did you splice single frames of hardcore porn into the reels like they did in fight club?
6 feet of 1"-5 & 6 feet of 3/4-6 acme threaded rod ordered for the x/y on the 24" x 48" cnc router project
oh, a router?
yeah.. pretty much. until i tweak the design and then i plan on reproducing the fram out of steel
i had enough of mini mill world :)
i got the bas for it dropped off today
it's sitting in the garage lookin pretty
table comes ina few days
then i gotta mount rails and the threaded rod on the base for x.. and start working on the gantry
why not buy a used series II bridgeport?
well.. i'm looking for one
but i still would liek to have the cnc router
they guy building me the frame works at a shop, he says they spent 30,000 on athier cnc router
he mention to his boss what i was doing here.. and he want's to chat ;)
sounds like a conversation where i make some money...
i'm glad i finally got off my a$$ and built that mini mill... if it wasn't for emc i would have never tried
it's opened up a lot of opprotunities for me here in dayton since people have been talking about it
do they really not use their router
they use it for making bar rails and such..
mouldings for cabinets
it's a cabinet shop
they were looking for a programmer for it a while back.. dude said if he knew i was capable of it i woulda had a sweet job payin like 20 bones an hr
but.. nobody really knew what i did at work thill i bought this project home...
you're a cnc programmer at work?
now i am at a different shop doing manual stuff
i built the mini mill to keep up on my cnc skills
in case i ever need to go back into it
but then... i got the hombrew bug
one thing i am stumped on...
what to use for the rails of the x axis
scraped box ways
it's gotta be at least 52" long and straight
and i need 2 of them
TWO scraped box ways
what to start from for the scraping though...
square tubes, angle iron
raw bar stock?
the more mass the more rigid
cast iron is the best
i was considering angle iron
but steel will work too
old car frame
and.. someone gave me the idea of fence posts
OLD BUS FRAME
but i don't know how straight they would be
or sturdy really
i peeked at my bedfram this morning as i was getting up. it's iron angle and it's plenty of material
are you like, planning on cutting anything even remotely accurately
you need mass, man
well for the most part i was thinking if i can do .010 that it would be amazing
you might be able to swing that
if i did .005 i would sh1t myself
not so much accuracy
i mean like, divots and stuff in the material
the wavyness in the cut
i figure i am going to be cutting wood... carpenters go by 1/6th at the most really
yeah if all my part's edges are within .010 -.015" of the programmed lines
i'm going to be happy
not edge tolerance
like, your edges are not going to be straight
the cutter is going to wander.
i suppose lapping it some
with like a power lap
i know i need to spend 30 grand to do anything worth selling to honda
might do some stuff
well like, i'm trying to think of ways
to get this thing accurate
but they don't buy stuff from a cnc router unless it's a front door sign
cnc wood router*
wonder if a round way is easier to deal with... start with somethinglike a 6ft peice of drill rod
can you use heavy gauge steel square tube
ok.. one idea i had was inverted aluminum extruded angle
and some skate bearings
round ways don't give you the rigidty
some smaller machines have used round ways in the past
the skate bearings would ride on the two angles ad 90 degrees
like, a pair of 4" round
stay away from bearings
just throw in a small crane in the order and you'll be set
well i guess bearings might work in this particular case
the skate bearings are tight
no it's not that
the interface between the rolling bearing and the way
for both accuracy and rigidity
picture an upsidedown v with bearings riding along each surface.. 90 degrees with all the force going down
*one of those on either side
how you position the bearings isn't the problem
it's that you have this teeny tiny contact surface
ok, i can see that
machine tools still use plain bearing surfaces
no rollers or any moving parts
like the ways on a bridgeport
some specialized machine tools use a roller way bearing
that is ultra ultra expensive
and most of those have hybrid plain/roller things going on
yes, like a bridgeport
the planer.. from 1913, that we have at work, it uses a rack n pinion gear to move the table
and it is suprisingly smooth
still has plain bearings
oh it does
they are v grooves on either side
that takes all the force
i left that part out when i jsut pictured it in my head
the rack and pinion is only there to propel the table, it's not accurate
so if i was going to do angle and the bearings.. i could put them on the inside of the angle instead of outside
doesn't matter where you put them
inverted vee ways work just as well as external
personally I would use a pair of cast iron tubes
file them down a bit
then scrape them in
with the carrages to match
I'd probably make a frame out of wood and pour concrete into it
i scoped out the cast iron at lowes, it is semi-straight... i even rolled it on the floor
can't scrape/file round cast iron
what i'm suggesting
is that for each "way"
you use two parallel square tubes
4" x 4"
and pour a base of concrete for it
let it settle for a month
go back and start scraping it in
that way you get very serious mass
and concrete is an excellent damping material
the only reason they don't use it in modern stuff is because it settles
yeah it tis, but i can't move that around in my garage without a forklift and i don't have anywhere to park one
you are planning on a 24x48 router?
