will do individual files...
[04:29:12] <dave-e> I know what to do for an iso but have not gotten beyond that
[04:29:49] <jmkasunich> dave: your box has no networking?
[04:30:25] <dave-e> no...the dhcp works fine on the macs on the network but linux seems to have to be hand configured
[04:30:45] <fenn> xinetd dhcpcd start
[04:31:13] <fenn> or pick an ip number at random and use that one
[04:31:28] <fenn> ifconfig eth0 up 192.168.0.*
[04:32:32] <dave-e> I plugged it in and tried (post install) and it tried to discover the gateway but no luck
[04:32:55] <fenn> it's probably 192.168.0.1
[04:33:03] <fenn> route add default gw 192.168.0.1
[04:33:15] <jmkasunich> can you ping the gateway?
[04:33:37] <dave-e> can't ping anything but localhost
[04:33:51] <jmkasunich> not even by IP number?
[04:34:43] <dave-e> brb
[04:37:24] <weyland> who was it that was asking about vmware and solidworks?
[04:37:36] <fenn> i'm somewhat interested
[04:37:57] <fenn> did you get it to work?
[04:38:11] <Phydbleep> dave-e: try '/sbin/ifconfig' and look for an 'eth0' line.
[04:38:24] <weyland> well, I'll letcha know when I'm done setting up vmware :) working on it right now
[04:38:35] <weyland> using the trial
[04:38:39] <fenn> cool
[04:40:19] <Phydbleep> * Phydbleep has had eth0 vanish mysteriously before..
[04:45:25] <dave-e> OK...I have a basic problem or two. router is 192.168.1.1 .... and I can ping that from this machine
[04:45:41] <dave-e> however sbin is not in my search path...
[04:45:56] <jmkasunich> it usually isn't for normal users
[04:46:09] <fenn> su - will include /sbin in PATH
[04:46:31] <dave-e> ifconfig reports etho as up ...displays MAC address but no IP
[04:46:46] <jmkasunich> no ip, but you can ping the router?
[04:46:49] <jmkasunich> that's odd
[04:47:14] <fenn> i'd say take eth0 down and bring it back up with the right IP address
[04:47:23] <dave-e> jmk...I can ping from this machine...not the machine
[04:47:31] <jmkasunich> oh
[04:47:35] <dave-e> 4.2 machine
[04:48:00] <jmkasunich> take eth0 down: /sbin/ifconfig eth0 down
[04:48:11] <dave-e> brb
[04:48:17] <jmkasunich> pick an IP for the linux box (one that your other machines won't use)
[04:48:26] <jmkasunich> like 192.161.1.200
[04:48:57] <fenn> 192.168.1.200
[04:48:59] <jmkasunich> correction
[04:49:11] <jmkasunich> yea, not 161 (typo)
[04:49:37] <jmkasunich> bring it back up: /sbin/ifconfig eth0 up 192.168.1.200
[04:49:44] <jmkasunich> try pinging the router and see if it works
[04:49:52] <jmkasunich> if not, you probably need to add routes
[04:52:31] <fenn> arg it appears redhat is totally busted regarding xorg-x11-devel and libMesa
[04:53:33] <dave-e> ping 192.168.1.1
[04:53:52] <dave-e> opps...wrong screen
[04:58:53] <dave-e> ok...lets get my story straight...
[04:59:11] <dave-e> can't even ping 127.0.0.1 !
[04:59:43] <dave-e> route shows 192.168.1.1 as default gw
[04:59:49] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: Broke the original multi-page halcmd help screen up into multiple sections, each less than one screen long. Added a help command to call up individual help screens.
[05:00:35] <fenn> gee I thought the original help screen was nice
[05:00:50] <jmkasunich> not on a 25 lines screen it wasn't
[05:01:12] <jmkasunich> do a cvs up and tell me what you think of the new one
[05:01:19] <jmkasunich> ones
[05:02:03] <fenn> erf i need to figure out how to do that with the developer cvs
[05:02:16] <jmkasunich> you are listed as a developer, right?
[05:02:49] <fenn> yeah
[05:02:57] <jmkasunich> cvs -z3 -d:ext:<yourname>@cvs.sourceforge.net:/cvsroot/emc co emc2
[05:03:29] <jmkasunich> must do export "CVS_RSH=ssh" first
[05:03:52] <fenn> add that to .bashrc?
[05:03:55] <jmkasunich> you'll want to put that export in your bashrc or some other place where it runs on login
[05:04:05] <jmkasunich> ;-)
[05:05:26] <dave-e> attempts to ping get 'Destination Host Unreachable'
[05:06:08] <jmkasunich> even "ping localhost" ?
[05:06:25] <dave-e> hosts has localhost declared
[05:06:34] <dave-e> yep
[05:06:47] <dave-e> either as 127.0.0.1 or localhost
[05:07:03] <dave-e> something is pretty well broken
[05:07:28] <jmkasunich> does "/sbin/ifconfig" show an "eth0" section and an "lo" secion? or "lo" only, or nothing?
[05:09:16] <dave-e> eth0, lo, sit0
[05:09:30] <jmkasunich> what is sit0?
[05:09:39] <dave-e> don't know
[05:09:51] <dave-e> let me see if I can ping it.
[05:09:57] <CIA-12> 03jmkasunich * 10emc2/src/hal/components/stepgen.c: removed some unused parameters from stepgen component
[05:10:01] <fenn> heh
[05:10:04] <fenn> $ ./halcmd help
[05:10:04] <fenn> RTAPI: ERROR: could not open shared memory
[05:10:04] <fenn> HAL: ERROR: rtapi init failed
[05:10:04] <fenn> halcmd: hal_init() failed
[05:10:04] <fenn> NOTE: 'rtapi' kernel module must be loaded
[05:10:21] <asdfqwega> about memory sticks: I use 'em all the time. AND it was nice of Paul to put the supermount-ng patch into the kernel on the BDI - just edit /etc/fstab, and it's ready
[05:10:21] <jmkasunich> do sudo scripts/realtime start
[05:10:48] <fenn> why is "help" linked to the rtapi module?
[05:11:26] <jmkasunich> help is just a command... before it executes any commands, halcmd needs to connect to the shared memory
[05:11:55] <jmkasunich> (if reading from a file, help could be the first command, or the last, or anywhere... )
[05:12:11] <jmkasunich> halcmd connects to shmem once at startup, then runs all commands, then disconnects
[05:12:55] <fenn> i think i screwed up installing rtai somehow.. it cant find any of the rtai modules
[05:13:00] <fenn> it worked last time
[05:13:19] <jmkasunich> last time? you reinstalled rtai?
[05:13:28] <fenn> no, before i rebooted
[05:13:46] <jmkasunich> that shouldn't keep it from finding modules
[05:14:07] <jmkasunich> you have a fresh checkout, right... and you did ./configure and make for that checkout?
[05:14:23] <fenn> modprobe rtai_up
[05:14:23] <fenn> FATAL: Module rtai_up not found.
[05:14:43] <jmkasunich> the up thing....
[05:15:07] <dave-e> see ya later...
[05:15:17] <dave-e> thanks for the suggestions
[05:15:26] <jmkasunich> there is some strangeness because rtai-ksched (or maybe it's rtai-sched) is a symlink to either rtai-up or rtai-smp
[05:15:32] <jmkasunich> bye dave
[05:16:23] <jmkasunich> that modprobe... was that part of the "realtime" script, or are you loading the modules manually?
[05:17:31] <fenn> manually
[05:17:42] <fenn> um, both? i forget
[05:17:54] <jmkasunich> rmmod any rtai modules that got loaded, then use the realtime script
[05:18:09] <jmkasunich> it has code that deals with the symlink and other strangeness
[05:19:52] <fenn> okay
[05:19:56] <fenn> it worked :)
[05:20:47] <fenn> i get this error when running scripts/realtime start :
[05:20:57] <fenn> insmod: can't read 'adeos': No such file or directory
[05:21:19] <jmkasunich> hmmm... this isn't a BDI is it?
[05:21:26] <fenn> heh no this is my fc3 box
[05:21:45] <jmkasunich> but you do have RTAI patches installed, right?
[05:21:49] <fenn> right
[05:22:04] <jmkasunich> and adeos patches?
[05:22:07] <fenn> hal appears to be working
[05:22:21] <fenn> rtai _and_ adeos?
[05:22:40] <jmkasunich> adeos is a low level patch that is used by rtai
[05:23:01] <jmkasunich> you may have it built directly into the kernel, so the realtime script can't find or load the module
[05:23:07] <jmkasunich> because it's already in the kernel
[05:23:14] <jmkasunich> so you can ignore the message if hal is working
[05:23:59] <fenn> okay
[05:24:10] <fenn> i thought adeos was part of rtai, that's why i was confused
[05:24:12] <jmkasunich> the "stack"... hal needs shared memory, which is supplied by rtapi, which runs over either rtlinux or rtai, and rtai runs over adeos
[05:24:57] <jmkasunich> technically adeos is distinct from rtai, but both are installed together so you can usually think of adeos as part of rtai
[05:25:12] <fenn> adeos is the nanokernel, and rtai is the interface between linux and adeos, right?
[05:25:32] <jmkasunich> something like that
[05:25:58] <jmkasunich> adeos mostly serves as an interrupt dispatcher
[05:26:18] <jmkasunich> hardware ints are detected by adeos, which calls either rtai or linux handlers
[05:26:53] <jmkasunich> in some cases, it pipelines them... first it askes rtai if it wants the int, if not then it calls linux to handle it
[05:27:06] <jmkasunich> gets compilcated, and I don't understand most of it
[05:27:41] <fenn> i wouldn't mind a little documentation written in english from the rtai folks :)
[05:27:57] <jmkasunich> don't hold your breath
[05:28:03] <jmkasunich> what do you think of the help stuff
[05:28:45] <jmkasunich> btw, you are the one who posted the halcmd feature requests for shell mode and the hex printout bug, right?
[05:28:51] <fenn> yes
[05:29:19] <jmkasunich> I was talking to ray today, and am planning some new halcmd commands that will make it more "program" friendly
[05:29:29] <jmkasunich> for example:
[05:29:42] <jmkasunich> list <type> where type is pin, param, etc
[05:30:11] <jmkasunich> similar to show, except it will give only the name of the object, all on one line, space delimited
[05:30:33] <jmkasunich> I should say "names of the objects", it's purpose is to determine what names exist
[05:30:54] <fenn> cool
[05:31:12] <jmkasunich> next: getp <name> returns the value of parameter <name>
[05:31:24] <jmkasunich> only the value, and always in decimal, not hex
[05:31:38] <jmkasunich> no type, owner, or other cruft
[05:31:50] <jmkasunich> also gets <name> does the same for signals
[05:32:02] <fenn> on the main help screen, could you add the list of command names only, right underneath "commands:"
[05:32:23] <fenn> loadrt, unloadrt, linkps, linksp etc
[05:32:25] <jmkasunich> I was trying to keep it under 25 lines
[05:33:01] <fenn> ah you've still got 5 extra lines
[05:33:12] <fenn> they dont need to be on their own lines or anything
[05:33:16] <jmkasunich> oh, I see what you mean, not one per line
[05:33:24] <jmkasunich> several per line, with commas
[05:33:27] <fenn> right
[05:33:27] <jmkasunich> I can do that
[05:37:22] <jmkasunich> I was thinking about changing the -Q, -q, -v, -V options as well
[05:37:48] <jmkasunich> the default is -q, so that doesn't really need an option
[05:38:14] <jmkasunich> I was thinking that every -q would make it one step quieter, every -v would make it one step more verbose
[05:38:21] <jmkasunich> and -Q, -V would be eliminated
[05:38:30] <fenn> how about -q2 -q3 etc
[05:38:52] <jmkasunich> blank would be the equivalent of -q
[05:39:05] <jmkasunich> -q would give you -Q
[05:39:10] <jmkasunich> -v would give you -v
[05:39:17] <jmkasunich> -vv would give you -V
[05:39:21] <fenn> dunno why you need -Q since you can just do halcmd &2> /dev/null
[05:39:58] <jmkasunich> perhaps
[05:40:10] <fenn> unless you aren't in a shell
[05:40:37] <fenn> that's getting out of my known territory
[05:40:38] <jmkasunich> it actually uses rtapi message levels to control what gets printed
[05:41:09] <fenn> so no error file descriptor?
[05:41:15] <jmkasunich> for instance, -vv (or -v) doesn't just print messages from halcmd, it also prints messages that are generated in the hal and even rtapi libraries
[05:41:23] <jmkasunich> no, it doesn't use stderr
[05:41:42] <weyland> fenn: it's looking good... farther than ever before... still installing tho...
[05:42:01] <jmkasunich> there are message levels, like DBG, INFO, WARN, ERR
[05:42:07] <fenn> reminds me of trying to install linux the first time
[05:42:24] <jmkasunich> (based on the way the kernel handles log messages)
[05:42:44] <fenn> ahh right
[05:43:11] <jmkasunich> because the rtapi_print_msg(level,message) function can be used from both user and kernel space
[05:43:41] <jmkasunich> so Q, q, v, V (or q, v, vv) simply set the level of messages that get displayed
[05:44:27] <jmkasunich> anyway, back to the new commands
[05:44:28] <fenn> what about just -v [debug, info, warn, err, quiet]
[05:45:02] <fenn> re-use the code for halcmd help
[05:45:16] <jmkasunich> I'm to lazy to a) write the code to parse that, and b) type those words instead of simply v or vv
[05:45:21] <fenn> heh
[05:46:00] <jmkasunich> anyway, I'm planning getp and gets, to solve the hex problem, and also allow you to get the value of a single item rather than the entire list
[05:46:12] <jmkasunich> the list command lets you get all the names
[05:46:21] <jmkasunich> was also thinking of a
[05:46:30] <jmkasunich> "next type name" command
[05:46:47] <jmkasunich> which returns the name of the next item of the specified type
[05:46:58] <jmkasunich> an alternate way to get the entire list, or part of it
[05:47:10] <jmkasunich> next pin <blank> will return the first pin name
[05:47:24] <jmkasunich> then next pin <first name> will return the second, and so on
[05:47:47] <jmkasunich> you could do "next pin stepgen" to get the first stepgen pin
[05:48:05] <jmkasunich> or "next pin s" to get the first one starting with s, and so on
[05:48:09] <fenn> is that how the hal data structure is set up?
[05:48:44] <fenn> like a linked list
[05:48:53] <jmkasunich> it's a sorted linked list
[05:49:22] <jmkasunich> in the refactor I plan to convert it to an AVL tree, so next would be very fast
[05:49:38] <jmkasunich> but even the linear list is plenty fast
[05:50:13] <Phydbleep> * Phydbleep goes "Woohoo!" and dances for joy..
[05:50:29] <jmkasunich> the "list" and "show" commands (and any command that prints data about multiple objects) have a weakness
[05:50:53] <fenn> they must run in one RT cycle
[05:51:14] <jmkasunich> no, halcmd is in user space, and doesn't affect RT at all
[05:51:36] <fenn> :( i thought you said halcmd got loaded into shmem, executed commands, and exits
[05:51:49] <jmkasunich> halcmd accesses shmem
[05:51:56] <jmkasunich> the linked lists and such are in shmem
[05:52:19] <jmkasunich> halcmd manipulates that info (or maybe just reads it, depending on the command)
[05:54:00] <jmkasunich> the actual pin and signal data is also in shared memory, but any halcmd access to that is inherently atomic
[05:54:14] <jmkasunich> manipulation of the list isn't atomic, but is protected by a mutex
[05:54:29] <jmkasunich> RT code accesses data only, so there is no risk of deadlock at the mutex
[05:55:16] <jmkasunich> for instance, newsig adds an entry into the signal linked list
[05:55:39] <jmkasunich> linksp modifies a data pointer, so that subsequent RT accesses will get their data from the linked signal
[05:55:49] <jmkasunich> setp actually modifies a data location
[05:56:41] <fenn> so if you do setp while running there is a chance of a collision
[05:56:46] <fenn> but everything else is fine
[05:56:59] <jmkasunich> everything is fine, period
[05:57:10] <jmkasunich> no hal data item is more than 32 bits long
[05:57:23] <jmkasunich> a setp writes to a parameter, but the write is atomic
[05:57:46] <jmkasunich> one executuon of a realtime thread gets the old value, the next gets the new value
[05:58:16] <jmkasunich> there is no control over exactly when it changes, but the RT code will always get either the old or new value, never some bastard mix of the two
[06:00:19] <weyland> fenn: have results
[06:00:23] <fenn> eh?
[06:00:28] <jmkasunich> damn... 2:00 am _again_...
[06:00:33] <fenn> mwahahaha
[06:00:50] <fenn> you can feel the power of the dark side..
[06:00:55] <jmkasunich> time for sleep... if you have any thoughts on the halcmd changes, either email me, or IRC me tomorrow
[06:01:09] <weyland> it runs, and appears to be nicely functional, but ...
[06:02:11] <weyland> system (linux) takes quite the performance hit while running vmware and the guest OS
[06:02:25] <jmkasunich> oops, almost forgot
[06:02:37] <weyland> gonna try some additional ram to see how it runs
[06:03:06] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: minor tweak to the first help screen - added list of all commands
[06:03:13] <jmkasunich> there... ;-)
[06:03:16] <fenn> thx
[06:03:29] <jmkasunich> goodnight
[06:03:33] <jmkasunich> jmkasunich is now known as jmk_sleep
[06:04:29] <weyland> I'm going 2 bed too
[06:04:42] <weyland> gotta be baq at the shop in two hours
[06:04:47] <weyland> night
[06:07:03] <Phydbleep> I just got given an Apple Pro USB keyboard with 2 dead keys to 'recycle' into a Panel>USB convertor.. :)
[06:07:33] <fenn> cool
[06:07:55] <fenn> will your panel have an LCD?
[06:08:15] <fenn> or some type of display screen
[06:09:13] <Phydbleep> Yeah, I have a Seiko G321D that I am going to tie to a serial port via a pic. :)
[06:12:44] <Phydbleep> http://www.eio.com/g321skel.htm
[06:14:15] <fenn> good luck
[06:14:55] <Phydbleep> What do you mean good luck?
[06:14:56] <fenn> * fenn always screws up stuff like that
[06:15:17] <fenn> where'd you learn electronics?
[06:15:37] <Phydbleep> Well, I do have a slight advantage.. I wrote that page 2 years ago. :)
[06:15:48] <fenn> yeah i just noticed that part :P
[06:16:21] <fenn> is that your scope project?
[06:17:04] <Phydbleep> That was just a general test-bed setup for testing the display/pic/code.
[06:17:44] <Phydbleep> I left enough of the code for someone to make a scope if they wanted.:)
[06:19:06] <fenn> mmm. basic.
[06:19:42] <Phydbleep> Yeah, I've been waiting for the pic-c to mature and get stable.
[06:36:58] <Phydbleep> fenn: Did you read the warning in the code?
[06:42:28] <fenn> yeah, i dont' get i-d-ten-t
[06:43:20] <Phydbleep> ID10T
[06:45:12] <Phydbleep> Hehehe.. It's a local EMS term.. Patient has exhibited terminal stupidity, Darwin Award is pending. :)
[06:46:31] <asdfqwega> Scope project?
[06:47:09] <Phydbleep> Spectrum analyser actually. 950-2150 MHz.
[06:47:49] <fenn> same thing
[06:48:23] <fenn> what's the limit on scopes these days? 100MHz?
[06:48:36] <Phydbleep> 500
[06:49:03] <Phydbleep> My old 7613 was rated for 200MHz.
[06:49:28] <asdfqwega> Heh, I don't know what I'd an o-scope for, but I'd love to have one, just the same
[06:49:46] <fenn> i've got a tektronix 502A if you want it :)
[06:49:46] <asdfqwega> Preferably a digital capture scope
[06:50:07] <Phydbleep> I have a little Philips 20MHz unit I keep around for audio/video/pwm stuff.
[06:51:17] <asdfqwega> An old fart I know always says "A man must have tools [Tim Allen grunting]"
[06:52:02] <fenn> asdfqwega: google sound card oscilloscope linux
[06:54:27] <asdfqwega> fenn: I've already looked at that - but for what I'd want a scope for, I'd prolly need 100MHz
[06:55:32] <asdfqwega> Besides, I could easily cook up a PIC circuit that would out-do a sound-card
[06:56:01] <Phydbleep> Yeah, I did it.. How hard can it actually be? :)
[06:56:05] <asdfqwega> But, that's a project for another day - and I'll just find a scope, instead
[06:59:52] <Phydbleep> asdfqwega: I have a source for Tektronix stuff at reasonable prices here locally.
[07:00:41] <Phydbleep> 7xxx series runs ~$250 with amp/timebase modules.
[07:01:19] <Phydbleep> That's 150 or 200MHz plugins.
[07:01:33] <asdfqwega> Phydbleep: Some other time...need to save some $$$ for Geckodrives
[07:02:12] <fenn> what's the big advantage of microstepping?
[07:02:30] <fenn> why not just bit-bang the parallel port with the step patterns, feed it to mosfets
[07:02:54] <Phydbleep> fenn: You can get really anal about where your shaft position is?
[07:03:12] <fenn> phydbleep!! i'm ashamed!
[07:04:33] <asdfqwega> ...and with that question, /me puts fenn on the ignore list
[07:04:45] <fenn> bah
[07:04:51] <fenn> steppers are for weenies anyway
[07:05:07] <asdfqwega> microstepping doesn't really make for more accuracy, just smoother operation through a wider speed range
[07:06:03] <fenn> wouldnt it be better to gear down the motors and use a really fast full/half-step driver
[07:06:13] <asdfqwega> Hmm. Too bad this thing doesn't have a setting for "Kick-ban with rotting badger"...
