good night all
good night alex
new monitor, yea!
eric_42" wide screen!!!
not free but is linux, http://www.graphiteone-cad.com/demo/quickstart.htm
, interesting, like MDT from ACad, i hadnt seen the 'subtract the path cut by this dia tool' before though.
'free' demo version may be difficult to get running on ubu ( made for suse )
well its swiss ,so it gotta be suse (Q)
has anyone heard of or used a config that included motion.adaptive-feed? Is it the same type of function that commercial machines have, where feed rate is reduced when spindle current increases?
adaptive feed was originally added for the EDM guys, who adjust feed based on spark voltage
I suspect so, and its awesome that EMC includes it. The commercial guys charge $$$$ for it
it could also be used to adjust feed based on spindle load
so it orig was a "torch height" for EDM :)
jmkasunich, did you work on the HAL config for "demo mazak"
in this line, "setp scale.0.gain [SPINDLE]HIGH_GEAR_RATIO", is a new INI entry shown? Is the proper syntax to add [header] and then VARIABLE_NAME?
(to the INI)
you can put anything you want in the ini file, and then reference it in a hal file
the part in the  is the section name, you can use an existing section or make one up
the part after the ] is the variable name - you can use a variable that already exists in that section, or make one up
sweet, I am going to set up a 0-10V spindle command, but I am not sure what step on the pulley I want to keep long term. I will use an INI multiplier to set the ratio
windy tonight ;-)
anybody know how much you can heat aluminum without ruining the temper?
planning to press or shrink a 1/8" thick aluminum wheel onto a 0.312" steel shaft, trying to figure out how much interefence I should have
I'm at 0.309 on the hole at the moment
try .002 interference
what is the wheel driving?
winding up a string
probably a 10 in-lbs of torque
the wheel is 3" dia and only 1/8" thick, so I want a tight fit
if you can press it on go with .003 to .006 then
freeze the shaft
I guess 0.003 will have to do (for 0.006 I'd have to make the hole smaller)
jmkasunich, did you get an answer about temper? I cannot fidn anything in Machinery's
I found something online that says to be safe don't go over 212F
I might throw the wheel in boiling water to expand it before I press the shaft in
and freeze the shaft
heating the wheel will be more effective though, aluminum expands much more than steel per degree
DanielFalck: thanks, just looked at that, but just dog-eared it :)
i thought it looked alot like Mechanical Desktop from AutoCad
there's also python under the hood
got it installed on ubuntu 6.06?
no I haven't attempted yet. But I looked at the tar source file
it has opencascade in there too
it looks like they are now licensing some Unigraphics source
which wasn't in the source file...
I could have sworn I had a 5-40 tap
but it appears I don't
DanielFalck: hmm, nice stuff tho. i put it away thinkin i hadnt given Weber system's freebie a chance ( they gave it away at a EMC fest years ago) and i hadnt used APT enuf ( it worked sweet )
yesterday i couldnt find a 5-40, but figgered I'd get by in alum by gashing the screw and screwing it in a slightly large hole... i succeeded in twisting the head off the screw without much effort :(
channelocks to the rescue
nice paper on linux used in control theory http://mocha-java.uccs.edu/CSL/ece4530.pdf
(a pid loop controlling a magnet suspended above an electromagnet with laser/opto distance feedback )
i need to get some channel locks
this is neat, it assembles the parts while still on the machine: http://www.youtube.com/watch?v=Ep0crJUFEIc&NR=1
hmm fenn what about swarf left in the whistle
i think it's run with flood coolant normally
wouldn't that blast away any swarf? or is it not reliable enough?
dunno but the video had some stuck to the engraving, most companies barrel polish before assembly
I can see the advantages
so, whats up guys?
* quiteBIGeye arguing over pick and place problems with Vq^
well, its a hobby I guess
as good as any :)
yo alex dood
how is married life treating you?
robin_z: pretty good actually :)
and the robots behaving themselves?
okie, im off to the gym
see ya later
what does motion.in-position do?
this one has nice angular thread milling http://www.youtube.com/watch?v=f2UtfHRdi68&feature=related
is motion.motion-inpos true after a move or during a move?
BigJohnT: only when the machine is exactly at the commanded position
BigJohnT: usually that happens only while "resting" :)
that's what I kinda figured
is there a "at commanded speed" pin or signal?
