looking at the help modules I don't see ttf...
I don't think there's a python ttf module installed by default on ubuntu. If there's an optional one, I'm not sure of the package name.
there is a tkFont module but it returns the family name not the file name AFAIK
whoo - we have flame
skunkworks_: like 1" flame or like 2 story tall flame?
so far - 1 inch
I should make a frozen pizza - just to see if the oven works.
don't smell gas - so that is good ;)
heh - no frozen pizza
[01:53:37] <jlmjvm> http://pastebin.ca/908014
jepler:what would the "net" be for these lines
also is setp still a valid syntax
my old configs were link syntax
setp is still correct, and has nothing to do with connecting signals
the net wyntax would be to add a name for each net, and then list all the pins (or some of them anyway) after the net name:
SWPLinux: how goes it at the shop?
net X-enc-phaseA encoder.0.phase-A parport.0.pin-09-in
skunkworks_: going fine - I came into the office because my feet hurt :)
it is hard to stand around all day
we've got the G&L X axis running at 90IPM back and forth
it goes up to 120 IPM, but we get a following error every once in a while
it's doing back and forth moves, 96" in each direction, and hitting indicators at the end
anyone know how to test to see if a igbt module is alive?
seems to be keeping to within a thou or better
SWPLinux: are you guys in Kansas?
oh boy working on Stuart's giant machine
yes, and it is a giant
can you guys take any pics?
the angle plates weight a ton or two each :)
SWPLinux: dual feedback?
I thought the angle plate I had was a monster
yep, got a few already, but haven't uploaded them yet
skunkworks_: yes, dual feedback
encoder for motor control, I term using the scale
Does it do what you guys where hoping?
did you guys set him up with his cadcam connection over the network too?
the tracking is very good, possibly excellent
there are some potential issues where you have to restart EMC to recover, but they're bad enough that restarting isn't the biggest problem
DanielFalck: no, the machine had networking already, we're not here to do that
SWPLinux: did you measure a step function to find the open loop acceleration time ? what kind of amps?
nope, old big allen-bradley velocity-mode 3-phase drives
It seemed like he was going to set up NCL to create gcode on the fly for emc over a network connection
net X-enc-phaseA encoder.0.phase-A <= parport.0.pin-09-in,shouldnt it have the <= also
do the amps really fight any position/speed changes? ( like belts are tighter than nuns buns ? )
the <= is never significant to halcmd, it's for user hints only
tomp: we don't really know ;)
you can feel that, but you gotta touch it
in fact, we don't really know the drive ratio - we got some weird encoder count ratio (24884 counts/inco or something)
the amps seem well tuned, and that may be a large part of the reason this is working well
if the low level open loop aint tight, the upper level position control loop will suffer
if EMC had to do the velocity loop for real, it might not work as well
oh, the motors stay still when we're not commanding motion
kick the table
the table weighs 10000 pounds or more
kicking isn't very good for the operator
kick the pulley
I don't know if any of us (or all of us together) can even overcome friction
? v & flat?
how is the way made ? v & flat, box way, linear guides with tractors?
uh - it's big
it's got hydrostatic bearings, and it looks like more or less box ways
they're covered my metal covers that are taken up on rollers as the table moves, so we can't really see them
i've used vertical hydrostatics ( Charmilles ) but not horizontal, so i dunno
this one is interesting - the spindle rides vertically, and is Y
there are some pictures of this thing online -- I think the URL is http://www.mpm1.com:8080/machines/GandL/pictures/
the table moves in X and W (X on top of W)
and the spindle is z (and is horizontal)
lookin, but net feed is sloooow
that machine is BIG....
yes. I was complaining that I need a wider lens to photograph it from inside the building
I'm glad that Stuart is converting it to EMC2
I think the spindle motor weighs more than an entire Bridgeport machine
it's in the same range anyway
i get the geomtery, i see how Z & W are like a BP knee/quill sideways
yep, sort of
X has 100 inches of travel, 47 on W, I think 100 on Y
and 18 or something insane on the quill
how fast does Stuart want it to rapid? I know you have it up to 90ipm now
90 IPM ;)
cheezmarie, yeh dont bother pokin at it ;)
he set up a nice pendant/control kiosk
yeah, it swings around nicely
I really like that
it probably needs some limits so you can't swing it into the work envelope
so.. Where are the youtube videos?
