so i was verifying this program the other night and in the middle of it i noticed where a 90 degree corner in the toolpath had been translated to a radius in the axis PATH display. every other program i verified the code in did not display the radius including axis' own GCODE display. is this something that has been observed before?
are you talking about the backplot or preview?
the preview was correct as per the program g code.. i was running it realtime with my motors turned off.. and the live toopath backplot was not correct...
i can share the code if needed
the backplot will show the rounded/blended corners but the preview will not.
as per planned?
also the backplot is less smooth than the real machine motion
but this was supposed to be two lines intersecting at a 90
I'm only guessing because I didn't see it...
and it cut a trace on the circut board i was milling right in half so it had me concerned
oh, so it was the real motion
i should have done a screenshot but i had to go to work...
yes real motion...
for pcbs do you use G64 Pxxx (tolerance mode)?
well, not real motion, it would have been real motion had my motor drivers been switched on
I bet you are getting more blending than you want. maybe because your accel is set low.
if you can do it again, I'd like to see a screenshot - I'm only guessing now.
i will run it again and take a shot.. i'll let it run tongiht while i'm snoozing
i might do the ink pen and paper test as well to see if it comes out in the actual motion
*should do the ink/paper test
the code i was running was from pcb_gcode and then i used codeshark to rotate the x/y plane 90 degrees...
the code looked right...
out of curiosity, how did it cut a trace when the motors were switched off? (or was that separate runs?)
if I understand correctly, the backplot line went through a trace unexpectedly
the virtual trace that would have been milled had the program been run with the motors on
a hypothetical trace if you will
the backplot is sampled at a much lower rate than the motion actually happens
if you lower the feed rate override near some corners, you may see that they're sharp
SWPadnos, you think it's just a resolution thing?
on my screen*
sampling, not resolution
ok.. that would make sense. maybe i shoulda waited until i did an actual motion test to bring it up.
AXIS samples 10 or 100 times/second, so if the corner is turned between samples, you'll see a cut corner
even though the actual path was square
(or mostly square anyway)
so it could be bad gcode [fixable], bad sampling [no problem], or excessive blending [fix with G64 Pxxx]
I don't want to guess which it is
or it could be an error, but that seems less likely than one or more of the above :)
yeah i just ran across the g64 when i was looking for related problems.
sure, or an emc bug I guess, but that's pretty unlikely
i was just wondering if anyone had ran across the behavior before. i wasn't making any conclusions yet!
I've seen all three problems at various times :-)
if anything i would have guess that codeshark screwed up the math when it rotated the xy plane
but i'll run the motors tomorrow with an ink pen ;)
the preview would generally show any bogus gcode
or slow down the travel rate near a corner and see if it sharpens up :)
well thanks for the input cradek, SWPadnos...
sounds like bedtime for me. good night
happy 4th n stuff
I hope I didn't screw up, I clicked on upgrade to 8.04 LTS in Software Updates :(
ah well, please tell and report the result
will do, I see that 115 updates are available now and the upgrade is not even done
I should quit clicking on things this early in the morning on a holiday
BigJohnT: you did that on a 6.06 + emc2 machine? let us know how that goes ..
looks like it will take a while
support for some applications ended but EMC is not listed...
I don't know whether you'll get a working emc2 after the update (probably not), but the other problem is that you can't edit the docs with the version of lyx in 8.04.
(it will write out a file that can't be read on 6.06 machines)
It just gave me a chance to bail out :)
I'd chicken out, given the chance
It said it would take 19 hours and 31 minutes to download 574 new packages and upgrade 918 packages
* BigJohnT says cluck cluck :)
good morning and happy holidaze!
Ill check my logs to check for BigJohnT swearing on Monday
jepler, recall you and others were helping find my Axis realtime delay error? i think it's related to classicladder somehow.
using the same configuration, i backed out all cl stuff and do not get the error.
i'm going to slowly add back in the cl and see where it starts to error.
i know there is a cl graphics error already. after running with the cl gui for a while, the gui dies.
archivist: I bailed out
YES and also the problem with lyx
I hates upgrades sometimes
apt-get segfaults on the debian box Im on now
Greetings. How much HD space is needed for a installation from the standard ubuntu emc2 cd?
