I'm trying to find it...
I see: Hardware step generator for 5I20
I must have imagined it. I can't find it.
It's okey, thanks though. I was RTFM the userguide to try and firgure out why I keep stalling
I SHOULD be able to get 150IPM, but stalls at 85.3IPM
Here's my current in (I'm only working on X axis atm) http://cpp.sourceforge.net/?show=38572
I find it VERY frustrating that there's a longer detailed explanation of HOME_IGNORE_LIMITS than there is for STEPGEN_MAXVEL
If someone has some ideas to an explanation for this, I'm all ears... I can jog in one direction at 150IPM wittout my motor stalling. If I jog in the other direction, it will stall almost at the exact same position.
oops... book on keyboard sri,
something's sticky or rough
JymmmEMC, running out of power/friction
In only one direction though?
some twist os screw difference
remember its harder walking up hill
This is X axis, there is no "up hill"
5TPI thompson ballscrew on 2:1 belt driven pulley
measure the torque then
feel by hand, string and weight and bar or best a torque meter
or a dc motor and measure the current for a fixed speed
or nice nifty tool a torque watch
The ballscrew is under the middle channel http://farm1.static.flickr.com/143/389868985_9590ce0735_b.jpg
In the pic, the pulley is removed there.
same end as where you see the disconnected X axis motor.
These were the same type of issues I was having under the Xylotex @ 24VDC, now on Parker at 48VDC (which is what mariss suggested for these motors, voltage wise)
add mass see if the working direction then fails
archivist: The whole gantry is attahced
[00:43:32] <JymmmEMC> http://farm1.static.flickr.com/166/424362252_1bd7946026.jpg
monitor the current across a shunt resistor ( to see difference in pos/neg direction ) or, some amps have a current monitor output ( eg: Copley ) this will uncover friction/extra load in 1 direction
The PS is 6A, my fluke does 10A. Gimme a few to rewire everything.
maybe it's crabbing
one edge of axis ahead of other ( when 2 guides/motive elements)
of the gantry itself?
the moving carriage
yes, is the picture showing longest axis as X ( l to r ) ?
whatever that is
does a light finger assist eliminate the problem?
let me try... have to stretch to reach it...
be careful please
I meant strect the ktbd cable
if the assist makes a difference , go back and see if the motive point (nut) is truly in center of the slide ( and in center in both direction in case the slide itself shifts off center when reversed )
sometimes loosening the joint (nut to slide) will relieve the problem ( not a solution but a way to see where the problem is at )
Ok... Not actually a"light" touch, but added force (not enough to move the amchien or anything), DID help it from stalling, and actually stalled it wehn I accidentally pressed the other direction.
ok, you'll have to tighten up the servo after you get it even in both dir's. servo or stepper?
hmmm, lag would indicate the difference also, but no lag in stepper
dont run at 99% of max power either as random failure will result
archivist: PSU wise?
hard to loosen the nut to slide connection? ( adds backlash, reduces crab just a troubleshooting trick)
tomp: ballnut to gantry you mean?
tomp: I can loosen it, but it's really fixed (no adjustments)
1 screw with a wide gantry?
one leadscrew, yes. 24" x 24" 5.5" XYZ
travel that is
sound like the size of the machine I have
the idea: float it (loosen it) , hope it removes the error seen ( you get balanced motion each dir) if the error reduces, then slowly snug it up and keep checking at each snug.
tomp: what, you thinking the X isn't exactly 90 deg?
err, parallel I mean
maybe the coupling face ( ballnut to gantry connection)
a dti to measure twist during travel
archivist: I only have one.
it's small compared to 24" fer sure, and so has to be very accurate ( or adjustable )
tomp: So, if the leadscrew isn't exactly parallel to the linear rails, you're thinking the when the ballnut gets to one end, it's getting skewed somewhat and binding?
no, thats snowplow or cosine error, i think the finger pressure suggests the gantry twists somehow
measure twist in a machine bed with differential levels
sorry for just jumping in, i will scroll up
motion twist not like bed twist, carriage on a lathe twist, not uneven concrete twist
FWIW... I can hear the balls turning... in the "good direction (L2R) they sounds smooth. In the bad direction (R2L), they sound rough.
what is twisting
i scrolled up but maybe i missed it in my skim
yes, let it tell you... thats the way to get at it, be sensitive to it and build a picture, test the picture, when the picture and the real agree, you know the error
can you hear the same on all bearings?
tomp: This is just the ballnut on the X, not on the Y (Z has acme)
so the ballnut sounds different in l2r r2l ?
that's very bizarre
yes, over the whole length of travel
that seems to narrow down the problem components :(
did you take the ballscrew out and run the nut over it
I had taken apart the ballnut a year ago or so, and made sure nothing was inside, cleaned up the return tube that had a jagged edge, and reloaded it all.
to completely eliminate everything else
there should be no diff on the ballscrew by itself l2r vs r2l, sorry sounds like bad ballscrew is probable.
tomp: nut or screw?
or gantry twist causing noise
its more the nut from the description ( trouble with ball flowing in 1 direction brings mental picture of a lip they fall over one way but hang on in the other )
but easy check, take belt off feel the screw by hand each way
archivist: I thought the noise came from then nut itself.
tomp yes but why
return was damaged and hand repaired, maybe not perfect, makes for bind in 1 dir
i agree simples tests should be done and that the ball screw itself shouldnt exhibit this if it is 'ok'
remove ballscrew& nut see if it's same both dirs, full travel. while its out see if the gantry moves nice both dirs fill travel, see if it crabs...
the gantry moves smooth
w/ leadscrew disenganged
ballscrews are preloaded, so it doesn't make much sense for it to be messed up in one direction
toast: No, this is not preloaded
So, if I slow down to 80IPM, no, but at 120IPM stalls in one direction???
torque reduces with speed
so reduce speed to increase torque
My motors: http://www.xylotex.com/StepperMotor.htm
* archivist notes a lack of proper spec there
sorry didnt know this was speed related. thought it was mechanical... is it a notch ( eg: ok at 80 , bad at 120, ok at 150 ?)
tomp: Anything above 85IPM
will cause it to slow in only one direction tomp
tomp: Just stalls more/quicker at higher speeds
rpm rating of ballscrew is ok with rpm at 85+ ipm ?
tomp: I believe these are the right ones: http://products.danahermotion.com/v41/ModelDetail.asp?PkgOrder=161&User=Screw&Rnd=308
JymmmEMC: i dont know, i just suspect the ballscrew/nut. if removes the nut should spin freefall from 1 end to other in both directions. wether that is a good test for >80ipm i dunno.
JymmmEMC: i saw the Danaher/Thompsom screws, a bit familiar with them, rolled threads... no spec on rpm ratings/suggestions/limits, but 80 ipm is a bit quick , its 2M/min... no need to think thats the problem for now.
tomp: prob is, neither do I. I've had problems for two years. Everyone said it was a Xylotex issue, so I went with PArker drives and doubled the voltage. I got better speeds, but the stalling has been the biggest issue
tomp: Hell, I'm not even sure how to calc the RPM at80IPm/120IPM
if 2 yeasr are into it, then pull the ballscrew and nut, hold it by 1 end hanign doen, make sure the nut free falls to the other end ( dont let it fall off ) then try other direction...
carp cant type
tomp: I did when I took apart the ballnut, all was good.
steppers do stall when you go beyond the spec
archivist: in my case, what would that be?
not enough info available to say
archivist: then what else do you need?
i wont join a yahoo group, but the pdfs might have that info from PacSci ) the motors are 'Pacific Scientific')
the pdfs want you to join
tomp: I can get them for you, but they are showing octagon in the pic on the PDF's
not sq like mine are
the right documents will have a curve of velocity and torque
maybe the pdf has the model number & numbers describing the speed/torque relation ( not pix )
tomp: hang on, I'll get the ones on yahoo for you... gimme a few
2:30 am /me ooorf to bed
alex_ is now known as The_Ball
So i've finally gotten all three axies working on my micromill, but now i'm having trouble finding good cam software
you ain't the only one
if we back up... its slow in 1 dir at >80ipm, the ballscrew&nut are verified ok, there's some thought that the motors may not be capable of going this fast , or the driver may not be putting symmetric waveforms at that speed
or is everyone else writing all thier gcode by hand too?
yeah many are
what parts are you wanting to make
for the latter , try swapping motor leads ( force other electronics to drive that direction )
cradek: well, at the moment, a fancy wooden box, it's simple enough a 2.5 d program would work well
tomp: flip aa# with bb# ???
