could someone help me tie M64 to hal?
okay. could someone help me convert my electric hot water heater to fusion power?
yeah, that's easy
you need a power transformer which increases the voltage about 20x
*puts in an order to digi-key*
owhite: so.. M64
so, suppose I was working with univstep_io.hal
I think I want to control something like ppmc.0.dout.03.out
you have 4 IO's available
its flood on, now, but I want to take over that bit and drive it by M64.
motion.digital-out-00 , motion.digital-out-01, motion.digital-out-02, motion.digital-out-03
those are the 4 pins that come out of motion and respond to M64 P0 P1 P2 P3
how about that?
close enough :)
but not quite
linksp = link signal to pin
you have 2 pins in that command
so you either A). use linkpp or B). define a signal first
A): linkpp motion.digital-out-00 => ppmc.0.dout.03.out
B): newsig oxygen bit
linksp oxygen <= motion.digital-out-00
B) looks more human readable.
linksp oxygen => ppmc.0.dout.03.out
your choice :)
great. let me do some poking. (BASIC joke)
I never got peek and poke
beat the hell out of learning addresses in C. :-)
owhite: is it working yet?
* alex_joni cusses beryl
I am WRITING! :-)
*tap tap tap*
actually I am trying to figure out which SSR I want to control on the USC board.
sometimes I wonder if I change hal if I'll end up with the 400 watt motors driving my table off the end of the ballscrew *starts up machine*
time to find out!
is M62 tied to motion.digital-out-00?
M64 P0 to turn it on, M65 P0 to turn it off
*works in MDI*
what? M62 ?
or M64 ?
I think M62 and M64 do similar things to the same output :)
doing some halcmd stuff.
M62 directly, and M64 synched with motion
oh. what's "synched with motion" mean?
M62 might happen at some time before the move starts
M64 should happen just before the next move
although I implemented the stuff in emc2, I can't remember any details.. let me look
crap. I have to take off for a little while.
he's not *that* round
darn.. needed someone to help me sort some deb foo
ah. well I know I'm of no use then :)
I installed beryl.. or tried to
now I ended up with a lot of upgraded packages which break dependecies :((
[00:57:00] <alex_joni> http://ubuntu.beryl-project.org/dists/dapper/main/
need to downgrade manually
was wondering if there isn't some apt foo to reinstall the latest version from the current repos
justin_ is now known as Twingy
alex_joni: you trying glitz or cairo, maybe svg stuff?
on a slightly related note, there was a program that was supposed to scan the installed packages and create a CD repository with those programs (or something like that)
there's probably a way with some shell script foo and apt-cache foo to get a list of installed packages that aren't the latest repository versions
is beryl not working for you?
The following packages will be removed: ubuntu-desktop x-window-system-core
no, not really..
bummer. too slow or other issues?
AIGLX starts ok
when I load beryl-manager it crashes X
so I fscked it for now
do you want some of the package revisions for X and stuff?
I can get those on my box if you like
right now I'm just trying to get emc2 to compile
I installed the magma kernel in the mean time
removed the nvidia binary driver
and I'm about 90% back restoring stuff :)
ah - ok.
[01:03:22] <alex_joni> http://wiki.beryl-project.org/index.php/Install/Ubuntu
I only afterwards read this:
Beryl is not supported on Ubuntu Dapper. Please consider upgrading to a newer version of Ubuntu instead.
I can try it on my edgy install then :)
I won't try it again, any time soon
I think I've installed xgl / compiz on edgy, but haven't done whatever I need to do to be able to use it
seems feisty has AIGLX in the xorg server
by default I mean
yes XGL might have been a better choice
maybe - I couldn't tell you at this point
Feisty's X.org 7.2 includes AIGLX, and enabling it is easy.
* alex_joni waits for grumpy groundhog
you just made that up
[01:07:05] <alex_joni> http://en.wikipedia.org/wiki/Ubuntu_(Linux_distribution)
"There are plans for a Ubuntu branch codenamed "Grumpy Groundhog", which has not yet been made available to the public. It is planned to be a permanently unstable development and testing branch, pulling the source directly out of the revision control of the various programs and applications that are shipped as part of Ubuntu"
ooooohhh - Ubuntu HEAD :)
we should enable RTAI packages from CVS magma
and emc2 from HEAD
now that would be a system
a s-s-s-system :)
now it works.. 5 minutes later .. kernel panic oops
I mean the panic oopses :D
oops - now I'm panicking
this reminds me..
I had this once:
"kernel BUG at panic.c:1521"
I still have a picture of the screen on my cell
"kernel BUG in header file at line 248"
that's a good one for a blog / archive somewhere
maybe one day
ok, seems I got all things ./configure needed
These are the top two error messages I have experienced: "Error handling Error" / "Nobody should ever get this error message, if you get this error you are seriously f*cked"
my favorite is "you don't exist. go away"
[01:15:28] <alex_joni> http://www.plethora.net/~seebs/ops/ibm/cranky37.html
"This is not a viral infection. It comes from programs like write, talk, and wall, if your invoking UID doesn't correspond to a valid user (probably due to /etc/passwd being corrupted), or if the session (pseudoterminal, specifically) you're using isn't properly registered in the utmp file (probably because you invoked it in a funny way)."
even ssh says the same thing
rofl, that guy has it rough
heh - I Like the idea of giving the store customer service number when they ask you for your phone number
yay, HEAD is running
I really should figure out how to get RTAI/SMP/A64 running on my machine :)
I'm slowly getting to the point where I was before my HDD failed :)
hope that's IA64 not A64
I can pick any two, I guess. right now I have SMP/A64 ;)
nope. x86_64 ...
how about C64 ?
nah - I'm an Atari man
ok.. nuff for today
hasta la bye-bye then :)
one minor thing I noticed :)
trying to fix it now
darn.. I know not how
[01:21:58] <alex_joni> http://linuxcnc.org/docs/2.1/html/install/installing_emc2/index.html
$ gpg -keyserver pgpkeys.mit.edu -recv-key BC92B87F
$ gpg -a -export BC92B87F | sudo apt-key add -
those should have been --keyserver, --recv-key, --export
in the wiki, or in your paste?
in the generated html
this is from the manual (lyx)
the lyx code is fine.. only the converted html sux
* alex_joni is checking the pdf now
pdf is fine.. so it's a problem with latex2html or whatever
not gonna try to debug that..