[07:06:43] <Phydbleep> * Phydbleep has a rudely shaped turnip and is not afraid to use it..
[07:06:45] <asdfqwega> Gearing down the stepper would still be a good idea
[07:07:11] <Phydbleep> Yeah, I go 10:1 for stepper drives.
[07:08:15] <asdfqwega> but the microstepping then allows the stepper to operate at the higher rpm, smoothly
[07:08:32] <fenn> Phydbleep: what do you use for a driver board?
[07:08:38] <asdfqwega> Talk to robin_sz about using stepper
[07:09:00] <fenn> * fenn wants the cheap answer
[07:10:30] <asdfqwega> I'm going to use Geckos for my mill (need 5A per phase), but for my next small machine, PMDX is looking alot better than the Xylotex boards I've used
[07:11:14] <asdfqwega> PMDX costs more than Xylotex, but has extra little features
[07:15:33] <asdfqwega> Like - intelligent wire connectors, extra protection, on-board test, and means to mount cooling fan
[07:15:34] <Phydbleep> fenn: Whatever I happen to build out of whatever's handy?
[07:16:09] <fenn> on-board test sounds a little too fancy for me
[07:16:22] <asdfqwega> [vikings] Junk junk junk junk junk junk junk glorious junk....
[07:17:26] <Phydbleep> * Phydbleep puts on his horned helmet and starts singing with the Vikings..
[07:42:50] <fenn> $6.10 for an l297??
[08:14:35] <Phydbleep> All right.. That's enough of the bloody Vikings..
[08:15:00] <Phydbleep> * Phydbleep mugs a 5.25" floppy drive for the head positioning stepper..
[08:25:05] <asdfqwega> Phydbleep: want more of those? I have about 7 from 5.25" floppy drives
[08:30:40] <Phydbleep> asdfqwega: I'm going to use this one for the bearings/mag-clutch for my jog wheel.. This one feels like ~56 steps/rev.
[08:35:34] <asdfqwega> Hm...mine are all Nema23, 200 step/rev, high-accuracy microstepping
[08:36:23] <asdfqwega> If I make a jog wheel, I guess I'll use the stepper I pulled from an old hard drive - 400 step/rev, perfect amount of dent
[08:37:09] <asdfqwega> Put a crash sensor on the head of the mill, and make some force-feedback to the jog wheel :P
[08:37:20] <Phydbleep> I need to machine a wheel for this one that weighs about what the stepper does.
[08:38:34] <asdfqwega> Dimensions?
[08:39:10] <Phydbleep> I can get about 4 revs out of it if I tape the leads down, hold the shaft and spin the body.
[08:39:10] <A-L-P-H-A> yikes.
[08:39:18] <A-L-P-H-A> I just got back ... stupid girl.
[08:39:24] <A-L-P-H-A> why does she have to live so far!!!
[08:40:57] <Phydbleep> asdfqwega: It's ~ 1.5" x 1.5" x 1.25"
[08:41:25] <Phydbleep> Looks like a 3/16" shaft.
[08:43:31] <Phydbleep> I figure ~ 3" dia and 1/2" thick plus another 1/2" dia, 1/2" thick collar in the middle for a mount..
[08:45:35] <Phydbleep> And I have a nice 1/2", 1/2 round, 2 flute mill to make the finger hole.
[08:50:40] <A-L-P-H-A> ??
[08:50:43] <A-L-P-H-A> what are we talking about?
[08:51:20] <Phydbleep> * Phydbleep was babbling on about parts for a jog wheel.
[08:52:06] <A-L-P-H-A> oh I see.
[08:52:45] <Phydbleep> And I'm pissed off.. I had the damn jeep lose an alternator bearing, Went and bought a rebuild kit, intalled it, Then I found the box with 4 more bearings just like the one I just bought.. :\
[08:52:54] <A-L-P-H-A> stepped out of my car, and I walked into a sauna that has big blue walls... it was gross.
[08:53:17] <A-L-P-H-A> Phydbleep. heh. couldn't be that expensive.
[08:53:29] <A-L-P-H-A> place take returns?
[08:53:31] <Phydbleep> $6/bearing.
[08:53:42] <Phydbleep> Already installed.:\
[08:53:50] <A-L-P-H-A> bah. pennies.
[08:54:22] <Phydbleep> Who knew that Precor treadmills used Delco alternator bearings to support the belt rollers...
[08:54:37] <fenn> well, duh
[08:55:43] <Phydbleep> I actually got the $13 kit and did the brushes/diodes/regulator while I had it open.
[08:56:40] <A-L-P-H-A> heh
[09:14:06] <Phydbleep> * Phydbleep wanders off to fall over..
[09:14:38] <Phydbleep> G'nite all. :)
[09:48:28] <jacky^> morning
[09:56:07] <robin_sz> meep?
[09:56:55] <jacky^> morning robin_sz
[09:57:01] <robin_sz> goodmorning
[09:57:11] <robin_sz> did you solve your EMC problem yesterday?
[09:58:10] <jacky^> uhm, no.. i'm testing a new cad to generate g-code, but the setup configuration maybe it's wrong
[09:58:34] <robin_sz> which cad?
[09:58:43] <jacky^> onecnc prof
[09:59:01] <robin_sz> actually, if it generates gcode, its cam, not cad .. but thats just a minor point
[09:59:36] <jacky^> yeah, onecnc already have cam extension
[09:59:53] <jacky^> seem pretty nice
[10:01:00] <jacky^> the problem is that i don't know hot to set many parameters in the configuration file for the NC output
[10:01:43] <jacky^> these values are atic aramaic for me :\
[10:01:46] <jacky^> hehe
[10:01:57] <robin_sz> they have a users club
[10:02:07] <jacky^> bah..
[10:02:09] <robin_sz> you could ask in there
[10:02:22] <jacky^> i don't know..
[10:02:35] <jacky^> i need to setup it to work with emc linux
[10:02:46] <jacky^> i think they use windows
[10:02:49] <robin_sz> http://eonecnc.com/6/ubb.x?cdra=Y&s=41260363
[10:02:55] <robin_sz> nothing to do with it ...
[10:03:17] <robin_sz> its just a question of setting up the software to output the gcode we need,
[10:03:32] <jacky^> right, yeah
[10:03:36] <robin_sz> it doesnt matter if that code runs on linux, windows, or a dedicated control
[10:04:16] <jacky^> infact, should be standard..
[10:04:31] <robin_sz> im sure they will have more idea on how to set it up than we will, I guess they have to set up every single installation they sell
[10:04:33] <robin_sz> yeah
[10:04:43] <jacky^> i need to learn more about initial string and other
[10:04:57] <jacky^> how is the weather today there ?
[10:05:01] <jacky^> sun ?
[10:05:04] <robin_sz> england?
[10:05:07] <robin_sz> dull and wet
[10:05:07] <jacky^> yeah
[10:05:23] <jacky^> here we have a strong wind
[10:05:35] <jacky^> but warm and sun too
[10:05:55] <jacky^> wheater is changing
[10:06:50] <jacky^> tv said wew'll have a rain in the next days,
[10:06:59] <jacky^> :\
[10:07:48] <jacky^> i hoped to go to fish in the next weekend
[10:07:52] <jacky^> ghghgh
[10:11:16] <jacky^> * jacky^ take another cup of coffe
[10:13:34] <robin_sz> I guess one thing we should do is create a reference template for emc,
[10:13:50] <ValarQ> robin_sz: reference template?
[10:14:07] <jacky^> hi ValarQ
[10:14:08] <robin_sz> showing the spec for the gcode
[10:14:20] <ValarQ> jacky^: 'lo
[10:14:29] <ValarQ> robin_sz: ah
[10:14:29] <jacky^> robin_sz: good idea :-)
[10:15:01] <ValarQ> robin_sz: i'm making a small crap2g-code filter right now
[10:15:08] <robin_sz> something someone like jacky^ could just give to his software vendor and say "it needs to follow this spec"
[10:15:15] <ValarQ> using the few g-codes i know (G00 - G03)
[10:15:28] <robin_sz> those are the easy ones :)
[10:15:32] <ValarQ> yeah
[10:15:53] <robin_sz> theres the G40/41/42/43 stuff
[10:15:54] <ValarQ> i think i have used G53 and G54 too...
[10:16:13] <robin_sz> yeah, workspaces
[10:16:44] <robin_sz> we could go on
[10:17:01] <robin_sz> I'll put one on the wiki perhaps
[10:17:31] <fenn> how does emc's gcode implementation differ from others?
[10:17:40] <robin_sz> depends
[10:17:46] <fenn> is it a huge difference or a lot of little nit picky things
[10:17:47] <robin_sz> on how differetn the others are :)
[10:17:58] <robin_sz> oh, its always nit picky things
[10:18:13] <robin_sz> there must be .. oh, 2 or 3 hundred different controls
[10:18:21] <robin_sz> each one with *slight* differences
[10:18:38] <fenn> they should just do everything with g1 and put a preprocessor on it
[10:19:03] <jacky^> ValarQ: is your machine ready to work ?
[10:19:04] <robin_sz> g1 is hard to do circles with
[10:19:33] <robin_sz> and its tricky to do tool length offsets with G1
[10:19:37] <fenn> why are circles harder than any other kind of curve?
[10:19:55] <robin_sz> sigh
[10:20:25] <robin_sz> ok, so if you do your circles with G1 you will need man many thousand lines for each circle
[10:20:33] <ValarQ> jacky^: no :(
[10:20:49] <robin_sz> roughly 1000 G1 moves for every second of cutting time
[10:20:49] <fenn> yay! add another thousand lines to the 50 thousand line program
[10:20:50] <ValarQ> jacky^: will be soon hopefully
[10:20:56] <jacky^> ValarQ: working on the HW driver ?
[10:21:10] <ValarQ> jacky^: no, thats finnished
[10:21:11] <robin_sz> and emc will not machine it anyway
[10:21:14] <jacky^> :)
[10:21:21] <ValarQ> jacky^: it's just some mecanical thingies left :)
[10:21:34] <robin_sz> so, anyway, the problems are not with G0, G1, G2 or G3
[10:21:39] <robin_sz> all those are standard
[10:21:50] <jacky^> ValarQ: stepper, 3 axes ?
[10:21:51] <ValarQ> jacky^: i played around with 3 stepper motors yesterday :)
[10:21:57] <jacky^> nice
[10:22:30] <robin_sz> this morning, I am educating my children
[10:22:32] <ValarQ> jacky^: yeah, i tested the backlash parameters in emc2 too
[10:22:52] <ValarQ> jacky^: got surprised it shows in the program
[10:23:01] <jacky^> wow, you switched to emc2 ? cool
[10:23:11] <ValarQ> jacky^: never tested emc1 actually
[10:24:15] <jacky^> ValarQ: how many Ampere the motors you choosed ?
[10:25:43] <jacky^> i'm actually using 2A stepper 200 steps, seem are quite strong
[10:25:43] <ValarQ> i believe these were 2A
[10:25:48] <jacky^> nice
[10:25:54] <ValarQ> 200 steps here too
[10:26:42] <jacky^> i halso have 3 other stepper 4 A .. i didn't used, no controllers schematic found
[10:26:48] <jacky^> too big
[10:26:59] <ValarQ> ok
[10:27:36] <robin_sz> 4A?
[10:27:40] <robin_sz> plenty of controllers
[10:27:44] <jacky^> robin_sz: yeah
[10:27:56] <robin_sz> xylotex, gecko ...
[10:28:42] <robin_sz> or did you want to build your own drive?
[10:29:19] <jacky^> i tried to search for a schematic, but can't find anything
[10:29:31] <robin_sz> I use gecko, I would build my own, but I cant really afford the extra cost
[10:29:39] <jacky^> i would like to build the driver from myself
[10:30:04] <fenn> jacky i've been looking at this all morning
[10:30:16] <fenn> http://www.pminmo.com/l297-8/l297-8.htm
[10:30:22] <robin_sz> I did build some 10A drives once
[10:30:39] <jacky^> i shipped a stock of IC's L298, but this seem are not for the driver i need
[10:30:44] <robin_sz> 297/298 is not enough to drive 4A
[10:30:52] <robin_sz> and its very low vvoltage
[10:30:53] <jacky^> is enough ?
[10:31:11] <robin_sz> not really.
[10:31:16] <fenn> 40V is low voltage?
[10:31:30] <ValarQ> fenn: yes
[10:31:33] <robin_sz> yes
[10:31:34] <jacky^> it seem to me the L298 can't go up 2A
[10:31:38] <fenn> what voltage are you running?
[10:31:52] <robin_sz> and anyway .. run a 298 at 2.5a, 40v and you'll soon have some smoke
[10:32:15] <fenn> well you gotta put a heatsink on it!
[10:32:16] <jacky^> the motor burns
[10:32:32] <robin_sz> yeahm I know you have ot heatsink it
[10:32:47] <robin_sz> but it gets close to the limits of the chip when driven hard
[10:32:56] <robin_sz> and i lost a few that way
[10:33:53] <fenn> that site has lots of other board designs btw
[10:34:06] <jacky^> fenn: nice link :)
[10:34:08] <robin_sz> sounds ok, but too expensove for me
[10:34:57] <fenn> yeah i was shocked at the price of an l297
[10:35:17] <fenn> well.. shocked is a little strong
[10:35:20] <robin_sz> I built some 10A drives once ...
[10:35:23] <robin_sz> http://www.redpoint.org.uk/photos/misc/front.jpg
[10:35:45] <robin_sz> they worked OK, but building them cost more than buying geckos
[10:35:58] <robin_sz> and they didn;t perform nearly as well ...
[10:36:13] <jacky^> nice
[10:36:41] <jacky^> no dissipator's need ?
[10:36:45] <jacky^> O_o
[10:36:49] <robin_sz> yeah, there was a big one
[10:36:54] <jacky^> ah, ok
[10:37:19] <robin_sz> it was a 297 driving some moisfet drivers etc
[10:37:37] <fenn> where'd the money go?
[10:37:50] <robin_sz> by the time I'd blown a few up, etc it cost more than a gecko
[10:37:53] <jacky^> nice, it also is sinngle face pcb :)
[10:38:17] <jacky^> i see few bridge on the upper side
[10:38:21] <robin_sz> yep
[10:39:36] <jacky^> variable resistor on right-bottom side is used to limit the current i suppose..
[10:39:36] <robin_sz> in the end, microstepping is what I wanted
[10:39:44] <robin_sz> yes
[10:39:49] <jacky^> cool
[10:39:56] <fenn> i think i need to give cnc a break for a while to let my brain rest
[10:39:58] <jacky^> seem ok
[10:40:18] <robin_sz> fenn: how long a rest?
[10:40:19] <jacky^> robin_sz: you tested it ?
[10:40:31] <robin_sz> of course,
[10:40:38] <jacky^> :-)
[10:40:44] <robin_sz> I built 4 and ran them on a backplane for my router
[10:40:57] <robin_sz> then I bought geckos ... :)
[10:41:20] <jacky^> :D
[10:41:23] <robin_sz> the microstepping is very VERY good
[10:41:25] <robin_sz> so smooth
[10:41:58] <fenn> geckos are $114 per axis though!
[10:42:15] <robin_sz> I did look at building my own microsteppign drives .. but it was not economic
[10:42:20] <fenn> how the hell did you manage to spend $350 on three little boards like that
[10:42:21] <jacky^> do you know vincenzov italian website ?
[10:42:38] <jacky^> it seem th have something of interegting too..
[10:42:38] <robin_sz> fenn: thats about 70% of what it cost to build my own non-microstepping drives
[10:42:46] <jacky^> http://www.vincenzov.net/eng/design/index.htm
[10:42:48] <fenn> no way
[10:42:50] <asdfqwega> *cough* whine *cough*
[10:43:03] <jacky^> there's also the italian versione :P
[10:43:05] <jacky^> hehe
[10:43:10] <robin_sz> fenn: sorry? how do you mean no way?
[10:43:44] <fenn> i mean, L297=$5, mosfet=$1
[10:43:57] <robin_sz> yes, thats what I thought
[10:43:59] <fenn> that adds up to $400?
[10:44:12] <robin_sz> it cost me 420 GBP for 4 drives
[10:44:16] <robin_sz> I added it up
[10:44:24] <robin_sz> by the time I finished .. thats what it cost
[10:44:29] <asdfqwega> Law of Life and Design: Horsepower = $, lotsa horsepower = $$, lotsa horsepower that doesn't blow up = $$$$
[10:44:34] <robin_sz> and they werent microstepping
[10:45:04] <robin_sz> well, for me, geckos have been better and cheaer, and I just plug them in and they work,
[10:45:29] <robin_sz> thus I can spend the time on other parts of the project, which is enough of a problem anyway
[10:46:02] <robin_sz> no point making more problems than you need
[10:46:24] <asdfqwega> robin_sz: Plus, you get a warranty!
[10:46:31] <robin_sz> true
[10:46:37] <fenn> well, that's somewhat contrary to how i live my life
[10:46:42] <fenn> but i see your point
[10:46:55] <robin_sz> well, yes
[10:47:01] <robin_sz> I like building stuff for myself
[10:47:06] <robin_sz> and .. I did it
[10:47:11] <jacky^> :-)
[10:47:30] <robin_sz> but, I now reallise it was not a good plan :)
[10:47:55] <fenn> btw did you salvage the working components from the boards that blew up?
[10:47:57] <asdfqwega> I've done that too...but after a while, I get tired of trying to re-invent the wheel
[10:48:32] <asdfqwega> Man-hours cost money, too
[10:48:41] <fenn> not if you can't get a job
[10:48:59] <les> morning (for me)
[10:49:05] <robin_sz> morning les
[10:49:11] <jacky^> hi les
[10:49:32] <robin_sz> * robin_sz plays his kids some music to educate them
[10:49:48] <les> big wash out here. tropical storm.
[10:50:00] <robin_sz> good day for sparying huh?
[10:50:03] <robin_sz> spraying
[10:50:14] <asdfqwega> "Hey kids, this is what we old fogeys used to listen to...when we could still hear!"
[10:50:17] <les> good day for spraying water!
[10:50:20] <jacky^> ehehe
[10:50:23] <robin_sz> asdfqwega: haha
[10:50:37] <robin_sz> Scorpions today :)
[10:52:11] <asdfqwega> How to feel old: Riding with friends of my youngest brother. "99 Luftballons" comes on the radio. Everyone else is surprised...they'd never heard it before.
[10:52:22] <robin_sz> heh
[10:52:35] <robin_sz> now, that was a good track :)
[10:52:55] <asdfqwega> I only felt old for about 5 seconds
[10:53:03] <jacky^> asdfqwega: germany ?
[10:53:04] <robin_sz> almost good enough to make me want to play bass :)
[10:53:06] <asdfqwega> Then I laughed at them and their ignorance XD
[10:53:28] <asdfqwega> No, I'm in Ohio, USA
[10:54:28] <jacky^> ah, ok but the song 99 Lufballons isn't of a germany band ?
[10:55:00] <asdfqwega> It's by Nina - and it was done both in english and german
[10:55:12] <jacky^> i remember..
[10:55:20] <asdfqwega> I prefer the german version
[10:55:24] <jacky^> nice
[10:55:32] <jacky^> do you remebrer Falco ?
[10:55:36] <jacky^> remember*
[10:56:15] <jacky^> Der Kommissar :D
[10:56:48] <asdfqwega> Rock Me Amedeus, Push Push, Viena Calling, Egoist
[10:56:55] <jacky^> :)
[10:57:09] <jacky^> Falco died in a car accident :\
[10:57:14] <asdfqwega> All favorites :)
[10:57:22] <asdfqwega> Recently?
[10:57:39] <jacky^> many years ago...
[10:58:04] <asdfqwega> Ah...I didn't really know
[10:58:08] <jacky^> 1998 in the Dominican Republic
[10:58:17] <robin_sz> np: Garbgae: Push it.
[10:58:20] <jacky^> yeah :(
[10:59:29] <les> Sigh. I give up on advice for home machinists trying to do ballscrew ends.
[10:59:40] <asdfqwega> Synthpop and 80s-ish is still alive and well
[10:59:40] <les> Repeat after me 3 times:
[11:00:06] <asdfqwega> I will not dutch-over my wife while she's sleeping.
[11:00:10] <les> The land surface on a rolled ballscrew is not a mechanical datum!!
[11:00:13] <asdfqwega> s/over/oven
[11:00:21] <robin_sz> heh
[11:00:28] <fenn> les: dont you think your sleeve would require a rather large bearing?
[11:00:53] <les> no it necks down
[11:01:00] <fenn> ah ok
[11:01:01] <robin_sz> * robin_sz nods
[11:01:16] <les> This is only for ballscrews that will not fit on the lathe
[11:01:23] <robin_sz> thats what I did with mine ... made an end piece and brazed it on
[11:02:22] <les> Well I have no problems machining screws up to 1 3/8 " diam
[11:02:48] <les> but I cannot offer it as a service for much less than anyone else
[11:03:17] <fenn> that idea helps me out with putting ends on acme threaded rod, actually
[11:03:28] <fenn> * fenn cant afford ballscrews
[11:05:00] <les> Well if acme screws have fully formed lands (usually) no problem
[11:05:11] <les> if your lathe is big enough
[11:05:14] <fenn> how the hell do you dial in a threaded rod?
