cutting 0.5, .005 at a time takes a while
but isn't cnc great
cradek: what are you cutting that needs such light cuts?
an extremely hard QC30 shell mill arbor
it's slowly becoming a renishaw probe mount on my awkward sideways lathe
grinding or cutting?
cutting with a carbide lathe tool
(I'm surprised it works)
using the bport, right?
I'm surprised you can't take a heavier cut
I haven't experimented...
the stringers are already black
stringers = chips?
* JymmmEMC hands cradek a bottle of coolant
what is your surface speed?
Jymm: that just breaks carbide
I started at 200 but was getting sparks
* JymmmEMC likes sparks
that seems odd
tool is still cutting, but I doubt it will have any life left in it
it's one of those brazed ones. I have an indexable I should have used
what is your feed per rev?
0.002 x 0.005 cut?
that seems _really_ light, even for hard stuff
it's just a 1/4 tool
I don't have any experience doing this - just want to get it done safely
how many passes done so far, and how many left?
it'll be about 100 altogether
very nearly done
the downside of many light cuts is tool wear - 50 passes on the first 0.010 of the tool spreads the wear out over twice as much tool as 100 passes on the first 0.005 of the tool
okay, so im in the math lab for a moment today, and im looking at some linear algebra, is thats what used to do most of the kinematics for emc2?
jmkasunich: I'll try more next time... but I don't care about the tool one bit
as long as it lasts thru one part - which it sounds like it did
gezr: kins for many machines can be as simple as a little trig.
cradek: ah cool, so your cutting some hardened shaft?
turning it down
if its an arbor, chances are its surface hardned, maybe you have seen a change in the material type?
err, not material type, but chips, the way they look
how does the probe attach? round shank in a hole, or something more complex
very complex, it
's adjustable for centering
so you have to do "very complex" in hard stuff?
the shell mill arbor has a couple flats in the right places
it's an excellent start
wouldn't it be simpler to put the complex stuff in a soft part
hm, tried to debur with a file - it doesn't cut
got a whetstone? or one of those little diamond files?
yay, one of the two hard cuts is done, and fits
hm, any good way to hold a round boring bar in a square vise?
drill a hole in a block and slit it
clamp 2 blocks together with paper between and drill down the line
ER40 collet blocks would be nice...
heh, I found a piece of scrap with a nice 1/2 hole in it
bizarre, the tool has a chip missing from the very tip, but it is also worn? flat on the end
not really bizarre
was it originally positive rake? (before the chip)
positive rake edges aren't very strong - the hard materal probably chipped it off early, and the result was a negative rake tool with a stronger edge that continued to cut but wore slowly
there are three or four different modes of fracture failure, some of which will shatter any edge no matter how negative, too
that sandvik book lists em
this was a 0.005 DOC x 0.002 feed per rev "lathe" operation, on an arbor in the bport spindle
(cutting the arbor, tool in the vise or some such mounting)
150 sfm, f-in hard steel (a shell mill arbor being converted to something else)
many many passes (like 100)
was it a fracture or was the tip just abraded
now for the hard part...
what's the hard part?
how do I see the last 'N' number of lines of conversation?
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-06-06.txt
a ER-40 collet block is pretty easy to make if you access to a CNC lathe
ds2: I wish I did :-)
made one out on the Haas at the JC
mine is too small for that kind of thing
well, make ER-16s then :)
an ER40 collet block at this point is at least one level deeper into the "making a tool to make a tool to make a tool" than I think cradek wants to go
but it would be nice to have one sometime.
I wonder how useless is a ER-32 collet block in aluminum
a square or hex collet block?
have an extra aluminum blank sitting around (threaded + bored out matching taper for ER-32)
parabolic collet block!
it is still connected to the 2" bar stock I started with
hex can serve as square I guess
no, hex can't index 90deg
but not the other way around!
ds2: with a height change it can
one cheater's way to make a collet block would be to make a block with a 0.998 hole in it, and shrink it onto the shank of something like this: http://www.maritool.com/Tool-Holders-Straight-Shank/c23_28/p43/ER40-1.0-4.0-STRAIGHT-SHANK-COLLET-CHUCK/product_info.html
if you are play games with gage blocks...
if anyone is interested in a ER-32 blank in aluminum...
jmkasunich: I could do that with one of the QC30 jobbies you got for me at HGR!
that would be cool (but huge)
heh... much rather get a block with a female MT3 taper in it
how about an index head with the spiral hole patterns?
I'd rather have a 5C collet block to be honest
sure, if I had a set of 5C collets
it will take a larger range of material, and you can pass the work all the way thru
The index head has a 3 jaw chuck on it
5C requires a lot of collets to cover 1 1/16 to near 0
jmkasunich: aren't they also just 1"?
oh 1 1/16? close enough
5c is far, far superior to er collets for production, but are a little dicy for one off stuff
but the gripping range is like +/- 0.002
ds2: true, I forgot about the ER ability to grab any size
ds2 they're a little broader than that
(this discussion started because he wants to hold a boring bar in the vise - I assume the bar isn't some weird size)
OTH, that funky ejector nut on ER's is a #@$%$!%$@#$%@
and that's already solved :-)
toastydeath: not by much
we have a "full set" of 5C and also 16C, holds any size from some tiny thing to the full spindle cap
there's probably 64 collets in the bunch
but i would readily admit most folks don
I can see only 1 reason to go 5C/3C... hex/square collets!
't want 64 collets of anything lying around
C collets and S collets both can auto-open
ER and TR series can't
at least not with any device i've seen, perhaps someone somewhere figured a way to do it.
I think a turret or a tool changer would be higher priority then an automatic closer :)
not in my experience, i've done a bunch of manual work with a 5c for small runs
a qctp is all you need most of the time
some deburring stuff, etc, where you don't even stop the spindle to change parts out
but a hobby dude isn't going to be doing that very often, so i can definately see the advantage to only needing 16 collets to hold every concevable round shape
fwiw, 1/16 to 1-1/8 5C by 16ths is $65 at enco, and a matching hex and square blocks are $40 (for both)
Oh you mean those quick release collet things that royal makes?
all i know is i can open 5c without stopping the spindle
I been burned by cheap 5C's... ones where the key slots won't fit the pin but the same size one from hardindge did
ds2 yeah, the enco aren't great - but for a hobby guy, they work
(that is what I have)
jmkasunich: is that just for the indexer or also on the lathe?
