qq- it might not be stable
or work with what i have
is there a way to debianize this thing? like dpkg it?
it is last stable
so I can remove it if I decide I don't like it
there are debs
well now why didn't you say so?
I'm all about aptitude install over here
pcncx, to learn things
I assumed they would be ubuntu/kernel specific
[00:02:07] <andypugh> http://www.linuxcnc.org/mozmck/
andypugh oh if they're not for Lenny then yes I'd be better off making my own
I need to learn how to make my own debs
pcncx, emc2 is difficult ...
theorb is now known as theorbtwo
pcncx, go compile for now , see what it act , and later if you are ok , see how to build a debian package
I have been compiling emc2 about 100 times a day for the last week
man for the old piece of junk this machine it is it still runs like greazed lightning!
see what/ see how*
hmm the tests file was rather underwhelming
hey there is a debian directory
I just wish there were some directions in it
hrm what kind of coke is this stuff smoking?
EMC2 requires the real-time kernel 2.6.32-122-rtai to run.
how I fix that?
the kernel version ir runs with is hard coded into the software?
pcncx, as said use your /home/paulcnc/EMC2/emc2-2.4.4
qq- yes i just did a ./configure help and see the switch now thanks
it just pointed me to install which I read
oh well I don't get it I've no idea what this switch is looking for
building emc what exactly is the --with-realtime= switch looking for in ./configure ?
I guess I'll have to sleep on all of this over here
oh well better luck tomorrow night all
[#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/,
wiki at http://wiki.linuxcnc.org/"
and now I do.
the price of Alibre is going up... I think they are fishing for sales by letting people know...
srpski is now known as Dannyboy
[16:40:19] <elmo40> http://www.youtube.com/watch?v=InM6xn0AJCU
Good morning to all :)
I had a productive night lastnight, had to chang computers out for my cnc project with one that would read teh emc install disk but got it installed finally!!
After that managed to get an hour of reading through the HAL manual, and start trying the intr to HAL programing...made it through the Halmeter section before I called it a night :)
Yea skunkworks, feeling pretty gooad about the progress, now I am wiring the control cabiet for my machine as I have had the components mounted in there for months now...just not time to move it forward hehe
"good" not "gooad" damn kb glitches ....need to re-read what I type to catch them
Nice work andy
Me? (Or is there another Andy?)
just curious, is there a graphical version of the halmeter so I can see the waveforms of signals rather than just a number?
sweet, thanks looking forward to tryig that when I get home :)
It's well worth getting familiar with Halscope, it is one of EMC2's "Killer Apps"
next, you need halAI, it guesses what you want to see and shows it to you.
At that point 'HAL' becomes an unfortunate name "I'm sorry Dave, I can't let you do that"
andypugh: you :)
I guess 3HP servos are getting on for the sort of size your machine needs?
[18:50:03] <cradek> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=330476132729
I suspect that they are trying to bypass eBay
is it for the one, or fore everything in the pic?
atmega: the one
there's hundreds of them listed (individually) for $10k each
must be 'free insertion fee' week
yeah, guess so
No, it is technically for none of them, the Ad is to get you to contact them to buy the individual holders outside of eBay.
[18:52:39] <skunkworks> http://www.youtube.com/watch?v=4g-SlFrxPWM
that one has 'make offer', could be legit if not obnoxious
skunkworks: wow that's an ugly-sounding machine
wonder what all the rattling is when Z moves
heh - didn't have the sound on
also - the spindle is interesting. Wish I could see the whole machine.
yeah some kind of centering contraption?
[18:58:34] <atmega> http://raleigh.craigslist.org/tls/2018012990.html
is that actually considered a mill or just a big drill press?
I smoked my 3 axis chineee ts6050 board using 24v psu
I think you'd call that a mill-drill
also a round column. not that great for cnc'ing
you can see dovetail ways on Y. it's meant for milling I think.
no, not that great for anything I bet - it's a compromise machine
lathe looks more useful
so they replaced it with a 5 axis variant ,anyone know if its the same driver?
Aye, the lathe is nice (but like all US lathes has the longitudinal handle in the wrong place)
cradek: could you cut aluminum/brass with any precision with it?
atmega: could I? maybe with enough care :-) could you? no idea...
(I offered $50 for one of his $10k tools...)
