open a file, write one byte, seek forward a gigabyte, write another byte, close the file
in unix, this file won't take up a gigabyte on disk
databases are often like this
is that little dc motor really 8KW? http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=160128276964&ssPageName=ADME:B:WNA:US:12
alex_ is now known as The_Ball_
and does "Tach: 9.5 V/Krmp" mean that you get 1000rpm per 9.5volts applied?
for about a second
skunkworks: do you interpret it to say that the continuous is 140V * 10.2A = 1.4kW?
but that is my opinion
The_Ball_: "tach" is an analog speed output -- it reads 9.5V per 1000 RPM
i just picked up two alu plates 320x220x32 and 320x220x10 of 5083. They will become a vacuum table for pcb milling
jepler, ah, of course a tachometer
so they'd need encoders before they could be used as axes
30,000 rpm. cool
jepler, centaur seems quite knowledgeable about filters, we might put some filters on the inputs designed for the pwm frequency of the pluto
hmm - maybe 10A stall is what ever voltage that would be.. Maybe less than 140v. still- pretty scary servos
skunkworks, buld me one of your pwm boards for 1.4kW thanks
The_Ball_: that would be great. is there anything about the pluto's pwm output I should clarify for him?
not really, the pwm frequency is 19.5kHZ right?
does it do 100% duty cycle or 99%?
it can do 100% duty cycle
problem with 100% duty cycle is that the frequency dissaperes
same for 0% of course but that is desirable
if you don't want 100% then set the max-dc parameter.
since the duty cycles (after scaling to internal units) go from -2047 to 2047, you would want to set max-dc to (2046/2047) = .9995...
and adjust gain by the same fraction
that is, if you want "the largest duty cycle less than 100%"
yes, that's perfect, then a narrow filter at 19.5kHZ would reject noise very effectively
the pluto has another option, at least for now: interleaved PWM
is that the PDM? I don't know how that signal looks
for "interleaved pwm" the upper bits in the pwm counter are rearranged before they are compared to the pwm duty cycle
for duty cycles less than 50%, the "on" times are 4 cycles of the 40MHz clock (so 10MHz/10ns) (except for at most one "runt" on time per 19.5kHz) and the "off" times are a variable number of cycles. For duty cycles above 50%, it is the opposite with off times being 4 cycles.
I was originally told a mode like this was useful for conversion to analog, but I don't know if my implementation choices (such as the number of bits to interleave, which determine the length of the mini cycles) were appropriate
on thing, is the pwm signal hardware generated or by the firmware. down the line if the firmware is updated will the pwm frequency change?
jepler, i am totally unfamiliar with this PDM mode
the signal is generated by the fpga firmware, so within the constraints of the chip resources it can be changed
do you have a machine that can run emc?
in simulator mode, even
not right now, im at work
OK let me see if I can make you a screenshot
emc's "pwmgen" in PDM mode is similar (but not exactly the same as) what pluto does in this mode
i have been meaning to do a 2.2 checkout, it might be time
I would thing the h-bridge should do 1.5kw.... 150v and 10a
yes, i guess with the proper fet's, what is the rating of yours?
500v 44a. but that is not really possible.
I am hoping for around 150v 20a with some peak power,
sure it is, and you get a nice pyrotechnics show
-- from top to bottom, pwm; pdm; produced from analog input
jepler: I found those motors you used for the etch a sketch..... $3/ea and ~50 available
new old stock
jepler, thanks, studying it now
Jymmm: don't buy too many before making sure they're adequate for even the toy needs
Jymmm: but, cool
backlash is a problem with those motors - isn't it?
skunkworks: not the servos, the "pancake steppers"
if Jymmm had found those servos for $3 I'd be first in line to buy more
The_Ball_: in pwm, you have to wait on average 50% of the pwm cycle (e.g., ~50uS) before the output goes the opposite way; over a full period, the average output is the desired value
Yeah, sorry... no such luck there. Though I did find some VERY nice enclosures for the new controller, bought the last two they had.