* skunkworks powers up the TI99/4a emulator to play parsec
not that weird
I liked parsec too
press redo or back
I meant the TI994a part =)
Don't make me break out the Coleco on your asses now!
was parsec a main-stream game? I only remember it on the TI
no pun intended
Brakout Atari 2600
it was just TI
never got into hunt the wampas
I never had any games for the 99/4A
unless you count the speech cartridge
TI basic was your friend
I have a friend that likes to call himself "Hawkeye" (like in MASH)
I had to work to get the thing to pronounce "Hawkeye is a twit" correctly
it kept saying "Hockey is a tuit"
Until I was in highschool - I thougth it said 'its basic' - must have been dyslexic
I used say+am
I think I ended up with "hockaye is a tewitt" or some such
you work with what you got ;)
plus some of the tone modifiers
it had a very limited vocabulary
the vocabulary was unlimited, but getting the pronunciation correct with the chip they used was difficult
always wanted the extended basic cartridge
oh it was text to speech?
it could do TTS, but you could also program it in phonemes or possible allophones
my first experience with speech synthesis was a table of phonemes
(that was an early speech chip, not the one in the TI, I think)
holy crap.. http://cgi.ebay.com/TANDY-POCKET-SCIENTIFIC-COMPUTER-PC-6-RADIO-SHACK_W0QQitemZ200067854578QQihZ010QQcategoryZ11713QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
I still have mine. still works
must be a trip down memory lane
it had a whole 8k of memory.. Still filled it up.
cool - http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=260071752551
I used to use one of those
heh - I was just looking at those ;)
we madea control system for something with that
hmmm - a photo processing machine, I think (can't remember now)
so, I think I'll be getting a lathe at the end of the month
that should be fun
it will be one day
better be - or atleast converted in the next mont
actually, it is already controlled by an NC dontroller
controler, that is
and it's actually being used in an aerospace shop right now
it's a Hardinge HNC, with GE controller
ah, a big machine
only 3000 pounds
where is this going? I thought cradek had the barn ;)
I may need to build one ;)
SWPadnos: use round parts for it
unfortunately, the photos from the eBay listing are gone, and the listing doesn't say much
it does say this though: "Machines in working condition. Always used with oil and still can hold .0002 tolerance."
I went there with a friend (a mech. engineer / machinist), and the machines are just amazingly clean
the tool plate has not one scratch on it
(and I think it comes with a spare)
this is what it looks like: http://cgi.ebay.com/HARDINGE-HNC-CNC-PRODUCTION-LATHE-S-Up-to-seven-avbl_W0QQitemZ200066336883QQihZ010QQcategoryZ97230QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
this one is at least as clean as the one(s) in that auction, and it's less than 10 miles from me :)
and it's $500
I was happy that they still have one (they started with four)
what's the uSTOR on the bottom?
so it looks like an 11" swing
ah - the CNC editing thingie
I'd say it's a lot of buttons to use on a retrofit ;)
I don't think the one I'm getting comes with that though
I don't see any tape - how is the program stored?
no position readout either
it looks like the uSTOR has a position readout or something - there's a red 0 under the white painted number
yeah it looks like it's 5x7 LED so it's probably alphanumeric
luckily, I have a jogwheel for just such an occasion
oh now this is funny: http://www.astratool.com/tooling/attlcat.htm
does yours have that tool changer?
they're selling a used tooling plate for $500
yes, plus an extra plate
also a powered collet puller
I don't think there's a barfeed though (not that my garage is big enough for one anyway)
and I can probably get them to hand over an air chuck also, if I want (they have a lot of stuff that they offered to sell with it if I want it)
"this machine can had automatic legs conveyour. the legs obtain difficulty when using the hexagon bar, when using the automatic conveyour due to the noisy sound and cause the speed couldn't increase. using the round could cut fast any angle but avoid the hexagon legs deficient."
sounds like the polygon turning video ;)
it's from there..
it describes why you don't benefit from the barfeed for haxagon bars :D
the barfeed is for to hexagon by rounding with the actuator
(I made that up)
cradek: if you read between the lines
good night guys
getting late here
I always like to avoid those hexagon legs deficient
good morning Alex ;)
it's all good
good evening jmk
03jmkasunich 07v2_1_branch * 10emc2/src/hal/components/stepgen.c: backport 'unobtainable step rate' error message
it shuold say "I'm sorry, I'm afraid HAL can't do that" ;)
thank you thank you. I've been here all week
you have been waiting to say that I bet.
err - no, no. not at all
I strongly suspect that I should not have eaten that Snickers bar
[03:16:56] <cradek> http://video.google.com/videoplay?docid=2545377385698242701&q=robot
that looks fun :)
except for the concrete floor
that looks insane
someone really really trusts that robot and its programmers
and the concrete floor ...
I cringed when the guy was upside down and his head was about a foot from the concrete - a typo, and his head is rammed into the concrete
yep. that was my concern as well
plus the couple of spots where his head was swinging close to the main pedestal
this is supposed to compare the 2.1 version and the head version:
cvs diff -u -rv2_1_branch -rHEAD src/hal/utils/halcmd.c
ignore the \, typo
it reports no differences
yet there are differences
(I have both versions checked out, and I can diff them with normal (non-CVS) diff
about 340 lines of difference
is the "-u" option supported in cvs diff?
yeah, use it all the time, when I'm comparing my working copy with the repo
two -r options is supposed to ignore the working copy and compare the versions in cvs
I suspect thats not what its doing
try it in both dirs, to see if the results change
it looks like it worked when I did it in the head checkout
so it's only taking the first -r, it seems
the man page specifically say:
With two -r options those two
revisions will be compared (and your working file will not affect the
outcome in any way).
oh the pain
cat was draped around my neck when the new puppy ran in
bleeding stopped yet?
and it still has all of its claws
the puppy tried to climb up onto the bed (over my face) this morning and made a huge Z shaped scratch on my forehead
I had to tell everyone at work that I lost a fight to Zorro at the bar
If I was on that robot - I would want a full cage.
I would want a barf bag
I love roller-coasters.
coasters are fine
I don't mind up and down and sideways, but I don't like spinning
or, it don't like me
i looked up what 1g acceleration meant... and the wikipedia talked about roller coasters, there's one in Edmunton that uses pneumatics to acclerate a drop to over 5G ( i thought that sounded deadly )
[04:42:51] <tomp> http://en.wikipedia.org/wiki/Mindbender_%28Galaxyland%29
jmkasunich: did you ever figure out cvs diff? either it doesn't work right, or I don't know how to use it
I gave up
well now I'm getting a diff
the results seem to depend on what dir you are in, even tho the man page says it doesn't matter
I've noticed before that -rHEAD and -rBASE just don't work
I thought it was fixed recently
(well recently in the geological sense)
maybe I'll just use regular diff on two checkouts for that kind of thing
cvs diff with one -r (against working version) has always worked
`HEAD' refers to the most recent version available in the repository
when I read what this says, not what I want it to say, the behavior might be right
(stupid, but right)
I think to get the head we want we can use -r1
how it get to be midnight already?
cvs diff -rv2_1_branch -r1 2>&1 |less
yes I think this works
do you have to be in any particular checkout to make it work?
(2.1 or head)
"HEAD" refers to the latest revision on the current branch in theRepository. The current branch is either the main line of development, or a branch in development created by placing a branch tag on a set of files and checking out that branch.
you can probably tell by reversing the order of the -r options
so we've been misusing the term head all along
because every branch has a head
I think this means if you are in a checkout with sticky tags (like you get when you cvs up or co -rv2_1_branch) that means HEAD is on the branch
the RCS Main Branch is called "the trunk"
yes we're using the HEAD term wrong, and so does the rest of the world, I think
and a lot of documentation is VERY fuzzy, like the thing I pasted above ("the most recent version available in the repostory") which is just plain wrong
hmm - when anders changed siggen params to pins (so he could control it with pyvcp) he commented out the param code
I'm tempted to remove it completely, then backport
only if the halcmd setp extension is also backported ...
