03jmkasunich 07TRUNK * 10emc2/docs/man/man9/encoder.9: add man page
03jmkasunich 07v2_1_branch * 10emc2/debian/emc2.files.in: backport: added man page for encoder counter
03jmkasunich 07v2_1_branch * 10emc2/docs/man/man9/encoder.9: backport: added man page for encoder counter
fenn: what is the machinerys-handbook?
it's all the standard threads and stuff, common formulas and tables, feeds and speeds
different types of alloys
all the sort of stuff you'd need to rebuild civilization from scratch, essentially
and have it interoperate nicely
what format is it inside the tar file you have?
best table for thread fits in print
Well, not quite civilization. There is no chapter on wine making as far as I recall. :-)
i can untar if you'd like
lerman Chapter 42
do you mind if I grab the .tar file? I can rate limit if you want
ch 41 is beer
43 is making your own stile
ejholmgren: not at all
well there is a chapter on quenching in brine.
that's why it's there
already doing basic rate limiting with wondershaper
just don't want to max out your upload for the time it will take to get 380M
umm... morally wrong. http://www.demonoid.com/files/details/540804/2389296/ http://www.mininova.org/tor/493894
but is it wrong if you bought the hardcover, and just want electronic searching/
Skullworks, technically it is wrong. But morally, I don't think so.
* Skullworks has a stack of 4 editions of machinery's handbooks.
it's like George Lucus, release VHS, get then an enhanced VHS, special edition, directors cut, and soon DVDs, or HDDVDs, or whatever...
well, consider that they also sell the C-ROM version, separately (or in a pacakge with the book) for extra money
CD-ROM, that is
I think if you own one... you should be allowed to own them all
so yes, they seem to think that they're separate products, and considering that they own the copyright ...
but I never lost the books.... due to a HDD or other failure...
and you still won't :)
i think it's morally wrong to charge money for electronic standards documents
jmkasunich: Cap bank jumps the voltage from 25.2 to 33.6!
I agree, but you can't exactly claim the moral high ground when you flaunt laws you don't like, while talking about those you do like ...
and funny as it may sound, a paper bookmark is faster than a search function.
and $100k's is beyond anything i could ever hope to pay for
(not that ISO policy is a law ...)
nobody enforces laws I like.
"One has a moral responsibility to disobey unjust laws." -- Dr. Martin Luther King, Jr.
Never thought I would see the day I would quote that man...
its harder to get into some bars than it is to enter the US.
try it as a foreign national one day ...
Pass, thank you :)
yeah or maybe just change your name to mohammed
especially with non-white skin
Wonder why that might be?
* fenn wonders how to derail this conversation
Don't signal and then move the switch :)
SWPadnos: you consider yourself self-employed right? how did you get started in all that?
example - extortion is not illeagal if done by a DA, because they call it "Plea barganing".
* Skullworks decides its time to pack up his soap box...
* fenn decides its time to go sit on the heater vent
Skullworks: Won't work, most people are too brainwashed anymore to have a concept of "illegal".
fenn: I co-founded a business while still in college, then left it to consult on my own when I got tired of dealing with my old business partners
Wonder what folks use to draw stuff and generate G code?
well more like, "Its anyone elses fault but mine."
though I'm not self-employed. I own all the shares of the corporation that employs me :)
Skullworks: I see you understand how things work now.
SWPLinux: That seems interesting...
all too well
Skullworks: There are still a lot of people who understand it, but for some reason you don't seem to find them in big cities (usually).
K'zan For alot of stuff I use a dinosaur version of BOBCAD V12 Gold which runs in DOS
Skullworks: I figure I should start thinking about that even though actual CNC is at least 4 months away at this point.
I have yet to attempt it under linux as it includes its own VGA driver.
Skullworks: Wine (not that I would bet on it working, but...)
But I bought it at the Westec Tool Show in Los Angeles circa 1990-91
yeah - wine - this is pure DOS - and it dosent love windows either.
so not sure what would happen under wine
Try it, worst it can do is fail to work...
BTW I used it back then on a Compaq Deskpro 286N (pumped up to 13mb ram) with a genuine Intel 80287-10
LOL, what would you really use more than 64k for? ;-)
Skullworks: what vga chipsets does it support? I'm lucky enough to have VESA drivers for my dos autocad
I hope they did delays right, if not it could be a real problem on modern machines
its a generic 640x480
have not found anything it won't work on.
basic 16 color vga will work right in all emulators
cradek: my autocad works fine with whatever drivers I have (when I can keep my lunch down long enough to run it :-).
right - its only 16 color
I'm perverse enough to kind of like using autocad 12
well - this version will import R12 DXF - just don't have any polylines...
otherwise it follows the path to the end, then comes back down the other side...
for 2d work I have not used anything faster.
due to the hotkey setup of the dos version
I have all the latest Win versions... great for 3d but way to slow and cumbersum for simple 2 1/2D work
Criss - are you the one using the pluto PWM driver on a lathe?
dinner - bbl
03cradek 07v2_1_branch * 10emc2/debian/control.in: needed for lyx to pdf
03cradek 07v2_1_branch * 10emc2/configs/stepper/standard_pinout.hal: examples to answer the most common config questions
03cradek 07TRUNK * 10emc2/configs/stepper/standard_pinout.hal: examples to answer the most common config questions
03jepler 07v2_1_branch * 10emc2/debian/ (configure rules.in): this automatically re-runs debian/configure if it is needed
03swpadnos 07TRUNK * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: Add beginning of a Stepper configuration quick-start guide
03swpadnos 07v2_1_branch * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: backport: Add beginning of a Stepper configuration quick-start guide
03swpadnos 07v2_1_branch * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: Oops - wrong version before
03swpadnos 07TRUNK * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: Oops - wrong version before
03cradek 07v2_1_branch * 10emc2/docs/src/Submakefile: new quickstart pdfs
Hmmmm is the O code behavior RS274 or EMC specific?
who's behavior does it follow?
Okuma uses someting simular - and claimes to be rs274d
EMC2 uses its own bastard implimentation
since it is not really part of the RS274D spec
Hmmm.... trying to figure out why I can't have an Oxxxx line by itself
03jmkasunich 07TRUNK * 10emc2/docs/man/man9/ (encoder.9 pwmgen.9 stepgen.9): fix pin naming conventions in manpage
must be a CALL oxxxx
the main program will not have an Oxxxx
is there a link for Okuma style G code?
not that I know of
HAAS/Fanuc wants a Oxxxx in the main program :/
I been programing for OSP5000, 5020 mill controls since 1990
want to put together a perl converter for HAAS
Okuma allows the main to have a Oxxxx
03jmkasunich 07v2_1_branch * 10emc2/docs/man/man9/ (encoder.9 pwmgen.9 stepgen.9): backport naming convention fixes
not questioning the correctness, just trying to figure out what to do with my existing code =)
sounds like the "correct" thing is to drop the O code unless subroutines are detected?
but for subs instaed of a M98pxxxx you would use a CALL Oxxx
and in place of a m99 it would be RTS
but thats OKUMA
M99 is what I learned for for HAAS
I run those daily as well
cool, how different is the HAAS flavor from EMC?
there are several specific G codes theat EMC@ won't have
G84 is floating only for now
G73 does not exist
trying to remember about G12/G13
HAAS made up a good package
stole the best of both FANUC and Yasnaq
'cept it makes it hard to goto another control when one has been spoiled by HAAS ;)
well... to be honest I like my Fanuc 18 better... but not by much
the only fanuc I have access to is a 0T and it is painful to setup compared to a Haas SL series
OK - thats my new mission
I'll do a haas comparrison
post it to the wiki
now - your looking for mainly turning code?
or milling too?
the local JC has all Haas machines and Ibeen doing their program
I work @ mincomfg.com - offically just a green button pusher.
that's what a lot of people in the program is training to be
but I am a trainer and setup lead, bumped down when they combined the weekend shift back into the weekday crew.
been programing CNC's since Bandit controls circa 1983
sounds like one of the instructors there
but is the EMC G code mostly modeled after the Okuma stuff?
yeah well - got bumped to a lower title, but kept the pay grade...:}
pay is good
Weekend lead has 15yrs with the company - so he kept the title
pay is OK
manufacturing jobs in the US is not a good carrer choice anymore.