The_Ball_: in pdm, you wait a much smaller amount of time, except when the desired value is most extreme (close to 0% or 100%)
jepler, yes, the pdm signal appears as a pwm signal at a much higher frequency
[02:07:48] <jepler> http://en.wikipedia.org/wiki/Pulse-density_modulation
hey, are GE faunic programmable controllers any good?
which is a good thing in analog conversion, less ripples
if so, what are they?
Jymmm, which series of controllers?
The_Ball_: sorry, no clue... black
Jymmm: beats me whether they're any good, but they are CNC controllers -- the same thing emc does
These were about 5" tall 4" wide and 6" long
we have a ge fanuc controller at work running the laser
pew pew pew
skunkworks: how do you program it?
or it is like a one time thing.
fanuc is uh
an amazing controller
It has ladder logic in it - but it is a machine controller.
The_Ball_: what series should I be looking for?
takes gcode and runs axiss
oh takes gcode directly????
they're what you see on the front of a lot of commercial cnc machines
well, shit.... maybe I need to head back and take a closer look
they're not compatible with emc gcode though
dunno why you'd want something that isn't compatible with rs274ngc
there's a reason fanuc g-code is like, an industry standard
Jymmm, i have some s2k controllers, they are fully programmable motion controllers, over kill for running something with emc, but very happy with them
(I'm trolling just a bit when I say that)
most controllers have a fanuc g-code compatability mode
The_Ball_: silly question, but why would you run them with emc?
they are motion controllers, they do not accept gcode
* Jymmm scratchs head
motion control and machine control are different
19:11:08) Jymmm: oh takes gcode directly????
(19:11:11) skunkworks: yes
(19:11:13) toastydeath: yes
so you program it to control a servo motor to do a specific task depending on the logical inputs, then leave it
his s2k controllers
are motion controllers
s2k != fanuc ?
we were talking about fanuc machine controllers and switched to motion control, i thought
or am i high
Fanuc's crap is usually labeled like "18T"
if you told me your bridgeport was made by chevrolet I wouldn't know if you were telling the truth or pulling my leg
ok, so motion control is what?
s2k != machine controller, yes
i don't know anything about motion controllers, just machine controls
Jymmm, a fancy PLC with a servo attached
The_Ball_: no drive in between?
the motion controller is the drive and a fully programmable PLC running some realtime os
ah, ok. making more sense now.
a programmable machine control is a g-code thingy that allows you to customize what it does
versus a dedicated mill or turn unit
Jymmm, here is a picture http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=180119980489&fromMakeTrack=true&ssPageName=VIP:watchlink:middle:us
could I use a motion controller between EMC and a servo
to preform additional error correction on an axis?
these support a secondary encoder yes, so you get servo stability and correct possitioning
have a look in the manual, it's on page 500 or something ;)
toastydeath: in principle, emc can use anything that can take a position command at the servo rate, and return a position feedback at the same rate -- but you have to write the driver if it's something exotic and new
jepler: i'm talking about multiple sources
The_Ball_: yeah, those ar enothing like these at all
yes, but it seems you guys were talking about machine controllers which is different
g-code accepting devices
a big ass control panel with wires hanging out the back
bunch of switches, etc
jepler, PDM seems to be the way to go for faster response in the PDM-analog conversion'
jepler, allows for a smaller RC delay
The_Ball_: I guess maybe I should talk to centar about how to make the PDM mode most useful for a digital-to-analog converter. However, the PDM mode might not be useful to you personally, because it has to be enabled for all PWM outputs or for none, and I seem to recall that one of your motors was PWM input and the others were analog input...
jepler, I am building a H-Bridge for the Z axis yes. The PDM mode switches to quickly for the FET's I guess
right now it is effectively 10MHz, way too fast for FETs
if centaur thinks a different rate (I could easily choose any 40MHz / 2^n) would be a better compromise for filters, I'll change to his suggestion
what's the --enable-run-in-place .configure parameter do exactly?
do you mean, what's the use? or how does it work?
what behaviour does it change?