I believe its already in 2.1
but I'll check
ah - port not too far back :)
if I have the diff right, the difference between the two HEADS is just `net'
yeah, its there
"just" net = 300+ lines
(I've been somewhat carefully inspecting the code I backport, hard to do that with a big diff
I'd be in favor of backporting "net"
ok, I'll bite the bullet and do it
I think net is a huge usability improvement
if we have net in 2.1, maybe we can deprecate linkpp in 2.2
makes sense to me
I hate linkpp, because it gives a bad name to the signal it creates
that would be nice
I guess that features that don't change the behavior of the interpreter or the TP (or NML ...) should be backported in most cases (even though there's a feature freeze)
I guess that means GUI improvements and HAL stuff
I'm not sure I think that's the criteria
I think it should be more lax for the "peripheral" areas, I guess
however I can't say exactly what it is either
hmm, a lot of unused code in siggen
some (new) things we're sure won't break anything (else)
like new drivers
or new commands in halcmd, for the most part
we must have solved the math library problem at least for the distros we care about
yeah I haven't heard of a problem lately
because my homebrew sin() inside siggen has been #if zero'ed out for a long time
I think I'm gonna remove that code
now is not the time to cleanup code on the branch
not even something that is inside an #if zero block?
umm I dunno
and has been for a long time - it has one of paul's auto-gripes on it
not long ago someone "cleaned up" an //#if 0 block and broke stuff
I'm all for auto-gripe removal ;)
well over a year ago he sent a script thru the code adding /*! \todo Another #if 0 */ to every if 0
03jmkasunich 07HEAD * 10emc2/src/hal/components/siggen.c: removed unused code, some left over from param to pin conversion, some ancient
I'll backport _after_ the farm compiles it on all systems ;-)
fading fast - I'll do the siggen backport, and the "net" inspection/testing and backport, tomorrow
err - hmm. maybe that's my cue also
03cradek 07v2_1_branch * 10emc2/share/axis/tcl/axis.tcl: backport AXIS home-all and pyvcp support
03cradek 07v2_1_branch * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: backport AXIS home-all and pyvcp support
03cradek 07v2_1_branch * 10emc2/src/hal/drivers/serport.comp: backport cut/paste error fix
man... I shaved today, at 4:30pm... and I have growth again, at 2am... I hate shaving.
so glad I have whiskers... and can get away with not shaving for a little while.
[07:38:51] <K`zan> http://cgi.ebay.com/ebaymotors/1912-Antique-Vintage-1-4th-Scale-Working-Collector-Car_W0QQitemZ270077056217QQihZ017QQcategoryZ6235QQrdZ1QQcmdZViewItem
Check it out, wow!
this is "fun" http://video.google.com/videoplay?docid=-7278899013907810847&q=robocoaster
a friend of mine is the son of the CEO of kuka sweden
what's with always adding corny music to videos like that
lol.. no idea
03alex_joni 07v2_1_branch * 10emc2/src/hal/drivers/ (pluto_servo.comp pluto_servo_rbf.h rbf2h.py Submakefile): backport pluto driver
03alex_joni 07v2_1_branch * 10emc2/src/hal/drivers/pluto_servo_firmware/ (8 files): backport pluto driver
03alex_joni 07v2_1_branch * 10emc2/configs/lathe-pluto/ (emc.tbl emc.var lathe-pluto.hal lathe-pluto.ini): backport lathe-pluto config
03alex_joni 07v2_1_branch * 10emc2/src/Makefile: backport lathe-pluto config
03alex_joni 07v2_1_branch * 10emc2/configs/sim/.cvsignore: ignore generated file
03alex_joni 07v2_1_branch * 10emc2/docs/man/.cvsignore: ignore generated file
03alex_joni 07v2_1_branch * 10emc2/docs/man/man9/.cvsignore: ignore generated file
03alex_joni 07v2_1_branch * 10emc2/src/hal/utils/.cvsignore: ignore generated file
A-L-P-H-A: so, just dont shave
03alex_joni 07v2_1_branch * 10emc2/configs/motenc/motenc_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/ppmc/ppmc_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/stg/stg_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/univpwm/univpwm_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/univstep/univstep_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/vti/vti_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/m5i20/m5i20_io.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
03alex_joni 07v2_1_branch * 10emc2/configs/dallur-thc/dallur-halvcp.hal: remove references to motion.spindle-incr-speed and decr-speed, these are deprecated
you guys have been busy!
* jepler looks back at the commits
jepler: when you code pluto-p driver modules, do you do that in something windowsbased?
ack. running in some vm-ware like virtualised environment?
lerneaen_hydra: I use the free version of quartus II in a windows xp running in vmware on my ubuntu laptop
lerneaen_hydra: I had the XP license and the fpga software is no cost, so it's only inconvenient
(there's a no-cost version of vmware too)
yeah, it's kind of sucky to be limited to windows though
they sell a version that runs on linux
but I think it's thousands of dollars -- I haven't actually checked
O.o , so they give away the windows version and sell the same program for linux for lots of money?
shouldn't that be the other way around? ;)
03jepler 07v2_1_branch * 10emc2/debian/control.in: we bundle yapps, remove it from requirements list
they sell a version for both linux and windows
they give away a free one for windows only
as far as I can tell
03jepler 07v2_1_branch * 10emc2/share/axis/tcl/axis.tcl: feature backport from HEAD: the "debug level" dialog
03jepler 07v2_1_branch * 10emc2/src/emc/usr_intf/axis/scripts/debuglevel.py: feature backport from HEAD: the "debug level" dialog
jepler: oh, a free crippleware version?
lerneaen_hydra: it seems to be adequate, but advanced features are disabled...
what's advanced? stuff that isn't needed for EMC stuff?
what I've noticed so far is some stuff to do with debugging and with estimation of device power consumption
oh, nothing very important then
I think maybe not all devices are supported either -- they have many products, all programmable from their devel environment
well, debugging maybe
oh, I see
I am sure their website gives a list of the differences if you are really interested in the details
03jepler 07v2_1_branch * 10emc2/docs/man/man3/rtapi_print.3rtapi: allow error messages from realtime
03jepler 07v2_1_branch * 10emc2/src/emc/motion/motion.c: allow error messages from realtime
03jepler 07v2_1_branch * 10emc2/src/rtapi/ (rtai_rtapi.c rtapi.h rtl_rtapi.c sim_common.h): allow error messages from realtime
breakfast time! bbl
hey - is this the mini-ITX board you have? http://www.mini-box.com/Hybrid-C7-1G?sc=8&category=99
You get that presentation and meeting successfully behind you.
behind me at least :)
this is Athalon64
I can switch if you want some testing.
no - just wondering if it's the same board
I'll be getting a (very clean anf perfectly functional) HNC + GE control in a month or so, and might want to put one of these boards in it when I retrofit
This is the via board.
hi rahy-via :)
rayh, that was
ray1 is now known as rayh-via
There we go
you were saying that the max latency was ~12000?
or I've got gaim!
you got gaim!
Right. That was a semperon
(you can't put in the 've - that would be correct phraseology)
ok, I thought it was the ITX board
Let me run it here and see what comes back.