JymmmEMC: wait i got no pdf but do have 3 popups :(
orpheus: do you already have any cad or cam software, or are you starting from scratch
tomp: disbale JS
no need i caught this one, looking now
cradek: well, I've used a couple of different programs with other people's equipment, but it's all commercial, and not having a windows box, or a desire to spend tons of money, makes those look like bad options
JymmmEMC: your's like page 10? got numbers?
cradek: so far, I've played with "cam.py", the wierdly-licenced script written by an mit prof
orpheus: we've got a page of links on the wiki - did you see it yet?
cradek: I also got opencascade built so i could try cam-occ, but it's actually not very far along and useless for writing gcode without further work
tomp: the numbers I have are in the pic here: http://www.xylotex.com/StepperMotor.htm
cradek: I started there, but eventually made my way to www.editthis.info's page on cam software
orpheus: I do my 2.5D with homebrew export from autocad - no good without autocad of course
there's very new work using xfig
recently posted on emc-users list
* orpheus looks
orpheus: vapt http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AptProgrammingForEMC
what kind of fancy is the box? fancy engraving, or just roughing a box shape?
even I haven't been brave enough to learn apt... and I like LaTeX
I like that there's a post on emc-users about 9-axis milling
ya know, for when 8 isn't enough
you won't cut with 9 axes, but they sure might be needed to move things around
JymmmEMC: i cant find your motor in that pdf, some are similar but not nema 23. BUT reversing the leads would force other electronics to drive in the (now) problem direction. that may be valuable to know. I dont know how to reverse yours ( the pdf doesnt show me)
tomp: Yeah, I was looking too, tried searching for the nM rating in the pdf
tomp: to rev the motor, do I switch the a pair with the B pair, or just one legt of each pair
cradek: facy box being defined as a wooden box, top and bottom, made to fit together, fancy because the sides (hopefully) won't be straight.
some cool arc's or something
JymmmEMC: dunno, i was hoping for a pic in the pdf to help
but yeah, i was going to rough the whole thing out of a 2 by 10 chunk I have laying around
orpheus: ah ok, so you want some tangent lines and arcs, stuff like that
apt might actually be decent for that
it will have a learning curve
tomp: ah, ok. I need to take a break, my neck is REALLY tense now. You gonna be around in a while?
otherwise, just get out your slide rule and trig tables and have at it :-)
i haven't actually looked at how apt code works, is it pretty powerful?
i was actually bummed to learn that I can't do trig in emc-gcode-math
i was about to
sure you can do trig in emc-gcode
really? i must have just had it wrong
JymmmEMC: if page 50 is applicable, then the whole set A is swapped with the whole set B ( to make diff electronics drive the problem direction )
but point is, it'd be ten times easier to lay out a shape in 2d, and tell some software to run a radius n endmill around it
orpheus: the manual for apt version used in vapt http://bytecolor.homelinux.org/~bytecolor/apt360-doc/HTML/index.html
that's what radius compensation is for! it's in emc.
that too was on the list of things i needed to learn how to use
does radius comp. take care of corners as well?
JymmmEMC: i'll let this thing logg, wrap a bag of frozen peas around your neck & take a break ;)
well, that shows you how little I know
well external corners. for internal (concave) corners you need to use an arc to represent the material left by the fillet
for an external corner it will generate a very nice arc to circle around it while staying in contact with the corner
ok... so on an inside corner (like if you were pocketing), it fails to do the right thing?
[02:00:59] <tomp> http://www.linuxcnc.org/handbook/gcode/diacomp.html
orpheus: have a look at the comp311_2.ngc sample
shit, this is an AWESOME idea. http://www.youtube.com/watch?v=GRcGfni70CI
orpheus: it does the "right" thing - but you have to draw the real part outline
a sharp concave corner is impossible to cut with a round tool, so you get an error if you try
now that makes sense
basically, it works great if you hand it the outline of what it's capable of doing
yes it works very nice
but if you were to hand it, say, and average vectorized letter of the alphabet to pocket into something, it would more or less die
yes it's not at all a pocketing algorithm.
can apt handle that?
I don't know honestly
yes, apt does pockets
'cause that's about the level of not-having-to-do-the-math that i'd prefer, even if there's a learning curve
of course, it's always nice to just throw an .stl or something similar at a cam program and let it figure it out, but oh well
I sure sympathize - but sadly that requires both windows and $$$
it looks like apt is really quite intelligent
yeah I think it is
after much effort I successfully made one program with it - haven't gone back since then
apt is like emacs/wordstar/vi most modern cad system use apt concepts or even apt code
maybe I should
vapt looks like a nifty tool too
I understand apt does 5 axis too
a pocket primitive used to build fancy pocketing macros http://bytecolor.homelinux.org/~bytecolor/apt360-doc/HTML/pocketing.html#figure103
i can certianly see why many cam systems use apt to do the math too
[02:11:38] <tomp> http://bytecolor.homelinux.org/~bytecolor/apt360-doc/HTML/pocketing.html
The APT processor can generate the cutter paths required to remove material from an area bounded by a polygon of up to 20 sides. This is done by using the POCKET statement, according to the following format: POCKET/ Re, c, f, f1, f2, f3, o, p, PT1, PT2, PT3, ..., PTN
i just have to ask.... i'm a debian guy... is apt.. in apt?
not that i mind
just that it would be funny
it's a fairly new port - not real easy to build yet either
but I don't know what they would call the package
after hand-editing opencascade files to get it to build, i think i can handle apt360
tomp: since you seem to know a fair amout about apt, here's a question
no, please goto http://bytecolor.homelinux.org/~bytecolor/
and see this help(?) page http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AptProgrammingForEMC
imaine you wanted to pocket out an hourglass shape
but the center of the hourglass was narrower than the width of the endmill
(my no is due to being a very slow typer , it was that there is NO apt repository for vapt aptos apt-360 )
can apt handle that?
surely you just get two pockets
I know we've seen this before... but damn. http://www.youtube.com/watch?v=oGq-9NNmr3o&mode=related&search=
the posting would give you 2 pockets
i'm just wondering if it knows how to do that on it's own
yes, it knows to not go in there ;)
so it'll actually make a pair of pockets?
and visit #cam when you can
yes, with a big enough cutter it ought to make no pockets ;)
what all goes on in #cam?
not much tonight, but it's for vapt/apt/apt-360
to be honest, opencascade is a nice framework for doing 3-d stuff
it just knows nothing about cam
and it sounds like apt knows everything about cam, and nothing about actual rendering
did i mention it (vapt/apt-360) can make valid 5 axis code? ( i have no way to test this tho ) and it'll simulate the cut , with offsets?