with --enable-run-in-place, you don't "make install". instead you ". scripts/emc-environment" in each terminal window before running emc, halrun, etc
ah, i see
it means that you don't need to install any files outside the emc2 tree you got from the cvs checkout
it's very useful to developers because we often have several source trees on a single system
it's also a good idea for anyone who wants to install a .deb package of emc and also compile another version, for whatever reason
yes, very nice indeed
(otherwise, "make install" will overwrite some files from the package [files in /etc] even though the package has prefix=/usr and the other version probably has prefix=/usr/local)
jepler, just to state the obvious, in pdm mode there is still one up and one down channel, right
if you wanted, you could set scale and offset with knowledge of the input range so that an output signal only appeared on one channel
you can also set the "pwm+direction" mode for pdm
i have never run emc on a quick machine, this is nice, time to upgrade the hardware i think
the limits bounding box in 2.2 looks very nice
03jmkasunich 07v2_1_branch * 10emc2/docs/src/hal/drivers.lyx: document revisions to the 5i20 driver (index fixes)
here is an off beat question
there might be an off beat answer!
ok, that's enough beating off
03jmkasunich 07TRUNK * 10emc2/docs/src/hal/drivers.lyx: document changes to 5i20 driver to make indexing work
can i have EMC move to a location X-Y-Z then call a user defined M function that takes over Z axis movement externally but watches the encoder count until a specified target value is reached - then EMC takes over again.
use EMC for positioning
including a G43 offset for electrode
then let the EDM handle spark control / flushing infeed
until Z axis has done the programed amount - then rapid out and prep for M6 of next electrode (assumes a multi burn to compensate for erosion.)
the only move emc can make that ends up in an undetermined place, and then recover from, is a probe move G38.2
maybe a do / while loop?
if you could issue a probe move, and then use adaptive feed to control the speed, and then hit the probe button when done, ...
jmkasunich_ is now known as jmkasunich
cradek: this would leave the encoder feedback live - just routing the motor control to the voltage sense circuit.
then once the Z target is attained the motor control switches back to EMC
the chosen GUI would show the Z axis movement.
kind like when the servos are in e-stop but encoders are powered and you manually crank a table - encoders continue to act as a DRO
yeah the only time emc will follow the encoders and not give a following error is when in "machine off" state
and I guess there isn't a way to isolate a single axis
I don't see an easy way to do what you want - if you aren't using emc's commanded position, you'll get a following error
is there no way a hal command can specify a position of a axis?
people have done this kind of thing in HAL but they have to "lie" to emc to pretend the axis hasn't moved
OK - just have to run dual Z axis
the DRO behavior on one axis is not possible
EMC would not be aware of the second slide
well, anything's possible, the program is yours :-)
but not everything is simple
in my mind - I was thinking of a loop that would check a bit. 0 would be go further in - 1 runs it out. The bit would be controlled by the setpoint of the arc voltage. each time tthru the loop - emc would advance or retract by an amount like .0001
but that might be too simple.
do you mean a gcode loop?
thats not a problem - the EDM has its own internal voltage comparator to handle infeed
gcode can't check inputs like that
this is sinker?
I thought you could set or check I/o
set, yes - check, no
well then - never mind :)
the hope was that I could use a single servo and read the encoder to know how far from the first "short" the infeed had gone
then at the defined depth the feedback loop for the edm is switched off and the servo retracted by EMC
where on the scale is 20LPM of vacuum? It's for a vacuum table to hold pcb's, but 20LPM doesn't mean anything to me
20 liter per min?
free flow rate
oh, per head, so it's 40 in parallel
[03:33:00] <The_Ball_> http://cgi.ebay.com.au/ws/eBayISAPI.dll?ViewItem&item=190126336923&fromMakeTrack=true&ssPageName=VIP:watchlink:middle:au
we are using a 1hp vane pump... To much? ;)
no clue on the specs
some millitary surplus
hehe, that should do the job
oh well - bed time. :)
what about some crazy Osub to handle the servo movement
a do /while loop
and a check of an input pin
incrament a G91 z#1
+/- .0001" per check, then reloop
neat part is you could actually improve retrack flushing
I think I'll write this up as a NGC file in proper syntax then post it for feedback
we went over that - read above :-)
(summary: you can't read hal inputs in gcode)
right - but this takes the output signal the would normally control the feed by the EDM and uses it to report a T|F condition to EMC, which in the do loop
will handle the movement
SO EMC is now doing all the servo movement
Skullworks-PGAB: EMC can't read the T/F condition
nada at the moment. Got some enclosures today though, so I'm jazzed =)
They're even rack mountable too
I'm trying to get around to machining a fibreglass mold..