I even dropped the trade and went into computers from 99-02
much better money and cleaner work
but that died off to - all that got outsourced overseas.
I work where I work because the company has process patents that make ther methods profitable
where are they/you?
the chinese can try to make the same part, and scrap 4 out of 5 for quality issues - while our process has much better yeilds
Colorado Springs, CO USA
so there are still manufacturing scattered all over
yes - but nothing like it was 20 years ago
maybe 5% of the jobs remain
as a Tool & Die man I used to make the Dies that form Craftsman screwdrivers.
now I make fuser rollers for HP, Canon, Ricoh, Panasonic, and Lexmark
and thats going to die out soon to likely
the shift now is towards full color hi-speed copiers and laser printers
because a B&W image printed with color appears sharper to the human eye.
so - I'm kinda planing my semi-retirement
build up a few CNC machines for my private shop
and do custom prototype work part time.
big jobs I just sub out to the best equiped job shop
wish they still have a T&D program at the local JC
they dropped all machining and most engineering for our JC
its a shame
what's JC stand for?
Pikes Peak Community College
I don't know what is still offered by UCCS
but I doubt its much good
good night all
soon the next excuse will be there isn't anyone left that knows how to do it
* Skullworks wonders how he got up on this soapbox again...?
Because you can see what the morons that killed it can'd...
no use in taking on students, if there will be no jobs for them....
the machinist training program here was seriously supported by local shops to provide a skilled workforce
most those shops have closed up
when you can fax a print to malaysia - and get the die FED-x in 30days at half the local cost...
'cept right now you have engineering students that can't do anything practical
this is true
there is only so much you can do on a ProE station w/o doing some building
the one regret I have was when I was in college they were just starting to use the LED calculators
so they never pushed the slide rule.
I never really got a chance to learn the "slipstick"
at the least they should know how to do it by hand and even that is being dropped
take a pick axe to the HDD
Our prior head engineer would only use Proe
now he's gone, and what a pain it is
Even Catia isn't as retro
well I'm done - stuff me in my box and call it a night - 3am will be here too soon.
chaos as usual
Jsut trying to get the garage to one layer deep so I can setup the machine.
It's been rough.... every time I turn around murphy has been ahrd at work.
fenn: drop me a message when you wake up
What is the status of any 3d path generating with a linux application, is there any good?
I have seen some stuff made in blender
Martin_Lundstrom: there's one commercial program that does it..
Martin_Lundstrom: though as for opensource/free stuff I have no idea
I can see some stuff at http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?Cam
"Chris Radek's TrueTypeTracer" your famous ;)
Anyone have used cam.occ? http://cam.occ.googlepages.com/screenshots
[12:37:57] <anonimasu> http://gcam.js.cx/index.php/Main_Page
how goes with the prototype anonimasu ?
Dallur: not good
Dallur: trying to find out how to get vtk to create objects without making them actors..
im looking at how to get opengl and python working togther :)
hi all, what's up?
and pondering about simulation
of a matchstick model ?
I'm logging. I don't understand 'bookamrak', anonimasu. Try /msg logger_emc help
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-01-22.txt
anonimasu: did you try anything in opengl?
awallin: still trying to make a window that work with pyopengl
i'll code a matchstick class afterwards..
we need something to get us form a matchstick matrix (possibly with irregular spacings) to a triangulation (traingels are easy in opengl)
seems like making a window is hard with pyopengl
anonimasu: you rang?
fenn: yeah, but I've forgotten what I had in mind
im trying to make a glut window.. but it dosent seem like it can render stuff.
there is a tkinter widget in python 2.2..
i didnt get anything to work in vtk
i gave up pretty early though
with the enthought installation on winXP, a few vtk examples did work
and on my old computer I had some pyopengl examples working - but I can't find them now...
was trying to display the array as a surface or a contour or with the dataset mapper but it didnt work mostly because i didnt know what i was doing
I find python programmers the laziest people on earth..
damn bastards too lazy to print proper docs.
same goes for pyopengl
its vtk that's missing docs
kitware wants you to buy their books!
i hate books
why is all api's crap.
awallin: "marching cubes" is the way to transform from matchsticks to triangles.
(Actually voxels to triangles, but matchsticks are just stacks of voxels.)
I'm not sure that matchsticks == stack of voxels
Other than details of resolution, in what way are they not.
in the matchstick model, I would draw the surface triangles between the tips of the sticks
that's a different surface from just replacing each stick with a "bar" of a certain area
Are you considering the sticks to just be points?
well, if we use native opengl..
we will have face backface culling and stuff automagically
Then just go down the rows one at a time, making triangles connected to their neighbors. I think openGL calls them 'strips'.
that's a triangle strip..
[14:19:51] <anonimasu> http://jdobry.webpark.cz/opengl/opengl_maximum_performance.html
At any rate, converting matchsticks to triangles (or quads or whatever) is pretty simple.
sticks are easier to make since you dont have to care about capping the end..
though for undercuts, we would use triangle strips for the walls of the sticks..
but, how that would work in practice is harder :)
Undercuts get you back to dexels. I did find a paper on converting dexels to triangles; but it looked like it would have to be a batch operation rather than an incremental one.
did anyone find a paper with the math worked out for how to cut the sticks with a certain shape cutter and a give G0/1/2/3 move?
I would characterize it as 'algorithm' rather than math; but yes there are a bunch of papers related to that. Also the patent I found is a pretty good reference.
so we have no problems then? :) somebody just needs to write it all up and test it...
I'd love to, but there is too much already on my plate.
anonimasu: just read that performance paper. Good info. Thanks.
is there any low prised 5 axis g code tools around exept rainea?
I bet there aren't. Do you have a 5-axis machine?
I assume you are asking for tools to generate g code; not interpret it.
the code is not for my machine
EMC does (in principle) interpret g code.
rainea is too basic
but not generate
correct. EMC has no code generation tools.
(Not exactly true. There are some python scripts -- more for demo than anything real.
anyone tried rainea for 5 axis toolpath generating?
anonimasu: how did you install pyopengl on winxp?
does anyone want to buy two dogs?
I'll sell them to you real cheap ;)
Are they a matched pair?
can't get anything done down in the shop anymore ... efing puppies
no ... different stepping codes
or will I need to add circuitry to bias them?
anyone know the wiki page listing cam and cad software?
I've got an off-topic question, but I want to ask you guys since this seems to be the most likely place to find this knowledge on IRC
I have a device with an ATTiny26L chip that I need to reprogram. I'm in the states, and I want to find a nice cheap AVR-ISP programmer like this: http://www.avrfreaks.net/index.php?module=Freaks%20Tools&func=viewItem&item_id=744
Any suggestions, or pointers on a more appropriate place to ask?
tomp_: thanks much
PWizard join microcontrollers
tomp_: I added your dial widget to pyvcp yesterday
thanks, i have tooltips now, but dont care much for 'em
when mouse is over a widget, a text from the xml pops up
right, that could be useful ?
<ttiptext>this is the button you want</ttiptext>
yeah, if you pastebin the code I can look at it tonite
doing it now
it will be a few hours before I have time to look at it
it needs 2 things, a timeout, and ability to change the text, and enable/disable (*3 things... spanish inquisition)
i just dont understand the DOM of python/tkinter... so i cant modify the thingy...
please could you do any further work on this relative to the pyvcp_widgets.py currently in HEAD
I renamed some variables to make the Dial code a bit clearer (I think)
anonimasu,awallin: Been thinking about a generic object model to store any/all geometric objects/shapes and I would appreciate comments/suggestions -> http://www.pastebin.ca/324619
awallin: didnt I use a base source from head? & renaming vars is fine
tomp_: the tooltip code is not based on yesterdays HEAD, but that's ok...