I have a collet block, and a collet chuck for the lathe (one of those sjogren speed chucks)
the sjogren was an HGR surplus deal
how much did the chuck cost?
[01:55:56] <jmkasunich> http://www.workholding.com/atscatalog07.html
I think it was about $100 - I've had it for 5 years or so now
what kind of lathe is it on?
my piece-o-crap shoptask 3-in-1
those things have a standard nose ? that link seems to suggest they are available only for standard nose/mounts
its just like any plain-back chuck - a register diameter and some bolt holes
it came with a backplate for some spindle nose that I don't recall
one of those ones that uses pins to let you quick change chucks
I ebayed the backplate and made a new one to fit my machine
the D mounts with the CAM?
yeah, thats the oen
vismach is fun http://imagebin.org/19577
using trivkins Z(guide) Wspindle XY to orbit B tilt A swing C rotate part UV translate part
jmk: I put the cinci and viper on my website this morning.
ds2: I take that back - it was a taper mount
if you scroll about 2/3 of the way down that page I posted, there are mounting drawings - I had the one on the left
the file is emc2-stuart.tgz
it is the tar file of the entire run in place directory
may be a little much but you will get everything
cradek: you got a Renishaw hal-ified for probing?
tomp2: the machine I will use it on isn't even running emc yet
getting it to fit the spindle is a big project, so I figured if I tackled that I'd be more likely to do the emc convert
maybe you need a grinder for the spindle, and maybe you can use one for a bit at fest
(for the taper , i imagine)
that's why I started with a shell mill arbor
that already fits my oddball spindle
Does anyone have a schematic for a dumb amplifier? I want to remove the SCR amps on the cinci. The SCR amps are rated at 33 amps. I cannot find anything to give me some direction. All the amps I can find do not accept pwm they all want a +/- voltage. The amps all have tuning capability. I don't want an amp with tuning capability. I want a dumb amp I can control with EMC and not have to superimpose the EMC tuning on top of the amps tuning.
stustev: skunkworks/sam was working on a very dumb one. not sure if it's big enough for you though
I spent the time to read most of the posts on cnczone. When he made the comment he thought he would just purchase a commercial amp I quit reading it.
wtf is a dumb amplifier?
* renesis_ can do amplifiers
goes along with a dumb shop owner
what I am calling a dumb amp is an amplifier that is JUST an amplifier. It doesn't have tuning/ramp capability. It just accepts pwm/pdm and amplifies the input signal
the 2+ kw amp from mesa looked 'dumb' ( no tuning )
are jon elson's amps too small? I think they are like you describe
yes jon's are exactly what I want. - too small
I asked him if I could parallel them. Not so easy to do.
wtf is an 'amp'
throw fatter transistors on it
7I29 from mesa "The 7I29 is rated at 22.5A continuous per channel at 165V maximum input voltage." 2 amps in 1 pkg
I am thinking I need something along the line of 3500 watts.
or just tie its output to some darlington transistors
some people have talked about Granite Devices, but I don't know anything about them personally
hmmm. too small it seems
oh - wait a sec, those are for Mach compatibility ;) (step/dir)
the fluxeon from granite is 1kW ac servo
stustev, what voltage?
the SCR amplifiers show 33 amp continuous - I think at 90-100 volts
ok. I have some Advanced Motion Controls amps, which appear to be 30A/80V
neat @ 2KW
what kind of load is it?
my thoughts here were along the line of bigger transistors. everyone I have talked to says this large of power requires different treatment than a 300 watt amp
the load is a dc servo
no like ohms, henries
how about autotuning Servo from Automation Direct 3kW http://web3.automationdirect.com/adc/Shopping/Catalog/Motion_Control/Servo_Systems/Medium_Inertia_(1KW_-_3KW)_Servo_Systems/3KW_Servo_System_(Med_Inertia)/SVA-2300
with motor near 2500$
I do not know. I can find out tomorrow. They are gettys dc servo motors
the problem with dumb amps as you get to bigger sizes is that they are too dumb
yes, there are different stages of power control devices (transistors, fets, SCRs, thyristors), and each has its useful range of voltages/currents
im just curious, really
jmkasunich, can explain further :)
if your PWM hiccups even a tiny bit, they go bang
its not filtered?
a true dumb amp like the skunkworks one is just PWM->power transistors->motor
with no protection
are the amps usually linear or switch mode?
that's exactly what I am told here
(I did the board layout for skunkwork's amp)
I am behind your current conversation - I was talking about the bang
renesis_: linear at kW power levels is just insane
but totally doable
efficiency for a linear rarely if ever exceeds 50%, and is usually well below that
i mean thats like pro audio amp ranges
which means you need to burn kW to make kW
yeah, 60% if ur good at AB
and actually they almost always do more than 50% now, good amp designs, especially chipamps
but yeah, never much more
haha theoretical max is like 87% or something
ask rayh, those AD drives and amps are good bang for buck, i use em too
and i use 'em direct coupled
what is the BCA amp from crown audio? is that architecture usable on a servo?
you mean the inputs?
audio amps are AC
yah, crown amp on emc, cool ( i want the brass box and glowing tubes :)
if you try and put filtered pwm DC on it, itll just die at any signal coupling cap
and pwm on AB is kinda a waste of power
might as well go full class-D
class-d is up to 100%, realistically above 80% for high power
nothing is 100%
* jmkasunich <- pedantic
and pwm on most AC amps will be voltage offset fucked by the last stage
heheh, common smps will do 95% now
we get 98% or so on VFDs
we lab tested some junk when i was in school, was hardly below 90% with any loads
500kW in, 490KW out, 10kW to dissipate - takes a darned big blower
like, you basically want a variable smps style psu
thats what a PWM amp is!
kinda, its bridged output usually
tho i guess you maybe need that for the servos
right - because a servo that can only go one direction is rarely usefull
cheap, old school... use hydraulics. a pegasus/moog valve can control lotsa pressure getting you speed and torque from a 24V 20ma signal.
but you cant use an off the shelf audio amp, almopst all of them are AC signal coupled thru a cap
itll kill your absolute DC values and float it around 0V
hello, fellow cnc'ers!
