03jepler 07HEAD * 10emc2/docs/src/l2hprep.sed: missing part to build html docs
03jepler 07HEAD * 10emc2/docs/src/Submakefile: fix moving html docs into place
there may be some downtime on cvs.linuxcnc.org later this evening. It should be short.
jepler_ is now known as jepler
jepler: Guess what???
is that the real robin_sz?
no meep, so guess no
could anyone please address me to some place where making axis to work is explained?
command does not exist in bin so I am supposed to compile again emc2
but I am not able to find the axis source to put in emc2/src
pier: for emc2.0.3 you need to get axis from an external source
axis.unpy.net, I'll get you a link in a second
in the mean time (after 2.0.3) axis is integrated in emc2, and will be released with 2.1.x
but I get a message
did you get the source package?
did you put it in emc2/src/ ?
yes and got it installed
did you reconfigure emc2? ./configure ?
did you recompile emc2? make ?
no I didn't do that
the configure process should print "Building AXIS from ..)
yes... I am absent minded :(
pier: no problem.. we're here to help ;)
checking for AXIS source... will build axis from axis-1.1.3
is that correct?
always getting the same error message
[07:50:59] <pier> http://pastebin.ca/147254
you should get a newer axis
1.1.3 is ancient
[07:56:44] <alex_joni> http://axis.unpy.net/files/downloads/01145931630/axis-1.3a2.tar.bz2
now going to the garage to get the machine working :)
does Axis need some 3d acceleration graphics?
it returns a Failure to allocate Z buffer
it should have some
but I think I heard about that before.. not sure what to do about it..
does it crash completely?
just shutdown and cleaning
pier_gar: try running glxgears
and see if that works
it returns the same message...
it used to run
pier_gar: it seems it's a bug with the video driver
check what video driver was detected
from googling around I see that it might be "SGI"
switching to "vesa" should make it work
very odd... glxinfo says direct rendering is enabled
pier_gar: hope my google skills are OK
I never managed to get that enabled before
pier_gar: yeah, but that means probably that it is enabled, and the driver doesn't support it
so it crashes
try switching to vesa, or maybe disable direct rendering (although I have no idea what that is :D)
how do I switch to vesa? xorgconfig?
I think so..
I usually look at /etc/X11/xorg.conf
but make a backup of that file
pier_gar: any better?
yes... thanks alex_joni
yet it is as slow as a slug >(
fumbling xorg.conf to see if any speed can be gained
with vesa you won't have hardware accel
but it will be working
not sure what to do to make it faster
usually hardware accel is not quite compatible with RT (at least the nvidia and ati fglxr drivers aren't)
alex_joni: setup the shared video memory in xorg.con solved
thanks for your help
I think I'll jot down some notes for slack users as soon as I setup the pin-out properly
* jepler waits for jymmm to show up
I think I have a solution to his problem
could be a little while. he's on the west coast
I don't follow all the comments, but having a binary protocol and forgetting to assign exactly one number for a protocol item seems like a problem. https://bugs.freedesktop.org/show_bug.cgi?id=5002
so Jymmm is using the xforward protocol - is that just X over SSH, or is it something different?
I tested using 'ssh -X' but I think he is using "xhost" on the display to allow tcp connections, and explicitly setting DISPLAY on the remote system.
Running an app that uses glDrawArrays() from a machine without graphics to a
machine with an NVidia card with NVidia's driver, produces a blank window.
... but I do this all the time I think
ho-lee crap - I just saw this on the DeskCNC site: "Requires same Screw Pitch on X and Y Axis for Circular Interpolation"
so you can cut ellipses by changing a leadscrew: it's a feature
could someone please give me a hand with machine configuration?
cannot figure out where I am making mistakes
I'll try, what's wrong?
I put the pinout in the standard_pinout.hal
linksp Xstep parport.0.pin-02-out
linksp Xdir parport.0.pin-03-out
linksp Xen parport.0.pin-16-out
but now I am stuck with it
browsed the manual and compared the difference with
the config working file I have with a BDI distro
but the axis won't move
I don't see Xen in standard_pinout, did you add that?
according to the mnual
so your X driver need a high on pin 16 to enable?
my boards need
what version of emc is this?
the last.. 2.0.3
on what OS?
linux ker 126.96.36.199 rtai 3.3
I am running it in place
ok so emc starts and runs fine but there's no output?
steppers are humming and still
what kind of processor and speed? I wonder if you have a plug-and-play parport that's not getting initialized
like me ...
