From $.50 -> $2.00 each
hmm... seems my cat can cut and paste. Crafty little critter.
so thats what that was...
talented aren't they
pesky at times as well.
dancing on keyboards, or newspapers, or whatever
just broke an endmill. Dang.
yeah, they have no respect.
And they hate the 4th of July. Whoa.
alex_joni: did you see my X axis conversion picture?
just got back home ;)
cradek: hi ;)
are you home for a while now?
[12:10:50] <cradek> http://timeguy.com/cradek-files/emc/DSCN6290.JPG
well.. depends on your definition for a while :D
have to get ready for work, bbl
I'll be around till next thursday, then I'll be in the mountain house for 3 days
and then I'm in sicily for 2 weeks
cradek: looks GREAT
what name should the library which contains hal_lib.c and XXX_ulapi.c have? someone's already taken libhal. libemchal?
what does it do?
it's the library of stuff necessary to make a userspace component
so that someone with emc2-dev installed can do so
hm .. maybe, since I'll be doing the same for realtime soon
no, I don't think I need to do anything for kernel, because there's hal_lib.ko
libemchal.so is fine too
care for some pictures?
I like pictures
[12:40:35] <alex_joni> http://dsplabs.utt.ro/~juve/blog/index.cgi/photography/01152273021
same here :D
ooh, time for free bagels for breakfast .. photos take second place to that, I'm afraid.
I can surely understand :D
wow those are pretty
Skunkworks: thanks :)
darn cant type
How you doing?
very fine .. thx ;)
been kinda away recently
and I'm going on vacation at the end of the next week
I have not kept up since fest.
To the mountains?
for a few days, then to sicily
for 2 weeks
Nice. Lucky Dog!
ray: was the trip uneventful?
For the most part it was.
I was awed by the interest in CNC.
* Skunkworks hasn't been around much
Nearly every little machine tool factory had prototypes or production machines.
Most had demo machines at the show in Beijing.
What you been doing Skunkworks?
mainly working on the house and my car
I need to get up to speed on where EMC2 is now.
I kinda just fell apart after fest.
cradek: that mounting looks great
I'm very pleased with it
how much space is there between the motor wheel and the drive wheel?
the foto is not very conclusive
it has a ball bearing and thrust bearings on both sides
I meant motor housing
10mm, the angle is deceptive
ok, thought it's closer
err maybe 6mm
heh.. I just noticed the keyboard in the back
if it's more than 1mm it's plenty
jepler commented on the cat hair stuck to the wrist rest
I 'overlooked' that
seen the flowers I shot?
rayh: this is kind of a big image but it's the X axis of my sherline servo conversion: http://timeguy.com/cradek-files/emc/DSCN6290.JPG
[13:27:51] <alex_joni> http://dsplabs.utt.ro/~juve/blog/index.cgi/photography/01152273021
* alex_joni heads home early today..
catch you later guys
rayh: were you around for the talk about the lathe tool shape compensation?
I did all of that after fest, maybe when you were traveling
cradek: I had asked yesterday - did you machine the mounts (outside and holes/slots) totally on you max?
both pieces were done on max
have you machined "bigger" stuff like that before?
sure, I've even cut steel on it
it just takes patience and care
I wrote gcode subroutines to cut holes and slots in multiple passes
a hole cut is just some helixes, and the slot cut is like that but arc/line/arc/line
does the slot routine do angled slots, or only axis aligned?
you specify the endpoints and width
[13:47:51] <cradek> http://cvs.linuxcnc.org/cvs/emc2/nc_files/useful-subroutines.ngc?rev=1.3
saw the code, but didn't really study it ;)
just load this file in axis to see the samples
one day, when I get back my emc development computer, I'll do that
I'll make a screen shot in a minute
I saw the motor mount though
wait - your not supposed to use o-words for "usefull" programs
whats the fun in that
now if only we had a #include in g-code
I was supprised to find mod
[13:51:39] <cradek> http://timeguy.com/cradek-files/emc/useful.png
ye - way cool
and they use radius compensation which is nice
show off. ;)
that is cool
a secret to cutting metal on a small machine is to never just plunge
that helps with non-centercutting endmills as well ;)
these all spiral down from the safety height
where you using an 1/8 inch mill?
that must have worked great
the slots are .190 (#6 screws)
the 2-flute mill I use for Al is about .121, I got it somewhere used and resharpened
I should get some more before I break it
hss or carbide?
probably just hss, not sure
should be able to tell by weight - carbide just feels heavy
hm, I never noticed that
comparable to weight to your pcb mills/drills
hmmm - does carbide stick to a magenet?
