if i have ti go that way ill get a 120 vdf and cnc the 1 slide
the 1st pic shows the nkob on the back
rayh... any thoughts on what the knob might be??
hi matt, ken
anyone on here who knows what the table3, table4, table5, etc. bits in the .PIN files in src/hal/drivers/mesa-hostmot2/firmware/5i20/ are for/do?
I've never looked at that stuff. (Hi, Matt)
me either until today
News from the home front. I'm making progress on GWiz.
oh, sorry, Hi Chris!
seb would know. he was on -devel earlier tonight.
mshaver: did you see kirk's rigid tapping on his HNC?
seb seems to be the architect of that directory
cradek: no. youtube?
The turret seemed very fast, but it looked like it turned in only one direction (or could be my eyes).
it does only turn in 1 dir - air motor...
yes - see emc-users
Quoting the list:
[00:59:57] <lerman> http://www.wallacecompany.com/machine_shop/HNC/Rigid_Tapping_3.html
[00:59:58] <lerman> http://www.wallacecompany.com/machine_shop/HNC/Rigid_Tapping_1.html
[01:00:00] <lerman> http://www.wallacecompany.com/machine_shop/HNC/Rigid_Tapping_2.html
i'll ask seb on the devel e-mail list about those bits
I think they may be for other step modes
swpadnos: that makes sense, I just can't eat up i/o bits for them
do you need those stepgens?
yep, need 4 stepgens would prefer cw/ccw bit step/dir is ok
you should ask seb if the unused pins are available
i want to output the step and dir to this: http://www.mesanet.com/pdf/parallel/7i34man.pdf
then you should be looking in the firmware/7i43 directory :)
8 x rs422 channels, so i............ thanks for that! you knowi'll have to look now, even though i know it's not there...
if it's not there, you need to update your CVS checkout
7i34, not 7i43!
it's pretty easy to make a customized version of the firmware, so if the other pins aren't available for I/O, that's a viable alternative
it looks like you make a .PIN file, then some magical process is invoked and a .BIT file is made
you make a pin file, plus an extra vhdl file that generates the components you want
then invoke the magic process
(which is a 1.7GB download, so start now)
oh, some xilinx thing eh?
heh - yeo
I don't know if there are makefiles for the FPGA stuff
a tutorial or recipe would do
yeah - that would be nice
BTW, could you scan me a copy of _The Book Of Thoth_ and e-mail me the .pdf? I've got some things I'd like to change around here ;)
too late, you will be judged
hey, that's eventually, this is now!
dude actually pulled it out of another dead guys tomb... damn!
on another note, did/do you make a little step pulse generator board?
with an AVR on it?
are there docs?
[01:28:25] <SWPadnos> http://www.cncgear.com/MPG/
I guess I should have cleaned that one before photographing it :)
that's it alright!
sure, I've got the build docs
I may even have one or two in a bin somewhere :)
2.25GB! Just to blow the fuses in some over-glorified PAL... <grumble>
oh, it's gotten bigger
there's an import tcl script in the 7i43/SOURCE directory. you should be able to run the IDE and import the project, which may make it easier to make changes (and manage generating a bitfile
why I bet this whole thing would fit in a 22v10 and we could just us CUPL...
heck, draw it by hand
good! hat might save a lot of head scratching
especially in the woods
(fewer ticks when you wear a hat)
i'm typing on silicone mt keyboad, can you tell/
dumped coffee on my otherone
I first downloaded it about 18 months ago, maybe even less, and it was 1.1G
bloat is nothing new, but the speed of that bloat is...
it's only twice as big
moore's law - 2x every 18 months
I didn't expect it to apply to software
but at this rate it will soon be too big to fit on the internet
no. it will be too big for anything _but_ the internet
even a DVD is only about 4G
yep, one more doubling and they'll be into Blu-ray
oh, I forgot about bluray
the internet also doubles
mshaver: I think I know how to run ise and generate the bit files; when I am confident in my directions I'll put 'em on the wiki
there is a makefile of sorts in CVS (I think)
jmkasunich: for the hostmot2 stuff?
yep, every time they update that program1
no, for the generic mesa stuff I was working on
I suspect it wouldn't take a horrible amount of hacking to make it "compile" the hostmot2 source
jmkasunich: that may be true
a tutorial would be good
back in a minute...
