Jymmm: yes it's interesting, but why did it make you think about me?
jepler In the way you turned a spiralgraph program into g-code using python
oh, are you trying to trick me into making a fractal g-code program now?
jepler: Heh, no not at all. lots of those out there. I'll reserve my trickery for far more valuable future needs =)
I have an interest in shapes and math, though I know little of the later.
Jymmm: I'd try to enter into a cnn studio and sit down for an interview :P
you could become a celebrity :P
did you seen goma video ? :D
Jymmm: I do have a sierpinski gasket g-code file, as a matter of fact
[00:16:18] <jepler> http://axis.unpy.net/files/etchcnc/etch-sierpinski.jpg
* giacus smoke last cigarette before go to sleep
jepler: =) I kinda have a peak interest in 3D shapes, hard to explain.
sorta a cube, but not.
jepler: sorta kinda likes these... http://www.physics.orst.edu/~bulatov/polyhedra/uniform_compounds/index.html
$ cat /proc/ioports
I think I did something wrong
found another one
Jepler: what does that last commits mean?
what is that aiming to do I mean.
telling linux that we're using the parport, so it won't let someone try to print their homework on the stepper motors
that makes sense
skunkworks: jmk observed that when you boot the mazak with ubuntu, the charge pump is activated and the power relay clicks
skunkworks: that's because linux has probed the parport at bootup
with this change, if linux has a parport driver loaded, 'loadrt hal_parport' will fail with an error message that says the I/O ports are in use
you have to fix it by making linux never load parport (actually, parport_pc)
and now your mazak servo amps won't power on for a moment when the computer does, and every day at midnight your mill won't go 4 steps to the +X (or something crazy like that)
:) - Now I have to figure out how to not load the parport_pc.
could - cradeks emc-ubuntu script unload the parport_pc? (if I am understanging it right)
or ask if you want to.
skunkworks: an upcoming emc2 package will do something about that
skunkworks: so everyone will get the problem fixed when they upgrade
great work. the brain power in that town must be emi or something I would think. :)
it is causing thunderstorms actually
I love this conversation when jmk and cradek got togather to solve one of my emc problems ( they where hooked into my computer )
<cradek> you know what
<cradek> I could reboot it into memtest86
<jmkasunich> and let him get the results tomorrow?
<cradek> it could run all night, you send him an email describing what to look for, and he could report in the morning
<jmkasunich> waitaminnit - how can you do that
<jmkasunich> reboot, yes, but reboot <foo>?
<cradek> I can use my powers only for good, not for evil
Dumb question, but does anyone know how to calculate the distabce of a diaganal of a 4" square (as example)? This is at 45 deg angle, but at various angles too?
the general answer is trigonometry, specifically you can use the pythagorean theorem
Dont know trig, any siggested reading? urls?
and the really specific answer is diaginal = sqrt(height^2 + width^2)
looks like dont know how to type tonight either =)
jmkasunich 45 deg only ?
thanks for simplifying that for me
for a square, height and width are the same
for a rectangle they aren't
so your 4" square has a 4 * sqrt(2) inch diagonal
03swpadnos 07HEAD * 10documents/lyx/emc2/hal_general_ref.lyx: First stab at canonical interfaces for analog input and output
Ok, I get it now... one other: What if you wanted a 1/2" wide slot to in a 4" square? any ideas?
like a dado as example
I don't understand the question
Lets say you setup your tablesaw with a dado at 1/2" wide. You have a 4" diamond (square at 45 degrees), since the diaganal is no long a thin line, but 1/2" wide, how would you calculate the length without exceeding the 4" material
btw, thanks jmkasunich: writing that one down.
the length to the points is still the same, so I still don't quite understand the distance you're trying to determine
the length of "full width" slot?
I'll draw it up. gimme a few.