[11:05:16] <asdfqwega> * asdfqwega sends an email and waits for weyland's mind to blow
[11:05:58] <les> fenn, you can put the indicator on ans set for threading with the same pitchas the screw
[11:06:02] <asdfqwega> ...and I blow my own mind while I wait, with some They Might Be Giants :)
[11:34:31] <jacky^> jacky^ is now known as jacky^afk
[11:48:46] <robin_sz> so les,
[11:48:54] <robin_sz> how is the new controlelr looking?
[11:48:59] <robin_sz> controller
[11:52:37] <les> good hardware, but it is not a turnkey controller
[11:52:42] <robin_sz> right
[11:52:48] <robin_sz> so lots of configuration to do
[11:52:54] <les> only does fours g codes
[11:53:06] <les> four
[11:53:07] <robin_sz> heh
[11:53:15] <les> so code to write...
[11:53:19] <les> lots.
[11:53:27] <robin_sz> cant you just put the emc interp on top?
[11:53:38] <les> So I still do not have a product
[11:53:46] <les> it IS the emc interp
[11:53:50] <jacky^afk> jacky^afk is now known as jacky^
[11:53:59] <robin_sz> well, that does more than 4 codes
[11:54:25] <les> the cannonicals for all the other stuff have to be written
[11:54:37] <les> oops one n
[11:54:45] <les> that is some work!
[11:54:45] <robin_sz> it does G0,1, 2 and 3?
[11:54:57] <les> 1,2,3,4
[11:55:08] <robin_sz> right ...
[11:55:25] <robin_sz> anyway ... regardless of all that
[11:55:33] <robin_sz> is the interpolated motion smooth?
[11:55:53] <les> I have not run it yet
[11:55:58] <robin_sz> ahh.
[11:56:00] <les> i suspect it is smooth
[11:56:08] <robin_sz> well, yeah
[11:56:12] <les> had to be on this piezo thing...
[11:56:18] <robin_sz> but people suspected emc was smooth ;)
[11:56:24] <les> ha
[11:57:02] <robin_sz> I guess you need to evaluate whether the tp is capable of smoothing up the motion
[11:57:17] <robin_sz> then decide whether its worth the time/effort to add the missing bits
[11:57:36] <les> Paul has done a good bit on a queue structure for quintic for emc2
[11:58:07] <les> but I sure like 90 microseconds for 4 axes on this card...and the price
[11:58:14] <robin_sz> yep
[11:58:33] <les> it just needs gui, more 274 support etc
[11:58:38] <les> can be done...
[11:58:47] <les> but it is some work
[11:59:03] <robin_sz> right
[11:59:04] <les> hence...
[11:59:14] <les> I have no product at this time
[11:59:17] <les> wish I did
[11:59:35] <robin_sz> ordered the iron yet?
[12:00:17] <les> no....if I have no control for the forseeable future....
[12:00:30] <les> what good is the iron to me?
[12:00:33] <robin_sz> here, have this chicken.
[12:01:02] <les> and egg?
[12:01:09] <robin_sz> * robin_sz nods
[12:02:07] <robin_sz> I still believe the hardware is good
[12:02:34] <les> I am trying to remember...you called the "dayton" router guys...
[12:02:44] <les> and they were clueless
[12:03:08] <les> still no machines for display here
[12:03:26] <robin_sz> thats a good thing no?
[12:03:36] <les> none in the country as far as I can see
[12:03:41] <les> yes good thing
[12:03:57] <robin_sz> I just see an opportunity
[12:04:31] <les> Well the things might sell...
[12:04:36] <robin_sz> might?
[12:04:40] <robin_sz> MIGHT?
[12:04:45] <les> just need proper sales and marketing
[12:04:53] <robin_sz> youve seen the price of the competitiors
[12:05:17] <robin_sz> theres plenty of margin
[12:05:25] <les> well the 4x8 iron for about 5k....
[12:05:53] <robin_sz> yeah
[12:05:55] <robin_sz> amazing
[12:06:06] <les> could aproach it another way....
[12:06:26] <les> just set up proper sales and maketing and sell as is
[12:06:38] <robin_sz> if it was me, (and someone in my country hadnt already got an exclusive)
[12:06:48] <robin_sz> I'd get the iron
[12:06:59] <robin_sz> put on a decent little control and evaluate it
[12:07:02] <les> the goofs on ebay won't ever get anywhere
[12:07:06] <robin_sz> quite
[12:07:13] <robin_sz> and mach2 will not sell it
[12:07:14] <les> but who is that?
[12:07:23] <les> is it the manufacturer?
[12:07:41] <robin_sz> doubtful
[12:07:57] <robin_sz> if it is .. then they just screwed up
[12:08:15] <robin_sz> who is going to distribute a product that the manu. undersells on ebay?
[12:08:25] <les> yeah
[12:08:35] <robin_sz> anywa, with mach2 its dead
[12:08:52] <les> let me check right now...the goofs were having a special "sale"
[12:08:59] <robin_sz> heh
[12:10:37] <les> http://search.ebay.com/_W0QQsassZusamarket999
[12:12:36] <les> "the power of mail spindle" hahaha
[12:13:05] <les> "the travel of working" hoho
[12:14:25] <les> " most of the cheap routers out there are junk"
[12:14:30] <les> hehe
[12:16:13] <robin_sz> well, at 3299 ...
[12:17:02] <jacky^> * jacky^ gcode is driving me crazy :
[12:17:05] <robin_sz> offer to collect it :)
[12:17:42] <les> 4x8 at 7999 is what I was looking at
[12:17:50] <robin_sz> ah yes
[12:17:52] <les> pacer sells at?
[12:17:54] <robin_sz> different market
[12:18:06] <robin_sz> I dont know what their 4x8 sells at
[12:18:25] <les> that one is on the pacer site now
[12:18:28] <les> painted red
[12:18:35] <robin_sz> thier 4x2 sells at 14K GBP
[12:18:44] <robin_sz> or whatever size the msall one is
[12:18:53] <robin_sz> waht that, 25K usd
[12:18:57] <les> holy sh#t
[12:19:18] <robin_sz> and thats with steppers
[12:19:27] <robin_sz> all the small machines are stepeprs
[12:19:31] <robin_sz> even the AXYZ ones
[12:20:03] <les> ok, so I have no control. What could I bolt on the thing RIGHT NOW?
[12:20:44] <robin_sz> g200x ?
[12:20:46] <jacky^> anyone have an idea what to choose for Circular movement g-code generator? allowed: Full, Half, Quarter, Arcs as Lines :(
[12:21:05] <robin_sz> did Mariss send you one?
[12:21:14] <les> he offered
[12:21:23] <robin_sz> jacky^: full, arcs
[12:21:28] <les> but it is not ready for prime time either
[12:21:36] <jacky^> robin_sz: ok
[12:21:59] <les> that's why I keep saying I have no control
[12:21:59] <robin_sz> les: wahts missing that you need?
[12:22:37] <les> some code I think
[12:22:47] <robin_sz> to do what?
[12:22:48] <les> front end
[12:22:52] <les> gui
[12:22:56] <robin_sz> you dont like the GUI?
[12:23:04] <robin_sz> I though it was OK
[12:23:10] <les> I had better check it again
[12:23:17] <les> perhaps I misunderstood
[12:23:39] <robin_sz> mine has been happily loading gcode and spinning stepeprs for months now
[12:24:14] <les> I have not seen the gui
[12:24:16] <robin_sz> all nicely clicky clicky
[12:24:21] <robin_sz> well, its all there
[12:24:30] <jacky^> last thing: maybe i finished to setup.. G02 G03 Circular: Incremental IJ, Absolute IJ, Radius Only, Incremental IJ+R, Absolute IJ+R ?
[12:24:48] <les> I guess I ignored it because it was stepper
[12:25:28] <robin_sz> increemental IJ I think
[12:25:41] <robin_sz> well, Id say get the hardware and try it
[12:25:43] <jacky^> robin_sz: thanks a lot !
[12:25:50] <jacky^> ok..
[12:26:48] <les> any site where I can look at screenshots?
[12:27:12] <robin_sz> of the GUI?
[12:27:17] <les> yeah
[12:27:19] <robin_sz> wait, I have it for dl
[12:27:23] <robin_sz> you can play with it :)
[12:27:29] <les> k
[12:27:55] <robin_sz> well, some version of it anyway
[12:28:11] <les> on winders box?
[12:28:39] <robin_sz> doze or linux? which you prefer?
[12:48:37] <robin_sz> right, now to set up VPN
[12:48:54] <robin_sz> anyone doen this with open VPN on a linux box??
[12:51:11] <jacky^> robin_sz: im not :(
[12:56:17] <jacky^> hi rayh !
[12:57:15] <rayh> How are you doing today?
[12:57:24] <alex_joni> elo
[12:57:31] <rayh> Hi alex
[12:57:39] <alex_joni> hey ray
[12:57:44] <alex_joni> what's up?
[12:57:44] <jacky^> rayh: i'm tryng to learn something of gcode
[12:58:07] <jacky^> hi alex_joni
[12:58:15] <rayh> Okay. I know just enough to hurt myself.
[12:58:23] <jacky^> :D
[12:58:36] <jacky^> haha :-)
[13:00:17] <rayh> jacky^: Have you done manual milling?
[13:01:07] <jacky^> rayh: the handbook pdf from yours website
[13:01:21] <jacky^> i'm reading at: http://www.linuxcnc.org/handbook/gcode/g-code.html
[13:03:20] <rayh> Okay
[13:04:57] <alex_joni> rayh: what's new?
[13:05:17] <rayh> Working to get ready for retroFest.
[13:05:21] <alex_joni> nice
[13:05:28] <jacky^> :D
[13:05:50] <alex_joni> how bout a fest ot build a wire hexapod ?
[13:05:51] <alex_joni> :D
[13:05:53] <alex_joni> to
[13:06:27] <rayh> I got the run the little NIST cable hexapod at NAMES a couple of years.
[13:06:40] <rayh> That is an interesting concept.
[13:06:44] <alex_joni> yeah I know
[13:06:49] <alex_joni> we talked about this a while back
[13:07:01] <rayh> Biggest issue is cable length.
[13:07:12] <alex_joni> and cable stretch
[13:07:49] <rayh> Yep. Even sagging will affect actual position.
[13:08:11] <alex_joni> sagging?
[13:08:37] <rayh> Gravity causes the cables to be not streight.
[13:09:16] <robin_sz> but .. thats predicatble?
[13:09:24] <rayh> It gets worse as you approach a singularity cause the load on that cable goes down.
[13:09:40] <alex_joni> robin: all is predictable
[13:09:44] <robin_sz> yeah
[13:09:46] <robin_sz> so ..
[13:09:49] <alex_joni> but.. you gotta take what's most important
[13:09:55] <alex_joni> and scrap the rest
[13:10:06] <robin_sz> knowing the weight of the platform and the weight per unit length of the wires ...
[13:10:17] <alex_joni> and normal stretch of the wire
[13:10:35] <robin_sz> it should be possible to modify the kinematics a teeny bit :)
[13:10:40] <alex_joni> right
[13:10:45] <alex_joni> that's what I was thinking
[13:10:53] <robin_sz> it would need Mr Maths to do it though
[13:11:01] <robin_sz> and hes just went for bnreakfast :)
[13:11:03] <alex_joni> but talking to jmk the other day there is a more simple solution ;)
[13:11:11] <alex_joni> at least mathematically
[13:11:15] <alex_joni> have 2 wires
[13:11:25] <alex_joni> 1 strong for support, the one you drive
[13:11:42] <alex_joni> and a lighter one for feedback
[13:11:47] <robin_sz> possible
[13:11:54] <robin_sz> but over complicated
[13:11:56] <robin_sz> and ...
[13:12:00] <alex_joni> robin_sz: I did a wiredrive once with a gecko
[13:12:13] <alex_joni> http://www.robcon.ro/emc/Resize%20of%20HPIM2892.JPG
[13:12:36] <alex_joni> I would use the same thing for a wiresupported platform
[13:12:36] <robin_sz> ooh, nice
[13:12:57] <alex_joni> don't laugh at that testboard on the top right ;)
[13:13:11] <rayh> I was thinking standing waves to measure wire length.
[13:13:12] <robin_sz> neat enough to sell :)
[13:13:22] <alex_joni> that's a 555 in there that does pulsing for the gecko
[13:13:26] <robin_sz> right
[13:13:40] <alex_joni> with a big CAP to o the start ramping :)
[13:13:44] <alex_joni> to do
[13:14:09] <alex_joni> I kinda like those motors
[13:14:11] <alex_joni> 24V DC
[13:14:16] <alex_joni> pretty tough
[13:14:22] <robin_sz> no encoder?
[13:14:23] <alex_joni> and with gearing
[13:14:28] <alex_joni> encoder is on the left
[13:14:33] <alex_joni> behind that black thingy
[13:14:44] <robin_sz> ahh,
[13:14:44] <alex_joni> below the transformer
[13:14:53] <alex_joni> separate than the actual driving
[13:14:56] <robin_sz> so you used TWO feed units
[13:14:59] <robin_sz> one to drive
[13:15:02] <robin_sz> the other to measure
[13:15:03] <alex_joni> yeah
[13:15:20] <robin_sz> I presume thats not "normal"
[13:15:20] <alex_joni> the gecko goes nuts if you open either of those ;)
[13:15:40] <alex_joni> well.. I didn't have time to reduce it mechanically
[13:15:47] <robin_sz> right
[13:15:48] <alex_joni> and it didn't matter
[13:15:58] <alex_joni> I've gotten some nice wiredrives
[13:16:08] <alex_joni> they have an encoder fitted in the pressing roller
[13:16:13] <robin_sz> ooh
[13:16:16] <robin_sz> nice
[13:16:18] <alex_joni> yeah
[13:16:22] <alex_joni> but not very cheap ;)
[13:16:28] <robin_sz> I have a *very* old Murex drive
[13:16:39] <alex_joni> anyways.. on this one I used a 2500 counts/rev encoder
[13:16:51] <alex_joni> on quadrature 10000 impulses for the gecko
[13:16:57] <robin_sz> it could do with an upgrade
[13:17:13] <alex_joni> so I could command it very nice for low speeds
[13:17:25] <alex_joni> it was for WIG welding
[13:17:30] <alex_joni> for driving the cold wire
[13:17:30] <robin_sz> the weld head is not so nice, but the supply is nice
[13:17:33] <robin_sz> yeah
[13:17:35] <robin_sz> I know
[13:17:42] <rayh> How large diameter is the measuring wheel
[13:17:44] <alex_joni> you don't need high speeds
[13:17:48] <robin_sz> I have a Transmig 500A set ...
[13:17:53] <robin_sz> thats nice
[13:17:56] <alex_joni> about 2.5 cm
[13:18:10] <robin_sz> can push 1.2mm wire into spray mode :)
[13:18:11] <alex_joni> so lots of precision
[13:18:22] <alex_joni> I'm starting with 1.4 wire soon
[13:18:26] <robin_sz> is precision that important?
[13:18:42] <alex_joni> on a platform thingy yes
[13:18:47] <robin_sz> right
[13:18:52] <alex_joni> not for WIG welding ;)
[13:18:58] <robin_sz> thats what I thought :)
[13:19:01] <alex_joni> the constant speed is a nice feature though
[13:19:07] <robin_sz> sure
[13:19:19] <alex_joni> I used this wiredrive along with a big welding automate
[13:19:21] <robin_sz> on robot welding, you you use dip or spray mode?
[13:19:29] <alex_joni> pulsed welding only
[13:19:34] <robin_sz> right
[13:19:38] <alex_joni> at least that's what I use
[13:19:50] <alex_joni> others prefer short arc on thin sheets
[13:19:56] <alex_joni> mainly because they are used to it
[13:20:08] <robin_sz> thin sheets?
[13:20:12] <alex_joni> I got some very nice power sources now
[13:20:15] <alex_joni> <3 mm
[13:20:18] <robin_sz> right
[13:20:20] <robin_sz> hey,
[13:20:24] <robin_sz> heres a thing
[13:20:30] <robin_sz> mercedes doors ...
[13:20:41] <robin_sz> they are 2mm at the bottom
[13:20:48] <robin_sz> 1.2 at the top
[13:20:50] <alex_joni> heh
[13:20:52] <alex_joni> nice
[13:20:59] <robin_sz> cant see the join
[13:21:05] <robin_sz> laser welded
[13:21:05] <alex_joni> but those are usually spot welded
[13:21:08] <alex_joni> or laser
[13:21:14] <alex_joni> or laser-Hybrid
[13:21:30] <robin_sz> and at the top of the A B and C posts to the roof, are laser welded too
[13:21:35] <robin_sz> cant see the join
[13:21:54] <alex_joni> nice
[13:21:59] <alex_joni> I did a lot of big welds lately
[13:22:04] <alex_joni> Tandem-MIG
[13:22:05] <robin_sz> big
[13:22:08] <robin_sz> or BIG
[13:22:08] <alex_joni> 2 power sources
[13:22:13] <alex_joni> not BIG, big
[13:22:18] <robin_sz> oh :)
[13:22:18] <alex_joni> about 300 mm^2
[13:22:26] <alex_joni> 25-30 mm sheets
[13:22:26] <alex_joni> ;)
[13:22:33] <robin_sz> not single pass?
[13:22:37] <alex_joni> nope
[13:22:43] <alex_joni> about 4-5 passes
[13:22:55] <robin_sz> root, 2 fills and a covr or two
[13:22:56] <alex_joni> single pass will get you into trouble at that thickness
[13:22:59] <robin_sz> yeah
[13:23:00] <alex_joni> right
[13:23:06] <alex_joni> root is single wire usually
[13:23:10] <robin_sz> yeah
[13:23:11] <alex_joni> to cloe eventual gaps
[13:23:13] <robin_sz> backed?
[13:23:16] <alex_joni> close even
[13:23:24] <alex_joni> theoretically backed ;)
[13:23:28] <robin_sz> heh
[13:23:32] <alex_joni> sometimes not :))
[13:23:41] <robin_sz> whatever works
[13:23:45] <alex_joni> yeah
[13:23:48] <alex_joni> anyways..
[13:23:54] <alex_joni> I got a new Tandem system lately
[13:24:01] <alex_joni> 2 x 650 Amps power sources
[13:24:04] <robin_sz> nice
[13:24:13] <alex_joni> wanna go a bit up with welding speed on that one
[13:24:19] <alex_joni> that's why I need the 1.4 wire
[13:24:27] <robin_sz> 1m a minute?
[13:24:34] <alex_joni> 1.5 for 5mm welds
[13:25:11] <alex_joni> or maybe even higher
[13:25:38] <alex_joni> on single wire you can go up to 50 cm/min for the same weld
[13:25:48] <alex_joni> around 30cm/min on manual
[13:25:52] <robin_sz> yeah
[13:26:00] <alex_joni> right.. now back to emc ;)
[13:26:05] <robin_sz> yep :)
[13:26:19] <alex_joni> was thinking on a plasma cutting wire suspended platform
[13:26:34] <robin_sz> possible
[13:26:43] <robin_sz> but might not make the accuracy
[13:26:43] <alex_joni> would pay up for big sizes
[13:26:52] <robin_sz> still need +- 0.5mm
[13:26:55] <alex_joni> yup
[13:27:08] <robin_sz> I had a part rejected on accuracy last week
[13:27:21] <alex_joni> bummer :/
[13:27:28] <robin_sz> 30.125 inches diamenter
[13:27:37] <robin_sz> I was 0.012 out
[13:27:49] <alex_joni> smaller?
[13:27:50] <robin_sz> got to get+- 0.005"
[13:28:07] <robin_sz> thats not going to be easy
[13:28:13] <alex_joni> right
[13:28:13] <robin_sz> its OK in hot rolled
[13:28:16] <alex_joni> thick sheet?
[13:28:20] <robin_sz> but im out in cold rolled
[13:28:24] <robin_sz> nah, 3mm
[13:28:31] <alex_joni> still ok
[13:28:41] <robin_sz> cold rolled has more internal stresses
[13:28:47] <robin_sz> it resahopes as you cut it
[13:28:48] <alex_joni> I had problems with 25 mm sheets, cut oxy-fuel
[13:28:59] <alex_joni> those would reshape up to 4-5mm
[13:29:01] <robin_sz> yeah
[13:29:08] <robin_sz> LoTS of heat going in with oxy
[13:29:15] <alex_joni> yeah
[13:29:33] <alex_joni> plasma would be better
[13:29:35] <robin_sz> my solution is simple though
[13:29:38] <alex_joni> maybe high def.
[13:29:54] <robin_sz> I know how much it is out, and in which direction
[13:30:11] <robin_sz> so I will just adjust the couts-per-rev on the CNC
[13:30:19] <robin_sz> to compensate for the ovality
[13:30:22] <alex_joni> yeah, but you need to place the sheet in the same direction every time
[13:30:30] <robin_sz> thats OK, it is same dir
[13:30:34] <robin_sz> and parts same dir
[13:30:43] <robin_sz> process evaluation :)
[13:30:48] <alex_joni> heh
[13:30:57] <robin_sz> get it right, dont change a thing
[13:31:18] <alex_joni> yeah..
[13:31:24] <alex_joni> till it gets warmer outside
[13:31:27] <robin_sz> well
[13:31:32] <robin_sz> I have that covered
[13:31:36] <alex_joni> heh
[13:31:44] <robin_sz> see, the plate warms up as you cut it
[13:31:51] <robin_sz> part 3 coems out different to part 1
[13:31:58] <robin_sz> ( 3 parts per sheet )
[13:32:07] <alex_joni> yeah
[13:32:13] <robin_sz> the solution is easy
[13:32:20] <anonimasu> dhello
[13:32:20] <robin_sz> pre warm the sheets :)
[13:32:26] <alex_joni> yo anders
[13:32:45] <alex_joni> rayh: still there?