and you want it to be light?
not light.. but not heavy as a truck
you're going to be talking over 2000 lbs for a solid router of that size
so far the base is made like a cabinet in a million dollar house.. out of 1" thick laminated plywood
it's heavy on it's own and it doesn't wobble
cabinets in million dollar houses don't cut anything
you got me there
we have a tiny router at school the students built
the base is 36x52
and even that thing is very very rigid
and very heavy
i plan on having a full 24 x 48 cutting area.. all the way to the edges
the design in my head anyways.. and we have done some sketches
i jsut haven't came up with a X way idea i thought would stick
other than your box ways idea
if your base is plywood
then just use angle iron
that may be the trick
do you have a surface plate
yeah.. a tiny one
and a fancy indicator and stand..
a fancy indicator?
yo know.. "jewled" as opposed to not
anyway if you get engineering blue
NOT dykem/marking blue
spread it on the surface plate
then drop your way on the plate
i found that in enco the other day
support the overhang
let it sit under its own gravity
persian blue or something?
but when you pull it off the high spots will be blue.
file those off.
then, once it starts to get kind of flat you'd have to switch to a scraper.
low spots nlue?
high spots blue
low spots not blue
oh because the ink is on th plate
i read a website about scraping once
but i never needed to do it
there's a roaming class
that I want to attend
just hasn't wound up in my area yet
we have an old carbide tipped thing i assume is a scraper at the shop
about 20" long?
has what looks like a blunt end on it?
if so, yep!
it's sorta half moon shaped and sharp
the kid was using it to clean out the tslots and i took it from him
those things cost like 90 bucks
i hid it in a cabinet
[02:10:41] <ohiopctechDOTcom> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=160111539378&ssPageName=ADME:B:WNA:US:12
<- small lathe
*it's ending soon
damn i always chek my email too late
i'm not a big fan of the small gear
the smal gear can help you build big gear...
lol not particularly
machines replicating themselves in a larger scale around the world till we have monster trucks n stuff?
or huge mining dump trucks...
larger mills and lathes were originally hand worked
i know i am being silly
i have seen the lathes and mills they use to build stuff like that
i once laid eyes upon a 35 foot bed lathe
and there is a 3 story mill at this shop i visit
presses sitting 10 feet in the floor because they are so tall...
we actually had some work done by a place that has a mill that the operator rides on the head of the machine in a basket
ima going to take some pics of the machine base and crash.. later folks!
[02:34:55] <ohiopctechDOTcom> http://www.industriamercato.it/eng/gantry-type-milling-machine/fresatrice-a-portale_multinorma_102037.html#
Q: Looking in the list archives & the docs, I didn't (yet) find any details about a preferred way to interface EMC2 to a spindle motor VFD. I am using an M5I20. Suggestions?
SWPadnos: is it possible to paralle two switching PS?
are you using 4 servo axes, or can you use the last "DAC" as a VFD output?
Unless they are designed for load sharing, generally not a good plan (IIRC).
Jymmmm, sure, as long as they're the same output voltage and you have good sized output caps on them (I think)
SWPadnos: I was planning on adding an extranl cap between teh power source and the drives
ok hw big 1F ?!