I hope the material you're maching isn't fiberglass
unless of course your masicistic, then GO FOR IT!!!
take all your clothes off and rub your body againest it!
I'm going to machine that blue foam stuff..
I hope you have a GREAT vac system
and HUGE tank =)
you'll need it =)
just a vaccum cleanre..
I'm not kidding.... the stuff goes everywhere.
why do you make everything so damn hard?
Hey, it's your crap, what do I care =)
I only make and sell foam cutters =)
I build machines and program plc
for a living..
and your point is?
Okey then, gawd damnit! LOL
and that was the whole point of it ;)
now to get this stupid device to sync
you just said you dont have a point, make up your dman mind already!
Eeeeeesh, kids these days
a windows ce computer..
[08:55:38] <anonimasu> http://www.cc-systems.com/en/onboard-pc/images/stories/CCPXLII.jpg
anonimasu: what is it, an lcd?
it's a computer..
anders, how is it up there? nice & cool?
today it's very cold :/
we're struggling with 36+ here :((
my body hurts like hell from this weather
it's been around 20 - 30c other days
alex_joni: done anything cool lately?
hello, Alexjoni: It seems there isn't a halui.program.stop, any chance for create a new one?
did I hear that EMC2 is known to run better on "older" hardware like P4's or is there no truth in that?
it all depends on how rt-friendly your chipset is..
The_Ball: P4 is now "older hardware"? That's news to me
my experience is that P3 machines are very good for realtime performance
jepler, sure isn't new ;)
oh, I assumed that was a typo
cradek, when you say "good for realtime performance", are you referring to the lowest guarantied latency?
or having acceptable max latency at all
some machines don't work with realtime at all like jepler says - that doesn't seem to happen with P3 machines
so newer hardware, even though it has more throughput, sometimes have higer latency?
Pentium 4 and newer offerings from Intel frequently use something called "SMI" for power and thermal management, which can cause long interruptions at regular intervals (e.g., 2ms every 64s on one developer's board)
there are chipset-specific ways to disable this, a collection of them is in the module rtai_smi.ko
bring back the 6502
my dual P3 system has a max latency of about 10us, or 5us if one cpu is isolated
6510 for me thanks
it's a great realtime system and cost me about $150
very nice, but is axis very slow with larger jobs?
it's acceptable for whatever I have done, but I suppose there is always a job large enough to make it slow
i was thinking of pcb's, they can have lot's of code
now try 3d contouring.
oh, yeah i guess, haven't gotten to that stage yet ;)
i picked up the alu for my vacuum table, 320x220x42mm... i think i went and did a overkill again
there are two tricks to know if you have large gcode files: First, put a comment (AXIS,stop) to make the automatic preview only display up to that point in the file. Second, hit the "clear backplot" button from time to time, and/or change MAX_POINTS in emcmodule.cc and recompile to have the backplot automatically trim itself more frequently
ah, very handy, i knew about clearing the backplot
42mm seems like a lot
yes, two plates, one 32mm and one 10mm, that way i can machine the insides of the table... but it's going to be a big lump
it'll take forever to machine ;)
i will complete the machine enclosure before i start machining that for sure
oh, I forgot this great intel thermal management feature: if the CPU is heating up, the clock is turned off for up to 3.0 microseconds at a time. (http://download.intel.com/design/Pentium4/datashts/30235104.pdf
jepler: atleast it dosent fry your cpu as much as the athlon thermal management ;)
ah but the AMD CPU will keep position right up until it fires the heatsink through the hard drive and sets off an atomic bomb explosion
wow, i assumed it reduced the clock frequency. i remember when the feature was introduced, they played quake at tomshardware and removed the cpu cooler while playing the game
now which is better, I ask you?