I think so
well that wouldn't be a test. Guess it doesn't matter
(we had grabbed a cutter we thought was carbide - tried to take a cut assuming it was. the cutter turned red and got shorter and shorter 1 inch diam.
quite the show :) sort of did some friction welding :)
PCB mills and drills should all be carbide
SWPadnos: yes they are
Skunkworks: cutting what?
it was a large steel backer plate for a sheer.
the type of sheer you would see in a junk yard.
for cutting up scrap
(this was on our big mill)
cradek: can you reverse the direction of the cut?
tecknically you are climb milling
yeah I noticed that but didn't change it
I should look at the cutouts and see which side has the better finish
just the largest hole left a piece
the main problem with climb milling is that the cutter could catch and pull the slide (if there is any play in the mill) which could break the cutter.
I'll try to reverse them... do you think you'll use them?
won't they ne climb on all sides??
SWPadnos: not on the cutout piece from the center
so the spiral sub goes one direction, and the slot the other?
no I think they both go ccw
I think I would use the program.. I think it would come in handy.
jmkasunich: by the way, it almost works to create multiple user components from one process, but it segfaults when one of them calls hal_exit(id)
jmkasunich: maybe with something as simple as a reference count of "components in this process" it could be made to work
cradek, why would the finish be different then? (if the round and slotted holes are all cut in the same direction)
ok back up a second
if you cut a hole ccw, there's a piece left in the center
ah - thanks
if you cut a (big) hole in either direction, there's a piece in the center :)
I just meant I should compare that piece to the other, and see which finish looks better
I always understood climb milling had a better surface
right - got it now - it's another piece (duh)
finish - but never really noticed
Skunkworks: these parts are a bit gouged up, I'm curious to see if the center piece looks the same
and with our big mill - there is not backlash - we ran both directions
"conventional" milling can also tend to bind the tool, I understand - maybe because the tool is trying to sink itself further into the work
conventional milling you are preloading the cutter into the lead screw
I'm thinking of the forces perpendicular to the direction of travel (though I haven't had a lot of coffee, so I may be thinking poorly)
I may be thinking poorly for other reasons as well ;)
I think it's traditional to take a very light climb pass at the end of the work
I think heavy climb milling is usually a bad idea like Skunkworks said - it pulls on the work
I have only seen a mill break once because of in on an old sloppy mill
(taking a heavy edge cut)
wow... well, the ulp for creating gcode from eagle files is sweet. Alas, I think I'm using the wrong tools with it :)
how goes it?
Alright I suppose. Shot an endmill into the wall yesterday. That was fun :)
How goes with you?
pretty good.. kinda busy unfortunately
I guess that's better than the alternative; bored or dead...
alex_joni: for a minute
wanted to ask you about hal create pin
I think I remember seeing some work from you regarding that?
umm .. I made a function that combines the snprintf() to create the pin name with the actual pin creation
jepler: is that in hal.h ?
extern int hal_pin_bit_newf(hal_dir_t dir,
hal_bit_t ** data_ptr_addr, int comp_id, char *fmt, ...);
ok, another thing :D
would it be hard to add a show world/joint to the Axis menu?
for the backplot & readouts?