[01:58:10] <jmkasunich> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/src/hal/drivers/mesa_5i2x/firmware/Makefile?rev=1.8
the .vhd: .rspec .spec step is specific to the flow I was working on
if you start with .vhd files I suspect it will work pretty well
"a dwell of 10 seconds must be programmed as G4/10.00; trailing zeroes and decimal point required."
ugh, anyone want to guess whether the ; is part of the gcode or punctuation in the docs?
I noticed that the G-code they wrote for that big EDM machine had semicolons after every line, and the guy always put the decimal point in
cradek: what system is this?
the decimal is always required for BOSS too
surely not emc
you haven't replaced that yet?
whatcha waiting for?
it works, don't break it ;)
steve stallings is making a cool retrofit board and I want to help with that, so I'm waiting for that
thats a better excuse than I have for most of my delayed projects :-(
yeah it's a great one
Sigh, I have come to the conclusions I am just too stupid, dense or something to understand gcode enough to pull off a NEMA23 mount. I would KILL for that program I had when I was working on the Moogs that you drew the part and spit out the gcode for it, sigh, SCREAM</frustration, heh>.
No idea where to go next.
the G-code you last posted is here: http://www.pastebin.ca/1057221
can you also pastebin your tool table?
That is WAY back now.
(I think you did that before, but hey - I'm forgetful :) )
the thing thinks it's smarter than I am
one thing I notice in that G-code - it's unknown where the tool position is when cutter comp is turned on
IT IS WRONG
I can get holes almost any place except where I want them and never in the right size, it is really funny (absurd, ridicules, discouraging and depressing :-( ).
you may want to G0 X0 Y0 Z0 just before the G43 lin
One would think that as common as this all apparently is there would be a gcode that says go HERE and make a circle THIS large.
it's called G2 or G3
Rather than having to go there, compute offset, determine arc, figure in tool size....
no no no
G0 X0 Y0 Z0
G0 <wherever you want the circle start point to be)
Yeah, I am so <pardon> fucking confused that this point it is pitiful.
moog... was it ez-cam??
EMC computes the offset
I don't remember it ran under dos and you drew a 3 view of the part and commanded PUNCH and took that tape to the moog and ran it.
you program the actual size you want, and you tell EMC how big the tool is, and it computes the path needed to get that size (for the most part)
Maybe my install is frucked... Dunno....
At this point I am willing to throw money at something rather than toss all this in the crapper.
Yes, my frustration factor is out of sight ATM :-).
Coffee with a hearty shot of jack, coming up :).
zephyr99 is now known as john_f
Gathering up my notes and stuff and sorting them down to the last pass, back in a bit...
I think the first thing I need to find is something more than a reference for the gcode stuff.
lets see what amazon has...
make sure it describes the flavor of G-code that EMC uses :)
ONLY relevent gcode reference = manual for the controller
fuck a textbook, and fuck the actual spec
oh except for the pictures of made up machine tool configurations
thats pretty neat, machine manuals dont have that
Well, the problem here is that *everything* I have found is a reference at best, they all seem to assume that the particulars everyone knows.
That NEMA23 mount seems like it should be plain dirt simple...
G2, G3 I J K or R Coordinated helical motion ("Arc feed") CW or CCW
helical? Seems to be linear...
the main point is that the G-code path should describe the *uncompensated* outline of the part
helical because Z moves while X+Y are doing an arc
so the path is line a spring
OK, question, let me set it up.
As I understand the problem...
Lets just take the 1.5" boss cutout.
Z ain't moving here unless I do it.
OK, for the part, I choose to make the center of that circle to be X0Y0
First I have to move it to the start of the "arc" (may mean same as circle?!?).
a circle is a 360 degree arc
So I guess I pick a point 0.75 off the center either x or y.
trick 1: when you use cutter comp, you have to enter the compensated move at a tangent (ie, already moving in the same direction as the first compensated cut)
The thing that is not clear to me is what the tool path actually is - tool centerline I suspect.
so instead of moving in a straight line to X0.75Y0, move in a half circle of 1/2 the diameter to the same spot
Ouch, at this point that makes NO sense to me at all. From that I assume the cutter is outside the radius.
