Houston, we have ignition!
Hi all - I thought I would share my latest little machining video, which makes mention of my use of EMC2
[04:06:33] <gfixler> http://youtube.com/watch?v=KI9_q2Y59_I
it was about 6 days of actual time on the machine, intermittently when I found free time after work
I'm still learning
getting faster with each little project
I'm also overly careful, so my loops will run for hours, just to trim something down in size on the lathe
I'm getting a bit more bold there, too
I've been machining (hobby, not pro) for 10 years, and I still didn't really start to get a feel for what the machine can do until I started doing CNC
how long without cnc?
I definitely wanted the cnc stuff right away
I got the machine running over xmas break
so I got it with the mill
steppers, or servos?
a shoptask 3-in-1
[04:14:54] <jmkasunich> http://jmkasunich.com/cgi-bin/blosxom/shoptask
whoa, they're pretty big, no?
about 600-700 lbs
big compared to sherline, small compared to bridgeport
it has so much speed and power in the video
compared to me
I can lift the sherline with one hand :)
they are handy that way
getting the shoptask in the basement was a project, and getting it out again is something I don't even want to think about
they make things for moving very heavy items, like gun safes into, and out of basements
nice oops on the boring head
very nice demo vid
then I read how they're supposed to be used, and felt pretty dumb
did you make that on a linux box?
Ubuntu 7.10, with Kino
I have to remember that
it's very non-intuitive, but once you get it, it's very handy
the help files were a must
and thankfully, a pretty quick read
the vid you saw is my one and only vid
I enjoyed it
video editing does rob me of more time to design, and code things for the mill, and lathe, though
nice fit on the part
yeah, I keep wondering how much of that was luck
did you install just Kino or the whole studio?
I have another machine with Ubuntu Studio on it, though
but it's mainly for recording work at my piano
that's a tiny Shuttle KPC
which I rate an extremely good deal
there's a Flickr set of the 7 days of work on the mill vacuum stuff
[04:22:12] <gfixler> http://www.flickr.com/photos/garyfixler/collections/72157604713088268/
actually - a collection of 6 sets
now that I look at it, it was more like 4 days on the machines
lol - augers
goin' for the helicopter of death there
I like to learn how to use machinery by endangering myself :)
did you use Maya in getting the gcode?
yes, and no
I took measurements off the maya mockup, but wrote the g-code in a text editor by hand
ah, just for dimensions, cool
I have used it to generate g-code for me
very nice, very pro ( keep your fingers in your pockets ;)
with MEL, it's embedded scripting language
thanks, tomp :)
jmk: what's the precision on the shoptask?
either 20000 or 40000 counts per inch, don't recall at the moment
the machine itself can work to 0.001 if you are carefull
10x microstepping (gecko drives)
40k per inch!
10 tpi screws, 2:1 timing belts, 200 step/rev motors, 10x microstepping
I'm planning to replace the screws with 5tpi ballscrews
are the threads precise enough for that, or is it all theoretical?
I've got the same resolution on my Bridgeport
1000CPR encoders, 2:1 belt drive, 5TPI
oh, I see - the machine drops all of that to 1000ths
slop in the ways, screw error, etc, mean you don't get anywhere near that resolution
if I'm quite carefull I can turn a press fit, but I can't just write a program and have it come out that good
I'm not sure what the microstepping of the Sherline driver is
I have to get close, measure, and then make a finish pass or two
I was fascinated to learn about that
the fact that a human can achieve better results than a machine
my whole life I thought machines were perfect, infallible, and amazingly accurate
heh, then you started using a machine tool
and I wondered why handmade cars were ever considered better :)
remember: the first tools were made by hand
yep, that blows my mind
it takes a brain to improve accuracy. machines can't do it themselves
I wouldn't say a human can achieve better than a machine - a human with a machine can do better than a machine alone
the only inkling I have of how has been reading up on the three-table method of creating a flat surface
remember hands made the first tools
heh - yeah, that's a cool trick
I wish I knew more of those
the Moore book on that kind of stuff is fascinating
I keep wishing some science channel would do a history of accuracy, precision, straightness, flatness, and related things
I'm going to look that up
I think its called "Foundations of Mechanical Accuracy"
its an expensive book if you can find it - I got it on interlibrary loan
I've scoured google a bit for such things, finding little bits here and there
the only mildly related text I have is the mechanical movements book
[04:35:52] <gfixler> http://www.amazon.com/Five-Hundred-Seven-Mechanical-Movements/dp/1879335638
great for ideas when building machinery, robots, etc
the Moore book is probably from the 60's or so - it partly reads as a boast about how good Moore jig borers are, but its also very interesting
[04:37:09] <jmkasunich> http://www.mooretool.com/publication.html
wow, that is pricey
do you feel reading through it has made you a better machinist in any ways?
thats why I don't own a copy
it inspired me to try hand scraping
which in turn taught me that hand scraping is a LOT of work
nothing has inspired me to actually try that yet
I did some on the Shoptask about a year ago
the ways were tighter at one end than in the middle
I was able to improve it quite a bit, with two days of hard work
that's good to hear
[04:39:47] <jmkasunich> http://jmkasunich.com/cgi-bin/blosxom/shoptask/scraping-12-17-06.html
I have a nice table saw with a miter channel that changes width
[04:40:08] <jmkasunich> http://jmkasunich.com/cgi-bin/blosxom/shoptask/scraping-12-18-06.html
which meant I couldn't use a miter bar with precision wheels, because it pinches on one end
hmm - a channel would be tough
I've been imagining a hone stone of some sort
worked back and forth in the channel
it really comes down to two steps - measuring the high spots, and removing them
woodpeckers just sent me spam on this thing today: http://www.woodpeck.com/sawgauge.html
you could probably remove them with fine silicon carbide sandpaper
precision ground pins
it answers a need I've had for awhile
I want the inverse
to measure the actual channel
something similar to that could measure width
straightness is a more difficult challenge
yeah, and I think straightness is the problem
I have a feeling the channel is about the same width, but bends off to one side at the back end
what's the closest thing you have to a precision straight edge?
clamp it down, and run the woodpecker in the slot, with the indicator plunger bearing on the straightedge
coincidentally, woodpecker's straight edge: http://www.woodpeck.com/sse2.html#1109
that's a good idea
I did use this to check the straightness
and was able to lightly rock it on one side, and not on the other
adjust the straightedge till the pecker reads the same at both ends, and the reading in the middle will tell you how curved it is
I'm going to do that
I could make a small, precision holder for my dial test indicator
one that runs along the straight edge
and could then see exactly where, and by how much it's curving
then you could indicate both sides of the slot, one at a time - and get both straightness and width
and then lots of work to straighten it up
depends on how far off it is
also coincidentally, I just bought some dykem from a small machine shop nearby
not for a reason - just 'in case'
the evil blue goo
yeah, I'm afraid to open it around anything I like
just open a fresh roll of paper towels and a bottle of some oil-cutting cleaning fluid first
but keep it away from your woodworking stuff - you sure don't want blue spots on wood
that'll be tricky
the entire barn is woodworking stuff
your surface plate - what material is it?
duh - I thought the "woodpecker" was their name for that saw gage- its the name of the company
does sliding it around damage it?
12x18, got it surplus for $50
littlemachineshop.com has a sale on a tiny one
not if you slide it on something that is already reasonably flat
[04:48:36] <gfixler> http://www.littlemachineshop.com/products/product_view.php?ProductID=2882&category=
more a stand than a plate
but pretty cheap at $30
I think enco sells pretty inexpensive surface plages
wow, 12x18 for $25
[04:49:30] <jmkasunich> http://www.use-enco.com/CGI/INLMK3?PMK0NO=909662
I'd have to read up to see what that means
they don't say which granite
I believe black granite is the best
the enco ones are black
hardest - least porous, so it doesn't pull in dirt, which can grind up the surface over time
I think when you are talking $25 plates for home use it doesn't much matter
blacks' most common, i think pink is harder, and more $$
ah, B is +/- 0.0002
yeah, I figured - it's in the specs at the link
that's still 5x tighter than my mill can achieve
of course I had no idea what grade mine is - it's twice as thick as the enco plates, with ledges
I had it reconditioned for I think $75, they said its now +/-0.000050
Hi folks, installed the latest(?) EMC and it does fine off the ethernet, but I will be using it with a wireless adapter (just arrived today) and apparently there is nothing to do with wireless in this install, should I just start over (nothing much done so far)? TIA
SWPadnos_ is now known as SWPadnos
5 *hundred* thousandths?
