03jmkasunich 07HEAD * 10emc2/src/hal/drivers/hal_ppmc.c: fix up a TODO
03jmkasunich 07HEAD * 10emc2/src/emc/motion/motion.c: fix up a TODO
03jmkasunich 07HEAD * 10emc2/src/hal/components/ (counter.c pwmgen.c timedelay.c): fix up a TODO
03jmkasunich 07HEAD * 10emc2/src/hal/hal.h: fix up a TODO
03jmkasunich 07HEAD * 10emc2/src/hal/utils/halcmd.c: fix up a TODO
03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot3_log.txt
jmkasunich: did you completely remove the old direction constants?
hal/halmodule.cc:562: `HAL_RD' undeclared (first use this function)
I removed them from hal.h, compiled, and fixed every error that popped up
does that mean the dependency for hal.c somehow missed halmodule.cc?
I don't know; I'll look into it
on my system, hal/hal.h is a dependency of halmodule.o
$ cat depends/hal/halmodule.d | grep hal.h
depends/hal/halmodule.d objects/hal/halmodule.o: hal/halmodule.cc \
/usr/include/python2.4/pyfpe.h hal/hal.h hal/hal_priv.h \
and if I "touch hal/hal.h" then halmodule.o is among the things rebuilt
03jepler 07HEAD * 10emc2/src/hal/halmodule.cc: get rid of old HAL direction constants
or was I a little premature in ripping them out?
strange, in sim/axis I am having trouble turning machine on
ignore last statement, for some reason your last sentences didn't show up
the button goes in just briefly and then pops back out
in fact sometimes it goes into estop
and ... sometimes it comes out of estop
yeah I keep poking F2 and it goes into all different states
cradek: I've seen something like that and ignored it
thankfully, tkemc does it too :-)
i've been trying really hard to ignore that, but you're not making it easy
crap I bet I just made it easier
03jepler 07HEAD * 10emc2/src/hal/hal.h: comment doc fixes: do not refer to old constant names
I did a make clean, and the halmodule.cc error _still_ didn't happen
(I haven't done an update, so I don't have jeff's fix)
maybe you fixed it, but didn't check it in?
03jepler 07HEAD * 10emc2/docs/man/man3/ (rtapi_release_region.3rtapi rtapi_request_region.3rtapi): these were documented in rtapi_region.3rtapi but didn't have links to these names
nope, cvs diff shows only my later changes to ppmc.c
jmkasunich: your system builds the axis stuff?
I thought it did
but now I wonder
is there lib/python/hal.so ?
you mean emc2/lib/python, right?
no .so's at all
in that directory
well that explains why you didn't see the build error
but replaces that with a more complicated question - why no axis build?
rerunning ./configure, maybe it will tell me something
don't you run emc?
remember - fresh dapper machine here
axis built and worked fine on my breezy install
checking for /usr/include/python2.4/Python.h... no
something tells me that is important
you would be right
no readline either
and no ncurses
sudo apt-get build-dep emc2 should get everything that is needed shouldn't it?
it's probably a good start, but those are the build-deps for 2.0
was axis still a separate package for 2.0?
you'd also want build-dep emc2-axis
(said the blind man)
looks like thats going to get the phython stuff
not readline or ncurses tho
[02:04:22] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2
are they considered optional?
what's the arabic? dish made with tomato, bell pepper, onion, garlic, and olive oil - served on rice?
sounds tasty tho
it was, I made some tonight
and I get to have red wine with mine - they're not allowed
omfg! he said lol!
omg we r n a chat rm
oooh my head - what happened?
(that was the taser bring me back to my senses)
jepler: it must be some kind of initial condition because once F2 works, it keeps working
turning on debug doesn't give any good clues
well I get this
wonder if it means something
jepler: were you running sim when you saw this?
I mean simulator
my box builds axis now - sorry about that
can you turn machine on?
you might be the only one with realtime right now
and aborting a running program dumps me into estop
something's really screwy
I'm going to try a full recompile
F1 brings out of estop, F2 turns on, and it runs the splash .ngc fine
escape works right too?