don't worry about the tool path yet
don't worry avout the radius yet
just think about making a 0.75" diameter half-circle to get to X0.75 Y0
Err half the diameter would be 0.375 - even smaller than what I have been getting.
half the radius is 0.375
which is fine
I ***HATE*** to be dense, but that makes no sense to me.
a 1.5" circle has a 0.75" redius
a circle half that size is 0.75" diameter, 0.375" radius
So I got to compute 1/4 radius and then the tangetnt of the the 0.75 radius then move the tool to there rather than circle center?!?
which made no sense? :)
wait a minute
Please, I have reached a point of terminal confusion, sorry.
don *NOT* think about compensation or cutter radius or gouging yet
it will only confuse you :)
Not even considering the compensation stuff at this point.
Trying to figure out how to tell it to cut out a 1.5" dia circle at a given xy
Lots of slop possible in the boss cutout, mounting holes are a whole nother problem'
ok, so a half circle that gets from X0Y0 to X0.75Y0 would be G2 X0.75Y0I0.375J0, right?
Lemme get that off where I can parse it...
I suppose I should fire up EMC to be sure this will work - hold on :)
G2 X0.75 Y0 - << makes sense...
No where have I seen the I and J parameters defined...
where have you read about arcs?
Gurgle the ref.
To my learning an arc is a part of a circle...
[03:42:12] <SWPadnos> http://www.linuxcnc.org/docview/html//gcode.html
click the G2 or G3 link
Where is I defined? Why I want a parameter half short I don't get.
I and J define the center point of the arc
it's in section 3.3.1 of the docs I linked
I is nowhere near that
when you are ON the circle, the center is somewhere else. the I is the >distance< from you to the center measured along X, J is the distance to the center measured along Y ( for circles in G17 the XY plane)
the edge of the circle is at 0.75, 0.375 is in the middle of the cutout
Err circle center changes - not in my world ?!??!
please read the manual section on arcs if you haven't already
G-code only has access to two points with XYZ coordinates: the current point, and the XYZ values you enter
a circle requires 3 points to be fully defined, so IJ are used to define where the center point of the arc is
Wow, sounds like the hardest possible way to accomplish that...
where's the logo interpreter for emc :)
BYW, read that way more than once and it makes little sense. parallel to the plane which is shown as two axis...
it's not possible to know what arc you want if there is no center point defined (or you don't have 3 points on the arc)
Center 0,0. Diameter 1.5". That just seems simple enough for a circle!?!?
sure, it is
Rather than breaking it down in ways that seem like pure masturbation.
doesnt say which way
Or CNC programmer enrichment :)
I'm trying to help you get something that can be used with cutter comp
Who cares, that is one more parameter L OR R
OK, I really am trying to understand this, really, I am NOT trying to be obtuse.
I believe you
Either I have finally run into something I am incapable of understanding or something.... Working anyway :-).
The I0.375 I do not grok, it seems totally out in left field...
you really need to know how arcs work :)
It isn't a point on the 1.5" diameter circle
you tell it where you want to go to, where to swing around, and wether to swing cw or ccw to get there. its just a convention.
I guess so,
ij = pivot point
it's like tying a string to the point defined by IJ
and drawing at the end of the string
Arghhh, the pivot point would seem to be 0,0?!?!
I am describing a DIFFERENT ARC!
Alien to me at this point, trying to understand it.
as I said in the first place, if you start at 0,0, then instead of moving to the edge of the circle in a line, let's do it in an arc of half the radius
methinks we need a diagram of a lead-in move
To me an arc has always been a part of a circle and the 0.375 seems to be in a world of it's own. Lemme go google on cnc arc and see what I find rather than abusing you folks.
K`zan: cutter comp isnt the easiest thing in the world..
i mean it really is pretty sucky, it's not just you
fenn: I am not even there yet
I just want to make a circle...
a circle is easy as pie :)
Middle of the cutter, inside or outside - doesn't matter. I would think I could change the diameter to 1.5-(1/2 cutter dia).
