I use the gcode ulp from here. http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
pcb-gcode is popular
Casainho: skunkworks_ let me see...
sorry for the delay....
ries: I had used pcb2gcode on KiCad gerber files
Casainho: what about the optimize utility I have been reading about?
ries: well, it's gerber files, should be equal for every PCB cad
ries: optimize utility? I don't know what are you talking about
ries: I don't know. I don't use Eagle and that piece of software seems just for Eagle
Casainho: ok, may I ask what you are using?? Last time I used something it was ultiboard and ultimate.... before that orcad :)
ries: 1st: I just use Free Software, 2nd: I just drilled one small board until now :-)
ries: the schematic/pcb program desgin I use is KiCAD: http://kicad.sourceforge.net/wiki/index.php/Main_Page
it can even make 3d renderings of parts
hows its part library going these days? last time I looked it was pretty empty
Nhaaa.. although it sounds cool, I don-t need that for my resitor and optocoupler >(
Ooo crap.. debian doesn't have gerbv dev :S
amazing... got it to work :)
ries_ is now known as ries
found a newer computer that runs emc nicely... ASUS P4P800-VM - latency about 12000... not bad as far as the P4 boxes go
[04:01:02] <LawrenceG> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Latency-Test
can somebody answer me about emc2 compiling?
giovanni; just ask
* frallzor strokes his electronics box
[12:24:21] <frallzor> http://pici.se/p/large/LujKdTUPh/
quite roomy I must say
frallzor: good day dude....
it get's let roomy over time....
SWPadnos: you snowed in up there? or are you down here?
it's snowing a little here, and it may pick up a bit before tomorrow
I'm more concerned about Chicago, where they're supposed to get 3-6 inches in the next 24 hours
yah - we are supposed to get about that also.
is it falling there yet?
Guest450 is now known as skunkworks
no - it snowed a little last night (about an inch)
ok, nothing a northern town can't handle :)
heh - hardly had to slow down at all. ;)
just rain here. quite a bit so far
I was impressed by how little we were affected by 20" of snow this weekend
then again, the PA turnpike was a disaster
I think the ubuntu team screwed up my audio again
I guess I should pay attention to what they are updating
if it makes you feel any better, the windows team screwed up my video
good evening.a question please.on lathe with emc can thread with more start ?
yes you can cut a multistart thread by offsetting the Z start
cradek, thanks please have a example the program ?
just use one g76 for each start
no problem, thanks for help
for 2 start, begin the second one over half a pitch
ok i understand , more thanks
Is there a way to command stepgen incrementally?
is there a nice demo that can be used to show off EMC on the EMC live CD?
ds3, you mean a simulation, not a real machine?
terrylm: yeah... a local club is showing a demo of Mach3...figure a EMC demo might be in order ;)
the EMC logo default doesn't seem to work well in the default simulator
ds3: what do you mean?
cradek: it tells me it exceeds the machine travel for one
this is just starting it up. I hit the power and release e-stop
you have to set a work offset
ds3, you could run your own part file in one of the simulations.
it even says this at the top of the gcode file
that's what they mean
for some reason I thought G54 offets were optional in the simulator
terrylm: my parts are not too interesting
will try that again
Anyone know the details of stepgen.n.position-cmd/fb? They seem to be absolute values, is there a way to use relative values?
sorry, relative to what?
incremental mode like.
relative to where it starts from.
can you back up and say what problem you are trying to solve?
My tool turret uses a stepper motor, which is commanded from classic ladder. I would like to tell stepgen to go forware or backward from where it currently is by n tool positions.
since the tool number is not incremental, what are you doing to make this an incremental issue? It seems to me that the tool number could just be sent into stepgen as the target position.
I had attempted to add/subtract this value from the feedback value, but that puts some of my compare statements in classicladder over 20 characters long and it seems to have a buffer overrun bug ...
Well, if one is on tool 1, and wants tool 8, it only need go back one position, not forward 7.