(leaves machine on)
maybe it's a simulator thing
03compile-farm 07BDI-TNG (2.4.18-rtai) * 10emc2head/: build PASSED
no change after a clean here
03jmkasunich 07HEAD * 10emc2/src/hal/drivers/hal_ppmc.c: revise EPP bus 'cache' algorithm to read and write ONLY the required bytes, and skip gaps in the register map
on my other machine with realtime everything is fine
maybe it is something about the simulator
but that isn't even realtime code is it?
cradek: I kinda suspect it is
something where the stat buffer isn't fetched proprely from motmod to "userspace"
or .. I dunno what
maybe the estop chain in HAL?
I got my "virtual jog wheel" to "work". It makes the servo feel like turning a weakly-energized stepper.
how many detents?
what's a virtual jog wheel?
its really a virtual detent
yes, what jmk said
it makes the servo+encoder snap to certain angles so it feels like there are detents
it's not really a very good idea
here's another one to play with
that would be cool too.
I spin it and it keeps going?
actually, I think that would require a torque sensor
make something feel like it has a lot of inertia
jepler: fwiw, http://pastebin.ca/235229
you apply a some torque, and instead of just accelerating easily, it resists, as if there was 10 lbs of steel on the shaft and you were spinning it up
when you let go, it coasts
emcStatus->io.aux.estop=1 has something to do with the problem.
"EMC Mortgage Corporation - Find the best mortgage for you" -- google ad on that pastebin
I didn't know there were ads - I don't see them
I think I've trained my eyes to avoid boxes of that particular shape
I see it - top banner
didn't notice it the first time
for some reason I don't have adblock on that machine
I don't use it much
cradek: do you know how to make the problem happen?
do you get "%"s when it does happen?
it always happens when I restart emc
no I don't get any %, I looked carefully
I thought I'd seen it happen, but I don't get it on my laptop
you're using configs/sim/axis.ini right?
I don't want it to be install2 -- that machine is realtime
you may have built sim on it though.
could be on india
wish I had a clue what to check, but I don't feel like I do
so you get it every time you run
yes lately every time (10 or so in a row)
I cleaned and rebuilt
03jepler 07HEAD * 10emc2/src/emc/motion/Submakefile: remove puresim/simmot, it is not needed
when you shut it down, is rtapi_app dead? does halcmd show anything?
yes rtapi_app exits, and halcmd shows empty after exit
while it's running, halcmd show looks normal enough
You could set the jog wheel up so that it matched the mass of your machines axis.
use spindle motor current feedback to provide resistance to the jogwheel when as the cut gets heavier ;-)
ooh your good
or make the knob harder to turn as you approach the machine limits
any vmware folks around?
SWPadnos or cradek ?
does it work with a RT kernel?
I got this during the install:
Trying to find a suitable vmmon module for your running kernel.
None of the pre-built vmmon modules for VMware Server is suitable for your
running kernel. Do you want this program to try to build the vmmon module for
your system (you need to have a C compiler installed on your system)? [yes]
you have to build the modules to match -magma
but if I do it will work?
you pretty much have to hit return fifty times, then it will work
my experience was different -- it didn't work on my laptop with the magma kernel. hard-locked when I started vmware.
hopefully that was some kind of laptop weirdness
this is a generic desktop
I guess I'll find out
does it look like the bitmap cache tracking on ppmc will help?
I think so
its in Jon's hands now
heh - should hve been some time ago, I think
looks like vmware wants inetd to be running
and looks like ubuntu doesn't run it
I don't think it's needed. are you getting a network error dialog when you start vmware?
haven't gotten that far
doing the install now
something like "vmnet0 could not be started" or something?
inetd: no process killed
Unable to make the Internet super-server (inetd) re-read its configuration
file. Please restart inetd by hand:
killall -v -HUP inetd
cradek: when emc is not running, does ipcs show its shared memory segments? Their keys are around around 0x3ex for the rcslib stuff and 0x6f for the buffer between motmod and task
the killall didn't work, and ps -A doesn't show inetd
interesting. I don't remember gtetting an error of that sort
It says hit enter to continue, I'll keep going and see what happens
looks like ubuntu doesn't have an inetd or xinetd installed by default
damn. I should have grabbed some DB25 breakout connectors when I was at Radio Shack today
I keo going back for one or two items. one day I'll remember to make a list first
I'm just shooting in the dark here .. but .. ipcrm -M 0x000003eb and the same for the other two in that range
while you're at it, try removing the semaphores too
that fixed it?