Dallur: I like the relational geometry concept which is used in MultiSurf: http://www.aerohydro.com/products/rg/papers/cad/rgs-pt1.htm
I can send you the paper later today if you remind me, I have it somewhere on my hd, just can't find it now
tomp_, ##microcontrollers seems a little quiet; any other channels that might be helpful? Thanks for the tip, btw.
Dallur: there's more at http://www.aerohydro.com/products/rg/rg.htm#papers
awallin: that would be great, I would really like to use an existing "standard" if one is available
that scheme would be good for a new CAD program, not neccessarily for a CAM program which would need to import all kinds of files (DXF, IGES, STL etc)
PWizard: i dont know offhand, but it's like google, just invent a name, and see if any group is connected to that name... else part and try again :)
awallin: if there is a format which can be fed any of those formats and can generate g-code from that format that would make life easier
tomp_, heheh, thanks
awallin: I think it might be simpler to generate a 3D object model and build the g-code from that and then de-couple the parsing entirely
is this correct way to get a fresh checkout of head? (not-a-dev) cvs -z5 -d:ext:firstname.lastname@example.org:/cvs co -rHEAD emc2.1-alpha0 ( i already hid my last version for safety )
Dallur: here are some notes on relational geometry: http://www.pastebin.ca/324632
tomp_: probalby only 'emc2' at the end
2.1 would get you the 2.1 branch
Dallur: what do you mean by parsing ?
reading in DXF or similar to a program
RG is nice when you get used to it (parametric), but it might be a bit awkward at first if you are used to Rhino or something similar
then it will depend on what kind of stuff is possible to define in DXF or IGES, if we want to be able to read in everything that is possible...
tomp_: You need "-rv2_1_branch" to get 2.1, not "-rHEAD"
awallin: true, but I think the first step is to define a a format which can store without loss any geometric object created by any CAD app
tomp_: if you want to name the directory that 'cvs co' creates, use -demc2.1-alpha0
in fact you never want -rHEAD
Dallur: the problem with that is "any", that might include a _lot_
jepler: thanks & thanks :)
Dallur: found any good specifications for DXF, IGES, STL on the net?
awallin: adobe has pretty good spec for dxf on their website
awallin: and pythoncad has a module to parse dxf, it's not complete but it's a reference
[15:30:18] <jepler> http://emergent.unpy.net/files/sandbox/gdepth.png
awallin: that's what I used, but I plan to cut it out, it's easier to rewrite
jepler: what are we looking at?
looks like spiral.ngc cut with a ball end mill
close up: http://emergent.unpy.net/files/sandbox/gdepth2.png
it's a 4x4 area rendered at 200dpi
this is a matchstick model?
about 1.2 million OpenGL triangles
what is the tool's dia?
cradek: .3 inches
jepler: help please cvs [checkout aborted]: no such tag -rv2_1_branch
jepler: so where's the code? ;)
tomp_: for the latest pyvcp you want HEAD ! not 2.1
tomp_: cvs -d:ext:email@example.com:/cvs co -demc2.trunk emc2
tomp_: or, for the branch, cvs -d:ext:firstname.lastname@example.org:/cvs co -demc2.v2_1_branch -rv2_1_branch emc2
cradek , does the latter give me what awallin wants me to use?
awallin: I'm not sure what "matchstick model" means
cradek: so is HEAD now called trunk ??
[15:35:39] <jepler> http://emergent.unpy.net/files/sandbox/gdepth.py
my commandline: python gdepth.py -t .3 -x "-2.1 2.1" -y "-2.1 2.1" -r 200 -o spiral.png emc2.cvs/nc_files/spiral.ngc
awallin: yes, every branch has a head, so we were using the term wrong
it only works for some of the emc2 nc_files and I'm not sure why
jepler: heightmap, for each x,y point in a grid, store the height of the stock z
awallin: yes, that's what this is
tomp_: no, you want the first one (trunk)
cool, you coded this yourself just now?
awallin: I have been looking around for the last couple of days, reading up on 3D modeling and the math behind the whole thing and from the looks of it any 3d object created by any CAD app can be represented by NURBS, what I propose is to handle straight lines and constant radius curves in a simpler more efficient way but to create NURBS from all splines, this should in theory be able to handle all geometries created by any CAD app
awallin: worked on it some yesterday then added the opengl preview today
jepler: psyco is some kind of opengl wrapper?
awallin: no, it's supposed to be a general speed-up for python -- it didn't seem to make a big difference
awallin: I used minigl, the opengl wrapper included with emc2
ah, the import command is halfway down the file...
it also writes the depth information as a .png file
jepler: in AXIS: File/Preview?
considering it won't be fast enough to use all the time?
cradek: it's not ready for that yet
but maybe someday
AXIS already knows the program extents
nothing about tool shape (in mill mode) though
i doubt i have all the imports to try it, how long did it take to render?
yeah, but I can get diameter but not shape
jepler: how long does it take to run?
awallin: that example takes about 2 minutes to prepare and then it can be rotated/translated/zoomed at about 5fps. I'm running it on a 933MHz machine, displaying over the network to my workstation with a fairly powerful nvidia card.
"GeForce 6600 GT"
err - hi :)
SWPadnos: you're going to tell me your card is better?
jepler: that is damn cool.
* skunkwork still hasn't played with the new debs :(
jepler: only a bit better :)
* alex_joni thinks of dual FX3500 in SLI
with improved lighting
jepler: very nice preview generation, & thanks, the new 'emc2-trunk is running now
[15:58:48] <jepler> http://emergent.unpy.net/files/sandbox/gdepth3.png
unfortunately I should spend the rest of my day doing my real worl
jepler, ooh, looks like sim?
We need to find someone to pay jepler to make this "real work"
jepler: Is that actual machine path? what is the input scale? (seeing the roughness of it) - is that what we are seeing?
lerneaen_hydra: yes, it's an approximation of the remaining material after milling "spiral.ngc"
what type of material sim is it?
skunkwork: no, it's the ideal path read from an .ngc file. that image is at 200 samples per inch
ok, so no undercutting, IMO good enough for 3-axis milling
that example (4x4 inches, 200dpi) takes about 2 minutes to create, navigates at about 5fps
that was longer than I had expected
sorry to disappoint you
what time does it take to run it in sim/AXIS ?
I don't see why it should take so long though
* alex_joni bets a bit more :)
essentially all you're doing is painting with a round brush in a spiral
jepler: since you have enough data to get vertex normals, can you do smoothing?
cradek: sure but it'll make it slower and lerneaen_hydra wouldn't like that
jepler: if i modify make_tool_shape(f, dia): so it described a formed electrode for sink edm, and just plunged straight down some depth, then the render would reflect the correct cavity!? it's amazing what you coded in those few bytes
oh, wait, was this running under the RT kernel?
or did it suck up 100% cpu?
933 MHz machine, remote display to computer with GeForce 6600
jepler, I think it's fantastic. great work
* alex_joni adds "as usual
oops - sorry. haven't had coffee yet ;)
I ran it under 'time', and when I closed the window it reported 59.9 seconds "user" time and 1.1 seconds "sys" time
5-10 fps sounds quite fast though
jepler, oh, I see
how often is the "brush" updated?
once every 1/200th inch?
see - it's twice as fast now ;)
lerneaen_hydra: the moves come from the EMC interpreter, so they are in theory ideal. only the stock model is discrete. if I understand correctly
you think that updating it in discreet positions would be faster?
say at the same res as the "image"
probably not, that would break up one G1 move into several tiny moves, I don't see how that would speed up things
I was thinking for the "render to image" bit
the papers by jerard/drysdale talk about binning the matchsticks and then only considering those under the shadow of the tool when doing the calculations
* alex_joni lights a laser at your shadow
alex_joni, can't you set up ciabot to say "Sharks! And Lasers!" every time someone says laser, but no more often than once per hour?
ciabot is not mine
it belongs to cia.navi.cx
what about logger_emc then ;)
do you know perl?