BigJohnT: not that I know
but that's because emc doesn't command speed
it outputs positions..
maybe at commanded speed = pid-error close to 0
can you tell if your ramping up or down from the F speed?
BigJohnT: I think there are some TP pins for that..
mostly for debugging
what do you need it for?
LOL I knew you would ask that
I still think that torch height would be better controlled by EMC as a THC has to guess as to what is going on...
If EMC knows that it is ramping up and ramping down then no Z adjustment should happen
to the F speed
on the X and Y axis
why shouldn't it adjust height?
as the torch slows down for a corner the arc voltage will change
that should affect your kerf too..
the more constant speed the better the cut..
but you have to slow down for a corner to change directions
BigJohnT: depends how you CAM it..
you mean if I radius the corners?
say you want to cut a 90 degree turn right
ok a loop de loop
you can do a loop :)
I've seen some hsm mills do that..
to avoid making sudden direction changes..
but I still can't imagine height changes much with minor speed adjustments :)
so if I have an ADC to convert the arc voltage to a BCD on say 4 input pins how could I use that in hal to set my Z?
BigJohnT: did you work out all
that seems to be a big problem with "corner dive"
I might be going 450IPM on thin material...
anonimasu not yet
EMF from the arc voltage signal?
it should be clean it comes from a pair of pins on the controller board not from the torch itself
I would do the voltage divide in the plasma torch and output 0-10 VDC
I'm playing with a cute ultrasonic sensor
4-20ma output analog
for distance measuring?
are you inputing that to EMC?
anonimasu: that has problems with what you're cutting
heat affects ultrasonic sensors though
if the workpiece is cut in two, and one part falls away..
they have a fairly wide cone
laser might be better...
5cm sensing area...
laser is sensitive to dust and stuff..
ultrasonic is not
(not nearly as sensitive)
back to arc voltage... I guess
oh yeah.. these sensors are probably not hobby stuff either .. :)
my work will pay for it (I own the business)
If this can be worked out it will make EMC very attractive to plasma users
I can't seem to find any TP pins...
it looks like Dallur has the EMC side worked out well from his chart. He is using an up or down signal from the THC
so it seems like I could use a BCD input and compare it to a set point via the pyVCP and move up or down...
weighted sum looks like it kinda does what I need...
the output will be an integer, you may need a simple component to convert that to a float (or maybe not - dunno)
or I could add a float output to the wsum. hmmm
I think an interger would work
but a binary to integer would be better...
the wsum output will be an integer - an s32 I think
what you put into it is up to you, just set the weights correctly
how do you figure out the weights?
is there an example of weighted sum somewhere?
off to google
the weights are whatever the binary inputs represent
they default to powers of two, so it converts binary input bits into an integer output
the mazak toolchanger had 5 prox switches but the last one represented 13 instead of 16, so the weights were 1,2,4,8,13
ok, so you can set the weights to different values
and by powers of two it is just like a binary 1,2,4,8,16,32...
it seemed that the most common use would be to change a bunch of input bits into a single integer
that's what I want to do...
should be pretty easy then :)
so is wsum.0.bit.0.in the first number in the list [,size,...]?
yes ...bit.x.in corresponds to the weight.n parameter
the output is the total of input 1 * weight 1 + input 2 * weight 2 ...
ok, thanks that makes sense to me now.
* BigJohnT is going to play with wsum for a while
I did a loadrt weighted_sum wsum_sizes=8 and that works
when I add addf wsum.0 servo-thread I get an error
function 'wsum.0' not found
either use halcmd to see what the actual function name is, or check the manpage section FUNCTIONS
I used the hal configuration
it shows wsum.0.bit.0.in etc
Is there anyone who can give me a hand trying configure a 5i20 card
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-02-10.txt
ok I got it, it should have been addf process_wsums servo-thread
BigJohnT, the tab key is very useful in an interactive halcmd session (halcmd -kf)
does anyone know how to change m5i20.0.enc-01-index pin to true?? I can't seem to find it in the HAL
tzak, I believe that's an output (in HAL), you don't set it to anything
the index pin tells you if the encoder is on the index (I believe), if you're trying to set up home-to-index or soemthing like threading, you probably want index-enable instead
its funny because I have the same setup for X and Z and they work fine .. just the Yaxis is different
what is the problem you're having?