yes I think so
if that's true it must be some kind of shutdown problem
3 successful runs in a row
0x000003eb 40075300 chris 777 8192 1
it shows a "nattach" -- is it possible that you have/had an old emc process laying around?
task or emcsvr or something
/home/chris/emc2.head/bin/io -ini /home/chris/emc2.head/configs/sim/tkemc.ini
03jepler 07HEAD * 10emc2/src/rtapi/sim_common.h: this comment corresponded to removed code
so was the whole problem caused by a pre-existing instance of EMC?
a leftover part of one I think
the old one must have had a bad cleanup of some kind
I wonder why the emulated disk is SCSI
I thought it usually matched the type of physical disk
easier or better API maybe
there's no reason to make it match the physical disk (I doubt vmware knows what the physical disk is)
vmware has the ability to directly use a physical disk or partition for a VM (instead of a file)
I'm _not_ doing that
yeah, that seems like a terrible idea
maybe I got my idea about the match from docs that referred only to that mode
I wonder how I should set up the guest X config?
it'll just work won't it?
(I'm doing a test install of BDI-2.20, because I happened to have the iso, and it will be a quick install
you can set the resolution to whatever is a convenient window size
the bdi install wants me to tell it what kind of monitor I have, etc
so "monitor size" = window size?
1024 x 768 then
I hate tiny screens
hmm, X isn't happy
the installer is graphical, and that works fine
you can set it up as an IDE drive, I think
don't you have a dapper iso?
but the installer lets you "test this configuration", which flails around a bit and then drops back into the installer again
I have dapper, but I bet this will go quicker
dapper is MUCH faster to install than brezzy
this is kind of a throwaway install, just to get the hang of things
* cradek hands jmk a freedos iso
very much so
what's up today?
same shit different day
same shit different day
*reading the mailing list*
it's about the same with me
anyone knows of a free 3D cad viewer?
one that can read ACIS sat?
hm, sorry I've got no idea
[09:35:46] <anonimasu> http://developer.hoops3d.com/
they seem to have a reference app for thir lib's..
I found a few apps online.. but mostly stuff I need to register for
no time right now for the hassle :)
if you've got time I can probably convert it for you when I get home
I converted it from acad 3d
but I wanted to check if it's ok ;)
and the pesky acad can export, but not import it
I could import it for yu
I just got a custom tool to make for work ^_^
alex_joni: if you mail me it I'll try to open it :)
03jmkasunich 07HEAD * 10emc2/src/hal/drivers/hal_ppmc.c: some minor fixes
congrats on making it back safely to your domicile.
Lerneaen_Hydra: do you still have the video up somewhere of your lathe cutting?
Nothing - at work
I have a google video version and an avi
which would you like=
hmm - the avi would be nice
my ftp is dead at the moment, do you have some mail I could send it to?
(or if you have an ftp server)
samcoinc at gmail dot com
and preferred codec?
na - I will make it work
(think windows machine though)
how about xvid at 3000kb/s?
hmm, at that framerate the filesize is quite large (90mb), gmail is max 20mb/mail, so you want it with a lower bitrate or do you have some server?
* Lerneaen_Hydra pokes skunkworks
this is odd.
this FEA stuff..
trying to calculate load on a 20cm bar..
with a 5kg load on the end..
going to draw up a robot again..
but im taking the oppurtunity to learn how to do fea at the same time
the built-in cosmo-crippleware?
the express version
oh, which version?
it's really nice
but, the results are odd..
rather I'm crappy at interpreting them
is the element size << dimensions?
there is a setting for element size, is it much smaller than the dimensions you're working with?
ie a rod 10mm dia should have an element size much smaller than 10
the "deformation scale" of my result ends up as 2.74625e^-010
I'm not sure how much smaller, but as small as possible as long as the calculation time doesn't kill you
makes the small deformation look brutak
skunkworks: do you have an ftp/whatever server somewhere?
or bittorrent or something?