RTD| -1399| -1399| -125| 7519| 7519| 0
RTD| -1398| -1399| -156| 6896| 7519| 0
is that with you messing with windows, launching programs, and running glxgears? :)
Looks like latency is better on the via 1gig than the semperon
that's not a huge surprise
Na just running about four windows.
full of stuff including gaim.
"full of stuff" and "doing stuff" are diferent
This does have the mesa card installed.
yep - I thought I remembered that (though one never knows)
but I think the wiki cautions that you can't run both emc and the latency test.
no, probably not
the idea is to stress the graphic subsystem while the latency test is running
that was fun
you can't run both at once. tried it.
glxgears on top of the other stuff.
then the latency test
right. and drag things around too
rayh-mini is now known as rayh-itx
ah latency test is still running after the logout and login
hi all, what's up?
I was playing with latency tests.
on an itx board? what processor does it have? passively cooled?
Small fan 1gig Ester
model name : VIA Esther processor 1000MHz
is it any good for RT?
Not to bad. The numbers turn out about the same as a much faster semperon here.
I wasn't able to get base_period much below 15000 without the display slowing down.
I'm off to make some front panels for 19" rack enclosures... will house the new servo amps etc. for controlling the mill... bbl
remember the cyrix proccessor
it would have been a 486 class iirc
I think I still have one in the shop
we had a cyrix math co-proccessor in our 386-20
[17:29:26] <SWPadnos> http://en.wikipedia.org/wiki/Cyrix
03jmkasunich 07v2_1_branch * 10emc2/src/hal/components/siggen.c: backport siggen param to pin change
03jepler 07HEAD * 10emc2/docs/man/man3/rtapi_print.3rtapi: markup fix
03jepler 07v2_1_branch * 10emc2/docs/man/man3/rtapi_print.3rtapi: merge rev 1.3: markup fix
03jepler 07v2_1_branch * 10emc2/src/hal/components/counter.c: merge rev 1.11: add velocity output
now I'm done with the enclosures for today: http://www.anderswallin.net/2007/01/enclosure-for-servo-electronics/
I need the transformer and the other psu parts to continue. Hopefully I'll have them by next weekend.
I wish there was a little more english information on the 130V 14A servo amp on this page: http://www.diy600.net/
though I can read the name of the chip it's based on, I suppose I'll have a look at that datasheet instead
it's the UHU controller, developed in germany I think, there should be plenty of info around... try searching on cnczone
what, "UHU-servo controller for spindle motor, 130V 14A." isn't enough for you? ;)
the logic stuff is done on a PIC I think
[19:45:41] <awallin> http://www.uhu-servo.de/
ha! it's an AVR
I think they sell kits, with the pre-programmed microcontroller included
but it's a step/dir controller, real emc:ers close the loop with emc!
I guess the thing to do would be to get one with the uC, and scope the output to the analog section
then make a 5i20 config to do that :)
probably not worth the trouble. I don't think the digital control side is that hard, it's the analog power electronics that's a "black-art"
jepler: so do you want a translation of what Sauli wrote on the DIY600 page
right - that's why you'd use the black art part, and skip the uC with the Mesa card
so you can use PWM instead of step/dir
and close the loop in EMC
rayh: you around?
SWP: yep. are you looking for a power stage right now?
got a NEC question ;-)
awallin: thanks for offering, but it's not necessary
"looking" may be too strong a word :)
now if skunkworks would finish he's H bridge
it's the power stage that interests me too, but not out of a real need
though I could probably use one that's a bit more beefy than the Mesa driver
and a bit less (for one application) than skunkworks thing
I should do a GPL'ed H bridge design one of these days
"one of these days"
I mean. sure. maybe Fest would be a good time to work on one :)
jmkasunich: would you make it scale:able?
scopes, people, motors, machines, MEsa cards. it's all good
or make it possible to drive big motors and stufF?
the problem is that scalable = too expensive for the low end
plus the boards are bigger due to trace/space widths, etc.
but it would be interesting to sit down and think about/write down all the tradeoffs and see what comes out of it
If I would find a big DC servo that would do about 6000rpm, maybe around 1kW, I would be pretty motivated to put together such a powerstage. I'd use the motor as a spindle
awallin: how's your budget?
did someone say EMC can close the loop? does that mean EMC can make use of a stepper with an encoder?
ds3: sort of
ds3: yes, but not many people are doing that
guess ray isn't around...
ds3: generally, servos with encoders
is there a write up on doing this? (got a good deal on a pair of steppers with US Digital encoders on there)
anonimasu: I could pay 3-400eur for such a servo, with an encoder
does anyone know if the US electrical code requires GFI on _all_ kitchen outlets, or just those within X feet of a sink?
(I'm installing one above the counter on the other side of the room)
jmkasunich: all I think
you can monitor the poisitioning of the machine with a stepper/encoder setup, but I'm not sure you could tune a PID to help the stepper "catch up"
except for fridges and other permanent appliances
I think they all have to be GFI protected, but they don't all have to be GFIs
SWPandnos: can it fault if it misses X number of steps?
that's counters though, any "normal" wall outlets don't have to be afaik
check them out..
ds3, yes, absolutely
they are not cheap.
ds3: sure, the following error would be adjustable
[19:56:49] <anonimasu> http://www.aratron.se/pub/1541/DC-motorer_SeriePENTA.pdf
you can possibly even do some slowing of the machine if a stepper gets behind, but that's a bit more complex
they just go to 3krpm..
but they are servos..
ds3: instead of hooking axis.0.motor-pos-fb to stepgen.0.position-fb, you would hook it to the position from your encoder
anonimasu: well, 400eur converted to your 1800-century currency would be maybe 3000-3500sek
ds3: once you've done that, a stalled stepper will quickly cause a following error
awallin: the 1.0kw continous was 700eur..
ds3, or you can possibly stick a PID on the following errors of the stepper motors, and send that input to the Adaptive Feed OVerride input, to slow down motion when the steppers are overloaded
anonimasu: ok, that was what I thought. so I probably need to find one second hand...
brushless dc motors?
how do you drive thoose?
beyond merely getting a following error when a stepper stalls, nobody seems to have a clear idea what to do with step+direction outputs and position feedback
like a stepper or a 3-phase AC motor, for the most part
SWPandnos: that is likely to do more damage; the problem I had before was trying to do a pattern of dots and loosing one of them really messed things up so stopping is best
anonimasu: hall signals usually tell you where the rotor is
I tried making an AC servodrive, but it's a bit complicated, so I sold my AC servos
I might be successful with a H-bridge powerstage though ;)
ds3, sure. monitoring and stopping is very easy with EMC. correcting for it is also possible, but as jepler said, it's unclear what the "correct" action is when a stepper starts lagging behind
awallin: I wish I undersood enough to build a AC servodrive..
heh - look familiar? http://www.oesincorp.com/servo_driver.htm
*nod* I can see correcting would be useful if you are doing actual cuts instead of pecking
ever seen a Geckodrive?
SWPadnos: I were just kidding
SWP: I wonder if it's sold by gecko, or if it's a copy
(like the G320 that's in one of the images in the page)
dunno - I'd hope it's an OEM
maybe I'll email Mariss about it
dosent the motor max current sound high?
or is that the new spec
that's not so high for brush DC motors
for a stepper drive, that would be high
ah.. it's the same for the 320 also..