* renesis_ does alot of audio junk
plus most audio amps aren't rated for continuous power
yeh that also
not looking to use an off the shelf audio amp I just asked if that type of architecture was useful
alot cant even do a sine at rated power
because normal audio is way less power
they skimp on transformers because of that mentality =(
the right answer for a big-ass dumb servo amp is an H-bridge with some protection circuits
at kW power levels, you'd skip the transformer - just rectify the line voltage
DC coupled class-d
speaking of audio... my sound isn't working on a fresh rtai kernel install with emc2... :(
there are more than 50 5 axis machines in the Wichita area. All of them have OOLLLDDDD controls. Most of them are mechanically kept up nice.
i gotta boot into the stock 8.04 kernel to use audio
heh, get a soundblaster
audio and realtime don't always play well together
if you want tunes while you run your machine tool, get an ipod
the realtime is hella robust
ive browsed web for docs during cnc cycles
emc is like whatever
chr0n1c: i had same problem, alex joni told me to install a package... fixed my prob, lemme see
but yeah itll prob chop the shit outta audio
i dunno i set this box up for my audio workstation.. then when i saw emc2 went 64 bit i figured i'd put emc2 on as well
i can still do audio if i boot into the stock kern
stustev, Advanced Motion Controls has more or less what you want
but! i jsut got the machine ini's tweaked and it's working just as good as it was on the old control box, which is, awesome!
they don't seem to have any PWM-input dumb drives over 20A, but they do have analog input drives up to 60A
chr0n1c: exact same symptom, still havent found the pkg...
only thing is i had to edit the ini's by hand.. i don't think the stepconf will work with my stepper driver
i love emce2 ... ;)
chr0n1c, what did you have to change byb hand?
i jsut got an email from a guy who wants me to engrave a zippo for him
it should be coming in a few days
make sure it's empty ;)
the zippo is empty? lol
SWPadnos: i changed the Yen signal
i couldn't find it in stepconf... and i already knew how to do it in a text editor....
was just quicker
is the Xen, Yen signals... amplifier enable?
it fits and works
chr0n1c, yes, they are
I thikn (without looking)
i don't know what you just did, but i'm excited for you cradek
he made a mount for a renishaw probe
ok - there's only one "Amplifier enable" output. did you need separate ones for each axis?
i found i only need one...
so i could have done it with stepconf then...
i just didn't wanna read ;)
jmkasunich: what app do you use for layout
I used eagle for the skunkworks board (thats the only board I've layed out at home)
i use eagle, too much time invested not to wtf
i figured out the jpg/bmp/gif/png/whatever graphic formats to .DXF with inkscape!!!
fucking 2 year learning curve on that shit
at work I've used Orcad (back in the day), Tango (a bit later), Pads Power PCB (recent) and Mentor Expedition (now)
now i just need some open source cam software that i can use and i am 100% linux
cant stand orcad
I was using orcad in the 80's and early 90's
ive heard ok things about pads, dunno mentor other than people mentioning
yeah thats what the gui looks like
jpg -> inkscape -> qcad -> ???(what cam software?)
mentor is big bucks
total legacy app, using that shit for a board hurt me
yeh, thats all i know about it
chr0n1c: beg twingy for dxf import
I liked pads - just getting started with expedition, haven't had time to form an opinion
i really like qcad since i been using it lately... sort kinda like autocad
in GCAM, he can do it hes just been doing other shit
[02:51:40] <cradek> http://timeguy.com/cradek-files/emc/renishaw.jpg
it should do pcboards eh?
ask me in a month - I have a board to do, in schematic capture at the moment, layout starting in a week or so
hehe, tried diptrace?
cradek: you did the bore and the facing?
i tried... OpenCam from sourceforge.. it crashed my computer.. had to do a re-install of the whole os
now I just have to make a drawbar
its like eda for kids, but its very simple to pick up, and pretty capable
ok, thats not as complex as I thought
1/2-13 to M8-1.25
i have to use it because someone i do jobs for does his schemas in it
is the arbor already drilled thru?
and there's a thread in the top for the drawbar my machine doesn't use
why threads on both ends? can't you just put a head at the top?
nothing except threads up there
or is the "head" gonna be a pair of locked nuts
the thread is already in it, that's what it will tighten against
the drawbar is male-male
* jmkasunich confused
there are female threads in the top of the arbor, right?
for a retention knob or drawbar
you use neither
is the thru hole bigger than the OD of those threads?
chr0n1c: check for linux-ubuntu-modules-2.6.24-16-generic linux-ubuntu-modules-2.6.24-16-rtai linux-ubuntu-modules-2.6.24-17-generic linux-ubuntu-modules-hardy-rtai one of those is what fixed sound on the rtai side for me
does anyone knoe if this has been ported to Linux? http://www.dakeng.com/ace.html
no it's about the ID
then you can't insert something from the bottom and have it engage those threads
no but I can put it in from the top.
* cradek is no rocket scientist
so it will have a fat spot threaded to match those threads (that's the 1/2-13?), then neck down, run down to the renishaw, and have the smaller thread there
how do you tighten it?
the pitches are different. it will tighten no matter what turns
I will make it narrow in the center so it's a little flexible
are you gonna turn the probe with respect to the holder? (eww)
it has to allow the alignment screws to work
omg what are you making
cradek: pretty, very pro equipment, congrats
omg ace converter, die
dang, shame the bottom isn't m6
if someone would smah qcad and inkscape together and somehow create the untimate cam package.. i would be so so happy
you could take a 1/2-13 setscrew, drill and tap the bottom for m6, and use m6 threaded rod for the long part
M8-1.25 is extremely close to 5/16-20. I was tricked for a while.
jmkasunich: that would be cool
* chr0n1c vows to type slower to prevent typos
tightening will be weird
you turn it CCW to tighten
(since that will back the 1/2-13 out faster than it unscews the M8)
I can easily put a slot in the top for a screwdriver
is the top of the toolholder flat?