Lerneaen_Hydra: I do but it has been so long since I have used it.. Have to remembere how to get there ;)
Lerneaen_Hydra: going to look
can't you guys use dcc?
does the java plugin allow dcc?
* cradek hands skunkworks a real irc client
that worked well ;)
cradek: how do you use dcc?
Lerneaen_Hydra: they are much smaller..
/dcc send somebody somefile
anonimasu: hmm, then I don't know why it's messed up
oh, that simple?
well it often doesn't work behind firewalls
what port does it go through?
0.19528158mm is the fine size..
anonimasu: oh, that should be fine enough for load-bearing stuff
I recognise that image from somewhere
the sender has to have a correct network setup for dcc to work
it's pretty forgiving of the receiver
which ports does the router/NAT have to pass through?
or does it vary?
hm, and the load becomes 50N
I don't recall the details of dcc
depends on your client. Some let you force the use of a particular port which you can forward through your firewall
dcc send says "I have a file for you. Call me back on port XXXX to get it"
Lerneaen_Hydra: a 20mm thick slab of alu should be able to take 5kg right ^_^
anonimasu: how long?
5kg is definetly doable ;)
I'd say more like 50
unless you've got a reallly really shitty alloy ;)
the stress plot are all blue.
and red=over yeild?
then what's the problem?
err let me get you a picture :)
the units are insane.
yeild strength = 2.750e+0.08
somehow I doubt that
that the load in the alu with 5kg at 20cm is 3.5*10^10
sorry - blue screen ;)
how do I tell if I am registered?
should be your modes ;)
it will bug you if you're not
skunkwork: didn't you register skunkworks ?
Lerneaen_Hydra: yeah it's odd
skunkwork: did you get the dcc offer?
Lerneaen_Hydra: could you try it again
skunkwork: I can mail you the file
it's shorter than 10MB I think
it's compressed with a much more lossy bitrate though
Lerneaen_Hydra: it might be my units that's a mess..
* alex_joni wanted to put it on linuxcnc.org anyways
5kg should be 50N
but alex_joni couldn't because?
anonimasu: yeah almost 50N
Lerneaen_Hydra: time & lazyness constraints
depending on where in the world you're located :p
alex_joni: oh, the usual
9.82/9.81 are the numbers I've seen most
but 10 is good enough for all practical purposes
on the equator you'll get a value slightly lower, and on the north/south pole a value slightly greater
skunkwork: test again?
Lerneaen_Hydra: one more time ;)
skunkwork: it could be my NAT that's messing stuff up
I set up my computer as the dmz.
hold on again
afai understood from cradek it's the hosts NAT thats the most picky
well that sucks ;)
if dcc doesn't work, the setup problem is on the sender's side
well - I get the offer and accept - then it fails after about 10 seconds
probably because I can't initiate it
skunkwork: wait a bit, I'm uploading it now
ok - thanks
do you have bittorrent/dc++/ftp/http?
I have a web site - but I have not played with the ftp part.
or played with it very little
have not played with gopher since the early 90's ;)
* Lerneaen_Hydra seems to have missed the entire gopher thing
ftp.electronicsam.com - login is email@example.com and password is emcstuff
if I did it right
* skunkwork is now just playing
Lerneaen_Hydra: will you be around for a bit tonight?
skunkwork: seems to work
anonimasu: yes although a bit AFK at times
skunkwork: I added an md5 too
is linuxcnc.org on dreamhost?
skunkwork: that linuxcnc version is 320x240 15fps
and lower bitrate
seems like yours is still uploading onto my site - cool
* Lerneaen_Hydra is AFK
* Lerneaen_Hydra has a name so long that he can't add _AFK to it
* Lerneaen_Hydra is slightly frustrated by that
* Lerneaen_Hydra is very AFK
/kick Lerneaen_Hydra AFK this
03cradek 07HEAD * 10emc2/src/emc/rs274ngc/interp_convert.cc: rapids used for exit should not be slowed down by FO
03jepler 07HEAD * 10emc2/docs/src/gcode/main.lyx: improve layout of table
cradek: I'm curious why feed override shouldnt affect rapids :)
because rapids are used to exit the thread
does that change apply to all rapids?