Bluntz: I only know of two driver ICs in use on those cheap boards. The other one is more fragile..
atmega: I reckon as long as you use the quill for vertical, not the head.
the screws are probably useless for precise work - you'd have to do surface plate layout for anything that needed to be in the right place
but you could do basic machining with it, sure
However, I reckon it is probably an order of magnitude better in all respects than my mill.
kind of pricey for an unknown
no room for the lathe
atmega: This is a cool combination machine: http://cgi.ebay.co.uk/SIOME-TC-5-RARE-UNUSUAL-Lathe-AND-milling-machine-/150509451835
(Though why folk advertise "rare" in relation to machine tools is a mystery to me. It means "no spare parts")
the book shows a little milling table, but I don't see it
I was thinking about purchasing an old used unimat lathe for a conversion
Any machine you get, you will wish was bigger.
That will happen on about the second job with a Unimat.
A Unimat is too small to even peel some carrots.
The hal pin name length limit is too short
what is the pin name length limit set at?
So, hm2_5i23.0.sserial.1.port.2.8i20.mac-current won't fit.
thanks, knowing that should save me some grief in my designing of my hal "circuit"
The alternative is hm2_5i23.0.sserial.1.8i20.1.max_current, and leave the user to guess which port it is wired into .
what happens if you increase it?
It's a #DEFINE in hal.h. I daredn't
I don't have 2 years to test every HAL component and combination
Is "serport.1.1" clear enough, do you think?
that is what users are for ;)
ok so the gantry design needs to accept 48" stock then I wonder if my 2.1 A astrosyns and this board are up to that
The size is less important than the cutting forces, and the forces will probably be typical.
But a big, heavy, gantry will move slowly with small motors
so a robotic arm design maybe?
andypugh: I changed it from 41 to 99 and everything seems to run
andypugh: would probably break configs that are nearly out of hal memory.
Does that waste shared memory do you think?
So, back to the sserial.1.1 idea?
you might want to increase the size of the hal shm block at the same time to avoid breaking any configs?
#define HAL_SIZE 262000
Well, 41 chars is already quite enough to be typing in halrun or the test-hal-command window.
if you'd increase this the same ratio it seems like you wouldn't break any configs.
that's sure true, but there is tab completion
Not in "Show machine config"
(It would be nice if you could at least copy-paste in that window)
(I guess I don't use that)
don't you click on the names in that?
You can type "setp" commands in the box at the bottom. That is often rather handy
I thought jepler added that.
And he might have done, I have not done a pull for months
never mind - I was thinking this was the same... r75ea5e76057b /tcl/bin/halshow.tcl: halshow: permit an unlimited number of levels
has anyone ever heard of emc2-2.4.4 disabling input shortly after it is run?
what do you mean by disabling input?
I mean I can move my mouse although the cursor inverts but I cannot click the buttons and the keyboard is completely unresponsive i have to power off the machine
I am at a total loss over here
Video driver? I think that they can be involved in mouse input?
I am using r128 and drm
I am a totaly nut job
the only thing I can do is move the cursor on the screen but it is like it isn't even really there I have auto raise on but wundows don't come up
and the mouse turns black
like the arrow
though before the rebuild it was like a big black spot
but I just rebuilt the whole thing again and notived it is arrow shaped again
and a loony too
I even tried to run emc in -d debug but it left me nothing
should i change the directory it leaves the log file in?
right now it tries to leave it in /tmp
Try changing "axis" to "xemc" in the INI file.
andypugh at this stage of the game I'm ready to stand on my head and spit wooden nickels if it gets me a bit closer
You should get a much less graphically intensive UI to experiment with.
pcncx: does "r128 and drm" mean you've installed the closed-source nvidia or ati drivers?
cradek no those are the open source plane jane GPL drivers
Only change it in the [DISPLAY] section, not the axis definitions :-)
cradek: r128 is open source radeon driver
but pretty old
pcncx: are you using the linuxcnc install?
as i remember
cradek no this is a bare debian system with rtai kernel
I mean are you using our custom ubuntu install cd, or some other os?
it seems to work otherwise
so we've got many many unknowns :-/
pcncx: try to use vesa driver first
do your rtai runtests work right?
cradek yes I'm looking for the one unknown that is messing this up
you mean latency run and stuff?
and simple gl prog like glxgears works?
do you mean the rtai latency test or the emc latency test (or both)?
far as I can tell rtai is here and working
when I run emc I get the config screen then it dies after that
like with the splash still on the screen
the other window pops up but you're done
it's true it might be opengl related, but that's only a guess - realtime starts at that point too
you could try the emc latency-test program to see if realtime runs (it does not use opengl)
ok how i do that?
latency-test at the shell
ok I found it in scripts
ok bbiab maybe
. scripts/emc-environment; latency-test
it is running
and I'm still here!