"Error 153 on card 74"
sorry that should have been in uppercase :-)
what tickled that response?
tomp: yeah, you said so. and i'm afraid my brain would explode if i looked at the code that implements the algorithms
i'm gonna go read un on about 12 different things
of course it's open source, but as opposed to emc, it grew out of IBM System 360 fortran :) and punch cards like in cradek's message
hey vapt runs, cool
where do I get the rest of apt360?
oh I see it on that wiki page
oughta be from bytecolor's page or here http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AptProgrammingForEMC
I should get the cvs apt360 and postp?
with all the python work you've done I bet you have all those 'libs' already
i found that the department of labor has a list
of apprenticeships by state
woot, there are plenty of tool and die places
looks like i've got my full ten year plan
so what's the relation between aptos and apt360?
that Apt looks very 'interesting'
orpheus: i dont know, ones the chicken & ones the egg, but dunno which... i use apt360 and i think aptos is unfinished (more unfinished)
in which ways
unfinished as in you cant get any output but can see soemthing happen ....
apt360/postp/vapt will create valid emc code
so aptos = unfinished, apt360/vapt/postp = working
use postp to create new posts ( EMC is one )
put a 'machine =EMC' directive in the apt file for it to generate EMC code\
what about say fanuc
i don't use any emc machines
fanuc and haas
i can't find the webpages waaaah
'machine=mill' where mill is mill1 thru3 or 4, one of those oughta be good for fanuckians
i dont have a fanuc ;)
one of my co-workers used to write apt postprocessors
maybe he'll remember it if i give it a shot
where's the homepage
fanuc is the model that most other posts are hacked from
i'm going by http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AptProgrammingForEMC
since it'll probably work for anything
not just emc
toast: ditto, use same page to find the others
par of me wants to know how apt works, and another pair of me is afraid to look at it, since those are algorithms i couldn't figure out how to do
i also kind of wish emc had a conversational mode
btw: the wiki shows how i manually added % and M02, the newest postp does that for EMC automatically
but that would be ridiculous to implement
toast: emc has conversational programming ( from JElson & RayH iirc) and gwiz and python scripts too, they all ask q's and output code
like, pockets and stuff?
like the cool thing about the conversational thing at work is that you can do custom contours and it will generate a roughing path
but i'll check that nonsense out
tomp: I got it all running, but when I load test.apt (partno spirograph) I see nothing in the vapt graphics display
and apt can be written with macros & python front ends to do the same...
tomp: I do get gcode generated though
cradek... thinking ( urmph umph urrgg )
tomp: that is hot
but like what i was thinking is a g-code/conversational hybrid
anyone remember the command to generate ./configure from a configure.in?
maybe one day when i am a millionare i will make my own control
* toast leaves for fairyland
i never remembber which of the auto<tab><tab> tools to use
cradek: did the vapt screen open?
yes and it (apparently) ran apt for me, because there's gcode in the tab
and apt output below
type 's' ?
just nothing in the graphics
(simulate the cut )
lamed: ./configure: line 4277: syntax error near unexpected token `fi'
cradek: a quick python question.... to execute a system command.... sys.exec("mplayer xxx.flv") ???
LawrenceG: sorry, not sure
np... I will find the docs somewher
tomp: that doesn't seem to do anything
cradek: ok, will set up and run it to see wahtsup
I don't see anything on the menus etc that would do the simulation
that souldn't have worked!
on line.. let's see.... 4272 of my generated ./configure, there was was appeared to be a spurious ','
i removed it, and all is well
what are the chances?
i mean, should i send in a diff? with just one character changed?
cradek: do you have the emc post? i bet not and your example remonds me, you gotta get the post from bytecolor.... try machin/pmill for now please
yes I get gcode out
cradek: right you musthave it, thats not the problem with no picture... i'm running the same code and it plots for me...
how did you load it?
i ran vapt, then loaded from menu ~/aptsuite/apt360/examples/spirograph.apt it had machin/dxf and machin/pmill you might have chgd it to machin/emc or gotten a newer spiro than i have
i edited a version to be machin/emc and it runs fine too
oh hey! that works
maybe the others really had no output
man they really ought to steal the mouse navigation from AXIS :-)
? a differnt src file?
I used test.apt before
cradek: the poor guy was always admiring axis
tilt swivel & zoom work for you? and 's' now ?
man, i just wish i could build apt
alt-R runs it again I think
can't figure out where to type "s"
me neither ( where'd i get that idea? )
ok when I add a FEDRAT to spirograph, I get gcode that works
under no condition should make fail on a first run, and succeed on a second run
cradek what fixed the 'no plot' in vapt?
I loaded a different file
test.apt and test2.apt maybe just don't work
yep, me too, test3 is ok ( fancy intersecting spheres ) pyopengl is nice
tomp: (I'm still taking a break - thanks for the frozen peas idea) Reversed the electronics, problem is still mechanical. ok L2R, stalls R2L.
and 's' simulate doesnt work, but i always do this, hammer on some fancy bit. i gotta get back to doing useful stuff in vapt & emc ;)
yes test3 is interesting...
JymmmEMC: so i got it straight,.. the symptom stayed the same after reversing the sets, so the err is mechanical. right?
right, L2R still bad after swapping leads... (had to scroll back to see the original symptom was good L2R , bad R2L )
argh, l2r still GOOD after swapping leads
it looks like apt is running postp without a full pathname on the input files, any advice?
"postp -pPMILL -n1 < cl.tap > test.PMILL1.tap"
postp: couldn't open '/usr/local/lib/python2.4/site-packages/postp/PMILL1.pst' file
i didn't mean to paste all that
easier to use vapt and it's file selector than to use the command line versions
how do you even get a segfault in an interperted language?
i mean, i'm sure it's some c-bindings
JymmmEMC: so the rumble in 1 dir does not depend on one side of an hbridge , the rumble is not due to the ballnut/leadscrew (removed & inspected ok ), the rumble is speed dependant... does the gantry move as well in both directions at >80 ipm (by hand)?
orpheus: does the error repeat?
orpheus: I dont want to clutter your system with dumps, but... is it reproducable?
and does the post exist? thats the error "couldn't open '/usr/local/lib/python2.4/site-packages/postp/PMILL1.pst' file"
JymmmEMC: that test would be w/o belt, w/o leadscrew/ballnut, just moving the slide by hand at approx > 80ipm ( 1.3" travel during "onemisissippi" )
err tomp: yes
orpheus: does the file exist?
also, the python-gl GLUT demo segfaults
"python vapt.py" segfaults
do you know if it uses GLUT?
well gl is common in both of those, so maybe a video issue, sorry
vapt uses pyopengl
i take that ack, actually, it's the "tom" set of demos for python-opengl
it does sound like it's python-opengl, not vapt
anyone playing with a lathe? I really want a know the best practice on how to TRUE up the cross slide on a lathe... my lathe is FULLY adjustable.
what do you mean by true up the cross slide
the cross slide or compound slide
it's a schaublin 102.
what do you want to true it to
hang an indicator off the toolpost
it's parallel to the part I'll be turning.
tomp: The last time I dosconnected the gantry and let it "free move", it was extrmely easy, one finger to slide the whole 24" in under 10s
that gets me one axis.
not the other.
you have a two axis compound
I can pull some photos... let me see if I have to take some or not
face a small part using the cross slide
then use that.
[03:40:55] <a-l-p-h-a_> http://lloydleung.com/gallery3/Lathe_Related/
JymmmEMC: if it felt nice in both dirs at that speed, then feels bad at (near) that speed with the transmission connected, it's in the transmission & i dont see where. you just said it was with the trans and no without it. ( trans = motor , pulley, leadscrew & nut ) I agree but dont know where :(
i only see a one axis compound slide
okay... let me get some photos...
let me get this right though
you are saying you have four axes
that you can move in?
your photo album makes me cry
carrage, cross, compound, and compound2?
why the grying at the album?