I perfer my meat hot but not fried.
finished bleeding yet?
it is healing well
i crashed a machine AGAIN at work
i really hate this control
it is too user-friendly
it throws me off
quit clicking the button labled "Crash"
i would prefer to stare at a cold, green CRT
with cryptic letters on them
not "mill line"
staring at 10 lines, with the exact same text on them
does not differentiate what you are about to do
03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: fix mouse jogging, maybe without breaking anything
this is one up on a hamster wheel huh? ;)
cradek: commit, then test?
how many ways do we need to jog anyway
just one more ;)
* ds2 waits for the "hamster barbell weights" commit
Mouse jogging? Sounds icky for some reason
hmm where do I fit the hamster to my hobbymat lathe?
archivist: You dont' want to know
it may improve said lathe enough to make it useable
archivist: NSFW http://www.realhamster.com/
archivist: When I say you don't want to know, I mean you REALLY DONT WANT TO KNOW! =)
03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: -/= jogging doesn't work for now
why hobbymat is crap, cross slide hand wheel calibrated 0-9, screw is 1.25mm pitch
was just taking it to bits as a cnc donor
I have hated it for a while
yes thats the ugly b.....
I have a long list of faults
So, you cna't print out a label and attache it to the wheels?
I actually have a program for printing rulers like that btw
heh that poor bloke thinks its good
I suppose he has never used a schaublin or southbend etc
archivist: And like you have </troll>
toastydeath: well, run TKEMC instead :)
Man I hit the jackpot yesterday, and all the odd ball stuff I've been looking for that (for some reason) is been hard to find; Enclosures, 15A IEA 15A filtered socket, 2.5" cap clamps
toastydeath: oh didnt read sorry
03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: back to v2.1's operation which worked best so far. too many ways to jog.
archivist: I was just playin, no need to take your ball and run home ;)
still attacking crap hobbymat
hobbymat is a fine lathe =)
contemplating is it worth doing anything with
do you have to re do the ways on it?
ds2: schaublin is a fine lathe..
"ways" its a flatted round bed
anonimasu: the hobbymat is better then the 7x10's out of thebox
archivist: then take it to the scrap yard for $5 (you pay them), then get on with life!
anonimasu, we have a few schaublins here but too good to attack
archivist: do you have the mill attachment for it?
might have been nice for a 4th axis
archivist: taking them apart is amazing :)
the hobbymat its no way sturdy enough to use for milling
archivist: it's made by retired watchmakers.. :)(must be)
sure it is... mill foam for lost foam casting
I like my swarf to fly over the tailstock
archivist: Ah, so you're just in a rant and rave mode, gotcha! (Must be PMS - Pre Machinable Status)... Carry On!
tee hee heheh http://rapidshare.com/files/39514465/Building_The_TESLA_Turbine_LQRS.pdf.html
03alex_joni 07TRUNK * 10emc2/src/emc/usr_intf/halui.cc: add halui.program.stop (it does the same thing as halui.abort, and helps only for consistency in naming)
has anyone machined HDPE?
03alex_joni 07TRUNK * 10emc2/docs/man/man1/halui.1: update manpage with halui.program.stop
test a peice with kerosene
if it doesn't mess the surface up, use kerosene.
as a lubricant
it will cut fine.
toastydeath: you and your kerosene
i'm not kidding dude
if you don't have a specific lubricant for plastics, kerosene works amazing for aluminum and plastics
so does diesel fuel
i've never heard that
pretty close to the same thing
but what are your questions about hdpe
Yeah, when you can't get a car on the towtruck, pour diesel fuel on the ramp, real slick then =)
does HDPE hold it's shape, or sorta kind "warps" distorts?