[07:18:06] <Jymmm> http://static.flickr.com/54/148050806_40f287c53e_o.gif
I hope that explains better.
ok yes it does
what you have is two cases of the same problem and you can figure them out the same way
well let me think how to explain it
take your time, grabbing coffee
now you have a smaller triangle with a width and height less than 4" and you want to calculate the diagonal still
so you know how to do that after you figure out just how much smaller than 4" the new sides are
the answer to that by also using the pythagorean theorem is .5" / sqrt(2) = .3535
so your new sides are 4-.3535 = 3.6464
so using the same formula again you find that your unknown length is 3.6464 * sqrt(2) = 5.1569
or you could just subtract the slot width from the point-to-point length
(5.6568 - 0.5 = 5.1658)
I'm not sure that would hold true for angles other than 45 degrees tho
in fact I'm sure it wouldn't
right it doesn't
hmm your answer is better, how did I add that much slop with my rounding?
maybe I did it wrong
one count in the least significant didit isn't much slop
I'd like to smack whoever decided to implement pause with feedrate override
ahhhaha.. I see ;)
Um... where did the sqrt(2) come from ? (just curious)
remember sqrt(height^2 + width^2)?
and square means height = width
so its sqrt(height^2 + height^2)
ah, funny that is also PP as well
which is sqrt(height^2 * 2)
PP 1.414, RMS .707
which is sqrt(height^2) * sqrt(2)
which is height * sqrt(2)
ok, I think I understand. thanks. One of these days I'll even figure out what 1+1 is =_
thats not a coincidence - the characteristics of sine waves are related to the trigenometry that calculates diagonals of squares
looks right on the money to me =) thank you both
I need to write that up under the heading "shop math"
03jmkasunich 07HEAD * 10emc2/src/emc/kinematics/ (tp.c tp.h): Motion controller part of adaptive feedrate control (for EDM and other applications). Still needs interpreter and canon code to turn it on and off.
03jmkasunich 07HEAD * 10emc2/src/emc/motion/ (6 files): Motion controller part of adaptive feedrate control (for EDM and other applications). Still needs interpreter and canon code to turn it on and off.
time for slee
you touched it last!!
time for sleep
looks like it works
(I hacked to turn it on)
maybe sometime tomorrow our edm friend can try it
used a joystick to control feedoverride ;-)
what are you doing awake anyway?
if you were going to stay up that late you should be out here coding!!!
what are any of us doing up this late?
thats what I keep asking myself
I don't know either
* alex_joni longs for his bed
oh, hi matt
hi form Argentina
alex_joni, fenn are you there ?
anybody knows how can i send a message to them ?
VNR: I am
i logged in only to tell you that i could get the EMC working, i promissed to do that ;)
VNR: that is such a good news ;)
VNR: how did it work out eventually?
i installed the packages listed in step 7 of the documentation (not all)
i have to say that the documentation is a little bit old, things doesnt work as it is said
VNR: I must admit it is pretty outdated, but we (I) are working on it, inbetween things
the documentation that says how to install emc without internet (step 7)
libc6 and all libstdc++* doesnt install, but EMC works without this files ;)
oh, right, those might be used for compiling
wait me, let me find the URL that i used
[12:18:29] <VNR> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?Installing_EMC2,
commands doesnt work too, i skipped all the commandss and used only one to install: dpkg -i
i want to know a litte bit about the mazak project, are you involved in that ?
alex_joni: i have to go, thanks for your help, bye
VNR: no problem, glad you got it working
is Debian testing ok to install Emc2 from source ?
found the answer on the wiki ..
going ro remove BDI now ..
jepler: it.po needs to be updated ?
I have a bit of time to play with it, today and tomorrow.. let me know
it.po for Axis I meant ..
giacus: make one for rs274ngc.pot too
debian testing should be ok for emc2
it's your choice though what kernel/rt you use
either the one from the BDI (sarge kernel patched), or the one from breezy (Ubuntu)
how come the pwm frequency of motor drivers rarely exceeds 20kHz?
it's expensive to make it very high
and you don't need such a high frequency..
for a motor even 10-15 kHz might be enough, but that's in the audio range of the human ear
with higher frequencies there are more transitions that dissipate more power
right, and motor inductions make it useless (it works very ok even at lower switching frequencies)
les_w: btw, I borked out some program once, and made a DC driver do PWM with .xx Hz ;)
err.. xx Hz
you could see the motor stutter
poor motor ;-)
I goofed up ordering inductors for power factor correction of the finger burner.....one decimal place off
that's ok alex, ty, now installing
les_w: which direction?
oh, too bad les
so I just sent the thing to chicago with the power factor off
32k pF shunt capacitance
a proper inductor would have neutralized it
anyway I have it lighting up cluster leds very brightly, and they seem amazed up north
which is silly
amazing what a silly light show will do
alex_joni: do you seen the guy goma video ?
if not, read this before: http://www.newswireless.net/index.cfm/article/2697
* alex_joni is a bit busy in the interp..
don't wanna lose my concentration, or I'll go nuts
and take a look at video http://www.youtube.com/watch?v=scvOLr0-CiY&feature=Views&page=1&t=t&f=b
bookmark it then ;P
wow giacus - that is funny
he's a star now LOL :D
I feel sorry for the other guy. That would be my luck - 10 seconds of fame and someone impersonates me :)
03jepler 07maint1_3 * 10axis/scripts/axis.py:
Fix problem reported by Marc van Doornik: in metric and an offset was in effect,
the displayed dimensions were wrong.
