03cradek 07HEAD * 10emc2/configs/max/ (max.hal max.ini): max now has an A axis and real spindle encoder
03cradek 07HEAD * 10emc2/configs/max/max.ini: fix ncfiles path, update threading program
03cradek 07HEAD * 10emc2/nc_files/threading.ngc: fix ncfiles path, update threading program
hey cradek. How is it going?
pretty dead tonight
it's been pretty dead since the big sf outage - I think we all got out of the habit of working on emc
this is reat - my "new" laptop agrees with rtai
I'm running my mill with it now
replacement keyboard is on the way...
yeah it was pepsied I think
I may not have to take a desktop machine to the workshop, which would be great
we had issues with old toshiba sattelites. they seemed to be a laminated diaphram - any liquid would wick in and stop them from working
I doubt most laptops like liquid much...
most things electronic are that way ;)
still making a mess with eagle :) http://www.electronicsam.com/images/KandT/eagleone.JPG
it isn't going to be pretty
cool... sometimes the autorouter doesn't give pretty
I always end up routing by hand to some extent
I don't have the patience for that - I think. :) - I will for the most part let eagle do it.
what kinda laptop?
huh - I have one of those at work
nothing fancy, but it has a real serial and parallel port, ethernet and modem built in
no built in wireless though
the serial port in it isn't supported in win 2000 :)
I found out.
this is like a 333mhz if i remember right
neither is opengl, which is the primary reason I have it
but who cares about windows?
I can't wait for vista to come out, because there will be great hardware for free everywhere
I have to :)
I had issues with turbocnc running on newer portable because of the power saving stuff I assume. Steppers would go tick tick tick as they where running.
stall at higher speeds.
some laptops seem to work for realtime, some don't, you just get lucky sometimes
my old laptop did not (would glitch when the processor fan went on or off)
so realtime has similar issues. Have to remember that.
it's just luck
try before you buy, if possible
just bring a live cd into best buy and boot it. (I suppose one stepper and drive also to test) :)
you could run the rtai latency test without hardware
you would see issues?
I think rtai even have a live test cd of some kind
another thing to remember :)
it's sure nice to see combined rotary/linear moves finally be right
nice work. You have a rotory axis on your machine?
yes, I've always had one but only recently made the driver for it
you really need to make some videos. :) threading and rotory axis
ouch, there are bogons in here
Issuing EMC_TASK_PLAN_EXECUTE -- (+509,+268, +88,g0a0,)
Issuing EMC_TRAJ_SET_VELOCITY -- (+205,+28, +0,0.000000,0.000000,)
Issuing EMC_TRAJ_SET_ACCELERATION -- (+206,+20, +0,0.000000,)
so, forget I said anything :-)
what am i looking at?
it needs to go somewhere in A, and decides to get there at zero velocity and zero acceleration
needless to say, it didn't get there
yes I'm sure it is
(gdb) p dy
$5 = 4.5102810375396984e-16
god I hate floating point
in other words, I found the problem
03cradek 07HEAD * 10emc2/src/emc/task/emccanon.cc:
this fixes a bug that made g0a90 set accel and vel to zero (because
the corresponding xyz move wasn't exactly 0 length)
again - nice work
Jymmm is now known as Red70sShow
Red70sShow is now known as Jymmm
Jymmm is now known as Jymmmm
SkunkWorks is now known as skunkworks
Jymmmm is now known as jymmmmm
03cradek 07v2_0_branch * 10emc2/src/emc/task/emccanon.cc:
this fixes a bug that made g0a90 set accel and vel to zero (because
the corresponding xyz move wasn't exactly 0 length)
jymmmmm is now known as Jymm
HELLO, i NEED SOME HELP URGENTLY
iS THERE ANYBODY TO HELP ME TO ACTIVATE MOTENC LITE CARD ON EMC2
hello mr dragan
I am new with emc, and i want to reconstruct one old cnc machine.
I already bought motenc lite card and want to activate it.
when emc starts i receive an error like motenc.0.encoder-read not found
Dragan: Do you have a working emc1 setup?
is somebody have an idea what can be the reason
i have ubuntu and emc2.install.sh installation passed well
did you check the wiki about montec?
do you have the Lite servo PCI?
do you use the montec config from the sample-configs?