it also depends on how you want to define "warping"
distort is a better word
it will be very suceptible to thermal effects and workholding
i don't know if there's an internal creep effect to hdpe, but i have not heard of anyone mention it
Hmmm, maybe I'll grab a piece of ABS instead
you're going to need to worry about those things with any plastic
especially if you are milling it
I have tooling made for plastics.
that doesn't affect your thermal load or workholding
well, i guess it does, but not in a substantial way
what size cutters
and how fast is your machine
they're designed to prevent melt back. 150MPH in 2.4s
uh, what tool is measured in mph
you asked how fast the machine is
spindle speed, duh
you want rpms, ips,
8k-25K RPM, and I can do 60IPM (for now)
cool, so you won't have any problem
technically 90IPM (under TCNC)
that 90 ipm might come in handy
JymmmmEMC: I thought you were going way faster..
anonimasu: not under EMC
JymmmmEMC: like 200ipm :/
zip zip zip
anonimasu: oh no, more like 120IPM but not stable under the current (xylotex) controller. I'm hoping that the new controller will take care of that issue though.
my mill is around 150 ipm.. though it'll be down to a bit..
around there after gearing too
I need more accel.. :)
anonimasu: how bif is yours?
just a sec..
This is a 3" round acrylic disc (Photo out of focus, poor lighting at the time) I carved http://farm1.static.flickr.com/84/206158222_cb5c34978c_m.jpg
it's 260 x 150 and 310 Z
where is the best place to get rack bar for pinion
10" x 6" x 12.2"
anonimasu: Ah, gotcha. Mine is 24" x 24" x 6"
Unit41, HPC in the uk can supply
it's a mill.. :)
though I need the accels twice as fast..
it's what makes it zip zip zip when contouring
anonimasu: YOU GET NU_TING! lol
how do you balance a servo?
anonimasu: Schultz, from Hogans Heros'
Does anyone know of a free CAM program that can import DXF, DWG, 3DS, and other formats?
never read/seen it
apparently servos, maybe analog servos only?
toastydeath: oh you mean offset adjustment?
have pots on them for something called "balance"
anonimasu: TV show from the 60's/70's
toastydeath: I guess you monitor the position of them and tweak it until your machine stops to creep
well, i mean, what does the setting do
on the motor itself
sounds like s/balance/tune/
toastydeath: offset 0-10v..
to make the servo stop moving around with no power...
like, the setting i am blindly referring to
which is probably what you are talking about!
affects how accurately a servo can do countouring?
jepler: Do you have any mroe of those etch-a-sketch PCB's by chance?
toastydeath: that would be the PID parameters..
this is something on the motor itself
toastydeath: toastydeath not on the drive?
toastydeath: He means so the motor won't keep turning do to the shaft being unbalanced (like an egg)
no, the shaft balance isn't it, it's not literal balance
i don't know why this guy called it balance
Then RTFM the spec for the motors =)
JymmmmEMC: I mean the electric balance..
that might be it?
anonimasu: shush, I'm trying to confuse him with my ignorance =)
toastydeath: that's off, it's usually on the drivers..
jymmmmemc: it was too obvious to work
atleast for everything I've seen but I havent seen mich
toastydeath: Ya never know =)
i wish i had a set of motors to mess with
and look at
the test for this particular setting on the drive/motor/whatever
was a ground disk, of an arbitrary diameter
in the spindle of a mill, or parallel to the ways of a lathe
and one takes a high accuracy measurement device at 45 degrees, and you command the machine to move in an arc
then, you kill it with feed hold when it is about at 45 degrees
and measure where the tool is versus where it should be
and this adjustment on the drive or motor or something will bring that number very close to what it should be
and maybe he was talking about on ancient equipment from the 60's
it sounds just like PID settings
and now, you guys do this inside emc or in some controller rather than on the physical equipment
yes.. that's right
still it depends on your drives..
some drives might have that and some dont..
regarding motion controllers
is there an advantage to having a motion controller between the servo drive/motor and emc?
or am i making new devices up
yes, yesterday some of the guys in here were talking about a distinct unit for motion control
you mean a motor driver?
yes, jymmmmemc has it
the fanuc uh, s2k?
toastydeath: *slaps youÄ
i don't know about this stuff, man
that's a cnc controller.. :)
nothing to do with emc..
fanuc makes motion controllers and machine controllers
all fanuc's machine controls are labeled ##A
like, 18T or 20M
or are you talking about motion control cards?
that was (and still is) my confusion.
they're standalone units of some sort
that control the servo drives?
anonimasu: a black box about 4" x6" x 4"
it piqued my interest because they can take multiple encoders
well, there are programable controllers for use in the industry I for machines..
