cool, my lathe works great, got both axes going
does it have a spindle encoder yet?
spindle speed control?
have to make another part for that
heh not yet
actually the spindle is sitting on another desk
03cradek 07HEAD * 10emc2/src/emc/motion/motion.h: fix possible mutex problem
for spindle speed control I'll need to use another parport, so I might put that one off for a while
but I need to get the encoder set up
it's so fast it should be forgiving about the exact spindle speed, unlike the nist lathe
cool, even with my metric leadscrews I end up with an integer number of counts per inch (9525)
how fast was the nist lathe?
26 in Z
also mine will be able to keep up with the spindle encoder at any spindle speed, so I should be able to thread at full speed
so.... 2000rpm, threading a 20TPI? wouldn't that mean you'd need to move at like 100IPM?
jepler, cradek, btw... thanks for the help lately with the RE stuff
I think the spindle tops out at about 1200 rpm in the belt setting I'll use most of the time (geared for power)
but more often I'll cut 32, 36, 40 tpi - smaller stuff
1200rpm @ 40 tpi, still requires 30ipm
that's till pretty fast.
that's about a m a min.
well... 30 x 2.54
Hi Alex, how goes it?
going on vacation today
for about 3 weeks :D
alex_joni: Ah, very cool. Where to?
Gonna see Jacky?
not sure ;) probably not
schön guten Morgen
I'm playing with GAIM... It's nice NOT to have 30K different chat clients. but not used to the UI
yeah.. gaim is nice
but I like irssi better ;)
I REALLY wish it had miltiple windows like mIRC does, stacked.
and it doesn't ?
It might, but I have NFC how to configure it for it =)
alex_joni: Enjoy your vacation! I'm calling it a night =)
alex_joni: I wish you a nice vacation!
alex_joni is now known as alex_joni_away
servo loop on the plc
03cradek 07HEAD * 10emc2/src/emc/motion/control.c: fix Volker's problem with teleop motion that was many times too fast
so the "factor of 10" was a good clue
yeah, but I didn't take it seriously (as an exact measure) and I didn't scope mine for some reason
so I found it the old fashioned way
cradek: that way being?
cradek: how is the lathe coming?
digging down through the properly-working things one layer at a time
the lathe is great, both axes are moving, and it's fast and strong
cradek: oh, right
cradek: so you can do 60ipm threading?
how was the tuning? what kind of folloing error?
I think I will be able to do that, yes, but it's untested
as far as tuning, I don't know what I'm doing, but after flailing for a while the following error is good (.004 mm during full-speed jog)
nice - what is top speed?
I have it set at 1500mm which is 60ipm, it could go faster but there's no need to push it
wow - that is very impressive - with emc doing all the servo loop.
that should be really fun to play with
not just that, it reads the encoders and makes pseudo-analog output with pwm
those where the little servos from the surplus place you sent me?
(a link to)
these are some that jmk gave me from his junkbox
ah - I remember now - at the workshop
the surplus center ones have gearheads I think, they would be way too slow
these spin FAST - my leadscrews are only 1mm travel/revolution
(still thinking about doing this for our big mill - (jmk would kill me ;))
cradek: the encoders are connected to the main axle of the motor, right? how many inc/rev are they? 200 or so?
skunkworks: heh, that would be wild
Lerneaen_Hydra: the encoders are on the motor shaft and are 500 line
are you using the divide by - for the encoders? the uP?
skunkworks: you'd need a lot of parallel ports to run the tool changer :-)
cradek: oh, sounds nice
yes an atmega8 for dividing the encoder
it will also divide the spindle encoder which is 1024 line
seems like I have been away forever :)
any videos of it yet?
I figure I can do it with 3 parrallel ports
not yet, I'll sure take one when it threads
skunkworks: these are the servos I use on the etch-a-sketch. http://www.goldmine-elec-products.com/prodinfo.asp?number=G15607&variation=&aitem=4&mitem=5
you could get a lot of IO for $100 worth of pci parport cards
do they have parport cards with 2 ports?
jepler had a link to a cheap i/o card a while back - like 70 dollars or was it 70 i/o
something like that :)
skunkworks: they're similar in size to the ones on cradek's lathe, but with an integrated 6:1 gearbox (cradek has 3:1 pulleys instead)
jepler - how much backlash in the gearbox do you think.
skunkworks: well there's the $65 PCI8255 with 72 I/Os: http://www.futurlec.com/PCI8255.shtml
that is the one I was thinking of
skunkworks: no idea about how much backlash is in the gearboxes. A heck of a lot less than is in the etch-a-sketch internals
skunkworks: and for only another $65 I'll write you an emc2 driver for the PCI8255
that's pretty cheap for what you get.
cheap - I would pay double
you mean, $65 for the card, or $65 for the driver?