I use montec config from sample-config, and I have Lite servo PCI
I can not find anything usefull on wiki
Do you have your board id set at 0?
Yes, I checked it was set by default
Just a thoght, since emc2 is not yet released it might be better to look up if there is montec support for emc1.
Which there is....
you are probably better off putting effort into emc2
robin_sz: dragan failed with emc2
emc2 will get there though
emc1 has "problems" .. its stable, but some bits are very icky
emc2 was written for a reason you know ;)
dragan: still there?
dragan: can you paste me the last few lines from dmesg ?
start emc2, chose the motenc config
i will try
then after it fails, open a terminal and run dmesg
MOTENC: Card is MOTENC-Lite, ID: 2
[ 4578.102138] MOTION: cleanup_module() started.
[ 4578.102144] HAL: threads stopped
[ 4578.102201] MOTION: cleanup_module() finished.
[ 4578.188585] RTAPI: WARNING: tried to delete task 03 while running
[ 4578.188597] RTAPI: WARNING: tried to delete task 02 while running
[ 4578.188602] RTAPI: WARNING: tried to delete task 01 while running
[ 4578.200249] RTAPI: Exiting
RTAPI: Exit complete
[ 4578.211448] RTAI[math]: unloaded.
[ 4578.274638] RTAI[malloc]: vfreed extent d0c0e000, size 2097152.
[ 4578.274760] RTAI[malloc]: unloaded.
[ 4578.374634] RTAI[sched_lxrt]: unloaded (forced hard/soft/hard transitions: traps 0, syscalls 0).
[ 4578.385852] Adeos: Domain RTAI unregistered.
[ 4578.385859] RTAI[hal]: unmounted.
[ 4578.397131] Adeos: Pipelining stopped.
ok, the problem is at the first line
Card is MOTENC-Lite, ID: 2
that means the card is not set up with id=0
either the jumper is not in the right place, or the contact is not good enough
I see, I will check the jumper again
Thank you very much for help, now I have to switch of the computer and check the card
the preference link on the wiki does not work... I can not ad stuff
chinamill: you can only add stuff if you are registered
dragan: absolutely no problem, that's why we are around
dragan: there might be another way, instead of changing the card to id=0, you can change emc2 to use id=2
aj: again You become a hero ;)
dragan: but that means that you'll probably have to do that in the future too, if there are any changes ;)
chinamill: yeah right :P
how, in some configurations files
chinamill: did you read the BasicSteps?
yes I read
dragan: yes, what config were you reading?
did you chose /etc/sample-configs/motenc/ ?
aj: yes, but the preference link is *ucked up
chinamill: works for me..
alex: i use this original files
chinamill: the first time, you need to add a username, and a password to get remembered
dragan: ok, the BEST way would be to set up a local copy of that (~/emc2/configs/mymotenc/)
~ means your home directory
dragan: how much *nix experience do you have?
I work in the past on the sco unix
ok, so you know your way around copying files ;)
I get: Could not get user-ID lock at /home/groups/e/em/emc/cgi-bin/emcinfo.pl line 3682.
copy all files from /etc/emc2/sample-configs/motenc to ~/emc2/configs/mymotenc/
dragan: then you need to edit all *.hal files in that location
inside the hal files you'll see some hal commands, you need to replace all 'motenc.0.' with 'motenc.2.'
and the next time you run emc, make sure you pick your local motenc config, not the one in /etc/emc2/sample-configs
ok, i will do it i need some time, this means that numbers after motenc means id number
what is the direct link to login for the wiki?
chinamill: you need to go to preferences
maybe you used the same username, but with a different password?
dragan: ok, I'm off for lunch, but I'll be around later.. hope you'll succeed
happy easter to the ones who celebrate it now
Thank you and goodby
Guten Morgen Alex
jepler: Did you get the mail with the swedish translation?
jepler: Updated in the CVS and all?
jepler: what was teh URL to your Gcode page (the most common gcodes used)?
[13:51:30] <jepler> http://axis.unpy.net/files/gcode.html
Have we got servo tuning guys around?