I thought about that the other day, sharks and lasers are not a good combo, the corrosive marine environment, a powerpack big enough for a 100W+ laser and the fact that the water would wreck havoc with the beam.... I think Eagles with Lasers are so much better
alex_joni, no, but if I did, would you implement it?
eagles are too small for lasers
lerneaen_hydra: I would use your patch
* lerneaen_hydra must learn perl
eagles can carry deer - I think they can handle a puny laser
not the battery pack though, unless it's something like a small scale nuclear fission device, like the ones on some space probes
SWPLinux: the problem is always with the fusion generator for the power
oh - nukular eagles might be bad ;)
not at all ;)
LH: I like that, Plutonium enriched Eagles !
just take something with a short half-life and use it subcritically (just use the passive radiation)
less volume possibly
LH: that's what they do with the satellites, just a few hundred grams of plutonium coupled with a TEC Peltier
the halflife is probably much higher than the average lifespan of the eagles, with all the radiation poisoning they'd get
tomp_: yes, if the electrode is a simple shape that can be described by a depth at each location
Dallur, oh, they're also passive?
so no active control of the reaction
better yet - just have the eagles drop rocks on people
LH: Yup, just use the heat, most only get 200-500W
make the eagles generate the power - tap into the wing motion and wind energy
damn, so that would be around 4000-8000W heat?
LH: Not sure, TEC's are probably pretty efficient when you have empty space around at just a degree or two kelvin
tomp_: a tool profile like <_> can't be expressed, because it isn't described by a single Z value at each location. A profile has to be like like V or U or |_|.
hmm, carnot's law is nice when T_lower is very low
jepler: cool! there would be no path in the way mills work, just a depth, and possibly a tiny circular path, say .010" radius at that depth, the corner detail would be of interest to
still, it must be quite hard to cool it down to ambient temp
as there's nearly no mass to give the heat to
LH: Well you want the heat around, if those things get to cold they stop working
jepler: i understand no undercut, and it was 'of interest to edm pepl'
LH: So it's all about controlling the amount of IR radiation
oh, how hot do they want to be?
LH: Not sure on the exact specs but they probably want to keep them above 200
sounds very possible then
(becuase if I say it's impossible, then there are no satellites)
LH: Imagine trying to make capacitors and resistors for sub 200 temperatures, all the rules change
<SWPLinux> eagles can carry deer
I think so
oh, ball mill by default ts = tool.shape, i was wondering why when the cmd line didnt specify it
I mean, if a sparrow can carry a coconut ...
our cat drug half of a ham out of the sink
SWPLinux: How about if two African swallows carry a coconut together by means of a string between them ?
tomp_: yep, default is ball. you can specify shape with -b, -V, -v, -e and radius with -r NUMBER
jepler: yes there's a 'tool crib' to choose from :)
Dallur, I don't knoooooooowww.....
red,,, no blue !
* eholmgren flings Dallur
hmmm. jepler, is it a bug that the tool makers all use a normalized radius? it seems the depth would be exaggerated
at least, for tools < 1" in diameter
SWPLinux: hm, could be
"I’ve seen Ravens pluck the eyes out of lambs. They land on the lambs head and peck out the eyes and as the lamb runs until it can no longer stand, they land on it and start eating it alive."
thanks, intelligent design
mean and nasty design is more like it
[16:44:13] <jepler> http://emergent.unpy.net/files/sandbox/gdepth4-cds.png http://emergent.unpy.net/files/sandbox/gdepth5-chips.png
jepler: those are extremely cool
heh - here I am trying to rotate that chips image ;)
it don't work so good in firefox
there's still some kind of trouble where files that use the tool table don't work
That is so neat.
yeah it's hard to go work on what I'm supposed to
jepler: applause! :)
bah, I can do that in mastercam or surfcam just as easily ;)
* jepler adds the GPL notice and uploads another copy of gdepth.py
awallin: there, now my software has a crucial feature that mastercam and surfcam match (at least, I assume)
they match the lack of the feature
jepler: I see your already in the year 2770! cool.
they match the lack of the featured feature
jepler: minigl is at sourceforge, but where are rs274.interpret gcode rs274.OpenGLTk ? ( new 2.1 trunk user )
I concentrated so hard on getting "7" into the number
tomp_: doubt that's the same minigl
"nearly a milennium ahead of its time"
tomp_: minigl and rs274.xxx are from emc2.1
tomp_: it helps to source scripts/emc-environment
if it doesn't accept your .ngc file, try removing/commenting all lines that refer to the tool table (e.g., t1m6).
I'll figure out why later
any of you gotten it to run yet?
not on 3dchips though
is there going to be integration in axis? (click a button, define billet size, tooltype (unless defined in tool table) and hit generate
lerneaen_hydra: A menui tem, probably
but .. that's still way ahead
if previous projects are anything to go by, my definition of way ahead is a lot farther ahead than yours ;)
jepler: wow, less than 2 minutes here, wicked!
tomp_, how long? and what processor?
<2 min ( only used the tray clock) and a p4 2.6Ghz
oh, quite a modern system
Dallur: Looking now
Dallur: is that for internal use?
Dallur: for toolpath's ?
lerman: anonimasu, did you see jeplers cutting simulation?
[17:16:20] <awallin> http://emergent.unpy.net/files/sandbox/gdepth5-chips.png
[17:16:33] <awallin> http://emergent.unpy.net/files/sandbox/gdepth4-cds.png
and the code
jerpler owns us :D
[17:16:39] <awallin> http://emergent.unpy.net/files/sandbox/gdepth.py
jepler = t3h pwnage
jepler: how did you make thoose bindings?
and the cmd line when at root of emc2 python nc_files/gdepth.py -t .3 -x "-2.1 2.1" -y "-2.1 2.1" -r 200 -o spiral.png nc_files/spiral.ngc
The program is too short. It must be a fake. :-)
* anonimasu is looking for something crossplatform
* cradek fears the units in gdepth.py
it can't be right unless there are 25.4s sprinkled everywhere
are there 25.4's in it?
I wonder where he got the opengl stuff from :D
anonimasu: What I'm trying to do is to define a geometric object hierarchy which can correctly store any object created by all CAD programs, this way I can parse any type of file (dxf/iges) into this object format, then create optimized cutting paths from this format
anonimasu: I wrote it myself for axis, basically
anonimasu: minigl is all my own. opengltk is a wrapper around ToGL, and opengl widget for Tk
Dallur: 2D toolpaths from a plasma that follow the geometry will be easy, 3D rough/finish for milling is a _bit_ harder
anonimasu, have I spammed you with www.pungkuk.com yet?
and awallin too for that matter
anonimasu: opengltk is similar to pyopengl's OpenGL.Tk, except I hear that OpenGL.Tk is deprecated and no longer included
jepler: that's nice
jepler: pyopengl has no way of creating windows
except if you download the context package too..