I cant get the encoder to count .. again X and z count no problem
SWPadnos: Thanks, I understand how wsum works now
tzak, does the encoder count then reset, or not count at all?
or toggle between two counts?
It seems as though the only difference between the XZ and the Y is the pin m5i20.0.enc-01-index is set to false the other two are true
if you spin the X or Z motor a little, index should go false
which it does.. and they both count on screen
the index pin is just the state of the index output from the encoder
I'm betting that there's a hardware or wiring problem
maybe a broken power wire to the Y encoder
ive switched the y axis cable and re-wired it to the Z channel and it works no problem
i do check the pins with a meter and sometime there is +5 and then sometime it changes??
I'm not sure where to go from here. I'm pretty sure that the problem is hardware
i am however taking the 5v from the mesa card to run the encoders is that a problem ?
you can check this by not running EMC, just use halrun -I to load realtime and get a halcmd shell, then load the mesa driver
I don't know - I guess it depends on the current draw and operating voltage of the encoder :)
what kind of servo drivers are you using?
oh. so the encoders only connect to the mesa
you could try connecting the Y encoder to enc-03 instead
you don't have to use 01
yeah that was my next option
it does sound like there's a hardware problem.
i hope its not the mesa
it's either the connection to encoder-01, the input channel (which could be configured wrong on the 7i33), the wiring between the 7i33 and the 5i20, or the FPGA on the 5i20 itself
is there a way to change the FPGA???
yes, replace the board ;)
thats not good
well i guess i can change it to 03 and try my luck that way
thanks for all the help SWPadnos i'll give everything you suggested a try
ok. good luck
SWPadnos: I got me one of these, and it works pretty good as a start/estop http://w2s.sti.com/stisf/ds/SFRBT/SFRBT2e.lasso?pcode=2583
did you get one or two? :)
SWPadnos: only had one, unfortunantly
at least your machine will be nice and safe :)
and it'll save a lot on wiring up the logic too
The only thing different is it has two inputs (so to speak), so if one is the BRB, you have to reset it, and the other one before resettting the whole thing.
do both inputs have to open and close to reset the relay?
well, if you want only one button, you should be able to just connect the inputs in parallel
to the single set of BRB contacts
But figure I can use a "reset" button, or make one an on/off switch.
Basically, it be a two step process to reset once estop has been hit. twist BRB, and hit reset button. Not a bad thing in itself.
well, there are other things that need to cause e-stop, like limit switches or charge pump output
those can be harder to release then re-contact
SWPadnos: At first, I couldn't figure out what the AUTO/MANUAL reset switch was for inside the relay. But I found if in AUTO mode and power is restored, it starts up the machine.
but only if both inputs are closed, right?
SWPadnos Not if it's in AUTO mode =)
uh - then I think it's defective
wait, I can't remember - took apart the test wiring already.
at no time should the output be on if both inputs aren't contacted
err - closed
(I'm saying that from a safety standpoint, not because I know anything about safety relays, because I don't)
if closed and power is restored, it'll start machine in set to AUTO RESET mode.
ok, that's expected
SWPadnos: Basically it works as it should, I tested various scenarios
ok, that's good :)
I've looked through the realtime components and kernel modules list but nothing I look at seems to convert an integer to a float. Clues?
GNieport_ is now known as GNieport
SWPadnos: The ONLY thing I found is that if you hit BRB really fast, it doesn't break the circuit. Thus latching BRB ftw!
BigJohnT, well, it's like this... :)
SWPadnos: I was using momentary buttons to test at the time.
excuse me guys, but does anyone have any 'online' reference material on this current 'c' programming. i have a couple of ancient books, but they don't really seem to help.
something like 'modern c for dumbasses'?
so is the wsum output a s32?
Roguish: I've been looking for that serious of books, haven;t found it yet =)
darn it there it is in my show window
it's an s32
glad i'm not alone.....
so a conv_s32_float should do the trick
hmmm, lets go see
Roguish: Might try ##c or http://www.iso-9899.info/wiki/Main_Page
thanks, i'll try those. i also found a few links in wikipedia that look ok.
cradek is the '5axis' guy, right?
it's confusing enough for a non-electrical engineer but when you change syntax in mid stream it makes you want to drink!
loadrt conv_s32_float uses an underscore
but the name is changed to conv-s32-float
so you have to use addf conv-s32-float.0 servo-thread
Roguish: The K&R C book is still the best one I know of, even though it's old. However, it only covers "the C standard", not POSIX, no GUI, no network, and so on.
once you've learned the "original" ISO C, you will feel comfortable reading the various online descriptions of new features in "C99"
* jepler wanders off
(the full "C99" standard can be read online, it's linke from the wikipedia article on ISO C)
net conv-s32-float.0.in wsum.0.sum gives me an error signal name must not be the same as a pin...