* robin_sz returns from starting the bike
Where can I create an account for the wiki? What url do i use to log in?
you don't need an account
you just need to know the super-secret admin password (emc)
enter your name in the name field, then enter the password into the admin password box, and the "edit page" link will be enabled at the bottom of each page
(you may need to reload to get the edit link)
[15:17:57] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?BasicSteps
Now it works like before... I tried to go to the link preferences before and it was broken somhow, but now it looks to be fine..
ok - that's good
there is something odd in emc2 traj planner/blending
velocity doesn't stay constant in line joints even on straight path
bug or bad tuning or something?
I tried code like this:
G01 X10 F500
sounds like it accelerates for short time at joints
G91 is relative motion mode, so that code actually moves 30mm in X?
is 10mm enough distance for your machine to reach its top speed and then stop again?
I think that code should behave like G01 X30 but it sounds different, I can hear the segments
what path mode is it in? (g61, G61.1, G64)
let me see what halscope shows
I blended mode
I've added a G64 P.01 to my test file
which signal should I plot?
In the simulator configurations, there are Xvel and Xacc (velocity and acceleration) signals. That's what I am going to plot.
I noticed the problem when running the 3d_chips.ngc
I didn't find any velocity or acceleration signals on stepper configuration
They're created by some lines in the .hal file for the simulator
[15:32:51] <cradek> http://timeguy.com/cradek-files/emc/colinear.png
so there is a small glitch in the acceleration
this happens because accel is high, allowing an accel to the given feed in very few trajectory periods (approximately one). That makes the accel and decel a bit asymmetrical
it's mostly a time quantization problem, very hard to fix exactly with the current algorithm
* jepler was still struggling to get halscope to trigger at the right time
ok, it's not a serious problem
you can probably minimize it by turning down your TRAJ period
but my acceleration is'n that fast. I have about 200mm/s^2
takes about 0.05 seconds to accelerate to 500 mm/min
I wonder what your traj period is
i'll try to adjust it
ok, 1 ~= 5
try .005 or .002?
or just ignore it, if it doesn't impact your part finish
just thought it was a bug or something
I see, I guess it sort is, but it's not the easy kind of bug
0.005 didn't help a bit 0.002 makes it unnoticeable
problem solved :)
well problem beaten with a stick until it hid, not really solved
gotta go make niceynice with the family, be back later
when in doubt, get a bigger hammer
hey Ray, you awake?
seems like changing traj period didn't fix it completedly. .002 just happened to be nice with F500, with other feed rates the problem stays
I didn't notice that lowering period would change amplitude of joint accelerations
why it's not possible to set traj period equal to servo period?
emc2 start fails when I do that
What's up skunkworks
was going to let you know the link is setup between the two buildings. I just need a router on the cable modem end. (that computer was hooked directly into the cable modem - I didn't notice that.
but I was able to get to both the access point and wireless router from both sides.
so far so good.
Fantastic. I've got to do something here.
I have seen it done with a cantinna in the focal point of a old directv dish. Suposidly able to get over 5 miles.
Really. That sounds like it might work.
I just need to do a little terrain analysis
The lake thawed out Wednesday.
Nice link. Thanks.
I seem to be having a strange issue (maybe not an issue) with the output from the parport. I have set the hal config to mode 2,2,2 (grey code) and when looking at the waveform on a physical osciliscope (not halscope) the all 1/0 (high/low) changes are locked to a 50µs resolution. I seem to recall that the RT part of EMC ran at 50µs, so is that what I am seeing? (It seems to perform a...
...psuedo-pwm, as it quickly hops back and forth between two time states in a way that looks like PWM)
Lerneaen_Hydra: it only can change the status of the parport at 50 usecs
if you demand another frequency it will interpolate
for 70 usec it will do sometimes 50 sometimes 100
if you get my meaning
alex_joni: Is that a physical limitation of the parport?
yes that is what I saw, the psuedo PWM
Lerneaen_Hydra: not of the parport but rather ISA, and the way the parport gets accessed by the processor
what PC are you using?
curently running an Intel copperming PIII (socket is similar to socket A) on an intel board. Integrated parport
what frequency processor?
ok, then you probably can run faster than 50 usec
I don't recall the fsb speed
try lowering the BASE_PERIOD
if you put it too low the PC will freeze ;) then you need to go up again
RT tasks taking over the machine ;)
thats rather.. frightening
no, but the RT stuff runs above the linux kernel
so when that takes too much time up, the rest can't run anymore
above being nearer of further away from the hardware than the kernel=
nearer to the hardware
ok, then that makes sense
what would a reasonable value for testing be?
and in which file is that parameter located?