G0 X0.75 Y0
Shutting up now :)
do not think about which end of the cutter is on the circle yet! :)
that gives you a circle of radius 0.75 (diameter 1.5)
That makes sense
ok, now if there weren't issues with gouging errors, you could do that with cutter comp turned on, and you would get a circle of the correct size - EMC would mae a smaller circle as needed
what does that mean?
sorry / communications stuff - habit - roger, roger (understood) - shorter to type.
ok - hard to tell if that's a growl :)
that's why I asked
when you on the circles cicumference at 3 oclock, the center is to your left (minus 3/4 inch )
* K`zan growl
sry if this is too late, here's what swp's trying to draw i think? http://fennetic.net/pub/irc/lead-in.png
or maybe rotated 90 degrees
I see what is happening - the move to the desired radius is 1/4 dia.
But why not just go to the desired starting point?!? (straight lines is the....).
yes, that's what I was doing (only on the X axis :) )
Seems like abuse of run time and tool wear/
to avoid the gouging error ;)
because historically EMC didnt do corners well in tool comp mode
but now it will happily do what you want (i think)
you can do the lead-in move above the part
gouging == over run at the desired point?
then do a spiral (helix) down
gouging means moving into an inside corner
(for hte most part)
that's why you can't just move in a straight line - the right turn is an inside corner, and gives you that gouging error
Arghhh circles don't have corners in my world?!?? Pardon, trying to understand the why so I don't have to do this to you again :-).
(the right turn is the transition from moving along X in a straight line and moving down as the start of the circle)
Ok, I think I am seeing something creeping up on my bleeding brain.... working....
It doesn't *know* we want a circle.
I would have thought G2 / G3 would imply that?!?
the gouging error he's talking about applies to any inside corner with tool comp on
So rather than moving directly to the point, we slide into it ?in case we are not doing a full circle?!?
it doesn't matter if you're doing a full circle or not
you have to approach along the same line of travel as the compensated move will be
what actually happens is the tool moves in a sort of squished arc because the compensation comes on gradually over the length of the lead-in move
if you look at fenn's diagram, what EMC will do is move the center pointof the small arc, and make it smaller, so that it ends up inside the larger circle by the right amount
then the larger circle will also be made smaller by the right amount, so the outer edge of the tool is on the line you programmed with the G-code
(line being path here, circles or straight moves)
OK, starting to make sense, but:
G2 X0.75 Y0 I0.375 J0
Would that not make a 3/4" circle?!?
On one side of the desired circle?!?
lemme think about this - that code makes sense in that drawing...
yep - it makes a half circle
circle is defined by the X0.75 Y0 positioning and because of cutter comp is going to be used we add the I0.375 to cause it to slide into the right place rather than just cutting a plain circle w/o cutter comp?!?
I has nothing to do with comp
we start at 0,0
we want to be moving down at the point 0.75,0
with you so far :)
so we make an arc that looks like a setting sun, going from 0,0 to 0.75,0
that's a half circle, since I can do that math in my head :)
emc2 user manual section 20.4.1 ( has a couple visuals for cutter comp) . for circle formats 14.3.1 ( no pictures... i hate cookbooks with no pictures, i must have dreamed JMK's pictures for this :P)
Beats HELL out of trying to figure out tangents :)
the pics are on the quick reference card
the center point of a circle with a diameter from 0,0 to 0.75,0 is 0.375,0
which is online, part of the docs
I don't think the back side is online though
Those might start making sense to me now.
[04:12:40] <SWPLinux> http://pastebin.ca/1061041
when i get my machine to home properly, all my parts are mirror imaged....
SWPadnos_ is now known as SWPadnos
mdynac_: are they normal before you home?
did that pastebin from me come through?
[04:16:03] <SWPLinux> http://pastebin.ca/1061041
mdynac_: then just put a minus sign before the SCALE value for the inverted axis
ok, there it is
k , as using stepconf to set it up
you could also invert the direction pin
no ay to do it in stepconf?
[04:17:48] <SWPLinux> http://pastebin.ca/1061041
i think stepconf has some invert checkbox
G2 X0.75 Y0 I0.75 J0 - makes a circle!
