i tried to burn a 3.7G dvd .iso (OSX for i386) to a dual layer media (ok for the dvd writer), got message saying "more than 50% space would be wasted" and GnomeBaker exited. Just installed k3b, same thing. is it possible to use big media and 'small' image?
Where did you get the iso from?
tomp: did you look for a "fuck off and stop trying to think for me" checkbox?
JymmmEMC: google eeepc and mac osx , you'll find it
jepler: sounds like Pascal, right
hokay, off to get some single layer dvd media, gnite
Heh, BIOS VIRUSES RULE!!! http://en.wikipedia.org/wiki/Extensible_Firmware_Interface
* chr0n1c just got back from fishing..
i only caught one... it was a largemouth bass about 10 inches long...
BTW... These are AWESOME drives... http://www.westerndigital.com/en/products/products.asp?driveid=357
better than bein' skunked... like i was last weekend ; (
oh man.. i saw a skunk... it was ready to fight but like 40 feet away.. hissing at me or something...
it mighta been the geese hissing... i dunno
the cool thing is the river is like 5 or 6 blocks away from my house...
skunks HISSING is NOT good... thats what comes just before Pissin'
lol, they are so cute though
a girl at work has a dog who killed 19 last lear alone
mi sisters 2 malmutes will do in 20-30 a year
wow that's a lotta skunks.. whre you at dmess?
dayton, ohio here...
the little miami river...
toronto Canada.. my sis is in Kirkland lake ontario
*which is waaay up today from all the rain...
we had a very HOT day today...
i heard a staory the other day about a three eyed fish... being in the industrialized city that i'm in.. i could almost believe it
it was up to the high 80's today after it quit raining
i could believe it...
my taskbar says it's 73 now... and the sun just went down a few minutes ago
all i know is im melting...
sorry to hear that... i almost melted yesterday! it hit like 92-93 around there and i was outback putting up my chicken wire fence around the garden....
the garden being a big 30 foot tall J (for my first initial) in my backyard that you'll be able to see on google maps satt. view when they update dayton ;) yay!
**right next to interstate 75 where lots of people should see it and ask.. "wtf?"
what you growin??
working on the freedom 35 plan???
tomatoes, bananna peppers, lima beans, sugar snap peas, radishes, eggplants
nice variety... cool
and some other peppers... and bush beans..
lol.. i need like a 10x bigger garden than i have
i was gonna do green onions but i forgot to get the seeds...
get some of the big chinese beans... 24-36 inches long
these lima beans are "big mama" from burpee...
spose'd to get 10-12 feet tall with 8-9 inch pods...
the beans are huge!
1 lima bean per spoonful, lol
that's what i said... i had to try them
i'll tell ya the rabbits loved the leaves before i got the fence up
and the family of groundhogs...
* chr0n1c is gonna play with some mastercam...
has anyone who's used emc also used mach3?
i hope that isn't a bad word here..
I think that some people may have tried it at one point or another
I've done some testing, but never run a machine (with either, really)
art is a pretty darn smart guy, so the software does what its supposed to from what I can gather, I think the only missing part to mach is the realtime bits
the others would know more specifics then I would
well, the whole feedback thing is completely missing ...
same with emc without a card though right?
oh thats right
you can use a parallel port to read encoders and do PWM
yeah, thats the uberness of hal right?
but the real thing missing from Mach is the abiilty to do actual servo control (with the software)
like, PID - "my input isn't what I think it should be, so I'll modify my output"
Also Mach is stuck in the step/dir rut - as if that were the only control mode.
that's almost completely true, but there is a driver for some $1000 Galil servo controller board
Mach still outputs positions to the board, which does PID on its own
(similar to the motion controller and HAL PID components
Only Galil I'd want is in 7.62mm :)
you should build a tiny 3 axis mill or something SWPadnos
or retrofit my Bridgeport
just put motors and gears on it and get a 100 dolla control card
shoe horn a BOSS next to it and gut it proper.
I have multiple Mesa cards, a USC, and a Gecko G100 - I don't think I'm lacking hardware ;)
I've got a question about my 8.04 install, but I don't have the machine right here right now. Should I stfu until I can get my hands on while I'm talking here? :)
i'm waiting for a bridgeport to fall from the sky and land in my garage
if you remember the problem, and can remember an answer, then ask away :)
i'm running 8.04 x86_64
that would be painful
chr0n1c: shot placement is gonna be very important there :)
well, I have a cute little 750 MHz 1U rack computer that seems to have a TYAN mobo
and I can't get at the video chip but I strongly suspect from info I have gathered that it is an i810
oooh Intel PIIX4 chipset maybe - the really good stuff
ignoring the latency-test results for the moment
the tcl/tk sim stuff and the curses interface all seem to work
chr0n1c, no biggie - those sometimes work OK, and if not, you can get a PCI or AGP video card for it, you should be golden
err - not chr0n1c , mmgood
OpenGL hoses up the screen, whether glxgears or the OpenGL/Python EMC2
* chr0n1c says his 64 bit mobo has a via unichrome video card built in
I can live with the tcl/tk interface for now
but since it's only a 1U rack box, the riser will only hold a single PCI card
mmgood: did you turn off desktop effects?
i'm using latencey settings from my 233 mhz pII right now on this 1.6ghz box i'm gonna creep it up little by little
and I want to use a "sacrificial" PCI parallel port
cradek -- nope
in my experience i810 works as accelerated opengl with realtime
I hope this Jaca IRC client is logging :)
desktop effects screw up all opengl applications that don't run full screen
not an emc problem
just a bad decision by ubuntu (effects are more important than GL apps)
fortunately you can turn them off
cradek: is the effect possibly a repetetive horizontal blitty sort of line?
not sure, it might vary
cradek: well that will be an easy thing to check
i uninstalled compiz today... it didn't do much on this unichrome driver anyways
it's probably only like a 4MB vid buffer
wasn't sure OpenGL could even fit in that
if all else fails, use the vesa driver. it will usually work fine but slow
well for now I think I'll just "punt" the OpenGL for this box
but the latency-test suggests doing OpenGL cr*p
so let me see here.... afk for a moment
and thanks btw
welcome, hope you figure it out.