Ive added a bit but 3543304 1k blocks used on my box (I have apache and mysql taking a lot)
eg my use dir is 1022852 of that
hmm, that sounds kind of large coming from a 700MB iso, but i guess installation onto a 2GB compact flash is a no go then
dunno Im likely to have selected more then the minimum during the install
Ive added some perl modules as well iirc
DaViruz, probably worth a try at that size
okay, i'll give it a shot then. thanks
you should be able to get an installation onto a 2G disk, but it isn't as straightforward
you'd have to start with the alternate or server CD, then install EMC2 using the install script
DaViruz, just remembered another large install on mine was the dev utilities so should easily fit
I'm not sure how much of the whole desktop environment you get from that (things like openoffice, firefox, etc)
and man pages amd languages , there a lot that can be thrown
SWPadnos: well, i don't need any of that on my cnc controller.. :-)
it looks like the package doesn't directly depend on any of the large metapackages (in fact, it doesn't directly depend on gnome or any UI)
SWPadnos: it does pull in python2.5-gnome2 libgnomecanvas2-0 libgnomeprint2.2-0 libgnomeprintui2.2-0
right. I assume those depend on some UI / gnome elements
(but they're not direct dependencies)
i have a quick question for some g-code guru, as i am quite the rookie....
mdynac: ask away
hopefully you'll find that there's a guru in the room
mmmm beets http://emergent.unpy.net/index.cgi-files/sandbox/img_7476-medium.jpg
how would i repeat a 3 line program over and over...
jepler: what kind of time is involved in making that pcb ( from end of pcb design to board milled & drilled )? I'm wondering the efficacy of milling prototypes vs breadboard vs commercial service.
breadboard isn't so great for SMT parts (though that board doesn't use any), and even a few hours is less than a day or two
which is as fast as you can get a quick prototype
tomp: oh it must have been less than an hour, and I still don't have "the process" down
great, that says a lot thx all
mdynac: let me see if I can whip up an example along those lines..
but if a complex board for production I would use a commercial service (pth and smt and small features)
loop examples using Owords http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Oword
( not the Oh*ll Osh&t Ocr@p words :)
oh - milling isn't suitable for production, unless you're producing for yourself and some riends
archivist SWPadnos I agree but never made pcbs for smt so far ( or poked myself in eye with pointed stick on purpose )
some smt parts are doable by anyone who can solder through-hole competently (SOIC); but small stuff gives me fits.
tomp PCB design was the day job a few years ago
ooh (AXIS.stop). nice! is there an AXIS.start and/or pause or color=green?
and re: pcbs' i agree i'm still ttl :)
tomp: well, there's kind-of a "start"; you can (AXIS,hide) and (AXIS,show) to suppress part of the gcode in the preview
thx, thats real cool
hand soldering smt 60 pin ics is "fun"
i had just done 4 of pminmo's stepper drivers each with a half dozen smt caps and res... too damn small for these old eyes ( but i did it and they work :-P
you haven't lived until you've done (a) 0603-sized dual LEDs and (b) 8-resistor respacks in 0402 spacing
mdynac: do you follow what's going on in that file? or is there a part I can clarify?
jepler: i get it, cool. is that standard run of the mill g - code, or is it unique to emc.....
mdynac: I think it's emc-specific.
well it is too cool, the more i play with emc the better it becomes...... and there is a good possibility that the emc wire edm may be back online....
that would be awesome -- keep us posted
tomp: flux is they key to smt stuff
cradek: have a link to that high-speed spindle you showed me the other day?
cradek: is it "wolfgang engineering"? http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=180258202157
I think I've heard that his spindles are good
but it could be someone else :)
SWPadnos: yeah that looks a lot like what cradek showed me
no motors on those though
or, $39 for a motor
yeah, he has an inexpensive motor though
I wonder if the 100W microcontroller lamp dimmer I recently designed works with inductive loads..
you could try it with a fan of suitable size (or a dumb dremel)
theres a pic of my new machine on imagebin.org
oh, it's kinda rainy here...
the big puddle in the middle of our shop floor was back today
i named it lake bragg after the boss
he thought that was hilarious!
which pic might it be dmess?
The 50 latest images added...
Up to 1000 images will be kept in our database for any one time. Additionally we will only retain the image for up to 7 days.
ahh i see your name ;)
a mini lathe?
i saw a bunch of square hot roll steel tubing hiding in the corner of the shop today and i think i need to build a lathe with it
* anonimasu nods
same with me
im thinking steel base and a linear slide for the carridge
with endplates to mount the spindle assembl
does anyone have a idea about what kind of cutting forces you can expect on a lathe?