for example, if you put a retention knob in, what does it seat on?
not really. there's hardly anything but threads
I don't think it is made to hold any knob
I think NMTB is drawbar only
what about this - 1/2-13 "plug", screwed in, and drilled say 1/4"
then M8 threaded rod, turned down to 1/4" for much of its length
1/2-13 is 12.2mm thread base. no reason not to use M8 all-thread
SWPadnos: It looks as if they have something that will do exactly what I want. thanks
top end threaded 1/4-20 or M6
with a nut
stustev, cool - hope it works out
[03:04:14] <chr0n1c> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?InkscapeHowto
<- just found that... interesting
cradek: how can that be
1/2 is only 12.7mm
now if only I knew the difference between these 30A8T and 30A8V drives ...
thread depth must be more than 0.25mm on each side
oh wrong number, dang
I must admit I was also confused, I was thinking about the 5/16" that is a 3/8-16 tap drill
.407 minor diameter - 10.3mm
still would be fine if it's well centered
you're gonna want to neck it down for flex, right?
cradek: you mentioned alignment screws - what are they for?
neck it down and thread the top for a more convenient size
I think its like the arrangement on an adjust-tru lathe chuck
stustev: they center the probe body to the spindle. they are on the probe itself
are you trying to adjust the probe tip to the center of the spindle?
jmkasunich: I may just turn both threads on a piece of brass or aluminum. I have a cnc lathe you know.
stustev: yes I want to allow that adjustment since it's built into the probe
which piece moves? there is an outer ring, and the inner part that is tapered and threaded
it looks like the inner one locates to the bore in your new holder, so it sits still and the main body moves around it?
the bore is loose around the center post on the probe
there are four setscrews - you can barely see the top one
yeah, I saw them
they used another scheme in their holder with some crazy cones and stuff that I don't want to replicate
so what locates the outer ring on your new holder?
you didn't machine any OD, I can't believe you got that lucky
yep lucky enough
spec is 25mm and my arbor was for a 1" shell mill
is it a nice repeatable fit, or just a kinda-close fit?
so I'll have a tiny bit less adjustment
one of us is very confused
let me assure you, it's you, but only because I can't explain
let me start over
on the probe, I see a shiny outer ring
the reference surface is the flat on top of the probe body, against the flat that used to be atop the shell mill
drilled for four radial setscrews
just mount the silly thing solid and use a probe calibration routine to determine the position of the probe tip relative to the center of the spindle
there is a collar in the arbor that fits loosely in a groove on the probe body
stustev: then you gotta put the spindle in the same orientation every time
cradek: I'm lookin at the pic
there are four setscrews that tighten against the outside of this collar to center it
there will be a drawbar to hold the flat reference surfaces together tight
that is true but not that hard to assure
my mistake was thinking that the outer shiny ring and the inner part with the taper and the threads were two parts that moved relative to each other to adjust
stustev: I have two possible orientations - the keys are the same
they are one part, and the screws bear directly on your arbor to adjust
got it now
ah I see what you mean. yes they are one solid piece.
stustev: I will have a spindle encoder though eventually... if I marked one key you are right it could all be done in software
is the face of your shell mill arbor ground? seems like black oxide to me
how accurate is that surface?
hard to say but it looks ground
but if I only have one probe length, it really doesn't even matter
the ball has to be on the centerline, that's it
well, you want a nice pair of mating surfaces
when used as a shell mill arbor, does the mill seat on that black surface?
there is a little burr where I pressed out one of the shell mill drive keys
yes it did
it must be ok then
I will have to grind off the burr with a tiny (dremel?) grinding wheel
it is black, but it's definitely ground
bet that QC30 toolholder was made from AISI 8620H (Almost positive if it was an Erickson/Kennametal)
it IS a real erickson
and you cut it up?
what's AISI 8620H?
jmkasunich: damn right I did
[03:17:23] <SkullWorks_PGAB> http://asm.matweb.com/search/SpecificMaterial.asp?bassnum=M862AS
8620 is a case-hardening alloy
heck even the taper is black, I hadn't noticed
that's obviously ground
it says RC61
back to drawbars.....
a long skinny bar that turns in the probe head thread is gonna have some torsional windup when you tighten it
many of the CAT40 holders are now Hard turned - only the bore ( for EM holders is ground).
you might want to consider a 1/2-13 plug with a unthreaded round hole and a flat top
then run a bar from the probe up thru the hole, and put a nut on top
what is the minor diameter of M8?
plan a won't work
I will have to make the M8 on something. I'd have to order M8 all-thread and have it shipped in. that's silly
plan b: use 3/8" roundbar - cut the M8 thread on one end
on the other end, drill and tap 1/4-20 (female)
then screw a 1/4-20 x whatever setscrew in and locktite it
so you can hold it with an allen wrench and tighten a nut on it - no windup
I'm about 800 words short of being able to draw a pic, I hope this is making sense
yeah that sounds clever
I use setscrews as studs a lot
or drill a hole in your table somewhere and use it to do an offset cycle like stustev suggested
(offset probe cycle)
we do it at work because we use alum electrolytic caps that have soft terminals - otherwise the threads gall and strip if you take things apart and together too often
[03:25:01] <jmkasunich> http://willepadnos.net/jmkasunich/pull-blocks-finished-1993.jpg
SWPadnos: it's tempting, but I don't see any reason NOT to align it, since there's a nice provision for it built in, and it makes life easier
heh - with a Dremel, the easiest solution is to "put a slot in it", which in that case would make the end suitable for a slotted screwdriver rather than an allen wrench :)
screwdriver slots are crappy compared to allen wrenches, and ones made with dremels are doubly so
cut a slot in it was the solution for a lot of things back when the only tool I had was a Dremel :)
if I had collet blocks, I'd get a slitting saw too
if your screw is soft enough to cut with a slitting saw, its soft enough for a screwdriver to ding up
(removing stripped allen-head SHCS, removing stuck wheel bearing tapers, making rings adjustable ...)
you've done a very pro quality job so far, don't dremel it
I don't admit to actually owning a dremel tool
Anyone know of English docs for dxf2gcode?
cradek: total change of subject: what is PIVOT_LENGTH
I'm pretty sure its the distance from the B centerline to some point on the spindle, but what point?