[03:42:43] <chr0n1c> http://potrace.sourceforge.net/
<- looks interesting here, i'm getting ready to try it out
Stuart said that he had to crank up the shared video mem to 256 to get good OpenGL
Skull: well that just ain'
t happening on this box. :)
it appears to be a Tyan S2420S mobo
ok - He was using a GTW Profile III
and lspci | grep VGA
returns 00:01.1 WVG compatible controller: Intel Corporation: 82810E DC-133 (CGC) Chipset Graphics COntroller (rev 03)
crud, I think I left my latency notes at work :\
Thats Intel's hardware version of Windows ME
but assuming I can remember the right order of magnitude...
a-l-p-h-a is now known as a-l-p-h-a
if the tighter (25uSec) timeslice has say 9000 to 16000 nSec of jitter...
and the looser (servo) (1mSec) one has a value lower than that...
i only ever encountered WinME once in the short life it had and i formatted the drive the first chance i had
is an optical drive attached (CD/DVD)?
toastyde1th is now known as toastydeath
might try shutting down and yanking that and run the jitter test again
though now that the 8.04 LinuxCNC distro is in place I could unplug it I suppose
It made a huge diff on my ITX system
I was getting about 9 uSec of jitter or less until I ran some other stuff
but that still seemed like a lot on the 25 uSec timer
I haven't figured all this out but jus as a percentage it seemed unacceptable
I'm going to redo the test with a 7200rpm drive (not the slow Bigfoot)
unplugging the CD drive is an easy thing to try
this little mobo has a SCSI interface (because it wants to be a server when it griws up)
then Ill post the new results for comparison
ouch - I love SCSI but not with realtime
I have used lots of realtime scsi machines and they have always worked perfectly
If I knew for sure exactly what I had here I'd post results but TYAN's site has orphaned the board AFAICT
they have all been P3
this is a sub-GHz P3
well maybe not lots, but at least 4/4
as always, the devil is in the details
i've got one ide channel and two sata plugs and NO sata drives... and a bunch of ide drives.
which I lack
may try my ASUS P2BS then - has a Tualatin Cele 1.3GHz on a slot 1 adaptor.
but this at least was not hiccuping with the horrible once-a-minute overruns of an earlier box I wanted to use
I've also got an nVidia 6100 (?) box that I'm typing on right now that 6.06 didn't like
I mught just try the Heron CNC livecd on that
anyway, thanks for the suggestions
I bet - only thing worse is a 6150 (with encrypted hardware HDMI Vista support)
what about XP on a 233 mhz pII with 384 mb of ram?
it seems that no one comes right out and says it, but are nVidia boards just a bad idea for RTAI?
video, I mean
i think the wiki says that somewhere
or said at one point
My TNT2 works ok...
well, it was worded sort of indirectly
mmgood: yes in my opinion
OK, I am off to try liveCD on this box
you can get by with the vesa driver using them though
thanks again for the suggestions
on my old gateway mobo i was using a tnt2 it was working great
matrox and i810 have always been best for me
how do I force the vesa driver?
put Driver "vesa" in xorg.conf
edit some x11 config file?
not sure if my 8800GT would be any better - all the gee wiz stuff isn't used.
need to remove anything else?
goes in the Section "Device" section
just comment out all others?
if any are active?
if there's another Driver "...", replace it; otherwise, add it
just vesa on its onw line?
sorry <--newb without machine handy
once again, my appreciation to you both / all
off to try this CD on my admin machine ... "Oh no, Mr Hands, he's gonna be mean to meeeee..."
"and you'll have to change it every time you boot the CD" ...
if you don't have a cd drive Gmount-iso works great... ;)
Got a question about Axis / tool radius compensation (G41, G42 (I think). Running my program and watching it in axis, having selected a tool with the right diameter, it looks like the cutter is running on the line rather than (in this case) inside of it. Having tried G41 and G42 the program errors out with something about cutter gouging. How does one get the cutter to cut on the desired side of the line (circle in this case)?
the line shows the tool path. it changes with the diameter
So what axis is showing is not what will actually happen?
you could run it once without comp, then don't clear the backplot, and then reload it with comp
axis IS showing the path the tool will take
As it sits the tool is straddling the line.
axis does not show your part outline. it shows the tool path
So the only thing to do is feed some material to it and see what happens to the part.
change your program to use tool 0, and reload it
sooo... is there any cool hidden bits of axis that isn't in the docs?
you will see it change
OK lemme give that a shot, thanks.
chr0n1c: my favorite feature is highlighting a line in the plot or gcode and seeing the corresponding line in gcode or plot
oh, i found that yesterday ;)
I bet you haven't read the docs then!
i was just randomly hitting all the buttons and clicking stuff to see what would happen
i read them a while ago
about 2 years ago
Huh, now I am getting "Cannot change units with cutter radius comp".