(also, why can't i type tonight?)
pretty machine, pretty album ( not so pretty led tie wrapped to indexing head ;)
tomp: I lost you on your last statment
JymmmEMC: it must be int the transmission, and i dont know where
a-l-p-h-a_: i think i see the other axis
it's not flash (adobe).
it might as well be =)
also i found a pic with your second axis
it's lightbox, which is just scriptaculous
know how there's a Z and X?
the Z and X angles are adjustable
tomp: Random thoughts: bent leadscrew, one end of leadscrew not square with the other end, ballnut mounted too high near one end, combination of one or more of the abive.
Z slightly... X, is fully (makes the tappers)
unless you have an interferometer or other complicated device, the best way to true it
is to do test cuts
JymmmEMC: can you move the gantry with the ballscrew and not use the motor? some rope wrapped around the pulley like an lawnmover pull ? to see if you can further eliiminate bits?
hmmm... okay... I've got something nice to face... after I get nice faces... I'll go for truing the X (so no tappers, unless I want to)
you are always, always, always goin to get a convex surface
if you get a concave surface, you have a bad X axis
the ways are warped if that happens
or there's sag
tomp: you suspecting the motor still?
doubt that, with the schaublin design... it's just a HUGE chunk of metal...
because with straight lines, you can't produce a concave surface
you have to stop
to do that
it'll either cut from the inside out, or from the outside all the way across the diameter of the part
I see what you mean.
JymmmEMC: no, suspecting we need to cut the problem in half and say the eroor is to the left or right of the cut. i suggest cutting at the belt and seeing if the error goes away or continues
also, you don't have to take the part off
just indicate with the off axis until both sides of the part indicate zero
tomp: hang on a sec...
i'll need to find some drill rod, or something to true up the X.
and use my TDI
not the car.
also when you do the face
it will ride a 0 for half the part, then take off at twice the rate of the taper
so when you make your adjustment, make it just like anything else- half the distance
hopefully it goes quick for you.
somehow I doubt it.
DTI dial test indicator TDI turbo diesel injected
I thought it was test dial indicator
had me goin
tomp: I also have other motors: http://www.applied-motion.com/products/stepper/motors/size23.php
it doesn't matter if you use a dti or a dial gauge
er dial gage
tomp: bbiab, heading to store.
JymmmEMC: too complex, try a rope just for giggles
as long as it's your most accurate doodad
now I'm wrong.
DTI it is
gage? gauge? i forget
gauge, i think i was right the first time
DI is the pluger, DTI is the wiggle one.
di == dial gauge
same doodad, different name
[03:54:02] <a-l-p-h-a_> http://www.littlemachineshop.com/products/product_category.php?category=1310310429&First=D&Last=D
I'm going by that.
widget and doodad
i actually like using dial indicators for truing stuff a lot of times
if they will fit
yes... I ghettoed up the spindle speed indexer. :)
bigger dial, more spring pressure
a-l-p-h-a: i do the same, make it work, then make it pretty
I haven't turned on my lathe in months.
wait, lathe in probably a 9-12 months, mill, in 7 months.
has anyone ever used a birmingham lathe
or one of their bed mills
I GUESS NOT
wooh... removed the ugly white border, that I didn't notice before. images aren't obscured anymore
anyone ever tried xilinx's ise webpack? what's the difference between a "lab install" and the regular one?
completely off topic: http://jmkasunich.dyndns.org/cgi-bin/blosxom/photography/euclid-beach-07-20-07.html
hope you're not annoyed at me.
heh, motion control in fixed point arithmetic
a-l-p-h-a_: no, not annoyed
just having a hard time focusing on coding right now
okay. :) I'm just antsy to get some faster speeds.
my lathe and mill are killing me.
they're so slow
is it a bad idea to let a blackfin (or any other dsp) to do trajectory calculations (in realtime)?
dsp are how many cnc controls work, so
nat bad idea
pmac (delat tau) ,pmd, aerotech, all dsp
a-l-p-h-a_ can you /msg me? or is that an unregistered nick?
a-l-p-h-a_: how slow?
jmkasunich_: are you avoiding me?
for the mill.
I'm avoiding everybody
* maddash goes back to programming his spartan 3
why do you want more than 30 ipm
on that lathe
toast,? what doc?
depth of cut, srt
jmkasunich_, I'm not registered...
oh... that's the MAX speed I get now anyways.
you can't turn at 30 ipm?
never mind then
jepler, I'll reg sec.
my other is... I'll ghost him off.
a-l-p-h-a_ is now known as a-l-p-h-a
so tpAddLine and tpAddCircle use tcqPut to add those 'geometries' to the queue. so the queue holds these fairly hi level ideas (line arc ). what divies them up into servo cycle chunks? tpRunCycle? ( using tcRunCycle to figger the new velocity? )
tomp: search for "Primary_displacement" inside of tpruncycle
nah, I ghsoted the other account... it's my workstation at work.
a-l-p-h-a: how deep are you cutting on your lathe
is it just me, or are 3 axes not enough?
try cutting deeper
rather than faster
or is that your limit
my motor's only 1/2hp
that you previously discovered
you should be able to go deeper, man
even on steel
that's my main stuff.
"you should be able to go deeper, man"
"that's my man stuff."
when i'm tired, sexual euphemisms spring from nowhere
a-l-p-h-a: .5 mm is way too shallow
(i'm not used to metric yet, i have to convert)
I'll try heavier cut next time.
maddash: thanks, i see it (Primary_displacement) and dont understand the idea ... if it's supposed to mean the dx dy dz that is added to where the system was last time, then someone could've chosen better names or explained it.
pew pew pew
tomp: Ok, so this rope thing of yours... I pull the rope and if it starts mowing the lawn, then what?
tomp: ot do you mean smooth balls vs rough balls?
;) to eliminate the motor, the rope allows yopu to approach the >80ipm, and wrapping it 'the othjer way' tests the 'other direction' meanwhile... you 'feel' if its better or as bad as with the motor
actually we have a rope-driven machine
(i am not kidding at all)
tomp: I'm a big guy, if it stalls when *I* pull the rope, we have bigger issues =)
we have a non motor non amplifier affected machine
JymmmEMC: me 'im 6'4" so, i can pull it at about 80 ips, so can you
80 ipm sorry
With these drives, I didn't realize they've been sitting idle.... the motor temp is 110F
steppers get hot when idle ( some have current reduction when not used ,its a function in the amp )
tomp: Yeah, but I have it disba;ed right now
amp disabled? hmmm, another issue, not related directly to the roughness.
no idle at 50% disabled (intentional)
what's going on?
oh, the ballscrew thing
noisey rough slide doodad
noisey bastid rough slide doodad
Ok, I disconnected the belt, if I turn the pulley 1/4 turn so the gantry moves L2R, smooth. If I turn the pulley so the gnatry moves R2L I can feel/hear the course/roughness in the ballnut
think removing the cap from a 2liter bottle,
JymmmEMC: good, the problem is in the components still connected, not in those that are disconnected
tomp: That's even a BIGGER issue... I have ZERO anything to fix it.