I would like to keep the 4th axis available for, well, a 4th axis. :)
Jymmmm, pretty big. I don't know exactly
just use the power supplies on different axes and share grounds
Chris_sub_1, well, a spindle is pretty slow, you could just use software PWM on an unused output pin
eric_u: share grounds??? Isn't that the problem to begin with?
no, it's fine
they are floating anyway
you have to take the ground off of one of the ps'
remove the ground from the case anywhere it may be touching
if they are ATX type***
a bench supply will haqve floating grounds. don't expect a PC supply to
no not PC PS
why would you want to do that?
but it is a switching PS
jymmmm, it depends on the swither design
in general you can't just parallel them
if there is an isolation stage, then you don't have to share grounds
he can just parallel them by using them for different drives
why not just use them for separate axis?
he's got 3 loads and 2 supplies
1 supply will power 2 of the loads, problem solved without driving the supply crazy
eric_u, well that is not paralleling them if the outputs are not connected together
what about putting a diode and a low value resistor in series at each PS and then parallel the output of that?
join with me in calling for sanity!
ds2, you will not get good load sharing like that unless all components are matched
in fact, one of the supplies will probably drive all three loads by itself
but there is no reason not to join grounds together, they are probably floating anyway
I've found a lot of those switchers are floating at 90v, which is annoying
sanity is something you'll never miss
at least that's what my mom used to tell me
oh my... they say there is pills for the condition you describe
How can pills help with load sharing? IRC is so confusing.
hey i found a decent webcam that i can focus... i can maybe do a worthwhile cnc vid now!
but no sound... everyone knows what a dremel sounds like right?
not sure, hum a few bars
hmmmmmmmmmmm hmmmmm hmmmmmmmmm hmm
[04:20:53] <Chris_sub_1> http://www.podshow.com/music/?artist_id=6521
this webcam is almost close to not taking crappy videos of the cnc. now that i have one that i can focus
dremel had me till he started signing
*if that's anyone in this room i apologize...
Dallur: are you around? any news?
alex_joni: wake up?
martin_lundstrom: I'm kinda here, sorry about yesterday, I wanted to work on emc but I had to work, linux box at a customers site went belly up and they couldn't fix it
Dallur: Come on you dont have to excuse yourself
Dallur: At least you have a paid job
Martin_Lundstrom: yeah, but it gets a bit hectic at times :)
Martin_Lundstrom: So I don't have much of a life ;)
If anyone have a tip for a small job, please let me know. (I have a registred company, so I can send invoice)
How did your plasma table work out?
though the plasma were too puny to cut what I needed to cut
so I run oxy-acetylene now.
anonimasu: do you have any pictures?
Dallur: So do I deare to ask you when you are going to make a cvs check in, maybe just make a new dir, so I can compile your still inofficial verion for test?
martin_lundstrom: the plan is to do it tonight
martin_lundstrom: as long as none of my customers mess up their servers
anonimasu: sorry.. was away
just got an info for alibre prof. for half the price ;)
thought you might be interested
[12:10:27] <alex_joni> http://sdm3.rm04.net/servlet/MailView?ms=ODUzNjgyS0&r=Njc2OTU5MTM3MgS2&j=Mjc5MTQ4NjAS1&mt=1
alex_joni: I'm way too busy to look into it
alex_joni: they need to fix their snap before I buy it ;)
anonimasu: kinda same here.. but I knew you were interested
alex_joni: maybe we should drop them a mail about it..
alex_joni: by chance it might even get there in a few versions ;)
I dont think we are the only ones with this stuff in mind..
anonimasu: I didn't like their 2D stuff
lots of better alternatives
though their snap is the real killer for me
heh.. got a new spot welding machine
question: I've apt360 installed, if I type the command: apt360 < example.apt | cat example.EMC1.tap I can see the gcode output generated by apt in the terminal
is there a way to use the axis filter feature in order to open an apt file and make Axis get the gcode in the output file?
could I do a script in paython for example that runs a command from command line?
AXIS expects filters that take the input filename on the command line, and generate gcode on stdout
if apt360 does that, it will work directly as a filter
no, apt generates a file that contains the gcode
that's too bad, then I think you'll have to make a script that cats the generated file
if I run the command: apt360 < example.apt
I cannot see the gcode
is difficult to do a script like that?
what file does it generate?
if I run apt360 < example.apt, it generates a file example.EMC1.tap
how does apt360 determine the output filename?
how does it know the file is named example?
form the input
example.apt -> example.EMC1.tap
I think you might be mistaken about how it's run then
if it's on stdin (<) it doesn't know the filename
I had run that command this morning: apt360 < example.apt | cat example.EMC1.tap and I've seen the gcode in the terminal...I don't know much about the command I used, I was tolded to use it...