[13:32:57] <robin_sz> right ... back to emc :)
[13:33:05] <robin_sz> actually, Im off .. later guys
[13:33:14] <alex_joni> later robin
[13:33:21] <rayh> you bet I am.
[13:33:32] <alex_joni> rayh: was planing to answer that task planner mail of yours
[13:33:45] <alex_joni> just didn't get enough time to do it :(
[13:33:48] <anonimasu> alex_joni: what's up?
[13:33:50] <rayh> Ah.
[13:34:01] <alex_joni> anonimasu: not much
[13:34:21] <alex_joni> rayh: didn't forget/scrap it... just didn't make it with the reply
[13:34:44] <rayh> I know that problem. I don't get done half what I plan.
[13:34:55] <alex_joni> I was thinking about emc, and the ability for macros
[13:35:01] <alex_joni> or smthg like that
[13:35:04] <rayh> Okay
[13:35:13] <alex_joni> and I think that task is the one that should support scripting
[13:35:21] <alex_joni> and interpret it
[13:35:29] <rayh> What's this macro going to do.
[13:35:33] <alex_joni> that way the macro could contain motion, io, etc
[13:35:42] <alex_joni> anything you would like
[13:35:47] <alex_joni> or anyone would like
[13:35:56] <alex_joni> say toolchanging
[13:35:58] <anonimasu> alex_joni: yep
[13:36:01] <anonimasu> agreed :9
[13:36:02] <anonimasu> :)
[13:36:04] <alex_joni> or batch processing
[13:36:10] <alex_joni> or whatever
[13:37:34] <rayh> I hadn't thought of it being a scripting interpreter but it certainly could.
[13:37:44] <rayh> Or at least a module of it could.
[13:37:48] <alex_joni> not only a scripting interpreter
[13:37:58] <alex_joni> but if it is the central piece
[13:38:07] <alex_joni> then it should be able to do smthg like that
[13:39:04] <rayh> I think "central piece" is the key to thinking about task.
[13:39:13] <anonimasu> * anonimasu nods
[13:39:17] <alex_joni> yeah
[13:39:53] <rayh> One thing we want to do is change a tool in mid g-code program.
[13:40:12] <anonimasu> pause & check tool and change ;)
[13:40:15] <rayh> My notion is that we just pause the stack of stuff coming from interp
[13:40:34] <anonimasu> ray yeah, and inject our motion..
[13:40:35] <rayh> store the necessary position and modal info.
[13:40:41] <alex_joni> short question: who commands the tool change?
[13:40:46] <rayh> Inject the manual mode stuff.
[13:40:47] <alex_joni> I mean who initiates it?
[13:40:53] <alex_joni> user & gui ?
[13:40:59] <anonimasu> either..
[13:40:59] <alex_joni> or g-code based?
[13:41:00] <rayh> txx m6
[13:41:13] <alex_joni> so g-code
[13:41:14] <rayh> Or manual tool change.
[13:41:15] <anonimasu> or a tool breakage detection algorithm
[13:41:20] <rayh> Same thing for either.
[13:41:43] <rayh> Yes we have some tool load canon commands.
[13:41:47] <alex_joni> anonimasu: heh
[13:42:01] <anonimasu> wouldnt matter..
[13:42:10] <anonimasu> but where does things go after the m6 ;)
[13:42:35] <alex_joni> right
[13:42:42] <anonimasu> pause the interpreter and inject our stuff into it..(from where)
[13:42:51] <rayh> That would depend upon the last auto state stored by task...
[13:43:12] <anonimasu> from a macro language parser specially for things like that..?
[13:43:15] <rayh> and the new tool and offsets commanded by the next interp message
[13:43:37] <anonimasu> or will we add all the functionality into the interpreter alone?
[13:43:45] <rayh> Those sorts of things would have to be rules based.
[13:43:48] <alex_joni> anonimasu: better not
[13:43:57] <rayh> No I'd give them to task
[13:43:59] <alex_joni> because you'll have to add it to all interps
[13:43:59] <anonimasu> alex_joni: my thought exactly :)
[13:44:11] <rayh> That way we don't have to ask the interpreter to compensate for
[13:44:16] <anonimasu> yeah better have a macro language that can do g-code to the interpreter..
[13:44:17] <rayh> other moves we've made.
[13:44:21] <anonimasu> err throw
[13:44:40] <anonimasu> for motion as a example..
[13:44:51] <anonimasu> moving back to the correct place and things like that
[13:45:03] <rayh> exactly
[13:45:20] <rayh> let the interpreter free run on the program
[13:45:31] <alex_joni> anonimasu: why throw it to the interp?
[13:45:39] <rayh> make the interpreter conform to that after other stuff has been done.
[13:45:40] <alex_joni> from task you can throw it directly to motion
[13:45:51] <anonimasu> alex_joni: but that would complicate things much wouldnt it?
[13:46:04] <anonimasu> if you want to be able to pass normal gcode as motion
[13:46:44] <rayh> That case, a pure motion move drops through task on the left column
[13:46:55] <rayh> It has only two gates.
[13:47:16] <rayh> One says, "hey this is pure motion."
[13:47:29] <rayh> The other says, "okay do it now."
[13:48:11] <anonimasu> you mean a priority gate..
[13:48:14] <anonimasu> or somthing like it
[13:48:25] <anonimasu> injected motion > other motion
[13:48:52] <rayh> Yes. The bottom gate schedules motion and IO in coordination with each other.
[13:49:08] <anonimasu> yeah
[13:52:43] <alex_joni> rayh: why are there 2 Bidirectional NML/SHMEM Service Modules ?
[13:52:48] <alex_joni> hey paul_c
[13:52:57] <rayh> phone brb
[13:53:24] <rayh> cause there may be more than one shmem process running on different processors.
[13:55:44] <alex_joni> right
[13:56:09] <alex_joni> but you need to assign some tasks to them
[13:56:12] <alex_joni> io on one
[13:56:17] <alex_joni> toolchanging on another
[13:56:22] <alex_joni> gear changing on one
[13:56:23] <alex_joni> etc
[13:59:04] <les> hi alex
[13:59:08] <alex_joni> hey les
[13:59:30] <jacky^> rayh: i would run this program: http://www.cuttingedgecnc.com/G-Codes/cncweb.txt, but emc say: 1 - offset index missing, any idea ? I deleted all lines before G21 G00 Z1
[13:59:34] <les> I give up on teching about ballscrew ends on cced list
[13:59:38] <les> teaching
[14:00:28] <jacky^> code is not compatible ?
[14:00:35] <alex_joni> jacky^: along with G43 you need a D1..Dx
[14:00:40] <alex_joni> based on the tool you want to use
[14:00:53] <alex_joni> G43 says it should use cutter length compensation
[14:01:00] <alex_joni> but it doesn't say what length to use
[14:01:11] <jacky^> mmmhh.. to hard :(
[14:01:15] <alex_joni> so you need to tell it to use tool #1 for example
[14:01:22] <alex_joni> or simply remove the G43
[14:01:24] <jacky^> ahh
[14:01:32] <alex_joni> if you don't need cutter length compensation
[14:01:39] <jacky^> how to say to use tool 1 ?
[14:01:44] <alex_joni> D1
[14:01:48] <alex_joni> D1 means tool #1
[14:01:52] <alex_joni> D2 means tool #2
[14:01:54] <alex_joni> etc.
[14:01:55] <jacky^> on the same line of G43 ?
[14:01:58] <alex_joni> yeah
[14:02:02] <alex_joni> G43 D1
[14:02:04] <jacky^> thaks :-)
[14:02:07] <alex_joni> also
[14:02:10] <jacky^> thanks*
[14:02:27] <alex_joni> you need to make sure that your tool file (usually emc.tbl) has the appropiate tool length defined
[14:02:36] <jacky^> so, D1 mean tool 1 D2 tool 2 and so on..
[14:02:43] <alex_joni> yes
[14:02:45] <jacky^> yeah
[14:03:00] <jacky^> my first tool in .tbl file is 3 mm
[14:03:06] <jacky^> should be right
[14:03:08] <alex_joni> the tool file (emc.tbl or generic.tbl has the tool number, tool diameter, and tool length)
[14:03:12] <alex_joni> then you're set
[14:03:18] <jacky^> :-)
[14:03:22] <jacky^> tnx a lot
[14:03:28] <alex_joni> you're welcome
[14:03:39] <alex_joni> les: people who want to learn.. should do it
[14:03:45] <alex_joni> teaching other won't help ;)
[14:03:56] <alex_joni> there is a nice saying in romania
[14:04:09] <alex_joni> people who know, do
[14:04:17] <alex_joni> those who don't teach others
[14:04:28] <paul_c> G43 H1 - Not D1
[14:04:29] <alex_joni> and those who can't do either usually organize
[14:04:56] <alex_joni> jacky^: what paul said (D is for diameter comp., H is for tool length) :/
[14:05:03] <jacky^> hi paul_c H1 instead of d1 , thnaks
[14:06:06] <paul_c> Also reduce the P word on the second line - 2,500 seconds wait is way too long.
[14:06:54] <jacky^> ok, thank you
[14:08:37] <paul_c> And watch out for those F words - A few F0. in there.... And several F.1
[14:09:00] <alex_joni> my mom always said not to use the F word ;)
[14:10:29] <jacky^> :)
[14:11:22] <jacky^> i'm going to remove the F..
[14:17:20] <alex_joni> hey Martin
[14:17:35] <alex_joni> paul_c: short question on M62,63,64,65
[14:17:47] <alex_joni> could those be supported through NML?
[14:17:52] <alex_joni> or is that too slow?
[14:19:05] <alex_joni> at least that's how it's done in emc1
[14:19:15] <paul_c> Two of those are coordinated IO with motion.
[14:19:24] <alex_joni> right
[14:19:39] <alex_joni> so motmod would need to have 2 more hal-pins exported
[14:19:52] <alex_joni> and should be able to turn those on/off like in emc1
[14:19:54] <alex_joni> right?
[14:20:33] <paul_c> This is where the HAL pin/signal stuff starts going tits up.
[14:20:46] <alex_joni> how so?
[14:20:54] <paul_c> The g code is "M62 Pn"
[14:21:02] <alex_joni> right
[14:21:13] <paul_c> where n is a valide index number.
[14:21:14] <alex_joni> where n is the pin number
[14:21:35] <alex_joni> ok.. so you might have a max_n number
[14:21:39] <alex_joni> say 32
[14:21:54] <alex_joni> and link the ones you actually use (or none if you don't use any)
[14:22:38] <alex_joni> the valid index number shouldn't be a valid parport index number
[14:22:39] <rayh> Why do you say that the motion module needs more pins for something that is IO
[14:22:43] <paul_c> The way it currently works in EMC[1]...
[14:22:48] <alex_joni> that would be too limiting in concept
[14:23:06] <alex_joni> rayh: because IO doesn't run synched with motion
[14:23:21] <paul_c> A P parameter is given - It need not be correct for the system (i.e. P might be 27)
[14:23:25] <rayh> It certainly can in the HAL
[14:23:37] <Imperator_> Hi Alex
[14:23:50] <paul_c> and if the index is outside the range supported by the hardware, it is silently ignored.
[14:24:01] <anonimasu> :)
[14:24:23] <alex_joni> paul_c: right, but the same can be for emc2 too
[14:24:29] <alex_joni> you have 32 valid pins
[14:24:43] <alex_joni> but you have only the first 3 pins actually connected to hardware
[14:24:54] <alex_joni> be it parport, or STG-io, or whatever
[14:25:20] <alex_joni> if there's a M62 P12, pin12 would be set
[14:25:22] <paul_c> So where would you translate the P word in to HAL identifiers ?
[14:25:26] <alex_joni> but that wouldn't do anything
[14:25:37] <alex_joni> have it fixed
[14:25:44] <alex_joni> 32 P words = 32 HAL pins
[14:25:59] <alex_joni> maybe 32 is a bit much
[14:26:08] <paul_c> 32 to much.
[14:26:14] <alex_joni> I'd have it based on a define
[14:26:36] <alex_joni> thus if anybody would need to increase it, it should be fairly easy to modify & recompile
[14:26:48] <alex_joni> it's a too narrow case for yet another ini parameter
[14:26:54] <alex_joni> I think we can agree on that
[14:27:22] <anonimasu> yeah
[14:27:22] <anonimasu> true
[14:27:23] <rayh> I don't agree.
[14:27:26] <alex_joni> say 8 pins would be a good default value
[14:27:32] <alex_joni> rayh: why not?
[14:27:43] <anonimasu> there will always be people that need 64 pins :)
[14:27:44] <rayh> A lookup table is much easier to deal with than a recompile.
[14:28:16] <rayh> Sure it makes the code a bit messier but is run time affected.
[14:30:23] <alex_joni> rayh: I'd agree with you in a general case
[14:30:33] <alex_joni> but emc is a configuration nightmare as it is
[14:30:55] <alex_joni> adding more config stuff is not very wise IMHO
[14:30:58] <paul_c> and HAL has done nothing to help in that quarter.
[14:31:17] <alex_joni> paul_c: that's debateable
[14:31:23] <alex_joni> but generally I agree
[14:31:45] <alex_joni> it makes it very easy/flexible for an somehow advanced user
[14:31:52] <paul_c> old case: EMC[1] had two core configs.
[14:31:56] <alex_joni> but for aunt tillie it kinda fails
[14:31:56] <rayh> IMO it is a config nightmare but there is no partial solution to that.
[14:32:10] <rayh> we will simply have to build a configuration front end.
[14:32:20] <paul_c> new case: emc2 now has three (min) configs, and optionally, many more.
[14:32:26] <rayh> and we might as well use it to make the system as flexible as possible.
[14:32:36] <alex_joni> paul_c: why 3?
[14:33:25] <alex_joni> emc.nml is smthg nobody touches (or very unlikely)
[14:33:39] <alex_joni> emc.ini is the same as in EMC[1]
[14:33:46] <alex_joni> and there's a hal file
[14:33:56] <paul_c> *.nml, *.ini, core_stepper.hal, standard_pinout.hal, etc
[14:34:06] <alex_joni> one .hal file would be enough
[14:34:25] <paul_c> No. One ini plus one nml would be enough.
[14:34:27] <alex_joni> but it's kept separately to keep it flexible
[14:34:39] <alex_joni> who changes the nml ?
[14:34:46] <anonimasu> I do
[14:34:48] <paul_c> I do.
[14:34:53] <alex_joni> anonimasu doesn't count ;)
[14:35:01] <alex_joni> paul_c: yeah, I do too
[14:35:01] <anonimasu> * anonimasu pokes alex_joni
[14:35:08] <alex_joni> but I meant users who do
[14:35:25] <alex_joni> very few
[14:35:25] <anonimasu> paul_c: are there a easy way to compile your rcslib apps with emc2?
[14:35:28] <paul_c> The nml is rarely altered.
[14:35:35] <alex_joni> anonimasu: what did you change in nml?
[14:35:36] <anonimasu> or is it easier to use emc1 when compiling stuff?
[14:35:47] <anonimasu> alex_joni: I added my app there..
[14:35:49] <alex_joni> emc1 is not easy to compile
[14:35:55] <alex_joni> anonimasu: what for?
[14:36:02] <paul_c> anonimasu: What is it you are trying to do ?
[14:36:08] <alex_joni> you could have used one of the many that are there
[14:36:19] <anonimasu> paul_c: I am curious if I can place my stuff in the emc2/src/ dir instead
[14:36:31] <alex_joni> paul_c: he's coding a feedrate override based on force measured on the spindle
[14:36:45] <anonimasu> alex_joni: some other day :)
[14:36:48] <alex_joni> heh
[14:36:52] <anonimasu> I still have the stuff for measuring load at my desk in a box
[14:36:56] <alex_joni> that's what you said ;)
[14:36:57] <anonimasu> shunt & opamps
[14:37:04] <alex_joni> anyways..
[14:37:06] <anonimasu> yeah but I kind of got sidetracked with other stuff
[14:37:10] <alex_joni> back to config stuff
[14:37:22] <alex_joni> emc.ini won't go away
[14:37:23] <anonimasu> paul_c: just trying to find a good way to stop being dependant on the emc1 sourcetree
[14:37:27] <alex_joni> I think we can agree there
[14:37:27] <anonimasu> :)
[14:37:42] <alex_joni> emc.nml isn't usually touched (so it might not get counted)
[14:37:48] <paul_c> anonimasu: Set up a src/utils/anonimasu directory, and create a simple make file there.
[14:37:49] <alex_joni> that leaves us with the hal stuff
[14:38:14] <alex_joni> paul_c: for hal there are 2 ways
[14:38:33] <alex_joni> 1. the current one (having a separate .hal file that gets parsed by halcmd or whatever)
[14:38:45] <alex_joni> 2. define stuff in emc.ini and have halcmd parse that
[14:38:57] <alex_joni> but it leaves you with a lot less flexibility
[14:39:14] <paul_c> crap
[14:39:52] <alex_joni> ok.. say we want to move it to the ini
[14:39:53] <paul_c> Most of the information hal would need is already in a "standard" ini
[14:40:08] <alex_joni> what if you use a stg ?
[14:40:20] <alex_joni> or a stg/parport combination ?
[14:40:21] <rayh> In the case of my multiple user space SHMEM modules you will have to touch NML.
[14:40:42] <alex_joni> rayh: that's smthg I don't want to go to right now ;)
[14:40:52] <alex_joni> that would complicate it even more
[14:42:26] <alex_joni> ok, taken from generic.ini
[14:42:32] <alex_joni> how would anybody use:
[14:42:32] <alex_joni> PROBE_INDEX = 0
[14:42:32] <alex_joni> PROBE_POLARITY = 1
[14:42:45] <alex_joni> probe_polarity is clear
[14:42:51] <alex_joni> but index refers to what?
[14:42:59] <paul_c> pin ID
[14:43:05] <alex_joni> what pin ?
[14:43:10] <alex_joni> USC pin ?
[14:43:11] <paul_c> YOU DECIDE
[14:43:15] <alex_joni> where?
[14:43:17] <alex_joni> and how ?
[14:43:37] <alex_joni> PROBE_INDEX = parport.0.pin-12-in ?
[14:43:43] <paul_c> PROBE_INDEX = par.pin.whatthefriggin..0
[14:43:46] <alex_joni> that would make sense
[14:44:11] <alex_joni> ok, now the step/dir stuff
[14:44:17] <alex_joni> that's a bit more complicated
[14:44:44] <alex_joni> would you have [AXIS_0] STEP = foo.x.pin
[14:44:52] <alex_joni> and DIR = foo.x.pin2 ?
[14:45:19] <paul_c> is that any worse than having it defined in some hal configuration file ?
[14:45:21] <alex_joni> what if stepping type gets changed ?, then STEP/DIR need to be changed too
[14:46:01] <alex_joni> having it defined in some hal configuration file allows us (devel.) to build standard hal files for all types that can appear
[14:46:56] <jmk_sleep> jmk_sleep is now known as jmkasunich
[14:47:08] <alex_joni> morning john
[14:47:14] <jmkasunich> morning
[14:47:26] <alex_joni> seems you crashed in at the right time ;)
[14:48:20] <alex_joni> brb
[14:48:23] <alex_joni> changing puters
[14:49:56] <jmkasunich> right time, wrong time... dunno
[14:51:04] <alex_joni> back
[14:51:21] <jmkasunich> didn't miss anything
[14:51:28] <alex_joni> right
[14:51:40] <jmkasunich> my arrival seems to have ended the discussion
[14:51:46] <alex_joni> now.. we started talking about motion coordinated io
[14:51:55] <jmkasunich> yes (I read back)
[14:52:00] <alex_joni> ok
[14:52:11] <alex_joni> what's your idea on that?
[14:52:35] <jmkasunich> about what you were thinking... hal pins out of the motion module
[14:53:12] <jmkasunich> I need to "follow" one of those commands thru the motion controller, see how they are handled
[14:53:34] <alex_joni> well..right now there's a NML command
[14:53:37] <jmkasunich> I assume they get queued same as arc and line commands
[14:53:43] <alex_joni> from interp/task to the motion
[14:53:52] <alex_joni> then the motion looks if it needs to do it now
[14:54:00] <alex_joni> or if it should get scheduled on tp
[14:54:27] <jmkasunich> do you know the name of the NML command?
[14:54:33] <paul_c> all SET_DOUT commands get put on the tp stack
[14:54:43] <alex_joni> EMCMOT_SET_DOUT
[14:54:59] <jmkasunich> found em
[14:55:00] <alex_joni> and there's a EMC_MOTION_SET_DOUT
[14:55:09] <jmkasunich> #if 0'ed out in command.c
[14:55:12] <alex_joni> not really sure where they differ
[14:55:14] <alex_joni> right
[14:55:35] <jmkasunich> I think EMC_MOTION_SET_DOUT is the NML command
[14:55:47] <jmkasunich> and EMCMOT_SET_DOUT is the motion controller command
[14:56:15] <jmkasunich> usrmotintf.cc converts NML to shmem motion commands and vice-versa
[14:56:22] <alex_joni> right
[14:56:42] <alex_joni> but it's in taskintf.cc
[14:57:10] <jmkasunich> ok (my brain's a little rusty)
[14:57:41] <Jymmm> * Jymmm hands jmkasunich a can of WD40
[14:58:39] <alex_joni> hello matt
[14:58:47] <mshaver> hi!