that is a good sign right?
not for us ;)
qq- help me fix this and I promise I'll go away
true that is a good sign. did you run some glxgearses at the same time?
no let me do that
ack i got the black mouse but my keybard is working
pcncx, i asked for help on almost same behaviors yesterday , non helped me , so ... now your tour
cradek you're bringing it right up to the edge
pcncx: with glxgears + latency or only latency&
the latency test is going and I started glxgears then the mouse changed color
but I still got a keyboard!
so either kill glx or change driver to vesa
while you still have keyboard :)
yeah, I agree with psha; you want to try the vesa driver
ah this machine is on a power strip easy enough to restart
what are your latency-test numbers?
11258 servo and 13443 base
though lastn ight fooling with run latency I hit like 16 and change with glxgears dragging iceweasel around
it never went past that spike
was like 16558
not so bad
I figure 17,000 setting and i should be good
I don't plan on making any surgical instruments
the servo-thread runs with 250k? so 16k is already 6% of it?
but til i get this working I'm not going anywhere with this project
I will use the stepper module
pcncx: replace r128 with vesa in xorg.conf and try again
OK so my next plan of attack is lose r128 and just run vesa?
another way is to install libgl1-mesa-swx11 instead of libgl1-mesa-dri
if it work you may try replacing DRI glx with software one
with r128 driver
but vesa is more safe way in most cases
yes i will try that first
software rendering isnt that slow after all
oh just hit a new high on my jitter
cpresser: rendering is not bad, but without page flips scrolling looks really bad
software rendering for EMC is the way to go
anyone using gladevcp here?
it always works right. it doesn't mess with latency. it's plenty fast enough for AXIS which just draws some lines.
wow when i killed glxgears my cursor came back
so it's defenitely glx + rtai issue
gonno just shut off latency now I'm going to try glxgears again
nope its black
there is something seriously wrong with mesa on my machine
r128 is pretty old and drm subsystem is changing very fast in last 2-3 years...
so together they're bad huh?
so nothing strange that there are some strange bugs with DRM
try vesa :)
well I got to say I am not too thrilled with r128
gives me maybe 300 fps in gears
so no great loss
yeah let me do the xorg.conf
one more question what do I do with drm?
I don't think vesa uses it
you mean LoadModule "drm"?
well in general do I need to get rid of it do i need it I don't know
AFAIK, X.org will load required modules based on dependecy automatically
I have Load "dri"
cpresser even if you have a conf file?
I thought the conf file was a show stopper for X
pcncx: conf file instructs what you want from X, but don't mandate what steps to do and what - not
like there is no way to explicitely tell X not to load something?
check the log-file. my X-installton loats a lot of modules, i did not explicit instruct to use
I could always del the module off my system
then see what X does about that
just try Driver "vesa"
maybe there is some 'blacklisting' but i'm not aware of it
better rename it :)
don't delete stuff first...
k i swapped r128 for vesa let me cp this conf to /etc
my server dies with oops :(
oops sounds like a serious problem.. does the backtrace help?
somewhere in xfs...
it's luck that i was logged in via ssh
so i've catched trace
k going to restart X here
but till tommorrow it's dead...
with my emc2.git public repo and patches waiting cmorley review...
so you dont have local access to fix it?
it's ~5-7km from me
and everyone there are sleeping already
maybe memory is broken..
this is going to take a while to get to work here
I'll let you know when I'm in X with a screen with vesa running
pcncx: problems with monitor-timings?
cpresser yeah looks like it I wrote down all the junk from xvidtune i got to setup modelines
is there like some config tool I can use?
boot single user, run 'X -configure', copy the xorg.conf.new to /etc/X11/xorg.conf, edit Driver to say "vesa", reboot
cradek yeah i did that
and what's the result?
but when i did my screen said mode not supported or soemthing
looked like out of range refresh
I got my modeline for vesa: start with legay-video-driver; run xrandr -q; copy modeline
is your video cable fully wired? screen made in the last decade or so?
its a flat panel
I had the same problem with my vesa-setup.. a lot of out-of-sync stuff
dvi or vga?
vga but it gets the dpms info
turn dpms off :)
least i think it does
you mean edid
there is an option for it
well whatever the card and screen do
check the xorg log and make sure it's reading it
pretty weird for it not to just work
yes i have a copy of the bombed log open now
story of my life
I'm used to it anymore
the only way is the hard way
(II) VESA(0): Attempting to use 75Hz refresh for mode "1280x1024" (11a)
I don't think my screen can do that
xvidtune said I was at like 60 Hz
hsync 63.98 v sync 60.02
so soon as i get this nonsense straightened out I should be good to go
then it will pick the right vesa mode for you
bizarre the xorg.conf that works has no modelines in it
maybe i should just take the ones out of the genned file?