Lemme try putting g40 back in there.
and i just brushed up enough to get this amd box running emc2
maybe you need to put g40 at the beginning
you should kill cc whenever you use it
i almost crashed the mill yesterday because of that =(
if you stop the run in the middle, execute a g40 before doing much
I thought that was what G40 was, but it sure isn't working. Same message.
toastydeath: that's good advice
huge g-code reference wall posters!
k`zan_emc: are you hitting g20/21 somewhere
Got it, had to run it in the MDI
With t0 it does the same thing, cuts right on the line.
yes but it's a different line
leave the backplot there, reload the program with T1 or whatever
the mill doesn't really care about your part, it cares about where it has to put the tool
so it makes sense that it shows you the path, not the cutter comp'ed thing
Err, it is? Looks the same to me, plots over the circle as before.
but is it the same circle or different?
and what's the radius value
So when I draw up a part I have to make the circle smaller than what I want to compensate for the tool diameter?
with cutter comp, you program the part size
cutter comp is g41/42
So if I want a 1.5" circle and have a .125 cutter I draw a circle .125 smaller to get an actual hole of 1.5"
* K`zan_emc confuzed as hell.
program it 1.5" and kick on g41 or 42
the machine will keep the cutter away from the line by whatever the radius value for that tool is
Must be something more I need to feed G41/2
check your tool table, make sure you have your diameters right
g1 g41 d1 x0 y0
then load one of those tools
I tried both in place of the G40 and it crashed the program.
One at a time.
did you give it a d word
T2 in the stock table is the tool I happen to have :-).
but did you give it a d word
all cutter offsetting commands need a value to look at
Er, no, didn't see anything that indicated it required an argument.
G41 to cut inside and G42 outside.
g41/42 need D*, 43/44 need H*
uh, 41 is cutter comp left, 42 is cutter comp right
g41/42 are left/right, not inside/outside
referring to which side of the line it stays on
as it cuts
if you do a g2 under g41, you get an OD
g3 under g41, you get an id
goodnight, my battery's about gone :-)
well in this case I with the direction the tool is traveling I want G42 which is inside.
cutter comp right
Thanks cradek es gn
do not get caught up calling it inside and outside
because you're gonna crash your machine someday doing that
Heh, someday :)
Someday I am going to figure out enough to consider actually making a part :).
So it won't just take the diameter from the tool table, I have to manually feed it to it?
no, you need to tell it WHICH value on the tool table to look at
I would assume it wants 1/2 dia to get it where it wants it/
so if you do T1 M6
G42 T2 ?
you need to do G41 D1
no, no T word
T is only for the toolchanger
D == tool number here as T == tool number to the T command?
OK, thanks MUCH!
D and H refer to the radius and height of the tool number
T is, on bigger mills, a command that will actually cause movement
so it's ONLY used for changing tools.
We'll get to height later, no idea in hell how to deal with that :-)
height is simple.
easier than cutter comp, for sure =)
Once I get real homes / zeros for this think it will make more sense. I hope.
The program I have is one generated out of qcad via dxf2gcode.
i don't use any of those programs so i have no idea what you just said
Needs a lot of work.
Takes a acad drawing and turns it into gcode.
height offset is actually more important than cc
Z0, for the machine, is the point farthest away from the spindle.
All the docs are in German so I am staggering around in the dark.
Have you seen what I am working with / making up?
hold a sec
[04:36:23] <K`zan> http://wrlabs.shacknet.nu/~vw/MyMachineShop/PipeDreamMill/
haha cool man
No laughing, it is a learning experience :). OH is it a learning experience :)!
Back to the emc machine...
i don't know if EMC uses true diameters or radius values
all the machines i use are radius
radius I believe
"Cutter is gouging with cutter radius comp"
it's likely you're trying to do "inside corners", which EMC is verrrrrry picky about
lol, and rightfully so
Tool might be too big, lemme try T1
(ie, it complains every time you try one ;) )
Tool table asks for DIa values - might be Radius for the revised lathe tool table.
any tool is too big unless you have an arc instead of a corner
Essentially making a mount for a NEMA23 motor
swpadnos: does emc support R words on g1
that's another way to get around the problem
oh - good point. those are diameters (oops :) )
4 holes for the screws and a 1.5" in the center...
didn't I say earlier that I don't actually run any machines?
R can be used but is not recommended
Well nothing from t0 to t2 gets me any other message. Sigh.
Maybe I should just drill those manually.
if you are trying to go around an inside corner with cutter comp on, you will get an error
k`zan_emc: you're not trying to drill with cutter comp, are you?
So I'll need to make a smaller size hole and forget cutter comp.
No trying to make a NEMA 23 mount with a 1/8" end mill.
4 mounting holes and one large one in the center for the stepper
so you'll "drill" the screw holes, and mill out the shaft hole, right?
"G0 G17 G40 G80 G90" - put a safe reset line in there between tools if needed.
Well, to be real honest I have not a clue what I am doing and it is getting less as time goes on.
Guess it really wouldn't matter if all the sizes were .0625 large.
[04:44:44] <SWPadnos> http://www.linuxcnc.org/docview/html//gcode_tool_compensation.html
i am really not sure what the problem is unfortunately
also you need to have the right "approach"
when you go into cutter comp.
otherwise, it won't work.
kzan - I can whip up a proper program for you - if you have a sketch of what you intend.
Yeah, I am quite confused ATM, going to read that section another 10 times and hope it at least starts making some sense :-/.
it won't until you start making parts.
you can read all you want, you gotta eventually sit down and try to get the physical machine to do what you want
SkullWorks_PGAB: VERY much appreciate that, but I have got to be able to do this myself else I will become a REAL PITA to you folks that do understand it :).
Yes, I need to throw some material at it and see what it actually does so I can figure out where in the process to compensate for what to get it right.
and then the questions become more answerable
because you are forced to tackle one issue at a time
rather than a whole part at once.
true - but seeing a program run - comparing the code to what you see it do will help you understand whats happening.
Well, silly me :-). Figured I drew up just what I wanted in QCad and ran it through dxf2gcode and that should have produced would be the perfect part :).
haha, if it were that simple, engineers would make excellent machinists
SkullWorks_PGAB: Yes, examples are indeed VERY valuable.
maybe - if the layers were set for different Z values
What I chose for an example was a plate to mount a NEMA23 stepper on.
SkullWorks_PGAB: milling depth we have not even addressed yet.
ok - maybe its time to stick with a spring loaded pen on a sheet of paper to trace out your moves.