I read a formula once that seemed to make sense
depends on how big of cuts you want to take and what materials you would be cutting
basically you take the spindle power, and divide by the surface speed
hm. for a small lathe :p
that would be the maximum
sandvik book has data iirc
1kw spindle.. or so..
1kw = 1000 N*m/sec
cutting speeds I only know in feet/min - ~100 for steel
I think that works out to about 0.5m/sec
looking for some data to give the fea program to calculate stiffness of this
so the cutting force would be about 2000N
also depends on tool geometry
* anonimasu nods
2000N sounds like doable with a square tubing.. and capped ends..
2000N or less
even without internal bracing..
roughly 400 lbs
that's the max the spindle can apply
a bit less..
true, so in theory 2000N is a nice conservative number
(I were thinking of getting fancy and adding a bar from end to end so you could tighten it to make the lathe bed stiffer..
but in practice, you want it stiff enough to not vibrate under an interrupted cut, etc
watch for springy ness/stiffness of a bed
preload does not make things stiffer
if you fix the endplates togther it should make it stiffer, or am I completely wrong?
anonimasu: are you building a lathe from scratch?
BigJohnT: I want a really tiny one..
BigJohnT: I've got one big and one really big and one small..
1kw spindle is not a tiny lathe
500mm of bed is a tiny lathe ;)
how many hp is 1 kw?
1.3 or so
fill tube to damp it
small but not tiny
[20:04:48] <jmkasunich> http://www.mini-lathe.com/
sand and silicone!
i'd fill mine with jelly
well, i like strawberry
anonimasu: why not one big box tube with a plate on top for ways
anon... my machine should end up with 150mm z and about 50mm dia. on motors... the crosslide can be moved but i dont have a tailstock yet .. if everyon could keep an eye open i'd appreciate it
but yes, i have some steel tubing in mind to build a small lathe
my idea is tho have a square tubing 100x200mm
as thick walled I can get..
then a linear slide on top
that could get $$$
* BigJohnT gets his calculator out
4 x 8
* archivist would get a lathe bed from a scrap yard
linear slides arent all that expensive..
Thats some big tubing
the problem with tubing like that is it is not square or flat
anonimasu: you'll very quickly spend more than the cost of a mini-lathe
well, I can fit it in the big mill of mine
dont buy a harbor freight mini lathe!
but wall section would be in the 3/8 ( 10 mm) range correct??
oh.. I had a small lathe..
i think that formula's wrong.. as cutting radius approaches zero, cutting force goes to infinity
I think 5-8mm
i went through 4 of them and didn't even get a full revolution of the spindle before they died
fenn: what formula?
you would be better off with a weldment then normalize then machine flat
er, not cutting force, but maximum possible force the spindle can exert
jmkasunich: power/surface speed
yes.. as surface speed goes to 0
normalizing isnt all that easy to do unless you have a nice big oven -_-
its not wrong
but NOTHING cuts at 0 sfm
as the radius approaches zero, you have to increase speed to maintain constant surface speed
a big oven or a big bonfire
increase speed = reduce torque
or a big lobster pot
the reduced torque and the reduced radius cancel out, and the force remains constant
heh.. I can get hot drawn/cold drawn steel..
like, when I buy it..
(im all into theorizing right now..
its imaginary... it dont happen in the real world
how hot do you have to go to normalize?
kinda depends on how hot you welded it but 500 works well
right up until it's about to go liquid... then freeze it
f that is
many aluminums need TIME to normalize.... you cant rush it
chr0n1c: not for normalizing
anonimasu: if you're going to mount slides to steel tubing, probably should read this: http://lmwatts.com/straightedge.html
for that you need a long T8/5 time
jmkasunich: there is a special variable used when talking about these things and it's calle T8/5
fenn: my 200mm facemill should make a long flat surface
yeah, ford used to let engine blocks rust in a field for years... i don't know if they still do
it means the time it takes the metal to cool from 800C to 500C
(probably doesn't quite apply to alluminum)
fenn: speaking of which.. I wonder how les is..