It is an 8 tool turret.
I see - if you want that, it does make it harder.
This must be a common issue on lathes, how do others do it?
my lathe's turret only goes one way. it has feedback that shows what position it's passing. when it hits the right position, ladder stops it.
what about that modmath hal component?
I suppose you could do the same thing if you run stepgen in velocity mode
works great for my mill's toolchanger
modmath? will look it up...
stepgen's feedback position -> modmath component or % expression in ladder -> logic to see if it's in the right place yet
terrylm: there's an example in the demo-mazak config
although it isnt in the most current format so it was a little difficult for me to decypher
yeah that's getting dated fast
i guess i can clean up my hal and ini a bit and put that in the wiki or something.
for the modmath stuff
there is absolutely nothing on that component in any documentation really
except for man modmath but it's a little vague
I was just looking for modmath, found nothing, will look at the man page.
yea, there is nothing except the man page
hmm i need to buy/make a chuck backing plate for my hardinge
i got a 6" 3 jaw chuck in a step chuck right now, and it's making me a little uneasy
I measured the shank on a 1/4-20 tap yesterday and it was about .254. Are these standard shank sizes and if so where can I find a list?
woohoo - Got my arduino and wings today!
Well, off to work I go... Thanks all, talk with you latter.
looking here: http://www.mcmaster.com/#2523a411/
it doesn't say the shank size
from the picture it looks like the shank on this one would be way under .250
some are made to clear the thread
i have about 6-8 different brands of 1/4" taps and most were .254-.257
I'm trying to figure out what ER40 collets would hold them - looks like 7mm
If that's 1/4"-20, the shank would have to be smaller
that's just a generic picture. i bet 4-40 looks the same
danimal_garage: thanks that helps a lot - do you have any 10-32?
I think that "nut taps" are made to clear the thread.
and 4-40? :-)
are you guys looking for shank diameter?
cradek: i have about 14 years worth of taps
so yes lol
.192-.194 for #10
Have you seen the special tapping ER collets that Hardinge do?
#6 and smaller is .140-.141
andypugh: no - do they have a square in them?
#4 is also .140? that's sweet if so
They have some with two setscrews, and some that incorporate float for not-quite-rigid tapping
all the way down to 0 in my drawer is .140"
we just found a 0-80 tap...
looks like 4, 5, 7mm will hold all the taps I regularly use then
0-80 is small but tps ok
i'd rather use a 0-80 than a 6-32 lol
seems like I use #4 or #10, never 6 or 8
i use 8-32 alot
i just found a 1/4"-60 tap
Page 4 at http://www.hardinge.com/usr/pdf/collet/2351.pdf
mcmaster has 1/4-80 but not 1/4-60
this one is 1/4-60
Number 10 Whitworth Instrument is 0.01 dia and 800 tpi
wth you gonna do with 1/4-60 ?!
Sorry, 400 tpi.
andypugh: neat, thanks
Jymmm: i used to work for a place that made automated inspection equipment... they used all sorts of odd sizes
probably for fine adjustment of something or other
they used a #12 tap alot too
I once spent ages looking for a pair of threads that would fit inside each other, with the same pitch, where I could get a left-hand variant of one of them.
The application was a little self-centring vice. (for holding microchips while they were being mechanically tested)
5/16-24 and 3/8-24?
I think it was, yes. I needed to use a slightly oversize tapping drill and an undersize thread.
But the design is actually in production.
i had a 3/8-28 tap once. i could never figure out why they would make one so close to the standard 3/8-24
Self-locking nuts :-)
Tonight I will see if my milling spindle can go slow enough and my rotary table fast enough to stay in synch for rotary hobbing. Otherwise I am tempted to make a dedicated hobbing spindle.
andypugh: sounds interesting. Video!
sorry for sucking, but what is hobbing?