I rmmod lp, parport_pc
what processor and speed is it?
as soon as I quit emc2 I must immediately tur off the board
otherwise axis begin to move
the same applies to bdi as well
not to dos cnc sw though
SWPLinux: any idea?
that baffles me...
me too :)
I had thought that pin 16 might be high by default, but that isn't the case here
pier_gar: what other pin is hooked to the Xen signal?
pier_gar: what happens when you first turn the computer on? do the motors move?
jepler: it's machine-on (comes from core_stepper)
cradek: oh ok
pier_gar: you should figure out what does that and fix it, it sounds dangerous
that only happens with linux and not under dos
that's not enough information for us to guess the cause
pier_gar: you mean that the motors move when you boot Linux, but not DOS?
except when he runs emc2, then they don't move at all
my advice: run emc2 when you don't want the motors to move, and exit it when you're ready to mill something
steppers are enabled but locked
jepler: yes that wouldn't be a problem
if I only could make those steppers to move
under BDI TASK is minimilltask
pier_gar: do you have access to an oscilloscope?
have you used a scope or meter to find out what voltage you get on the pins -- for instance, is Xstep or Xen just floating when emc2 is not running?
ah - so on BDI you can get it to work?
ok - no scope, BDI works
with a simple tester I noticed some "noise" out from some pins
with a simple c program I am able to enable and disable steppers
then you're one step ahead of me, because I can't even get a pin to change on my port ;)
but there is always a noise on the step and dir pins...
without a scope, how are you telling that there's noise?
so I am not able to increase/decrease the frequency
I know it is a crap method but there is some kind of signal that can be revealed by means of a tester with frequancy-meter
what frequency do you get?
and the steppers obviously move
when running slackware with realtime kernel, but not under dos?
then run emc2
does your system require use of 'probe_parport'?
as a metter of fact with and without rt kernel
before you unload 'parport_pc' is the noise present?
with and without parport module
I am getting mad
have you tried adding pull-up resistors (e.g., 4.7k on each line)?
no.. as I don't have such problems wunder dos
perhaps a TTL too
odd thing is that bdi works
and now I was positive it would have worked with emc2
pin 16 in particular can be totem-pole or open collector in different parport modes
it works under DOS and BDI, so I suspect something in Ububtu is done differently
SWPLinux: it's not ubuntu, it's slackware
err - that's what I meant ;)
I'd suspect EMC2, except that there are issues before it's started
perhaps you could hav a look at my lsmod output
pier_gar: can you try booting with an Ubuntu/emc liveCD, and see if the problems still exist before you run EMC2?
I switched again to slack because I wasn't able to get axis move with ubuntu too
ah - didn't realize that
have you tried emc2 on BDI?
but there was a problem with the headers
do you have the problems with motors moving before EMC2 is started?
someone in the list told me
probably jepler or cradek - they really know about that stuff
yes... if I switch on the board now (emc not running) the steppers move (the three of them)
this is booted into BDI?
no.. I am under slack now
but I can reboot
board is on and the stepper X locked, Y Z free
ok, so on ubuntu, the port seems to act the same as DOS, before running EMC2
no really... the steppers stopped after loading stepper_mm from the main menu
then I quit EMC and the steppers remained still
now I am looking for the configuration file
the board is on but the steppers still
one sec - I'm on the phone
SWPadnos: it's gas!
ok. off the phone now
it's gas? does that mean it works? :)
hmmm. I have a suggestion ... ;)
I'll try and copy the config files
what about the motors moving before you start EMC2?
does that happen on ubuntu as well>?
at the baginning
then after loading stepper_mm
and remained still after leaving emc too
can you cange the parallel port pin that's used for enable?
mmm that would require a bit of rummaging with wires but it might be possible
perhaps there is some process accessing parport under slack
I guess if you can prevent the "normal" parallel port drivers from loading, there may not be a startup problem
I'll have a go
probably some combination of CUPS and the 3 kernel parport drivers
(lp, ppdev, parport_pc, I think)
ok. good luck
thank you very much
you're welcome. (did I help? :) )
speaking of problems always helps to pinpoint the possible solutions
in any case looking at the steppers moves did really give me a bit of hope... now it is almost a question of honor :)
a duell with this disto
slack I mean
ok thank i am going to reboot
...a difference I noticed between ubuntu installation and slack in rtai modules is that the first loads adeos... could that be a clue?
depends how you built your kernel -- adeos can be a module or it can be built-in
lerman_ is now known as lerman
ooh, here's a DIY project for you: http://blog.modernmechanix.com/2006/08/24/build-a-fan-motor-television-receiver/
(it appears that you keep the picture in synch by constantly adjusting the fan speed with a rheostat )
... you will find plenty of amusement and a chance to do a bit of experimenting to further the cause of television, which is bound to improve with time.
that's a clever way of making the scanning disk
the wire spiral? yeah, agreed
anyone still around?
* ValarQ is
hi all (and the ones not here)
* ValarQ goes to bed
yeah, I plan that too in a minute or 10
had an financial audit today :/
SWPadnos: greatly, got home at 2130
personal or at the business?
business, for some EU project
not even big money :)
takes longer - they actually look at the decimal point ;)
hi jepler !
long time no see
you've been missed (oddly)
long holidays :P
how are you ?
the holyday in it ?
all ok ?