JymmmEMC: got a cap right here, gotcha
no spare parts, no mill, no lathe, nada
JymmmEMC: still you've narrowed it down.
no access to a mil/lathe either
I've never touched, seen, felt, a ballscrew, so I never knew if turning the opposite direction should be equally smooth or not.
these last 24 months
JymmmEMC: both directions should be identical, at least, your hands shouldnt feel any difference ( some fancy gauges might )
tomp: BIG difference just making 1/4 turns
is there a gib in the table?
have you ruled out a loose gib
do you think that if you >could< reverse the ballscrew ( end for end ) that the symptom would be in the other direction?
tomp: the ballscrew has been machined to fit one direction only
yes, i want you to imagine
a-l-p-h-a2 is now known as a-l-p-h-a
ok, i'm thinking the err is in the ballnut, but avoid that, and test this: try to do the 1/4 turn without the nut attached to the gantry... same feel? different feel?
tomp: Ok, that's a lil weird todo. there is a machined block the nut attachs to, when you turn the leadscrew this block flops around
gotta hold the nut while turning the ballscrew 1/4 turn cw,ccw
is this possible? hold the block in one hand? turn wit other?
gimme a few, gotta lift the machine up to disconnect the gantry from it.
maybe you dont need to disconnect the gantry from the machine, just the nut/nutholder from the gantry, and while the system is laid over, just hold the nut and turn the screw...
tomp: With the gantry disconnected.... There is no differnce in hear/feel turning cw or ccw no matt the speed I turn the screw
holding the nut in left hand, turning pulley with right.
tomp: I have a special jack I use to hold the machine up.... it's called a 10" 2x4
JymmmEMC: ok, the err is NOT the motor/pulley/amp, it's not the ballnut/screw, it IS the ballnut/holder/gantry. if gantyr move nice by itself and ballnut/screw move nice by themselves,...
then if the ballnut/gantry dont work well together... the answer is not obvious but is in that combo
what can it be, misalignement of nut centerline to travel vector? maybe .. any change of changing that alignment?
tomp: Ok, if the left half of the leadscrew is mounted slightly too low on the left side.... could that casue the ballnut when moving left cause the screw to bow/flex and explain the things we've seen this far?
tomp: To your other question, no. Ther eis ZERO adjustments at this point.
other than the tension of the ballscrew itself
ther's a nut on each end I could loosen/tighten againest the bearings it's in
but that is more or less leadscrew play than anything else
ok no adjustments... and if the ballscrew dives down to left ( -z at -x ) , yes that would cause problems, you might free the end and let it float for some more 'feel' tests
ballscrew misalignment causes a whole ton of problems
(not sure it would cause exactly the symptom you described, but yes a problem )
nothing to free... all machined blocks
remove the left block?
only support 1 end of ball screw for a test
tomp: I have a suspicion that the left end mounting block might be too high, pressing the screw down
your suspicions get better as you get to know the machine more :)
it's awefully hard to toss any kind of guage in there being a ballscrew thread
no gauge, the 'feel'
(use the force luke)
tomp: the leadscrew will "pop out" the right side block
say that again? you got a nut on each end? i imagine pulley blockwithbearing screw nut screwcontinued blockwithbearing
tomp: I have to hit my other computer to see if I have photos... gimme a few and I'll find/take and post some if I can
or... (the "try this" approach ) shim the right block, or shave the left block ( shim is less permanent ;)
no, the error will be the height of nut center too gantry vs the ballscrew center to gantry, and shimming would just bow the screw
so dont do either... maybe shim the nut... in general make all 3 center the same height from top of rail and same distance from edge of one rail, and both rails equidistant.
fenn_ is now known as fenn
fenn: hi, hope you are well, longtimenosee
tomp: This is the Y axis, but same thing on the X http://farm2.static.flickr.com/1032/862245359_e867e2c759_b.jpg
got thi pic ok
hideyho.. just passing on a song for alex
Jymmm: i see the nuts you spoke of, i was thinking ballnut (doh!)
i got a sysadmin job :\
tomp I understand
fenn: great! hope it pays enuf for your cnc hobby
fenn poor bastard, at least it isn't a tech job
Jymmm: any guestimate on the drop from ballnut to end of ballscrew ( how much it dips )? and any way to shim the ballnut mount by say 1/2 that ? (up to all of it )
btw: the machine looks real nice
| the end mounting blockers with bearing and end nut
/// == leadscrew
==========this is the fixed frame=====================
|/| == ballnut/block
ok, makes sense
Jymmm: and you suspect this...
==========this is the fixed frame=====================
where left block is differnt than right block?
looks like you might shim the right block ( and not the nut )
close, more like....
==========this is the fixed frame=====================
dots replacing spaces...
==========this is the fixed frame=====================
ok, then shim the left block down , seem like it may work?
tomp can't, the gantry mounting block is fixed too
somethings gotta move to make their centers in line
hold on, uploaging photos
and that seems to be the bore in the left block.. ok will look
tomp: many places just line bore that kind of thing
yes, the prob here is the bore aint a line ;)
it's a lightning bolt ;)
unproven lighting bolt
i'm just saying why it would be unadjustable
Jymmm: true, something led you to think the left end 'went down' tho
toast, yes unadjustable would be good
only because when I move R2L is when the stalling occures
nah, most big stuff is adjustbale
nonadjustable is bad
material ages and settles
sometimes the machine gets bumped on the table and the bore is off
Jymmm: right ( i needed reminding ), and that could be checked by a block slide between 'fixed frame' and 'screw'
toast: i meant it'd be ok if the line bore was good, but yes, things change.
pew pew pew
Jymmm: got it...
do you have drill rod that will fit in the holes
the ones that mount the ballscrew
there's a millwright method to tell you the angular and linear error
between the bores
in the middle? in the ends?
but it only works somewhat close
like, a tight fit
like sort of a gage pin
a looong guage pin
I dont have any drill rod, some SS rods
how much travel is there
if you get like, two 14" spans that fit dead nuts into the holes
and go close to one another
and that you can fit dial test indicators on
you can measure all the errors
are you taking remove the leadscrew completely and tossing in a drillrod in it's place to measure?
two drillrods, actually
one for each hole!
that is how you check shaft alignment
i see the whole leadscrew assy mounts onto the extruded aluminum strut, and than can be shimmed (block to strut) to align the screw centerline with the rails
oh, the C channel you mean?
also looking at that the diameter/length ratio is probably too long
sorry, i mean c channel to extruded, shim betweenthose 2
toast yeah, nothing to align againest.
tomp inverse what you just said
Jymmm: no no, i mean checking it
not aligning it
just seeing if it's out
tomp I *TIHNK* when the ballnut mount is attached to the gantry plate, the gantry plate is puching UP on the leadscrew causing it to bow on the left side, which is lower than it should. - at least that's my theory. the thing is, even if that was the case, I have no way of fixing it.
tomp This POS I think is too tall http://farm2.static.flickr.com/1040/863196994_17db5696e6_b.jpg
is the machine level
if thats too tall, then in http://farm2.static.flickr.com/1213/863197056_71deace0d0_b.jpg,
shim between 'gantry plate ( thingy with nice webs) and the vertical side plates... thismakes gantry plate 'lower'
Jymmm: i can easily be wrong but it looks like shim to me ( to lower the plate stretching rear to front that is undercarriage of gantry )
and shim can be undone easy
toast: yes it should be level, ways straight and axis square
shim Jymmm :)
Jymmm: hope it works out, you have really narrowed it down to howto from wtf. i gotta crash, gnite!
tomp: If I shim the leadscrew mounts on (each) end, that will put down force on the ballnut
you CAN mount a machine that size without stressing it
it takes some fabrication
and is probably not worth it
mount i.e. shim
oh i misread
i'm sorry, nevermind
pew pew pew
Good morning! Can somebody point me to a supplier of cheap milling tools (for PVC, PE, Plexi-Glass), diameters 0.5mm~3mm?
My mill will work in a couple of weeks so I start searching for tools. Austria and surrounding countries preferred.
Can someone point me to online shops, please?