[20:22:26] <andypugh> http://www.youtube.com/watch?v=Yulhx_Z0ZUI
(That is how it is done commercially)
Hobbing starts at 57 seconds
My half-a-plan is to wire the axes together in HAL, so they stay permanently in synch at the set ratio.
how different is hal from something like a driver for a webcam or sound card?
whoa thats nuts
I can do HAL, I wouldn't know where to start with a driver.
[20:41:44] <cradek> http://www.youtube.com/watch?v=x2Hwe6-EXug
I have never figured out how they do the double helical gears without a gap in the middle.
andypugh, endmill the path
I have a david brown book with a picture
OK, inside David Brown is where I have seen them. In fact the reason that I want to hob my gears is that my dad brought me up with the concept that form-milled gears are an abomination.
I am pretty sure my dad doesn't have an opinion about that
i used to work in shops that did alot of hobbing and gear cutting, but sadly i never got to run one of the machines
generated gears rule
yea, my dad can tell you how to fix a buick, but i dont think he has an opinion on hobbing either
some forms cannot be generated though
[20:49:12] <andypugh> http://www.zakgear.com/WN.html
clock form is one that cannot be hobbed unless you use the modified form with round root
andypugh, for interests sake one day I wrote the gcode to fly cut a generated involute form using rack form cutter
and for the double helical here is a vid http://www.youtube.com/watch?v=FkzqwIjjBhs&feature=channel
he spoils it at the end though
gear to what comment?
video should be explanitory /me has to go awol for half an hour
Does anyone now what the number-of-channels specifier is for the "invert" hal function?
I have tried "num_chans" and "cfg" next try is "count"
(It is "count")
Grrr! that last fauliure didn't unload hal properly. Time to restart
yyhgh; op.,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, iooooooooooooooooooooo///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////]\['?""
my son's first chars on irc ;)
not his best performance
well, he's 7 months old..
well to be fair, I've seen more inane things on irc
I have typed more inane things
And my conclusion is that I probably can't spin my rotary axis fast enough to follow the spindle.
it's pretty much the same what he's saying right know
so I think he got the spelling right
So plan B is for the hobbing tool holder to incorporate a big pulley to be run from an auxilliary motor.
(then I will run with the milling head in neutral)
I have a question about computer requirements.
What do you have?
Homebuilt, 700Mhz PIII
EMC is not demanding of CPU or memory, but the realtime-latency is important
works ok, get about 15000 for latency figure
That's pretty good.
Should be fine
question I have is about video cards -- with linux I've always been a fan of nVidia with the restricted drivers...
should I NOT use them? The Getting started manual suggests not?
If you get OK latency then no reason not to.
Ah, okay, perhaps I'll try the restricted driver. So the test is, does it hurt latency?
Did you run the latency test with glxgears running fullscreen?
Exactly. As long a your latency is reliable, nothing else matters much
no, ran lots of other stuff -- including recursive searches for files (as I know my hard disk is old on that machine) but don't have glxgears installed.
Booted off the LiveCD?
I thought it was part of the standard EMC install
I booted from it and installed it directly. Didn't know glxgears was even on there ;)
I think EMC is amazing, what a piece of work.
just try typing the word in a terminal.
sure I'll rerun latency test tonight with glxgears running.
For the moment, it looks like it will all work fine.
great, thanks so much. First timer here with machine control, but strong on Linux and electronics, microcontrollers, etc.
Mainly need machine to make PCBs and small robotics parts, as my designs don'tfabricate well anymore with FeCl :)
Sounds like you have the difficult bits covered.
yeah, it's a blast. So much to learn! thanks for your help, I'll give it a whirl again tonight.
I often see a latency spike when starting a GL program but after its started theres no problem
with the nvidia driver and a SMP kernel
do you have multi cores cpu
Valen: If you are talking to jhelmick, he has already gone..
yeah, I know, I was just saying
acemi I used to have an actual twin CPU machine, dual Xeon 3.06
now I am using an intel atom board
is the last cpu/core i isolated
My EMC machine is a dual Xeon.
ours got hit by lightning
dual cpu or dual core?