horz boring mill, worked on one in taiwan, had a cage for the operator to ride in on the quill
youtube, of course
one of the other machines here is a gantry-type with several bed extensions
when it's put together, it'll have 100 feet of travel
5-axis I think
not as tall or wide though (IIRC)
yeah it was under 6' wide iirc
Nice. Hmmm - a new emc-fest location? maybe a winter - then summer at rolands? ;)
winter would be better in Florida or something
skunkworks_: I'll let you be the one to ask stewart
yeah, maybe he'll bring some machines to Florida
i didnt believe it but yeah, hydrostatics are used for some really big masses
our mill has recirculating roller bearings for the ways. (kneedle bearings)
that would only be for the knee ;)
[02:24:48] <tomp> http://en.wikipedia.org/wiki/Gas_bearing
or gneedle for Ubuntu
It has been a long day. And my spelling isn't good on a good day
pro-flex gas pipe is pretty nice. easy to install. Just ran about 40 ft of gas line thru the house.
interesting. is that legal in your neck of the woods?
yes - I assume so.. that is what was used to hook our furnace up. ;)
I'm fairly sure that the gas company isn't allowed to connect to pipes that were installed by someone without a license or certificate or something
It is funny.. It says right on the tubing.. 'requires cirtificaton to be purchased'
or maybe the gas company just doesn't like losing the installation business ;)
and yet home depot/lowes carries it :)
it is easier to put togather than most coax ends.. :)
did that guy who was trying to install ever come back?
yes - got it at the local menards
heh - Menards. what a funny thing
very popular in Wisconsin, I think they don't exist elsewhere (except maybe Minn or Il)
yes - pretty local
Iowa and California too
they're in nebraska
I think my wife used to design their sale circulars, until she got sick of it
we have a home depot also in town though. Usually just a little more expensive
oh, I didn't realize they were that widespread
(Menards - I know HD is)
they have a nascar.. ;)
did a leak test.. shut the gas valve off - wait an hour.. turn burner on. runs for 5 seconds.. same as when it was just shut off
The vavle was verified not to leak.
SWPLinux: what is the interface between pc and amps/io ? I dont recognize the rack of 'eurocard sized cards'
Jon Elson's PPMC system
i thought i saw big dc motors, like sem's
cool, a stepper on that beast
no, they're big-ass Allen Bradley AC servos
which beast are you talking about?
PPMC, not USC :)
the 5inch dia 20 inch long motors
(not even UPC)
hmmm. I don't think there are any motors that narrow ;)
oh, not aware of Jon having non-stepper motor controls...
i just guessed from photo
oh sure - in fact the USC more or less turns a stepper into a velocity servo
ok, so servo is the norm
the USC takes a velocity command and generates steps for that speed
does the motor rotary encoder get back to emc or just the linear scale
hmm, elim blash
the encoder is used for a PID that has FF0, P, and D
the scale runs to a PID that has only I
they're summed, and that's sent to the servo drives
2 PID's for 1 joint/axis (cartesian), bleeding edge!
yes, and it's not without problems
if one encoder loses some counts, they diverge a bit and eventually you get a permanent following error - you need to restart EMC to clear it
though that's a hardware problem anyway ;)
but you hope for great position control with the linear
I think we tried using the scale only, and it sucked very badly
yes, it's a 1 micron scale
100" 1u = BIG$
there's a few thou of backlash or springiness or something, so that's a big problem
I'm sure, but they're original so no extra $
a few thous is some Nx25 in microns
so lotsa u
great work tho, you are an adventurer!
but if you lose quadrature counts (from either) they'll diverge
there have been so many questions, we just wanted to try it out
and luckily Stuart has about the biggest machine ever made, with the right feedback setup ;)
That is awesome.
what brand scales, i have a Heidenhain tester for the 11uA style
1 u - .0001?
25400 counts / inch
oh just 3
I think they go down to 0.5 actually, but they're set for 1 right now
1um = .000039yadayadayada inch
yikes. this http://www.electronicsam.com/images/KandT/conversion/accupins.JPG
will give us .0001
.0001 inch is pretty coarse for a scale
look carefully at the requirements of your resolver
we took a look at the AD2S1200 development board here, and it won't work with just any resolver
yes - I knwo
i don't know if it will work or not with these resolver like pins
you need to find out the allowable excitation frequency and signal levels
sounds like it was pretty fruitfull there.