I wonder how it's interfaced to under linux
this is the link cradek had sent me a while back http://www.surpluscenter.com/item.asp?UID=2006071311472077&item=10-1834&catname=
skunkworks: well just send the $65 in the form of one of these PCI cards to: <my postal address>
bill203: It's a PCI device with a block of I/O ports .. the 8255's I/O registers appear at base_address + 0xc0, + 0xc4, ..., + 0xec
yah - that went right over my head the first time. time to eat I think
bill203: so I suspect it's quite simple: one outb() to set up the direction of each port, and then a few outb()s and inb()s to read/write each port.
more information here, if you can read thai or pick out the english words and numbers and make sense out of it: http://www.etteam.com/download/10PC_INTERFACE/1008/Man_ET_PCI8255_V3.pdf
alex_joni_away is now known as alex_joni
whoa. I just came back and said hi the same moment you did, after I had been done for an hour or so
ok, was afraid they're cooking you
alex_joni: did I see that dapper is officially supported now?
cradek: uh, random EUQ, how difficult is it to set up EMC with G33 and optical sensors? (not encoders, the fork-like sensors used in mechanical computer mice for instance, with an index and stream)
alex_joni: oh, is there a new kernel for dapper, or is it the same one I've been using for a while?
same old ;)
I think I got mine a bit after the semi-pubic beta release
ooh, sweet :)
2.6.15-magma aj4 iirc
better typos are hard to find ;)
wonder how a pubic kernel looks like ROFL
*does not want to think of that*
this is even bash.org worthy ;)
haha, almost is actually ;)
reminds me of my other bash quote
or well, the one I submitted
[17:19:01] <Lerneaen_Hydra> http://www.qdb.us/62090
* Lerneaen_Hydra hint hint, vote it up ;)
already did that
[17:22:36] <Lerneaen_Hydra> http://www.kotaku.com/gaming/retro/watch-a-chimp-play-ms-pacman-186031.php
oh my, that was actually quite surprizing
wanna read my favorite quote?
you can see by the actions that he can tell the difference between "run away" and "eat them all" he also can control the game quite well
[17:24:05] <alex_joni> http://126.96.36.199/irc/irc.freenode.net:6667/emc/2006-05-07.txt
20:20:01 <SWPadnos> ni Ray. maybe modem # 45 will work better than #41
20:20:21 <SWPadnos> s/ni/hi/
20:20:31 <SWPadnos> I am not a knight who says "ni"
alex_joni: haha, and then a minute later, you all revert to discussing sane things... :D
Lerneaen_Hydra: did you see the movie?
no, it's Monty Python
Monty Python and the Holy Grail (1975)
oh, yeah, of course! you do you take me for?!?!?! :p
[17:31:49] <Lerneaen_Hydra> http://video.google.com/videoplay?docid=-3934788900154749704
And that is why the earth is shaped as a bannana
* Lerneaen_Hydra prods cradek
[17:44:35] <alex_joni> http://bash.org/?664845
LOL : < alx> wow irc feels like television right now
< alx> 75 channels and not a thing worth watching
alex_joni: I thought you were on vacation
jepler: in the mountains, in bed.. wasting some time ;)
alex_joni: I didn't know you had internet there
I'll still be in .ro till sunday, so I do have some net available
alex_joni: how's the weather?
jepler: outside you mean? :D
right now 17C
* Lerneaen_Hydra pokes cradek again, just for fun
which is way nicer thna 30+ we have at home
37 here today
cradek: I was poking you about the question I asked above: uh, random EUQ, how difficult is it to set up EMC with G33 and optical sensors? (not encoders, the fork-like sensors used in mechanical computer mice for instance, with an index and stream)
It's a tad cold here though
it was nice and 25-30-ish
but its gone down to 18 or so
oh I didn't see it
that's what I made the counter hal module for
alex_joni: one of the puns in that video that I was able to understand. http://video.google.com/videoplay?docid=-3934788900154749704
cradek: thanks for fiing the teleop problem
guess I never cared about the right speed for jogging with my tripod
cradek: is it simple to set up?
Lerneaen_Hydra: doesnt sound very complicated
it's just a matter of getting the right hal connections
alex_joni: sure, no problem
so now I just need to do hardware stuff
alex_joni: it errored right away with stepgen limits
cradek: then I must have used a borked config ;)
I didn't see anything
03cradek 07HEAD * 10emc2/src/Makefile: sample kinematics for Volker's system with the table rotated 45 degrees.