I think I'd want to build a glass display case for something that accurate
its funny, by pure coincidence it was reconditioned and certified exactly 31 years after the date on the original cert tag
traceable to NIST, the mother of EMC :)
I guess 50 millionths is more accurate than 5 hundred thousandths
because it adds in that extra 0
K'zan - how is EMC using the ethernet?
I got to watch them do it - it was interesting
grinding wheels, right?
? :) i just dont think in hundred thousandths, both are same thing tho
they flipped it upside down on a much larger plate, and slid it back and forth with diamond dust between the plates
slide, turn, slide, turn - by hand
but you'd have to say 5.0 hundred thousandths to really mean 50 millionths
to stress that extra digit of precision
then take it off, clean, and check using a very sensitive dial indicator and a precision straightedge
what makes their original plate so incredibly flat?
just checked the tag - its actually 60 millionths, not 50
where's the divine mover? :)
I think its partly that three-plates averaging thing
oh, well that's no good then ;)
I wonder if there's an XY machine with a precision reader, and grinding tip
that can run around checking every point, and grinding it down
if you slide one thing on another thing, constantly rotating them so the pattern is nicely random, the two things wind up matching
right, but you need the third table in there
or they may perfectly match a spherical shape
they generate flats by using three
I wonder if three run against each other in that way will automagically flatten all 3
although there are techniques to do it with only 2
so you don't actually need to check anything
1 on 2, then 2 on 3, then 3 on 1, etc
that's pretty impressive when I think about it
the moore book goes into that kind of stuff
using randomness to create very non-random things
moore does it by scraping three cast iron plates to match each other
I'll just copy my emc directory over to the server and reinstall and see what happens.
I bet the $25 surface plate from Enco needs $40 in shipping
if not more
still, I had no idea I could get a surface plate for such a low price
I've had in my head they were likely hundreds of dollars
ones that aren't made in China are
is Enco a kind of Harbor Freight?
just about cheaper than floor tiles!
but a grade B is plenty for home use
yeah, maybe a notch better
they're mostly mail order - I wonder if they have any stores
the cheap surface plates are only really cheap if you can go pick it up yourself
the smaller ones by enco qualify for free shipping
I've held off on certain bits of precision, because they were either too expensive to consider, or too cheap to believe
get a plate, make a nice cover for it, and it will last you forever
I probably only need something this big: http://www.littlemachineshop.com/products/product_view.php?ProductID=3068&category=1438824943
I can only make parts 4"x8" on the Sherline
and that's pushing i
for the sherline mill & lathe , a 6"x6" might be enuf
I took a piece of masonite and put a quarter-round frame on around the bottom edge - it sits on the plate when I'm not using it
yeah, though I like the 9"x12" to have room for tools
what are the ledges?
it's 4" thick for most of its length, but for 2" at each end its only 2" thick
ah, so it can be sunk into a table?
so if you wanted to clamp something you could get the clamp underneath
oh, I see
never used the edges on min
I want to find a few more uses for a surface plate before I invest
I still don't know how necessary one would be for me
I have no personal experience with Little Machine Shop, but from what I've seen, you can do better at regular industrial suppliers like Enco, MSC, etc
I'll be sure to check them out, and compare
there are some things that are specific to small machines that only LittleMachineShop will have
but for example a height gage is probably less elsewhere
good to know
I also seem to keep coming up with ideas that need precision ground pins
so I need to source those
dowels - mcmaster car
wow, half a million products
you name it they got it
I have some 1/4" dowel pins here (1-3/4" long) - they measure 0.2502
+/- 0.0001 which is about as well as I can measure with my best instruments
I don't know if they're made, but I was wondering about a kit, like a gage block set, of just every measurement
0.001, 0.002, etc
yeah, they are made
for quickly checking hole sizes
trying to get an ID with calipers is next to impossible for me
however unlike gage blocks that you can stack, if you need 0 to 1.000" by 0.0001 that is 10000 pins
(only 36 or so blocks)
the other thing I wondered about
some kind of caliper-like thing that would let you widen a pin
you could measure it with calipers, or a micrometer
and turn something to widen it to whatever ID you needed
widen a pin?
you lost me there
like an adjustable reamer?
in a sense
that will widen a hole
more like 4 panels
widening a pin means adding material to it
itself gets wider
that can widen out from a center
the reamer gets wider
I don't mean actually widening a pin
this would be a device
that could maintain some sense of roundness
but be widened
there could be intermittent space around it
I'm trying to think of an example
like a 3-jaw chuck, but in a reamer configuration
yeah, I suppose
like a honing tool for a cyclinder bore?
like so: http://www.use-enco.com/CGI/INSRIT?PARTPG=INSRAR2&PMAKA=334-1114&PMPXNO=946759
the idea would be that it would have 3 points of contact
are you talking about just _measuring_ inside a hole?
ah, does that push the sides outward as you tighten it?
bore with the lathe, back out, and test its size quickly with pins
are those the t-handle deals?
pins if you have a limited range of sizes that you need - else the 10000 pins problem
small hole gages for under 1/2"
they expand inside the hole, you lock them in place, then measure with a caliper/micrometer
yeah, I think I know them
telescoping gages for 1/2" to 6"
just saw them in a glass case, too
and inside mics for over 6"
oh, I don't know the small hole gages
I know the telescoping ones
lots of pins :)
is that what it is?
because that was my idea - I just wanted pins of every size :)
[05:14:26] <jmkasunich> http://www.use-enco.com/CGI/INSRIT?PMAKA=615-6610&PMPXNO=947860&PARTPG=INLMK3
pins are so much better then a telescoping gage
cheap bore gages contact on 2 points, are a split ball with a screw. put into hole, adjsut screw till it almost binds, withdraw and mike
yeah, those are the ones I saw recently
I couldn't imagine using the telescoping gages very easily
an intramike is a reasonable compromise
[05:15:00] <jmkasunich> http://www.use-enco.com/CGI/INSRIT?PMAKA=615-6250&PMPXNO=949268&PARTPG=INLMK3
ds2: what is an intramike?
let me find a pix
gfixler: you are right - telescoping gages are hard to use
oops, those things were small bore gages at ENCO
I wanted something with 3 points of contact, but with some length to them, so they'd hold the pin vertically in the bore
I do fine with small hole gages, but I suck with telescoping gages
so you didn't have to worry about if you were in at an angle
maybe 2 sets of 3 points
hmmm no good pix
at different depths on the pin
and you could turn something to push them all out
in micrometer fashion
the problem with 3 points is that you can only mic it if it has two points directly across from each other
jmkaunich: they are 3 point contact micrometer for measuring the inside diameter
and then pull out, and measure that
but 3 points is a pain, because it's hard to caliper
you turn it and it extends the 3 contact points
you are describing bore gages
unless you do something like the 3-wire method, with tables, or calculations
but aren't those 2 points?
intramikes are direct reading :)
the problem with bore gages is again, limited size range, and high price
on a T handle?
there are two and three point versions
2 seems really difficult to get just right
if you have the rod at an angle, it's going to give a diagonal reading
thats the skill factor
which is why I suck with telescoping gages
no good pix for an intramike
I don't want skill :)
I want automation
3 would require skill, too, unless you had another 3 higher up
they are nice for boring unlike the @#$@$@$#@$@ telescoping gages
then they'd hold the whole deal perpendicular
I just got one of these: http://www.emachinetool.com/tooling/temp_top_three.cfm?FamilyID=S305185&Source=PTC
you still rock it around in the bore to get it perpendicular, but the indicator tells you when you are there
do you feel you're really nailing the reading?