I don't care much about "cross platform", so I don't mind that this runs only on Unix+X11
jepler: I do :)
if it's possible, the more platforms the better :) imo
yay for the gpl then!
awallin: I don't think I will do the 3D toolpaths, but I want to try to at least do the file parsing and be able to render the objects, then I will start to worry about the 3d toolpath generation
I'll be bacl in a bit
having a bath
jepler: it's 39 sec on 2.6G P4 w 512meg
oh, that's not that bad
Dallur: you probably want to make the geometry definition independent of rendering platoform (OpenGL, vpython, vtk, opencascade) etc. etc. so that it will be most useful
awallin: yup, by storing only storing the vertices that means that you can render it any way you want :D
Dallur: I would suggest the Relational Geometry we talked about earlier, I will try to dig out the paper...
awallin: I will need to duplicate some of the vector work already done in vpython but no biggie
awallin: the only issue I noticed with relational geometry is that it deals with surfaces, which in my mind is a bit silly, surfaces should really only be involved in the rendering process imho
well, before you draw a surface you need to somehow define it
awallin: any surface can be defined as a set of lines joining the vertices of the polygons, which for toolpath generation is what you want in any case
awallin: might just be terminology
yes, that's STL, but that's a very inefficient way of storing a parametrized surface (Nurbs etc). Wouldn't it be better to store just the defining points and the formula for the surface
but at some point, when toolpaths are generated an STL representation might be needed
awallin: it's the lowest common denominator from what I gather, ie. any format can be changed to that format without any loss of information
STL?? surely you loose information in an STL representation. it's just a question of how much
awallin: and when dealing with isosurfaces you can replace the lines with NURBS curves and get the same result
stl is to raster as nurbs is to vector, perhaps
hmm does STL even handle any splines ?
well, the most common implementation
the format technically has support for other shapes, but they're never used
awallin: what I am talking about is not STL, it supports NURBS curves fully, which means that it supports any type of spline, but rather than grouping them as a surface you group them as a solid
picked up a ~3' 19" rack at the surplus store for $15 last night ... woot!
I have this urge to fill it with old Sun hardware
but that would probably be pointless
alex_joni, you there?
bought another welder today
very VERY nice machine
awallin/anonimasu: I made a small demo with the data-types, I don't know if the python code is all that good but I think it demonstrates how any 3D form can be fitted into a hierarchy and broken down into vertices
jepler: if i use python nc_files/gdepth.py -t .3 -x "-2.1 2.1" -y "-2.1 2.1" -r 200 -o 3D_Chips.png nc_files/3D_Chips.ngc i get a rect prism, not like your .png (mm vs inch? got G21 in .ngc file)
and the ngc is fine in axis
jepler, is gdepth.py dependant on EMC or can I run it on my edgy box?
lerneaen_hydra: it needs minigl, gocde, and rs274 modules from emc2.1.
oh, the src files from head or compiled?
used to be called HEAD :)
so the newest version of cvs is trunk?
or is trunk the 2.1 split?
TRUNK is what HEAD used to be
TRUNK is the proper term for what we called HEAD: the most experimental development version
that can still be three things
the reason is that (according to CVS docs), every branch has a HEAD, which is the most recent version on that branch
there is only one TRUNK though
The branch called v2_1_branch is where we are putting the finishing touches on the version that will be 2.1.0
either TRUNK or v2_1_branch should be OK for gdepth
oh, so the 2.1 branch has its own head
and the newest/most unstable one is trunk?
yes - it was a terminology error before
and does it need the src or the compiled versions of those files?
is there any place to grab them?
[19:47:06] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Why_aren_t_there_packages_for_Ubuntu_6_10_Edgy_Eft
oh, do I need a working install?
sim works fine
SWPadnos: on edgy?
not that I've tested
that's what lerneaen_hydra is using
yeah, I feel like using bleeding edge stuff :p
* lerneaen_hydra is running XGL too
I'm sure a clueful person could build emc2-sim on edgy or any other modern system
is there any reason to believe that sim will/won't work on edgy?
it will certainly work
(ie, other than "dunno", is there any evidence either way)
there were some dependancy issues with the .dev
naming conventions and the like
but please don't ask us to help with all of it
some package names were different or the packages were unavailable in edgy
ok, but getting source and compiling should give you a working sim emc2 ...
so the stuff in debian/ is wrong
I forget which package names lerneaen_hydra reported a problem with when he tried to install dapper sim packages
hmmm. I wonder if I tried that when I had an edgy install ...
oh you mean the package?
something with python2.4
edgy uses 2.5, doesn't it?
I thought you meant "can it be compiled" to which the answer is surely yes
cradek, I don't plan to spam all too much ;)
cradek, yeah - that was my (mis)understanding as well
jepler: i tried 3D_Chips and got an untouched workpiece , yet 3D_Chips ran fine in trunk, any idea?
anyway, this was more of an idle question more than a "help plz"
tomp_: I had to slightly modify chips and cds by removing lines that referred to the tool table. not sure if that's the trouble you're having .
cradek, so don't get all too worried ;)
jepler: it was a 1 line fix, now, so why does 3D_chips run faster than the spiral ( amazed!)
tomp_: dunno. the runtime will depend on the area, the resolution, the length of the cuts, and the tool size
robin_sz: am now
jepler, bigger tool = longer time?
becuase there are more pixels to apple a transform to?
what's the difference between a NET application and a win32 application on windows?
different execution machine
is NET slower?
not by much afaik
afaik yes it can be, mostly it's about the same
bought a pulse welder today :)
coo.. what flavour?
not exactly well known
oh, I know them
nor top line
not exactly the best rep. :)
well, we'd looked at a migatronic
you're hurting me :D
* lerneaen_hydra hurts alex some more
anyway, it puts down nice welds for the money
lerneaen_hydra: you don't know how
robin_sz: lucky you then
you dont rate them?
fortunately I've never used one
I would have bought a Kemppi but that was lots more $$$
kemppi starts to sound like an ok machine
maybe esab, lincoln, fronius or cloos :)
but I'm biased :D
fronius we looked at a while ago ... very compley control
how much k was this?
* robin_sz nods
50k sek for a crappy mig welder?
haha 4k5 ^^
robin_sz: prolly 1-2k more for the other brands
lerneaen_hydra, thats not a lot for a pulse mig
pulse mig, what are they based on?
robin_sz: why go new ? :D
alex_joni, not a lot of good quality used machines out there right now ...
hmm.. I know of a particular one :)
lerneaen_hydra, you are familair with mig welding?
not sure if there are any in england though
standard mig sure
and yo uknow about the 3 basic trasnfer methods?
you probably only know dip transfer, where the wire dips into the weld pool
robin_sz: give the guys from Wolverhampton a call
the classic "buzzing bees" sound
robin_sz, yeah, only dip
alex_joni, the fronius people?
ask for PATS
and then things fuck up and you get spatter
too high voltage say
they might have some used ones
lerneaen_hydra, well, you know when you wind it up a bit more, you get a lot of spatter?
or too far away
PATS is 7+ years old, but it wasn't matched so far..
right .. well, if you turn it up even further, say to 350 amps?
how's it behave then?
now switch to pulsed, and you get that over the entire domain :D
does the magnetic field guide it correctly?
more or less
so basically, pulse mig does 40 amps background current
and then 600A spray
at about 100hz
with a varyign duty cycle
AC or just varying DC?
the avergae current is still 130A or whatever yo unormally use
but its in pulses
the result is amazing
well.. there is AC now
AC would fuck things up badly, right?