Lemme scroll back and see if I got that right...
yes - use stepconf to invert the direction pin
<SWPadnos> G2 X0.75Y0I-0.75
load that pastebin and play around with it in axis/sim
what do you call it in stepconf?
drawings for circles and cutter comp
K`zan: remember that i,j are relative to the current location
look at the pics, they should help
jmkasunich: thanks i was looking for that
the key point is that you always program the uncompensated path
there is no "direction" pin in stepconf
do you have servos or steppers?
then you probably have signals on the parallel port setup page that are names "X Step" "X DIrection" ...
in pastebin xmpl: does the comp turn on thru the 1st 180 degree move, and is fully comped for 2nd 180 degrees? making a sort of spiral, i think
invert the one that says Direction for the axis (or axes) that move the wrong way
yes look at jmkasunich's good pix
methinks according to jmkasunich's pdf, that the compensation commands begin at where you are and are fully in or out of effect at the end of the gcode line that requested the compensation/or lack of compensation.
moral: be fully comp'd when you get on the finished edge, and turn when you leave that edge..
i like to leave the edge before blowing comp
just to be safe.
make that 'turn it off AS you leave"
we used to use 'the worm' just before entering the contour. like fenn's curved leqdin but just larger than the tool. http://www.drahterosion.com/english/frames_e.htm
i like a lil' arc with some hot -z action
for pocketing, anyway
i guess Z doesn't matter for a perhipery
does this sound familiar to anyone? rtapi/sim_rtapi_app.cc:75: error: ‘strtol’ was not declared in this scope
also missing remove_copy srand48 lrand48 and exit
okay nowmy machine moves properly, but it homes in the rong direction....ho do i fixthat?
change the sign of the home search velocity
thanks guys....it works...
sigh.. need to download 155MB to satisfy build-dep
* fenn grumbles about lyx
gee now the emc goes into pause and won't move
Is it possible to comment comments - Sometimes I need to comment out a block of code?
K`zan: you could put it in an o-word loop (use if statements)
just use "O2000 sub" at the start and "O2000 end" at the end (no if needed - the sub is never called so it won't execute)
Thanks! I assume <ulp> that the 000 is incrementable if one needs to do so more than once?
Recoding my program...
K`zan: load this one up in Axis: http://pastebin.ca/1061076
rr, copied it out and will try it here in a few.
On the desktop here - no EMC on this machine. Got to go back and fire up EMC box as I got tired of being scrunched up working in there :).
install sim and play around with it
Would be nice if it would work here.
rather than running a machine all the tiem
yes, the emc2-sim package
Not running machine, just using axis to see what it does.
if you're on ubuntu
checking now, thanks!
;-( adept doesn't find it :-(
you probably need to add the emc2 repositories
Do I need a special respoitory?
[05:48:48] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2
there are all sorts of sim build instructions there, but I think there are also sim packages
ah - you need to add emc2.2-sim to the repo list
I think: http://linuxcnc.org/dapper/
where the instructions say emc2.2, make that emc2.2-sim
[05:51:16] <SWPadnos> http://www.linuxcnc.org/dapper/emc2-install-sim.sh
Jesus, couldn't they just list the repo rather than having th read through all that??!?
why would anyone do that
Got universe, still not there.
when it could be hard instead
For those of us who are not into masturbation for the sake of masturbation...
EMC2 is not in any repository for ubuntu :-(.
well, there are lots of people who have never heard of a repo, so the instructions are meant for them
toastydeath: Same issue with uncopyable dialogs...
did you look at that script I linked to?
Open with what, lemme copy it and go to the browser manually.
run it in a shell
or look at it to see what you have to do :)
OK, got it, lemme try it.
4,623 open windows :)
i wonder why lyx thinks it needs texlive-full
emc2-sim: Depends: libpth2 but it is not installable
Sigh, easier to just go fire up the box, back on that in a few.
damn. this is something we've seen before
K`zan: sudo apt-get install libpth20 libpth-dev
i mean, libpth2
or maybe its just libpth now
oh now it's libpth20 in ubuntu too, yay!
are there sim packages for hardy?
i must admit i'm confused by the repo organization (or lack thereof)
it doesn't look like there are
it changed, but we couldn't move things around once people had installs pointing to the old location
or something like that
K`zan: are you installing the sim packages on hardy?
looks like it.. i dont see any way around the silly libpth2/20 issue besides rebuilding the package
especially with packages.ubuntu.com dead
SWPadnos: VERY useful example, thank you VERY MUCH!