Are they high-priced?
you're doing plastic so you can use a cheap endmill
two flute, high helix
if you can
Are for plexiglass special tools required?
however kerosene or wd-40
will be an amazing lubricant/coolant
It always was melting, then "click"... :-/
that's the problem
you need to run it faster, with a lighter feed
and you need to run kerosene (preferred) or wd-40
right to the tool
and a two flute mill
Oh, forgot to mention that I cannot use liquid coolant.
also what tool were you using
Mill has nothing to cath the coolant.
uh, i'm saying why can't you use coolant
plastic needs coolant to cut well
otherwise you deal with warping and melting and all horrible other things
i personally would use a two flute endmill with kerosene
I currently use a borrowed mill and I found out that low rpm and only 0.5mm cutting depth + slow advance work fine.
right because you keep the heat down
and the mill doesn't really melt it any
anything will cut well if you pretent it's air hard tool steel
Well, I am in no rush, it can work slowly...
but a two flute endmill will give you tons of clearence
in the gullet
so the tool won't bind
high helix increases the chip removal
which is incredibly important in plastics
There are fishtail tools available. What's with them?
i'm not familiar with that particular phrase
I only have a name: "Think & Tinker, Ltd."
i can't find anything that really stands out
oh, it refers to the point somehow
i have no idea man
three flue is more like
i would cut plastics with one flute index endmills
but you probably don't want to for the money up for that
bascially the idea with plastic is to get it out of the cut as fast as you can
when you are hogging it off, that is
otherwise you go slow and try not to generate any heat
which is what you've been doing, as you said
[07:52:34] <xantipe> http://www.thinktink.com/
<-- This is the link.
i get it
There go to the tools (right), then choose endmill cutters...
yeah i'm already there
like, you can use those?
but you are at the very, very extreme
of the cutting spectrum
in terms of plastics
those endmills are to help the folks at the very other end - the very extreme of speed
to help them go even faster
because the bottom of the mill rubs on the plastic
as it springs back
and melts it
like, once you get into cutting at 80-100 inches per minute
then you're going to need those mills
Doesn't the tool for soft media (http://www.thinktink.com/stack/volumes/voli/store/endmill.htm)
prevent exactly this?
like i said man
No flat tip, so material is not rubbed and melting?
you don't have that problem
you are not pushing the right tool fast enough
to encounter what they're talking about
you can use that tool though
no ill effects
you're sitting right before this weird middle zone
Also, they appear expensive (for my needs).
where the plastic isn't being removed from the cut fast enough
or with enough coolant
plastics need a coolant
or else they'll weld to the tool
bind, and break the mill
Air-cooling (compressed air) is no option?
but it's not really the problem
you can try a TiN
or a TiAlN
What is TiN and TiAIN?
but really for what you're doing you don't need TiCN/TiAlN
all these are coatings
it's like a solid lubricant
it decreases surface friction and "stickyness"
and also increases the amount of heat a tool can handle
TiN is gold
Ok, got it.
TiAlN is like, a brown color
and TiCN is blue.
but air won't help
that's not really the issue
also go full hog into it, you need to really throw the chips clear
and you may completely snap a couple endmills trying to find the right parameters
but you really need a two flute endmill
and i would not try to use an endmill under a 1/2"
and spin it 3000 rpm
or more, if you have it
I thought that air would help getting material out rapidely from the cutting zone.
that helps it clear after it's been cut
but doesn't help the actual cutting processs
i'm talking about when the material is still partially attached to the block
trying to keep it free during that process
that's when it builds up
But you are right, less cutting fins will be most important.
it's like trying to cut aluminum with a 4 flute
same kind of deal
Well I have up to 18000 rpm
Well I have up to 18000 rpm
what size endmill are you planning on using
if you have a vortex tube that might help the heat
This is a converted engraving machine from gravograph, Model VX92.
This is the problem. Shank not more than ø3.2mm.
Besides: Does MSC ship worldwide or only USA?
Besides: Does MSC ship worldwide or only USA?
i keep disconnecting here.
not sure - i believe worldwide, they're a huge company
Strange internet connection you use...
it's just the past couple minutes
and it's only this server
everything else is fine
but yeah dude, 18000 rpm, a vortex tube, the fewest flutes
MSC has a office in Europe, too?
whatever their website says
What is a vortex tube?
compressed air in, frigid air out
folks have been picking them up where wet machining is unacceptable
Don't you think this is a little bit expensive? ;D
no, not really
compared to other materials it's pretty cheap
not sure about the endmill
vortex tube i think is under 150 bucks
but they never break
but if you can't use the cheap solution - wd-40 or kerosene
then that's the only coolant option you have
where was i
you're probably talking under 200 bucks
... funny ...
for the vortex tbe
and that is cheap
Not so bad.
at work we use boring bars that cost 2000 bucks each
but like, that tube
will do you good
especially for milling work
If I use liquid cooling, where is usually the tray located? Below the monting plate?
you just spray it on?
for wd-40 anyway
no real catch plate
and for kerosene, you use a brush
and just brush it on
or put it in a small squirt bottle
and squirt some on
which is a better option
and just squirt intermittantly
you only need a tray if you have constant-flow coolant
the coolant will coat the coolant and get kind of sucked up, no puddle
coat the chips
So this is no permanent splashing coolant?
it makes a tiny mess
as opposed to a huge mess
you're throwing just enough on to keep the cutter lubricated
rather than trying to keep the cutter submerged in coolant
and wd-40 doesn't make much mess at all
comes in a spray can
and is a great way to just test stuff out
Ok, this might be possible. I could use a flat tray below the T-slot plate (don't know the right name) to collect what is rinsing down...
it won't rinse down
So less is to be used?
you're not COOLING stuff
a-l-p-h-a2 is now known as a-l-p-h-a
you're just lubricating the cut
most of it smokes away or clings to the chips
there's nothing left to drip and cause all sorts of nonsense
i use messy coolant for pocketing only
Oh, I see. I thought you meas a strong flow of coolant...
that's for very heavy machining
you are nowhere near that yet.
At school (20 yrs ago) I manually milled, but n other dimensions... Therefore I associated this dimensions with here...
just a quick spray of wd-40 every so often.
depending on how it's cutting, which will take some practice to learn
wd-40 is oily?
wd-40 is a lot like kerosene
it's light petroloum distillates
it's a bit oily.
So glueing after cutting will become probelmatic?
not if you clean the surface
which you should be doing anyway
can you use rubbing alcohol on the plastic?
do not use benzine if you can help it
use rubbing alcohol
and that will give you a very clean surface for the glue to adhere
the only reason i ask is some people are using plastics that will melt with rubbing alcohol
so yeah man
give it a shot, wd-40
then clean up after.
should come out real pretty.
Now I must only get my mill finished. I just download EMC's iso now. Must complete 2 additional stepper drives and the power supply, then put everything together.
cool man =)
anyway it's about 4:30 am where i am
so i am going to bed, goodluck
A lot of work. -- What? You are still up at 04:30? -- So you are from the States?
Hello, is there a guide for installing emc on other distros than ubuntu, especially OpenSuSE 10.2? I don't really like to start now with ubuntu and its different features.
Did nobody try that?
me? I wish
you would actually have to do something to be hung over.
Skullworks-PGA1 is now known as Skullworks-PGAB
And well, Just too stressed/frustrated/discouraged right now to drink
hmm was there a small mushroom cloud in your recent past
Skullworks-PGAB: Went thru a 4-6 hour "narrow down the issues" session with tomp yesterday, Seems to be in the ballnut/mount/gantry which 1) I dont have any spare parts for, and 2) I don't have a mill/lathe (or access to one) to fix even if I wanted to
that is what was binding and causing the lost steps?
Skullworks-PGAB: Well, binding at least it seems. stalling, not lost steps
guess you need to find a small shop that will do little odd jobs on the 6pak exchange medium.
Still waiting to see, but it looks like my motors *MIGHT* have speed/torque issues too.
Skullworks-PGAB: The difficult part in all these is there's really no way for me to measure what needs to be changed. So even if I did find a shop, not sure exactly what to tell them.
Skullworks-PGAB: I really dont have the precision measuring tools
you have a 0-1 mic?
I have a 6" .001" dial calipers and a .001" DTI
caliper would work
Skullworks-PGAB: not easy at all to measure distance of a ballscrew though
could you shim up the ball screw end mounts to match the ball nut mount - or if they are too tall, shim the ball nut mount - shim with sheets of paper an measure final thickness that works.