I was stunned I have to say
but robin_sz doesn't have the bucks :D
alex_joni, true enough
3K for a welder was a lot
never mind more
hmm isn't pulse mig supposed to be able to weld aluminium to galv. steel ?
thats mig brazing though
special silicon copper wire
alex_joni, what do you reckon a cloos PATS comes out at? 6K euro?
robin_sz: certainly less
alu to galv steel?
so like a jigsaw type connection between the two?
it hasn't been produced for 8 years?
something like that :D
I saw a sample from audi/VW factories which looked pretty sweet :D
older machine.. but it works like a dream
robin_sz: 603 PATS
built as a freakin' tank
2 water pumps (one for torch, one for electronics)
robin_sz, what were the other two types of mig welding?
lerneaen_hydra: he told about all 3
lerneaen_hydra, that was them ... dip, globular transfer and spray
dip, intermediary and spray
dip is sometimes refered to as short-circuit arc or similar
theres the thing lincoln try and get people to believe in too
so globular is several mm?
and spray is sub mm?
robin_sz: there is also CP (Cold process) now
but that's AC for thin metals
what do you mean sub mm?
the 14k I mentioned before
as in the particles are smaller than a mm or so in diamter
oh, much much smaller
or what is "spray"? gaseus metal?
spray must be almost a vapour
globular is just messy ;)
lerneaen_hydra, do you have a mig?
my dad does
if its capable of 300A + try it someday
turn the voltage up to about 36v
I don't recall the max current
and the wire feed 3/4 of the way to full speed
using say a 0.8mm wire
it just goes silent ...
but do it on thick metal :)
did that when we make our boiler. just painted the wire on.
robin_sz, this is steel you're talking about, right?
alex_joni, the welders loved it
oh, I only have alu gas ATM
* robin_sz nods
if you have a welder powerful enough for alu, it shoudl spray OK
quite fascinating to see really
it's a bit special, as the welding nozzle has the spool on the.. err.. thing you hold, rather than in the PSU
to put it in perpective, a modern mig pulse welder compared to a traditional mig, is like comparing plasma to laser ...
yes they both cut metal ...
so pulse FTW?
that much better?
two 5mm plates ... 8mm fillett weld, perfectly flat, at 45 degrees, no spatter, smooth smooth surface
robin_sz: core for a salesman job?
pulse welders are not common around here at all
I'd played with a fronius, but it was hard to use
this had one dial for material thickness, dial it in, weld.
SWPadnos: yes, the tool shapes were pretty screwed up
ok. I wasn't sure if I had missed something there
robin_sz: also, you cant compare the price of a plasma to the price of a laser.
awallin: a .net application requires the framework
awallin: good framework btw
the only good thing microcrap has done.
anonimasu, well, true, but in terms of results, thats the difference between standard and pulse mig .... looking at them now, standard mig is just ,..well, faulty!
anonimasu: I'm looking at c++ compilers. Found wxdsgn.sourceforge.net for winXP, uses wxwidgets for GUI, so should cross compile under linux also
anonimasu: just in case I actually get something useful done and others would want to use it
awallin: what are you working on?
quick terminology question, what are the '%' at the beginning and end '%' called?
ds3: start/begin of tape ?
alex_joni: nothing much yet, but I'm interested in the cutting simulation, parametric CAD modeler, and CAM algorithms. So it would be nice to have an OpenGL+GUI environment in which to try out different things
ds3: they are historical reminescents
alex_joni: Oh, and since I've spammed the freesteel blog with my comments, Julian T promised to release some CAM algorithms GPL which would be nice to try out
from back when programs were on tape or paper I think
now with multiple tools based on g-code comments: gdepth7-pcb-multitool.png red shows where you've cut all the way to the "bottom" of the stock.
e.g., to select a .1" ball-end: (TOOL,b,.1)
awallin, ooh, nice
jepler, tsk tsk, what kind of units are you using anyway... ;)
jeff: can you paste a full link to that?
[21:01:09] <jepler> http://emergent.unpy.net/files/sandbox/gdepth7-pcb-multitool.png
jepler: that is very cool
jepler, ooh, nice :D
oh - yay! I see that version also allows movements that are "off the grid"
or out of bounds, I guess you could say
* robin_sz points to a very tired kangaroo
get your boats
random question: how's EMC under KDE/xfce/fluxbox/blackbox ?
lerneaen_hydra: doesn't matter
so no gnome calls or anything?
lerneaen_hydra: the only thing you might be missing is the menu entry
only that on xfce you probably don't get a menu entry automatically
yeah, that's easy to add though
nice, one less thing to worry about
KDE might already work
03alex_joni 07TRUNK * 10emc2/configs/demo_mazak/demo_mazak.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/demo_sim_cl/demo_sim_cl.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/etch-servo/etch.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/lathe-pluto/lathe-pluto.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/m5i20/m5i20.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/demo_step_cl/demo_step_cl.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/hexapod-sim/ (README core_sim_6.hal minitetra.ini): beautify configs as on the 2.1 branch
are you all running gnome?
03alex_joni 07TRUNK * 10emc2/configs/dallur-thc/ (README dallur-advanced.ini): beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/motenc/motenc.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/nist-lathe/inch.ini: beautify configs as on the 2.1 branch
I use icewm and always launch emc from the commandline
03alex_joni 07TRUNK * 10emc2/configs/max/max.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/sim/ (8 files): beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/halui_halvcp/ (README halui.ini): beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/ppmc/ppmc.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/stepper-xyza/inch.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/univpwm/univpwm.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/univstep/univstep.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/stg/stg.ini: beautify configs as on the 2.1 branch
03alex_joni 07TRUNK * 10emc2/configs/vti/vti.ini: beautify configs as on the 2.1 branch
by commandline do you mean tty1 or a "terminal"?
03alex_joni 07TRUNK * 10emc2/configs/stepper/ (README sim_inch.ini stepper_inch.ini stepper_mm.ini): beautify configs as on the 2.1 branch
aah a flood!
* lerneaen_hydra is swept awaaayy....
rxvt-unicode to be specific
lerneaen_hydra: a terminal within X, not a text console
03alex_joni 07TRUNK * 10emc2/configs/m5i20/m5i20_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
03alex_joni 07TRUNK * 10emc2/configs/motenc/motenc_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
03alex_joni 07TRUNK * 10emc2/configs/ppmc/ppmc_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
03alex_joni 07TRUNK * 10emc2/configs/stg/stg_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
03alex_joni 07TRUNK * 10emc2/configs/univpwm/univpwm_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
03alex_joni 07TRUNK * 10emc2/configs/univstep/univstep_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
yipppeeeeee skeleton G code translator done
who wants to machine skeletons?
that's so morbid
Got my old acadlt installed on the 32 bit box, for some reason I can't get acad13 to install (yet), does anyone know if there are .DWG to gcode converters out there?
none I know of in the $affordable range (I think you can only do 2d or 2.5d dxf for $cheap)
I found some links earlier today
if you look at the log..
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-01-22.txt
I have the url at the work laptop
ah - gcam
NP, not like there is a rush going on here :). I may figure out what I am doing wrong getting acad13 installed...
I found a free one earlier
dwg2dxf or something like that
worked ok for me
I found some gnu cam stuff..
[21:40:53] <SWPadnos> http://gcam.js.cx/index.php/Main_Page
no, I found a whole page with a list of like 10
oh - I think I know the page you're talking about
but I don't know where it is :)
had cad programs lined up too
there's one with all sorts of free CAD/CAM programs - not just mechanical but electrical and PCB as well
anyone who tries to read DWG files directly is being silly. DXF is the portable openly documented format to use when moving drawings between applications
Found it, thanks!
If one has something that puts out DXF, which I may or may not have, just covering the bases...
alex_joni: did you need a drawing of a machine?
03alex_joni 07TRUNK * 10emc2/configs/vti/vti_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
oh my.. that took quite a bit to appear
alex_joni: did you see my question?
yeah, but didn't get it :D
I thought you were looking for a model/drawing of a mill
anyone know where to get one of these milling heads? http://www.macromould.se/images/frasmaskiner_swedenpurappl.jpg
a 5 axis head?
Martin_Lundstrom: mind if I ask you how much you want to spend?
ooh, it's chinamill!
[22:00:51] <Martin_Lundstrom> http://www.macromould.se/images/batbranchen_friggofraspulpet.jpg
anonimasu, too much ;)
Martin_Lundstrom: 80k usd?
anonimasu, you know somebody who makes that stuff?
ofcourse you can always make your own.
yeah, but I cant find the link right now
I would love to have a look at it
if this is a router it's possible that you can build one pretty easily
lerneaen_hydra, how is your studying coming along?
anonimasu, you think so?
nah, it's going quite well actually
anonimasu, do you have any good ideas for the A and B axis?