look for the low reading?
much better than I did on telescoping gages
I can see that
when rocking "in-and-out", look for the minimum
see, my concern is in getting a pin that mates a certain way with a bore
gfixler: how many different size bores are you interested in?
you can have a 3/8" shaft that doesn't fit in a 3/8" bore
though they both read 3/8"
if you do a lot of the same size, pins is the way to go
because one is 0.0001 off
that's the biggest problem
likely I'll need whatever size pin I don't have :)
[05:21:00] <jmkasunich> http://www.use-enco.com/CGI/INSRIT?PARTPG=INSRAR2&PMAKA=890-9739&PMPXNO=5215207
I can certainly imagine constraining myself
press fit, shrink fit, slip fit.... you just need a box of adjustable holes
e.g. - maybe I just design with even numbers
most everything I do will be my own designs - not trying to match something else
if you are designing stuff, you could probably limit yourself to a half dozen sizes
3/8, 1/4, 3/16. 1/8, 1/16, 1/32
get ordinary steel dowel pins
I've thought of machining my own pins
mic them with the best mic you have
but you need to precision grind, really
they will probably be a tiny bit oversize
and I don't know how to do that
yeah, I have aluminum bars here in some of those sizes
and they're all a bit over
then polish off 0.0001 at a time with fine emery
that's a big skill thing
polishing evenly along the length, and around the circumference
you could make yourself a set of dowels, one is 0.0002 over nominal, one is at nominal, and one is 0.0002 under
wow, my digicals only go to 0.0005
spin it in the lathe - that will make the circumference problem go away
how then to keep the measurement along the length, though?
I've been wondering about that a bit recently
calipers aren't good to much better than a thou - the jaws bend, and you can't control the pressure consistently
I have been looking for an excuse to buy some calipers
I mean a micrometer
this would be a good reason
reaming with water versus oil gets you .0001 diff, enuf to make one snug and one slip, there's lotsa tricks ( even a shim along the reamer side )
that's an impressively fussy difference
tomp has a good point - if you are interested in making precision holes to fit pins into, don't worry about measuring them
just get the appropriate reamers
for example, for 1/4" dowels (which turn out to be 0.0002 oversize), I have two reamers
you drill, then ream?
one is 0.250something, and makes a nice smooth slip fit, the other is 0.2497, and makes a force fit
right, drill 17/64
reamers are sweet
I like that they can impart different fits
17/64 seems a bit too big
SWPadnos: are you a robot?
crap - 1:26am, and I still haven't walked the dog
I wasn't last I checked
ah, east coast
it's 10:30 in LA
yes, here too. it may be time for bed
cleveland OH actually
well thanks so much for all the discussion
SWP is more east - connecticut
LA - how's the weather this week?
I picked up a ton of great info
Vermont, thank you!
it's very hot
it was fun
connecticut. how rude
I'm a cold-weather person, so now begins my months of suffering
[05:28:23] <tomp> http://steampunkworkshop.com/lcd.shtml
i want this terminal!
my wife and I were in LA (Santa Monica actually) last week, and it was warmer here than there
we got out just before the heat wave
move to cleveland, you'll have plenty of cold
it was crazy this weekend
my portable A/C unit couldn't even keep up
now it's raining/snowing here, of course
that's my ideal weather, believe it or not
I live in the wrong place
tomp: make it!
I like it cool as well, though continuous rain or snow isn't ideal for me
continuous rain and snow makes me feel better about being inside all the time
anyone happen to know about how much power one could expect for a laser printer laser?
enough to put your eye out?
but enough to melt plastic? engrave stuff maybe?
it works by ionizing the plastic to cause a static charge
I was just wondering how much power :)
it has a really big heatsink on it, so I thought it might be plausible
Sigh, looks like EMC Ubuntu doesn't deal with wireless adapters. Weill root around in it a bit more tomorrow and then get ready to figure out how to route over 100' of ethernet to get to the shop...
:( night k'zan
gfixler: 6 points of contact to measure a bore is overconstrained. you need 4 (two degrees of freedom: rotation and movement along the bore)
so 2 'bars' at 90s to each other, and the shaft?
you know, I was thinking of another option
i dont know how to actually do it, that's just the theory
like calipers, or telescoping bore gages, but tiny, hardened, ground, perfect balls
are bore gages ball-tipped?
the small ones
I almost feel like they need to be
flats wouldn't hit IDs properly
problem is, once you start adjusting it, it's no longer spherical
there'd be a little arc
oh, you mean the t-handle thingies
alright, new take on the idea.
jam some putty in there, let it harden, pull it out, and read that :)
i dont really like the t-handle things.. i have a pair of old-school calipers i ground the end of into a sort of rounded shape
so i can adjust with the adjustment screw
then lock them down?
different kind of calipers.
I've found it to be hard to pull calipers back out of a hole without changing their reading
sure, if you were using vernier or dial calipers you'd lock them
SWPadnos_ is now known as SWPadnos
we have the day off over here.. so I'm out to enjoy some sunshine :)
catch you all later
interesting, today's a holiday ehre too...
archivist: we had monday off too
we get next monday off though
oh, catholic holiday
prez here made it a long weekend
which means no stepper controllers for me until monday :(
woot a freebee rotary encoder just arrived in the post
archivist: i hate when that happens :)
hehe, I know its a return from a job, I saw it about a year ago and it had dirt on the scale
2500 gray absoluts after a google
sucessful test run of the plasma table with the machine torch last night
anyone know if the o word call will open a file in pre 2.3 HEAD and what the syntax might be?
I know there was some disscussion on it a while back
anyone know how small of a time the P will use for a G4 dwell?
I'd imagine at least one servo cycle
but I'm not sure what happens if you specify P0, for example
I'm needing 0.1 seconds at the mininum
I don't remember if the number is seconds or milliseconds, but I'm sure the manual will tell you :)
the number is seconds
the manual says seconds
there you go
but I didn't know if it would do 0.1 second
I suspect that there is no guarantee that the pause won't be longer than the number given, but it won't be shorter.
I put it in as a test but could not tell if it was 1 second or 1/10 second of actual delay
a little test I ran seems to indicate that "g4p0" actually pauses for about 20ms
it's for the pierce delay for my plasma table
btw, I had a good run last night with the machine torch and using touch off to find the material height at the start of the cut...
we await the video
btw video creators, use a tripod, so there is less scene change per frame, that should improve the user experience (been watching a few this dinner time)
I don't have a video camera but use my still camera that has a video capture. I do have a tripod left over from 35mm days...
archivist: that is a good itea
I just have a still still camera that does vids
the bottom of the camera almost always has a threaded hole, damn close to 1/4 20, tho i doubt it's exactly that , anyway a cheapo tripod/steady is cheap to do
ah! why such threads on a 'world' device?
its a very old standard
i had whitworth tap adjusters on my mg midget
witworth - what the hell is a witworth?
I used to get whitworth threads on stuff from Tiwan
a buck 2 eighty
wo! the steampunk guy wants to do an ECM EMC2 machine http://steampunkworkshop.com/cnc-ecm.shtml
steampunk is nerd cancer :v
lewin1 is now known as lewing_
I need to demag a part after picking it up with a dc magnet. afaik an ac current on the magnet that decreases in voltage will do perfect. I can pulse the ac through the magnet but was wondering if there was some simple way to decrease the voltage from 24 down to 0
pull the object slowly away from the field.
or a variac
the object is stuck to the magnet and is real small
this is a high speed operation
that I must control with a plc
that would be outside my experties..
thanks for the try
pulsing the ac seems to work in testing but always looking for the best solution for any problem
if a screwdriver gets magnetized, you can whack it, and then it's not magnetized
though sometimes you need to multi-whack it :)
"I'm hell at wacking!" John Book in the movie witness.