I like the face he do before say good morning :P
and that he isn't out of bed even
some people shave in the shower
too much information
Does anyone have an idea where I can borrow a 400V AC conection to test a plasmacutter that I'll bring along with a small compressor?
In France, around Nice
no idea ;)
websys: mind if I /msg you for a while?
I have a few questions about synergy
websys: I also have a question
I know, peoples have to live ..
but there's any hope to see a complete open source vers. of synergy in the next time ?
If I remove solid modeling I could make it open source but the solid kernel isn't mine to give away
03alex_joni 07HEAD * 10emc2/src/emc/task/emccanon.cc:
Added NML message for ADAPTIVE_CONTROL mode inside the motion controller.
This message comes from emccanon.cc and gets sent to task.
03alex_joni 07HEAD * 10emc2/src/emc/nml_intf/ (emc.cc emc.hh):
Added NML message for ADAPTIVE_CONTROL mode inside the motion controller.
This message comes from emccanon.cc and gets sent to task.
* giacus Get:7 http://solaris.cs.utt.ro
sarge/extras kernel-image-2.6.10-adeos bdi.92 [13,9MB]
giacus: you shouldn't use solaris anymore
that's a deprecated machine
it only holds stuff from a few months ago
what then ?
I should take them offline
try the ubuntu repository
[14:48:03] <alex_joni> http://www.linuxcnc.org/index.php?option=com_content&task=view&id=21&Itemid=4
* alex_joni goes home
if that doesn't work , the BDI debs are somewhere on ourproject (google bdi4emc)
paul seems to pretty aggressively remove old debs from his repository
jepler: space problems?
I think it's a terrible practice
jepler: hi btw
I don't know why
is jmk up?
we had agreed to meet for breakfast at 10AM, which is in 8 minutes. but I doubt he'll be online tilla after breakfast
jepler: if you see jmk, let him know I added most of the stuff.. but can't continue right now, the rest should be fairly easy
ok we appreciate it
you checked in what you were able to do?
* alex_joni needs to hurry to his massage hour, and then prepare for a small b-day party
and I even tested a bit..
thanks, see you laer
whose birthday, yours?
oh, best whishes :D
03jepler 07maint1_2 * 10axis/scripts/axis.py: backport: fix metric offset display bug
Now an emc2-install.sh icon will appear on your Desktop. Right-click that icon ...
what icon ? I don't have any icon on my desktop !
I'm using Ion3 :P
jepler: sigh !
E: Impossibile trovare emc2-axis
giacus: if you're not using the default desktop we assume you know how to run a shell script
thats script seems do not work fine on my system
giacus: you get that error when you don't have the ubuntu deb repositories installed. fiix it in synaptic or by editing sources.list
(this can appen when the network is not available when you install ubuntu, and it's happened to so many people we have to do something about it )
sudo: apt-key: command not found
have to check a couple of things ..
I'll be back in an hour or so to help
but for now, it's breakfast time in USA
have a nice breakfast ;)
btw, it doesnt work for me:
The following packages have unmet dependencies:
emc2-axis: Depends: python (>= 2.4) but 2.3.5-2 is to be installed
Depends: emc2 but it is not going to be installed
Depends: python (>= 2.4) but 2.3.5-2 is to be installed
Depends: libgcc1 (>= 1:4.0.1) but 1:3.4.3-13 is to be installed
Depends: libstdc++6 (>= 4.0.1) but it is not going to be installed
E: Broken packages
* giacus cross the foot fingers, hands fingers are busy ..
my latest mill-project pic: http://photos.xmission.com/bill/v/machining/IMG_0296.JPG.html
I like steel
I'm afraid to mount the rails.
cheap and good
I may replace that plate with some 1/2 inch aluminum, to keep the weight down.