* chr0n1c just tried out cam.py from the wiki with a dxf from inkscape and it left out the arcs :(
the ini file has three values (two commented out) - the uncommented one is zero
jmkasunich: the point from which you want to measure tool lengths
can be anywhere
what point do you use?
some reference tool that's probably no longer in its holder
lol, cradek, i have a dremel for the spindle in my engraver
(I'd like to be able to use the same g-code and tool table on the sim as the real mill
leave it at zero if that's how I have it then
I really don't remember
IOW, we'll figure it out in galesburg
the vismach model uses TL=0 as the spindle nose
I guess I'll do that, so I don't have to change the tool rendering code
that means the pivot len is the dist from b axis to spindle nose
if you leave it at 0, no tool loaded (TLO=0) means the table won't move when B moves
pivot len will be 45mm - that is the distance from B centerline to spindle nose
(in my model)
if I put in a 45mm tool, then the tooltip will be on the B axis and the table won't move when it tilts
cradek: does your TTF converter work on 64bit?
64 bit machines...
chr0n1c: as far as I know
omg you have a font converter too?
do i need to build it? or is there a binary for 64 bit?
its just C, if you compile it on 64 bit it should work fine
I don't know of a binary for 64 bit. but it is a trivially easy build.
i'll try, i am a type make and hope it works kinda person when it comes to code though ;)
ISTR the makefile is something like 1 line
on a 5 axis machine with a gimbal (two axis) head an attempt is made to make the center lines of rotation orthogonal and coincident. This is never successful. It can be very close and usually is close enough to allow the machine to cut material with the necessary accuracy. The pivot length is the distance from the pivot point of the rotary axis(axes) to the gage point of the spindle taper.
the gage point is a theoretical specific point. In reality it can be exactly as cradek stated - anywhere you want it along the spindle axis. Your post or your machine config will need to know where the point is and any tool set device will have to coordinate with that point.
the coordination will allow the distance from the rotation center to the tip of the cutting tool. that is the important number.
to be known
is there any convention for gage point? or is it truly "arbitrary"
thinking faster (maybe slower) than typing
there is a convention
there is a gage line on the taper
I think it sits right on the nose
ok, so it is more-or-less at the nose
I think machinery's handbook shows the convention
* chr0n1c wishes for an gpl mastercam clone
for my purposes its good enough to know that it is near the nose, not at the tip of some arbitrary reference tool
the convention is not important as long as any tool set device allows the calculation of the length of the tip of the tool from the rotation center point.
I doubt max-nc's spindle taper is in MH anyway ;-)
heh probably not.
good night all
ehh someone needs something looked up in the book?
i have a 1943 version here
I hate emc
or rather, I hate the hoops you have to jump thru to configure the interpreter if you have angular axes
axis can jog xyzbc just fine
but G0B10 does nothing, and G0C10 moves B
(if its any use, axis jogs and runs gcode fine on this xyzabcuvw http://imagebin.org/19577)
G0B doesn't move A, but G0C moves B? - that sounds strange at best
* chr0n1c dreams of the day i get a rotary table
I'm doing trivkins at the moment, but that should still pass things on to the respective joints
xyzabcuvw isn't missing any joints
I think the proble is that I haven't defined a joint for "A"
trying that now
it used to be xyzacuvw and was fine
I think you will get 6 joints, and you only want to connect 0-3, 5 and 6
0-2. 4 and 5
uh, yeah - damned off-by-one problems :)
I have [TRAJ]COORDINATES = X Y Z A B C
define the one you don't want to use as a slave of another axis and don't hook it up
I also tried it without the A
chr0n1c: this has nothing to do with slaves
jogging works perfectly, but the interpreter (MDI) does not
its an interpreter configuration problem
but you need to define it to not crash?
so far I haven't had any crashes
B should connect to axis.4.*, C to axis.5.*, nothing to axis.3.*
doing that makes the jogs work right
I'm watching the axis.0-5.joint-pos-cmd pins, and they are just fscked
for example, when I do g0B10, none of them change
and yet you do get errors when you try G0A-
uh, no, you wouldn't with A in COORDINATES
but then I command G0X0 (mind you, it was already at zero), and _then_ axis.4 changes to 10
crap. I just realized that I have to shut down my laptop tonight, and I have about 50 windows open
terminals, browsers, editors ...
ok, I just did G0X0Y0Z0A0B0C0 and everything is zero
X10 does what is expected, did X0, again did what is expected
same for Y and X
is there a nonzero feed rate?
A10 - nothing happens (no movement on axis.3.joint-pos-cmd
G0 doesn't need a feedrate
I know - and yet I still ask the question :)
there was some other problem related to not having a programmed F word, maybe 6 months ago
SWPadnos: Pull the plug and yank the battery... shut down complete.
well, the issue is that of getting back to where I am in all those 68 windows (I counted)
and maybe not saving some changes, you know
SWPadnos: One at a time, one at a time.
SWPadnos: Had this been a MBP, you wouldn't have to deal with it all.
as even window positions are saved.
freaky - G0B10 doesn't move, but G1F45B10 does
ok - that sounds suspiciously like whatever that other problem was
this is really screwy
other rotary moves may work now, even without F or XYZ words
G0B10 doesn't move the _first_ time
issue it again, and it moves
issue G0B20, nothing
second time after startup, no F word specified?
issue it again, and it moves
or every second time?
seems like every second time, but there could be another pattern here
I have been issuing the same G0 twice in a row
wonder what happens if I issue different values for the two commands
ok, I just did a G0<everything>0 to get back home - had to do it twice
I guess it was more recent: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2008-03-08.txt
then I did G0B20 - the DRO velocity display became non-zero for about the right amount of time, but the DRO axis numbers and the HAL pins are all still zero
start around 22:32:00
may be related, may not be
I think the F word thing is a red herring
I believe this behavior was happening before I ever did a G1 or F anything
gonna restart and see if that is true
yeah - the other problem was for spindle-synchronized moves
ok, all homed, all zero
but I wonder if something similar happens at other times
G0B10 - vel becomes non-zero for a while, then goes back to zero, no DRO changes, no hal pin changes
G0B20 - vel becomes non-zero for a while, then goes back, DRO reports B is at 10 (should be 20)
G0B0 - B goes to 20
its like moves are being queued somehow
G0X10 - X went to 10 as it should, B went to zero (as the previous command told it to)
that sure sounds like the motion never happens, but the motion controller thinks it did, so it re-outputs the expected start position
or something weird like that
right - something too weird to look into at 12:18
g'night to anyone who's still up
Anyone know of English docs for dxf2gcode?