during the threading cycle, yes
hmm, maybe that's not right though
it's more right than none of them
but some are moving to the start of the next pass
it's possible those should be overridable
hm, I'm not sure you should be able to FO rapids..
emc has always been that way, but there seem to be lots of different feelings about it
as they should be as fast as possible :)
can you feed override over the MAX_VELOCITY for a axis?
it's an interesting perspective to think that rapids don't have a specified feed so there's nothing to override
that's "the argument from semantic pedantry"
03jepler 07HEAD * 10emc2/configs/sim/ (lathe.hal lathe.ini): simulated lathe configuration (uses AXIS)
cradek: I wonder what's the most implemeted way in commercial machines
it's a little too bad that all the sims have to share a tool table
hmm I guess they wouldn't, would they
anonimasu: I'm sure I don't know that
they don't have to, they just do
yeah.. they do..
[17:25:26] <anonimasu> http://www.haascnc.com/cncmag/answerman_article.asp?VolumeNo=-1&IssueNo=-1&ArticleID=548
interesting -- FO affects homing speed
yes, a very nice feature I think
well for testing at least
not for the later steps of a multi-stage home sequence
err - not nice then :)
03jepler 07HEAD * 10emc2/configs/sim/ (lathe.tbl lathe.ini): include a lathe-specific tool table. change tools at a plausible location.
SWPadnos: I don't understand what you mean
hmmm - in fact, it may be bad for the initial step as well
you mean a different velocity affects when the switch switches?
well, for home-with-index, you want to be going at a known speed
yeah going faster is often a problem
it affects how far the table moves when trying to stop
slower is probably always? fine
and it may affect the switch action as well - but I don't know about that
for some switch types, moving too fast may make you run past the switch as well
jepler: I get spindle +- buttons, but no on/off buttons
and spindle override starts at 0%, which can't be what we want
cradek: starts at 100% for me
why does jog speed go up to 3600ipm?
03jepler 07HEAD * 10emc2/configs/sim/lathe.hal: link something to spindle-forward to get the buttons in axis
hmm, I got a 100% start that time
03jepler 07HEAD * 10emc2/configs/sim/lathe.ini: fix jog velocity
I sometimes get 0% and sometimes 100%
I haven't gotten 0% yet, but I've seen that sometimes with FO
Issuing EMC_TRAJ_SET_SPINDLE_SCALE -- (+233,+20, +3,0.000000,)
I got it that time. Something issued this.
File "/usr/lib/python2.4/lib-tk/Tkinter.py", line 3035, in tag_remove
TclError: invalid command name ".bottom.t.text"
when exiting by clicking the WM's X
really? not here.
did you do it just as it was starting up?
it *is* AXIS sending the spindle scale command
*** set_spindlerate '0'
Issuing EMC_TRAJ_SET_SPINDLE_SCALE -- (+233,+20, +3,0.000000,)
somebody knows what is
03cradek 07HEAD * 10emc2/nc_files/g76.ngc: g76 demo
spindle speed needs to be non zero in order to enable
nobody can help me ?
Didier: In emc 2.0.x, you can't issue a command like M3 S0, or M3 without a previous S value.
maybe you can just write S with some value you don't care about
wow - I can download at 325KB/s from my web site.
WI, US 11/03/2006 6:33 A.M. OUT FOR DELIVERY
You got it. :)
so.... who is going to buy the core2 quadro?
I just hope AMD has a quad-core upgrade path for socket 940, eventually
but I'm not holding my breath
03alex_joni 07HEAD * 10emc2/docs/src/common/tkemc.png: updated to latest version of tkemc (so_slider and optional_stop_button)
mtedad_ is now known as mtedad
alex_joni: fo_mode is a char, but you assign 1.0 to it when initializing
oops, wrong channel
03alex_joni 07HEAD * 10emc2/src/emc/motion/motion.c: meant unsigned char not float (thanks chris for spotting this)
I've got a mill running using steppers and it works pretty well (thanks for some awesome software), however my steppers came with encoders attached (that I am not currently using). I have been trying to figure out how to use them without buying a dedicated encoder board. It seems to me that if the fast HAL thread can generate step pulses real time then it should be able to count the encoder pulses as well. It would be nice to have real feedback and not to have
spat74, reading encoders on the parallel port (or other IO card) is easy, there's a HAL component to do that
is there ?