I think i need to take a step back from all of this for a bit here
before I wear a hole in my HDD where xorg.conf is stored!
at least I know what the problem is now
whoah it works!
What did you do?
I renamed the kernel r128 module then rebooted
so now it can't load the danged thing
I tried to rename the xorf r128 driver as well but X wouldn't start
and now axis is up running and not flaking out
hope all your gremlins are really gone
well my gl performance is even worse
but hey it works right?
I thought it was going to take me days of fiddling around it still may
like why my machine won't use the vesa driver I've no idea
too bad we don't get a plain old xorg.conf anymore these days
monitor stays black and popus up a floating box that says input not supported
well I'm running a basically empty xorg.conf now
cradek: I was under the impression you could still make an xorg.conf
I have a full one in my 10.04 myth box
Valen you can
don't mean it is going to work
you can ... and then it also reads other conf files ... and then it also autoconfigures everything
yeah X does seem to have more of a mind of its own these days
yes more and more - it's often a good thing - not always
this is the rule you can hurt X in the conf file or X can go off and hurt itself
its a good thing when it works
when it doesn't work your screwed
like if you screw up your conf unless it is something simple like a fontpath you may be doomed
no, you just ssh in or drop to a console
exactly doomed to fixing it
I'd rather be able to fix it than having to do a windows and say, it didn't auto detect go buy a new video card
the though of getting a ned video adapter did cross my mind and still is with all of this
if i had another one lying around it'd be in there let me tell you
they are pretty cheap really
cheaper than me beating my head against the keyboard for days
you have some ATI card dont you?
yeah an antique rage 128
whats the host PC?
I can see how it got the name
a gateway E-3400
i mean PCI, PCI-E
ahh, they are getting rare
yeah I've no good spares lying around now
honestly I'd be tempted to blow $100 and one of the intel dual core atom mbos
are you using mesa?
I believe so
as in servo hardware
I just got plain stepper drivers
so parrallel port?
you know step and direction
yes parallel port
there was something mentioned about a firmware bug (that could be worked around) on the intel mbos
but apart from that it might be a cheap $100
you get a working system out of the box with a 10.04 install
and ~4000-6000 latency
2 cpu + running with isolcpus is sure a big win
if latency is important to you (software stepgen)
for servos or hardware stepgen, makes no difference
cradek: makes me feel better ;-P
also I run the servo loop faster
do you have a really high accel machine?
this box I don't think I'll do better than 20000 latency
that's really quite fine for the majority of systems
well for me for now it is going to have to be
when I retire my other PC maybe it'll run EMC better?
but if i spend anything on a PC it won't be for EMC right now I'm afraid
it'll be for a new main machine
cradek: our machine has some nonlinearity to it
the higher servo loop seems to help
even $100, thats 2 end mills
not where i buy my end mills it isn't
if I'm lucky i can get 100 endmills for $100
[22:31:58] <Valen> http://www.newegg.com/Product/Product.aspx?Item=N82E16813121399
I'm guessing your not talking about 12mm coated carbide
I'm talking about whatever I can score on the used market it varies
I buy them with the cosmoline still stuck to them or whatever that brown crud is
we use carbide
my mill isn't rigid enough for carbide
so I'm better off with HSS
honestly i can't keep up with all the grades of carbide and neither can 98% of the people running carbide tooling
I've researched it and it is quite involved
pcncx, so which X driver use now ?
qq- same old same old but I ditched the kernel module
it was the thing KOing me
so now use what ? this > r128.ko ?
I just took it out so it cannot load
it was doing something
because now glxgears gives me about half the framerates it used to
but axis isn't locking my system up
00:36<qq-> pcncx, so which X driver use now ?
so there is something going on there
same X driver just took out the kernel driver
there are two different drivers
X has one in some modules/drivers directory and there is one in the kernel modules
grep Driver /var/log/Xorg.0.log , says ?
and apparently X can still use its even if the kernel one isn't loaded
it is going to say r128 but I'll look
(II) R128: Driver for ATI Rage 128 chipsets:
i see, so no using vesa , thats what i needed to know
when i built my kernel I thought it was a good idea to include the kernel module in the configuration
so I did
and it worked up until I tried to run emc
well I tried to get X to use the vesa driver and it wouldn't
well it did but I got no display
yeah there is still stuff going on here I don't fully understand
what I do know is axis is up and running and my system isn't frozen solid
but there is one more thing I want to check out quickly with all of this that might be a factor
nope not a factor I'd started emc sued to root
works as my user as well
dang that kernel module is just poison to axis
ha ha I got Pc speaker sound effects!
that is too cute who did that?
is there a way to turn the sound off? or at least lower the volume?