SkullWorks_PGAB: you want me to put up the .dxf and the program that got generated by dxf2gcode and hacked to work (such as it does)?
k`zan_emc: you're probably best trashing that program and figuring it out.
Heh, it is the only liferaft I have ATM :).
Atleast we can see any errors you may be fighting.
It does do something that sorta kind looks like it is in the right places (depth not considered at this point). With this thing, assuming I actually make the part, it will take a few passes assuming a 1/8" 6061T6 plate...
lemme stuff them up on the site, just a sec...
Looking for the dxf...
i am kind of interested in seeing this program myself
* SkullWorks_PGAB reaches for the laptop with the DOS version of BobCAD from 1992 to open a DXF file.
QCad is what I did it in...
OK, got it all together :) - http://wrlabs.shacknet.nu/~vw/MyMachineShop/tmp/FirstTryDXF2Gode/
I hacked stuff out of the gcode to get it to work. Need to change some things in the .dxf2gcode config file...
i am looking at this program and i am totally confused
Remove some things.
Welcome to my world....
do you just like
have a drawing
in mspaint or something
or is that the dxf
I think QCad will export the drawing to something useful, lemme see...
this program will crash your machine man
it's got you rapiding 1 inch through the part
and then tries to feed to the same spot?
maybe solidworks can deal with dxf
oh, acad can
exported it to a .jpg in the same place, that work for you OK?
nah i gots autocad
BRB, coffee buffer overflow...
what's with the weird bolt hole pattern
Err, that should be standard NEMA23 ?!?
just hand program it
i was just askin'
ds2: <blush> if I had a clue how it would be another story.
it isn't that hard
Some magic on the holes with that G2 stuff
if you are just doing 1 of these, it is easier to MDI it :)
Well, sigh, bear with me on this...
ds2: he isn't fluent in gcode
What I am trying to accomplish here is to be able to draw up what I want and make it :)
you ar doing full circles so G02/G03 is easy
My previous experience (10-15 years ago now) was with Moogs and I had a graphics program that spit out a paper tape of the (I assume gcode) which got fed into the reader on the moog.
Interestingly enough I wasn't allowed out of the office and didn't even know gcode existed until I started this little (heh) endeavor...
Well, I knew the word...
short summary: predrill so you don't have to plunge; g00 to center of hole about 0.1 above part; Feed into either desired depth or 1 diameter; G01 over to one side; G02 X/Y of the other side, I0.0 J(Radius of circle), once more time with the sign of J flipped
Sure appreciate that program now though.
compensate for tool diameter as needed and use G02/G03 as appropriate for particular machine
ds2: we've been through this though
write it out, run it in axis and if it looks good... run it for real... less pain then figuring out posts, etc.
he isn't familiar enough to do what you are asking
toastydeath: I am sure.... but I am procastinating... :)
So do you want to spiral pocket the center or just cut it
Just copied all that to my notes file...
IMO, the only time CAM programs are really useful is for 3D surfacing or doing parts with a bazillion features
Wow, I am embarrassed to say this, but "duhh, drool, I wanna NEMA23 mount", Sigh, life is NOT fun being this horridly ignorant...
As hard as I am trying to dispell that ignorance...
heh... actually I need to make one these too
About this time I read enough to think I know something, heh...
k`zan_emc: you need to throw your program out.
make a sketch up on a napkin.
the four holes, what size drill you need
Kzan - did you real want .27" dia mounting holes?
I already got the nappy skeptch :)
the center of the holes, diameter
and just figure out how you're going to get the machine to move there.
write it out.
Err, I think so, NEMA23 drawing is on other box.
or just manually do it
drill out the center to 1/2 or larger if you got S/D bits
then put on a boring head
ds2: i'm sure you are trying to be helpful but uh
toastydeath: what is wrong with manual for single parts?
a couple things: he's trying to learn how to make his machine move
um - all 4 holes are at different locations from the center bore...
oh... sorry.. thought the exercise was in getting a NEMA23 mount done
and doing it manually kind of defeats the purpose
i remember he was working on a CNC refit of a micromill
4 each 5.1mm for the mounting holes...
Well, at this point the PipeDream makes the uMill look like a bridgeport :).
since they are supposed to be symmetrical just give me 1 hole location and dia.
According to my conversion .20"
IIRC, NEMA23 motors used a #8 screw on the corner
Mounting holes are spaced 1.8559" (47.14mm)
really? hmmm I better remeasure
either that or I don't have NEMA23s on the taig :/
Lemme check the data sheet again.
SkullWorks_PGAB: I that with a nut?
Well, nowhere does it say NEMA23 but that is what the guy I got them from said they were.
And yes #10 screws fit just fine.
guess I better go do what I am suppose to do :(
Nut sits just so you don't need a wrench on it (stepper body fits the flat).
OK I grabbed a SEC Nema 23 - holes on it are.180"
center boss is 1.498"
Sounds right, I figured 1.5" was close enough not to have to pound it in with a hammer :-).
mounting holes are 1.875"
Seems a 1/8" end mill should have worked there...
If you need some place to paste - /join #vmmst
for the center - you going to pre-drill a hole or just mill a circle?
or spiral out
Possibly out of ignorance I was just going to mill it all with the .125 end mill.
Assuming for sake of this example, 1/8" 6061T6
OK - we do that - in 2 parts
Just for what it is worth, this is my first attempt at actually doing something on my own ...
this to be cut on your pipe dream or what?
That was the plan...
I assumed a while back that the hardware end of this was going to be the easy part :). Looks like I was right...
ok - limited rigidity - do it in 4 passes (hope thats small enough!)
now i don't know the machine but seriously?
Where I never got to here was once I had as much as I had to essentially duplicate the holes and move Z down appropriately between them.
four passes for .125?