what's harder is dealing with the information you get from the encoder
what is it called ?
but you can't assume the same frequency as the counts output
wonder who came up with that name :D
Charles Babbage or somebody from that era, I bet ;)
spat74: how many encoder lines / rotation ?
not sure just yet, All I know right now is that there are 4 extra wired from the back of my steppers.
judging from the lines on the encoder disk, they can't be too different than the 1.8 degrees the stepper steps through
ok - that's good
not _very_ good
but it could be worse
the software can only read the encoder at about the same rate as it can output steps (or microsteps)
SWPadnos: half I read somewhere
a little less actually, considering jitter issues
Well, I am not above limiting my step rate to get feedback :)
no, I'd think double if anything, since only one transition is needed for a quadrature encoder, but two are needed for a step
The driver exports variables for each counters inputs and output.
It also exports two functions. "encoder.update-counters" must be
called in a high speed thread, at least twice the maximum desired
count rate. "encoder.capture-position" can be called at a much
slower rate, and updates the output variables.
ok - that's the jitter problem
spat74: can you say exactly what your goal is? People have different ideas about steppers + encoders
it's better to look at the encoder lines significantly more often than you expect transitions
I really just want to detect missed steps
are you having a problem with that?
only when I mess something up
I see, like a crash?
if I set my feedrate to high, or a crash
you want it to stop and not lose position in that case?
I really should have more powerful steppers, but I got these cheap :)
that would be perfect :)
just generating a following error would make me happy
right I see, that does seem promising
encoder on the back of the stepper
there was some talk about this on the emc-users list very recently
I'll have to look through the archives, thanks
sadly they're broken (on sourceforge)
well, even if all you could do is point me to the name of the hal module I'll do my own homework from there :)
if you can get some hard info about your encoders (number of lines) we can help you do the math to see how it might work out to read them in software
it's called encoder
spat74: you came at a time that some of us are interested in this setup, so it would be great if you can get the hardware end set up and help us
Jymmm: network issues.. ssh broke down
I was planning to just set up the counter software, then do a bunch of moves with the stepper onloaded (off the mill) so no lost steps and get the data from there.
you load it with a command like "halcmd loadrt encoder num_chan=3" (or however many you want)
alex_joni how did that happen?
Jymmm: pesky connection at home (cable)
it's just called encoder ?
spat74: good idea, you could also just turn it one turn by hand and read the counts. It'll surely be a power of two or ten.
Jymmm: irc client is at work, I ssh in
alex_joni ah. doesn't auto-reestablish?
Jymmm: ssh doesn't
or 360 ;)
alex_joni so you're not using keys huh?
I meant power of two or MULTIPLE of ten
Jymmm: not usually
From the look of the encoder wheel I'd bet it's not more than 2 counts per full step
do you mean two edges or two lines (eight edges)?
spat74: do you microstep?
nope, just full steps
I have the steppers output geared down going to the mill'
spat74: for clarification: one slot on the encoder converts to 4 counted edges
I would say one or 2 solid color stripes per step
so 4-8 edges
to be able to count them all, you'll need to go slower with the max pulse rate
4-8 times slower
or you can get a small circuit to scale the encoder counts down
actually my max pulse rate is limited by the stepper's not losing steps right now, so the software has plenty of headroom
spat74: oh, ok.. that's ok
if you can set your stepper driver to do half steps or microsteps it might boost your steppers performance
if you're lucky enough to have a parallel port that can do control port inputs, you have enough to work with
or a cheapish second parport
I can do either, but I thought you got your highest torque from full steps
yes always use at least half stepping if you can
spat74: what drivers?
on my mill, full stepping can barely move the machine
I'll probably rumage around I've got some extra parallel port boards somewhere
good to know, I'll have to do that.
do try half stepping, you'll be surprised
cradek: perhaps I should try halfstepping my machine instead )
actually - as of yesterday - the sourceforge email archive is up to date.