* Twingy just finished building vacuum table for his cnc
i wonder what kind of machine this MAZAK project is about
should be some pics about, there is a youtube vid of it solid tapping
3 axis, with tool changer iirc
hmm i´ll look for it on the project page there is no picture of the machine :(
big machine :)
sorry afternoon i mean
pjmcabbb is now known as pjm
pjm is now known as pjmm
hmm nearly dinner time
drinking my first cup of coffee
* archivist needs tick coffee boy for another
wee, just thinking about it has caused the coffee boy to start work
or more correctly, the boss arrived and the coffee boy is brewing for him
i found some nice 48v 3.5a switchers in the skip this week
archivist i got one of them steppers from that company
very nice indeed
I have a 20 switcher here 28V
yeah thats a nice size
i'm just working on the mechanics for my Z axis at the moment
so its some thick old lumps of ally plate to machine up
[12:07:32] <skunkworks> http://www.cnczone.com/forums/showthread.php?t=59150
odd why doesnt he use axis , it already knows the 4th axis
has anybody tried to use drawer guiders for XY table? i wonder how they would perform for some light milling
likely too much play
preload to reduce
dont know look here http://www.ladegeleider.nl/N_art_345.html
these seem to have a sort of ballbearing on 4 sides
would be intresting as they are so cheap
just folded metal, dont expect accuracy
may be suitable for an engraving set up "if" shortcomings can be dealt with
probably need 4 per axis as the balls are a only a short length
i mainly want to experiment, i just altered a exsisting mill as my essay assignment. now i want to build one from scratch
ow yeah i didnt think about that
better to home brew a bar system with cheap ball bearings
there is a web site somewhere with some pics
raid the local scrap yard
my mill is a built from scrap and reused parts
nice what did you use for spindle? a dremel?
a lathe headstock
XY comes from a tool setting machine, vertical col is lathe bed, headstock now on slides
archivist: any new pictures/videos?
well - heck ;)
actually since a server crash even the old pics arnt up yet
very slooooowly restoring
im buying gear cutters -_-
grinding my own if way out of my skill level.
we self grind up fly cutters for odd gears
I dont have any good enough grinder to make them..
matching old shapes
I'll just buy the cutters..
unless they are $200 each
good ones are not cheap, depends on type
heh, im making delerin gears..
I'm atleast going to inquire seco about them
we have a special indexing grinder to resharpen the clock gear cutters
seems like seco dont make gear cutters
what sort are you looking for
the day job sell one type
[14:26:26] <archivist> http://www.clock-works.clara.net/cata/wnpc.htm
yeah that's the kind
but note only clock gear form not involute
that's not as good :/
maybe I should just mount a air tool and grind them on the mill.
I could live without involute gears..
often its cheaper to buy a standard gear and machine to fit
the problem is that I cant get gears very easily
[14:32:57] <archivist> http://www.hpceurope.com/index.php#
hell, cutting a tool out of steel would probably suffice for making the gears I'd like
if sharpened enough
hpc is a stockist/gearmaker
I'd like a gear cutter then make 20 of a few gear sizes to keep in a drawer for then I decide to build stuff :)
are each cutter can only make a range of sizes
except hobbing cutters :)
[14:36:43] <anonimasu> http://www.youtube.com/watch?v=8UBxNPbg0ls
<- actually that would be the nicest to have
looks like reprap's got a decent l297/l298 board design, though unfortunately it's made to hook to a PC +5/+12V power supply instead of the 40V that an l298 will work best with.
the bicolor LEDs across the motor windings are an interesting touch, though I don't think it'll be that useful in real life
[14:38:18] <jepler> http://reprap.org/bin/view/Main/Stepper_Motor_Driver_1_1
[14:38:51] <anonimasu> http://www.youtube.com/watch?v=UqnqswBl6HA
why is it limited to 12V? there's no regulator on it
it's limited only by the connector style (and maybe by the capacitors they specify, though obviously you can substitute)
it's too bad they don't have 5V on the small pinheader
archivist: seems like thoose hobs are what I want
too bad the part where he cuts the teeth is missing
ah but it looks like the pitch on the PC power connector is .200in so you'd probably be able to use something else with a similar footprint (5.08mm terminal blocks)
certainly something to keep in mind when people with tiny machines ask about stepper drivers
archivist: got any idea how he machines the flutes on thoose?
just watching dunno yet
after part4 it ends -_-
[14:57:02] <anonimasu> http://helicron.net/workshop/gearcutting/
hmm hobs are screw cut!!
thoose are just gear cutters
there are a few that are confused!!
look at that last page
it looks like a much better one
he is indexing so getting flats when milling
I can do that kind of thin without a problem
could "run in" with grinding paste as a second op
* anonimasu nods
anonimasu: there are more parts, but I couldn't figure out the ordering, something is screwy
the other page seems to have more info about then
I know the videos are for beginners, but I spent a lot of time wishing he'd get to the point so I could just see his setups and results
* anonimasu nods
same with me
the interesting part is cutting the teeth
here's the blank, here's how I cut the teeth in the cutter (plunging a 20 degree tool), here's me hardening it, ...
it makes it seem horrendeously hard to make thoose cutters
if someone is learning about gearcutting, you don't have to tell them how to square up a vise or face off a blank
I think I'll just grind a single point cutter in the big mill at work.
get a gear, grind cutter till it matches the forn near enough, go for it
we do that for escape wheel cutters and turret clock sizes
I'll grind it when I go to work later
the stuff I do now gets / /
I'm going to go do it.