I assume we're talking about steel weldments here...
fenn: or I can throw it at the grinding shop we've got in town
high strength steel pre- h/t 350 F post h/t BUT pre- shot-peen 500 f... post s/p 375 max...
I may be making one in the future
anonimasu: cheater :)
fenn: whatever get's the job done :)
jmkasunich: not necessarely welding..
steel treatment in general
i w as thinking i'd weld up a tube bottom frame, throw it on our 40" thompson water grinder to get it all square and flat then start bloting stuff to it
I put them in an old oven to normalize
it's usually done to relief stress after welds..
chr0n1c: that's a plan
the main reason for building a machine like this is that it would probably be easier then buying a normal little lathe
you heat it up to 800C+ then let it cool down slowly
BigJohnT: kitchen oven?
fitting ballscrews is a pain :)
800C is fsckin hot for home stuff
jmkasunich: baker's oven is probably better
not the one in the kitchen :)
i boiled Al DASH 8 nose gear parts in PEI a few yrs ago to avoid the 18 day trip to oakville for stress relief...
kitchen oven won't get you that high.. but still better than nothing
I only go to 500F
i like the backyard bonfire idea...
for hot roll weldments
if you want the steel to get really hard you can do the reverse: heat it up, then cool it in oil or water
fenn: I'd really like to learn how to scrape ways but learning it takes a shitload of time
throw the frame in there for about 25 minutes
i normaliz all my home made tooling in the ove.. leave it in for a month of cooking.. my angle plate has NEVER moved .0001" since it was ground...
fenn: and lots of hand work.. and well, lots of physical labour dosent work very well for me :/
I scraped the ways on my shoptask before I put it back together after moving it - not loads of fun
the guy that taught me told me if I don't crank up the heat too much when welding the oven will do the trick
bond fire is too uneven itl twist to shit... for sure
we just need a way scraping robot
BigJohnT: that depends what you want
im right here but i AINT cheap..
jmkasunich: how long did that take?
most of the time you don't want to crank up the heat in the weld
two days (a few hours each day)
besides good penetration, you also get a hardened heat-influenced zone
i always welh HOT
that's not too bad
and that one usually cracks
you can use lower temps and longer time works the same
alex_joni: hows that?
mig?? or Tig??
BigJohnT: lots of energy into the workpiece will heat up the area around it
if the part is big heat will be dissipated fast
and the temp will drop fast
so that works like a regular hardening in air (heat it up, allow it to cool fast)
yes.. thats what i do... pour it in and MOVE on... get outta there
anyone got a better idea about
the molecule needs a longer time change forms at lower temps that all
how to make a nice base?
maybe I should just buy a grinder :P
BigJohnT: this mostly refers to the 3-4mm area around the weld pool
talking cross-section here
anonimasu: how precise does it have to be?
(lots of places sell them here)
they usually do nc grinding nowdays so there are lots of manual machines
but sometime the dont make the curve... and actually transform when welding slon and cool
BigJohnT: sorry.. it seems it's called heat affected zone in english
[20:22:55] <alex_joni> http://en.wikipedia.org/wiki/Image:Welded_butt_joint_x-section.png
anonimasu: i'm thinking you could use shims and moglice potting compound to level the rails
instead of grinding the surface flat
hm.. a grinder would be 200eur :p
that looks like it's not in crap condition with a magnetic talbe
(lots of places have closed here during the last years)
BigJohnT: heat treatment usually gets rid of that, so your oven is mostly usefull :D
[20:25:11] <anonimasu> http://www.blocket.se/vi/15787583.htm?ca=1_12_s
alex_joni: I don't see the burned finger in the picture, I always have one :)
BigJohnT: I'll send you a pic of my eyes, when I do some welding next time :/
BigJohnT: did some nice Tandem welding this week (basicly 2 x MIG)
I got burned once looking away from the welder when I worked in a shipyard
sideways is the worst..
we were inside of the aluminum uptakes all shiny
ouch :D .. that's why I got myself a nice pair of goggles
I started to get used in dealing with that
couple of eyedrops, a vile of painkiller .. and a good nights sleep
and if you do wake up in the night.. the most important thing is to keep your eyes closed
no matter what..
just pee carefully
(or the wife gets annoyed)
SWPadnos: just pee around
works like a sonar
oh, happy belated anniversary (I think)
well, that works once you get to the bathroom
yeah, well.. it would suck to use the sonar on the way
scanning for the bedroom door is not recommended
this one time at welding camp...