a person told me "use this command" and it worked...
you typed it with the | just like that?
he called it "pipe" if I remember...
it is very surprising that your command works
ok, that command doesn't make a lot of sense - we need more information on how apt360 is run
it is not surprising that you got bad advice online
if the file "example.EMC1.tap" already existed when you issued that command (from a prior run of apt360) then it makes a little more sense
well, the problem is that I don't have apt installed here so I cannot try in this moment
but I can take a look in the wiki, there is a page on the apt
well the answer to your question is yes - you should be able to run apt360 as an AXIS filter
but it gives no gcode output...it generates a file...I don't know how he knows the output filename, but it generates a file.tap, I think it uses a postprocessor for EMC and generates that file
in the wiki it does not use the "<"...just apt /apt360/examples/squareTJP-1.apt
andt the output is squareTJP-1.tap
that makes much more sense
supposing it is the correct way (maybe there was a file previously created with the same name so the command worked), how con I make a python script that executes a command from the command line?
[14:40:18] <jepler> http://pastebin.ca/477188
this shell script will copy the given file to a temporary location and run apt360 on it. If apt360 is successful, it "cat"s the contents of the output file. When it's done it removes the temporary files.
jepler: exit $?
won't bash exit with the exit code of the last command (apt360 && cat) ?
many many thanks, I will try it
jepler: hm I don't know
sh -c 'false&&true'; echo $?
ok you'd think I'd know that
and...I name this scripts for example apt-reader.sh and in the Filter section of the .ini add that .apt files should be opened with that script, right?
ok, I will try
you would also need to make the script executable (chmod +x apt-reader.sh) and place it in a directory on your $PATH or in the emc bin directory
hmmm.. I have a 1ghz dual processor machine I can play with :)
I would like to build I little hal system I have to use at university...simple things, but I would like to give it a little graphical interface using pyVCP
I would likt to use it stand-alone without EMC2, I've seen the dro example in the wiki
it says taht the hal command to launch pyVCP is: loadusr -Wn fred pyvcp -c fred pyvcp-dro.xml
man halcmd, look for loadusr
fred is the name to give the component .. the example might be better if it used a more meaningful name, such as "dro"
(and I think it is also the initial part of the HAL pin names)
"loadusr -Wn <name>" means "wait for the component to become ready; it is called <name>". "pyvcp -c <name>" means "use <name> for the component name and pin name prefix"
aaah, ok, so it creates a new hal components that has all the pins correponding to the widgets
anyone ever heard of featurecam? i found a demo of it in my garage that i never opened... they rep for dayton was tryin to sell it to me but i didn't feel like learning a whole other cam prog.
just wondered if anyone used it before*
I used feature cam one day....
featurecam sounds like progressquest
it worked...but after I switched to edge cam and I've not used it anymore
they got a whole fancy brochure and real cd's.. they must be making some money
oh see, I've it still installed in mt PC
i wonder if the demo will let me save files?
i may just check it out
I don't know...I used the complete version
maybe you cannot generate the gcode...so the demo is completely useless...
*i'll check into it someday...
in the students version of edgecam you cannot generate the gcode
i found this one demo of a program, it wouldn't let you save the g-code you created but you could copy and paste it to a text file.. lol... ;)
then in the new version of it.. they had the copy n paste fixed.. it was some sort of jpg to cnc prog
ah, third question...I'm remembering all the questions... :) I've read that modal codes M101-M199 are user defined codes...
but how to define them??
i've not came across that info yet either
i think it's up to whoever is writing the control software?
I've for example a fourth relè (like the spinlde, flood and mist), how could I define two Mcodes to use that relè in my gcode?
map it to the pin with hal
so I should change the EMC source files in order to get this code recognized by interpreter
i'm not so sure about the details i have not gone that far into the config yet, you should def. get asecond opinion on that
Look at the m101 example in the nc_files directory
[15:13:20] <skunkworks> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/nc_files/
well, I didn't understood nothing...:)
where do I put that script?
and how does EMC2 interpreter know that I've defined a new code?
sweet! you can have g-code run programs on your box.. say when the program is done it could load up the inspection database software
xemet, emc doesn't know - it just tries to run an external program whenever it encounters M100-M199 codes
Is uspect it gives you an error if the program isn't found
I suspect ...
would you put the file named "m101" in the ini folder maybe?
or "m102" ?