BigJohnT: the first argument to "net" is the name of the signal you want to connect between the two pins
dammit I forgot, thanks
something like "net mysignal conv-s32-float.0.in wsum.0.sum"
the guy who invents a computer that will do what we want instead of what we typed will make a fortune
I'll be the first to buy one
I hear that
hello mr Kasunich
is there some common way to automate toolbreak checking when using emc2?
Yea! it works thanks
if you have some kind of tool length checking device (a switch that the tool tip would activate) you could probably do it
stop spindle, move to switch location, poke at switch with tool, see if it activates
maybe do all that in a g-code subroutine
i was thinking about something like that
is it easy to have a O-code subroutine from a different file?
I don't know - I'm just learning g-code myself
my focus up to now has been inside the motion controller and HAL
having my machine actually running is forcing me to learn g-code ;-)
jmkasunich: You have no inclination to use a CAM program do you?
jmkasunich: I've noticed that, may I ask why?
I don't run windows, except for easycad, and I don't buy expensive software
i'm planning to do a higher order control to wrap around the emc2 backend so i can always generate partial programs to execute
something similar to MDI commands
jmkasunich: But if there was a gpl CAM pkg you might consider it?
hi all im back... locked up solid
I've been actually writing g-code and machining parts for about one month
so far I haven't attempted anything that needs real CAM
welcome back mr Messier
jmkasunich: Ah, gotcha.
[18:02:45] <Jymmmm> http://www.dealextreme.com/details.dx/sku.474
I write a program to write my gcode for me, who needs cam
If I want to compare two floats and see if the first is higher or lower than the second what would I use? I see comp but it only tells me if there is a match...
comp does higher/lower
archivist: thats the spirit :)
between the 2 floats
Vq^, gear cutting so its easier that way
i've been thinking of putting together my small code-generating functions into a cam-library
by using the hysteresis?
would I need one for higher and one for lower?
[18:07:28] <jmkasunich> http://www.linuxcnc.org/docview/html//man/man9/comp.9.html
that's the page I'm reading
or trying to LOL
"With zero hysteresis, the output is true when in1 > in0"
archivist: is your program to write g-code not a sort of CAM?
archivist: what language do you use for generation?
that man page could be better
* jmkasunich fixes it up
and I know Jymmmm knows php so no excuse
jmkasunich: thanks for fixing it up
archivist: PHP == Pretty Horny People ???
i used to generate code from python but i found that Haskell was a very good fit
Ive started a gerber to gcode
20 lines or so and i had my own DSL for generating gcodes :)
Domain Specific Language
i pasted my heidenhein to g-code converter here a while ago, it used my small g-code library
[18:14:36] <Vq^> http://arda.no-ip.org/h2iso/
[18:14:56] <Vq^> http://arda.no-ip.org/h2iso/NCode.hs
<- g-code representation
BigJohnT: is this better (pin descriptions):
comp.N.out bit out
Normal output. True when in1 > in0 (see parameter ’hyst’ for
comp.N.equal bit out
Match output. True when difference between in1 and in0 is less
yes thank you
will be in version 2.2.4
that is much easier understand for me, thanks
BigJohnT: did you check out the wiki page yet? could you look at it and maybe do some edits to it?
looked good to me, what did you want to edit?
if you saw anything that needed some tlc
no, it looked good to me, heck even I understood it LOL
heck, I even understand comp now!
now if I can make a green light come on in my pyvcp when the arc voltage is equal to the setpoint I'm going to take a nap... wish me luck
equal is a trick concept with something as noisy as arc voltage
equal to within one millionth of a volt?
equal to within my hyst
I suppose, I'm just simulating at the moment
Maybe someone has some ideas.... If you're to have emergency battery power, why go with a series of gel-cells as opposed to a marine battery?
it would be in whole volts anyway after it goes through wsum
cradek: my import program will be heavy modified
assume somewhat portable
Jymmmm: its not wise to connect gel cells in parallel
to allow 5 axis gcode generation
jmkasunich: That's how most big UPS's work.... parallel 12v@7AH gel cells.
are you sure they parallel them?