I'm just worried that I wont get them on straight.
you'll get bimetallic strip effect
steel is nice :)
you mean if I use aluminum it will expand/shrink at different rates?
no, alluminium sometime have some bad effect
that's what I know
and, is expensive too ..
yeah, it's not cheap.
that plate that the rails are sitting on was $22
Had a reply back from Paul about including emc2
03jepler 07HEAD * 10axis/extensions/gcodemodule.cc: stub out new messages from emc2 interpreter (messages for adaptive feed rate)
Let me paste the reply.
Given a suitable emc2 package (or packages), it would take an hour or two to
generate an ISO - Even with the inclusion of custom images and/or help
screens, the build time would be about the same. Space has always been
allocated in the hope that emc2 be usable one day..
and then goes on to say
that last bit is pure fud. I'm looking at 4 other people who use emc2 every day.
The current package rules need to be thrown away and a new set written, a
couple of hours work at most.. However, I am unwilling to put my name to such
a package because of concerns over (probable) copyright infringements with
respect to two sets of files.
I guess your idea to pay paul to package emc2 is not gonna fly then. that's too bad.
but emc2 is gpl software, so unless he intends to violate the gpl I don't understand what "(probable) copyright infringements" he means. Is it just more FUD?
The last line.
In parting, I can build a BDI-4 based CD for you complete with emc2,
bigAl: the emc on the BDI is in violation of quite a few copyrights, so I see no reason for paul to concearn about emc2's copyright
alex_joni, Are you sure, and can you provide specific details ?
well I for one don't understand what paul's saying.
and I will ask the same question of Paul.
bigAl: I am sure, but no, I cannot (or rather) will not provide specific details, because it's not my copyright, so I guess I shouldn't care
personally, I believe that paul has failed to respect the GPL's requirement to include "the scripts used to control compilation and installation of the executable", e.g., the bits necessary to make 'dpkg-buildpackage' work for the emc and python-axis packages.
perhaps it's a violation due to ineptness rather than a deliberate will to violate the GPL
that does not amount to "quite a few copyrights", just one
they dont actually have to be included, but must be supplied if you ask.
which may be an oversight on his part.
bill203: jepler asked quite a few times
bill203: paul censors people who ask for it on his mailing list
ahh, then that's no good.
bill203: when I asked the same question when he was in irc, he quickly left.
who actually holds the license on emc? nist still?
emc1 is Public Domain
bill203: emc is copyright all the contributors
or, rather, each contribution is copyright its contributor
err, what jepler said
so anyone who's contributed can (try to)enforce thier license.
if you're corporate you'll want to be a lot more careful about license issues than some random individual
03jepler 07HEAD * 10axis/tcl/axis.tcl: add halscope to menu
[17:30:25] <giacus> http://rafb.net/paste/results/iR68p060.html
here's my issue
any idea ?
I suspect binutils ..
that's Debian testing
apm, acpi & co are disabled ..
gcc version 4.0.3 (Debian 4.0.3-1)
GNU Make 3.80
a lot of the hits for that error string are for 64-bit machines
[17:38:31] <jepler> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=328707
(but that's about kernel 2.4)
Linux emc2 2.6.8-2-386 #1 Tue Aug 16 12:46:35 UTC 2005 i686 GNU/Linux
I think binutils..
I could try with a more recent kernel source
have to check rtai-3.2 ..
giacus: what patch did you use?
just looking here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?RtaiSteps
some package is different in Debian, but should be ok
giacus: use the Adeos patch from gna.org
03jepler 07maint1_3 * 10axis/extensions/gcodemodule.cc: backport canon interface change
03jepler 07maint1_2 * 10axis/extensions/gcodemodule.cc: backport canon interface change
alex_joni, Your email address please.
bigAl: alex.joni AT robcon.ro
or firstname.lastname@example.org (as you can see on SF)
mail sent to both of you.
bigAl: didn't get anything yet
alex: you there?
jmkasunich: on and off
trying to figure out whats left to get G50/51 working
chris is looking a the interp side
I was gonna work on the other side
03jmkasunich 07HEAD * 10emc2/src/emc/nml_intf/emc.hh: added taskintf.cc function for adaptive feedrate
03jmkasunich 07HEAD * 10emc2/src/emc/task/taskintf.cc: added taskintf.cc function for adaptive feedrate
03cradek 07HEAD * 10emc2/src/emc/rs274ngc/ (8 files): g50/g51 should be its own modal group, checked in for alex to finish for me
2.6.16 seems to compile ok until now ..
if I get it running I'll write an howto:
How I did it. (by victor giacus frankenstein)
there was some issue with the 2.6.10 & debian
scaling and mirroring?