I thought it was lower..
nope. the servo drives have always been 20A
I just wish they would build a brusless drive :
I think Mariss is working on it, but it's third in line (or thereabouts) for him to work on it
third in line?
I believe his list is (1) the unstallable stepper thing, (2) a CPLD version of the G320, and (3-ish) a BLDC / 3-phase AC drive
unless something else comes up first :)
is the grex already working with mach3 ?
mostly, but that's not his problem
that's on Art and Steve Hardy
my geckos is shipping
started to build the frame for the plamsa machine today
jepler: I see stuff about 2.1 on the wiki, is there a set timetable?
awallin: "when it's ready"
going to weld the supports tomorrow :)
awallin: the v2_1_branch was created some time ago and the plan had been to add no additional features there (only fix bugs)
awallin: but some of this recent stuff (hal_pluto, pyvcp) are so important that the board decided they should be backported to the 2.1 tree -- I think you've been involved in this discussion as well
awallin: since if we keep up the current pace, 2.2 will take another year
jepler: a bit...
I think pyvcp would need about a month to stabilize
otherwise it'll be more work to backport changes if an unfinished pyvcp is released with 2.1
I think vcp was marked as deprecated in the manual, slated for removal in the 2.2/2.3 timeframe
awallin: one goal I personally would set is to have emc 2.1.0 released by the time of "cnc workshop" (a yearly US meeting of cnc people) which I think is in June this year
so I think pyvcp should be there (if it isn't already)
Hm, I dropped a "I hope you will develop AC servodrives sometime in the future." in my mail to them :)
jepler: June sounds fine.
nice people to deal with
[20:15:50] <anonimasu> http://www.practicalmachinist.com/cgi-bin/ubbcgi/ultimatebb.cgi?ubb=get_topic;f=13;t=004497;p=0
awallin: if we do it like emc2.0.0, then before 2.1.0 there will be a series of "testing" releases, starting as soon as we're confident the branch is in good shape
EMC, say no more.
To get EMC running correctly on Linux you need an arse shaped like a penguin.
gotta love all very educated answers.
anonimasu: the problem might be that there are a _lot_ of different brushless motors out there. trapezoidal vs sinusoidal commutation. number of poles, commutation based on hall+encoder or only encoder etc etc
awallin: trap/sinusoidal dosent matter does it?
awallin: dosent the drives use either thing?
* anonimasu wonders how rutex is for brushless
anonimasu: fwiw that "advice" about the velocity of Z vs other axes needing to match *is* true for emc1, but not for emc2
anonimasu: trapezoidal commutation is what I think is called brushless DC
jepler: I cant remember that when I ran emc1
sinusoidal would be called brushless AC (permanent-magnet AC), synchronous AC
I get it :)
jepler: though I never ran that much emc1
anonimasu: If cradek were here he could fill you in on all the details
jepler: I kind of got scared when my table took off during cutting..
for some weird reason..
awallin: i think brushless is trapezoidal and synchronous is sinusoidal
fenn: yes, but sometimes synchronous AC motors are called simply brushless...
anonimasu: in emc2, the max accel and max vel for a move are computed taking into account the projection of the move into each axis, and that axis's maximums. in emc1, the calculation was much more simplistic, and I think that a 5:1 velocity difference was one of the things that regularly triggered behavior like stalled steppers
anonimasu: or left you unable to use the velocity and acceleration that one axis was capable of
jepler: ah that one I've seen..
I'll be back in maybe 30-40 min. bye.
* jepler wanders off too
* alex_joni just got home
anonimasu: no, was out
bought a new dvd writer :P
[20:38:46] <alex_joni> http://www.pinetree.net/humor/threewishes.html
wow - only $599!
jepler: is this "orgone" the same as our blog software?
they say others call it "aether" :P
"your money will be in your bank account, orgone" :)
[20:42:00] <SWPadnos> http://www.toolsforwellness.com/ww100.html
"In Willard Water when the catalyst micelle is added to the regular water it alters the molecular structure of the water. " ROFL
SWPadnos: I dare you sell one to your mother :D
I dare me to not bother :)
I wonder if her sracastic mind or her chemical mind would hit me first?
I wonder what would hurt more :D
Hello. I have a milling machine with servo drives that support torque, velocity, or voltage feedback inputs (but no simple step/dir). Can anybody tell me if I'd be able to use those drives with EMC?
most probably yes
I only say probably since they may need 300V control signals or something :)
owad: yes, but you need an interface card
I won't be able to use the parallel port, you mean?
they'll also require feedback that EMC can understand, like encoders (which may already be there)
correct, the parport won't work directly
it does have encoders (though I'm not certain how they're wired up)
what's controlling it now (or before)?
an Acramatic 2100
and a few custom interconnect boards from Light Machines
it's a Benchman 5000
[20:49:59] <alex_joni> http://www.sea.siemens.com/machine/product/mcA2100.html
heh - well, I've never heard of any of those, btu that doesn't say much
they're really expensive and everything's locked down
I just bought the setup used and the control has a half dozen problems, so I figured I'd probably be best of scrapping it.
I'm browsing EMC's 'supported hardware' page right now. Anybody have any specific recommendations?
it's very likely that the drives use a 0-10V or -10 to 10V control voltage, since that's an industry standard
owad: the motenc has been suggested as good
but a lot of people bought the 5i20 from mesanet lately
yep - 10v
Thanks alex. I'll take a look at those.
owad: i've seen them used for electrodes, looked nice, stiff enuf, what i saw suggested it'd be suitable to rerofit with emc2
was yours enclosed?
automatic tool changer, too
rack tool changer as i remember?
meaning machine does most of the motion
the tool changer is mounted on the far end of the table
there may be caveats about using emc with rack tool changers. someone please comment before you spend money ;)
Here's a description I got of how the tool changer works:
"You just program the table to a XY location and lower the Z, trigger the air collet drawbar and pickup or drop a tool."
then you return to the program youinterrupted ( thats the catch)
right now, the toolchange code in EMC2 doesn't support "arrays" of tools, I think
but it's certainly compatible with G0 / M64Px / M65Px to move to the right spot and activate / deactivate the drawbar
Would I need an array…?
jepler: earlier you said "if 2.1.0 was released in time for the CNC workshop you'd be happy"
even a row is an array, it's just one-dimension :)
array (a line of tools)
I'd be very dissapponted
I'm aiming for less than a month from now
yeah - I was thinking at least a couple of months ahead of CNC workshop, so we can have some stuff to work on at Fest :)
What's CNC workshop?
[21:02:09] <SWPadnos> http://www.cnc-workshop.com
no, it's work (shhh - don't tell the IRS it's fun :) )
Anybody going to Cabin Fever next weekend?