Skullworks-PGAB: s/shim/shave/ I think
if you look at that pic, you can see the disconnected ballnut/mount that connects to the gantry underplate.
random question: which CAM apps do most people here use with EMC? that are reasonably cheap (sub 1k USD)
I THINK it's too tall, and the left side of the ballscrew is slightly too low, so when the ballnut is moving towards the left, it binds and bows the leadscrew upwards.
unrelated.. Though, I think I found where some backlash was coming from; the leadscrew had some side-side play, tightened up the end nut now.
* Skullworks-PGAB uses Bobcad - but thats mainly because I got into it back in 91 and I KNOW it.
jym - was that a kit?
Skullworks-PGAB: No, hang on...
[15:00:30] <JymmmEMC> http://k2cnc.com/Machines/CNC-router-KG-2525_detail.asp
so you bought is as a turn key unit
Skullworks-PGAB: No electronics
or - less motors/
(that option is no longer offered, I was the last one)
they should have fitted motors and tested it prior to shipping
Skullworks-PGAB: I shipped the whole machine back to them with my motors. They shipped it back to me saying "all is well"
* Skullworks-PGAB notes K2 must have low standards
Skullworks-PGAB: I have a box of machined pieces that they had to re-do as the the maching was off. I must have at least 4x on some of them.
Mind you, I have never touched a mill in my life, but even I know better than to clear the swarf off before clamping down raw stick so there isn't a beveled edge so it's not a 90 deg right angle
shit like that
another point is you can't trust 8020 type extrusions to be straight or flat
Skullworks-PGAB: the extrusions aren't milled flat?
the 80/20 is the least if the issues, it's the milled pcs.
though they'll warp if you machine them...
Guest657 is now known as richarda
JymmmEMC: keep the investigation simple, you know what feels right and what feels wrong. try loosening up the undercarriage cross piece and re-feel. if it's no different, retighten and try loosening left end of channel, if no better, retuighten that & try right end. one at a time, simple tests, not much time.
tomp: What I suspect will end up happening is finding that the left end leadscrew moutn is snafu'ed. I'm just not sure how to go about fixing it once I've confirmed that to be the case.
maybe loosening the bolts can confirm it
tomp Well, what I expect I'll have todo is remvoe the entire X axis, drop out the C channel then replace ans THEN I could shim up the end and ballmounts.
I need to confirme the leadscrew is true first though.... anyone got a granite flate then could send me =)
you know best, it looked as if shimming c channel ( and other bits) only rqd loosening and inserting the shim. to check the straight... a machine table is pretty flat, with 2 Vblocks and a DTI you could check ( or a lathe way )
mshaver: hi matt!
tomp: I have nothing like that available to me, other than the DTI.
DanielFalck: I really liked your apt for emc stuff
it's not really mine, but I try to help
an extra rail like what you used as x slides? ( they are very straight )
use one thats mounted, you just want to lay the screw on it and look for daylight/feel with gauges
tomp: good idea, hoping that IT isn't warpped either =)
well, it is the thing you have to be same as, and if it is warped, then find out now. ( a machinists square's edge will show that up )
an indicator mounted on one bearing block can be slid along it's rail. measuring the other rail ( checking deviation between the 2 rails )
an indicator mounted on one bearing block can be slid along it's rail. measuring the same rail ( checking deviation along its length )
again, the bank waits for no man .... bbl
JymmmEMC:can you improvise vblocks - clamp two rules, metal strips, whatever together approx 90 deg?
skunkworks: hi, how is it going?
richarda: I might have some 1" angle aluminum around here, (maybe U channel)
following on from what tomp was saying,
if you have a plain journal at each end of your shaft
support it on said vblocks and rotate
if you use the al angle/channel, put a spot of oil at the contact point
humm, not even an echo in here...
I opened my mouth and everyone stopped talking!
chuckle, been there, done that
make one feel like he farted in church or something.
what you playing with at the moment?
a routine that can only be done simply if while's could be nested, but they can't. Its a pita.
I'm lazy and let the CAM spew hundreds of lines
but then I don't consider myself an expert, it may be possible but I can't grok this job without that ability
who's got a CAM?
it's a commercial one; onecnc
ahh, no got.
the more I do, the less I use it!
I need to drill a hole, 1.3" in diameter, thru a 2" thick block of alu, with a cnc controlled micromill that only has enough moxie to cut about .010" at a time.
So I was going to write a loop to cut from the center to the edge in increasing arcs, and wrap that inj a loop to push z down by .010 every pass thru the other.
Thats a hell of a lot of typing without being able to nest it.
with nesting its about 30 lines of code.
you need a G13
circle pocket macro
lemme look that one up
Humm, not in my copy.
emc2.16, is my doc book out of date?
anyone here rig up a pendant with EMC?
I need to port it - make an O-sub to do it
or at the very least... a jog wheel?
Ziegler: Pseudo pendant using an external NUM_PAD
Ziegler: on a very LONG cord =)
Are you saying I can make the inner loop's while function as part of the o100 sub and getaway with nesting it that way?
Ziegler: It allows me to be at the machine at least.
might - I have not tried it
I'll give that a shot, back in a few.
Ziegler: here is my jogwheel setup http://www.anderswallin.net/2006/11/jogging-emc2/
richarda: I unmounted the X axis, soaking it in charcoal fluid right now to get rid of all the grease.
I have written a G13 macro for Fanuc - since it was originally a Yaznaq command (also hijacked and expanded by Haas)
very nice awallin
G13 won't normally work on Fanuc - but by writting the macro and naming it O9001 I was able to direct G13 to O9001 and pass all the augments
I think I can do the same for EMC - just not as smoothly
not sure if I can pass augments in the interp
jepler: hows this going? http://axis.unpythonic.net/01174426278
might be able to pass variables
How come whenever you look for a clamp, you have a zillion of the too small ones, and can never find one of the right ones?! eeeesh
If you guys had $900-$1K to spend on a 3 axis system... what driver and steppers / servos would you be looking at?
at least 400 oz/in ish
speed... the more the merrier
(at least 80 ipm on rapids)
* Skullworks-PGAB wants a bigger budget
the cash.... is only for the driver and steppers
ha, yeah right
I doubt it for servos
gene: emc2, nc_files, examples, useful-examples.ngc - chris' helical hole milling any good?
$1000 for machine - $1000 for 3axis ballscrews - $1000 for Servos, amps and PWS - and I'd still likely come up short on cash.
Specifically, I have been "given" up to $1500 to spend just on the driver and motors
Ziegler: NEMA size?
that is flexible
NEMA34 it is
[17:45:49] <JymmmEMC> http://cgi.ebay.com/Qty-2-NEMA-34-425oz-in-200step-Stepper-motors_W0QQitemZ110150073544QQihZ001QQcategoryZ71394QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
Surplus center has a bunch of Ametek servo/generators in dual shaft config which can accept an encoder
Ziegler: brand new http://cgi.ebay.com/3-NEW-Steppers-Motor-NEMA-34-85BYGH450B-665oz-in-CNC_W0QQitemZ320134284348QQihZ011QQcategoryZ78196QQrdZ1QQssPageNameZWD1VQQcmdZViewItem
good for about 400 oz-in
665 oz in
at 1:1 you could get 200 IPM rapid
add Us digital encoders
[17:48:30] <Skullworks-PGAB> http://www.surpluscenter.com/item.asp?UID=2007062621461319&item=10-2351&catname=
get a PICO systems servo amp pkg
build a 20amp 24V unreg PWS
your all set
what is a good way to gauge the torque capability of a servo?
Ziegler: place your foot under steel spike, apply voltage to motor. if you can still walk, trya again with another motor.
Hey, could someone help me with the math to determine RPM's @ 150IPM?