BigJohnT: variable/saturable core transformer?
fenn: hmmm let me look that up...
2 atoms come out of a bar, one sez "i think i lost an electron in there', the other sez "are you sure?", the first one says "im positive!"
lets you amplify/control an AC signal with a DC signal
a growler aka de-magnitizer
tomp2 it has to be done with the dc magnet that catches the part...
is this 1 time or like a chip conveyor full of stuck parts?
it's a DC magnet? what tha heck does that mean?
every 6 seconds 24 hours a day
ok, automated unsticking, you need a field pulsed i suppose
so you're trying to de-magnetize it or pop the part off?
this guy 5698K211 at mcmaster carr
not a yes/no question.. which one?
Sounds like a job for emc, a pwm out, and a transistor or two.
pop the part off i just caught
keep a layer between the magnet and the part, thin, but non metallic, increase the distance between the layer and the magnet, the part falls off
yes to both
then I need a slide or cylinder and there is not much space to work
i was thinking if you change the magnetic field fast enough there will be an eddy current in the part which would repel it
really gotta whack the magnet with a high voltage for that though
faster than 60 cycles per second
ok I used a door bell transformer at 16vac and it did pretty good...
after using my 24 vdc on it to pick up the part
I just installed this 6.06 ISO
at least :)
alex_joni, it was a hard decision :)
assargadon: heh, why?
I mean, should I try dist-upgrade or this installation
um, just reinstall
why would you have anything important anywhere other than /etc and /home and /usr/local?
I only hope my other two systems will start noramlly
renesis: some people keep things in /root
okay well theyre doing it wrong
and /var/spool/mail is also interesting
ha fuck local mail
fuck mail daemon in general
I have a choice now: should I try my other systems work
or I should begin to gain EMC experience just now :)
renesis: also cron is in /avar/
yu can cron from /etc or /home
god i hope you mean var and ubuntu didnt make up some /avar in /
renesis: I usually back up /etc /home /root and /var
and /usr/share/samba (which i put there..)
var is just a usless place distros argue over
I have ugly awfull russificated interface in EMC :)
How can I switch myself to pure english?
also you should put anything outside of /home except in /usr/local =)
j00 are not distro builder
er, shouldnt put anything
i hate it when they stick shit in var, like httpd roots
gee thanks guys thanks for making me look for it and giving me an extra place to backup
assargadon: you know how IB< always says.. "Never touch a running system"
var is almost as bad as opt =(
hey alex_joni: the HOME_FINAL_VEL works like a champ on 2.3
assargadon: export LANG=C, before doing emc
who IB is?
In every case, he is wrong
* assargadon feels his system is much slower than 7.10 installation
assargadon: crap.. can't type.. IBM I mean
assargadon: is it a dual-processor PC?
It's P3-500 machine
try latency-test before you try emc2
(although the P3's should be very good at realtime)
omg fuck that
i will not do emc below 1GHz ever again
gee renesis, tell us how you really feel
EMC is bloat in terms of sub GHz boxes
especially considering the distro choice
its doable, but its mad work
while turboCNC just works
renesis, it sounds strange
ut sucks and its dos
but itll push steppers on a 200MHz machine
different beast, of course
trying EMC on a 200MHz machine is a joke
as far as most production controllers really much, much slower that 1 ghz
yeah but EMC cant do slow machines
renesis: why not?
and alot of it is ubuntu
you're right - it's harder to get things working on slow ma
um, because its time critial realtime on a OS not made for that
what do you concider slow?? I have run dapper on 400mhz machines..
DOS is almost machine level
install a server machine, without X
then emc2 on it with keystick
the OS sucks but its prob as close to the hardware as you can get
im using EMC right now, i love it, but ill never ever try to put it on a subGHz box again
turbocnc is nice.. until you want blending.. and a mouse.
respectfully, it is failure compared to turbocnc at sub GHz
or a network connection
this aint about computers
this about making parts
CNC machine dont need a net connect
hey guys...I was critic several last days :)
And when I beat myself
and install EMC at all
EMC is awesome
its a dog on sub GHz boxes
YOU beggining criticise it :)
maybe you guys dont run microstepped drives, i dunno
* skunkworks_ started out using turbocnc.
no dude i have experience with all three popular software controllers
and EMC on several boxes, abnd machines prob 50 circuit boards and dozens of metal parts
so...what I should do now, when EMC installed?
Should I read Manual? :)
with EMC, and i can feel pretty secure saying its a dog at sub GHz compared to turbocnc
turbocnc sucks, but it can push microstepped open loop systems at 200MHz
EMC on a system like that is fail, totally and utterly
granted, EMC does more (im speaking in combination with axis gui)
but the real killer:
EMC variable system is complete fucking ass
total shit, unless it got 300% better in the last couple versions
thank god it got away from the #numbered variables
so yeah, sub GHz, and if i needed to write a program with variables (i just dont now, because EMC is so bad at it and i use EMC) i would use turbocnc
i dont give a fuck about linux vs microsoft anymore, linux fails horribly at CAD and EDA
eagle runs in linux, buts its considerably slower, and GEDA is useless for a few more years if they do everything right from now on
qcad is functional commercial leftovers from ribbonsoft
not exactly what id call ideal, so i have to run a windows machine for CAD and EDA anyway
assargadon: yes - read the manual. cover to cover.
emc is the best thing going if you have money for a dedicated GHz+ system and dont want to work with variables in gcode
What manual? Users Guide or Integrators Manual?
start with the users guide
assargadon: if you want to use it, then User Guide first
your so yeah, thats how i really feel
Is there any king path to the EMC? :)
what is PID?
"...output is based on PID compensation...."
a control algorithm: Proportional, Integral, Derivative
[18:40:34] <awallin> http://en.wikipedia.org/wiki/PID_controller
assargadon: running a stepper machine?
then you really don't need to worry about pid.
I don't know renesis I've been programming variables since 1996 without major issues.
skunkworks: ah, in this therms :)
skunkworks: then no, my machine is servo
skunkworks: I just recognize you question incorrectly
are you using step-to-servo drives, like Geckodrives?
skunkworks: I thoughts you asking me if I have theoretical interest or running something
and you asked about steepers or servo ;)
rayh: I think it's hard to guess how to improve on a system that is reportedly "complete fucking ass" - so ignoring complaints like that is probably best
SWPadnos, I don't know for sure what is step-to-servo is
my old soviet robot has DC motors
motor drives that take step and direction input, but run servo motors rather than stepper motors
Uh Yep but I couldn't resist.
and optical encoders
11:44 < rayh> I don't know renesis I've been programming variables since 1996 without major issues.
with #'s, apperently
SWPadnos, then no for sure :)
i mena fuckit why not just code in raw memory pointers
DC servo drives can take a number of different inputs, including analog velocity or torque commands
all computer can made is start motor and stop it
wouldnt be any harder to double check
Hell I wrote the first manual on variables while still running RedHat 5.1.