I'm in naples again now
travelling from a point to another ..
heh.. nice, siracuse was great
the highwat from Regio to Naples not that great :D
how I sayd .. :(
but not that bad.. only slow
got some photo from there ?
yeah, slow ..
there are work on progress there
oh. emc 2.0.3 is out, great :P
I'm from laptop, but the damn m$ sharing connection system work bad from k4ts pc .. :(
losting the wi-fi connection very often
giacus: on my site (http://dsplabs.cs.utt.ro/~juve/blog/
hello pier !
holydays over :(
yeah.. few days yet for me :-)
how is emc going there now ?
found the way ? ;P
banging my head against walls
guys in the list helped a lot
e.03 is running
yet not working
discovered something weird
I compiled the kernel without the adeos support
and with parport whic seem to hassle emc2
so I decided to compile for the tenth time
not able to find adeos support in menuconfi option
pier: what kernel?
you won't find adeos
it's called IPIPE now
o god... I think I am going to faint
pier: this happens to people who want to do it themselves :D
I am telneting to the garage to save my marriage
* alex_joni is kidding
I know.. but you can do more over ssh
patched once more rtai
running make gconfig
alex_joni: I think that we all people here like to do things otherwise we would all end up using wingate
alex_joni: IPIPE ought to be enabled?
I can send you a 2.6.15 config if you like (to look at, not use)
is it related to adeos?
I think it got renamed
nice very kind of you
[21:45:00] <alex_joni> http://wiki.linuxcnc.org/uploads/config-2.6.15-magma
out of curiosity, are you using the ADEOS patches from the RTAI patchset, or from the Adeos site?
RTAI is what I used
bye all :)
right - I remember when I was trying an upgrade on my Gentoo box that there was some concern about using the RTAI patches instead of the ADEOS ones - that's why I asked
I have been browsing to find out if applying adeos patch would be advisable
SWPadnos: but the ones from ADEOS are equally good I think
I should have worded that the other way - ADEOS = bad, RTAI = good
no - there were some differences
SWPadnos: hmm.. I remember it the other way around
but I eventualy back up to plain rtai 3.3
get rtai, get patch from gna.org/adeos
heh - well - I suspect you know more about it than I do, so I'd go your way on this :)
but I think it was the other way - that the ADEOS patches with RTAI were the ones to use
read about not applying adeos patch
SWPadnos: either way, the patch failed for me :D
had to bend and twist it a bit by hand
heh - from RtaiSteps: "On no account use any patches from www.adeos.org - The RTAI team modify the adeos patches and have introduced some Incompatibilities."
(on the emc wiki)
maybe that was the problem :)
SWPadnos: That's what I bumped into!
ok - so my memory isn't as bad as I remember ;)
[21:50:17] <pier> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?RtaiSteps
that's the page
hmmm - alex, do you know if the ARM port has support for the AT91SAM7x?
don't think so
I mostly have seen ARM7 support
bummer. I'd hate to have to add it :)
those ARM chips don't have MMUs do they?
* alex_joni goes to bed
hey - maybe we can take a look at RTAI-Lab, and make a HAL configurator ...
* alex_joni runs to bed
oh- nevermind. it uses Scilab.
* alex_joni hides under the bed
at least it's not LabView (hate)
I'm not here anymore
SWPadnos: still hate labview?
why yes, yes I do hate labView
that even caused a netsplit
* alex_joni seriously goes to bed
only Dan Falck was dropped here
well this sucks. my nice new PC Power & Cooling power supply ( http://www.pcpowercooling.com/products/viewproduct.php?show=S61EPS&view=techspecs
) won't power up my H8DCE motherboard.
I know the 750W version works, because I have a friend with two like that
and the only difference I see is the 12V rail has an extra 11A drive. but both of these models are higher current on 12A than the original in my case
geez 12V 49A .. and that's not the high-power one?
heh - nope :)
all to draw a few polygons?
the real beast: http://www.pcpowercooling.com/products/viewproduct.php?show=TC1KW4E&view=techspecs
what the hell do you do with all that power? Draw a few triangles and perform a few floating-point divisions?
their supplies are excellent as well - they'll do those power numbers over the full temperature range
well, I don't use most of it. I bought that supply because it has the correct connectors for my motherboard, and is about 100 times quieter than the original
the whole machine takes ~350W, including the two Dell monitors, running continuous compiles and GLXGears
(not that GLXGears is a real strain on the GPU)
I am also becoming more interested in quiet computers
performance is lower on the list
yeah - the slowest computer you can reasonably buy these days is about 5x the speed of the one I use daily
but mine is 5x as loud :)
I can't even hear the doorbell half the time, and I'm right at the bottom of the stairs, with my office door open
it's not quite as bad as cradek's big box though
ALPHA.. you done those parts yet...???
12 hrs wrk 12 week turnaround sounds like the landing gear buisness....LOL
send jymmm I think I may have found a workaround for your glx error. Try this in the terminal where you are about to start emc2: export LIBGL_NO_DRAWARRAYS=1
* jepler tries again, talking to memoserv this time