I've been running full stepping.
skunkwork: good to know, thanks for pursuing that
anonimasu: my understanding (and experience) is that full stepping is the worst way to run a stepper
cradek: why didnt you tell me this before?! ^_^
and i"m planning to add some other stuff that will require output pins
anonimasu: you didn't ask
cradek: that's really nice :)
I might be able to get a bit more speed
* spat74 checking archive
spat74: then a second/third parport might be the cheapest solution
yeah that's a good way to get a lot of inputs for very cheap
hey, I might only have 2 axis CNC'd now, but my goal is 4 :)
anonimasu: you were using a G201 ?
plus seperate limits and homes for all axis
spat74: emc2 will drive what you can come up with :)
spat74: that's a nice luxury
anonimasu: set it to 10 microsteps :)
that's what I love about it :)
alex_joni: hm, that's steps_now * 10
well, thanks everybody for all the help
anonimasu: err.. the G210 is still a 10-microstep drive
alex_joni: ah, yeah right it's just my input
spat74: you're welcome, keep us updated
try getting a response like that from a commercial vendor's tech line...
anonimasu: guess you can't drive it better than that
spat74: our pleasure
Jymmm: rubber duck ?
there's a simple solution for this it's called servo.
alex_joni ask anonimasu
lots of pasta.
for ½ a year..
water and bread
bread's probably more expensive then pasta :D
not over here
you still need to boil the pasta :D
I could do that over at work :D
that's additional water & energy
and at least some salt
update on my problems and I've read the manual several times-EMC2 will come up with the sim-axis config, will not come up with univpwm because I don't have their card installed, which I understand. It will not load the nist-lathe config from this address http://cvs.linuxcnc.org/cvs/emc2/configs/nist-lathe/.
Can someone check out my problem with this one is. Thanks
that one doesn't match your installed software
Nick001: you cannot mix configurations from the stable version you have and from the development (probably still buggy) version
someone gave me that address so I could come up with a lathe type page with the proper type of axis offsets
Nick001: the 2.0.x versiond doesn't have much lathe support
most of the lathe development is quite recent, and will be included in the next release
any suggestions on how I use EMC2 to control a Hardinge HCNC. With the Allen Bradley control, one uses fairly simple G coding for axis and spindle control with single point threading.
Nick001: EMC2 can do all that too.
cradek: 2.0.x too ?
if you don't want to wait for the 2.1 release you will have to use the cvs (development) code
instructions for building it are on the wiki: wiki.linuxcnc.org
I'm apt-get source emc2 and its getting v 2.0.4 dsc & tar and I assume I will extract and install them. How much of lathe capability will this have, when is 2.1 release due out and how much will it support a lathe
2.0.4 source you are getting is exactly the same one as teh installed one
you won't have any improved functionality
the 2.1 is probably due out this year
any recommendations on what and how do I control this lathe with
I'd recommend you start uding emc2
to get the feeling of it
it can still control the lathe even if you don't use some advanced features like lathe-type tool compensation
Nick001: we can't talk you through doing a whole conversion of your lathe; it's just not practical
Nick001: I think you're going to have to dive in on your own, or hire someone who knows EMC to visit you and help
Nick001: if you want to do that, then it's best to ask on the Users mailing list (and please include your place of residence)
my usage is x and z axis with offsets in increments of 0.00005 which a 2000 line encoder will yield with the pico board. The spindle of course has to be controlled and constant surface speed would be nice. The tooling that I use is single point and basic in nature.
I don't think constant surface speed is in emc2 ____ yet :)
no, it's not
I have already tried the Ajax/Centroid control and it does not work. I already have the encoders mounted and the dc spindle controller is in place. I've contacted pico and he says his cards will work on this machine with the servos and encoders attached that I have.
lathe support is fairly recent
CSS I can wait for
Nick001: don't expect it to be a simple operation to plug the card in and it's done
that's usually the fairly easy/fast part
I know that. I have a lot of learning to do yet and I have to start somewhere.