Might be optomistic, really do not know at this point.
its a dremel machine
Well, a foredom, but yes.
Generic foredom :)
just giving Toasty an Ideal of your spindle rigidity
15K is good!
Until I put together a PWM of some sort to change RPM, but that is a way down the road at this point.
TRO on it sucks, but I got another handpiece that is much better.
also you might want to figure out the whole z axis thing
and setting tool length offset
I run 1/8" carbide at 24K - because I can't run the spindle faster.
that's a lot better than the 4k i run mine at =(
toastydeath: I'd ask you to convert that to newbie but we'll hold on that for now...
k`zan: tell your machine where the tool is on the Z axis
because you can't cut anything if the machine is way up in outer space
In the chuck...
yeah, but how far away from the part?
it doesn't know.
that's what tool length offset (tlo) does
I thought you set that with "touchoff"?
that's the command to set it, yeah
the technical name for how long the tool is is TLO, tool length offset.
I put a 1/4" gauge on the part and run Z down until it is fairly firm and then set 0.250 in touchoff.
so if you hear that said by folks, now you know what they are referring to.
Thanks. I have hope here, I am so damn confuzed at this point I must be learning something :) LOL.
don't worry man
you'll get there
Problem is, at this point, "Home / zero" is a very variable constant...
home is the machine's home point
zero is a floating value.
but to confuse things, sending an axis home is sometimes called "zeroing an axis"
Yes, it is wherever it is today. More or less the front left of the table, which for this thing may not be the best place, but hell, I dunno.
well you can move one of the work offset zeros to the part
like say, to the corner
or to the middle of the part.
As long, apparently, as I don't offend limits...
Which I blew out a lot since I kept running up against what I entered as really there...
What would be panultimately cool would be to make stepper mounts for the uMill on this and then give it to my roomie for the lapidary stuff as well as some engraving...
Just need to make x and y as rigid as I finally got Z.
Yes, this IS fun if frustrating at my level of ignorance...
I think I can..... I think I can.... I think I can.... :).
Minor crises, bbiaf
k`zan: if you are hitting limits
after setting your work offset on the part
you are doing somethin' wrong
toastydeath: Beyond ***ANY*** doubt :).
when that happens check the +/- of the numbers.
If you ain't screwing up something you ain't either doing or learning nothin' :-).
Ole Southern Saying(TM) :-)
it's better that you learn on your own dollar
my first couple major screwups were on the company dime
kzan you using the 1/8" EM as T01?
t2, if i recall
ok - and I started being lazy (using an Inc routine for the holes) but I figure that would jus add confusion... so I won't
SkullWorks_PGAB: EM? Just used the second entry in the default tool table becasuse it was 1/8" dia...
Damn inconsiderate of me to pile up my stuff where they kitties want to jump up :-).
Got it cleaned up though and in a better place now.
anyone here familiar with milling stainless steel
is dmess awake
he would know
303 is butter
I did tonnes of 303 for Synthes
i was facemilling some this week
dry I hope
and i am trying to bump by feedrates up
should i do it dry?
i've got stainless specific inserts
less thermal shock - better tool life
just never done it before
on stainless, anway
I often used Kennametal 725M for everything
is .013 too much per tooth, on a 4" facemill?
i lost an edge but i think it was probably thermal shock
i'll kill the coolant and see what happens
the inserts are rated from 200-660 sfm
i am down near 200
maybe i'll bump it up.
can maybe even go to .017 depending on depth of cut and available hp
20+ hp on a geared head horizontal
doc is light at the moment, we just wanted the job running
should purr right thru
it was purring, which lead to my confusion regarding the lost edge
may have just been a bum insert
that's what i'm thinking, it failed in a really weird way
like a little notch
in a nonobvious spot
cool man, thank you.
I've had batches where the cobalt binder was not uniform throughout and the inserts had very random life cycles
that would certainly give someone who has never done stainless before, like me, pause
it was so strange, it was just a little notch that was gone
right ABOVE the corner, which is what i expected to be missing
under a magnifier the coating looked like the measles sent them back for analysis and got full credit for the lot.
Folks, appreciate the guidance tonight, but I'm *beat* and am going to call it a night, catch y'all on the morrow.
gn es tnx agn!
i'm getting black on the inserts, is that normal?
I'll have code for you...
we do like, 95% aluminum and i have been running the slow/bigger machines now
with the steel and stainless jobs
SkullWorks_PGAB: Thanks so very much, I am just too beat to last much longer, got keyboard key prints in my forehead.
so it's a learning experience
with coolant its likely a sulfer component
run dry and you won't see that I bet
it was weird because it didn't look like BURNING
it looked like suit
the coating was immaculate
with a little bit of edge wear
no coolant would make my life simpler, too, since this machine has no enclosure.
just check the insert specs - I bet they show a preference to run dry - except maybe for finishing passes.
won't get soaked
i gotta go look them up then, all i have is the package and it doesn't help
just sfm suggestion.
been there - done that, burned out the wet/dry vac...