BigJohnT: what I also do is take a wet handkerchief and put it on my eyes
together with an eyeflap or whatever that thingie you use in airplanes is called
that grinder that costs 250 eur has a bad cylinder(bad seals) from the sounds of it
I keep my hood down and long sleeve shirt on even if it is hot
well.. I learned to do that oo :D
especially when playing with 2 x 3-400A
thats some amps there
yeah, it has the advantage no-one needs to hold that torch
and the robots don't complain .. much
when they do you fix them :)
I got a 4" adjustable wrench
roughly 1.5m long and 20 pounds heavy.. when I turn up with that, no robots dare to complain anymore :D
that's a big one
I have a 48" pipe wrench if you need help
a bit smaller than this one: http://www.robotmonkeylab.com/Monkeyshop/BigWrench/Props%20-%20Bigass%20wrench.jpg
alex_joni: is that you?
if so, you damn very strong
(I know it's not)
anonimasu: it's a prop
I know ;)
the wrench is real though
i've seen open ended wrenches taller than me
well, when you need cranes to move them they are big.
the biggest tool I've seen was a 5m caliper :D
that's a big caliper...
wasn't very heavy..
2 people could nicely work with it
the biggest caliper i've used is 48"
hm, im buying that grinder.
if it's still there tomorrow
alex_joni: so normalizing is for the heat affected zone only?
BigJohnT: no.. for releaving metals of stress
that internal stress might be from casting, welding, bending, whatnot
for welding from the weld pulling as it cools?
that and the heat affected zone
actually the whole part starts to bend after welding
depending on the sheet thickness and weld geometry it's more or less visible
I welded some 1/2" sheets the other day
yea, I've seen that a bunch
and used lots of current/small speed
it curved about 1" up :D
try that on stainless and it will curve up 4"
the plates were roughly 2-3 feet long
I haven't seen much 1/2" stainless :D
I do some stainless structural from time to time and it really moves
yeah.. I did welds on .5 - 1mm sheets of stainless
there it's extreme..
a T joint will warp the heck out of the tube
I also welded some 100mm sheets :D
yeah, hard to call it a sheet :D
but if it comes in a 2x12m size
that reminds me ..
I saw a nice application lately
they use metal sheets to make some beams
but the sheets come in rolls, not flat
and it's 1" thick
how many amps do you need to weld 100mm?
fenn: not more than for 20mm
uh.. how much is that
you just need more passes and a wider weld area
3-400A for manual welding
you can probably weld it with 200A and a lot of time and passes too..
BigJohnT: so.. long rolls of 1" thick metal
goes through some machine to straighten it out
I know how that works :)
then through another that cuts it (mechanically) into stripes (I think 10 pcs at once)
that's some big iron doing that
why dont they just do that at the mill?
BigJohnT: yeah, I've seen it done a lot of times.. but not in 1" thick.. that was really sick :)
fenn: steel mill?
that's couple k km away
how long are the beams?
and you seldom have the same widths you need
fenn: sometimes up to 18-19m
but usually up to 8-9 m
in Tiawan I went to a window blind company that made the slats from roll stock. They had a 20' deep pit after the slitter because the different thickness across the roll would wind up at different speeds
i can delete stepper.var and emc will re-make it on startup?
or uh... can i?
i think my g5X's went screwy somewhere
but you can copy another one from /etc/emc2/sample-configs/stepper
oh, so what happens if i set everything to 0.00000 like i just did?
it seems to corrected my strange behavior
hmm... is there a something somewhere that would tell me what all the var's in that file are?
one changed back to a 1.00000 after i loaded emc2 from 0.00000
chr0n1c: yup.. it's called a manual :)
the one with 1.000 describes the currently used coordinate system
ahh, i usually don't read manuals until i get stuck...
i have read quite a bit of the emc manuls at this point ;)
but not that part
that's the best part
I always sleep so well when I try to read it :D
chr0n1c: take a look at G92 in the manual
ahh one line down ;)
good night all
good night alex
g92 is a deadly SIN...
here goes the ink and paper verify of the midibox_core module...
i'll let 'all know how it turns out in a couple hours.. lol
so, the ink cartridge from inside the ink pen fits in my spindle collet... shoulda tried that a long time ago