Im just pricing a hobbing job for 200off model loco gears
cradek, I wrote a widget to calculate speeds and feeds and would like to send it to you to check it out to see if should be put on the wiki what do ya say? 487kb
als: cool. if it's got a free license, it would be nice if you'd put it up there. I'll have a look if so.
what do you mean free lic?
I wrote it
interesting program for creating optical encoders: http://www.mindspring.com/~tom2000/Delphi/Codewheel.html
als: when you give it out, you also get to choose the terms. putting a GPL statement on it is a way to be sure everyone knows that they are free to use/modify/improve/study/redistribute it
als there is one on wiki now but it don't do a lot
GPL prevents someone from taking your work and making it non-free
cradek, pm your email and I send it your way,and you can let me know what it may need
hm, I chickened out on the grinding I need a proper grinder first.
I was just planing on sending the executable
BigJohnT, the one for drills?
yea I think that is what I did it for
I never really finished it all the way for cutters etc
this one gives you a range for SFM type of cutter, max machine ,rpm carbide hss material type
this is what I use all the time. it is on the wall next to the mill. http://timeguy.com/cradek/01192416516
sandvik give away a calculator sometimes (must find new batteries for mine)
wedevåg has 2 programs one for drilling and one for milling
[16:02:01] <OoBIGeye> http://wts.wedevag.se/Download/
this one outputs M03/4 Sword Fword
it would be cool if the interp could take diameter/surface speed and set S,F
I guess it does, indirectly, in lathe mode
but it gets the diameter from the X position, not from the tool table
what do I need to do to release it GPL
[16:10:10] <cradek> http://www.gnu.org/licenses/gpl-howto.html
can't I just put the executable up there and they us it as they will?
an executable alone is not as useful as source alone or source+executable. It's not easy to improve an executable. An executable may or may not still run on our computers in 4 years.
gcc is haveing trouble compiling the release version,it compiles in debug form?
internal consistency failure looks like from geometry.h
pjmm: you might be interested in some encoder drawing tools i've collected here: http://fennetic.net/machines/index.php?encoders
fenn ok thanks a lot, i will check that out
cradek: how did your boring operation work out?
* archivist reads some more of the 5000000 counts per rev saga on the mailing list
oh I saw the start of that this morning, 2 posts was the count, I havent opened my email.
cradek: oh execlent man
yeah, very good, those arbors, are like 45rc or so
hmm nice toy
going to do the email thing
gezr: we decided it was RC61 since it was an Erickson
my file wouldn't cut it, just slid over the edge when I tried to chamfer it
heh efffing hard
cradek: no way
cradek: but Im probably wrong :
seriously, it doesn't file
stupid ) key
really, did you try filing on a secion you turned?
ie not a case section
the corner I wanted to chamfer was a newly turned surface meeting an original surface
you just fail at filing
and you used that brazed carbide hu?
oh, I'm up to speed on the 2,000,000,000,000 ppr encoder stuff, facinating
someone needs a cluestick
I looked at his web site today
or an old timmy speedometer, the sort with a friction disk and stuff
friction, only ever seen magnetic speedos
correct me if I am wrong, on a 4tip ballscrew, at 100rpm, that thing would give, what, 0.0000000001 positioin resolution?
maybe magnetic, I dont really know
something silly and beyond sense and temperature control
yeah, I didnt think in terms of temp, but I did consider measurement methods
is his application a home brew, or in clean room? I dont remember reading that bit.
well website implies garage mechanics
well, I hope he gets his moneys worth
how much would that be anyway, those things are 700 each?
speed reducers are going to be 300 or so
measurement equipment to make sure its all working right is going to be what, 3 or 4 mil?
I'm logging. I don't understand 'thislog', anonimasu. Try /msg logger_emc help
I'm logging. I don't understand '!thislog', anonimasu. Try /msg logger_emc help
archivist: I dont know, I dont know enough to understand what the goal with that is. sounds interesting at the very least dont you agree?
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-06-06.txt
gezr either interesting or complete missunderstanding
heh - the first picture on his site has the mess he brought to the workshop.
[18:08:22] <skunkworks> http://conceptmachinery.com/
(in the background. (not the mess in the foreground)) ;)
omg nice plastic flap on the PC case
I thought that as well :)
i kept going to conceptmachine.com and wondering what the big deal was
i guess you gotta start somewhere
I started out digging ditches and burying horses and what ever else I could get to bring some food to the table
then you figured out eating the horse had a better ROI?
yep, now I only dispose of the bones
$50hr for backhoe work unless I have to hold my nose then it's $150
so no cespool digging then eh?
at least your not having to carry sheet with 2 buckets and 1 stick
BigJohnT: how you doing today?
cesspools are clean when you start
FDH! here gezr
Fat Dumb and Happy!!!
ah, :) cool, I got my bike parts in yesterday, did some integration by parts today, so I dont have the mental aptitude to really be working on my bike right now but im going to anyway
its rather depressing to me, at the moment, sorta perplexing
and on top of that, the longer I put it off, the longer I can wait to see if it will run
try sorting some nuts and bolts, always works for me
I bought a Honda 150 Dream once for $25. The guy said there are two in that pile and in those boxes just take one of each...
yeah, I pulled the part bags out. In a way, its a lot like legos
lego motorcycle - pure genius
legos are the basis for everything
fiberglass 10x scale legos would be great
the technics ones with the holes
legoservomotor with a 4 point snap in place legoservoencoder, and not to mention the legoio-block
you know they make those right?