200SPR * 10TPI * 5x microstepping = 10000
150ipm and 10 tpi?
doesn't that seem low?
so 1500 threads passed, ie revolutions, per minute
150 inch/min X 10 REV/inch
look at the torque cruves http://cgi.ebay.com/3-PCS-NEMA-23-DUAL-SHAFT-STEPPER-MOTOR-425-OZ-IN_W0QQitemZ120142062953QQihZ002QQcategoryZ78196QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
1500rpm with a stepper, not good
pps? pulse per second?
goes to shit faster than 3000 PPS
I'm assuming that's what it is lerneaen_hydra
so shitty after 15rps, 900rpm?
Ok, lets back up a bit....
IMO it's shitty once it's even started moving ;)
HOW do you calc RPM's at 150IPM ?
150 inch/min X 10 REV/inch
* or / ?
150 * 10
1500 RPM's ?
the 10 being your screw pitch
Ok, now the PPS (if that's what it is?
isn't that quite fine pitch for a high speed screw?
200 steps per rotation?
Ziegler: motors, yes. 5x microstepping
* Ziegler used an 18 pitch screw
ok so 5 x 200
leadscrew are really 5TPI on 2:1 pulleys == 10 TPI
1000 pulses per rotation
JymmmEMC: ah, ok
Ziegler: need PPS, not PPR
5 x 200 x 150 = 150,000 pulse per minute
2500 per second
at 150IPM ?
Hmmm, tha't s JSUT within the boundaries of this motor
5 x 200 x 150/60 = 2500
.5 Nm ?!
(5 microsteps) X (200 steps per rev) X (2.5 Rev per second)
2500 Pulse per second
man, these really drop off at higher speeds
0.35 N-m (50 oz-in)
so .7Nm ~= 100 oz-in
ever used googles conversion?
[18:11:14] <Ziegler> http://www.google.com/q=100+ounce*inch+to+pound*inch
google was 404??
that link was you gave at least
[18:12:59] <Ziegler> http://www.google.com/search?q=100+ounce*inch+to+pound*inch
now try Nm
buggers.... google is stupid
hehe it will do pound-in to Nm
[18:15:20] <Ziegler> http://www.google.com/search?q=100%2F16+pound*inch+to+N*m
there... just divide the number by 16
[18:15:40] <kanzure> http://search.yahoo.com/search?p=
also works ;)
but they use 'p' as the variable rather than 'q' hehe
[18:16:28] <Ziegler> http://search.yahoo.com/search?p=100%2F16+pound*inch+to+N*m
I wonder if that works
didnt for me
I've been using Yahoo search more frequently recently, actually, because they let me connect with their web API :)
yahoo doesnt seem to have a convert feature
Google dropped their web search API. And they had a 1k queries/day limit.
anyone have an opinion of FlashCut ?
oh yuck... its USB interface
a-l-p-h-a2 is now known as a-l-p-h-a
ah this was just too funny - http://ar15.com/forums/topic.html?b=1&f=5&t=599655&page=1
advantages of bipolar steppers ?
very funny Skullworks-PGAB
Ziegler: never a dual moment ?
* alex_joni is back home
alex - I have a Open Office .odt that might be wiki material - possibly under homing.
Skullworks-PGAB: wiki allows uploading
does it have to be feformated?
hmm - ok
Skullworks-PGAB: I guess who really wants, can convert it
you could paste it into the wiki.. but it depends what it is..
(maybe export it as pdf if it's schematics/pictures.. it's more universal)
no pics - but there are some active links to pdf's
pdf's of schematics and hardware maps
Skullworks-PGAB: what is it?
It details the CMC Servomate SM1
a 3 axis servo amp used in some Bridgeports and Hurco machines
If I install ubuntu server, then later on want to add X, will it add all those dependancies like desktop usually does?
it has an issue which makes it difficult to work with EMC
that being limitswitch and homing
I wonder where awallin is.
I am hosting the PDF manual for the Servomate and wrote up a description of how it works when homing
this is so people can brainstorm a workaround.
* alex_joni thinks anonimasu has psychic capabilities
anonimasu: there was a nice swedish jazz band at a festival last week
* Skullworks-PGAB is 1/4 swede
thought you might know them 'viktoria tolstoy & band'
[19:58:52] <alex_joni> http://bp2.blogger.com/_6A-p9iEvES8/RomQY0jov0I/AAAAAAAAAZo/VZTxrSNcCLY/s1600-h/bollards-dublin.jpg
Im a bit dense at this moment... why is it funny?
Ziegler: nope, look again
or may be imagine the workflow, when they finish, pack their things and try to drive away
maybe they can still get out lower right
I guess the point is not to allow cars in there
wait never mind.. just moved the pic over
so I doubt they can drive around
well, I hope for them, some of those bollards release
anyone respond to my ubuntu server question?
are we well?
I finally stopped responding to morons on cnczone
$wife has taken kids to Scotland for the week, thus I am free ... FREE!
well, cnczone, its where they live
JymmmEMC: easy to add X
* robin_sz wonders if that moron on cnczone he was responding too got the hint
Ziegler: Right, but I DONT want to add all the extras that desktop usually adds.
$wife was also out buying that sodding harry potter book at midnight
Ziegler: No sound card, so dont install all the player crap (as example).
just install x
or install X-ubuntu
hardly has anythign with it
I mean, I like a good story, but there has to be some minor level of reallity in it to make it believable ...
kids who can do magic ... well, maybe
flying on broomsticks .. wellm pushing it a bit
but a ginger kid with two friends? ... you must be kidding
* Ziegler blinks
robin_sz: Hey, ppl fly off the handle all the time (in reality), just make it a broomstick handle for dramatic effect
JymmmEMC just "apt-get install Xorg"
Ziegler: I hope that xubuntu isn't anything like kubuntu with the flaky installer that's borked.
didnt "bork" for me on a 450 mhz 64 megs of ram system
JymmmEMC: I installed xubuntu several months ago and it worked
Ziegler: k or x you installed?
was lighter weight for an older machine
I was spekaing of K borked, not X
[20:20:26] <Ziegler> http://mirror.anl.gov/pub/ubuntu-iso/DVDs/xubuntu/6.06.1/release.1/
I know JymmmEMC.. and I was letting you know X didnt die on me
I chose the alternative install
JymmmEMC: I don't think so
[20:21:41] <Ziegler> http://mirror.anl.gov/pub/ubuntu-iso/DVDs/xubuntu/6.06.1/release.1/xubuntu-6.06.1-alternate-i386.iso
that it pulls in the desktop stuff
alex_joni: you know the dependieces I'm talking about?
Or DDT.... Diet Dr Thunder (walmart brand of Dr Pepper)
alex_joni: robin_sz http://www.kelinginc.net/KL23H286-20-08B.pdf
JymmmEMC: yeah, I guess
you'll get some dependencies, but hopefully not a full bloat
* Skullworks-PGAB needs more Dr. Thunder - have gone below my 4 case reserve limt.
how's the weather over there?
* alex_joni is melting here
we had close to 110F today
now it's 94F and it's close to midnight
[20:37:52] <Skullworks-PGAB> http://weather.yahoo.com/forecast/USCO0078.html
in the UK there is only 1 word to describe it
may was wet
june was wet
july is wet
flooding, rain, more flooding
for an Island they do seem to have drainage issues...
OK a really big island
[20:41:57] <alex_joni> http://www.weather.com/outlook/travel/businesstraveler/tenday/ROXX0040?from=36hr_fcst10DayLink_business
seems it's getting 'colder'
[21:00:40] <richarda> http://video.google.com/videoplay?docid=-5437209871684663077
is that cool or what?
hmm - wonder what kind of MPG it gets...
a pair of Schrekling turbines? Or wrens?
jmkasunich__ is now known as jmkasunich
pew pew pew