I think it needs to be in the nc_files directory (as specified in the ini)
and I think it needs to have a capital M
yeah linux is case sensitive
so let me understand....m101 is the name of the script...
look in nc_files, and see what the sample M101 and M102 are :)
when the intepreter encounter M101 it search in the nc-files directory for a program named M101
...now...what kind of program could it be? the example is a bash script
kinda like a sub-program
anything that's executable
it could be any type of executable?
it could even be more g-codes in the file
yes, and it gets two numbers, from the P and Q words in the G-code
ni, it can't
SWPadnos , no?
the file is not interpreted by EMC, it's exec'ed by the system
so if I want just to get a hal pin high with M101 and get it down with M102 (think of activating a relè) I've to wirte a script that does that
and EMC doesn't even get any output from the program, it just does its thing, then EMC continues on
for example a python script that use hal library
yes, and the script could be as simple as "halcmd sets my.sig.name off"
with the appropriate #!/bin/nash at the top, of course
I can give halcmd commands with python?
there is already a pin for spindle on/off set-up in the default config i believe xemet, you should jsut need to cange that to your relay pin
err - bash :)
no, that would be a bash script
ohiopctechDOTcom, that depends on whether the relay is connected to the spindle or not :)
so just I create a script that execute a halcmd command
SWPadnos, of course...
in fact, the script could execute "halcmd -f <some_file.hal>", to do a lot of HAL stuff :)
now maybe it is the time to learn the bash language
you could set M101... to delete the g-code file after the program is complete.. like mission impossible. "this g-code will self destruct in 30 seconds..."
remember that you don't get any feedback to EMC - I'm not sure that EMC even knows that the program executed successfully (ie, it could have had an error but EMC may not know)
is there a way to read message into emc from that script with a pipe maybe?
or.. you could jsut have the script pop opena terminal window so you can see what's going on?
you can have a terminal window, I think, but you can't pipe anything back to EMC
even if I don't know bash language....let me understand, I can just create a text file named M101 that has on the top #!/bin/bash and for example only one line halcmd -f example.hal ?
as long as you chmod +x, yes
make it executable and put it in the nc_files directory and it will work?
that's a very good thing :)
look at the M101 and M102.c that are there already for examples
they also show how to get the P and Q values
[15:28:05] <SWPadnos> http://cvs.linuxcnc.org/cvs/emc2/nc_files/M101?rev=1.3
[15:28:09] <SWPadnos> http://cvs.linuxcnc.org/cvs/emc2/nc_files/M102.c?rev=1.3
I'm more interested in doing hal things with Mcodes created by me...so for the moment the bash script with halcmd is all what I need
I've read the two examples they are good
yep. simple and to the point :)
Interesting, that part scan finally finished, and I was going to see if I could speed up the display update, which obviously lags the machine by over a second, and found some errors on the shell screen when it had stopped
what kind of errors?
* CMS Status = -1 (CMS_MISC_ERROR: A miscellaneous error occured.)
* Recent errors repeated:
SHMEM: Can't take semaphore
* BufferLine: B emcStatus SHMEM localhost 10240 0 0 2 16 1002 TCP=5005 xdr
* ProcessLine: P emc emcStatus LOCAL localhost W 0 1.0 0 0
* error_type = 5 (NML_INTERNAL_CMS_ERROR)
what does that tell me
I need to leave for a bit, so email me, firstname.lastname@example.org with your comments please.
emc uses "shared memory" to communicate between certain parts of the program, including between the user interface and the part called "task". Programs can only read or modify one piece of information at a time, so for instance if you are updating the X, Y and Z coordinates of the machine from "task" at the same time as the user interface is reading the X, Y, and Z coordinates, you might have a situation where the user interface sees X1, Y0, Z0 if "task
the "semaphore" is the way that emc prevents this from happening. It is a way for the user interface and task to negotiate access to the shared memory, so that it only seen when it is in a consistent state (either X0, Y0, Z0 or X1, Y1, Z1)
in this case, one program (either task or user interface) waited for the semaphore too long. If this happens consistently, it can indicate a software bug (because one part of the program is not releasing the semaphore)
on the other hand, because they are not realtime processes, it is not guaranteed that the semaphore deadline will always be met, so it is possible for this error to be printed even if there is not a software bug.