Martin_Lundstrom: are you thinking of Rab Gordon? http://www.rainnea.com/cnc_machinery.htm
SWPadnos, not at the moment
something like it
I saw a design at cnczone that looked nice
but it was complex..
more specifically this: http://www.rainnea.com/cnc_5axis.htm
seemes alot of design work, or?
if it's a router you dont need the drive to go through the head
03alex_joni 07v2_1_branch * 10emc2/debian/emc2.files.in: install help files to different location (-/help/ from the path), that's where configure expects them
lots of machining
there are others using two satellite positioners to get the rotary axes
03alex_joni 07v2_1_branch * 10emc2/src/Makefile: install help files to different location (-/help/ from the path), that's where configure expects them
wow, the rainea stuff is neat!
yep. he also has software called the CNC Toolkit, which is (at this point) free
where to get those sirveling things?
it generates 5 axis toolpaths from GMax models
I tried the software
I never got it to work right ...
hmm what shall i search for in ebay, any keywords?
the rainnea design is nice.
but that was GMax, not the toolkit
it works ok
it's a cheap design..
Martin_Lundstrom: head for 5 axis machine http://www.doughtydrive.com/
but there is alot of handson work, too much for some toolpaths, I think
tomp_, Is that arm sterdy?
Hmmm, GMAX comes with m$flight sim 2004 (fs9) IIRC.
Got it installed (but never used someplace).
Got it installed (but never used) someplace.
the doughtydrive seem to be directdriven from steppers...
Martin_Lundstrom: dunno, all things have some limit
Martin_Lundstrom: depends on what you need
it seems easier for a smaller mill for the spindle to be fixed and the workpeice to moce
yep, this is why Im fond of snail gears, no limit
like the cradle/rotating table config
[22:14:23] <anonimasu> http://video.google.com/videoplay?docid=1304521895942190362
i'd beef it up with some gussets, that upper plate is a bit thin
lerneaen_hydra, in this case i can not use that
[22:15:12] <anonimasu> http://video.google.com/videoplay?docid=-923619711177709820&q=5-axis&hl=en
oh, too large workpiece?
they sure run the carbide cutter hard
got your boats?
I wonder what they are making
flood coming up
what's with the pecking action when roughing?
I have a rubbermadrass ;)
they do hsm loops at the end :D
5-axis is pron dammit
[22:17:53] <anonimasu> http://video.google.com/videoplay?docid=-2565514988992667488&q=5-axis&hl=en
what's with the varying rpm?
that's the servos
noise, must be some gearing
at which movie?
the last one
I think the rainnea 5 axis construction is really nice
03alex_joni 07TRUNK * 10emc2/configs/ppmc/ppmc_io.hal: remove references to spindle-incr-speed and spindle-decr-speed
[22:20:27] <anonimasu> http://video.google.com/videoplay?docid=8267565192144365360&q=5-axis&hl=en
lerneaen_hydra: look at the end
looks quite simple to implement
or so :)
lerneaen_hydra, but where to get the pieces?
Martin_Lundstrom: make them?
Martin_Lundstrom; yeah :/
that's probably it
anonimasu, You want to help me?
Martin_Lundstrom: preoccupied with work stuff.
anonimasu, no probs ;)
Martin_Lundstrom; with what?
the 5-axis thing?
I'm going to bed soon :/
worm gears and stuff should be easy to find
can you save my soul?
* lerneaen_hydra saves Martin_Lundstrom's soul
[22:23:56] <anonimasu> http://video.google.com/videoplay?docid=-6442965485163123253&q=5-axis&hl=en
Martin_Lundstrom; seriously though
* lerneaen_hydra makes world peace
lerneaen_hydra: look at that machine
* anonimasu is amazed
* lerneaen_hydra wantsss it!
lerneaen_hydra, are you interested?
[22:25:27] <alex_joni> http://video.google.com/videoplay?docid=8267565192144365360&q=5-axis&hl=en
look ma.. no chips
lerneaen_hydra, I want to make the BC axis part
did you see how fast it drilled the holes!?!
anonimasu_get is now known as anonimasu_
Martin_Lundstrom; oh, hmm
my other client died.
what did I miss?
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-01-22.txt
the workpiece changer/part inspector by Festo is neat http://video.google.com/videoplay?docid=-3434713554178934645&q=5-axis&hl=en
(turn DOWN the volume)
Martin_Lundstrom: are you swedish?
anonimasu_; Martin_Lundstrom = chinamill :D
I cant remember where he's from :D
anonimasu, yes, I used to live there
ok, not now?
Now I live in Nice
that's too bad, it's always nice with a userbase in the same country
anonimasu_: I know :D
alex_joni: move here
1 / year is enough for me :)
lerneaen_hydra, Are you interested in helping out with a BC axis milling head?
alex_joni: aw, comon it's nice with the new goverment and all..
too cold for me :D
it's just -22c
outside right now
If I ever make a 5-axis mill it would probably be the cradle/rotating thingy type
it was +9 here today
hey, -10 here
it was +10 yesterday
-8 for the weekend though
rotating the spindle is hard..
lerneaen_hydra: go puma :D
lerneaen_hydra: a cheap robot
I found a ASEA for 2000eur :D
he needs force for milling
so the robot is the XYZAB table?
100kg payload min. :D
lerneaen_hydra: no, just stick a spindle on it's tip
lerneaen_hydra: way cheap
I'm hoping to keep my 3-axis mill + retrofit under 20k SEK
a new one is 80k :D
[22:33:06] <anonimasu_> http://www.blocket.se/vi/10617921.htm?ca=1_12_s
ie 2k eur
anonimasu_: that new :D
the mill is 1.3k€
anonimasu_: it's asea all right
asea.. it was :D
before ABB was founded :D
the stability is probably not good enough though
so.. not the newest :D
I'd still like it
[22:34:44] <anonimasu_> http://www.blocket.se/vi/10467882.htm?ca=1_12_s
how about that?
the first one is better
[22:35:20] <anonimasu_> http://www.blocket.se/vi/10453648.htm?ca=1_12_s
75kg is not crap
Anyone, what keywords to use to search for the swivel thing in http://www.rainnea.com/cnc_5axis.htm
the first one?
where did you find 75kg?
the last one
first one is 6kg which is about right for that size
the ones I use are 15kg
and 5000mm working range :D
well.. think I'll head to bed
CIA is not putting my messages through :/
good night all
they should have some 5 axis stuff
'night everyone else
hmm but type that rainnea used looks neat
if you can make the parts yourself.
the commercial heads are not,,
maybe the one tomp posted..
but the real ones are several $10 000
it seemes alittle sloppy
hm, for that kind of cash difference I dont think it's a big deal :D
Considering getting these for thrust bearings for the uMill, assuming they are 8x16x1?mm (got a query into them). Comments anyone? http://www.skatebearings.com/etoad/index.php?preq=viewproduct&productId=164
[23:07:25] <K`zan> http://www.skatebearings.com/etoad/index.php?preq=viewproduct&productId=164
hm, why not..
I'm considering skipping my duplex configuration..
and going for a easier design..
Seems to be what others are using although no one ever IDs them as other than "skate bearings"...
it feels overkill..
Anything has got to help :-).
Cheap :-) too.
my bearings are like this > | <
I'm going with all the examples I have see other people do... :-/.
and with a nut for the screw..
Seems simple enough...
Martin_Lundstrom: the video looks pretty good http://www.doughtydrive.com/images/DD2a.wmv
it's a mess :)
though it works really well :)
Will order them once I get the actual dimensions then.
those skate bearings.
[23:17:18] <K`zan> http://www.skatebearings.com/etoad/index.php?preq=viewproduct&productId=164
these.. they are metric sized
what dimension do you need? I think I have one around I can measure
From what I gather 8mmx16mmx1mm
What others are using on the same mill for the same purpose...
8mm is the inside dia for sure
they are thicker then 1mm
let me get mine
I need 6 of them for the 3 axis.
they have thicker, but that would require replacing the lead screws...