I thought they put them in series
jmkasunich: I'd have to open up a couple to confirm
jmkasunich: Why did you say parallel was bad idea?
sharing of load, and sharing of charging, can be tricky
when in series, all cells or batteries see exactly the same amp-hours
Jymmmm: i think the reason is that marine batteries go to hell, if your not charge/discharge cycling them, they dont like to have small limited draws on them
in parallel, the one with the lowest voltage can actually take charge from the one with the highest voltage
gezar: Ah, so you need a smart charger to use them properly
there is nothing to ensure they discharge or charge evenly
jmkasunich: ok, that makes sense. assumeing one is snafu'ed
after probably 10 5 minute power outages, when the 11th comes for 6 minutes your out of juice
micges: what does your program do?
murphy's law says it will be
Jymmmm: I think the way to pick batteries is as some questions:
1) what voltage do I need? do I have a choice or not?
if you have a choice, run as high a voltage as practical, that will reduce wire size and resistive losses
if you have no choice, then that decision is easy
once you know how many batteries in series (voltage), then you pick the size of the battery based on amp hours
Well, I have two cases (24 units) of 12V@7AH new batteries already, plus what I've already had in the past. 12VDC is what I need mostly, then an invertor for 110VAC occasionaly
Weee I got a green light, thanks for the help
I also have a solar panel I'd like to use as well as 12/110 to recharge the battery bank
jmkasunich: it's way cool, thanks
Jymmmm: then I guess you have no choice but to run them in parallel
what are you powering?
jmkasunich: Primarily comms (two-way radio), maybe a laptop on occasion.
radios are 12V input
Yeah, technically 13.8VDC, but close enough.
how much current do they need?
and how long do you need to run them?
* BigJohnT going to take a nap now
jmkasunich: let me looking at the current draw, just a sec..
jmkasunich: Ok, on radio A, 11.5A draw
On radio B, 2.1A draw
radio A - is that during transmit (maximum), or all the time?
That's TX, max draw
per the specs
how much do you expect to be transmitting?
if average draw is a lot lower, that would be good to know
If on battery, I'd like 2hours potential.
That's basically CYA scenario
so if you transmit continuously, it would be (11.5+2.1)*2 = 27.2 A-hr
= 4 batteries
Yeah, which will never be continous TX for 2hours
(gawd I hope not anyway =)
do you know the rx load for the radios?
On Radio A, Rx drain is < 1A http://www.universal-radio.com/catalog/fm_txvrs/4116spec.html
Radio B = http://www.universal-radio.com/catalog/ht/4211spec.html
so even if radio B is 2.1A all the time, if A is rx only, you have (2.1+1)*2hrs = 6.2A-hr = 1 battery
looks like radio B is less than 0.1A on rx
Yeah, B is a HT, A is a mobile rig.
thats under 1.1A for both radios in recieve mode
probably 5 hours or more on a single battery if you don't talk
to be honest, unless you absolutely need zero downtime, I'd hook up a single battery, with connections that let you quickly swap it out
run on one at a time, and swap when it gets low
carry as many or as few as you want to, since you are only using one at a time, you aren't constrained
the problem is that when you need it, you better make sure the batterys are charged before hand.
So, having them 100% charged all the time is important, and then add the ability to (solar) recharge them in emergency use.
My solar blanket (at full sunlight) raw outputs 24V@1, I still need to add a smart charger in there for the solar aspects. It PWM the gel cells to prevent crystalization
charging is another issue - as I said before, paralleling does NOT ensure uniform charging
in fact it almost guarantees non-uniform charging
if you have 24V, charge two in series
if the solar cells give you 1A, then 7-8 hours will charge two batteries in series from flat to full
The one charger I have, will let me output 12 or 24, but at only 6A max draw.
That charger/supply is intended for CCTV systems with a 12@12AH single battery.
Maybe I'll have to see if they offer other charge/supply that allow multiple batteries connected to it, so it'll even charge all of them.
jmkasunich: Yeah, I'm not sure the solar charger outputs 24V, that's just the RAW output from the solar panel itself.