Nick001: I bet Jon Elson (the pico guy) will be a big help
heh.. picoguy :D is there a nanoguy?
that's his big brother
bet you can't see the femtoguy :)
* skunkwork still only knows to pico.
probably out chasing the fems he's busy
03jepler 07HEAD * 10emc2/src/emc/usr_intf/axis/scripts/axis.py:
set AXIS_NO_AUTOCONFIGURE to disable auto-simplification of the GUI
Point the About dialog to linuxcnc.org
03jepler 07HEAD * 10emc2/share/axis/tcl/axis.tcl:
set AXIS_NO_AUTOCONFIGURE to disable auto-simplification of the GUI
Point the About dialog to linuxcnc.org
Lerneaen_Hydra_ is now known as Lerneaen_Hydra
SWPadnos: was the video playable?
that was skunkwork I think
this is a film about a dramatic relationship between man and fish
Yes - worked great - thanks
the fish doesn't think... because the fish knows.. everything
* Lerneaen_Hydra has no idea what alex is going on about
it is friday and it is past alex's bed time.
not on fridays it ain't
skunkwork: oh, right. of course
Try this address for searching the logs: http://www.linuxcnc.org/cgi-bin/search.cgi
jepler: raining there? :)
skunkwork: no, I was making some changes in my "server" machine
trying to monkey an additional disk into this silly SFF case
ah - I suppose it is the only thing you like about winter? No rain to take out your internet?
frustrating -- I got it in but the sides of the case will never go on now
snow can do it too, at least when it's melting :-P
did alex_joni go to bed? I want to complain that the search URL doesn't work -- it's 404'd/
jepler: there's no search.cgi on DH
I don't trust my cgi enough to put it there
I should probably remove the search tag from the logger
that would be best
there's a google search on linuxcnc.org
if it already indexed the irc logs it should work better than normal google
I was thinking that it would be nice if a link to the search page or the latest log could be stuck in the /topic
it's pretty frequent that someone asks the logger for the URL
[21:51:36] <alex_joni> http://www.google.com/custom?q=+site%3A126.96.36.199
03alex_joni 07HEAD * 10emc2/docs/src/gui/tkemc.lyx: first draft of tkemc documentation
03alex_joni 07HEAD * 10emc2/docs/src/gui/ (tkemc_interp.png tkemc_mdi.png tkemc_override_limits.png): some needed pictures for tkemc.lyx
03alex_joni 07HEAD * 10emc2/docs/src/motion/.cvsignore: ignore some more generated files
03alex_joni 07HEAD * 10emc2/docs/src/config/.cvsignore: ignore some more generated files
03alex_joni 07HEAD * 10emc2/docs/src/gui/tkemc.lyx: fix some oddlooking pictures
03alex_joni 07HEAD * 10emc2/docs/src/gui/tkemc.lyx: some ispell fixes.. (wonder where the rest are hiding)
[23:15:33] <alex_joni> http://www.robotpower.com/products/osmc_info.html
<- ever seen this one?
Q on cl, how do get more outputs for classicladder to show up in Halshow?
mtedad: you need to specify it to use more outputs on the line where you load it
it's usually in the .hal file
loadrt classicladder_rt numRungs=50 numBits=50 numWords=8 ... numPhysOutputs=40
there are quite a few options you can set there
check the sample config /etc/emc2/sample-configs/demo_mazak/demo_mazak.hal
parm: numSections:Number of sections to allocate (int)
parm: numArithmExpr:Number of arithmetic expressions to allocate (int)parm: numPhysOutputs:Number of outputs to allocate (int)
parm: numPhysInputs:Number of inputs to allocate (int)
parm: numMonostables:Number of monostables to allocate (int)
parm: numTimers:Number of timers to allocate (int)
parm: numWords:Number of words to allocate (int)
parm: numBits:Number of bits to allocate (int)
parm: numRungs:Number of rungs to allocate (int)
found it...same file i was working in...is there any relationships to the amounts to be put. like numbits a multiple of rungs/
that depends on what you need I think
* alex_joni heads to bed
good night all
thank you alex.
mtedad: no problem
i still have a problem though, when i add a signa like i-1, then add another i-10 they all get linked together. i.m using like i-a for i-10..went hex