[14:58:59] <jmkasunich> hi matt
[14:59:02] <mshaver> i sent that card to your cousin
[14:59:14] <alex_joni> yeah he's got it
[14:59:24] <alex_joni> will get here next month probably
[15:00:02] <mshaver> oh ok, jon elson may send you some stuff too
[15:00:44] <mshaver> universal stepper and/or universal pwm controller cards
[15:01:05] <mshaver> hi john!
[15:01:16] <alex_joni> is it christmas ? ;)
[15:01:32] <mshaver> in june...
[15:01:49] <alex_joni> yeah.. kind off ;)
[15:01:56] <Jymmm> alex_joni : Yes it is, where is OUR presents!!!
[15:02:50] <mshaver> he'll be committing them to CVS sometime after he gets the cards...
[15:09:40] <paul_c> jmkasunich: When are you going to fix slot 4 so that it uses the correct PLAT definitions when compiling EMC ?
[15:10:22] <jmkasunich> today?
[15:10:32] <jmkasunich> can I just copy them from the rcslib script?
[15:11:13] <paul_c> would have thought so.
[15:12:43] <jmkasunich> USRPLAT=nonrealtime and RTPLAT=realtime?
[15:13:27] <jmkasunich> looks like we also added a ,/configure step to rcslib
[15:13:46] <alex_joni> rcslib/etc/configure
[15:14:00] <jmkasunich> right... is there something similar for emc?
[15:14:07] <alex_joni> it's the same
[15:14:14] <alex_joni> emc uses the results from rcslib
[15:14:34] <paul_c> No need to run configure a second time - That is only needed for rcslib.
[15:14:35] <alex_joni> so you need to use the same PLAT RTPLAT for emc as you did for rcslib
[15:14:57] <alex_joni> nonrealtime & realtime for the ./configure sorted-out stuff
[15:15:04] <jmkasunich> so I shouldn't need to do "cd etc ; ./configure ; cd ,, ; cd src ; make"
[15:15:08] <alex_joni> nah
[15:15:14] <alex_joni> you do ./configure once
[15:15:22] <alex_joni> on the first compile
[15:15:27] <alex_joni> after that only make
[15:15:32] <alex_joni> both in rcslib and emc
[15:15:52] <jmkasunich> this is the compile farm, I do ./configure every time, in case the config script has been changed
[15:16:00] <alex_joni> pretty small chance for ./configure to change in the future for emc1/rcslib
[15:16:05] <jmkasunich> true
[15:16:10] <alex_joni> but.. if you do, leave it like that
[15:16:17] <alex_joni> run ./configure, then both makes
[15:16:21] <alex_joni> for rcslib and emc
[15:16:28] <alex_joni> with the same PLAT's
[15:16:33] <jmkasunich> ok
[15:17:03] <jmkasunich> the rcslib farm build script runs configure then make, so the emc farm build script should only need to do make
[15:17:28] <jmkasunich> only way to find out....
[15:17:30] <jmkasunich> * jmkasunich tries it
[15:19:55] <jmkasunich> revised script loaded on farm, it's doing the cvs up now
[15:40:54] <steve_stallings> steve_stallings is now known as steves_logging
[15:46:54] <CIA-12> 03paul_c * 10emc2/src/emc/ini/ (iniaxis.cc initraj.cc): todo tags added to all FIXME comments so that they get highlighted when auto-generating docs.
[15:47:09] <CIA-12> 03paul_c * 10emc2/src/emc/ (12 files in 3 dirs): todo tags added to all FIXME comments so that they get highlighted when auto-generating docs.
[15:47:19] <CIA-12> 03paul_c * 10emc2/src/emc/ (4 files in 2 dirs): todo tags added to all FIXME comments so that they get highlighted when auto-generating docs.
[15:47:41] <CIA-12> 03paul_c * 10emc2/src/ (19 files in 8 dirs): todo tags added to all FIXME comments so that they get highlighted when auto-generating docs.
[15:47:47] <CIA-12> 03paul_c * 10emc2/src/rtapi/ (rtl_rtapi.c rtl_ulapi.c sim_rtapi.c): todo tags added to all FIXME comments so that they get highlighted when auto-generating docs.
[15:48:19] <alex_joni> heh..nice ;)
[15:50:28] <CIA-12> 03paul_c * 10emc2/src/emc/ (kinematics/trajectory.h rs274ngc/interp_convert.cc): todo tags added to the remaining FIX-ME comments so that they get highlighted when auto-generating docs.
[15:50:31] <jmkasunich> darn... the slot 4 build failed
[15:51:25] <paul_c> PLAT is still wrong !
[15:51:47] <jmkasunich> ? those are the ones used for the rcslib build
[15:52:04] <paul_c> running "make clean PLAT=nonrealtime"
[15:52:19] <paul_c> unning "make make_dirs PLAT=realtime"
[15:52:33] <paul_c> (both from the rcslib log)
[15:52:41] <jmkasunich> yes
[15:52:53] <jmkasunich> isn't that what you told me to do?
[15:52:58] <paul_c> Yes.
[15:53:19] <paul_c> EMC on slot 4 is still using PLAT=linux_rtai & rtai
[15:53:31] <jmkasunich> no it's not
[15:53:37] <jmkasunich> (refresh your browser)
[15:54:45] <paul_c> got it...
[15:56:34] <jmkasunich> java: Command not found.
[15:56:42] <alex_joni> lol
[15:57:58] <CIA-12> 03paul_c * 10emc/src/emctask/emccanon.cc: It would appear that slot 4 doesn't like line breaks in the middle of strings..
[15:58:23] <paul_c> Let's see if that cures it.
[15:59:17] <jmkasunich> ok, I restarted it
[15:59:24] <jmkasunich> takes about 20 mins tho
[16:01:40] <paul_c> well.... You can use the time to explain how you plan on integrating the trajectory planner that I'm working on in to emc2.
[16:01:50] <jmkasunich> lol
[16:02:04] <jmkasunich> the planner that I don't really know anything about?
[16:02:12] <alex_joni> budge it in
[16:02:37] <paul_c> the one I've been committing files to emc2 for.
[16:02:57] <jmkasunich> sorry, haven't been studying those commits
[16:03:09] <jmkasunich> does it use the same API as the existing planner?
[16:03:45] <paul_c> currently, yes. But it is an ugly kludge.
[16:04:39] <CIA-12> 03alex_joni * 10emc2/src/hal/ (hal.h hal_lib.c hal_priv.h): added locking types and functions to hal_lib
[16:04:43] <jmkasunich> well if its the same API, then the choice of planner is a link time thing... we can link two motion modules (motion.ko, and motion_old.ko or whatever)
[16:05:02] <paul_c> streamlining the data passed to the trajectory stack would yield improvements (and require changes) in some of the other sources.
[16:05:44] <jmkasunich> suggestion... keep the api compatible until we have significant testing on the new planner, then improve the api and drop the old one
[16:05:56] <paul_c> and getting rid of the existing tp/tc structs would save significant memory.
[16:08:31] <paul_c> The existing API does not allow for stuff like NURBS
[16:09:25] <jmkasunich> I would like to have both objects compiled and allow the user to select between them from the ini file, at least for a while
[16:09:45] <jmkasunich> so folks who are simply making parts don't become unwitting beta testers
[16:09:57] <paul_c> ini selection for the TP ???
[16:10:08] <jmkasunich> old vs. new
[16:10:39] <alex_joni> imho it's not really necessary to clutter the ini for this
[16:10:47] <alex_joni> the people who want to test the new TP
[16:10:56] <jmkasunich> just a mo
[16:11:01] <alex_joni> probably will know how to do it even without ini params
[16:11:13] <jmkasunich> there is no "cluttering" involved....
[16:11:39] <alex_joni> you mean having motmod (and motmod_new) ?
[16:11:41] <jmkasunich> the ini file already contains an [EMCMOT] section, and it alreay has a variable EMCMOT=motmod
[16:11:49] <alex_joni> that could be ok
[16:11:54] <jmkasunich> so have motmod and motmodold
[16:12:11] <alex_joni> but not smthg like TRAJECTORY_PLANNER = old | new
[16:12:15] <alex_joni> ;)
[16:12:17] <jmkasunich> no
[16:12:21] <alex_joni> ok
[16:12:31] <jmkasunich> the make would build two modules, motmod and motmodold
[16:13:05] <jmkasunich> or perhaps motmod and motmodnew... depends on which one we want to be the default
[16:13:25] <alex_joni> for now motmod & motmodnew
[16:13:36] <alex_joni> at a later stage motmod & motmodold
[16:13:51] <jmkasunich> yeah, and eventually drop motmodold
[16:13:59] <alex_joni> right
[16:14:33] <jmkasunich> if the api's are the same, then motmod and motmodnew are strictly linker options
[16:15:00] <jmkasunich> if not, then we probably need to split the C sources so that one copy can be modified to use the new TP api
[16:15:08] <paul_c> Can I rely on scales & other axis data being passed via the command handler ?
[16:15:23] <jmkasunich> scales as in steps per unit?
[16:15:40] <paul_c> or units/step
[16:15:59] <jmkasunich> I wonder why the TP needs to know about steps at all
[16:16:05] <paul_c> also need cycle times for the servo loop
[16:16:40] <paul_c> TP needs to know about steps so that it has a fuzz value to work with.
[16:17:44] <jmkasunich> I thought the output of the TP was a float/double in inches or mm or whatever
[16:19:37] <jmkasunich> re: periods... emcmotController() is called with the period in ns
[16:19:39] <paul_c> if (output < 1/steps.per.unit) { /* Not an error */ }else{ /*flag error*/}
[16:20:41] <jmkasunich> and the first thing it does is compute a couple of doubles, servo_period (in seconds) and servo_freq (in hz)
[16:21:45] <jacky^> i found that cad i'm using to generate g-code insert a line at the begin of file: T# G43 H# D# (..MM END MILL), running a program it stop with a msg: D word with no G41 or G42, i think should i change the line format to: T# G43 H# G41 #D1 (..MM END MILL), should be right ?
[16:21:58] <jmkasunich> as far as steps... right now the float -> discrete conversion is done in the stepgen module (for steppers) or the encoder feedback module (for servos)
[16:22:01] <paul_c> Don't need the frequency of the servo loop, just the dT
[16:22:24] <alex_joni> jacky^: yes
[16:22:37] <jmkasunich> then don't use the freq, the code provides both, in some cases you might be able to change a divide to a multiply using freq
[16:22:39] <jacky^> alex_joni: thanks
[16:23:11] <jmkasunich> back to scaling... the scale is in the ini file, so it could be passed down to the TP if needed
[16:23:50] <paul_c> along with max accel & vel. on an axis & global level ?
[16:24:00] <jmkasunich> all in the ini file
[16:24:03] <paul_c> or are they now a HAL variable ?
[16:24:57] <jmkasunich> the axis ones are read from the ini file by halcmd and used to set some HAL parameters like the stepgen accel and vel limits
[16:25:07] <jmkasunich> but the "master" value is the one in the ini
[16:25:35] <jmkasunich> (note: peteV's core_servo.hal currently defines scaling within the hal file, but that is going to change soon)
[16:25:54] <jmkasunich> oops, not core_servo.hal, motenc_motion.hal
[16:28:44] <paul_c> as long as I don't have to go pissing obout trying to find a hal pin/param for the data I need.
[16:29:28] <jmkasunich> if the data is needed in both hal and emc's core, it should be in the ini, and the hal should be the one that needs to go pissing about for it
[16:30:33] <alex_joni> * alex_joni goes pissing
[16:34:19] <paul_c> so... We can get rid of great gobs of shared memory, and if people want the "old" tp, the stack can be in kernel memory ?
[16:34:40] <jmkasunich> the 200 entry motion queue you mean?
[16:35:02] <paul_c> That is the tc stack, but, yes.
[16:35:32] <jmkasunich> I'm not sure about the terminology....
[16:35:39] <jmkasunich> /* append it to the emcmotDebug->queue */
[16:35:40] <jmkasunich> tpSetId(&emcmotDebug->queue, emcmotCommand->id);
[16:35:44] <paul_c> use tp/tc stack
[16:35:48] <jmkasunich> is that what you are talking about?
[16:36:06] <jmkasunich> (from command handler for EMCMOT_SET_LINE)
[16:36:21] <alex_joni> it's that tcqAdd & co, iirc
[16:36:55] <alex_joni> /* TC_QUEUE_STRUCT definitions */
[16:36:55] <alex_joni> int tcqCreate(TC_QUEUE_STRUCT * tcq
[16:37:01] <paul_c> doesn't matter where the pointer is stored. It can be in shared memory, or a static var in emcmot.
[16:37:02] <jmkasunich> anyway, I see no reason why emcmotDebug->queue can't be moved from shared memory to kernel space
[16:38:05] <jmkasunich> oh, I missed a line... the actual write to the queue is:
[16:38:08] <jmkasunich> if (-1 == tpAddLine(&emcmotDebug->queue, emcmotCommand->pos)) {
[16:39:42] <dave-e> jmk awake??
[16:39:47] <paul_c> hey Dave - Have you been bitchin' on the dev list yet ?
[16:40:08] <jmkasunich> hi dave
[16:40:10] <dave-e> no not yet I have to have something to bitch about first
[16:40:17] <dave-e> hi jmk
[16:40:39] <dave-e> I finally got the .hal files is the right place...
[16:40:39] <paul_c> that bug ain't going to get fixed until you do...
[16:41:10] <dave-e> but system fires up and dies because motmod is not in /proc/modules
[16:41:48] <jmkasunich> dave: are you using a BDI or something else?
[16:42:00] <dave-e> 4.2
[16:42:11] <dave-e> with current cvs
[16:42:17] <dave-e> like yesterday
[16:42:41] <jmkasunich> can you give me more details? what exactly do you mean by "system fires up and dies"
[16:43:16] <alex_joni> motmod shouldn't be in /proc/modules
[16:43:16] <dave-e> let me go get the actual text or as close as I can remember across the room
[16:43:23] <dave-e> brb
[16:43:25] <paul_c> * paul_c is away: "Got other things to do."
[16:45:47] <jmkasunich> looks like adding the todo tags to iniaxis.cc busted things
[16:46:04] <dave-e> OK...the .hal files and ini are from Ray
[16:46:15] <CIA-12> 03alex_joni * 10emc2/src/hal/hal_lib.c: added locking check to hal_lib functions
[16:46:24] <dave-e> 'Module motmod does not exist in /proc/modules'
[16:46:45] <dave-e> 'modinfo: counld not find module motmod.o
[16:46:49] <dave-e> could not
[16:47:04] <jmkasunich> when do you see this? when you run emc.run?
[16:47:13] <dave-e> yes
[16:47:26] <dave-e> I get starting emc and then the error messages
[16:47:34] <cradek> I bet the actual error is earlier
[16:47:45] <jmkasunich> usually emc.run prints lots of stuff
[16:47:52] <alex_joni> dave-e: it's actually pretty simple
[16:47:57] <alex_joni> you need to edit the ini
[16:48:01] <alex_joni> and look for motmod.o
[16:48:05] <alex_joni> it should be motmod
[16:48:16] <dave-e> OK
[16:48:18] <alex_joni> on bdi-4.2 there should be a motmod.ko
[16:48:21] <jmkasunich> oh, that... the .o vs .ko thing
[16:48:40] <alex_joni> but emc.run should figure that out by itself
[16:48:43] <jmkasunich> where did you get your ini from? that was fixed in cvs weeks ago
[16:48:49] <alex_joni> given that motmod is in the ini
[16:48:56] <alex_joni> [19:44] <dave-e> OK...the .hal files and ini are from Ray
[16:49:25] <dave-e> hmmmm got to run for awhile be back in awhile...and we'll sort this out
[16:49:54] <alex_joni> ok
[17:04:04] <CIA-12> 03alex_joni * 10emc2/src/emc/ (ini/iniaxis.cc task/iotaskintf.cc): fixed some small problems with doxygen tags inside other comments
[17:06:21] <CIA-12> 03alex_joni * 10emc2/src/hal/hal.h: added HAL_LOCK_RUN to lock running/stopping of the HAL threads
[17:09:10] <alex_joni> alex_joni is now known as aj_eating
[17:09:49] <jmkasunich> builds on BDI-4.20, I'll kick off the compile farm and see what happens
[17:15:23] <anonimasu> *yawns*
[17:18:18] <aj_eating> anonimasu: stop yawning and do smthg usefull ;)
[17:18:25] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: added an include that is needed for BDI-2.xx (I'm surprised it didn't cause at least warnings on other systems)
[17:18:37] <aj_eating> like me ;)
[17:18:47] <aj_eating> * aj_eating eats :))
[17:22:23] <robin_sz> meep?
[17:22:26] <anonimasu> aj_eating: I'll slack a bit go and buy a drink then code some
[17:22:41] <aj_eating> don't drink & code ;)
[17:22:50] <aj_eating> unless it's highly caffeined
[17:23:04] <anonimasu> fortunately that is the case
[17:23:13] <aj_eating> nice
[17:23:20] <anonimasu> I
[17:23:22] <robin_sz> are you suggesting that all programmers are caffeine addicts?
[17:23:24] <aj_eating> * aj_eating is running out of coke
[17:23:28] <anonimasu> I've slept 2 hours earlier :)
[17:23:42] <aj_eating> should be enough
[17:23:53] <aj_eating> :P
[17:24:40] <robin_sz> * robin_sz returns from sticking his nice new sign on the wall
[17:24:49] <aj_eating> got pics?
[17:24:54] <robin_sz> yeah
[17:24:58] <robin_sz> on the phone ...
[17:25:12] <aj_eating> pics on the phone?
[17:25:15] <aj_eating> wow ;)
[17:25:26] <robin_sz> dont all phones have cameras these days?
[17:25:38] <aj_eating> ahhh.. you call those pics ?
[17:25:50] <aj_eating> I call those waste of money
[17:25:51] <robin_sz> of a sort
[17:26:05] <robin_sz> shrug, they are free
[17:26:12] <aj_eating> not really
[17:26:13] <aj_eating> ;)
[17:26:19] <aj_eating> you do end up paying for them
[17:26:23] <robin_sz> arent they?
[17:26:26] <robin_sz> how?
[17:26:28] <aj_eating> just you don't notice
[17:26:34] <aj_eating> carrier costs
[17:26:55] <robin_sz> mmm, not sure I understand
[17:27:06] <aj_eating> well.. you pay for calls, don't you?
[17:27:11] <robin_sz> yes
[17:27:26] <aj_eating> you know how much profit a big company usually has?
[17:27:35] <robin_sz> billions
[17:27:38] <aj_eating> 6-10% is pretty ok
[17:27:44] <aj_eating> I meant %wise
[17:27:54] <robin_sz> ok, well, yes
[17:28:16] <aj_eating> don't know about uk, but over here the local cellphone operators have about 50-55% profit
[17:28:21] <aj_eating> I call that stealing
[17:28:21] <robin_sz> basically, ALL phones have cameras these days
[17:28:42] <robin_sz> they HOPE you will send them as multimedia messages to people at 0.50GBP a time
[17:28:44] <aj_eating> well.. most do
[17:29:02] <robin_sz> I just d/l by infra read to the slaptop
[17:29:42] <aj_eating> usually they take some moneay back even from calls you have
[17:29:47] <aj_eating> not only MMS�s
[17:29:49] <aj_eating> MMS's
[17:29:54] <robin_sz> shrug
[17:30:22] <robin_sz> well, its the same price regardless of whether your phone has a camera or not
[17:30:28] <robin_sz> for calls
[17:30:36] <robin_sz> and the phoens are "free" anyway
[17:30:57] <anonimasu> hehe
[17:31:00] <anonimasu> got pictures?
[17:31:11] <robin_sz> yeah, I need to boot into windows though ...
[17:31:16] <aj_eating> maybe robin will send you a MMS
[17:31:18] <aj_eating> :))
[17:31:22] <robin_sz> linux doesnt work well with phones
[17:32:30] <anonimasu> aj_eating: I think that would cost him more then the servos he found at ebay ;)
[17:32:58] <aj_eating> aj_eating is now known as aj_feasting
[17:42:24] <dave-e> is jmk still on?
[17:43:38] <aj_feasting> aj_feasting is now known as alex_joni
[17:43:45] <alex_joni> * alex_joni is still around
[17:43:50] <anonimasu> hm..
[17:43:51] <alex_joni> dave-e: sorted that motmod ?
[17:43:57] <anonimasu> alex_joni: where do you find rcslib in emc2?
[17:44:06] <anonimasu> or that -I$(somthing)
[17:44:11] <anonimasu> s/that/what
[17:44:15] <alex_joni> rcslib in emc2???
[17:44:17] <dave-e> instead of the ini from Ray I edited the generic.
[17:44:31] <alex_joni> dave-e: hope that is emc.ini not generic.ini
[17:44:40] <alex_joni> there is no generic.ini in emc2 afaik
[17:44:51] <alex_joni> only in the emc on the BDI, but that's not emc2
[17:44:55] <dave-e> no there is no emc.ini in emc2
[17:44:58] <anonimasu> alex_joni: or whatever you use for nml communication
[17:45:04] <anonimasu> yes there is a emc.ini
[17:45:20] <rayh> Hi Dave.
[17:45:30] <dave-e> well I didn't find it on a 4.2 with yesterdays cvs
[17:45:31] <rayh> You fussing about the stuff I sent you?
[17:45:33] <dave-e> hi ray
[17:45:37] <anonimasu> dave-e: I recompiled emc2 yesterday it's there
[17:45:45] <anonimasu> configs/emc.ini
[17:45:47] <alex_joni> dave-e: you sure you're not using the BDI4 tree ?