* alex_joni imagines it's just the emcstatus variable that changes
how would you backplot the joint position?
only for diagnostics reasons :D
as if it were XYZ...
for the backplot, the coordinates are taken from motion.traj.position at around line 1391 of emcmodule.cc
seems to me like the backplot is not useful for joint mode
maybe it should just be disabled
cradek: hello, indeed it's hard to picture that
* alex_joni has an idea though :D
what are the 'stat' buffer fields called for the joint positions? they may not be available to pyhton yet.
for AXIS 2.5...
cradek: ideally you'd have bargraphs for each joint (minpos .. current pos .. maxpos)
for joint mode I mean
jepler: I think it's this: emcStatus->motion.axis[axis].input
cradek: at least for the present ..
the future might involve knowledge about kinematics, and display a nice robot in 3D :D
ok, enough of me nagging.. go home, enjoy the weekend :D
G43/G49 in axis would be cool too...
yeah we're going to add that soon, since I'll need it for my lathe
The position shown isn't the commanded position, but... I dunno... the top of the tool with G43 on I guess.
it just ignores tool offsets (as if they're all zero)
I made modification to the ULP to use G43 tool height compensation, since I can't align my tools all the same like you do :)
The G0 in the prologue got me though ! :)
I don't understand
G0 honors G43 right?
alex_joni: is .input like actual and .output like commanded position?
Yup, if it's on. :) The prologue torns it off with G49, then seeks to the 'safety height'. The safety heigh without compensating for the tool, though, was under my bed :)
jepler: may I admit I have no idea? :D
that'll teach you to run my gcode wihtout checking it
alex_joni: it looks like those fields are already available, but not in a convenient list representation
(hope it didn't hurt anything)
yeah... mistakes happen :)
Nah.. at that point it was my code, not yours :) I modified it to do G43. You were happily not using bit offsets. I broke my bit and put a nice gouge in my clad, but that was all. And bits and clad are cheap compared to other things :)
I'm getting pretty rough cutting though. I dunno if my bit is just poor, or if I'm going too fast, but it basically is tearing the copper traces off the board at this point.
are these the thinktink tools?
Nope, not yet. The thinktinks are on their way... This is a HSS dremel bit. I can only get it spinning about 5K rpm, so I've dumped my speeds down to about 1/5 of what you were doing....
you may need to rig a different kind of spindle
jepler: it seems it's busted for usage:
stat[axis].output = 0.0;
stat[axis].input = joint->pos_fb;
the chips (dust actually) fly out of the cutter now on mine and I get a nice clean cut
so only joint feedback is available, output shouldn't do anything
even when I was at 8-10krpm it was not like that
I tried a constant stream of air thinking chip removeable might be a problem, though it didn't seem to help.
I bet the right tools will be better.
Probably so. Should be in early next week...
I may try another go or two at slower speeds and see if that helps. Maybe I'm trying to take too much off too quickly for the bit.
at 5krpm you should be going no faster than 5ipm
Yeah, I'm doing around 3-4 ...
seems like that should be ok
alex_joni: so in effect I can only get the actual joint position, not the commanded joint position.
seemed reasonable to me :)
alex_joni: I'll commit something now to give python easy access to the joint positions
alex_joni: but it's only one step on the way
jepler: no need to rush it for me..
but thanks anyways
I'm trying to figure out what goetler is doing..
talked to him a bit, he's the guy doing the 45 degrees tilted workspace
he's demonstrating how badly kinematics really is :)
that's what he's doing
that reminds me, I should get back to working on ckins
fenn: I think his math is borked
fenn: my kins worked great with emc2
well, he didn't really tilt the axes, but rotated them
if I had one of those tripod machines I'm sure I'd get around to proper nontrivial kinematics support in axis
want me to mail it to you? :)
jepler: what are you working on?
fenn: axis can't jog in joint space
and it won't display joint positions in the DRO
I think those are the main two problems
alex may know better
jepler: I don't have any stepper drivers and interest to keep working on it
ah so like if the tool strays off course you'll see that on the backplot?
jepler: bet you can build one easily
how do you jog joints at the nml level?
normal jog commands
so you just switch modes somehow?
and in teleop you set carthesian vectors
alex_joni: how hard would it be to ship? I can come up with the drivers, unless they're unreasonably big motors
cradek: EMC_TRAJ_SET_TELEOP_ENABLE emc_set_teleop_enable_msg;
jepler: this really is a TOY
3 HDD steppers
I ran them at 5V
alex_joni: that sounds nearly ideal .. any fool with 12 transistors and perfboard could put together a driver
the only problem is that it's a bit big
Hmm... maybe my 'step up' from the bridgeport for CNC should be a tripod....