(in selected direction of course)
okay so youre not anything close to a normal avg user
youre a dev
or at least involved
and it can asquire robot, if encoter "enabled" or "disabled" (shadowed) now
ok I'm an average user
Asd you are correct that g-code is not a very "computer programmer" friendly language.
gcode is totally easy to read
gcode is the easiest language
um, out of curiosity, what does TurboCNC do with variables that EMC doesn't?
okay you guys argue with me that #'d variables are totally awesome and acceptable
im pretty sure you can do words, and not eccentrically
But it does work pretty well for guys used to making chips by turning cranks.
i dont remember using the word variables was any easier, first implimentation
looking at a manual for TCNC v4.01, I don't see any ovious difference between EMC and TCNC, except for some restrictions on where you can use vars in TCNC
obvious, taht was
right, words vs #s
no, there are no words in TCNC
and i remember there was something weird about words
at least first implimentation
up to 10000
im pretty sure there are words
[18:48:34] <SWPadnos> http://www.dakeng.com/man/turbocnc.html#_Toc90515777
that says there aren't
i dont believe theres docs youd have to demonstrate
or show me in program docs
i actually dont like turbocnc at all =(
* cradek rolls his eyes
that was their program docs, by the way
that thing has major bad errors
tcnc and emc variables (and expressions) look very similar to me
no im saying their in program docs
i dont trust that project at all
but is it complete fucking ass?
weird maybe it was one of the controller at school did word variables
\i dont care about the competition
that's good to know
you guys should be better
so i can tell people who fucking awesome EMC is
if you have a proposal for an improvement to gcode, you should write it up in a careful and thoughtful way
so what are your (realistic) requirements for the Awesome rating?
yeah thats fine, one day when i have time
I mean, iGHz step pulsing on a 100Mhz CPU is obviously unreasonable
label = value
and literal insertions in code
"you should be fucking awesome instead of complete fucking ass" is useless and just makes people ignore you.
err - 1 GHz
with maybe some sort of variable marker prefix
just like assembly language
like #thisvar = 10.0 ?
if assemblers can handle it, i dont see why a gcode interpeter cant
being that gcode is a higher level language
or #thatvar = [#thisvar * sin[#theothervar]]
yes i think thats maybe how it works now, but i remember there was something eccentric about the EMC variables in the version i used
I want gcode to look like this: http://cvs.linuxcnc.org/cvs/emc2/nc_files/flowsnake.py?rev=1.1;content-type=text%2Fplain
or maybe this http://axis.unpy.net/files/01167315565/example.gcl
Um nothing higher level about gcode.
yes there is
assemblers have the advantage of multi-pass parsing
can someone please add that feature to emc?
its an intepreted langiage
youre not feeding a processor raw gcode in prom
no it isn't
um,. yes it is
there is no intermediate representation of G-code
assembly language is lower
gcode is an intepreted language
now you guys are bothering me
wait, that's a stupid conversation
gcode is a higher level language in context of raw processor machine code
G-code can't be done the same way as assembly because it's executed inline, there is no pre-parsing of variable names
and there's no fix-up pass
okay why not
oh wait -- gcode *can* look like that, when you configure axis to use program filters
scan the header for variable definitions
replace any variable occurance with the literal value of the variable
cmon i could code that on a microcontroller you guys cant code that on EMC?
are you looking for the equivalent of #include files?
But then it isn't really rs274xx
when then add a suffix to it
renesis: you clearly know a lot of this. Why not volunteer your time and write it? emc2 is improved only when people contribute improvements.
heh, there are a lot of text in this manual, heh :)
call is rs274-nbetter
i dont do higher level code
sorry im not trying to waste my time on software
i have far more interesting shit to hack
then why waste ours?
fuck crashing computers, its all about crashing tools and circuits
well, if you'd like to write up what you'd like to see, people can then look at it and critique the ideas
if you think criticism of your program is a waste of time, that says alot =)
yes well, when i have time
criticism in the form of "your shit sucks ass" is a waste of time
EMC is not what i have decided to do with my life at this point in time
you shit sucks ass is an abtsracted way of saying it can be better
renesis: so why are you here then?
criticism in the form of "named parameters would be nice" is not a waste of time
because i love EMC
I have no problem with constructive criticism
okay well you have a prob with bad words?
renesis: usually yes
i said that
because numbered parameters are complete fucking ass
sounds like my ex could not please her no matter what you do
dude is an assember can turn mnemonics and variables into hex, EMC should be able to do it too
gcode is a higher level language than asm
step 1 in constructive criticism: look at the documentation first, to see if your criticism is still applicable
asm is literal, gcode is interpreted
asm is not literal
step 2: see if the latest version of the software still exhibits the problem
i dont see how reading variables in a head and replacing them in the following lines is hard t code
There is a very old part programming (APT) language that might meet your interests.
There are a couple pages on it in the wiki.
maybe it breaks strict rs274 compliance
but really, who cares
step 3: if and only if both of those tests fail, make a constructive suggestion about what you want to see and why
and as jepler pointed out you can always use a filter on your files
so why not do it
as long as you can feed it rs274 and it works, whats the prob with extra functionality
gcode spec is like from the 70s
yep, and any extension still needs to support those old flavors unmodified
that's the hard part
theres absolutely no characters you can use to escape a variable?
i dont even see how this would break strict rs274 compliance at all
there is a variable escape character, it's "#"
What circumstance would you want to escape a variable.
you can use named parameters with the latest emc2
if numbers dont how do words?
but I guess trying it out would be too much
so your super strict rs274 interpreter knows its a variable
and not gcode omg
i did try it
the sample code I posted earlier (with named variables) should actualy work, it wasn't hypothetical (though it may have had syntax errors - I didn't check it)
i remember it being odd for some reason
so im not 100% what im bitching about isnt implimented
x#1000 is a perfectly good use of a variable
and don't try it on EMC2 2.0.1 or some BDI, get the latest liveCD to test
that's also legal now, I believe
* awallin thinks SNR has dropped a bit on this channel lately...
yeah i remember trying it and being pissed off by something abut implimentation
awallin: it's negative
(which is way odd for a ration)
you may need to use  to enclose the variable name
G-code by definition removes all spaces and is case insensitive
what does that matter, whitespace doesnt do anything
thats why there are words
also, every letter except E (and maybe E now) is used as a programming word
SWPadnos: E is used too
okay so dont use a letter as a variable prefix obviously
HELLO is an H word, an E word, two L words, and an O word
identical to H E L L o
What's a "Named parameter not terminated" message mean?
yeah what does this have to do with the topic
trust, i know this shit
I don't trust that, since you've been complaining about missing features that actually aren't missing
anyway, you guys have a problem with a user who loves your product telling you a feature is ass, thats cool
says alot =)
and you've said that TCNC can do things which its own manual says it can't
well i may be mistaken
so feel free to troll me for that
could be (I could be also, I only know what I saw in the manual)
maybe it was another controller im remembering
but i really dont trust the project, so i dont see a manual as proof of anything
renesis: the minute you provide usefull feedback you'll be taken seriously
anyway, its got some issues
alex_joni: i have
renesis: I did say _usefull_
sorry if you cant see thru the creative use of the language
proof is in testing, which you should do with EMC2 before swearing at us
yeah right, now whos being pointlessly critical
so you code the app?
or you just adding your two cents
renesis: yes we do
i have tested variables
it was hellish
found it "Named parameters work like numbered parameters but are easier to read. All parameter names are
converted to lower case and have spaces and tabs removed. Named parameters must be enclosed
with < > marks.
yeah i tried them
ok, so my syntax was wrong
how long ago?
there was something weird about them, but maybe it was a first implimentation thing, i dunno
how long ago?
like i said, as soon as it happened
i dont know youre involved in the project not me
you tell me when it happened, you prob know where the changelogs are
I'm asking you how long ago you tested
docs are in like 3 diff places i never know which is the authority
im saying when it was implimented
so prob within the last year
renesis: that's why I mentioned _usefull_
so far you haven't provided any concrete information about EMC or TurboCNC which hasn't been contradicted by either documentation or examples
Ha it works g0x#<point1>
and ive said its maybe been sorted, i dunno
alex_joni: drop it man
thats the 3rd time now
if someone has a prob with what im saying, bans happen
you can even do G#<movementmode>X#<point1X>
renesis: all you did so far was wasting about 1h
renesis: we don't ban people here..
i have to flip a board brb
Nah. renesis did something useful -- he made me rtfm.
anyway, if im wrong and its fixed, fine
well this is boring
or if i was just using the first implimentation wrong, fine
Now to setup the variables <first_base>, <second_base>...<home_plate>
but the fact that theres word variables now obviously proves that numbered variables are less useful
Nope. I found the opposite.
which was my main point, which people tried to argue was perfectly fine, which i think is really silly
ok, so one of your major complaints has been addressed already, what's next?
i dont care is turbocnc is the same, turbocnc sucks, emc should suck
It takes more keystrokes to do named than numbered.
see my examples above
yeah try and keep track of 30 variables without constantly rechecking
I was actually contradicting your statements about EMC missing named variables (and TCNC having them) usefulness is user-dependent
because like, i dunno i dont have many convos in pounds and numbers
one thing you can do with numbered parameters is use indirection - not possible at this point with named variables (since you can't store a name in a var)
and if it wasnt turbocnc it was some other controlle,r maybe something at school
so you can do things like G0X##20, and it will use the value of the variable number stored in #20
It would appear as if the # "escapes" a variable and <> escapes a named one.
so each has its uses
okay so if something works with numbers and not letter variables, im thinking you guys still arent done
so ill wait longer to bitch i guess (i dont see why they would be implimented diff and have diff usage and if you guys wanna argue that, by all means go at it)
which is maybe why i was going nuts trying to do shit i was doing with numbered variable with the word variables
Darn SWPadnos's ## works.