[17:45:52] <dave-e> yes, apparently it ini is out dated.
[17:45:57] <alex_joni> anonimasu: there's libnml
[17:46:22] <dave-e> there was no configs...i had to create one and stuff in the .hal files
[17:46:48] <alex_joni> dave-e: can you paste the cvs command you used?
[17:46:52] <rayh> Okay Dave. You don't have the head of emc2
[17:47:10] <dave-e> I used the stuff off the wiki
[17:47:25] <alex_joni> that's BDI4 stuff, not emc2
[17:47:25] <rayh> That is BDI not EMC2
[17:48:13] <alex_joni> use:
[17:48:14] <alex_joni> cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc login
[17:48:41] <alex_joni> cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc checkout -d emc2 emc2
[17:49:13] <dave-e> cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc checkout -Pr bdi-4 -d emc emc2
[17:49:31] <alex_joni> notice the -r bdi-4 in there?
[17:49:46] <rayh> But you will want to put that in separate places in your directory structure.
[17:49:50] <alex_joni> that's the bdi-4 branch paul_c has set up to accomodate the emc sources
[17:50:11] <dave-e> OK so pull that and start over
[17:50:18] <alex_joni> dave-e: right
[17:50:20] <alex_joni> cd ~
[17:50:27] <alex_joni> mkdir emc2HEAD
[17:50:32] <alex_joni> cd emc2HEAD
[17:50:39] <alex_joni> then the cvs -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc checkout -d emc2 emc2
[17:51:47] <dave-e> sound good to me...be back later today...maybe...trying to pack for leaving on tues...with wire, tools, computer, etc for fest
[17:51:57] <dave-e> 3K Km to go.
[17:52:19] <alex_joni> ouch :)
[17:52:34] <dave-e> 3 days plus
[17:53:01] <dave-e> thanks for the help
[17:53:19] <alex_joni> np
[17:53:35] <anonimasu> hm..
[17:53:37] <anonimasu> makefile mess
[17:53:43] <alex_joni> anonimasu: where?
[17:54:11] <anonimasu> in my emc2 ;)
[17:54:28] <alex_joni> heh
[17:54:39] <anonimasu> trying to find libemc and all stuff..
[17:55:01] <anonimasu> if it is there.
[17:56:45] <anonimasu> compiling it under the emc1 branch seems nicer atleast I'll get some code donw
[17:56:47] <anonimasu> done
[17:58:07] <anonimasu> atleast for now
[18:00:16] <jacky^> Cutter gouging with cutter radius comp, machine is stopped, anyone know what it mean ?
[18:00:37] <jmkasunich> you tried to use cutter comp on a sharp inside corner
[18:00:52] <alex_joni> yes, or tried to move to a location that is inside the cutter comp.
[18:00:54] <jmkasunich> or a curved inside corner with radius less than tool radius]
[18:01:07] <alex_joni> say you were in 0,0 turned cutter comp, then you tried to move to 0,0
[18:01:17] <jacky^> thanks
[18:01:25] <jacky^> i'm confused
[18:01:34] <alex_joni> jacky^: we all are
[18:01:35] <anonimasu> alex_joni: that shouldnt break it or should it?
[18:01:41] <anonimasu> ;)
[18:01:50] <alex_joni> anonimasu: from what I witnessed it does
[18:01:51] <jmkasunich> cuttern comp is definitely confusing sometimes
[18:02:04] <alex_joni> you should move to a position outside the cutter diam
[18:02:08] <jacky^> it seem stop at line: N60 M03
[18:02:10] <alex_joni> then you can move back to 0,0
[18:02:25] <jacky^> this line become red
[18:02:33] <alex_joni> hmmm.. isn't M03 end ?
[18:02:44] <jacky^> alex_joni: manually move ?
[18:04:10] <anonimasu> m30
[18:04:15] <anonimasu> :)
[18:04:17] <jacky^> if i don't disturb too much, i can paste the few lines
[18:04:33] <alex_joni> ok
[18:05:00] <jacky^> or.. mybe have a better solution, sorry..
[18:05:49] <rayh> With tool diameter comp, I tend to run it first time with a 0.0 diameter tool
[18:06:14] <rayh> that gets past the problems and lets me see finished size
[18:06:20] <jacky^> alex here : http://www.rafb.net/paste/results/M3GkfX59.html
[18:06:20] <rayh> and no cutter comp errors.
[18:06:22] <anonimasu> hm my first qt app compiles
[18:06:27] <anonimasu> I think
[18:06:36] <rayh> Next plot with tool diameter and watch.
[18:06:45] <jacky^> its only the first part of program i'm running
[18:06:54] <jacky^> i think its enough
[18:08:14] <jacky^> program stopped at line 7
[18:08:29] <jacky^> N60
[18:14:49] <LawrenceG> has anyone got gplot ( from the axis package ) running? I am running bid-4.20 and gplot is looking for librcs.so which only exists under emc1 builds
[18:15:33] <alex_joni> LawrenceG: hello
[18:15:38] <alex_joni> didn't see you in a while
[18:15:53] <alex_joni> I did some work to make axis run on emc2
[18:16:00] <alex_joni> check axis.unpy.net
[18:16:13] <alex_joni> can't recall if gplot was included or not
[18:16:15] <alex_joni> lemme check
[18:16:29] <LawrenceG> Hi alex, been pretty busy
[18:17:08] <LawrenceG> I tried loading the deb.... it seems to be missing librcs.so as a dependancy
[18:17:27] <LawrenceG> or it could be included in the package
[18:17:40] <alex_joni> there is no librcs.so in emc2
[18:17:53] <LawrenceG> or bdi-4.20...
[18:18:08] <alex_joni> try compiling axis from scratch using the patch on axis.unpy.net
[18:18:14] <LawrenceG> lg@speedy3:~ $ gplot
[18:18:14] <LawrenceG> Traceback (most recent call last):
[18:18:14] <LawrenceG> File "/usr/bin/gplot", line 42, in ?
[18:18:14] <LawrenceG> import emc
[18:18:14] <LawrenceG> ImportError: librcs.so: cannot open shared object file: No such file or directory
[18:18:24] <LawrenceG> is the error I get
[18:19:21] <LawrenceG> can you type "gplot somefile.ngc" ? and have it run
[18:21:09] <LawrenceG> I am working on getting qcad -> dxf -> gcode -> emc path running on linux and need a gcode viewer to see if my conversion software is making any sense....
[18:21:29] <alex_joni> I don't have an installed gplot
[18:21:54] <anonimasu> my first qt app compiled :)
[18:21:57] <anonimasu> neat
[18:22:17] <LawrenceG> I may have to do a full emc2 build and try the axis patches...
[18:22:37] <anonimasu> :)
[18:25:21] <cradek> LawrenceG: gplot is defunct. Just load your file in axis to see it.
[18:26:11] <jacky^> rayh: it work with a tool 0 ! thanks
[18:26:49] <jacky^> 0.0 diameter
[18:26:52] <rayh> Welcome. Watch the plot again and see where if goes into an inside corner.
[18:27:11] <rayh> That is where the error is.
[18:27:19] <LawrenceG> Chris: that works, but running the rt kernel really slows down the deveopment box... was hoping for a non rt solution
[18:27:37] <jacky^> :-)
[18:28:17] <cradek> LawrenceG: with emc1, rt is not required. This is an emc2 problem.
[18:28:30] <jmkasunich> LawranceG: running a RT kernel shouldn't slow down the box, unless you are actually running RT code
[18:28:48] <cradek> LawrenceG: if you want to resurrect gplot, we will happily accept patches.
[18:29:22] <cradek> jmkasunich: to run axis, emc has to be running, since it uses the interpreter to generate the preview
[18:29:57] <jmkasunich> right... but I haven't seen emc2 slow down a computer unless period is set too fast
[18:30:05] <cradek> jmkasunich: agreed.
[18:30:17] <jmkasunich> if he's getting a noticable slowdown, then something isn't right
[18:30:21] <cradek> jmkasunich: and if you only want the preview, you could set PERIOD pretty slow.
[18:30:32] <LawrenceG> ok... emc2 it is
[18:30:34] <cradek> or, use emc1/sim
[18:31:47] <LawrenceG> anyone have librcs.so for a 2.6.10 pentium build they could email me?
[18:32:32] <cradek> even with a working emc1, gplot doesn't work
[18:32:39] <cradek> like I said, gplot is defunct.
[18:32:46] <alex_joni> cradek: it worked a while back ;)
[18:32:51] <LawrenceG> ok, cancel that thought
[18:32:56] <CIA-12> 03jmkasunich * 10emc2/src/hal/hal_lib.c: removed locking from hal_init() and hal_exit()
[18:32:59] <cradek> [chris@max chris]$ gplot
[18:32:59] <cradek> Traceback (most recent call last):
[18:32:59] <cradek> File "/usr/bin/gplot", line 38, in ?
[18:32:59] <cradek> from rs274 import execute_code, ArcsToSegmentsMixin, parse_lines
[18:32:59] <cradek> ImportError: cannot import name execute_code
[18:33:10] <alex_joni> same here on emc2 ;)
[18:33:13] <robin_sz> alex_joni: you wanted a sign picture?
[18:33:22] <alex_joni> sure
[18:33:25] <robin_sz> http://www.redpoint.org.uk/photos/upload/misc/Image071.jpg
[18:33:26] <cradek> gplot was never updated when we ditched the homemade gcode parser
[18:33:34] <cradek> (at axis 1.0)
[18:33:43] <jmkasunich> nice
[18:34:05] <anonimasu> neat!
[18:34:15] <robin_sz> ta :)
[18:34:20] <robin_sz> should have made it bigger
[18:34:26] <alex_joni> nice
[18:34:32] <robin_sz> it looked big inside ... but not so big onthe wall
[18:34:44] <alex_joni> seems robin_sz has still got it
[18:35:02] <robin_sz> *it* ?
[18:35:06] <anonimasu> * anonimasu kicks the QT developers
[18:35:10] <LawrenceG> Chris... so the fastest way to get axis up and plotting is to get and build emc2 and the latest axis code and emc2 patch?
[18:35:13] <alex_joni> the ability to produce nice stuff ;)
[18:35:16] <alex_joni> *g*
[18:35:19] <robin_sz> heh, thanks ;0
[18:35:27] <cradek> LawrenceG: do you have working realtime or not?
[18:35:39] <LawrenceG> yes 2.6.10 adeos
[18:35:44] <anonimasu> why do they have to use UpperCaseAndLowerCaseLikeMad
[18:35:51] <cradek> LawrenceG: then that might be the best way
[18:35:57] <alex_joni> LawrenceG: you have bdi-4.20 ?
[18:35:59] <cradek> LawrenceG: I only use it with emc1
[18:36:06] <alex_joni> axis should be included on the BDI
[18:36:21] <alex_joni> so you shouldn't need to do neither of the steps if you don't want the bleeding edge
[18:36:34] <alex_joni> just change tkemc to axis, and you should be set
[18:36:39] <robin_sz> ooh, theres a picture of my winding a toroid too
[18:36:50] <robin_sz> http://www.redpoint.org.uk/photos/upload/misc/Image049.jpg
[18:37:28] <alex_joni> that doesn't count.. it's not manual :D
[18:37:33] <robin_sz> heh
[18:37:42] <alex_joni> got a pic of yourself?
[18:37:45] <robin_sz> ahh .. and my 3 axis router making handgrips ...
[18:37:51] <robin_sz> http://www.redpoint.org.uk/photos/upload/misc/Image046.jpg
[18:37:58] <LawrenceG> I have tried to build axis several times... no luck... the axis on the bdi-4.20 had some display problems here and I was trying to build from source so I could debug it.... Nothing quite works as it should...
[18:38:14] <cradek> what display problems?
[18:38:18] <robin_sz> finished grip just on the left hand side .. looks like an apple core :)
[18:38:34] <alex_joni> LawrenceG: you probably will have problems compiling axis on the BDI emc
[18:38:42] <jmkasunich> precision apple eating?
[18:38:44] <alex_joni> that's why I worked to get it working for emc2
[18:38:50] <robin_sz> heh,
[18:38:57] <anonimasu> :)
[18:39:19] <cradek> in my opinion, the BDI emc is an unsupported dead end.
[18:40:03] <cradek> unfortunately.
[18:40:38] <LawrenceG> ok.. that helps .... there seem to be too many choices at times... will try the stock emc2 build.
[18:40:38] <robin_sz> I fear you might be right ...
[18:40:39] <anonimasu> * anonimasu made a pushbutton
[18:40:40] <anonimasu> yey!
[18:41:04] <cradek> LawrenceG: yeah that is probably your only option. Your BDI can't even build emc1 so your hands are tied.
[18:41:27] <anonimasu> well, emc2 is great :)
[18:41:35] <cradek> LawrenceG: however, if there are problems with the display, it's probably your OpenGL setup and not axis at fault
[18:42:17] <cradek> LawrenceG: on the other (third?) hand, there are lots of bugfixes in axis cvs that you don't have.
[18:42:39] <cradek> can you run another opengl program like glxgears?
[18:42:55] <robin_sz> is axis available as a .deb ??
[18:43:09] <cradek> robin_sz: not from the official channels
[18:43:12] <LawrenceG> it could be my display... I only have one box here that has enough horsepower to run axis... runnning a matrox g400 display card that should be pretty well supported by the gl driver...
[18:43:18] <cradek> robin_sz: don't know if paul has debbed it or not.
[18:44:07] <robin_sz> cradek: I suspect, eventually, we will get to a situation of letting users do a standard, debian install, adding a line to their sources.list and apt-get install emc axis foo
[18:44:08] <cradek> LawrenceG: can you give any more details about what problem it has?
[18:44:24] <LawrenceG> glxgears runs... about 4161 frames in 5.0 seconds = 832.200 FPS
[18:44:36] <cradek> robin_sz: that would be nice.
[18:44:48] <alex_joni> robin_sz: axis is debbed on the BDI
[18:44:51] <cradek> LawrenceG: that's good news
[18:44:56] <alex_joni> python-axis iirc
[18:45:17] <alex_joni> robin_sz: that works now
[18:45:24] <alex_joni> that apt-get install emc
[18:45:31] <cradek> IMO it's not the axis build that's difficult - it's getting your python setup right in the first place
[18:45:33] <robin_sz> alex_joni: right. thats a very good thing then
[18:45:36] <alex_joni> with paul_c's deb repository
[18:45:41] <robin_sz> right
[18:45:48] <alex_joni> it even pulls a rtai patched kernel
[18:45:52] <robin_sz> excellent
[18:45:53] <alex_joni> so it's pretty stright forward
[18:45:57] <alex_joni> yes
[18:45:58] <robin_sz> thats awesome work
[18:45:59] <cradek> robin_sz: that's a terribly old version of axis though.
[18:46:05] <cradek> robin_sz: pre 1.0
[18:46:17] <alex_joni> 1.0b2
[18:47:08] <robin_sz> presumably you could get the source, patch it up and submit a nice shiny new .deb if paul_c doesnt have time?
[18:47:21] <cradek> I don't have debian
[18:47:22] <alex_joni> hmmm.. I could
[18:47:27] <alex_joni> I have debian
[18:47:34] <alex_joni> but I don't know about debs
[18:47:52] <alex_joni> but an emc2.deb along with an axis.deb would be great
[18:48:01] <robin_sz> you just do apt-get source install axis (I think)
[18:48:07] <cradek> I have no desire to work on the dead end that is emc/BDI4
[18:48:18] <cradek> and that's what the debian is for.
[18:48:36] <alex_joni> cradek: one day emc2 will be on the BDI
[18:49:18] <robin_sz> well, I see debian as the one distro we could support
[18:49:19] <cradek> alex_joni: of course, and I hope it has a modern axis
[18:49:30] <robin_sz> redhat, gentoo etc are all just too hard and fragmented
[18:50:46] <cradek> supporting debian has the advantage of them going years without a new release
[18:50:58] <robin_sz> well,
[18:51:05] <robin_sz> they just released sarge :)
[18:51:09] <cradek> hmm.
[18:51:19] <robin_sz> went stable last week
[18:51:28] <cradek> how long was it since the last one?
[18:51:30] <robin_sz> but yes, it does tend to be ages
[18:51:40] <robin_sz> 2 years I think
[18:51:42] <cradek> (seriously, I'm not being sarcastic here)
[18:52:08] <robin_sz> stable is genrally on a 18month/2 year lifecycle
[18:52:20] <robin_sz> testing is pretty OK for non mission critical stuff
[18:52:48] <LawrenceG> I am running ubuntu.... very nice debian distro... bdi-4 was a piece of cake to install
[18:53:03] <cradek> ubuntu is sure catching on
[18:53:06] <robin_sz> ubuntu is debian derived,
[18:53:27] <robin_sz> not sure what advantages it ha over debain
[18:53:44] <cradek> pictures of attractive people on the CD
[18:53:52] <jmkasunich> I think they strive for a faster update cycle
[18:54:00] <LawrenceG> one needs a high speed network connection... the focus is installing from the network although cdrom images can be used.
[18:54:09] <robin_sz> they dont offer a full security updates service though, like debian stable does
[18:54:15] <jmkasunich> "give the people what they want", in this case new shiny baubles to play with
[18:54:31] <cradek> Ubuntu is released regularly and predictably; a new release is made every six months. You can use the current stable release or the current development release. Each release is supported with security updates for at least 18 months.
[18:54:37] <CIA-12> 03alex_joni * 10emc2/src/hal/utils/halcmd.c: added lock/unlock support to halcmd
[18:54:43] <robin_sz> right
[18:54:56] <robin_sz> anyway, I'm with debian and staying there :)
[18:55:17] <cradek> same with me and redhat (for desktops) and freebsd (for servers)
[18:55:36] <alex_joni> I was a big SuSE fan
[18:55:39] <robin_sz> I lost it with redhat sometime ago ..
[18:55:51] <LawrenceG> I was a redhat fan until they messed that up by dividing their forces...
[18:55:55] <robin_sz> the 7.2/7.3/8.0 etc time
[18:56:00] <alex_joni> but SuSE isn't the way to go anymore
[18:56:05] <alex_joni> since Novell took it over
[18:56:11] <robin_sz> * robin_sz nods
[18:56:13] <cradek> I'm still using redhat 9 everywhere - a great release
[18:56:24] <robin_sz> debian is my current bunny
[18:56:25] <anonimasu> hm, slack 7.2 was the greatest..
[18:56:27] <anonimasu> :D
[18:56:39] <anonimasu> it was the last non bloated release..
[18:56:54] <robin_sz> I have a set of dedrat 5.1 disks, my first ever linux :)
[18:57:06] <LawrenceG> rh 8 was easier to get working on hardware than rh 9... in 9 they broke a bunch of network and display driver...
[18:57:48] <cradek> robin_sz: you're a young pup then
[18:58:12] <robin_sz> 42!
[18:58:38] <anonimasu> I have redhat 4 on disc somwhere
[18:59:12] <cradek> http://en.wikipedia.org/wiki/Yggdrasil_Linux
[18:59:19] <cradek> my first installed linux distribution
[18:59:35] <alex_joni> I started with RH52
[18:59:38] <robin_sz> now, where do I kow that name from
[18:59:40] <cradek> ... on a 150MB ESDI disk
[18:59:59] <cradek> 4M of RAM on my 386 of course
[19:00:20] <LawrenceG> has any one installed built emc2/axis successfully using wiki pages as a guide? That will be todays task...
[19:00:31] <alex_joni> I did ;)
[19:00:37] <cradek> I think only alex_joni has built it, ever
[19:00:49] <alex_joni> heh
[19:00:57] <cradek> alex_joni: I'm serious :-)
[19:01:00] <robin_sz> no, I built it too ...
[19:01:03] <alex_joni> cradek: I think weyland or another one has too
[19:01:09] <alex_joni> was it fenn?
[19:01:12] <alex_joni> might have been
[19:01:18] <cradek> alex_joni: with your patch?
[19:01:40] <LawrenceG> I think fenn was having trouble with axis build... wiki page is incomplet
[19:01:56] <cradek> **sigh**
[19:01:58] <alex_joni> yeah
[19:03:28] <LawrenceG> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CompilingEMC2
[19:10:26] <robin_sz> fenn was talking about moving the whole wiki to some other random wiki package
[19:10:44] <alex_joni> pukifoo
[19:10:54] <robin_sz> someting like that
[19:11:15] <robin_sz> its published under a creative commons license, so he can if he wants
[19:11:33] <alex_joni> LawrenceG: it worked for me like described there
[19:11:56] <alex_joni> only had to use :
[19:12:13] <alex_joni> env EMC2ROOT=/path/to/emc2 sudo python setup.py install
[19:12:23] <alex_joni> to be able to copy the stuff to /usr/bin
[19:13:41] <LawrenceG> EMC2ROOT=/path/to/emc2 is what on your system... source dir or binary or what?
[19:14:22] <alex_joni> there is only one emc2 dir
[19:14:28] <alex_joni> source dir = binary dir
[19:14:40] <alex_joni> the dir where you checked out from CVS
[19:14:56] <LawrenceG> like ~/emc2HEAD/emc2 if I have downloaded the cvs stuff under emc2HEAD
[19:15:21] <alex_joni> right
[19:16:23] <LawrenceG> ok ... just cleaning up HD to get get rid of old axis/bdi stuff
[19:16:44] <alex_joni> well.. you still need the old rtai kernel
[19:16:50] <alex_joni> rtai patched kernel
[19:17:31] <LawrenceG> cant remember is the adeos patch an rtai kernel?