Gotta rehack mine based on that, should be a lot shorter this time :-)
I don't know why that last small arc has a jog at the end
Only need the arc on the lead in and out I see.
note that the axis preview shows the travel as +/- 1.38 (on my computer anyway)
Tool diff ?!?
the tool I used was a 0.125 cutter, so the travel is 1.5 -0.125 = 0.375
Hacking merrily away now :)
With a bit more than a vague clue :-)
if you comment out the G42 line, the size in axis will change
heh - glad it helped. it's bedtime now
Noticed that when I put it on at one time.
actually, it was 2 hours ago, bit oh well
Rest well, you certainly earned it tonight :-). Again, thanks VERY much!
Well, it made sense with a 0,0 ref, things get crazy from there. BUT, making progress I think. Brain is a bit fried tonight, will jump on it again tomorrow. Amazing the math required to make a simple circle.
Hey guys long time no see. I was wondering where I can look to see the cutter comp routines. I'm having an issue where I can't use cutter comp and need to right my own I was hoping to see how it was done in EMC for some inspiration.
Is this the place to ask a question?
there's one :)
A question about emc2 of course.
yes, this is the place
assuming you're talking about the machine control, not the backup system
Ok How would I go about changing the Rapids default speed?
rapids are done at the maximum velocity, which are set in the .ini file for each axis
rapid speed is defined as "as fast as possible". are you looking for rapid override?
* cradek kicks SWPadnos
Not overide but the actual g00 speed.
did you use the stepconf wizard to generate your configuration?
Yes stepconfig was used
ok, then you can run it again, load your existing configuration, and change the maximum velocity setting for each of the axes to what you want.
are you trying to speed it up or slow it down?
Slow it down . I'll try the stepconfig wizard
patlbs: remember the velocities in stepconf are per second, but F word in gcode is per minute.
so if you want your rapids at 120 inches/minute you will put 2.0 inches/second in stepconf.
mm works the same way if you are metric.
metricF is in meters per minute?
or mm per minute
feed is always units/minute
metric is mm/second in stepconf, mm/minute in gcode
if you've defined your units as meters, then it's meters/min
* cradek kicks SWPadnos again
gcode is in inch/min or mm/min
* SWPadnos blocks
oh yeah :)
anything else is crazy talk
I think I mentioned somewhere that I need more coffee
i dont think ive ever metric code
and shin guards?
those silly G20/G21 codes
no, I'm a black belt
well, I have a black belt anyway
hehe, so yesterday i reindexed the tools and flycut the vacuum table spoilboard
and when the operator went to cut shit, you could see cut marks in the spoil board, but could not feel them
I'm wearing mine now - it holds my pants up
the last guy who setup the machine is a failure
yeah. I think I have to fix my buckle for them to work that way
in the immortal words of Donald Trump - "You're Fired"
i have a nylong web belt
with chrome buckle, the strap is black with little jolly rogers
Okay very well. Another question?
don't ask to ask, just ask
that's the primary rule of irc
some chans would just ban you for that, heh
may I ask a question?
people will answer when/if they can, sometimes it's much later when they see your question.
may I ask another one?
* cradek kicks everyone
* alex_joni goes home
s/goes/crawls after cradek's kick/
patlbs: we're always like this, sorry
that's not true.. sometimes we're worse
actually its quite tame here, be happy you didnt ask for help on efnet
normaly very civil in here compared to ##php or ##c
you would get it, but itd likely be wrong, probably on purpose
That's okay. I'm new to the irc stuff and EMC2 to also too to many two's
i has a lag
How do you home an axis after losing steps on that axis?
you can home an axis as many times as you want by pressing home
dont go to fast to lose steps
But I run out of travel on the machine.
I think you are going to have to describe your question more specifically to get a correct answer
homing can mean many things. in general it sets up a reference point (origin) for an axis. there are lots of ways to do that.
machine at works homes slow
Okay when I jog to a position then hit home it zeros. If while running a program I lose steps I cannot go back to the original position it will only go to the zeroed position which is incorrect.
without home switches that is what you get
you need to either (a) reduce the speed so you don't lose steps or (b) add home switches so you can re-reference after losing steps
SWPadnos: is fast on the keyboard
(a) is preferable, since your part is already ruined when you notice you've lost steps, so preventing that is best
Yes reduce speed but I need to start somewhere. I'm only testing
on my tiny mill, the gib tension is alot of it
No real parts invovled
is like a tolerance vs speed tradoff
i can go 60ipm consistent if i leave the table a tiny tad little bit loose
but yeah with everything snug its more like 45
and for actual work i never set it over 30
obviously if you have a real CNC you scale those numbers up, hopefully
open loop taig fairie mill is an adventure
patlbs: move to where you would like to be and hit home on the gui
Emc won't let me jog past zero
i think he means how its not an absolute home
patlbs: for now you will have to restart EMC to "unhome". In EMC2.3 there will be the ability to unhome.