I think rayh may be
It's about a 15-minute drive for me, so I'll definitely be heading over.
only 8 or 10 hours from here :)
nice html there ;) ... very patriotic?
heh - yeah
there actually used to be some information on the site :/
How big is CNC-Workshop? Is it worth the drive out?
lots of CNC stuff, EMC and other
The web page barely even describes what it is
many of the EMC developers will be there
well, I drove 22 hours each way last year, and I'm planning going again :)
yeah, I really don't like the new website
What goes on?
as a script kiddie might say, it is teh sukc
Are there workshops/classes?
there are classes and demonstrations on a range of subjects
man, that website _really_ sucks
I just registered, and it still sucks
Likely to be anything on learning CAM?
it has only a miniscule fractiuon of the info that used to be there
someone needs to tell Roland that nobody wants to register just to look at the information that should be in plain sight
I think I have his phone number somewhere///
asp .... ewwwww
his number should be in the contact list from last year, if nothing else
here are some pics from two years ago at the workship
[21:09:00] <jmkasunich> http://www.cardinaleng.com/emc_mazak_project.htm
Is a servo amp different than a servo drive?
(cardinal engineering's own website isn't all fscked up)
Hi jmkasunich. You had a question for me?
wow - so I should bring the truck and trailer? ;-)
who are the 4 people in that middle picture?
rayh: I was wondering about the requirements for GFI on kitchen counter outlets, and I know you are a licensed electrician
yes, whether you want to buy a lot of junk or sell it ;)
somebody answered the question for me (the answer was yes)
(but may have been wrong :) )
A gfci is required on all kitchen countertop general purpose receptacles.
just bought one
GFCI protected, or a GFCI at each position?
taking a break from fishing romex thru the walls
I normally put a gfci in the first box and feed the other receptacles from it.
SWPadnos: irrelevant for me, its gonna be a GFCI
rayh, that's what I thougt
rayh: this is old work, an outlet on the opposite side of the room from the sink and main countertop
(above a short counter between stove and fridge
If you're being inspected you'll need two circuits for countertop.
it will be fed from a different circuit than the others
there are already two above the main counter, on two circuits
this will make three
sounds good to me.
this branch only feeds two other outlets, both intermittent duty
outlets are one of the (few) things the students really did right when designing/building this house
one in the basement ceiling, one on the front porch
Does the front porch have it's own ground fault receptacle?
Wait that's not outside is it?
yes (under roof and in an outdoor duty box with the flaps
sure. I was trying to remember your place. Remember the garage and basement better than the entry.
you bought a student built house?
it's been here for 20 years, so it's unlikely they made any catastrophic mistakes ;)
just lots of annoying ones
rayh: yeah, its in a box with flaps, but I pressed the test button and no click
those things don't like outdoors
flap actually, one big one because its a gfi
the feed to the front porch passes thru the box that has the basement ceiling outlet in it
I could change that to GFI, and daisychain the porch from it
That motenc is a bit pricey.
rayh: does code now disapprove of putting the actual GFI outdoors?
If you've got enough gfci's for the job there is no need to daisychain.
No that is fine.
But those monstor covers are a real nusance.
I use that outlet about twice a year
How's the weather there now.
electric leaf blower in the fall, and christmas lights in the winter
That is one to many times.
I'd have to do without the leaf blower.
weather here? (or alex?) rainy and 40ish
yuck. I'd rather have cold than near freezing.
don't have to shovel rain
"monster cover"? do you mean the in use ones?
this isn't that big
The cover that fits over a plug is a bad design.
just a flap that covers the outlet (but it covers the whole thing, not one hole for each receptacle
The little covers were nice.
kept bugs and things out.
since its well under the porch roof I'm not worried about the "in use" part
rayh: warm here
awallin: you still around?
alex_joni: hi, just came home.
rayh: is roland likely to be at the shop on a saturday afternoon?
awallin: hi, I think 2 weeks is more realistic, then a couple of months for 2.1.0
I think I'm gonna give him a call and tell him (nicely) that the cnc workshop site stinks
He is there most everyday between 8-5 central time
alex_joni: right. well if people are reasonably satisfied with the current state of pyvcp, then it can be included in 2.1
awallin: I think that would be enough for now
awallin: I'd like to include whatever you have (that works) in a few weeks as the 2.1.0 version of pyvcp
it's much more than I had on halui for 2.0.x
jmkasunich: , alex_joni, that sounds ok. anything I need to do, or could you handle the cvs magic for me?
awallin: there's no real cvs magic..
checkout 2.1, copy the files over, add to cvs, commit
the only thing is the submakefile / makefile needed foo..
roland says he's working on the website a little every night, and in two weeks or less it should be pretty much done
he should get joomla :P
awallin: if you'd like to continue working in HEAD and have one of us do the backport, that would work
just let us know when it is in a state that you are comfortable releasing, we can copy that version over to 2.1, then you can continue in head with newer stuff
jmkasunich: that sounds good, by the end of the month it should be in a reasonable state
that sounds like a lot time
end of the month? 2 weeks...
if everything else is ready at that point, we do the backport on feb 1, and release on feb 7 after some quick testing
jmkasunich: yeah, I was planing for a release by then
jmkasunich: re GFI on kitchen outlets. Kitchens are required to have at least two 20 amp circuits. So, if the GFI is in a box in the circuit, it must be a 20 amp GFI.
lerman: not in the box, individual GFIs at the outlets
(the box is an old federal pacific and its bulging at the seams - I'd love to replace it, but that is beyond what I'm willing to tackle, and plenty of $$ for a pro to do it
That's what I meant by box. (Sorry.) I haven't seen the outlet type in 20 amp versions.
anyone used lightscribe on linux?
When I moved into this house it had 60 amp service and 4 fuses (plus a range/dryer pair of fuses).
Converted to 200 amps myself.
over here you're not allowed
well, with only 4 circuits to start with, the job is a lot easier
you need to be certified to work on 230V
this one has to many to count
In Connecticut, you can do anything on your own house. No matter who does it, it has to be inspected by the building inspector.
its a 20 position box, and many of the positions have duplex breakers in them (two independent circuits on the same phase, not a 230v two pole breaker)
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_led.png: a screenshot of a led
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_number.png: a screenshot of Number
The quick way is to add a new box with a circuit feeding the old box. Then move the feed to the new box and you are done until you want to add more circuits (to the new box) or move some from the old to the new. That's what I did (even with only four circuits).
03awallin 07HEAD * 10emc2/docs/src/hal/vcp.lyx: document Number
what happens when you want to move a circuit from the old to the new, but the wires aren't long enough to reach
You need to add a junction box -- or do a lot of work. If you are moving all from the old to the new, you can use the old box as a junction box.
see - it quickly escalates to "more than I want to tackle" ;-)
That's only if you insist on moving circuits from the old to the new. You can just leave them in the old box.
I want to get rid of the old box because its such a mess
an outlet box is a mess? that seems rather hard to accomplish :)
main breaker box
that would definitely be "more than I'd want to tackle" :)
jepler: are you around?
another reason why I would NOT keep the old panel as a subpanel: http://www.inspect-ny.com/fpe/fpepanel.htm
What do you think of Pmw?
lerman: I've never used it. I think awallin tried for pyvcp, but ran into trouble.
It seems to have lots of good "mega widgets".
lerman: I tried it, but did not do very much with it, don't remember exactly why
awallin: something to do with problems subclassing?
Does someone have a good file open selector dialog widget I could use?
(Pmw doesn't seem to -- although I haven't looked real hard)
jepler: ah, yes, since I am a lazy programmer, pyvcp depends critically on all widgets being subclasses of tkinter widgets. vcpparse in its current form did not like a pmw widget
lerman: you don't like the built-in one?