DIN rail mounted
who was talking about them
no, but who had the s2k's
that's as far as I know about them..
and that you can program them to do stuff ;)
there are programable motion controllers(cards) too..
no, these were not cards.
and those are what emc uses?
though you could ;)
but thats another story
well, that would be benificial
[20:53:47] <anonimasu> http://www.galilmc.com/products/accelera/dmc18x6.html
because as i understand, emc doesn't do the whole "bajillion encoders" thing
while these mysterious motion controls DO
toastydeath: theres a misnormer with that..
toastydeath: emc dosent do multiple encoders because nobody has cared about doing it.
toastydeath: it's not a limitation.. it's just not implemented because nobody has had a need for it..
well i can't program
but like, if these motion controls can do it out of the box, and return a pre-corrected location to emc
that would work brilliantly
you can read many encoders into HAL and do whatever you want with them
toastydeath: I think you are fooling yourself if you think it'll be better..
and pipe the output into emc?
you could nest pid loops or whatever
toastydeath: you would still need to have the controller do what you want..
sure, whatever the "output" is
toastydeath: better idea would be to learn about hal.. :)
anonimasu: yeah, but what i'm saying is that these controls seem to do exactly what i want
but if hal can handle it, that's brilliant
toastydeath: you have no idea what they do really :)
didnt you say so ;)
i have a vauge idea, and hopes
random musing: would it be possible to have a gcode section that probes the current tool, and when an input is received it uses the Z distance for a tool offset? (auto tool measurement)
lerneaen_hydra: yes, I did that already, didn't you see it on the list?
cradek: you did?
toastydeath: better would be to learn hal and implement it yourself ;)
must have missed that
anonimasu: i'm basically trying to stumble my way from what i know about kinematic accuracy
into actually producing an accurate positioning stage
lerneaen_hydra: he meant he implemented it..
it's easy for users now :P
oh, right :p
yeah it works fine
it's just one of the many things you can do with probing
toastydeath: im not sure it's all that well, since you end up with dual pid loops..
it would seem to be quite simple to implement with hardware (just get electrical contact)
i don't want to use two encoders to try and drive the same axis
i want to use multiple encoders to correct a master encoder
cradek: got a page describing it?
toastydeath: well, you really need to get specs for thoose little black boxes..
toastydeath: or just figure out how to do the correction..
the correction is easy
subtract the value from encoder 2 from encoder 1
then give the result to emc
lerneaen_hydra: you use probing (G38.2) to get the length, then set a transient tool offset G43.1
cradek: got a gcode + hal file sample?
hal file = hook the button to the probe input
can you stuff that into a macro?
oh the standard probe input?
yes the probe input, it uses the probe gcode
sure you could put it in a subroutine
it's very simple
oh I would put it in my CAM app ;)
toastydeath: I'd suggest playing with hal.. :)
toastydeath: or well, call GE about thoose things
yeah man, it seems that way
but im more into beliving in emc ;)
haha, i thought hal would do it but someone told me earlier it wouldn't
*comes from the guy that havent looked at hal*
would do what?
hey, i did look at hal
i looked over the docs for it and whatnot
cradek: correct encoders against eachother
cradek: or position that really is..
cradek: I assume using touch-off to get the work material with the right coordinates wouldn't break anything?
if the problem statement is "use many encoders to do ... something" then HAL can certainly do it
lerneaen_hydra: I don't understand the question
lerneaen_hydra: yes it'll still work..
cradek: setting your workpiece height
um, a coordinate offset done manually beforehand via touch off for instance wont mess up when using the autoprobe
I think the comments in the gcode explain that
so no nasties like reverting to the base coordinates
you touch off the first tool, and then the program measures it and remembers the length
lerneaen_hydra: I think we got away from that with emc1..