you might be setting home when you want origin?
soft limits are keeping him from getting back to where the origin is supposed to be.
the limits for each axis define how far you can move
like workspace origin, instead of machine home
yeah i hate that there should be a disable
renesis: there is in emc2.3
so i liked axis gui
but now im using some NUM controller bullshit at work
640x480 dos on xp with vid drivers only go down to 800x600
so now i really really like axis gui
stupid comercial pc based controller is worse than the three free options
Thank you and goodbye
come back anytime
but not too much you will get addicted
cradek: I have found the home/unhome of a single axis very useful on my plasma cutter
it would be neat for setting mavchine limits for fixtures and clamps, on the fly
or maybe if there was a dialog for soft limits
that would prob make more sense than changing machine origin for alot of things
I've hoped for configurable soft limits for a long time, but never done anything about it
heck maybe they could even honor TLO. now that would be slick.
i have to change the laundry to the dryer
so im doing 4 hours of bus everyday for 5 hours of work
half of the work is figuring out wtf my job is because the guys on the floor design all this shit
can you go to fewer but longer workdays?
they work 7:30 to 3:30
and i havent been there long enough to be asking for a key
hopefully i can get a car this month
to get like 8 hours, id have to wake up at 4:30
for the last 5 years thats like around my bedtime
* BigJohnT goes back to structured text programming for his panasonic plc, lots more fun than ladder
is plc like fpga or wat?
or wait is plc the big controller module things
non program-counter based text logic scares me
programmable logic controller
my mind always wants it to be Chip, no Controller
this one is very small about 3" x 4" x3"
lots of io still
and very fast with two 100Kh outputs
heh, thats not fast!
20 inputs 16 outputs
fast for a plc
or i dunno maybe for plc but i do asm code for speed at 20MHz, heh
ops forgot the M
i read a datasheet for one someone was talking about, was like a little tank thing
that was much bigger tho, i dont think it had many more io ports, tho
some dont believe, that you can IRC and machine :))
jepler: I got the tapping head mounted and working on the mill last night. a few days too late but nice anyway.
archivist_emc: you can irc from the cnc, heh
the realtime shit is really good about keeping the cycle going, i can browse web and the cnc keeps going like whatever, sometimes web goes a bit slow
can do quite a lot while the cnc is running
heh, wasnt so much the case with mach
I have an XML doc build (mysql docs) going on on the box as well
Is it needed to download CD with integrated EMC for Ubuntu 8.04?
[16:58:25] <SWPadnos> http://www.linuxcnc.org/hardy/emc2-install.sh
Hi guys... I have a little problem with my A axis: when I home it, it doesn't go to 0, like my X Y or Z. I use g92.1 a lot, but when I close EMC and start it again, it remembers the A position. Any idea what's wrong? thanks
do you have a home switch?
No I don't... my ini file section for the A axis is here http://pastebin.com/d281b8bf2
ok, then homing means "don't move, but mark the current position as HOME_POSITION"
you probably have a POSITION_FILE listed in the ini, which is why EMC remembers where it was
SWPadnos: any idea how I can reset it to 0 without unrolling all my turns?
so what does it do when you try to home? move toward 0, not move at all, set the location to 0.0?
I don't see a POSITION_FILE in my ini...
I don't know if you're supposed to set the HOME_*_VELs to zero, or just leave them unset
SWPadnos: When I home A (or all), A stays at its current position
and the position display doesn't change to zero?
The display stays the same
does the homed symbol ever show up?
the homed symbol shows up
oh - you have no HOME_POSITION defined for A
SWPadnos: thanks, I'll try to set one up... just a sec
err - maybe that's HOME. I'm not really the one to answer this question :)
Sigh, got the boss working (short of making an actual part) but the mounting holes I am still trying to figure out, thought I had it about 0200 this morning.