The charger prevents reverse flow when there is no Sun
[19:26:34] <Jymmmm> http://www.morningstarcorp.com/support/Why-PWM/why-pwm-1.shtml
This is the smallest unit they have... http://www.morningstarcorp.com/products/SunGuard/index.shtml
the voltage will go down with current too
It's a bummer, as the chargers never expect any other source of power to recharge... Like on a boat, you would have the alternator, then solar as a backup.
Sure, I could hook up some relays to do what I want, but I'd rather have some intelligence in the device.
Plus I dont like the idea of a relay being constantly pulled in
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2008-02-10.txt
anyone know the dpkg-reconfigure xserver-xorg switches to accept defaults?
ok, xserver is going to make me grumpy
well, I just went through it twice, and it comes up with a blank screen
guess I'm going to have to use vi
monitor doesn't complain any more, but ubuntu doesn't display anything once it turns on X
hopefully I can log in remotely now
I guess my mga driver doesn't like my monitor, every resolution is out of range
* alex_joni wants one of those..
* quiteBIGeye want's one too!
I think the next evolution would be a rubberband clip..
I want one
gets stretched automatically
* archivist is in linux and cant watch
Well, shit...now there's a use for all those skate bearings I bought!!!
bottom side would somehow pick up the rubberbands and load them
I'm running linux (ubuntu/firefox) and I can watch it just fine
Vulcan 20mm rotary cannon, with rubber bands
skunkworks: i wouldn't waste much time on that :) build a real pneumatic gun instead...
I havnt managed to get any sense yet on video playing
several hundred rubber bands ;-)
archivist: this is flash, not video playing
archivist: anyways.. till then, here are some photos: http://disintegrator.co.uk/photos.htm
Hmmm, I wonder if there's some way to add a proximity sensor for tracking purposes to it
like a phalanx
[21:41:03] <Jymmmm> http://www.fas.org/man/dod-101/sys/ship/weaps/mk-15.htm
anyone ever skied Windham mountain NY??
new york usa...
never been to usa...
Vq^:.. your the heidenhein to emc guy ... right>>
to g-codes, thats correct
not recently :)
are you using the heidenhein converter???
no, not really, why are you asking=
as in a 5XX series controller
bcz it'd be easy to scab.... and i'd be ok with it.. as the mazak one too
i write some stuff in heidy.. email to a friend and have g-code backe in a few minites somme times..
I have a fan in one of my computers going out, and its mooing like a cow
does it give any milk??
im not going to look for an utter
gezar, light oil under the sticky label
shuting the machine down
sewing machine oil... works well
* archivist uses the nearest whatever it is 3 in 1 or clock oil
for about a month..
dmes1: were you interested to use the converter?
clock oil should last longer as its a non drying oil
i like all converved perverted converters... pls may I
[21:55:22] <Vq^> http://arda.no-ip.org/h2iso/
Vq^: perverted converter, i cudnät have said it better myself, "crap converter"
dmes1: be sure to mention limits, i'm very interested in improving it
have you got cycle 9xx implemented
no, whats that?
tilted plane management
sounds interesting, shouldn't be that hard to implement
have yo got all the manuals??
keeping a linear transformation matrix in the machine state should do the trick
nope, i don't have any manual, which really makes it a bit hard
* skunkworks has tried to fix a lot of fans.. It isn't worth it unless you can't get a replacement right away. Just a temp fix
dmes1: do you know where i could retrieve the language specs?
yes. in MY cd case
checking if it'll mount..
no all siemens 840D stuff wierd... i wasnt drinkin hard then...
sorry im not finding the cd right now but i will..
good night all
heidenhain manual available on website, by control series, includes programming and operator guides, discludes plc and most parameters, 'secret codes'
alpha1125 is now known as a-l-p-h-a
tomp: i would really prefer pure language specs
all better now
I just hope the pb blaster stuff i used dosnt go bad, ill probably have to buy a small can of 3in1 oil
lewin1 is now known as lewing
3 in 1 may get sticky too
yes 3 in 1 is terrible as an oil
what's pure language specs???
* archivist_emc finishes another axis and plays daisy
sounds better as one adds parts
who was mentioning that 1.5L engine that could run on bearing grease and sawdust???
works until you get it off the used car lot anyway
make 400hp or so??
archivist_emc: need a video...
how many axiss are you up to?
www.archivist.info/cnc bottom link has some of the first gears cut
skunkworks: 3 fitted and working
how can i kill a window??
handball... really... like the HURCO's had??