CC [M] drivers/isdn/pcbit/edss1.o
CC [M] drivers/isdn/pcbit/drv.o
* giacus ZzzzZZzzz
2 hours compiling.. :/
just forgot to remove 7-8 modules from config ..
03alex_joni 07HEAD * 10emc2/src/emc/rs274ngc/interp_internal.cc: tiny fix to make cradek's work perfect
giacus: 2.6.16 works nice with debian
Hi alex. Rayh here
I guess so :-)
I guessed so :-)
i figured you would
Well, I tried it.
Using an Ubuntu apt pool on a Debian install
In my (oh so very humble) opinion, it is A Thing Not To Be Done While In One's Right Mind
I'll stick with compiling tarballs...much safer.
asdfqwega: wot happened?
asdfqwega: talking to me ?
I added the emc2-axis apt pools into my sources.list, updated, and then attempted to install emc2
The dependency scheme is completely different
I think I would have broken my system six ways from sunday, had I let it go through
I'd like to play with Ubuntu sometime, but...not right now.
I run it on my desktop at home.
it's pretty slick, but it sucks a lot of memory
[22:11:21] <bill203> http://video.google.com/videoplay?docid=4468957986746104671&q=500kv
video of a 500KVA line being disconnected, under load.
huge electrical arcs..
I've tried using a 'user-friendly' disto before (Mandrake), and it couldn't survive my usage.
Put on a hard hat and flash gear, come on over, and I'll show you what happens when you drop a screwdriver into a live 480VAC switchgear
asdfqwega: I'm in love with debian and ion3 as wm
* asdfqwega is currently working as an electrician :)
I used evilwm for some year
however, ion3 has floating workspace
giacus: Haven't tried ion3 yet, but I got disgusted with KDE, and now have a fluxbox the way I like it
I also tried mdk years ago ..
Mandrake, now Mandriva
User-friendly, but brittle when upgrading
but I have to say, Ubuntu is nice
just I'm a debian user
Heh, now I'm going to have to give ion3 a try
is cool ..
I'm just not a typical user, I'm picky AND like to fiddle
Heck, I don't even care for paul_c's BDI...I just use the .deb files in a debian install
Lady? Who? Where?
I seen paul included evilwm in BDI
if I read right in the ML
asdfqwega: I' m lady
hm, did he include emc2 too?
ok good night
s/ /o /
03alex_joni 07HEAD * 10emc2/src/emc/nml_intf/emc.hh: final bits for G50/G51 to reach the motion controller
03alex_joni 07HEAD * 10emc2/src/emc/rs274ngc/interp_convert.cc: final bits for G50/G51 to reach the motion controller
03alex_joni 07HEAD * 10emc2/src/emc/task/ (emctaskmain.cc taskintf.cc): final bits for G50/G51 to reach the motion controller
include/linux/thread_info.h:52: internal compiler error: Segmentation fault
ouch, dunno how many time I'm resuming make modules..
giacus: try the debian way
I should be near to end yet..
<- if you dare
yes, nice, already seen that page
some day ago
the 'hard' way is to patch a kernel the debian (ubuntu) way
yep, I just tried to ask on #gpe about RT pacthes for the 2.4.19-rmk6-pxa1-hh37 (the ipaq running gpe) and cross-compilation, got no luck
and I've another issue on the PDA, few space
the damn SD card won't work on 39XX serie :(
03jepler 07HEAD * 10emc2/nc_files/cds.ngc: remove some g0 moves that go right through the piece
losting the hope to get emc2 running there ..
03alex_joni 07HEAD * 10emc2/src/ (config.h.in configure.in): make G50/G61 appear in the list of current g-codes
03alex_joni 07HEAD * 10emc2/src/emc/rs274ngc/ (interp_convert.cc interp_write.cc): make G50/G61 appear in the list of current g-codes
this is new: www.emc2.it/
03jmkasunich 07HEAD * 10emc2/src/emc/kinematics/tp.c: fixed a bug I introduced last night - should not pause/abort until synchronized move(s) are complete
Linux emc2 2.6.16 #1 PREEMPT Wed May 17 23:00:27 CEST 2006 i686 GNU/Linux
run baby run :D