<catching up> The issue with the D510MO is that the BIOS doesn't properly declare EPP mode. So it is fine for normal parallel port setups and for PCI card setups, but anything which uses the p-port for comms (Mesa 7i43, Pico PPMC, Pluto) needs to be persuaded to work. (EPP is perfectly functional, but is not detected as such).
is there a way to mute axis sim?
Plus points for the D510MO are that is is tiny and runs fine completely passively cooled.
oh are you running stepper/sim? run sim/axis instead
that is a cute trick, and nice for testing, but would get irritating after a while
cradek the sound effects do make it very realistic
its almost done I'll live
ha ha my last one didn't do that is it new?
it's not new
huh because I've run the sim in 8.04 packaged release
and I'd have remembered if it made that ungodly noise
um what config did you select?
does this help? Machine configuration file is 'axis.ini'
I picked the first axis sim off the menu
it shouldn't be making noise then...
stepper/sim_inch.ini makes noise. sim/axis.ini doesn't
yes the inch I think how can I tell for sure?
let me restart it and pay more attention
which one doesn't make noise?
Sample Configurations / sim / axis
not / stepper / sim*
OK got it I think let me run and see
program exceeds machine maximum on axis Z
read the beginning of the program
yeah i don't know too involved for me at my level keeps saying exceeds machine maximum on axis Z
your simulated machine has its Z all the way up
you need to move it down and set the origin on your pretend workpiece
jog down a few inches, touch off Z, 0
hmm it is running now
and no noise
I can see the argument that this is a good simulation. I am not convinced it is the most confidence-inspiring intro for noobs.
andypugh I had no idea my PC speaker was so loud!
I never heard it more than beep til that simulation
I was meaning the mandatory touch0off
yeah i did it 4 times and I still don't know why the 4th time was a charm for me
and if cradek hadn't been egging me on I'd have given up
now strangely I miss the horrible noises
some people are never happy...
yeah if you'd have told me i nthe middle of that last sim that I'd be missing those sounds I'd have looked at you funny
I think this machine has an on board PC speaker and I can't monkey with it to muffle it any
I remember one old system I took the PC speaker off the case and threw it into an empty glass wrapped in paper towels
Doom on that machine with PC speaker was a bit much
it sounded cool in the stuffed glass though!
this rocks I think I'm up and running over here thanks!
put a resistor in line with it
Valen its a little piezo unit right on the mobo
hit it with a hammer
fix it good
wouldn't be the first time I fixed a PC with that technique
I put a 28 oz Estwing through one once hit it so hard I cracked an HDD case inside it
never had a lick of problems out of any of the ther PCs that were around when I did that
the rest learned run or else!
I think that is when I began to realize that sometimes i just have to walk away from some things for a while
but I'm not kidding I worked on that problem for 3 days straight and when i mean straight i mean straight
I mean 72 continious hours
when i finally figured out what it was I was so pissed off I cracked the machine
but that was back when I ran Windows and there was only so much you could do you know?
Hello everybody. I am on a AMD Phenom II X4 955 Processor stepping 03 with RTAI from magma, 22.214.171.124 kernel, all the latest bleeding edge stuff (doesnt work on Ubuntu 10.04 live cd either) and latency-test shows up as all 0s
NTU_live well now that sounds good to me
also i am on the arch live cd atm, with gnome, and the EXACT same image, it works finee on my PC
i am on my dads PC atm
both 785 boards except i am on a 3-core +1 unlocked core where as this is a true 4 core CPU
NTU_live steal your dad's CPU he'll never know ...
* pcncx grows horns |_|pcncx
well I'm off to go veg out in front of the tube for a while
Anyone using a 5i20 +7133T and a +5v -5v
output of bash -x latency-test http://pastebin.com/EpT2gWHp
also when i start EMC i get: USRMOT: ERROR: command timeout and its repeated like 10 times then says /usr/bin/emc: line 358: 4652 Killed $EMCTASK -ini "$INIFILE"
i attached gdb to the PID of latency-test and here is what i got for bt: http://pastebin.com/SJibPLuV