Now if I tighten down the cranks, it locks up or is too tight to turn...
those bearings are 8x22x7 mm
Not sure how that would work, probably bigger than the blocks.
SWPadnos: is there some trick to decoding 608?
Doing this slow and easy to the budget :-).
As I can and all that.
I googled for "608ZZ" :)
the ZZ means double shield, I think
Yes, should be.
lemme see if I can find an example of what I am doing here....
the numbers are pretty standard. if you look at MSC or McMaster, you can find the standard number for the bearings you have
K`zan: are you in the bay area? I got a bunch still in the package that i found cheap
Oh...heh thought there might be a system to the numbers
then look on eBay - there are loads of bearings for sale cheap (like $19.95 for 100)
[23:22:52] <K`zan> http://www.theblossers.net/index.php?action=view_gallery&id=3&module=imagegallerymodule&src=%40random44022cc7ad591
for a bit longer :).
Oh n/m... shipping will kill it
Yep, appreciate the thought MUCH though :-)!
cuz I can probally let you have 6 of them for maybe $3ish
no - the only standard thing I know is that 6000's are standard, 7000'sa re angular contact, and I think 5000s are extra thick or something
(that covers my cost; they been sitting around for a long time so...)
Pretty sure they are going to be too big and too thick :-(.
Lemme go mike that...
thrust bearings aren't the same thing as the 608's
heehe.... what I would like to really find out is the official specs on those 'trailer' angular contact bearings
SWPadnos: Heh, I only need 6 of them :-), 100 I would have no idea what to do with :).
are they all the same? .... :)
you have the Harbor Freight micro-mill, right?
ds3: Unf the 7mm width would outrun the shaft :-(.
the parts list doesn't seem to indicate that there are any bearings
see this for what I am planning on doing:
[23:34:27] <K`zan> http://www.theblossers.net/index.php?action=view_gallery&id=3&module=imagegallerymodule&src=%40random44022cc7ad591
Seems to be the "usual" fix for the uMill...
In one form or another...
K`zan: you might be stuck with buying something from MSC/McMaster
ds3: Could, I'm sure they have something, but most people seem to be using skate bearings for this...
8 for $9+shipping...
skate bearings aren't the same thing that person used
No, but a lot of people are from what I gather and successfully.
sure, but the mount design would be totally different
Anything is better than what is(n't) there now :).
8x16mm is 8x16mm :).
those definitely are not 608ZZ's
No, but the idea is the same...
no - 8x16mm radial bearings != 8x16mm thrust bearings
it's the difference between a tire and a lazy susan
Hummm, wonder why so many people use skate bearings ?!?!?!?!?!?!??!?!?!?!?!?!????
they use different bearing blocks
SWPadnos: those are thrust bearings?
Could well be, could WELL be...
for a radial bearing, you have a (very tight fitting) hole that holds the outer part (like a tire), and the inner part turns
didn't know they make them like that
a thrust bearing allows one side to rotate independently of the other, like a turntable
well, that's what the guy said, so I'm believing him ;)
As I understand it (and fully realize that I could VERY well be confused), they use the bearings to snub everything up and still have it turn.
for something that rotates so slowly, why not go with sintered bronze bearings?
err - the blog K`zan poisted, not the 608's
tomp_, anonimasu I discovered that the rainnea design is not a snail gear type...
I can tighten up the hancrank wheels to where I have no backlash, but I can't turn them at that point (well, not easily).
skate bearings? http://www.electronicsam.com/images/KandT/zaxis.JPG
Because, it seems, that the handwheel is flat against the block - the skate bearings would still allow it to turn.
skate bearings in that usage might be bad, a large enough load may pop the alls out
no they won't
this keyboard is getting non responsive :(
no, from what I gather they are in a "dished" holder.
with like washers on either side.
but that's not how a skate bearing would mount
skate bearings have a center part (central like an axle would be) that you would be clamping from both sides. you would still be able to turn the outer race freely though ;)
a skate bearing is held by the outter circumference/race
| == skate bearings --- == shaft(leadscrew).
I'm sure as hell confused...
More confuzed :-).
this is where a white board app would be useful
follow the link from that blog to the mcmaster part number
when it's loaded, let me know
ok. see the first diagram
Problem right off the top is that the cost went from $10 to $60 :-(.
where it says "shaft dia." "thk" and "OD"
that's only there for the pictures :)
I'm not suggesting that you buy from there :)
so - you see how the bearing goes together? flat ring on top, rollers in the middle, flat ring on the bottom
(well, flat with a groove for the balls)
Skate bearings are , in theory anyway, same thing with a couple washers, but I see the point about thickness now.
no they aren't - that's what I'm trying to tell you
lemme go look at the skate bearings again.
wait - they have them on McMaster as well
search McMaster for 5972K45
OK, I see what you are saying now- skate bearings are attached to the SHAFT, not just on it.
for a thrust bearing the two sides are independent, fora roller bearing the center and outside are independent
What I want is the ones that just sit on the shaft and are supporting the bearings against the blocks.
if you can use something bigger -
[23:48:54] <ds3> http://www.goldmine-elec-products.com/prodinfo.asp?number=G13613
SO much for a cheap solution :).
if you bore out the block a little more, then you can use a skate bearing, and it won't even take up any extra thickness
1.27 OD @ $1.00/each or $0.50 10+
Unf way too big :-(
boring out the block is the problem at this point....
or keep an eye out on the surplus vendors, small thrust bearings do show up everything now and then
The main reason I like the way the link did it was no boring needed...
Easy enough to do once I have CNC working.
what's the problem with boring it?
you must lose some travel then, since the thrust bearings add thickness ...
No 4 jaw chuck for the lathe to begin with.
$100 for that...
boring only needs a drill press (and a lot of patience, if you have a low power spindle)
that is not a problem
more or less.
yes, you can bore!
I have that lathe... it comes with a face plate
no need for a 4J
err, ah, how does AHA, didn't think of that!
Yes, it came with one.
NOW I see!
BUT before you begin
THANKS THANKS THANKS THANKS!!!!!!!!!!!!!! :-) :-) :-).
face something and check for flatness
Use it like a 4 jaw chuck, sorta - kinda.
mine was out of alignment by at least 0.010 :(
I can face it once I get it mounted.
no, face a dummy peice
probably should anyway.
otherwise you will wind up with a dish or a cone if it is out of alignment
all that will do is transfer whatever is out to the dummy?!?
it is out
I'd take a piece of aliminum, thread a few holes (for mounting a workpiece), drill a few more (for mounting to the faceplate), and turn that down for a good flat surface
Why not square up the faceplate in the first place ?!?
yeah, my headstock was not aligned with the bed
well, yuo can do that too ;)
* K`zan ignorance is surely showing :-/.
no - probably mine, I'm not (yet) a lathe guy :)
faceplate is 1 - cast iron; it is suppose to be dirty stuff; 2 - face plate is something you paid for, best not to waste it
Wouldn't that make more sense in that you only have to sqare up the faceplace rather than everything you do with it?
it isn't the face plate, it is checking to see if the headstock is square wrt to the ways
Ah, I think I see, a face place for the face plate....
you can put a large peice of stuff in the 3J chuck and face that
that's what I did
Heh, I don't think anything on those is cqure...
and the bed is not of consistant thickness either... really makes it hard to use the camlock mod on the tailstock :(
Hummm, back to being lost and confused again - at least I didn't waste $10+ doing it though :). Thanks!
The $60 for the bearings is starting to look better and better...
either you pay all at once or you pay peicemeal :/
Yeah, maybe I should just byte the bullet and get the CNC-Fusion mounts...
Expensive as hell, but plug and play...
[23:58:49] <K`zan> http://www.cncfusion.com/micromill1.html
I have never thrashed so much on any project in my life :-).
you got the LMS Y axis extension for it?
Not yet, but it IS on the list. Probably before the CNC stuff.