too expensive though. gotta subsidize those bicycles you know
seen the things they use at the cement mixing yards? those concret blocks looks a lot like a giant lego
Ide live in a lego house
hell, if it was free, ide wear a lego guy suit,
and even hold my hand in the ever gripping the round pole C
hmm, that sounded bad
darn 5/16 SHCS won't fit in a .257" hole darn engineers... oh wait I'm the engineer
I love it when people have ideas about what you've designed and they have no clue that you designed it
"This should be like this the current solution is no good, who the hell designed it"
"I did and the reason is that the other way wont work properly ever"
then you have a big ass motvation to why not if they still feel like debating :p
or "uh yeah I drew up the bolt and then I designed the rest around it"
designing around stuff in stock syndrome
that's not true :)
not for me atleast
Im contemplating servo motors in parallel, I have a "few", erm 100 ish
oops 0 missing
must take a pic of the pile
made by canon precission 75gram-cm at 4K rpm and 12v
puny little things hence put in parallel
were printer head traverse motors
archivist: ah you must be referring to "in situ resource utilization"
its the latest NASA development process
as in crap in a plastic bag the bog is broken
* BigJohnT likes unloading trailers full of electrical panels in the rain
transimision is almost completely together now :
BigJohnT: As long as they're not connected to electric source, your okey =)
no, they were brand new
but arc nicely when plugged in and wet
brand new and no plastic packaging? what is this world coming to
old is the new new
my qemu quit working :( stoopid automatic upgrades...
emc2 still works though!
ummm yah tolerance and repeatability of about .001" http://www.cnczone.com/forums/showpost.php?p=460435&postcount=12
skunkworks: really nice work
skunkworks: I think his calipers are borked
maybe ;) (I think it has been a long week and I am a bit crabby)
maybe its brand new wood
speaking of wood... a woodchuck aka groundhog ate my lima beans
while i was playing with my cnc** (see, it was on topic kinda)
woodchucks are the future of CNC
* renesis said it first.
* chr0n1c seconds that motion
skunkworks: just take the rest of the week off...
BigJohnT: thanks :)
* archivist looks up at clock, its playtime
* BigJohnT wonders where you get black iron pipe that straight to get 0.001"
archivist: I thought it would have already been playtime for you.
impressive - huh?
not exactly been struggling to work all week BigJohnT
I like the old socket set tool box electrical panel
* archivist feels idle
* BigJohnT is going to town to get a 6x2 and some carriage bolts
archivist: what's it called when you engrave a sharp corner using a conical bit?
within .001" over 1" maybe
until the humidity changes
90 degrees right now in dayton, ohio
but he has wooden calipers won't they expand at the same rate as the machine?
90 degrees and 100% humidity here in bloomington, indiana
my wooden calpers are made of granite
okok so i went through almost ALL of the cams progs on the wiki page last night and none of them did a whole lot of anything that i could cut... i think it may lie somewhere in the inkscape or qcad dxf output... part of it may be me, who knows? so how does anyone get dxf to gcode to run with emc on linux?
i know a few way to do it with windows...
what kind of part are you trying to make?
like say an engraving on a zippo from a jpg... but i need to be able to edit the dxf... and have control of depths and such in the gcode
i know i can tweak the code by hand...
but how do i get there?
how did you get from jpg to dxf?
traced black edges with paths... then saved it as dxf
it looked ok in qcad.. with a bunch of extras....
cool. if it can do that, it could export gcode directly.
then i deleted the fuzzy lines... saved it as a dxf again...
then tried that dxf in all a bunch of cams from the wiki
that's what i was saying yesterday...
inkscape and qcad would make a sweet cam prog mashed together
i have no clue where to begin that journey
other than a stack of books i've been dreading to read
i have seen....
some sort of dxf output for inkscape, i think on the wiki
but it still didn't do a whole lot of anything to cut
actually the files always ended up empty excpet for a few lines of code to init the machine
Hello, If some has a minute to help with a Pluto, I'd appreciate it.
[20:41:14] <chr0n1c> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?InkscapeHowto
that's to output dxf from inkscape
or i mean gcode
kirk_wallace: whats up?
(I have played a 'little' bit with the pluto)
I get insmod error. Iam trying to copy the message but having trouble there too.
I will throw some thing out there while you're fighting with copy/paste.. How is the printer port setup in the bios? (it need to be some flavor of epp)
I have it set to EPP
and at 378
I have in the hal "loadrt pluto_servo" and very littile else.
That should work.. what is the error?
Still trying to paste
heh - use http://pastebin.ca/
insmod: error inserting '/usr/realtime-2.6.15-magma/modules/emc2/pluto_servo.ko': -1 Input/output error
are you using a cable between the pluto and the printer port?
You mean it has to be connected? Just kidding, it's plugged directly in.
The pluto driver defaults to 378 - so that should be ok.
how are you powering it?
* skunkworks powered it off of the usb port.
I tried a NetMOS with an address (0xC000) and no luck.
I have a 5 Volt (switcher) module.
Oops. I measured on the board and got zero volts. Getting another module.
Oops, another mistake. I wasn't on Pluto's ground. Now measuring 5.4 Volts.
5.4 is a bit high
I think it takes up to 9 volts or something like that.
it has its own regulator
is the input high enough to regulate
I measured before the reg., It's supposed to be a LDO.
yes - it is like 3.3v fpga
(I have run it on 5v no problem)
i may have just figured out some of my inkscape to dxf problem with persistance...
The EMC driver loads the flash, so I don't need to anything outside of EMC to init it?
kirk_wallace: can you try loadrt pluto_servo epp_wide=0
yes - emc loads the fpga
Yahoo! EMC came up. Thank you.
Should the LED be lit?
I think the led is the output of pwm 0 out.. iirc - so it shouldn't be.
Okay, now on to the next steps
fenn_ is now known as fenn
haha. toasty where u live at?
how old are u
u build ur own machine?
I thought even toasty was too smart to answer that question
why wouldn't i answer that question
no i haven't
are you even contemplating building a machine?
can a Hummer be repurposed into a machine tool?
i bet you could probably make it into a lathe of some sort
I remember engines being used somehow
there's some ridiculous project that changes an engine block + other parts into a multitasking machine
I'm just thinking they will be cheaper than chinese mill-drills soon
a lot more rigid, too.
eric_U: nah, there will always be people with more money than brains
(and with egos as big as their wallets)
well, it depends if they want to save the environment or if they want to save $$$ on gas
I think a lot of those sales were financed by home equity loans
because the hummer has a lower energy and pollution cost than a prius does over the life of the vehicle
but the gas is insane
no it doesnt