[19:18:01] <alex_joni> adeos was part of rtai
[19:18:07] <alex_joni> now they split or smthg like that
[19:18:09] <alex_joni> but yes
[19:18:22] <LawrenceG> ok... I should be able to run with that...
[19:21:22] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: extended the halcmd 'show' command to allow the user to specify a single item or group of items, rather than always printing everything
[19:22:31] <alex_joni> hehe
[19:22:37] <alex_joni> who's frobz ?
[19:22:39] <alex_joni> ;))
[19:23:12] <jmkasunich> my memory may be rusty, but that or something like it was in the old zork (or pre-zork) adventure games
[19:23:25] <alex_joni> nicee
[19:23:40] <alex_joni> * alex_joni has a task for jmk
[19:23:54] <alex_joni> there is smthg that you missed
[19:24:21] <jmkasunich> what did I miss?
[19:24:34] <alex_joni> updating the halcmd manfile ;)
[19:24:48] <jmkasunich> ewww... that is badly out of date
[19:24:51] <alex_joni> *g*
[19:24:57] <jmkasunich> as is the hal_introduction doc
[19:25:00] <rayh> jmkasunich: How soon will we have an upgrade to the halcmd part of the pdf?
[19:25:16] <rayh> for those new commands you put in there just now?
[19:25:20] <jmkasunich> when the dust settles on this round of changes
[19:25:26] <jmkasunich> (still have others to add)
[19:25:54] <jmkasunich> getp, gets, list, next (as we discussed yesterday)
[19:27:37] <jacky^> ugh ! i've seen an impressionant video of a laser cutter on cuttingedgecnc.com :\
[19:28:05] <rayh> Okay. I'll wait a bit before I try the stuff.
[19:28:19] <jmkasunich> ray: the built in help is up-to-date
[19:28:37] <jmkasunich> "help show" will tell you how to use the new show syntax
[19:29:44] <jmkasunich> question about the list command, since you are here...
[19:30:14] <jmkasunich> should it accept a "pattern" like I just added to show, or should it always print a complete list?
[19:30:28] <robin_sz> jacky^: you didnt like tha laser cutting?
[19:31:00] <jacky^> wow !
[19:31:23] <jacky^> incredible :\
[19:31:28] <robin_sz> yep
[19:32:07] <robin_sz> mine are 1500(equivaltent) and 1700watts, not 4000 ... but still quite quick
[19:32:37] <jacky^> O_O
[19:32:51] <jacky^> wowowow
[19:32:55] <jacky^> too cool
[19:33:05] <robin_sz> too expensive ;)
[19:33:18] <jacky^> first time i seen something like that
[19:33:22] <robin_sz> that machine is close to 1m euro
[19:33:54] <jacky^> ugh !
[19:34:21] <dmess> high all ... ; )
[19:34:32] <robin_sz> you seen the "Aerostell" mill mpeg yet?
[19:34:43] <dmess> alpha available??
[19:34:54] <jacky^> robin_sz: no..
[19:34:57] <dmess> not i
[19:35:18] <dmess> i need a deburring tool..
[19:35:21] <A-L-P-H-A> ?
[19:35:31] <dmess> need to chat
[19:35:35] <jacky^> robin_sz: have a link ?
[19:37:01] <A-L-P-H-A> dmess? did you need something?
[19:37:50] <dmess> i wanna talk to ya ya
[19:38:33] <dmess> could emc1 hand a robotic deburring tool with a follow probe attatched
[19:39:05] <rayh> rayh is now known as rayh_phone
[19:39:30] <A-L-P-H-A> I don't know. emc2 probably could, if you got it (emc2) to work.
[19:39:47] <dmess> yeah so i hear..
[19:39:56] <A-L-P-H-A> I doubt emc1 could.
[19:40:08] <A-L-P-H-A> emc really needs sub routines. heh.
[19:40:37] <dmess> im thinkin cheap and simple stupud philosophy...you know aircraft stuff only..
[19:41:19] <dmess> programmed and accepted processes only
[19:41:49] <dmess> we could make a mint in this area in metal finishing
[19:42:05] <A-L-P-H-A> deburring tool, like a hook thing, that you run over it?
[19:42:41] <dmess> no a belt or air driven burr and the likes
[19:43:00] <dmess> remove all toolmarks
[19:43:17] <A-L-P-H-A> sorry, don't know.
[19:43:23] <dmess> and radius and chamfer to dwg specifications
[19:44:11] <rayh_phone> rayh_phone is now known as rayh
[19:44:20] <dmess> i think its doable..
[19:44:47] <rayh> jmkasunich: You had a question about halcmd commands a bit ago, could you repeat that.
[19:45:09] <alex_joni> [22:28] <jmkasunich> question about the list command, since you are here...
[19:45:10] <alex_joni> [22:28] <jmkasunich> should it accept a "pattern" like I just added to show, or should it always print a complete list?
[19:45:45] <dmess> i can design/make the air tool end of arm pcs but wil emc handle the pc input from a REAL cad cam program
[19:46:07] <rayh> I'll get the latest brb
[19:46:35] <anonimasu> dmess: well, it all depends on your post
[19:48:02] <dmess> ill custom buid the post..
[19:48:11] <rayh> wah. That's a lot of updating.
[19:48:25] <anonimasu> emc works nicely with output from cam programs unless you have large programs with lots of segments
[19:48:27] <alex_joni> people've been busy ;)
[19:48:42] <dmess> gonna be HUGE..
[19:48:52] <dmess> sorry.. fact of life..
[19:49:47] <robin_sz> dmess: ther is another way ...
[19:49:48] <dmess> what causes the issiu with the larger programs
[19:49:55] <anonimasu> dmess: the tp.
[19:50:09] <anonimasu> dmess: it's not the program size it's short segmented paths
[19:50:10] <alex_joni> not larger programs
[19:50:23] <alex_joni> but programs with very short segments are the problem
[19:50:24] <robin_sz> dmess: waht about, cutting the parts using a method that doesnt cause a burr in the first place?
[19:50:34] <dmess> so whats the OTHER way i havent needed in the last 23 yrs
[19:50:42] <anonimasu> robin_sz: you know of one?
[19:51:05] <robin_sz> I presume we are talking sheet aluminium here, since its aero stuff?
[19:51:09] <anonimasu> build your program for slower then HSM speeds ;)
[19:51:14] <dmess> youve never turned a sprocket or gear..
[19:51:32] <robin_sz> yeah
[19:51:35] <anonimasu> I'll be buying a 5kw motor for the mill tomorrow
[19:51:45] <robin_sz> gears are almost always finished by eltro machining
[19:51:55] <dmess> negative ALL 300m or super hy-tuf.. aermet 100
[19:52:04] <anonimasu> ~400$
[19:52:16] <anonimasu> 2860rpm one
[19:52:30] <dmess> very little aluminum ..only in the smaller regional jets..
[19:52:46] <robin_sz> right, but sheet stuff?
[19:52:59] <dmess> no forgings..
[19:53:12] <anonimasu> hm, I was under the impression they machined duraluminium.. or aircraft grade alu..
[19:53:18] <anonimasu> for airliners
[19:53:22] <jmkasunich> sounds like landing gear or something
[19:53:34] <dmess> most heat treated after iitial m/cing..
[19:54:06] <dmess> yes langing gear.. is something... its US
[19:54:26] <jmkasunich> robin was probably thinking airframe and skins
[19:54:40] <robin_sz> forgings huh
[19:54:40] <robin_sz> this is for de-flashing them I presume?
[19:54:45] <robin_sz> I was ..
[19:54:51] <robin_sz> and waterjets :)
[19:54:52] <dmess> forged steel... machined all over... finished within an inch of their life..
[19:55:24] <dmess> nope to remove kellering marks during m/c
[19:56:15] <alex_joni> nice hexapod : http://cuttingedgecnc.com/Movies/ffmm.wmv
[19:56:27] <anonimasu> I love that laser :)
[19:56:32] <jmkasunich> how do you view wmv on a linux box?
[19:56:35] <dmess> even sheet stuff is not nice..has a life of its own...
[19:56:35] <anonimasu> alex_joni: seen the toolchange?
[19:56:39] <alex_joni> yup
[19:56:43] <alex_joni> not much to see ;)
[19:56:47] <anonimasu> that's the speed I crave for ;)
[19:57:30] <dmess> i need clean and repeatable fro what i want.. speed is a bonus
[19:57:47] <anonimasu> yeah, but the toolchange speed ;)
[19:58:10] <dmess> you ever seen a CHIRON change toos on the fly??
[19:58:15] <anonimasu> no
[19:58:32] <dmess> 2.6 seconds chip to chip...
[19:58:42] <anonimasu> :)
[19:58:49] <anonimasu> now you are talking speeds
[19:58:49] <anonimasu> :)
[19:59:42] <dmess> goto 3.4" with the longer of the 2 ofsets... the one in the spindle or the one coming n and when you get ther its done.. rock and roll..
[20:00:14] <anonimasu> hehe
[20:00:24] <anonimasu> I dont need thoose kind of speeds, but I love to see it done
[20:00:28] <anonimasu> click*finished*
[20:00:59] <alex_joni> * alex_joni clicks on anonimasu
[20:01:03] <robin_sz> sigh .. got outbid on those servos
[20:01:08] <alex_joni> that didn't work.. :(
[20:01:09] <robin_sz> shame :(
[20:01:16] <alex_joni> * alex_joni double-clicks on anonimasu
[20:01:29] <alex_joni> that's better
[20:01:30] <alex_joni> :)
[20:01:33] <dmess> all the tools reside in arms around the spindle with bearings.. an air cylinder drops the whole assembly to retact tohe tool out of the spindle
[20:02:07] <anonimasu> hm, so you use the toolchanger as a retainer?
[20:02:19] <anonimasu> to keep the tool seated in the spindle?
[20:02:19] <dmess> then each tool/arm has a cylinder to push the next tool into pos'n as it retracts the next..
[20:02:42] <anonimasu> very nice
[20:02:50] <dmess> no only for t/c... had a springloaded drawbar too
[20:03:04] <anonimasu> hm ok
[20:03:23] <dmess> mine were twin spindle m/c's
[20:04:17] <dmess> 600mm apart with 160 mm cutters swapin in and out every 5 minutes or so is hard in mechanicals...
[20:05:21] <alex_joni> robin_sz: isn't it usually bad to look at a cutting laser?
[20:05:26] <dmess> siemens 310m controls..
[20:05:40] <alex_joni> especially a 4kW one ?
[20:06:10] <dmess> probably... very bad
[20:06:18] <alex_joni> right :)
[20:06:51] <dmess> DONT LOOK INTO THE LIGHT>>>>????"""" im blind
[20:07:15] <alex_joni> what light?
[20:07:19] <alex_joni> I don't see no light
[20:07:31] <anonimasu> hehe
[20:07:32] <alex_joni> now that you mention it.. why is it so dark in here ?
[20:07:35] <anonimasu> robin should know
[20:08:41] <dmess> sheite they got you too alex
[20:09:11] <alex_joni> * alex_joni is commiting blindly
[20:09:12] <dmess> it was a fly by shooting..
[20:09:32] <CIA-12> 03alex_joni * 10emc2/src/hal/utils/halcmd.c: added a status command for halcmd, right now it displays HAL locking
[20:10:24] <dmess> if hal locks her up... your daughter's safe..
[20:10:33] <alex_joni> lol
[20:10:44] <alex_joni> but.. you need to be root for locking her up
[20:11:11] <dmess> doesnt root make you swallow the key too...
[20:11:22] <alex_joni> SHA-key ?
[20:11:38] <alex_joni> now there's a thought ;)
[20:11:42] <dmess> i found just locking her up for a few wors well as GOD
[20:11:52] <alex_joni> generate a SHA key on HAL load
[20:11:57] <alex_joni> anyone with that key can unlock
[20:12:04] <alex_joni> so you gotta remember that one
[20:12:07] <alex_joni> say.. 256 bits ?
[20:12:12] <alex_joni> or more?
[20:13:23] <jmkasunich> cool, 1 trivial conflict, already fixed
[20:13:25] <dmess> she could be there forever... with that...you do want them to move out eventually..
[20:13:54] <alex_joni> jmkasunich: think this is it for me.. getting late
[20:13:59] <jmkasunich> ok
[20:14:12] <alex_joni> lemme know if there are still other issues
[20:14:23] <dmess> by then the alzyhamers and stuff... you know.. you wont remember HOW to get rid of her
[20:15:00] <alex_joni> get rid of who?
[20:15:24] <alex_joni> * alex_joni forgot
[20:16:14] <dmess> the daughter
[20:16:32] <dmess> man i cant work off you... LOL
[20:16:36] <alex_joni> heh ;)
[20:17:08] <rayh> The halcmd stuff looks good.
[20:17:24] <rayh> I can specify a pattern and get only that info back.
[20:17:36] <anonimasu> :)
[20:17:36] <anonimasu> :)
[20:17:40] <anonimasu> err my lag is horrid
[20:17:41] <jmkasunich> which brins me back the my question
[20:17:45] <dmess> me and a guy from work are gonna do a YUK-YUKS open mike nite soon...
[20:17:51] <anonimasu> I think I caught some of that alhzeimers
[20:17:55] <jmkasunich> we talked about a "list" command that would list only the names, all on one line
[20:18:09] <rayh> Right.
[20:18:13] <jmkasunich> should 'list' use patterns, or always print all names?
[20:18:23] <dmess> its cunt-ages stuff
[20:19:06] <alex_joni> what's a YUK-YUKS ?
[20:19:09] <rayh> What value would it be to list "pattern"?
[20:19:14] <alex_joni> sounds yucky
[20:19:20] <rayh> I can see list pins
[20:19:21] <dmess> comedy club
[20:19:34] <alex_joni> ahh
[20:19:34] <rayh> list sig
[20:19:37] <rayh> and such
[20:19:41] <alex_joni> list lists
[20:19:46] <jmkasunich> rayh: dunnt, that's why I'm asking.... I expect you will be using list to get all the names, so not much
[20:20:15] <alex_joni> list developers
[20:20:18] <rayh> Okay.
[20:20:21] <jmkasunich> oops... it wouldnt' be "list pattern" it would be "list [type] [pattern]"
[20:20:21] <alex_joni> that would be nice ;)
[20:20:35] <jmkasunich> for instance, "list pin axis" would list all pins starting with axis
[20:20:37] <alex_joni> list [type] would probably be enough
[20:20:40] <rayh> Then after a list pins we would ask for the type of each or such.
[20:20:43] <jmkasunich> "list pin" would list all pins
[20:20:59] <jmkasunich> well list only gives you names
[20:21:16] <jmkasunich> if you want the value, you would use "getp name"
[20:21:34] <jmkasunich> for other info (type, etc) you would use "show param name"
[20:21:53] <rayh> using something like show.
[20:21:59] <alex_joni> halcmd show message "night people"
[20:22:19] <jmkasunich> goodnight alex... thanks for all the halcmd work
[20:22:27] <alex_joni> no sweat ;)
[20:22:40] <anonimasu> night alex
[20:22:47] <alex_joni> I closed that Feature request
[20:22:54] <alex_joni> hope you don't mind John
[20:22:56] <jmkasunich> ok, thanks
[20:23:07] <jmkasunich> not at all... a closed tracker is a good tracker
[20:23:25] <alex_joni> yeah.. if it doesn't work a bug-tracker is more appropiate anyways
[20:23:39] <alex_joni> night guys
[20:23:41] <alex_joni> nice day ;)
[20:23:54] <jacky^> <hi alex_joni
[20:24:10] <jacky^> nigth
[20:31:02] <rayh> jmkasunich: The list, get sounds like a good querry bidirectional model.
[20:31:39] <jmkasunich> just about done with get... one or two more tests and I'll commit it and move on to list
[20:39:03] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: added halcmd getp and gets commands to get the value of a parameter or signal
[20:39:24] <mshaver_away> mshaver_away is now known as mshaver
[20:39:34] <mshaver> darn, just missed alex
[20:41:24] <jmkasunich> hi matt
[20:41:39] <mshaver> hi john!
[20:45:28] <paul_c> * paul_c is back
[20:45:34] <mshaver> Q: i've been poking around in the rs274ngc & task dirs whilst getting a standalone interp running & i was wondering if anyone knows why the canterp.cc file defines functions like emcTaskPlanExecute() that are normally defined in emctask.cc.
[20:45:48] <mshaver> brb in 2 mins
[20:46:48] <ValarQ> oh, halcmd updates
[20:48:08] <jacky^> hi ValarQ
[20:48:53] <paul_c> mshaver: I'd say....
[20:49:06] <jacky^> is there any file ngc ready to test backslash error ?
[20:49:10] <jacky^> hi paul_c
[20:49:41] <paul_c> rs274ngc* calls a;; the canonical commands that the canterp.cc version is parsing out.
[20:51:07] <jacky^> paul_c: are you talking to me ?
[20:51:33] <paul_c> nope ; mshaver
[20:51:38] <jacky^> ah, ok
[20:52:11] <paul_c> mshaver: most of the functions in canterp.cc are stubs.
[20:53:22] <paul_c> and verilly will it suck if/when it comes to adding support for some of the extended low level functions.
[20:53:38] <mshaver> yes, but why stub them out?
[20:53:57] <mshaver> why not just link in emctask.o?
[20:54:15] <mshaver> like we do with rs274.o...
[20:54:28] <paul_c> 'cos the build will fail if you try to link it to an executable
[20:55:10] <paul_c> If you link canterp.o and emctask.o, you will get conflicting symbols
[20:55:39] <jmkasunich> Valarq: yes, halcmd updates ;-)
[20:55:54] <jmkasunich> hopefully they will address some of the issues you had with reading hal data
[20:56:19] <mshaver> yes - I mean why not get rid of the stubs/alternate defs in canterp.cc, and then use the regular ones in emctask.cc?
[20:56:45] <jmkasunich> although I just realized that u8 still prints out in hex and decimal
[20:56:57] <jmkasunich> need to fix that for getp and gets
[20:57:20] <paul_c> mshaver: emcTaskPlanExecute() are totally different. It's one or the other.
[20:58:48] <anonimasu> hm, now lets see if I can load a svg as a icon
[20:59:18] <mshaver> i guess i'll have to read this code - but i would assume that since both interps emit canonical commands, that the emcTaskxxx functions wouldn't know/care what interp is calling them...
[20:59:45] <mshaver> or at least i would hope that they were interchangeable!
[21:00:06] <paul_c> Look to see where emcTaskPlanExecute() is called from.
[21:00:35] <mshaver> ok...
[21:02:58] <paul_c> Much of emctask.cc is a wrapper around rs274ngc functions.
[21:05:13] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: modified halcmd getp and gets command output format... decimal only (except for bits), and no leading or trailing whitespace
[21:05:55] <mshaver> emcTaskPlanExecute() is defined in emctask.cc, and called from emctaskmain.cc - it's NOT called from anything in src/emc/rs274ngc...
[21:06:46] <mshaver> the 2 interps need to be made to operate the same (except of course for the language they interpret ;) )
[21:07:19] <mshaver> you agree?
[21:07:19] <paul_c> Right - And emctask.cc is the api wrapper
[21:07:52] <robin_sz> woo hoo .. my 2m/144mhz radio works :) ...
[21:08:02] <mshaver> i'll put that on my "things to do list"...
[21:08:02] <paul_c> emctaskmain calls emctask which calls rs274ngc_*
[21:08:38] <mshaver> that brings up another thing i was thinking about earlier today
[21:09:08] <mshaver> the fact that the interp func names begin with rs274ngc...
[21:09:25] <mshaver> that's ok when you have 1 interp, but...
[21:09:30] <jmkasunich> doesn't the new interp class address that?
[21:09:42] <mshaver> no
[21:10:01] <jmkasunich> ok
[21:10:08] <jmkasunich> * jmkasunich goes back to hal ;-)
[21:10:09] <mshaver> it makes it myinterpname.rs274ngc_xxxxx ;)
[21:10:16] <jmkasunich> oops
[21:10:27] <jmkasunich> gotta fix that IMHO
[21:11:23] <paul_c> The rs274ngc_*() functions are the external, public interfaces.
[21:11:29] <mshaver> yes, & make canterp into a class like Paul did with rs274ngc (nice work too Paul, I like what you did, now that i've had a chance to dig around in there)
[21:11:50] <paul_c> The Interp:: class already isolates the namespace.
[21:12:26] <jmkasunich> but won't a STEP interpreter still have to implement methods called rs274ngc_xxx?
[21:12:40] <jmkasunich> seems they should just be called interp:xxxx
[21:12:47] <mshaver> yea, what john just said...
[21:12:50] <paul_c> We can drop the rs274ngc_ prefix
[21:14:00] <mshaver> you're right, it would work with rs274ngc, it would just look odd... , so dropping the prefix is a good idea
[21:14:07] <paul_c> BUT you would need to re-write canterp around the Interp class
[21:14:34] <jmkasunich> rewriting canterp around the class seems like the right thing to do to me
[21:14:40] <mshaver> yes, i think that's needed too
[21:14:52] <jmkasunich> any future inteps would also use the class (STEP, etc)
[21:15:18] <mshaver> that way all interps have the same api (what john just said)
[21:15:22] <jmkasunich> and the standalone interp should just invoke the class as well, so standalone canterp, etc, becomes possible
[21:16:28] <mshaver> yep, that's where i was coming from - i wanted to make a standalone version of each interp, without having to re-write the code for each one
[21:17:13] <mshaver> we can only hope that the output of a standalone canterp would exactly match its input ;)
[21:17:36] <jmkasunich> yes... and that would be a good way to test it
[21:18:22] <mshaver> :-| - TEST!... we're going to TEST this stuff!