CNC knee mills great for small one off proto types - but messy
thankfully we don't have too many of those
i have a long way to go with this whole machining thing man
ty for your assistance
300 series tends to stick to the insert with coolant - and can lead to cratering - but the way the tooling engineer put it - cutting dry the chip floats on the coating like a drop of water on a hot fry pan - actually cutting easier
yeah, the solid lubricant coating
i am just trying to get over the "USE COOLANT" peer pressure
303 has Selenium added IIRC to help make it a more free machining
to be fair our other inserts are like, bargain basement variety
yeah it cuts so nice
so, so nice
don't matter - this applies to simple TiN coated carbide
i'm just a wuss
i'm mostly talking about the grade, not the coating
the grades we have are low hot hardness
you could mill dry but the insert will wear very, very fast
lol, they don't even have a brand
they're like "import" brand inserts
in paper boxes
somebody bought like, hundreds of them a long time ago and we're just now running out
and starting to buy real inserts
I don't have my notebook (with the secret feeds & speeds) but maybe a lower SFM with a higher chipload will carry the heat off in the chip and give you better tool life.
eh, i don't have to worry about those
the machine i am running gets good tools/inserts
i just have to figure out what the heck i'm doing, which is the real problem
those inserts get used on the aluminum machines, and they last for forever
We have some tube we turn - 1025 steel - I had the engineers buy me new tool holders to use Trigon (WNMG) style inserts - same price as the CNMG insert but now I get 6 edges vrs 4 - so it costs them less.
over a 2 yr time I've got tool life up from 7 parts per edge to 60 part per edge
lots of testing
just keeping logs and stuff?
used ceramincs dry for awhile - now back to a simple carbide base with a gee-wiz majic coating.
TiZAlNQRX coating that does your laundry!
what SFM do you start to drop the coolant?
there is supposed to be something even better coming - but I have not had a chance to use it yet.
or do you pretty much go by the mfg's recommendation
well many of these tests were done with free samples from the mfg
Our Iscar rep want the business so he plays ball and is sure we get exactly what we need.
"give us a box of inserts"
yeah - they just hand us 2 packs (20 total) and say give them a try.
that's pretty hot, we should try that with our rep
I run them at HIGH SFM too
i doubt we'll get anywhere
1900 rpm on a 2.25 dia
haha holy crap
i wanna do that =(
yup - small mushroom cloud when you open up the door
and that is with a big torit running
i hope i get to stay on this whole stainless thing
aluminum is kind of boring and not very interesting
"how fast does the machine go? well, okay!"
it's like there's real strategy and forethought with the steel
I'm having trouble finding the right place in the docs/wiki for a go/no go regarding my current configuration. If latency-test is giving be jitter results on the close order of 15 to 16 uSec, what's the likelihood I'll be able to drive a stepper directly at a reasonable rate?
as opposed to driving a stepper controller with just direction + "step"?
I'll idle here 'til I'm kicked off, and I can check the published logs if anyone responds after that :)
It _is_ nearly 2:30 AM Pacific time, after all.
I'll mention what I'm wondering in more detail, for the record.
[09:29:19] <mmgood> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TweakingSoftwareStepGeneration
this talks about latency, but not jitter.
The doc says "If your "ovl max" number is less than about 15-20 microseconds (15000-20000 nanoseconds), the computer should give very nice results with software stepping."
Well, cool. But how does the "jitter" figure reported by latency-test relate to the /usr/realtime*/testsuite/kern/latency tool output?
mmgood: it measures the same thing basicly
although latency-test gives slightly bigger numbers in praxis
hi to all, is there anyone?
hi i have an emc question
can i talk ?
just ask your question!
well, sorry for my bad english i hope you understand me
i use tkemc interface
i would use in g code some variable
the question is : how i can load a variable value from external program ?
you want to change the value of the variable as the program is running?
yes i mean that
what is your application? it might be possible with a HAL pin maybe?
no, whit hal i think no possible. Now i explain you..
first i running emc
second i load a g code whit a code, example a square
in a external program i load a variable , example x=100 y=200
the value will be in a #1000 and #1001
next i run the machine
you understand me ?
ok, and the x=100 and y=200 values will affect how bug the square is or something
and you want to change x and y while the program is running?
but this when emc is running and machine stop
in top of gcode program
isn't it possible ?
can't you just edit the g code file and reload it?
I'm not sure...
the g code program is a code to produce a standard form and
it is load by default when i running emc2
only four parameter need to change
AFAIK you can not pass a variable to a g code program
can you just place the parameters at the top of the file?
yes that is possible
but how can change they ?
no, from xternal program
a file editor?
or a dedicated program in pyton or c
I don't use tkemc so I don't know with that interface. In Axis you edit the file then reload it...
in tkemc is the same but i would not edit the file
in axis there is a pyton program
having custon M-codes returing values back to the g-code file where the call was made would be nice, I don't think it's supported (yet?)
that's what I understand too awallin
who produce a gcode and then load it in gcode windows of axis
however in 2.3 you can pass a variable to an outside file in the form of a call
you do a
o<myprog> call   I think
is it the emc version ?
yes the TRUNK
not released yet but you can install it
i have the 2.2.5 version
where i can find it ?
[13:20:36] <BigJohnT> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2
look at getting the source with CVS
It is the bleeding edge so YMMV (your machine may vaporize)
you still have to edit the calling file
isn't it a stable version ?
I'm using it ok
is this a machine were the operators are just button pushers?
yes it is
this is the problem
how can i pass the variable exactly ?
in that case I would write a python script to edit the file.
that way the operator only sees the 4 parameters on his screen
yes i mean that
change the number hit save and reload then go
can you program in python?
if there was something in g-code to read a HAL pin then that could be a pyVCP number...
now i go
awallin: that would be the best way but I don't know if it is possilbe
thanks to all and sey you later
see you later
awallin: that is an interesting real life problem...
tou can do something with analog inputs, but I'm not sure that would help
but could you insert the values into a g code file
I'm not sure, but I think the value goes into some parameter
like a 52nn parameter?
yes, but I'm not sure how/if that happens :)
* BigJohnT hears the breakfast bell
A M66 only puts into a 52nn parameter a value indicating a timeout or not
according to the docs, M66 E- reads an analog value to #5399 http://linuxcnc.org/docs/html/gcode_main.html#sec:M66:
I believe this is implemented in 2.2.5
jepler: I missed that part when I skimmed over the docs and skipped right to the end
Gamma-X: you there?
renesis_ is now known as renesis
good afternoon genius cnc'ers!
anyone ever try to share files between XP and ubuntu 8.04?