bulky is bad when you need to ship overseas
to ship via airmail
yah that 1 square meter steel plate seems kinda bulky :)
jepler: my cousin is coming over this summer, maybe I'll convince him to take it with him
it's hardly 1 square meter
* fenn squeaks
more like 40cm x 40 cm
jepler: but I bet you can build one in 2-3 hours
[21:07:25] <alex_joni> http://dsplabs.cs.upt.ro/~juve/emc/mytoy/Resize%20of%20HPIM4955.JPG
you can see all details here
hmm... does it take a tool?
it's just a toy
zwisk: I used a pencil
you could probably jam a pencil in it
a very small one
ahh. You could put a pen in the .. right..
that's a washer :D
sorry.. I'm clumsy around my keyboard today :(
ooh... I like the work on the guides from which the string comes out...
those are MIG current tips
but.. why mess around with a tripod when it's practically just as easy to make a hexpod?
fenn: always easier to build 3 axes than 6
but you've already got it all set up
and you need to get it perfect so it'll work
no you dont
I doubt I can do that for 6 axes
Ooh... ST h-bridge chip.... :)
alex_joni: there's no redundant constraints in a hexapod.. any way you slap it together it will work
hmmm.. I see
fenn: maybe one day I'll build a second tripod :D
and mate them
jepler: if I put the parts in a box?
without the metal sheet?
alex_joni: is it fishing line?
I wonder how useful a hexpod using 'string', like alex_joni's toy, would be...
alex_joni: maybe .. I would have to find 3 matching stepper motors otherwise
jepler: I can do that ;)
but you'll have to mount them on a table
did you add a cylinder to the motor shaft or is that the bare shaft?
and measure the distances :)
those were steppers that moved the HDD head
you can see there's a small pointy thing coming out of the shaft
what are the guides that the fishing line goes through?
do you really mean HDD?
yes, the older ones
no, hard drive :D
jeff should go to surplus center instead I bet
* alex_joni hmmms
what are the guides that the fishing line goes through?
cradek: welding current tips
about 1-2$ / piece
I don't know what that is
zwisk: i dont think you can really get a sense of how rigid it is without actually poking at it
fenn: Seems to take a litlte bit higher garage than your typical gantry though.
cradek: on MIG welding you have a welding wire
which passes through the welding gun
maybe when I send fred's lathe back someone else can try to talk him out of his portable hexapod
in order to get current on it, so it melts (from the arc produced), there is a small piece called the current tip which does the energy transfer
ok I see
[21:24:44] <alex_joni> http://en.wikipedia.org/wiki/Image:MIG_cut-away.jpg
surplus center does have some 12V stepper motors for $4
alex_joni: so I measure Bx, Cx, and Cy and edit tripodkins.c ?
eventually (when we'll care enough) those should come from the ini
eventually those should be determined by a ball-bar test
the comments indicate the controlled position is in XYZ .. does it "move" in Z by creating slack on the wires?
jepler: I used a metal rod placed between the tips, to have the same distance between the three tips
but you need some weight for that to work properly
as I've seen it there is about 1-2mm tolerance on Z where nothing seems to be happening ;)
so the shape the wires take on is less like a catenary and more like a line?
(is that the word I mean?)
what's a catenary?
so the shape is less like the fishing line falling off the motor shafts
[21:29:39] <jepler> http://mathworld.wolfram.com/Catenary.html
a catenary is the shape a slack wire takes :)
well.. actually it looks like crap when you slack it
because the fishing rod has been tightly run around the motor shafts
which causes it to bend
I'm imagining hanging this thing off the ceiling of a dark room, with an LED and a weight at the controlled point
and when you slack it (at least without significant weight) it twists & curls
a ghostly light moving around in 3 dimensions
jepler: that might be cool :D
and time-lapse photoing it?
you can make 3D pictures :D
by varying the LED light
30 minutes exposure in a pitch black room :D
jepler: sounds like a great idea
the only problem will be homing :D
well I'm trying to talk cradek into building one of these when he's done with his lathe
jepler: this one is soo easy you can build it yourself
i call it.. the tilliepod
weell, good afternoon
* alex_joni heads to bed