03cradek 07HEAD * 10emc2/src/emc/kinematics/rotatekins.c: sample kinematics for Volker's system with the table rotated 45 degrees.
alex_joni: if you had traj period and servo period equal, you would have not seen the bug
oh, then that might be the answer
somebody should re-do this page with emc2/halscope: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?PWM_Servo_Amplifiers
hmm.. makes me REALLY wonder
that's more than one year old, and I have never seen it before
I haven't either
oh, it's only linked from emc1 stuff
look out for CIA-8
03jepler 07HEAD * 10emc2/src/emc/motion/motion.c: if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
03jepler 07HEAD * 10emc2/src/hal/classicladder/module_rt.c: if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
03jepler 07HEAD * 10emc2/src/hal/hal_lib.c: if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
he'll flood any second now
03jepler 07HEAD * 10emc2/src/rtapi/ (rtai_rtapi.c rtapi.h rtl_rtapi.c): if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
* cradek ducks
03jepler 07HEAD * 10emc2/src/hal/components/ (16 files): if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
03jepler 07HEAD * 10emc2/src/hal/drivers/ (11 files): if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
03jepler 07HEAD * 10emc2/src/hal/utils/scope_rt.c: if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
03jepler 07HEAD * 10emc2/src/emc/kinematics/ (genhexkins.c rotatekins.c tripodkins.c trivkins.c): if MODULE_LICENSE is not provided by the kernel headers, provide it ourselves. get rid of #ifdef tests elsewhere for it.
* jepler waits for the compile failure...
03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot2_log.txt
there it is
* jepler looks at the error and scratches his head.
objects/rthal/classicladder/calc_rt.o: In function `StateOnLeft':
/home/John/farm/emc2head/src/hal/classicladder/calc_rt.c:37: multiple definition of `__module_kernel_version'
objects/rthal/classicladder/module_rt.o:/home/John/farm/emc2head/src/hal/classicladder/module_rt.c:181: first defined here
apparently the mere act of include <linux/module.h> creates some symbols in an object file
.. on that kernel
but is MODULE defined for that file?
I assume it is
* alex_joni wonders where from
EXTRA_CFLAGS = $(RTFLAGS) -D__MODULE__ -I$(BASEPWD) -I$(BASEPWD)/libnml/linklist
in our Makefile it's defined for everything in the kernel
03jepler 07HEAD * 10emc2/src/hal/hal_lib.c:
provide MODULE_LICENSE based on kernel version, not whether it's defined already.
reorganize to include rtapi.h before using MODULE_LICENSE
03jepler 07HEAD * 10emc2/src/rtapi/rtapi.h:
provide MODULE_LICENSE based on kernel version, not whether it's defined already.
reorganize to include rtapi.h before using MODULE_LICENSE
03cradek 07HEAD * 10emc2/lib/.cvsignore: I guess this is needed
03compile-farm 07BDI-2.18 (2.2.18-rtl3.0) * 10emc2head/: build PASSED
03jepler 07HEAD * 10emc2/src/rtapi/rtapi.h: remove #warning
libnml/os_intf/_sem.c 72: semctl(35422212,0,0) failed: (errno = 1) Operation not permitted
I get several of thse when I exit emc now
when did it start?
today I think?
well, I hadn't updated for a while
on current head
I don't, when running sim/axis
I wonder if it's anything to do with us both trying to use the same machine?
I just tried again and didn't get it
ok ignore that
I ran a few times and haven't seen it yet
cradek: so is axis' handling of tool offsets done, or should it do more?
I don't know
sometimes the preview and plot don't line up - I think the interpreters can get out of sync regarding g43/g49
with a diagonal line after the tool is changed, or with a mismatch at the beginning of the program?
I would sometimes get an offset for the whole run
I didn't study it
m2 doesn't remove the tool but an interpreter reset would...
maybe that's it
other than that, it works, when you load a tool the cone position jumps like it should (to the new tooltip)
also I think "reload tool table" isn't quite right - both interps need to do it and only the emc interp does (I think)
I expected there to be a "diagonal line" bug, where the preview plot would have a diagonal line but the backplot wouldn't.
I don't follow
with no tool loaded, program g0z1x0y0 / t1 m6 / g0 x1 / m2
the "g0 x1" move will also move in "z" an amount related to the offset for tool 1
I see what you mean
so the preview has the diagonal?