I think so
"if you lose quadrature counts"... dont go nuts before looking at the output with a real analog scope, see that the magnescale output is constant over length, and the rotary is constant over a few revolutions. ( and directions, no sloppy reader heads/pickups )
tomp: unfortunately, nobody brought the test equipment that we would have done that with
everyone *did* say, "oh, I could have brought a scope"
though actually we found a working scope (10MHz though) I don't know if we'll have the time or motivation to do that
we have a b&k 10mhz scope. It definatly has fixed quite a few things
we did back and forth moves over 96", and got repeatability (as indicated by - er - a pair of indicators) to within ~1 thou
they have different counts/inch
that is pretty snazzy for 96"
yeh 1 in 10000
I am sure you know this - your site is down.
[03:18:11] <skunkworks_> http://www.electronicsam.com/images/house/range.JPG
little messy - but getting there.
ooh pretty, shiny
and yes - a goblet of wine.
enc_core_stepper.hal:32: pin 'encoder.0.position' does not exist
gmail said one of the files I sent to someone has a virus in it
will not let me download
usually any exe will be treated as a virus ?
must be, but I'm trying to think of why there would be a exe in there
cant check because they don't let me
it might have naked pictures of the president playing with a rubincs cube in there of national sekurty
if I had those, you would have seen them, or I wouldn't be typing this because I'd be at gitmo
actually, gmail doesn't let you send executables, can't imagine why they decided it has a virus
was probably a xls virus or something
something in ms office or maybe it just had a weird format naming
gmail doens't allow zip files either
you have to rename the extension.... .EXE to .XEX as example
I thought you could have zip files, but no exe in there
I usually use tgz
[06:13:47] <SkinnypuppY34> http://imagebin.org/14101 http://imagebin.org/14102 http://imagebin.org/14103 http://imagebin.org/14104
Tonights EMC fun
Used a 1/4 " endmill pretty good tolerance the slots came out .505 instead of .5 dead on . Not bad for ~ .010 b.l. on acme screws
SkinnypuppY34: hint hint nudge nudge http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=4181
SkinnypuppY34: is that steel or aluminum?
12 L 14 off ebay . Thanks !
SkinnypuppY34: is that steel or aluminum?
Leaded steel , screw machine stuff
much better finish than 1018
I'd been using a can of koolit to keep the tool sharp damn chips wouldn't brush off
How could they, you dont have a chip brush left =)
Oh that one .... it's more of the nasty broom
There's a ballscrew on fleebay I can use for my spindle, I'm determined to win it if I have to eat shoe laces for a while ... I gotta have some Z
Just remember to boycott ebay for a bit http://money.cnn.com/2008/02/15/smbusiness/ebay_strike_update.fsb/index.htm?cnn=yes
I won't be selling anything there for a while...
I wish craigslist would do a free store alternative to ebay that would attract more listers there
[07:08:57] <SkinnypuppY34> http://cgi.ebay.com/Quick-Change-Tool-Post-Taig-Sherline-Atlas-7-9-lathe_W0QQitemZ220202812985QQihZ012QQcategoryZ45019QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
How nice of em to include the prints ...
BigJohnT:were you able to get the tlo.py to work
jepler:can you take a peek at this,im not getting any errors,emc will turn on and can see the bits change in hal show
[14:49:40] <jlmjvm> http://pastebin.ca/908487
but it gives an immediate joint following error,alredy swapped the a and b phase,same result
im pretty sure before i could move the motor and see it on the readout when the drives were off,no change on the readout now,but i can see the a and b phase lights change in hal show
Guest339 is now known as skunkworks
heh - say - taking a quick look at your hal file - it doesn't look like your encoder counter is hooked to anything.. (except the physical encoder pins in the printer port)
do you have your original hal file that worked in 2.1whatever
it was a 2.1.7 file
i can paste it
[15:16:47] <jlmjvm_> http://pastebin.ca/908514
* skunkworks isn't a hal expert. I just play one on tv.
the old file was for the z axis,im using the x axis now
22. linksp Zpos-fb <= encoder.0.position 114. linksp Zpos-fb => axis.2.motor-pos-fb
your not doing the similar with net X-encoder-fb <= encoder.0.position in the new hal file
got any ideas
i changed it to net X-pos-fb <= encoder.0.position and same result
it should be Xpos-fb
unless you changed it further down
* skunkworks is winging it
i appreciate your help,been winging it most of the night
net X-pos-fb <= encoder.0.position,is that what you meant
yes - should be xpos-fb
down in your hal file is this line net Xpos-fb => axis.0.motor-pos-fb
which is what your trying to hook the encoder postion to. (if I understand it) and X-pos-fb <> Xpos-fb
[15:53:01] <jlmjvm_> http://pastebin.ca/908547
thats what i have now
stepper with encoders?
anyone attempted to make laser beam shape compensation on EMC ?