>>> import tkFileDialog
lerman: jepler might disagree, but the most capable tookit for python could be wx (www.wxpython,org)
>>> f = tkFileDialog.askopenfile()
Where do I find decent documentation on all of this stuff?
03awallin 07HEAD * 10emc2/lib/python/pyvcp_widgets.py:
add FIXME texts all over the place
hint: if you are python literate and have free time, take a look!
there's not a lot of great documentation on tk
[22:01:36] <awallin> http://www.effbot.org/tkinterbook/tkinter-index.htm
Will EMC be able to trigger the air collet bar and respond to limit switches through a 5i20 board, in addition to to servo control and reading the encoder?
theres a lot of documentation on tk, it's just not great :)
owad: sure, but the 5i20 will not give out analog +/-10 V by itself, you need an addon card for that
ok, but through
the add-on is the 7i33
Thanks, jepber. I just tried tkFileDialog.
through EMC I can program it to do other random stuff I need besides just motor control, right?
owad: there is a ladder program included
the other I/O pins are available for use with the motion controller, classicladder (a built-in PLC), or external programs (M100-M199)
... or for new HAL components you write in Python, C or C++
Is there a breakout board for the 7i33? I can't find one.
SWPadnos: ever tried lightscribe?
owad: what would the breakout board do?
alex_joni, nope. are you talking about the CD/DVD marking thing?
yeah.. just got a drive that supports that
[22:09:25] <alex_joni> http://www.daqstuff.com/50_pin_breakout_500013.htm
the 7I33 has a 50-pin header. My motors and encoders are all separate.
alex: that's exactly what I was thinking of. Thanks.
03awallin 07HEAD * 10emc2/lib/python/pyvcp_widgets.py:
make fillcolor and bgcolor work for Bar
fix the FIXMEs ;)
has the options for the items on canvas (seems to be correct so far)
tomp: thnx, I'll look at that.
I think my spindle is controlled by a 0-10v signal, with 10v being its max speed of 10,000 RPM. I'm not sure how I'd interface that to the 5i20?
awallin: all options for tkinter dont seem valid for pyvcp tho.... on the pyvcp_widgets scale these dont work...
owad: pretty simple
the 5i20 has at least 4 servo outputs
tomp: no, the allowed xml elements are specified in pyvcp_widgets.py
tomp: but maybe that could be removed?
one of them is probably free... so you use that for speed setting
activebackground, repeatdelay, relief.... many more
and route the 0..10V analog to the control
oooh what a powerful stepper board this is: http://futurlec.com/Mini_SMCC.shtml
yes, they should be 'native'
tomp: OK, so that's one more FIXME
hmm… I was hoping I could get away without using that, in case I wanted to add a 4th axis
awallin: are you familiar with 'introspection' in object oriented languages? (no insult intended -- I haven't read you code)
owad: I think there is an 8 axis config for the 5i20
but I think that also reduces the number of general IO
Given a class, you should be able to determine what options IT knows about.
lerman: no, I am pretty much a beginner in oo (being an academic I have browsed through Gamma et al etc....)
lerman: yes, that's the sound way to do it, I just don't know how - yet
awallin: now coding an analog knob ( the jogwheel wont output a fraction even tho the container is a float ), and maybe same for scale widget
I'm not sure how one does that in python. But I believe I've seen enough to know that it is possible.
tomp: don't code new widgets, add a scaled output to jogwheel!
tomp: and the resolution for the scaled output could change when shift, alt, or control is held down
awallin: jogwheel is integer even tho the container is float, it's in many places forced to integer values (+=1, int() etc ), it cannot yield 0.5
the low level up & down insist onit
tomp: yes, that mimics a hardware encoder, it outputs an integer number of counts also. but the virtual jogwheel could have a float output
I would not like to change count
right, i'm making whats like a pot
I agree with awallin
I would add a pin, so there would be jogwheel.0.count and jogwheel.0.scaled-count (or whatever)
if you insist that the widget output a number that will never have a fractional part, call it "count" and make it an S32
if you are going to output a float, you should copy the setup used by knob2float, and the motion controller jogwheel inputs
with pins to control scale (and possibly enable)
the code has integer from the beginning ( floats with no fraction), i only "insist" that floats be allowed fractions :)
it's nice code, but wont tune gain
then you should have a "scale" pin that specifies the amount of change per "click" of the "knob"
I still think it's better to add to jogwheel than to code a new widget
which mean the 1 turn can be non integer (again not allowed)
jmk: that will be done by holding down shift/alt/ctrl, no new pins needed
tomp: just add a new pin to jogwheel, and make the up/down calculations for it float
awallin: I guess I should have been yelling at you, not tomp? ;-)
I thought that jogwheel did output a "real" float
no problem, consider it for my own use then,
do post your code somewhere, or commit if you have cvs access, I'm sure we can agree on something in the end
such a widget should have scale (output change per click), min value, and max value "parameters", whether they are specified in the xml, or made hal pins so they can be changed on the fly
of course, you got it
if you want the existing behavior, set scale to 1.0 and you have it
tomp, why must one full turn result in an integer? consider 100 ticks at 0.0001 each ...
i never got it to give < 1, example please?
tomp: the current pyvcp jogwheel will only give you integer counts
k, i'll send the code to you , i like whats been done but need some analog stuff, thanks
make a new float-out pin, multiply integer_counts by scale, and you can get <1 output ...
tomp: now that I think about it, there might be a logic for a separate pot widget. Jogwheel will turn and turn with no minimum or maximum limit. A pot would have a minimum value and a maximum value.
it's lower than the output stage
SWP: but that requires HAL code. we can easily add to pyvcp and make the hal coding unneccesary
lerman, maybe you can add something like this in your code i'd just see this in cnczone http://cnczone.com/modules.php?name=Links&file=viewarticle&id=428
I was talking about within the widget, not in HAL
Is EMC uncommon in supporting commercial motion cards? I notice Mach3 and TurboCNC are both parallel-port only.
bbl - gotta run and eat
SWP: ok, sorry
owad: parallell port is probably what 90% of hobby cnc'ers use
Would I have an easier time with programming/control if I replaced my Copleys with Geckos and used the parallel port?
owad: not a lot easier
yeah - I notice the price of the interface boards is approaching what I'd pay for three geckos
any advantages to _not_ using the parallel port?
depending on the resolution you want
you can probably get about 50kHz pulses from the PC
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_meter.png: screenshot of meter widget
at 40000 pulses/inch, that's close to 1PS
of course those numbers are all made up
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_bar.png: screenshot of bar widget
In the back of my Benchman, I have a large board labeled "Encoder / Tach Board". It has three connectors coming out of it, labeled X/Y/Z that go to my Copleys. It also has three smaller connectors, with 8 or 10 pins, that are also labeled X/Y/Z, and that I presume go to the encoders. I presume the Encoder / Tach board would go, right? Would those encoder lines get hooked directly to the 7i33?
jmkasunich, maybe this chip can be usefull in your future Gpl servo driver http://www.onsemi.com/pub/Collateral/MC33035-D.PDF
owad: also your system will be much nicer with a dedicated board..
owad: you dont need to dedicate all your cpu time to pushing out pulses
owad: I'd stick with the copleys :)
les runs thoose drives on he's big router..