Suspect the end mill size is an issue for this. Going to have to change the hole size and just do a circle - maybe.
Or just use it as a spot drill and do it on the drill press.
what is the end mill size?
#36 - 0.1065
ok, and you're making a 0.201 hole or some similar?
G2 X1.0255 Y0.925 I0.1005
Makes for not a happy camper :-)
I think that should work, but probably nor if you try doing half-sized entry moves like that example
SWPadnos: I added HOME_POSITION = 0 to the [AXIS_3] section and restarted AXIS... no change
yes - you can tell that because I (the radius in this case) is < the tool radius
no, it's called HOME - my bad
you don't need entry moves for every hole, you only need them once at the start of the compensated path
Looks that way. Error message would take way too long to type in since it isn't copyable (grrrr) but yes, so just using it as a spot drill and then doing it on the drill press would make sense here.
That is what I thought I was doing - just the lead into the circle.
one lead in to the first circle, then no lead-ins for the other circles, then one lead-out on the last circle
Lemme comment that out and see if I get the gouge message
are you trying to cut all the holes in one program, or making separate programs? (I'd suggest one progrma, if it's possible to hold the work)
Nope, same messaage
Could do that if the end mill isn't just too big.
SWPadnos: the HOME = 0.0 was there already... I wonder what's wrong
dunno - I'd have to read the manual on homing to be able to help ;)
K`zan_emc, can you post your code to pastebin?
OK, thanks for the suggestions
sure. too bad they didn't help
by the way, POSITION_FILE is in the TRAJ section - dunno if you looked for it there (or did a search)
do the other axes also retain their positions?
SWPadnos: I have nothing in the [TRAJ] section about a position file. It looks like I have garbage in the value for the A axis.
hmmm. can you post your entire ini?
SWPadnos: here you go: http://pastebin.com/m70ca6f82
SWPadnos: Yes, just a sec
well, I don't know what's happening. sorry
OK thanks... I'll try the demo configs to see if I get the same behaviour
unless there's a problem in the hal files, which I doubt could cause the effects you're seeing
K`zan_emc, gotta run for a bit. I'll take a look at the code when I get back
SWPadnos: rr thanks, takes pastebin a while but: http://www.pastebin.ca/1061463
take out all the G40 and G42 lines except the first G42 and the last G40
and leave all the entry moves except the first commented out
now I'm leaving :)
SWPadnos: I found that if I use "92.1" in the MDI window before closing EMC, then I get A to 0 when I start EMC again, and homing sets it to 0. thanks
off topic sorry I have a three phase motor that only has L1 L2 L3 it shows a wiring diagram for 230/460 but only has three wires not the typ 9 ? anyone
is this a motor with jumper bars
K`zan_emc, ignore my previous advice. I'll be back later with better info
?? don't see any
do you have U V W terminals
Hugomati1, OK, I had thought about G92 offsets, but didn't think they were the problem
just L1 _L3
how many wires does the diagram show?
ok, do the wires come out of a junction box?
could this motor have been rewound for one voltage
open the junction box, and you should find the 1-9 wires
unless you opened it already and didn't find them :)
the box is open only three wires
and they go into the bowels of the motor?
I guess you're SOL then
trial and error
SWPadnos: Thanks, I am, YA, out of ideas :-(.
I forgot that you still need to be careful of inside corners
SWPadnos: I think the right way to go about this is to simply spot the mounting holes and then drill them manually on the drill press.
Join1 following error. Only thing I changed was the feedrate from 10 to 100, hummm...
Something about that last G2 didn't like it, but it wasn't really necessary (I think).
fix your config.
check the accels of your axes.. and what the stepgen max accel is
K`zan_emc: your machine should NEVER EVER EVER end up with that condition :)
unless you have a closed loop system and crash your spindle into the table.. :P
[20:19:55] <BigJohnT> http://imagebin.ca/view/4LKmSIG6.html
[20:20:10] <BigJohnT> http://imagebin.ca/view/QuJWVyOg.html
BigJohnT: what does that machine do?
it tests a running lawnmower engine
we test the starter switch, the stop switch, the stop time, the brake function, the static governor setting, the RPM all in less than 5 seconds
we built 3 of them at the same time
so if you make one mistake you made 3 :)
It is fun work most of the time
good night all
yahoo.. my lathe is finally in the basement...