in that case, emc would continue working properly
oh I love showing off how much I know :-P
I could tell that it was a shared memory problem, but what it means (to the user), I have no idea
now (thanks to you), I know that it either means there's a problem (bug) or it doesn't :)
Looks like one of my lines got cut off. It should have read: ... you might have a situation where the user interface sees X1, Y0, Z0 if "task" is in the process of updating the position from X0, Y0, Z0 to X1, Y1, Z1 (even though the machine was never actually "at" X1,Y0,Z0)
I used to, but then it got small enough folk had to look with a microscope.
then you can tell them all about the right kind of microscope to use :)
That's about all that's left to tell.
cradek... you have right... with 220.127.116.11 emc2 works great... now i have regular 18.104.22.168, 22.214.171.124 with rtai3.5 for realtime and 126.96.36.199+rtai3.5 for emc2 :)
Nice work zelva.
188.8.131.52 is just a crazy number imho
heh... I wonder if that may have had anything to do with the tail -f I had running on the output file?
i cant agree more :D
gene: no it shouldn't
gene: can you describe under what circustances you got that?
you mean the file that emc is writing probe data to? no, that would not be the cause.
ok now i have to burn DVD with that and try it :)
Aat this point, thats what was left on the screen after I'd exited, 2.1.5
jepler: it's emcStatus.. so that's pretty widely used
note that by default "tail" waits 1 second before checking the file contents, so that would introduce an appearance of lag in showing probe points..
Next, do we have a program that can visualize the data in the put file without needing a lot of massaging to do it?
put s/b output
that lag was about 300 probes at a time in practice
file system granularity I'd assume, no biggie
yes that makes sense
300 probe lag doesn't come from 1 second tail lag
but from flushing things to disk
The reason I asked about visualization is that one end of the piece has a just barely visible flexure and I accidently pushed on it once
I don't know of a program that is specifically for showing the probed points
gene: any pictures? :)
So I need to know if I have to go through it logitudinally and adjust the z values a thou or so.
Not yet, of the probe you meam?
probe, what you scanned, resultant code :)
jepler: maybe someday you can import a cad drawing with fixtures drawn.. for collission avoidance..
that's not a bad idea actually
Since much of it lengthwise is a straight line, anomolous data should be easy to filter.
I can put that on pastbin I suppose.
code is at <ttp://www.pastebin.ca/477338>
output .txt file too big for pastebin.ca
the last few lines are in this format:
3.530000 0.460000 -0.080866
3.530000 0.470000 -0.082929
3.530000 0.480000 -0.084721
3.530000 0.490000 -0.087118
3.530000 0.500000 -0.090289
with the first value obviously the x, next is t etc
next isy etc
I thought, while waatching the code run, that I needed to do a mirror od o100 because it was 4 or 5 seconds doing the retrace move each time
Humm, do we have gnuplot installed by default?
no, but we can install it with apt
or adept I'd assume
no only apt
lemme see if I can find it
no, apt can't find it
yes it can =(
what repo then
i'm on a windows box
apt-cache search gnuplot
i've installed gnuplot via apt a bazillion times
and that is a perfectly accurate number.
you may have, and so have I on fedora boxes, but:
gene@shop:~/emc2/nc_files$ sudo apt-cache search gnuplot
bonnie++ - Hard drive bottleneck testing benchmark suite.
calc - An advanced calculator and mathematical tool for Emacs.
devscripts - Scripts to make the life of a Debian Package maintainer easier
now I ask you, what the hell has all that got to do with it?
[16:32:43] <toastydeath> http://packages.ubuntulinux.org/edgy/math/gnuplot
didn't change a thing
i am blaming you in some capacity.
Just because I have broad shoulders (and gut etc)
how do I enable the universe repo, thats where its at?
i always edit /etc/apt/sources.list
and include the thing manually
BUT APPARENTLY YOU'RE NOT SUPPOSED TO EDIT ANYTHING BY HAND ANYMORE
I'll give that a shot, bbl
adept is pulling it in now, thanks
BUT, when push comes to shove, the cli rules...:)
okay so does anybody understand twitter
wth is adept?
a better gui front end for apt, possibly a little smarter but not sure abouyt that
its the default package manager for kubuntu-6.06
wth is kubuntu?