We've got pointers.
but yeah thats obviously user error, not knowing theres two diff parallel variable systems working in EMC
* rayh gets caught in a user error.
rayh, try ### some time :)
yeah btw sorry for wasting all your times talking about the program
hmmm. I wonder if you can do stuff like #[#10+#11]
ha damn i just pulled the tool out instead of flipping the board
Yep it (###) works
I reading and reading...and I know all this stuff :)
2L .005 tip 30degree pyramid cutters = totally awesome
I think you can go 10 levels deep or something like that
fuck single flute half rounds
do *NOT* tell us how you feel about single flute half-rounds
and only $12
well, actually i love them
renesis: do you feel some special pleasure when swearing?
ive done most my boards for this project with a broken one
at .015 DRC
acemi: yeah man fuck off what does have to do with shit?>
im from efnet
im actuaqly being really nice right now
how nice of you
yeah i try
alex_joni: do you take special pleasure in making personal comments that have nothing to do with technical issues?
cuz shit like that in #electronics makes me wanna stab
It's been fun. Now I've got to get out the four wheeler and run down something for dinner.
* renesis envy
renesis: this is not efnet, nor is this a channel for bored teens
and im only board because the machine working =(
youre making trolling comments
please stop, thank you
renesis, in my opinion, youhave been trolling nearly the entire time you've been posting this session
Hey, gooroos, can I look trought parts like "stepdir configuration"
ah well thats a matter of opinion
without carefully read them?
yes, it is
i have made no personal comments on anyone
so if im trolling, im trolling EMC code
I got the idea, but I have completely different situation
and im def not trolling EMC coders, because i have stated several time that i <3 emc
assargadon, if you won't be using step/direction, you can skim that section
I want :)
only alex_joni has taken this to a personal level
I just haven't :)
well, you're chatting with the emc coders, so saying that the code (our code) sucks ass is kind of a personal insult
it's kind of equivalent to "you're a bunch of idiots"
well, because its true
i think that part of the code is ass
that doesnt mean i think the devs are ass
just means that they prob orking on more important stuff
big project yo, lots of pieces =)
<SWPadnos>it's kind of equivalent to "you're a bunch of idiots"
<renesis>well, because its true
not that youre a bunch of idiots
I understand :)
But it's looks funny :)
my opinions on the app are true
i <3 the devs, god how many times i gotta say
what's more, you are making complaints that are now irrelevant, and you obviously haven't refreshed your evaluation of EMC2 before complaining
I get it
I assumed our comments crossed in the mail :)
actually my way bigger complaint was EMC distro choice and speed requirements
the variables thing was just a tangent
you can run EMC on any Linux + RT kernel
the distribution is for easy installation
also if they dont act similar, the numbered and worded variables, then that may have been my issue
and i dont see why there would be two variable systems
Alex suggested using a server install and keystick, which would look a lot like TCNC, but should work on a much lower spec PC
ther eis only one variable system
yeah sorry man
im not trying to hax0r linux for my cnc
but variables can only contain numbers, so indirection doesn't work on named variables (or at least can't point at named vars)
i did that for years and years for fun, and now i got other shit to do and way less time
why not, just number the variables in the preprocessor and then make them interchangeable with the numbers
if you install ubuntu server, you can still install EMC2 packages for it
there is no preprocessor
ubuntu server isnt much lighter
i believe it still runs mad pointless daemons
tho its been a couple years since i tried it
you certainly don't need LAMP for EMC
anyway its usually fine once i go thru and kill service init and switch to blackbox
gnome is an issue
you can use xfce if you like
EMC can install onto xubuntu too AFAIK
i can use blackbox if i like
and that's just for the pre-built packages
i have, axis is fine
and i get a real window manager
you can also get just the RT kernel, and compile EMC for yourself
not some explorer shell wannabe
yeah i really dont want to
or compile it all yourself - lots of options
i know i can, but thats not what i do with my time anymore
yes i know, but you chose a default distro, and im not particularly happy with the choice
although i do see the point
ubuntu is the shit for non-vanilla binary distros
I don't believe there is any distro at the moment that will work on a low-spec PC *and* has drivers for modern hardware
i cant stand deb but ubuntu installs always go nice and smooth and hardware detection is ace
slackware, gentoo, prob arc
the idea was to make installation as painless as possible for average users
i think arch
yes i know
which usually includes connectivity with Windows networks, since as you mentioned CAD/CAM on Linux is a bit lacking
and i believe that was a bad choice for a RT system =)
but yeah you guys succeed, it is easy
and machinists not always lunix nuts
well, it's RT + GUIs + networking
so yeah, understood
ubuntu detection and setup is as good as anything
at some point, I may make a "small system" distribution
ive had less probs with ubuntu than SuSE
since embedded RT is my field
which is the only other biggy distro ill use
last time i tried doing a custom rolled EMC is was for a 300MHz box
that didnt go so well, and was tons of work to have it fail
were you trying to get EMC to work on some ARM machine at some point?
i broke the GHz computer, i newegged a new system, for like $140 (open box mobo yay)
so its 1.9GHz now, i should try again
hmmm. ok - I was thinking of someone else
just an old amd K6 i used for turbocnc
im doing SAM7 ARM stuff right now
SAM7 or SAM7X?
but not really planning on doing gcode interpreters with it
I tried to install EMC+hardy on a old Duron 700. Its to slow. what cpu suggest you? 2GHZ Atlon works fine
gefink: i have a 1.9GHz sempron
ubuntu didnt detect any of the hardware really
but it can push the parport fine
and the SD card reader works
im going to try the new EMC on ubuntu8 when im done with the current projects
current emc/ubuntu just too old for my setup, but its doing RT fine
45W proc, too, for $30
newegg has a ton of open box shit lately
ubuntu 6 worked with Duron700 slow and safe
getting alot of returns as they getting popular, i guess
if 6.06 works on the Duron machine, and it's fast enough, I'd just use 6.06 there
but ive never had anything missing from anything open box, my scanner had a beat up box, the mobos were fine exzcept no shrinkwrap
heh, i wiped from dried water flux from the back of one mobo with alc, but i doubt that would be any kind of problem to cause a return
(tho water based flux can be conductive a bit after drying, heh, supposed to be for pin dipping i guess)
SWPadnos: that do i. but wondering about ~30 usec latency and lost soundinterrupts
ok, that's a problem then :)
yeah what kinda of system are you driving with the duron?
just about any modern PC should be fast enough
i wasnt microstepping under i got a GHz athlon XP
the problems will likely be with onboard video
get teh nvidia gf
so plan to add in a cheapo video card
gf + lunix + nv or binary drivers = lub
i think rt doesnt like the binary drivers or something, tho
i never bothered trying them
(i hear people bitch, tho)
gefink: wie geht's?
alex_joni: prima. meine ubuntu8 versuche auf meinem duron sind gescheitert
i did fuckup my flip note coordinate =(
gefink: we should probably take it private..