THEN it uses that to find the offsets for any subsequent tools
touch off uses g54, right?
also it's better if you can to set tool 1 to 0,0,0
also, you can touch off with a tool offset in effect, and get the right results
and have all your tools go off that
so they work across multiple work offets
yes ideally you have SOME kind of repeatable depth, but even if not, this scheme will work, since it measures and remembers the first tool (the one you used to touch off)
remembers it outside the work/tool offets?
does subroutines work in emc?
is it hard to make a sensor thats got repeatability to say 1/100th mm? (3 thou)
lerneaen_hydra: I got that with a regular old microswitch
[21:09:02] <anonimasu> http://www.gefanuc.com/za/ProductServices/MotionControl/Standalone/index.html
lerneaen_hydra: i have a sensor that will do that
paper from my notebook =)
.01mm is not 3 thou, it's .4 thou
toastydeath: looks like you need to program them manually still.
just about anything you use will trigger that repeatably
inches are not my thing :p
anonimasu: is this re - motion controllers?
[21:10:12] <anonimasu> http://www.gefanuc.com/Downloads/en/flexiblemotionsolutions_cutsheet_gfa611.pdf
so they're kind of hal in a box?
I guess control in a box for like machine tasks..
think production lines..
but from what i understand re: hal
you could implement a production line with that, correct?
this is more like PLC motion controllers..
or is my preception of what hal is so far out of whack that i should never talk again
does that mean never talk again =(
look at the application for thoose things
btw, any one of a utility (perferably command line and runs in Linux) that will take a STL file in and a series of DXF files describing slices of the STL object?
any one know of...
toastydeath: I guess you could..
emc sewing machine!
toastydeath: thoose motion controllers have their own OS..
[21:16:05] <anonimasu> http://application.gefanuc.com/infolink/manuals/gfk1848a.pdf
i am noticing that
that should help
that is a gigantic manual
* anonimasu waves a HAL flag
just hope it dosent set anything dangerous.
* alex_joni brings out the NML gun
i have EVEN MORE QUESTIONS, except these are unrelated to the other annoying questions i have been asking
* archivist avoids downloading manual and gettin distracted
does EMC have an MDI mode?
it's called ...
when i installed the emc environment thing
that was just emc
i didn't see one
which is why i ask
toastydeath: what are you talking about?
perhaps we don't know what you mean by "MDI mode"
manual data input?
all the GUIs offer an MDI mode or tab
toastydeath: each of the GUIs have a MDI mode
this means only one thing
in axis, press F5 then type in your gcode
i am far, far more retarded than i once believed
but there is also a standalong mdi program
ah, that would explain it
well - in axis it is also a tab..
yeah, activated by F5
but also labeled MDI
now if there was only an offset editor....
clearly, actualy looking at the gui is for sissies
[21:22:26] <skunkworks> http://www.electronicsam.com/images/KandT/Dapper.png
[21:22:32] <alex_joni> http://linuxwiki.de/EMC/EMC2_GUI
older - but similar
good night al
alex_joni: see ya soon =)
hey... is there a way to pic up were you left off in the middle of some g-code
or pick a specific point to continue from?
Just us chickens
is there any way to easily pick up in the middle of some g-code after an e-stop?
Ziegler: Yes, click on the line.. and press set Next line..
err the line where you want to resume at..
set next line?
where is that in axis?
I cant remeber on the same menu as run is..
ahh I see it cool
then just run?
but make sure you are at the right place before if it's a arc or such :)
and be careful if you have tool offsets and stuff..
you should ask someone that knows better then me..
about how that works with estop, but I think they are modal so it shouldnt matter..
I have been doing it the hard way... ie.. starting over
this isnt long code
but I have 10 of these signs to make, and they take about 8 minutes a piece
it should help you out lots