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-02-18.txt
micges: are you working on some CAM routines for linux?
ok, sorry was thinking of someone else
jlmjvm_, then allow enough error to avoid joint following error
thats not the problem,the display isnt showing encoder movement
ah then connection problem
or hal file
im sure its in the hal file
connection is good,i can boot up in mach,move encoder and see dro move,did that to double check the connection
can also see the pins in hal when i move it
but its not getting to the dro
SWPLinux:could you take a look at this if you get a minute http://pastebin.ca/908547
line 24 and 114,115
Wow - the gangs all here. :)
you guys are still there?
yep, but not for much longer
normal life calls
jlmjvm_: two problems with those lines:
1) the newsig command on line 114 shouldn't be needed, since the signal is created on line 24
2) the name is misspelled on line 115 - it's X-pos-fb on line 24, and Xpos-fb on lines 114/115
<skunkworks> which is what your trying to hook the encoder postion to. (if I understand it) and X-pos-fb <> Xpos-fb
its definately working
now i can start straighting up the cabinet,was looking pretty good yesterday,looks like a birds nest right now
really appreciate the help guys
SWPLinux:didnt you tell me once upon a time that i could possibly have latency issues with an onboard parport and a pci parport
i can get a dual pci parport and turn off the onboard if you think it would be better
I think it has been shown that pci printer ports are a little bit faster than the internal ones. but I am sure it is computer dependant.
seems like the dual could possibly work better since it would be matched,but figured i would ask first
it will take more time to read/write multiple parports
one thing you should try to do is use one port for input and another for output
that might save a couple of accesses (and it might not be all that important anyway)
do you think it would be worth 25 bucks to try the dual pci and turn the onboard off
should have another computer put together tomorrow,it may work better
will try it first
it will be a different chipset and a faster cpu
will holler back later,gonna hook up the other encoders and fix this birdnest,thanks again
SWPLinux: Hey, do you have any P3 slot1 > 600MHz and => 512 cache?
[18:24:58] <skunkworks> http://www.electronicsam.com/images/866mhz.JPG
what the new scope or your feet or the pc
your supposed to do the math.. :) 40khz
shoes are old
maths!!!! hell no /me watching his first steel pinion being cut on the machine
heh - cool
too much vibration to have camera on it at the moment
I had to cut one to see possible problems :))
a guy just asked for a tapered hole 23 feet long in a forged tube that is 12" OD with a 9" to 3" ID. he wanted to WEDM it, and hold a .01" tolerance over the 21 foot. I said no can do. the deflection of the wire just from the discharge would be more than .01"
one of many things that just wouldnt work over that len
I wonder what he thinks he needs that for.
wouldnt say ;)
I bet it's something that he didn't really need that for.
is it just me or does that sound like a really big gunbarrell?
guns dont have tapers like that as far as I know
yeh, tapered bullets :)
sounds more process oriented
he sent a print with info covered up, and sez he can get a contract for 'several' just too wacky
* quiteBIGeye thinks it sounds mad...
I saw some silly small WEDM at a show last week
wire electric disharge machining
ahh.. we got one of those, but didn't know the englisk name for it... pretty difrerent in swedish :O)
woot gear to size /me can go out to lay now
whats the word in swedisk: like erodieren? funk erodieren?
"gnistning" seems to be a more common name
tnx Vq^ '
how goes it?
it goes sotoing at the bar waoitng for food and having my brakes worked on
I can't look atm, oncellphone
ircing that is
oh - then never mind. just a picture of the new range in the unfinished kitchen
I was wondering why your typing sucked ;)
ah very cool ill look when I get home
also - a 40khz pulse train out of emc on a 866mhz pentium III http://www.electronicsam.com/images/866mhzyum.JPG
it has been running all morning.. no latency errors
DanielFalck: actual computer http://www.electronicsam.com/images/866mhz.JPG
some compaq of some sort
deskpro - one of those mini cases
what did you have to do to get to that point?
sacrafise a kid
nothing - other than turning off 'power saving' in the bios. Just a dapper install and emc 2.2.3
what is the max lat?
used the step_conf - which worked great by the way - thanks jepler.