Mine? Dual CPU.
Old-style Xeon .
on the xeon I was getting 2000 latency, on this new atom I'm getting 4000
hmm very nice
It's a 1U server, SE7501CW2
I get 6000 latency, which is more than adequate
Made in 2003
oh that was with the nvidia binary driver and a 6600GT video card
so axis ran really smooth with a large tool path
1U might make things nice though ;->
mine was a full ATX size
Yeah, the Computer, drives and VFD all fit in a 19" rack that I made out of extruded ali profile and MDF
Now, I can think of three solutions to the fact that my rotary axis can't keep up with the milling spindle for hobbing.
can you just slow down the spindle
1) Make a faster rotary axis. I have an ER32 chuck on a ground 3/4" shaft, so that could be as easy as a pair of plummer blocks, a pulley and one of my steppers (or even a servo)
2) Slow down the spindle by lifting the motor and inserting a secondary reduction.
3) Make a hobbing cutter holder that incorporates a pulley, and drive that with a motor mounted somewhere else, through a cobbled-up reduction.
4) Just get on with form-milling the gears.
Incidentally it was quite good fun having the lathe-spindle encoder output wired directly (in HAL) to the A-axis position.
whats your rotary axis now?
The new rotary axis is probably the best solution, and the most work. It would save me figuring out a way to tilt my existing rotary axis, though.
It's a stepper-converted rotary table.
ahh, just convert it into a servo driven one?
see how easy that was to say
I am not convinced that my servos have any more "oomph" than the steppers.
And it would involve rather more HAL and electrical work.
I ought to try it, I guess. I have the parts.
servo will likley have a higher speed though
(Well, I would need a new coupling spider as the serve shaft is smaller than the stepper shaft)
[23:08:55] <alex_joni> http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/catid,38/id,1813/
I hadn't realised that Axis had UI flaws until I looked at that :-)
UI toolchange is nice.
no pyvcp is a drawback.
(or is it an option?)
andypugh: you can keep pyvcp as a standalone program
My pyvcp is just 5 text boxes. I use them in lathe-mode for target-x, target-z, surface speed, cut and feed. I then have individual G-code programs for turning, boring, facing, grooving and threading that use those inputs. It works quite well.
I am not sure that switching the rotary to a servo will help much. The drives top out at 2500rpm, and I think that my stepper is currently happy to 1800rpm
(The actual servo can do 6000rpm, I would need to fiddle with the driver board components to get that though, and I don't understand them)
understanding is usually cheaper than silicon ;->
[23:31:39] <Valen> http://www.vapourforge.com/russell/mill/rotary_table/11012010455.jpg
is soon to be ours
just needs to be connected to emc
I can buy silicon, I can't buy understanding. My brain is full.
2048 line encoder > 100:1 harmonic drive > old rotary table
sorry 400W motor in there too ;->
That motor looks to be right in the way....
Well, tricky to use it as a C-axis
which one is C?
Parallel to Z
ahh its for a horisontal axis
dad made http://www.vapourforge.com/russell/mill/clamps/05022010509.jpg
the other day
must be the sexyest clamp heads I've ever seen lol
they came out really nice, lovley finish ;->
Nice, but why?
the plastic ones shattered in about 4 seconds
I suspect you might find the second-weakest part in 16 seconds.
the handle part seems not too bad
Keep an eye out on eBay for some of these http://www.lawson-his.co.uk/scripts/products.php?cat=Carver%20rack%20clamp
(but note how very quickly they get horrily expensive as the size goes up)
good thing about these ones is they are fast
Those are pretty fast, and strong.
[23:47:51] <andypugh> http://picasaweb.google.com/bodgesoc/Gibbs#5397761057273088882
(I was inexplicably proud of that setup)