Any idea what they cost to replace?
owad: what kind of motors do you have?
24V, 50 oz. in. (35 Ncm)
owad: i use copleys suited to brushed dc motors, they get power, analog command, and tacho, they output power to the brushes. meaning the other cables are encoder and copleys have no use for that info
ok, so the copley's would be getting their analog commands and tacho directly from the 7i33 board, right? And the encoders would also be connected directly to the 7i33?
no. the tacho would come from the motors..
and the encoders would go back to the 7i33 board..
and command would go tho the baord also..
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_spinbox.png: screenshot of spinbox
owad: got a pic of your whole machine?=
Looks just like this one: http://web.archive.org/web/19980212043518/http://www.lmcorp.com/bench5k.htm
ah tiny :)
and it's as described on that page, except with a 10,000 RPM spindle and an ATC
nice machine :)
Thanks. I was working on a CNC conversion to a mini mill when it came up for auction cheap. So I ditched the mini mill and went with this. :-)
owad: wow, me wants too!
owad: define cheap
One just went for $5,000 on eBay, though, which I thought seemed pretty good, too.
owad: how many hours has it done?
Is there a way to tell?
From the notes I have, I don't think it was run much.
some controls store that
and it hasn't been used for a couple years
actually, there is a log on the Acramatic
but the Acramatic isn't working
an error log, that is
maybe a usage log too
I'll be back in an hour or so. I _really_ appreciate all the help.
owad: no problem :)
Thanks, all. 'Later
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_scale.png: screenshot of scale
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_jogwheel.png: screenshot of jogwheel
can I insert comments, that won't show in the output, into lyx ?
awallin: not that I know of
but you can put a footnote or sidenote, that's less intrusive
03awallin 07HEAD * 10emc2/docs/src/hal/ (pyvcp_hbox.png pyvcp_vbox.png): hbox, vbox screenshots
03awallin 07HEAD * 10emc2/docs/src/hal/vcp.lyx: basic pyvcp description now in place
there. now the docs are in reasonable shape.
did jepler solve the latex2html problem ?
no, I haven't looked at it
lerman: still around?
jepler: how often are the pdfs and html generated? every hour?
couple of hours apart
awallin: something like that -- I just started a build so they'll be updated in <5 minutes
even if it's couple of hours, they (so far) always had the state I needed
but it terminated with an error:
make: *** No rule to make target `../docs/src/hal/pyvcp_label.png', needed by `depends/Master_HAL.d'. Stop.
jepler: wait, I'll update the image
awallin: alex_joni: you can make a paragraph have the "comment" style, which will not be shown in the output
220.127.116.11 in UserGuide (in my version of lyx anyway)
03awallin 07HEAD * 10emc2/docs/src/hal/pyvcp_label.png: screenshot of label
jepler: oh, nice :)
now it should build
I'll probably never use it though
alex_joni: lerman is back
lerman: I wanted to ask you something..
about the changes you did to present messages to users from comments
how extensive/notorious are the changes?
Not much. In one file: interp_convert.cc (I think).
awallin: a note about showing multiple lines of stuff in LyX-Code: use ctrl-enter at the end of each line. it reduces the extra space between lines.
awallin: you can tell whether you used ctrl-enter by a red symbol that looks like the one on the enter key
jepler: ok, thanks, I'll do that when I work on the docs the next time
now the HTML seems to be updated. thanks.
awallin: yes the second update finished
jepler: still the missing </end> tags... but it's not consistent, sometimes they are missing, sometimes they show just fine
awallin: the generated html is plainly broken
Yes. I believe so.But there were really two steps. First I had named params with trailing $. Then I fixed that to #<name>
[ <halpin>''my-button''</halpin> ] (default is button.n)
lerman: would it be ok if that would be part of 2.2.x ?
we are currently backporting some non-intrusive features to 2.1 before 2.1.0 is released
but mostly things which don't affect people that don't use them (like new drivers)
Well, it would be nice if it were thoroughly tested; but it shouldn't break anything. That would mean that named params would be part of 2.2.x also -- and that's fine with me. -- since I'm not using named params because I'm not using emc2 yet :-) it won't affect me either way.
lerman: glad you feel that way
I DO want to get named Owords done soon, also o<name>
someone mentioned sub-programs, is it possible to call sub.ngc from a main.ngc ?
My main thing is to NOT have to merge my changes with anyone elses; but since I'm the only one doing stuff on the interpreter (it appears), that shouldn't be a problem.
That would be in the category of "include files". I haven't really looked at that yet.
subprograms would be an thing for the interpreter
tomp: you still there?
Yes. We could add a feature that caused a file to be included into the current file as if the file were copied there. That would let us have libraries of subroutines that were just included where needed.
tomp: got anything working?
tcltk does, tkinter is the job at hand... expect at least a day for me to finish it
really python ( i got a pointedStick in it:)
alex_joni: I added "canonical drivers" to the 2.1 page in the wiki, but maybe that's not relevant?
[23:41:23] <alex_joni> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Emc2.1.0
awallin: just modified that
00:41 < tomp> really python ( i got a pointedStick in it:)
damn these video codecs, now acemi has posted some new videos, but I can't watch them...
* jepler curses at latex2html
awallin: If you are interested in python introspection, see: http://www-128.ibm.com/developerworks/library/l-pyint.html
this simple program has a stray literal < in its html conversion: http://emergent.unpy.net/files/sandbox/l2hbug.tex
awallin: yeh, the fab@home stuff same thing
* robin_sz has never really been certaint hat introspection was a good thig
rather liek reflectio in Java
lerman: thanks, I'll bookmark that for later reading
it *can* be useful
but there is a tendency for newbies to use it in crazy eays that break the basic tenets of OOP
awallin: you may want to try using Ctrl-" to insert an actual quote character in your python code examples, rather than just " which inserts a special (smart-quote-type) character instead.
awallin: that's "more correct" and seems to avoid the latex2html bug too
jepler: ok, thanks for that lyx tip
can anyone play the new videos by acemi, links at http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Videos
alex_joni: maybe the better videos from the wiki video page could be hosted on linuxcnc.org
i prepared it with ffmpeg, it's mpeg4
awallin: works for m
maybe you could remind me sometimes earlier in the day?
I have standard windows xp with all updates, and the the latest divx. what am I still missing?
awallin: funny, should be working
Need some help using an 8 wire stepper unipolar if someone has the time
unless divx is being a bitch and only playing files with the 4CC "divx"
Ziegler: there's a great doc at geckodrive.com
Ziegler: about the wiring of one :)
Doesnt talk much about unipolar
awallin: my advice would be to use ffdshow
let me read over it again
awallin: prefferably a new build
FOURCC code "FMP4"
[23:54:37] <lerneaen_hydra> http://www.free-codecs.com/download/FFDShow.htm
I'm not sure but I seem to recal the commercial divx decoder being bitchy with the fourCC code
ok, I'm downloading that now. so are acemi's videos any good?
the resolution/bitrate left a bit to be desired ;)
if you can live with larger files
ok, now it's working, with ffdshow. thanks.