I would have to check.. It was between 10-15 iirc
with the 'power saving' on - it would climb to 50us
I have a dell dsktp with nothing on board and I get about 16
what are you hoping to get?
going to try headless and see what it gives
under 10 and I'd be happy
that is hard.
I don't see many under 10 - that I have played with. (I don't really 'tweek' them though)
I know but I might be able to headless
maybe.. I have never tried it.
doubtfu but I'm going to try
This one stays under 10 for a long time.
but I know it goes over after a while
we really need a hw db
of what mobos work good or bad for that matter
there is a place for it on the wiki.. but I never remember to add to it.
I always have trouble adding to a page that already exists
i have a question about the traj planner do i ask it here or in som other like "dev" chanel if there is such
ask - some one might know. Right now a lot of the developers are traveling.
okay, well my question is this: if i have an axis that is much slower than the rest, how do i make the machine go full speed when that axis is not in motion.
my machine at work tops out at 12m/min one axis but i can go mych faster if i use 2 or 3 ""word... axeses?? ""
that doesn't make sense... Each axis can be congiured independantly
oh nm me
but each axis needs a "headstart" over the traj planner... right?
so if i have like speed 2 on one axis and like 100 on all other i would have to set the traj to below the lowest...
Yes - it will go as fast as the slowest axis - (on multi axis moves)
isn't that based on what gcode its running?
emc takes care of it. it will obey the constraints of each individual axis.
but if the traj speed is set to higher speed than the slowest axis i get at lot of faults, some thing like "axis following fault"
that makes sense
its been configured todo one set of settings and it sounds like you are telling it to break the speed limit
so to speak
exactly, so.... i have a sneaking fealing that the z axis on our next machine will be quite slow... and i don't wan't to run the whole machine on half speed...
can you improve the z on the other macine?
then there is something wrong.. It should not be a problem.
well, yea, but that is not the point... as soon as you ues more than one axis it's mad to not use full speed...
The traj section is usually set to sqr(velx^2+vely^2+Velz^2)
you can't ask granny holding hands with 3 5year olds to run as fast as them
hmm, will have a talk with Vq^ about that, because we can't get it to run stable if we have traj set to more than the slowes axis...
im not sure
you should be able to set the traj limit to infinity and still have a stable machine
i got the impression that the traj-planner ran ahead of the axis then
or something very very large, like 100000
SWPLinux: ok, then we have done something else wrong
thats good to hear
coordinated motion is limited by the requested feedrate, the traj limit, and eaxh axis limit
quiteBIGeye: not that good, we have still got an error
I missed the part about what actually happens :)
but it says something like "to avoid axis following error, the axises have to have a little headroom over traj"
i read somthing like 5 - 10%
your probably talking about stepgen headroom. the stepgen vel has to be a little higher than the axis velocity for that axis
Vq^: is that what i'm going on about and not realizing it?
since each axis can move as fast as its own limit (in [AXIS_n] sections), but stepgen may need to "catch up" with a new velocity command, it needs to be able to go a little faster than the fastest rate you actually expect to need
"We have discovered that it needs to have a little "headroom" over the accel by the trajectory planner, otherwise it can fall slightly behind during accel and later overshoot as it catches up."
it's more like you have to reduce the limit in the AXIS sections, not that you have to add a little for stepgen
that's the axis accel though (or TRAJ, whichever is lower)
accel - thats it :)
one of these days ill have tp figure that out
quiteBIGeye: guess we will have to try and increase the traj-plan vel
yea, like maad :)
Vq^: feel free to do your *1000 increments :P
is it acceptable to have a higher traj acc as well?
Jymmmm: been running for a good hour or more.. 11us jitter
does that work the same? I wonder - trajaccel = sqr(accx+accy+accz) ?
from a purely physical standpoint, yes
then that is what I would set the traj section to.
you can set the TRAJ limits as high as you want, the AXIS_N limits will supersede the TRAJ limits
assuming there are no bugs :)
thanks Vq^ for gnistning ( handy for web searches, if you know the process name in various languages, I find lots more with elettroerosine & funkerodieren than with just 'edm' )
good night all
* alex_joni will be travelling the next couple of days..
Would anyone be interested in some P3 motherboards? http://www.intel.com/support/motherboards/server/l440gx/
Would anyone be interested in some P3 motherboards? http://www.intel.com/support/motherboards/server/l440gx/