I pobably should counterbalance the head before adding the last stepper
needed a knack.. to work it out... was 3 axis trackball... when you held the correct button it worked ok... if you didn't it usuasly crashed a tool..
i hated that machine
and i played QUIX as a kid so if you know the game ...
just got a new -old jethro tull album on-board.... ; ) things look up///
* skunkworks loves jethro tull
locomotive breath is one of my favorite
archivist tha'ts a nice idea I never considered, a webcam
Jymmmm, it will be for machine centering when I get around to it
archivist machine centering?
archivist like zeroing out the tooling to the material and such?
the cutter has to be on the rotary table centerline +- bugger all
and each is different
Hey, that gives me an idea... I've been using a flashlight to zero out the tool to the material, Maybe I could add in a camera and automate the process.
my problem is I dont know where 0 is on any tool, without a lot of calcs and measuring
archivist why's that?
monitor problems are gonna make me grumpy
Jymmmm, some are very old cutters not to spec
eric_U: you tried vesa driver yet?
that was my next trick
eric_U: try a liveCD and see what it finds
knoppix is a good one for that
everything ids just fine
the way I read the log, it starts just fine
eric_U: DL and burn knoppix, it's worht having one around handy for troubleshooting
anyone know how to restart X?
what if it's running?
sudo reboot worked:)
of course, I still have to walk down to the basement to see if there are graphics
how about sudo /etc/init.d/gdm restart
or just sudo kill X, which gdm will then restart for you
says there is no mode of the name that I put in my xorg.conf file
is it lying?
dpkg-reconfigure finds the mode
oh, that's a good sign
would be if it worked
is the mode line in the correct screen / driver section?
looks like I'm gonna have to find a vga cable and go back to nvidia
having trouble with an LCD?
oh. I don't know the incantations for the open source driver
but I think it's the matrox g550 I have it on
for the nvidia proprietary one it's something like Option "connectedmonitor" "DFP"
or if you have two, it's "DFP,DFP"
what's the DFP mean
Digital Flat Panel
im not up to following the dynamic machine control email set, any of you able to figure that out?
well, he wants to - um, well, move stuff, and uh, measure stuff
he needs a translator
do you get the impression that he is wanting fully autonamous machining?
i have wireless can runs thru vlc with its own web server...
he does need a translator - English isn't his first language
he's actually doing quite well I think
I cant quite figure it out, I dont reply to 99% of the emails but I do read them all
I understand what he is wanting, its just a massive rework I guess
he does want some measure of automatic machining, but I think it's more about measurement and confirmation of the machined part, not so much automatically going from one part to the next
dmes1, a user who has posted to the user list
im not bashing or making fun of here at all, its just starting to seem really complicated to me
who where what???
it is complicated, and I'm not sure I know quite what he wants
he may also not quite know exactly what he wants
okay, so im not alone
oh no, not in the slightest
and im thinking a simple cad/cam solution up in my head
not a full blown system but a single path generator
how more could it be simplified other than the EMC2 manuals...
I'm reasonably sure that he wants the machine controller (EMC2) to machine a part, take some measurements (maybe of the stock, aybe of the part), and then feed that back to the CAM app so it can generate ne wtoolpaths for the finish pass(es)
APT.... if its mathematical... its done.
too machine tool intensive...
well, to some, machine tools are supposed to be extensions of cad systems,
and cam and verification... every time you post and chk its different... even a little
I dont know anymore, to me machine tools are just that, and they are all full of limitations
gezer: i still play the game and what we would have don in 3 hrs now takes 12...; (
RULED from across the pond
im not putting anything down im just saying, I dont expect a whole lot out of any type of machine
ie, can have a great machine that does perfectly, then someone asks you to machine a part thats biger then the work envelope,
i can do almost anything on most any machines....or control
post processor SHOULD handle that
have you seen the linux apt 360??
no I havent
i pulled the b-777 upper and lower torque links in to have a visual for the first time...
chk out #cam channel
its dedicated to that.... or VAPT visual APT
BTW good tool path generation routines can track remaining stock... for re-machining purposes
oh yeah thats for sure
what have you used gezar??
all im thinking about is a single line path, say either ---- or a | type move, where the user inputs specifics to the path, yeah, you have to combine all the lines for a part sure, but
bobcad a little, the rest its been by hand
and give 1 program to finished part
text... = APT360 then g-code thru the post processor
so it works with a vga cable, imagine that