Alex, I wasn't sticking out a leg for you to pull, really :)
* alex_joni stops pulling
anyone using rtai-lab on debian ?
now I gotta go study on gnuplot, I had it put the docs in too.
zelva: not likely
only ubuntu ?
rtai-lab in general
aha i see :)
Humm, gnuplot only does 2d, grrrr
I think it does 3D too
it does 3d too :)
it does 4d if you enable it at compile time
(i am making things up again)
Gnuplot supports many types of plots in either 2D and 3D.
So apparently I'm going to have to write a filter as it barfs on line one of the output.txt file
[16:55:50] <alex_joni> http://gnuplot.sourceforge.net/demo_4.2/surface1.html
gene: of course you need a file for gnuplot
wow that's some fancy graphs
how to interpret the data it gets from you
gnuplot> -0.000000 -0.500000 -0.171521
"floor-plate-scan.txt", line 1: invalid command
lol srsly dude
you can't just shovel data into it
[16:56:30] <alex_joni> http://gnuplot.sourceforge.net/demo_4.2/surface1.8.gnu
that's probably what you want
prop charsets grr
Is that part of the insstall?
now what, feed it to gnuplot as the first file?
to be honest i don't know
that didn't work error in line 7
i never got gnuplot to do what i want
gene@shop:~/emc2/nc_files$ gnuplot surface1.8.gnu floor-plate-scan.txt
"surface1.8.gnu", line 7: undefined variable: offset
I use it under an alias, called amplot, part of the amanda package, works just fine
oh i have no doubt it works fine
i'm just saying i'm incompetent
you have company, now row your end of the boat :)
same here, out of coffee
gene: check your installed gnuplot for scripts
might be a diff. version
I wonder if anybody read my message about tests with Mach3 and EMC2 using G41
[17:04:30] <danielbr> http://article.gmane.org/gmane.linux.distributions.emc.devel/443
I think you have to go back to chris and rays posts. It would be a 'substantial' re-write of the cutter comp code.
I read that skunkworks, thanks
i've used plenty of controllers that flipped out when you gave it a curve it couldn't trace
yes, but when you have two G1 block?
from what i understand you are saying
when you give it two g1 blocks that are at an angle
it flips out
this is a mach3 plot http://imagebin.org/8507
[17:08:54] <danielbr> http://imagebin.org/8508
[17:08:59] <danielbr> http://imagebin.org/8509
what are you trying to show
this mach3 g41 comp
when you give it a cutter radius larger than the programmed radius?
that's natural behavior for a lot of controllers
it's letting you know you did something retarded
see image 8509
i'm looking at it
yeah, but is that really retarded?
if you want to cut a square with a arabity sized cutter?
[17:10:57] <danielbr> http://www.bmed.mcgill.ca/dept_resources/machine_shop/machinery/mill/Fagor/8040M_USER.pdf
this a Fagor doc see pag 300-312
then don't use cutter comp?
or lie to the machine.
toastydeath: how do you get around to doing the g2/g3's at the corners?
the purpose of cutter comp is to give you the same result regardless of the cutter
not to give you a different result based on the cutter diameter
wich is why several commercial controllers freak out when you do this
toastydeath: well, then cutter comp is crap..
not from a money making standpoint
or, it's only good for compenasting down to 0.xx..
you need to use g2/3
most cnc controller i know work like mach3 or fagor example
toastydeath: I guess we should all just order out own copies of mastercam..
i dunno about you but I write most of my code by hand
i thought that's what we're talking about here
this is from that fagor doc http://imagebin.org/8506
what is "most cnc controllers"
do you have a list of specific models
because i've seen it freak on fanuc controllers and a haas control
hm, heidenhain craps out I think..
though there's a chamfer command you can use ;) and just throw in your cutter diameter from the plc..
you could write a fillet macro on fanuc and haas
all italian cnc woodworking controllers we work accept g1 intersections and round corners
what is your application
that you need to use intersections
instead of defined radii
what is your application
what part are you producing
that you need to use an intersection
I think this is more like a being lazy thing..
but if you read that post you can see my problem
for me I know it is..
i don't see a problem
i see you trying to get around fully defining your part
it make this easy
and to be serious, i think you should not use cutter comp
if you really want to change the part based on cutter size
we never had cutter comp - our machines where too old. Always offset in our cad software.
or by hand
I don't think I have ever used it.
if you have some good cam is OK
here is why some (expensive) controllers freak out about this radius thing
you are cutting a 12000 dollar block of tool steel.