i had it working for like 3 seconds.. then nothing
start samba on nix
or share drive on xp
share xp drive is easier
well i'm at the point where i just did a ftp server on the win box... samba is getting uninstalled ;)
it's too touchy for me
one boot it works the next it doesn't...
that's why I said just share a drive on the XP box and connect to it from nix. No need for samba.
that uses the samba client
From nix, click on PLACES, NETWORK SERVERS, and enter in the ip/hostname of your xp box
smabd can be a pain to setup, but samba as a client is pretty straight forward.
there is nothing going on today, nothing at all
yes... i kept the nix samba browser, i'll play with it
i got stuff going on! i planted some sugar snap peas in the garden ;)
and i got xp running on a 233 mhz pII gateway with 384 mb ram.. and it's snappy! i'm surprised
it installed last night while i was sleeping ;)
i think it took like 3 hours
**was my old cnc control box... now to be my dedicated windows Cad/Cam box
cad/cam on a 233?
yeah.. the only thing that's slow really is loading huge dxf's
i've been playing with it all morning
**the computer, lol
I ended up sleeping till about 45minutes ago, some neighbors decided that 3am, was a perfect time to go outside and have a fight
i got some neighbors that are wanaba scrappers
I got to hear someone call one guy Eric, and all the other names I heard were exactly the same, maybe 10 people in total,
they tear apart dryers, ac units, refridgerators at like 8 in the morning in the driveway
with sledgehammers and sawzalls
... i don't they they eveer heard of a screwdriver
thats not bad though, at least the sun is up
lol, i guess ;)
i got a drunk neighbor who comes home at 2:30 am all the time revving his engine and yelling at the whole neighborhood in his front yard
**he told me he was gonna kill me one night when i told him to shut it
other than those two nieghbors.. the rest are awesome neighbors
BigjohnT, you out there? questions on your m5i20 and spindle setup????
what problem are you having?
Roguish: ask away!
thanks. just looking thru your website.
i want to use the 4th inputs on my 5i20 to run a spindle motor.
i see in the wiki you have done just that.
would you mind sharing your hal files?
I use the 4th encoder input for the spindle, but I use general purpose IO for controlling the spindle
yeah, that's ok.
it's hooking up the 4th encoder to the spindle signals and parameters and such I am trying to sort out.
ini and HAL are here http://www.anderswallin.net/2008/04/emc2-test-run/
that's not the final version but pretty close
looking and saving as files. thanks. it will take a while to read through. gotta start simple first and build it up.
ok, if something is unclear I might remember what it does :)
oh, i see in some examples, like the mazak, a reference to a [SPINDLE] section in the ini file. is that documented anywhere ?
i don't find it in the current docs.
maybe i have missed it.
If you define new stuff (a variable) in the ini, you can use that later in the HAL files. That way you don't have to hard-code calibration parameters etc. in the HAL file
by the way, your little boats are pretty cool. i used to race on large ones, 20' up to 80'.
80' ? big time racing!
so, i can actually add whatever i want to the ini file? (within reason)
spinikers were 4000 sqr feet.
where was this?
40 tons of boat.
hawaii, sardinia, san troupe
I raced on a 28' on Lake Erie for about 10 years - got to do one race on a 45' which was about as big as we had around here
for people with LOTS of money.
so, should add a [SPINDLE] section and whatever i wish to pass?
you can reference values in that section from your hal file
i don't like hard coding much of anything that could conceivably change.
like 'setp scale.0.gain0 [SPINDLE]SCALE'
Roguish: my HAL file has mostly hard-coded values, so perhaps it's not a model example of a 'nice' emc2 setup
i want the 5i20 4th axis to run a spindle. just on/off, speed/direction for now. then, maybe orientation.
do you have a VFD, or a servo on the spindle?
that's ok. the mazak demo has lots of softcoded lines.
right now it's a servo on the bench. yaskawa, just like the 3 axes motors.
ok, that should be pretty straightforward then if you already have the three other axes running
oh, yeah. my test setup has 2 m5i20's and is running 6 motors.
thanks guys. i willl try for a while. but i may be back for some more dumb questions.
I wish firefox would stop crashing
eric_U: which version?
dunno, trying to get it to start right now
i've been using the firefox 3 beta on xp and linux and it's working super!
here 126.96.36.199 dont crash
all versions since can't remember have crashed
friefox 3 beta with addblock plus and firebug
oh. its time to update
I just have addblock
firebug is the bug reporter?
it's for tweaking html/css code
or inspecting really...
what antivirus use you?
I've never had it crash on windows
this is a bit old
and full of problems
I am wondering why fedora is using it
try to disable powermanagemant
i don't think desktops should even have power management on them unless they were buil to run on solar pwoer or something
they probably should
i always turn it off when i first turn on a new mobo
why waste power?
i turn my pc off when i don't use it anyways..
just to test if its lead to problems
so when i am using it i want as much juice as i can get
you aren't going to get that though, the processors throttle w/heat
maybe I should build a kernel
I don't feel like updating the whole os
actually, just installed a new kernel, should reboot and see if it fixes anything
87 degrees / 52% humidity in dayton, ohio
Colorado Springs - Partly cloudy with a slight chance of thunderstorms. A few storms may be severe. High 68F. Winds S at 10 to 20 mph. Chance of rain 30%.
Natures Swamp Cooler working overtime.
too bad swamp coolers dont' work in Pennsylvania
Poplar Bluff, MO Breezy and 94
the stinkin' thunderstorms don't usually help much either
MO when reduced to 2 syllables is Misery. ( I had Family in Pacifica for many years)
what part of MO is that in?