+alex_joni ich verlier sogar soundinterrups
i put the pad diamater instead of the X value =\ =\ =\
gefink: ich hab dir ne private nachricht geschickt..
alex_joni: gute frage wie das geht.
gefink: schau nach andere Tabs..
alex_joni: ich glaub ich muss mich von meinem komandozeilen irc verabschieden
was fuer ein client?
er heisst nur irc
sonst keine info
gefink: /join #emc-de
I rebooted in "normal" linux :)
to read some docs...
They should be somewhere at EMC CD, isn't it?
assargadon: you mean the regular kernel?
the docs should still be available from the menu
if you have a different PC< then getting them from the LiveCD is tricky.. because they are in the squashfs (and it's a pita to mount it..)
regular kernel and 7.04
ie, not the liveCD
assargadon: but anyways, you can find the docs online at http://www.linuxcnc.org/dovciew/
Heh, really :)
It some kind of shadow over my mind :)
heh, all 3 OSes works normally
and it's funny :)
assargadon: good to hear
even more manuals in the web :)
5 instead of 2 ;)
It some kind of RTFM kingdom :)
assargadon: look at the html/
[20:16:24] <alex_joni> http://www.linuxcnc.org/docview//2.2/html/
you have all you want there :D
but PDFs looks more funny and can be read in offline...and they consist from 1 file
assargadon: suit yourself :)
I'll use both :)
I can't find chapter
"How quckly write driver for your machine and force all this staff works in 5 easy steps"
Can't you give me direct link?
hmm.. you probably want the programming_manual for that.. but that's really lacking (low on content)
skunkworks_: he's trying
seems like it.. persistant :)
I have a hunch that it's a german guy I'm trying to drag in here
anyways.. he had a pretty odd error with 8.04, so it would be great if he could make it in here
work again tomorrow..
today's been a nice change (day off) :)
I'm free tomorrow but I'm going to work anyway
but I got some machining to do so it's nice :)
anyone know the syntax for calling a file with a subroutine with onnn call?
there is no special syntax
BigJohnT: didn't o100 call work?
just didn't execute the subroutine
if the appropriate onnn subroutine isn't in the current file, then the nc_files_dir will be searched for a file names onnn.ngc (I believe)
that's lower case o, not uppercase
not nc_files.. I think the PROGRAM_PREFIX folder
that is what I did
yes, PROGRAM_PREFIX - that's why I menmonic-ized it to nd_files_dir ;)
the o- call should also be in a .ngc file; at least at one point it did not work to execute o- call in mdi
so similar to M100...
except lower case :)
be back in a sec ok
or a minute I'm slow today
ok, that should only work in TRUNK
o-call in mdi is in TRUNK
but I think BigJohnT is on pre-2.3 cvs
and I have 2.2.5
the o-call in MDI was after MDI o- calls I think
and you need the LAZY_CLOSE option somewhere in the ini file
(for MDI O-calls)
I want to do it from a .ngc file not from MDI
none of the aforementioned changes are in 2.2.5 ;)
you'll need TRUNK, from what I can tell from the CVS logs
you mean pre-2.3 CVS TRUNK?
yep that's what I have...
be back in a minute
oh, well that's not 2.2.5 ;)
SWPadnos: he has both
I have both installed on this machine
is that a problem?
it-s pre-4.8.5 CVS TRUNK
BigJohnT: no, we're just fooling around.. don't mind us
I thought so...
yeah, it's hard to tell when we're serious
it happens so seldom
your just making fun of me cause I didn't call you sir...
of course I do
I'll test it now
alex_joni: Now i have a new irc client (firefox plugin)
gefink: klingt gut :)
yep firefox plug in
ok when I try and run it I get AXIS error
BigJohnT: our mindreading skills are a bit rusty
EOF in file:home... seeking o-word: o1361208336 from line 1
I was typing it...
* alex_joni learns patience
well then - you type too slow.. ;)
I'm at step 5
I told you I was slow today
too much 3-cad work
the o value is a bit off at 1.4billion
BigJohnT: works here
I also have o250 call in the o250.ngc
ok first or after
(o250.ngc still needs to be valid)
so I bet the M2 is the one missing for you
yup.. still works without the o250 call
but you need M2 at the end of the file
or any other closing Mx
ok I added an M2 to o250.ngc
hmmm not yet here...
BigJohnT: I have this
both files are in ~/emc2/nc_files
same as I have
and I opened stepper_inch (copied by the config picker to ~/emc2/configs/stepper)
I'm in emc2-trunk and using axis sim....
check yourconfig.ini : PROGRAM_PREFIX
I bet PROGRAM_PREFIX might be off
PROGRAM_PREFEX is ../../nc_files
BigJohnT: for me it's /home/juve/emc2/nc_files
but I tested on 2.2.5
let me test it on 2.2.5
BigJohnT: so either it's broken in TRUNK, or it doesn't like relative paths
I think the latter
I'll change it and test
I get the same error on TRUNK
seems like bustage
aha.. in TRUNK it only works with named subs
^ that works on TRUNK
and that works for me as well
should I try and update the manual in trunk?
BigJohnT: no, this is a bug
[21:38:21] <alex_joni> https://sourceforge.net/tracker/index.php?func=detail&aid=1955715&group_id=6744&atid=106744
BigJohnT: but if it's missing from the manual, then please do
yea, it's completly missing
* alex_joni is off to bed
good night all
[21:43:44] <renesis> http://darkertechnologies.com/image/SSPX1123.3g2
.008 DRC on the taig is working omg
heheh, im average fab house technology spec, now
chinapcb says they can do .004 space/trace now, tho
Just got your bug report. Try it with a file name of 123.ngc instead of o123.ngc.
I'd guess that will work. Notice that you needed "file.ngc" NOT ofile.ngc. If that works, I would just call this a spec change.
yep, that works
Sorry. Didn't mean to be so abrupt. "Hi Alex" :-)
but the error is a bit obtuse :)
What did the error say?
EOF in file:/home/juve/emc2/nc_files/foo.ngc: seeking o-word: o145160720 from line: 1
(not sure where the o145160720 comes from)
Whoops. I probably didn't change the code that does the error message. the o145... is probably the address of the string 123.
yeah, it changes..
lerman: ok, feel free to close my bugreport
I'll have to update my version of trunk and fix that.
if you remember to fix this other issue (the o145..) I'm happy :)
OK. But in that case, you should generate the obtuse message bug report... Otherwise I'll forget it.
or I can update this bug
and update the info..
I'd appreciate that.
I've got to run... My Kempo (martial art) class starts in a while.
I'll open another bug for this.. thanks
Thanks for your diligence.
Ok, I've updated the user manual now what?
lerman thank you ( this is realtime bug fixin :)
fast as fast can be...
BigJohnT: was that on TRUNK?
cvs diff -u > manual-change.patch
(in the folder with your changes..)
then send me the patch somehow (either email, or pastebin.ca)
BigJohnT: you saw the spec change for TRUNK.. right? (123.ngc is the filename)
I'm still up a couple minutes.. so if it's ok, I'll just commit it
does the cvs diff take long?
now it's rolling
depends where you ran it
less manual-change.patch | wc -l
what do I pastebin?
(the content of that file)
[22:02:15] <BigJohnT> http://pastebin.ca/1004390
looks like it has a couple of diffs
one for g61 and the one I added
but both are ok
BigJohnT: join #emc-devel
I reading and reading and reading
and I still cannot imagine, what I should to do :)
Well, what to do, the 6.06 version of EMC/Ubuntu doesn't deal with wireless. Run 100' of ethernet or go back to sneakernet.
* assargadon trying to recognise...
Wonder of the new/planned release with 8.08 will work with wireless...
assargadon: where are you lost?