But you can see the issue. emc would have to look an infinate number of moves ahead to figure out the new cutter path.
your cutter is too big for the radius.
you want that thing to alarm out before you start cutting.
mach2 have g41 like emc2 work after some demand art change the comp code http://www.cnczone.com/forums/showthread.php?t=14537
now with mach3 g41 is useful!
but i can not do what i need with mach3
for other good features emc2 have and mach3 do not
toastydeath: that's a good point too
it's there to save your rear end when you're making a part that matters
and i can fully understand your frustration if you're just doing a couple wood parts
and just want to hack out a fast toolpath to get the job done
but i don't believe changing a controller to assist the fast hack toolpath way of coding is the right thing to do
[17:26:08] <danielbr> http://imagebin.org/8510
also maybe you should look into an apt postprocessor
apt is a geometric machining language
you define lines and stuff and tell the machine to do things, "go to this point, on the left side of this line"
i don't have any at the moment, sorryt
has a link somewere.
they are working in a visual apt clone/frontend
that sounds cool
i have never been able to find any APT stuff
for my aplication the better way use tool diameter comp like Fagor or Mach3
is anyone can change the comp code i cam pay for this work
ok - so I change the "ati" to "vesa" correct?
seemed to work.. the ati driver was causing overruns
When using a SSR to control a load, what side to you connect> hot or neutral?
you should never ever switch neutral
do you mean which side do you switch? normaly it is the hot side. Less chance of accidents.
let me rephrase that: you should never ever *EVER* switch neutral
unless you're also switching hot
03petev 07TRUNK * 10emc2/src/hal/components/at_pid.c:
-Modified PID calculations to calculate PID values appropriate for EMCs
PID equation. The PID equation they were intended for is slightly different.
-Improved accuracy of period calculation.
if your radio has a loud buzz, try turning the plug over in the outlet
cradek: Since you mentioned it.... on polarized plugs, which is hot, the wider one?
let me rephrase that - less chance of killing someone.
ask the internet, I'm not sure
I searched for "polarized plug hot" and it looks like the answer is behind the first link
extern int usrmotInit(char *name);
the source code for emc is hard to follow
it should have been further split into subfolders
you are absolutely correct :)
~/emc2-src/src$ find -type d | wc -l
yes, a few more folders are clearly what is called for
or directories even
subdirectured ?? :)
where's the go code ?
in the rs274 directory
why doyou need rtapi in linux ?
RTAPI is a thin layer over the underlying realtime sytem (which can be RTAI or RTLinux)
it allows the rest of the upper layer code to be independent of the lower level RT system
so its extra processing overhead?
priority based should work aswell ?
who do i submit axis bugs/feature requests to?
renesis: there is a tracker on sourceforge which can be used for that purpose.
renesis, not really - if there are direct mappings tfrom RTAPI to RTAI (for example) then the RTAPI call can be a define
gringos, no, the normal linux schedulers (even Ingo's "real time patches" aren't real-time enough for machine control, and aren't hard RT
Ingo's patches may be getting there now though, I haven't looked in a while
but don't let our naysaying stop you from writing an rtapi layer above whatever technology you think might work instead of rtai or rtlinux -- if it works we would welcome the contribution.
i got nothin
do have some feature requests though
4 corner touch with a probe to know your peice levelness
or multipoint for bumpy surfaces
4 axis would be great for that
5 better though
where can a person even get a 5 axis driver board ?
having problems finding the hardware io code ?
no microstepping ?
the step waveform looks the same no matter whether the stepper controller uses microstepping, half-stepping, or full-stepping
the distance moved per step changes. that distance is set in the inifile, item [AXIS_#]INPUT_SCALE
Dallur: Are you around?
Dallur: Now I am ;)
Dallur: Any news?
a-l-p-h-a2 is now known as a-l-p-h-a
[22:52:32] <ohiopctechDOTcom> http://www.youtube.com/watch?v=IBW01i_fIwA
<- new cnc mini mill vid