Lerneaen_Hydra: I ran down to 5.5 usec on my Athlon 1600+
it is located in the ini file you are using
BASE_PERIOD = 50000
that's the default value.. (50000 nsec = 50 usec)
start with 20000, that should be ok on your machine (I hope)
then go down to 15000, 10000
and on another track, I've seen differing reports as to how much current one can safely draw from the paralell port, some have said 20mA, and some 200µa, do you know anything as to that (at the moment most places seem to have reported 4mA)
although 10-15 usec might be problematic (jitter)
is there any parport limitation?
usually it depends on the implementation
but for most any application you should have something between the parport and chips that need current
either transisters, optocouplers, buffers, etc.
but running an optocoupler as-is is ok? one doesn't need a pre-amplifier for the optocoupler?
no, you only need to turn on the diode inside the optocoupler
and that is usually around 1-3mA
which should work
[17:01:56] <jepler> http://www.fapo.com/1284elec.htm
Lerneaen_Hydra: cool.. look at the output
even if it works it might jitter
it seems to behave correctly
Lerneaen_Hydra: ok then ;)
looks the same except the timing window is 20µs instead of 50
I seem to get jitter at 15000 though
it jumps between three states somewhat sporadicly when moving through timeing windows
ok, then 20 is your value
03alex_joni 07HEAD * 10emc2/src/emc/usr_intf/xemc.cc: initial version of xemc.cc, brought forward from emc1
03alex_joni 07HEAD * 10emc2/src/Makefile.inc.in: added xemc to the build process, need to check for Xaw in ./configure next
03alex_joni 07HEAD * 10emc2/src/emc/usr_intf/Submakefile: added xemc to the build process, need to check for Xaw in ./configure next
Jymm is now known as MrAsshole
03jepler 07HEAD * 10emc2/app-defaults/XEmc: resource file to give XEmc the proper appearance
MrAsshole is now known as Jymmm
03jepler 07HEAD * 10emc2/scripts/emc.in: set XFILESEARCHPATH so that XEmc is found for RIP
that's untested for run-installed, but if run-installed puts it in one of the default locations everything should be dandy
03alex_joni 07HEAD * 10emc2/src/ (Makefile.inc.in configure configure.in): proper way of compiling xemc (in case configure finds Xaw-dev)
03alex_joni 07HEAD * 10emc2/src/emc/usr_intf/ (Submakefile xemc.cc): proper way of compiling xemc (in case configure finds Xaw-dev)
jepler: can I add that to make install?
alex_joni: you mean copying XEmc to /etc/X11/apt-defaults? Makes sense to me to add it.
03alex_joni 07HEAD * 10emc2/debian/ (control emc2.files): added libxaw7-dev to the list of deps, note: emc2 will compile even without
03jepler 07HEAD * 10emc2/configs/sim/xemc.ini: simulator variant to run xemc
03alex_joni 07HEAD * 10emc2/src/Makefile: install XEmc by default
03alex_joni 07HEAD * 10emc2/src/ (Makefile.inc.in configure configure.in): added ./configure tests for ncurses (needed for keystick)
03alex_joni 07HEAD * 10emc2/src/emc/usr_intf/Submakefile: added ./configure tests for ncurses (needed for keystick)
will the freqgen pwm mode output a pwm signal without feedback (encoder)
if it does is it proportional to velocity of some kind?
skunkworks: the number of "up" cycles is position-scale (the parameter) times velocity (the pin)
so for instance if you want to have a velocity input that goes from 0 to 1, with 1 giving a 100%-up signal, set freqgen.X.position-scale equal to freqgen.X.maxfreq
Jepler: thanks again. I will play with it.
what are you trying to drive with freqgen?
I wanted an easy way to test the driver bridge without hooking up feedback yet.
yeah that should work fine
that is the plan then :)
jepler: You familure with EPS format at all by chance?
Jymmm: I know a tiny but of postscript, but not much
I can't be held responsible for the typos. There's a cat sitting on the laptop
Tiny butt is good.
Weneger is besser.
I think I might build a laser tub cutter
started on the design a bit
probably a 4 axis machine
the 5 axis ones are too complicated, and dont really offer any advantage
ideal job for emc by the look of it