I think I got the main idea of HAL, but I still not sure what should I do
I have a hardware (3d-positioning system)
with control box
this control box connected to computer via LPT
it takes 2-3 inport and outport to send the command or to asquire the data from robot
despite the fact this commands a quite easy
is this for a tool changer?
(put the motor forward/backward/stop)
uh, M3, 4 & 5 control the spindle
I failed to understand what toolchanger is
it's just robot with 3 perpendicular axes
for automatic tool changing in the spindle
powered by servos
ok, not a mill
a catesian robot?
yes, sort of.
just moves X, Y, and Z
it have some kind of "hand", which can be activated or deativated
I'm guessing here but there is some different kins for robots
but it useless now :)
"just moves X, Y, and Z" is correct
but each one is rotating not linear?
each one is linear
should be the same as a mill then...
yes, it same as mill - logically
but most mills have e little different construction
but same movements
yes, and same control
so I can:
X +- Y+- Z+-
it's servos, not steppers
so X: forward/backward/stop
speeds/feeds suggestions for acrylic pockets?
and same for others
3/8" carbide, but the machine is only 80lb desktop
and you can obtain the status of encoder
for every axe
as far as status of endstops
both endstops and encoder is just 1 bit
for home position and limit switches
every encoder have 1 channel only
so you should keeep direction in mind, not measure it
that's beyond my skill level at the moment
problem is finding starting point
I'm kinda up to speed on steppers more than servos
start with one of the example configs
is it correct: I need to implement HAL module
which represents my control box
that part is beyond my skill levels at the moment
perhaps some servo experts could step in now...
so pity it beyonds my skillevel too :)
* renesis only know how to do open loop stepper insanity
did you try some of the example configs
BigJohnT, depend on what you calling "try"
yes, I started some kind of simulation :)
calc says i can do 1600rpm/33ipm for .005" cuts at 54fpm diameter speed
DO WE DARE
haha my steppers are 30ipm max this config, im pretty sure
and the cone on the screen writing the letters by the program
axis is nice
axis gui is why i wont use mach or turbo anymore
the feedback and persistent paths makes it easy to catch manual coding or setup positioning errors
put the blocl in the freezer and cut it slow
i was gonna use baby oil
freezer tip sounds neat, ill try that
heh, i should do that now its 1" acrylic
take awhile to make a difference
ive cut acrylic and neopene the same way.. speeds and feeds
i wish i had a million dollars so i could just buy delrin and fly thru it
try to get it cut while its still cold.. so it'll chip nicely
they need to figure out clear delrin
but its mad pockets, itll prob be warmed up anyway by the time im at the bottom
ill just make a pool of baby oil and spoon out the shaved ice lookin chips
works on delrin too...and many other ones.. minlon too
baby oil worked cutting out the stock with a jogsaw, and ive welded cuts behind the blade with that jigsaw
haha, really sucks when you spend 5min going thru a block and it just doesnt fall =(
yeah i bet on delrin its great
well assargadon: did you pick the most suitable config and try and run your machine?
that stuff already cuts easy
go near full depth and air vortex the chips out
dmes1: thanks for tip, def sounds smart
well, ill go deep
my machine is pretty pussy
the 3/8 wont break, itll just skip the X/Y steppers
BigJohnT, I cannot imagine how any config can know the protocol of talking to control box
slow plunge rate the roll.. and BLOW
near full depth just asking to east it =(
it's not just rewiring...
do you know what the protocol is?
is it PWM or something like that?
they make step/dir servo drivers, no?
yes, and I even write something like driver
that would be neat i could just throw them on my config and try em out
BigJohnT, I don't know what PWM is, but I doubt
ren what have you got now?
ok, I'm just guessing at this point
tho im kinda amzed the open loop setup is as trustworthy as its proven to be
dmes1: the big taig mill, with a xylotex 4ch
It's very specific...not hard, but specific
for a couple years now, is pretty hardcore
chatter getting too much to follow so I'm headed for the porch and a glass of wine
but yeah, its only 80 lbs, so its not gonna rought out pockets at 1" deep well
pwm = pulse width modulation!
pwm = easy!
the eastern block ran on stepper technology exclusively until the '90s
(heh, i code alot of microcontroller stuff, its like the first thing you figure out on a new chip)
ive gotten some tools from that area
take 2 1/2" passes and get over it
enco import stuff, itll be marked yugoslavia or something
yeah okay ill try
i have a stupid amount of 1"
my exgf just had it sitting in her back yard
leftover from a job her bro did i was like, fuck thats prob $600 in acrylic
i need to go .835 deep, so ill try .3" passes
ive had 3x33x7" deep pockets to deal with you dont have a problem.. your just inexperienced and afraid
hehe, stackup chassis w00t
well, yeh and my machine is little
like, i have a cnc degree but im not a machinist, im an electronics tech
machine tools are scary!
but your work fit its envelope if you work smart
yeah im alot better now
yeah, i have an AS cnc programming from a public college in los angeles
its neat, my instructors mostly did mil and aerospace in the 70s and 80s
associates degree, 2 years
its like mill and lathe mostly, 4 CNC programming courses, 3 manual machining courses, a hellish GD&T inspection course, bit of cad/cam
and some gen ed
good lord you can hang that where mine is.... in the outhouse at camp...i have one from George Brown College
the computer fucked up and i took all 3 CNC courses in one semester
instead of three
i was learning to program machine code on an 8085 proc, tho, same semester
so gcode was kinda easy compared to that hell
yes.. i feel your pain what yrs did you go to college
im done like now
im not a machinist, tho
the idea of working in a production shop scares the shit outta me
and then i met alot of machinists, and they all confirmed my fears
oh ok i finished my mess in 1989....went to work
and theres the pick of the mideastern dude wrap around a lathe...
yeah i like doing fast electronic prototype stuff
so the CNC stuff kinda related to that more than mass production stuff
then go into a prototype makers apprentiship.... i'm one of those
id do CNC tech for a prototype house, but besides that, its too dangerous too much going on, not for me
yeah toolroom machining and prototype stuff seems interesting
but im really alot better at electronics than machining
then go for machine tool dealers as a techie...
servo and plc is hard to find...
well, right now i dont want to go anyplace
because once you leave here, you never come back (los angeles rent inflation makes rest of the country look silly)
if you have the NUT i'll hand your resume to the maintenance manager... were in Ajax ontario canada
and JBL and cerwin vega are local
so im going to try and scare a vench or design tech position at once of those places
*secure, omg hi, mr typo here
go for broke....
yeah if that dont work
i got it
then theres a job offer in maryland
but thats insane, because its mil contractor stuff
for research lab, like dream and a nightmare, way things are now
insane.. i have 1 in northern RUSSIA...
but yeah if that dont work out, im just gonna try and get a CNC tech job
because theres lots of that locally
haha northern russia
sounds cold and barren
been thinking 3 months... there is still snow there
im pretty much finishing school
i was full time, but they fucked me on aid on a technicality
good coin... good company.... but its a 6 year contract...
so to them i wasnt even halftime, so instead of $100, or $500 or whatever for halftime, i got $0
so im pissed, im like fuck you guys, i already got 2 degrees from you haha, im going to get a job
learn more doing tech work for someone else, right now
i know the tech, i can work fast, im disorganized as fuck tho
do you get it done??
me owning a business would prob be a fun two years and then omg one month just into the dirt
but i could do it better
like, my goals now are more to learn technique more than finish projects, so when they come out nice its kinda a bonux
that all the BIG guy sees.... Find the big guy and impress him fast
actually im wasting time right now
i have like 5 circuit boards to mill
then that acrylic thing
oh yeah fuck some aluminum heatsinks
then pick up my parts from ups (OMG THEY FAILED AGAIN)
do the plastic in the am ... put them in the freezer oernite
i leave at like 6A
tho, bus for san jose
this is like some nightmare joke turned into 24 pcb project
all built into altoids mint tins
so cut them at 4 am
they NEED to cool
but im planning these fine DRC circuit boards to fuckup