Next door to Six Flags over Mid America - 40miles west of St Louis
in the big city
Country - 400acre ranch - breeding horses for jumping or what ever the "Horse" version of motocross is.
i'd like to see someone do a superman off a jump with a horse ;)
horse motocross that would be fun
Motorcycle side car polo - thats a real riot
that reminds me i haven't played atv offroad fury on my psp in a while ...
more fun than donkey basketball
but not as much fun as cow bingo, let me assure you
I'll stay clear of sports where competitors might kick dung up in your face thank you.
SkullWorks_PGAB: I wrote some notes on the spindle setup: http://www.anderswallin.net/2008/06/spindle-details/
there is another good sport... DungBall
Horses and I have an understanding - We each stay on our own side of the fence and everyone stays happy.
awallin: - Great - Thanks!
awallin; why the 1:1 ratio on your spindle rather than a 2:1 or something like that?
how often do you use low rpms?
i just checked out your site awallin
... what did you use for a base on your machine?
i went as far back in the posts as the first emc test...
i think i have hpv.... i thought that was dormant in males lol
Gamma-X: Um females actually.
or you said dormant, nm
lerneaen_hydra: now we have 0-5400 rpm with 1:1, more would be nice, but torque would be a problem with the 40mm face mill if we would go 1:2 to get 10krpm
oh I see
I don't quite recall but torque is about constant for an async motor independant of speed right?
chr0n1c: the machine is a heavily modded Opti BF-20, it stands on a base that's sold for a bigger opti manual mill, I think the BF-45 stand
i like the spindle setup....
it's got me inspired to build a mini version of it
lerneaen_hydra: if the vector VFD does its job right then torque should be ok also at low rpms
i'm using a stinky knockoff dremel tool right no for my spindle
higher torque at lower speed?
or just the same?
So its still using the Seig 7x lathe spindle?
ok right, that's what I was thinking
SkullWorks_PGAB: yep, the spindle is a spare part from littlemachineshop
lerneaen_hydra: but ofcourse you would halve the torque if we would go 1:2 to get 10krpm
i think i have hpv or herpes...
you want the P/N for the taper roller bearings for future ref?
maybe you need vagisil Gamma-X?
sure why not
awallin; yes of course
i hope its hpv not herps lol
Its actually a common wheel bearing...
the motor from a vaccum cleaner might make a good mini spindle motor... hmmm...
awallin; how have you found the stability of the base and main column of the bf-20 to be? my experience with it has shown it to be relatively weak, deflecting up to 0.05mm for modest forces
30206 Taper Wheel Bearing 30x62x17.25
how can you run a 40mm endmill and not get bad surface/tolerance
anyone know why my vmware server with ubuntu 8.04 starts up without eth0 sometimes and with eth0 sometimes? does it automatically disable eth0 if there's lots of traffic at bootup and it can't get a response quickly from DHCP?
VXB sells them here for $6.95 for bearing and race
lerneaen_hydra: here's the 40mm face-mill taking a 1mm deep cut into alu: http://video.google.com/videoplay?docid=-6222166294337898920&hl=en
you would have to design a seal as they don't have a rubber seal like ball bearing units.
how's the surface after the cut? smooth? how are the dimensions?
the surface is good provided the spindle is tuned carefully to be exactly vertical. that takes some time to get right...
I'm real fast at tramming the head on my bridgeport... that comes from doing it a lot :(
dimensions are within 0.02-0.03 mm mostly for small parts, maybe up to 0.05mm for bigger ones. It helps to keep the chip load down for the finish cuts. Sometime we run the finish path twice
awallin; yes, that of course assumes the spindle is perp to X and Y
hmm, that's pretty good actually
I may have to do the same - put liner bearings on the Z axis as the X3 mills tend to have head rock in the Y axis plain.
oh ok, to allow for springback
on my bridgeport cnc i could eyball the tram when i had to move the head and it would be within .005 ;)
my ex bridgeport*
I run a free pass or Spring pass on a huge Mori - Its often the tool flexing - not the machine itself.
awallin; oh, is that 0.02-3 mm after the finish cut or the roughing?
lerneaen_hydra: we only measure after the finish pass(es), roughing usually leaves 0.2-0.3 mm material for finish cutting
and I take it the roughing has never gone past this offset?
awallin: i just joined the monocam mailing list/group so i can follow along... i'm afraid i won't be much help other than beta testing though.
from deflection you mean? not that I recall. otherwise it would just be a following error
yeah from deflection
ok, sounds good
chr0n1c: there's been very little activity on the C# coding lately, but I'm in the process of coding it in python/C++ now so there should be something to test in a month or so
lerneaen_hydra: if you have the dovetail ways they need to be tuned just right for accuracy vs. friction
awallin; yeah, true
also the stock BF-20 spindle might have quite a bit of runout
mine's pretty good actually, under 0.01mm
i just placed a dial indicator against the spindle housing and pushed in some directions :p
* SkullWorks_PGAB just had another interior decorating inspiration - Replace the coffee table with a nice big surface plate - Question remains - Pink or black granite.
don't forget to reinforce the living room floor
only about a 4" thick
can a plain old potentiometer be used for things like jog wheels ?
Still have a 36" x 24" in the create bought when Maxtor shut down here.
chr0n1c: not really
hmm... does hal read voltage on pins?
i need to do some reading
analog to digital converter hardware is needed to read voltages on pins
(on physical pins)
the hal driver puts those readings onto HAL pins so you can do stuff with them
ok, that makes sense
heh... the difference between automotive folks and machinist... Just heard on the a chevy commerical... body tolerances as tight as 1mm... uh... errr....
sheet metal - bend to fit, paint to match
building dies for the sheet metal.. that's wher ethe fun is... ;)
where the best work is from the half retired cranky guy ;)
I got the cranky part down - unfortunately the half retired part is a ways off
anyone have a sec for dumb question about setting up xylotex board?
rebooting to stock kernel... i miss my sound.. cya!
there's a gigantic HMC on machinetools.com
i am going to put it on my christmas list
jmkasunich_ is now known as jmkasunich