[21:18:38] <mshaver> i thought if it compiled...
[21:19:06] <jmkasunich> well, I know it's a radical idea...
[21:19:25] <robin_sz> hes probably had too much sun
[21:19:44] <jmkasunich> * jmkasunich is hiding in the basement, it's hot and humid outside
[21:20:20] <robin_sz> best not come out, that pesky daystar is bad for you
[21:20:37] <mshaver> anyway, glad you guys concur (it's humid here too - went swimming earlier - soon have to cut grass - ugh!)
[21:20:43] <jmkasunich> I was out earlier, cutting down weeds
[21:20:52] <jmkasunich> grass was yesterday
[21:21:42] <mshaver> btw, should all the standalone interp source files go into src/emc/task, or somewhere else?
[21:25:11] <CIA-12> 03paul_c * 10emc2/src/emc/task/emccanon.cc: Remove a duplicate function (had been ifdef'd out anyway..
[21:26:21] <rayh> mshaver: You get the sputterer running?
[21:26:27] <mshaver> i see with all these commits i need to update before i start editing again - that's a good thing!
[21:27:09] <robin_sz> w00t! .. bought a vacuum pump :)
[21:27:24] <mshaver> yep, all except the rf supplies - that's tomorrow - 3p converter is not taxed at all by todd's stuff!
[21:27:51] <jmkasunich> what are you sputtering?
[21:28:53] <mshaver> mainly TiN, TiCN & (believe it or not) PTFE!
[21:29:07] <robin_sz> coo,
[21:29:14] <robin_sz> you can sputter ptfe?
[21:29:58] <robin_sz> i can only imagine that takes some doing
[21:30:15] <rayh> So the phase converter advice worked out. Glad to hear that.
[21:30:29] <mshaver> yep - I think my buddy todd is the only one who does it, but he uses a teflon target (cathode) & it's deposited as a light brown coating
[21:30:43] <robin_sz> weird
[21:30:49] <mshaver> nitrogen process gas
[21:31:05] <robin_sz> 'k
[21:31:37] <robin_sz> and the coating sticks?
[21:32:18] <mshaver> rather well actually - it's used on some valve seats, guitar strings & medical industry parts
[21:32:38] <robin_sz> ooh, teflon guitar strings :)
[21:34:14] <mshaver> rayh: converter is great! They need to do a prettier paint job, but electrically it's really nice - i'd buy one myself if i needed it
[21:37:26] <Phydbleep> Did anyone here want Crydom 240V/40A Solid State Relays ??
[21:39:16] <jacky^> hi Phydbleep :-)
[21:39:52] <Phydbleep> Hi jacky^ :)
[21:40:40] <Phydbleep> * Phydbleep is testing a stepper for useability as a jog wheel bearing/sensor.
[21:41:05] <jmkasunich> sensor? looking at the winding voltage?
[21:42:02] <Phydbleep> jmkasunich: Hehehe.. Enough to light a red led if you bump it. :)
[21:42:12] <jmkasunich> sounds interesting... might be hard to dermine the direcion of rotation tho
[21:42:46] <Phydbleep> Nope, I'm getting a really nice quadrature output. :)
[21:42:59] <robin_sz> cute
[21:43:35] <jmkasunich> just need a couple comparators to square it up, and presto... I like it
[21:43:42] <robin_sz> might even have more "feel" than a opto too
[21:43:46] <Phydbleep> 2 led's, power drill.. Plug the leds's into the connector and spin the whole thing with the cordless. :)
[21:44:17] <jmkasunich> but how does it behave when you turn it slowly (one or two steps/sec)
[21:44:22] <Phydbleep> jmkasunich: Schmidt trigger then J/K flip-flop. :)
[21:44:46] <robin_sz> ahh, yes :) .. well, theres the rub :)
[21:44:58] <Phydbleep> Just nudging the thing through 1 step will flash the leds.
[21:45:27] <CIA-12> 03jmkasunich * 10emc2/src/hal/utils/halcmd.c: added list command to halcmd... prints a list of HAL object names, primarily for use by other programs rather than by humans
[21:45:34] <paul_c> jmkasunich: gcc-2.95 -c -Wall -g -I/home/paul/EMC/emc2/include -I/usr/lib/realtime/include -I. -DULAPI -O2 -DLOCALE_DIR=\"/usr/share/locale\" -DPACKAGE=\"emc2\" -DXTHREADS -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/X11R6/include -I/usr/include/atk-1.0 -I/usr/include/pango-1.0 -I/usr/include/freetype2 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include meter.c -o /home/paul/EMC/emc2/src/.tmp/meter.o
[21:45:34] <paul_c> In file included from /usr/include/glib-2.0/glib.h:52,
[21:45:34] <paul_c> from /usr/include/gtk-2.0/gdk/gdktypes.h:32,
[21:45:34] <paul_c> from /usr/include/gtk-2.0/gdk/gdkcolor.h:4,
[21:45:35] <paul_c> from /usr/include/gtk-2.0/gdk/gdk.h:30,
[21:45:37] <paul_c> from /usr/include/gtk-2.0/gtk/gtk.h:31,
[21:45:39] <paul_c> from meter.c:59:
[21:45:45] <robin_sz> ever connected two steppers back-to-back?
[21:45:49] <robin_sz> thats kinda fun
[21:46:15] <jacky^> O_O
[21:46:30] <jacky^> back-to-back ??
[21:46:36] <Phydbleep> Yeah, It's a great way to tick off a 'know it all' too.
[21:47:00] <jacky^> hehe
[21:47:05] <jmkasunich> paul: I got the block of text, but not the actual error message
[21:47:30] <Phydbleep> jacky^: Spin one and the other will spin.. Use mis-matched sizes and you can apply more power to the input than someone can hold on the output. :)
[21:47:48] <jacky^> uhm
[21:48:09] <jacky^> interesting
[21:48:09] <CIA-12> 03paul_c * 10emc2/src/Makefile: Lost the rs274ngc_ prefix from all the public interpreter calls - This will allow canterp & other (as yet unwritten) interpreters to use a common API.
[21:48:14] <CIA-12> 03paul_c * 10emc2/src/emc/ (10 files in 2 dirs): Lost the rs274ngc_ prefix from all the public interpreter calls - This will allow canterp & other (as yet unwritten) interpreters to use a common API.
[21:48:18] <paul_c> mshaver: Hold on to your hat.
[21:49:09] <jmkasunich> hey... it even compiles!
[21:49:09] <paul_c> Tat was a quick search'n'replace on the interp.
[21:49:11] <robin_sz> I guess that will make, say, an excellon drill interp pretty easy to add in?
[21:49:16] <mshaver> all right! i was thinking, "i'll have to start on that tonight..." - guess not...
[21:49:36] <jmkasunich> paul: that paste you sent me...
[21:49:40] <paul_c> mshaver: canterp & standalone is your problem.
[21:50:00] <paul_c> jmkasunich: yes ?
[21:50:01] <jmkasunich> I suspect it's a gtk 2.0 vs. 1.2 problem, but you didn't send me the complete error message
[21:50:02] <Phydbleep> Hehehe... When you do the led's/drill trick you can let the wires/led's whip and see the quad pattern. :)
[21:50:15] <mshaver> yes sir! i'll look at it tonight after grass mowing!
[21:50:28] <robin_sz> Phydbleep: hehe.
[21:50:33] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:431: warning: `visibility' attribute directive ignored
[21:50:33] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:432: warning: `visibility' attribute directive ignored
[21:50:33] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:433: warning: `visibility' attribute directive ignored
[21:50:34] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:434: warning: `visibility' attribute directive ignored
[21:50:34] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:435: warning: `visibility' attribute directive ignored
[21:50:34] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:436: warning: `visibility' attribute directive ignored
[21:50:36] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:437: warning: `visibility' attribute directive ignored
[21:50:38] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:438: warning: `visibility' attribute directive ignored
[21:50:40] <paul_c> /usr/include/glib-2.0/gobject/gtype.h:439: warning: `visibility' attribute directive ignored
[21:50:51] <jmkasunich> all warnings then, no errors?
[21:51:22] <paul_c> It's your code that isn't compiling cleanly.
[21:51:44] <jmkasunich> actually it's GTK code that isn't compiling cleanly
[21:52:19] <jmkasunich> sounds like I need to apt-get gtk 2 and test it... works fine under GTK 1.2
[21:52:47] <paul_c> All the Gtk hal stuff generates the same ream of warnings.
[21:53:10] <paul_c> scope, meter, etc.
[21:53:29] <jmkasunich> yeah.. and I'd bet a program consisting of #include <gtk.h> would do the same
[21:54:28] <jmkasunich> well, I can't duplicate it here until I get GTK 2, and right now it's dinner time
[21:54:34] <jmkasunich> jmkasunich is now known as jmk_food
[21:54:54] <robin_sz> <homer>can;t code now, eating</homer>
[21:56:11] <robin_sz> mshaver: I presume this latest pluggable interp will help with things like excellon drill fiels and the like?
[21:58:49] <mshaver> robin_sz: yes an excellon interp would be easier to write
[21:59:07] <robin_sz> thats kewl, I can see great potential for this on future
[21:59:14] <robin_sz> not necc. for excellon
[21:59:20] <robin_sz> but various personality modules
[21:59:20] <jacky^> Phydbleep: i've a lot of display's 7 segment here, could i build something of nice to use for the cnc machine ?
[21:59:22] <paul_c> jmk_food: Yes. Just including gtk.h generates the same warnings.
[22:09:50] <CIA-12> 03paul_c * 10emc2/src/emc/task/emctask.cc: Convert rs274ngc error printing to a more generic form.
[22:24:45] <anonimasu> hmnight guys
[22:25:53] <Timbo> anybody got any recommendations for CAM packages on linux?
[22:26:58] <mshaver> mshaver is now known as mshaver_cutting_
[22:27:41] <mshaver_cutting_> mshaver_cutting_ is now known as mshaver_yard_wk
[22:28:27] <LawrenceG> qcad has one but it costs... not sure what it is about.. http://www.ribbonsoft.com/camexpert.html
[22:28:48] <Timbo> i had a look at the demo
[22:28:59] <Timbo> it doesn't seem particularly sophisticated
[22:29:03] <LawrenceG> I'm still trying to determine if qcad will work for my aclt replacement
[22:29:17] <Timbo> more or less straight dxf->toolpath conversionn
[22:29:19] <Timbo> -n
[22:29:45] <LawrenceG> are you running emc2?
[22:29:49] <Phydbleep> jacky^: You could build an entire DRO (digital readout) from TTL chip..
[22:29:59] <Timbo> no, not yet
[22:30:00] <paul_c> Do you need full 3D parametric modelling or just 2.5D ?
[22:30:08] <Timbo> 2.5D
[22:30:11] <Phydbleep> jacky^: But it would be the size of a pc case.:)
[22:30:48] <jacky^> ugh ! :-)
[22:31:00] <jacky^> i'm lookin at http://www.discovercircuits.com
[22:31:07] <CIA-12> 03paul_c * 10emc2/src/emc/kinematics/tp.c: Add a comment to a FIX-ME.
[22:32:08] <LawrenceG> paul... just built emc2 from cvs... it seemed to go ok but having problems starting... any ideas what RTAPI: ERROR: could not open shared memory points to?
[22:32:26] <LawrenceG> permissions or missing device or ?
[22:32:51] <paul_c> Are you running udev ?
[22:33:11] <LawrenceG> good question... is that part of ubuntu?
[22:33:42] <paul_c> check the contents of /dev
[22:34:01] <Timbo> ubuntu is udev yes
[22:35:05] <Phydbleep> jacky^: This is what I'm looking at for readout on my machine. http://www.eio.com/g321skel.htm
[22:35:30] <paul_c> LawrenceG: ls /dev/rtai_shm
[22:35:42] <jacky^> Phydbleep: wow ! cool
[22:36:09] <LawrenceG> ls: /dev/rtai_shm: No such file or directory
[22:36:34] <Phydbleep> jacky^: A G321 with the backlight is $50.
[22:36:43] <asdfqwega> Phydbleep: Hm, name looks familiar...
[22:37:03] <Phydbleep> asdfqwega: Shhhh. :)
[22:37:15] <jacky^> Phydbleep: nice, i was thinking more..
[22:37:36] <paul_c> LawrenceG: You need to pick on jmk_food and find out why emc2 doesn't support udev like the bdi-4 branch does.
[22:38:07] <asdfqwega> jacky^: Get several million LEDs and make a large colour display
[22:38:31] <jacky^> asdfqwega: i know..
[22:39:08] <Phydbleep> jacky^: For led's I'd use something like http://led-displays-leds.shopeio.com/inventory/details.asp?id=531&cat=Leds&sub=LED%20Displays
[22:40:15] <jacky^> doh !
[22:40:21] <LawrenceG> /home/lg/emc2HEAD/emc2/rtlib/motmod.ko base_period=50000 servo_period=1000000 traj_period=10000000 key=111
[22:40:21] <LawrenceG> Version: 1.34
[22:40:21] <LawrenceG> Machine: EMC-HAL
[22:40:21] <LawrenceG> RTAPI: ERROR: could not open shared memory
[22:40:22] <LawrenceG> HAL: ERROR: rtapi init failed
[22:40:38] <LawrenceG> etc
[22:40:41] <jacky^> i've about 20 of this
[22:40:49] <jacky^> the same
[22:41:07] <Phydbleep> jacky^: Like the display in this. http://68.35.143.93/lathe/061105001.bmp
[22:41:23] <asdfqwega> in the commit logs, jmk changed the shmem_key in emc.ini - would that cause the problem?
[22:41:50] <paul_c> No. It's because rtai_shm is missing from /dev
[22:42:11] <jacky^> Phydbleep: very nice :-))
[22:42:12] <Phydbleep> jacky^: That panel box is about 400mm x 400mm
[22:42:21] <LawrenceG> I think it loads, but script cleans up everything when it exits
[22:42:36] <paul_c> This had been solved wrt udev in the bdi-4 branch.
[22:48:19] <jacky^> Phydbleep: found the circuit right for my know: http://www.allaboutcircuits.com/vol_6/chpt_7/9.html
[22:48:22] <jacky^> LOL
[22:48:27] <jacky^> :)
[22:48:49] <jacky^> i will try this
[22:49:24] <robin_sz> hmmm ... if i just cat a .tif | lpr?
[22:49:30] <robin_sz> will I get an image ..
[22:49:37] <robin_sz> or a million sheets of text?
[22:50:28] <robin_sz> * robin_sz goes to the printer to wait ...
[22:50:35] <jacky^> :\
[22:50:57] <jacky^> maybe cat .tif > lp ?
[22:51:07] <jacky^> if lp is ypur printer..
[22:51:47] <robin_sz> its a remote printer
[22:51:51] <robin_sz> running lpd
[22:51:52] <Timbo> that'll create a file called lp
[22:51:53] <jacky^> ah..
[22:51:56] <robin_sz> cool .. it worked :)
[22:54:42] <Timbo> http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&category=64815&item=7519014958&rd=1
[22:54:44] <Timbo> i love ebay
[22:55:40] <robin_sz> nice
[22:55:50] <robin_sz> about 20% of normal price, if that
[22:56:07] <Timbo> yeah
[22:56:31] <Timbo> not sure how much i'll use the 1mm though :)
[23:00:20] <robin_sz> sell them on eBay :)
[23:00:39] <robin_sz> my ebay purchase today was:
[23:00:41] <robin_sz> http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=7521620557
[23:05:16] <Timbo> coolant pump? :)
[23:28:45] <robin_sz> vacuum pump
[23:29:08] <robin_sz> for my vacuum .. errm .. pick and place system
[23:29:50] <robin_sz> its picking and placing 250kg sheets of steel, so it needs a bit of flow capability
[23:29:56] <robin_sz> just in case :)
[23:31:02] <robin_sz> that will probably end up emc driven too, as HAL gives me the opprtunity to do cute things to interface it with the sheet storage
[23:36:13] <A-L-P-H-A> Is it safe to set the axis cycle_time to 0.00012 if the period is set to 0.000024?
[23:36:49] <robin_sz> no eyed dear
[23:40:11] <A-L-P-H-A> heh
[23:40:19] <A-L-P-H-A> well, time to test out of I'm gonna miss pulses.
[23:40:22] <A-L-P-H-A> hORRAy!
[23:41:04] <A-L-P-H-A> wtf? uninstalling microsoft office required me to close thunderbird.
[23:41:06] <A-L-P-H-A> weird.
[23:44:16] <robin_sz> you know the first mistake you made?
[23:45:05] <robin_sz> installing MS office in the first place
[23:45:07] <A-L-P-H-A> open office wasn't good then.
[23:45:11] <robin_sz> ahh
[23:45:15] <robin_sz> its pretty good now
[23:45:26] <A-L-P-H-A> yeah... hence, UNINSTALLING :D hahahhaa
[23:45:33] <robin_sz> yeah
[23:45:44] <A-L-P-H-A> No more office. :D
[23:45:51] <robin_sz> hurrah!
[23:46:05] <A-L-P-H-A> thunderbird, firefox, (sunbird is shit still though), and OO.o installed.
[23:46:53] <jacky^> A-L-P-H-A: try also latex and emacs
[23:47:08] <A-L-P-H-A> latex? I don't do functions.
[23:47:14] <dave-e> hi- I'm back and need some more help
[23:47:17] <A-L-P-H-A> or am I thinking something else?
[23:47:28] <jacky^> right
[23:47:31] <A-L-P-H-A> emacs?... I dunno... seems way about and beyond what I need.
[23:47:35] <robin_sz> latex and emacs?
[23:47:45] <A-L-P-H-A> I don't even like VI.
[23:47:45] <dave-e> does one build emc2 like cvs on 4.2?
[23:47:58] <jacky^> emacs is not an editor, it's like an O.S. :D
[23:47:59] <A-L-P-H-A> bdi 4.20?
[23:48:09] <dave-e> yep
[23:48:17] <A-L-P-H-A> dave-e, if you do... can you document it? and perhaps share your experiences?
[23:48:19] <jacky^> yu can send and recieve mails
[23:48:20] <robin_sz> * robin_sz tries to imagine what would happen if he asked his secreetary to use latex and emacs for wp.
[23:48:25] <jacky^> connect irc
[23:48:26] <dave-e> ha!
[23:48:36] <jacky^> send sms
[23:48:38] <A-L-P-H-A> jacky^... I want something EASY to use.
[23:48:41] <dave-e> if I don't then jmk gets to do it at fest
[23:48:41] <jacky^> and more other..
[23:48:53] <robin_sz> A-L-P-H-A: yeah, exactly ...
[23:48:53] <jacky^> A-L-P-H-A: it's easy
[23:48:57] <A-L-P-H-A> I don't want a text editor to be come my OS thanks.
[23:49:01] <jacky^> also have a nice tutorial
[23:49:16] <jacky^> you can say goodbbye to the mouse
[23:49:30] <jacky^> enjoy the power of keyb
[23:49:36] <robin_sz> A-L-P-H-A: latex and emacs can do it all, and more, but the UI is not suited to normal people
[23:49:49] <dave-e> nah! have her design a form in postscript
[23:49:52] <jacky^> it's a time question..
[23:50:01] <robin_sz> jacky^: that worked OK in the 1970's, its time has passed
[23:50:17] <jacky^> if you start right
[23:50:57] <jacky^> robin_sz: in the 1970 can't send sms from your editor..
[23:51:04] <dave-e> lyx still works for documentation and the export is easy
[23:51:05] <jacky^> can believe me
[23:51:13] <robin_sz> neither would I ever want to
[23:52:01] <dave-e> dave is out for food .... back later
[23:52:08] <jacky^> one more powerful thing is you can use lisp (emacs language) to build your own functions
[23:52:19] <robin_sz> OK, heres the deal.
[23:52:40] <jacky^> :\
[23:52:41] <robin_sz> you explain lisp to my secretary and have her do soemthing useful with it
[23:52:53] <jacky^> i'm laso hard working to learn g-code
[23:53:07] <jacky^> i know it's is importante for me
[23:53:11] <robin_sz> sure .. but you are a techie
[23:53:16] <robin_sz> you can learn it easy
[23:53:24] <jacky^> can't use a cad without know how gcode work :\
[23:53:30] <robin_sz> true
[23:53:35] <jacky^> :-)
[23:53:35] <A-L-P-H-A> brb... testing new gcode.
[23:54:07] <robin_sz> but ... people should be able to use a word processing applicatrion withuot learnig millions of keystrokes
[23:54:21] <robin_sz> OpenOffice is the tool for "normal" people
[23:54:22] <jacky^> also right
[23:54:50] <robin_sz> us techies can do it our own way :)
[23:55:22] <jacky^> i made my little experience also with the wm
[23:55:37] <jacky^> after tried all: from kde gnome & co.
[23:55:46] <jacky^> i switched to ion
[23:56:09] <jacky^> after few mounth of hard work, i found my peace :D
[23:56:53] <jacky^> i concluded that Linux is not for ordinary use
[23:57:49] <robin_sz> well,
[23:58:14] <robin_sz> I set my wife a box with gnome, and shee uses it fine
[23:58:37] <jacky^> yeah, right
[23:58:42] <robin_sz> I dont really use my wm much at all ...
[23:58:59] <robin_sz> im sure it does LOTS more than I use it for
[23:59:30] <robin_sz> all I ever do is open a few xterms and a couple of browser windows
[23:59:52] <jacky^> right