Follow up question from yesterday. I am receiving this errror Joint '0' on limit switch error. Hal shows this pin as false and when i toggle the limit switch it doesn't change. I have it connected to a hobby cnc board pin #10. Has anyone experienced this error ?
seems obvious that emc isn't talking to the port
I agree I believe that the port is bad. Any idea what voltage you should see on the port #10 pin to ground ?
is there an external pull up?
IIRC, a properly initialized parport will be at logic high
who knows what voltage that is nowadays
Yes a 10K resistor is on pins 10, 12, 13, and 15.
what voltage are you measuring
Logic high is + voltage I measured 24mv.
sounds like it's being pulled low
When I activate the switch it goes to 0.
24mv is logic low
it should be 3.3V or 5V
I can't get the port to output anything more than 24mv. I have tried ECP and EEP. Looks like i need a new port.
Thanks for the help :)
does lspci -v|more mention any parports?
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-03-05.txt
I didn't check but the jog functions well and the steppers move.
nvr mind then
Good evening, gentlemen.
I have a question regarding break out boards, drivers, etc.
fenn knows everything
I am just starting down the path of building a machine. It will be a Joe2000chvy 4x4 Hybrid. It is popular on cnczone.com.
I'm a Linux guy so I want to drive it with EMC2.
I have not yet selected the steppers/servers/electronics.
I'm guessing servos (600~1000 oz in) would be appropriate for such a gantry.
I'd like to run fairly new computer hardware but I could probably scare up a board with a parallel port if I had to.
emc has a worksheet that calculates the cutting force and the required stepper size. Its on the wici site.
What should I be looking at for interface/driver options. Is there something common that everyone uses?
same as the mach guys, pretty much
Zylotex seems to be the most popular with them. Same here?
for steppers anyway
too hard to know the answer to that
connection reset by lack of spring clip in ethernet cable
what's this chevy hybrid you mentioned? wood router?
Yes. More or less.
fenn, I thought you knew everything
judedude: so, as i was saying, pin 10 is in input mode
Aluminum gantry (80/20), some wood components, v-bearing.
so turning it on probably wouldn't do much
he should be pulling it high though
sounds like it is pulled low, or his external pullup is blinked
Its a hobby cnc board and it checks out good.
tombrown_: you gotta link to this machine? my search-foo is weak
I dunno about that, it's not pulling the port high
does it pull the pin high when disconnected?
which btw, I have no idea if that's a good idea :o
No the pin is low. I checked the cable and it passes the signal low. At the back of the computer the signal is still low. That is with the invert check or not.
[03:01:05] <tombrown_> http://www.joescnc.com/
like i said, pin 10 is input
right, and thus emc is not controlling that pin
you can see which pins are input and which are output here: http://linuxcnc.org/docs/html/hal_drivers.html#r1_1_1
hm that pic is sorta screwy (half transparent)
OK I guess I am not so good electically. If 10 is input how can the signal be high.
you are supposed to be pulling it high externally
is there a jumper
if you wire it to logic high the pin should go to logic high (but use a resistor, not just a wire, so you dont break anything)
No. According to HobbyCNC the pins are a pass through pulled up with a 10K resistor
which board is it?
Hobby CNC pro board
heh i wonder who thought up that name
you know these guys?
no, just remarking at the irony of "hobby pro"
"patented chip protection circuitry" -- I smell something
judedude, did you solder it yourself?
no online manuals == fail
Yes. I pretty good a soldering been trained in wepon spec soldering.
I see jumpers all over that board
one picture shows a label that says "no pul.... aieeeeeee, thump"
They did provide a schematic for home switch hook up its the same as on the EMC site.
what is the function of J1?
yeah how hard would it be to get a pic of the whole board at once
Pins 1 and a few other do not have pull ups
he's very protective, there are missives of hate all over the internet
i know he wasn't much help on this so thats hhow i found you guys.
i wonder why there's a bare pad marked +5V just sitting in the middle of the board with a circle around it
can you disconnect and measure the resistance from pin 10 to 5v?
Dave claims its for the pins that have no pull up. Should i be using this?
eric_unterhausen: 1 14 16 17 are always configured as output, so it wouldnt make sense for them to have pullups
Yes I can measure it may take a few minutes.
fenn, most parports were open collector
meaning there is a pullup somewhere
dunno about nowadays
well, "nowadays" could be anything
the great thing about standards is you get to ignore them
there is no standard
measures .427K Ohms
I'm always really unhappy with products that don't have online manuals
.427k? like 427 ohms?
that's a strange number
I thought it would be simple.
it is simple
that was disconnected from the parport?
and the limit switch?
because 427 ohms != 10k ohms
No I didn't disconnect it.
you have to disconnect it
OK be right back
otherwise the 24mV on the port (or whatever voltage) will screw with the measurement
is that Dave Rigotti?
that sounds familiar
there are so many questions that could be answered with a 3 minute superficial reading of a manual
or even a schematic *gasp*
I understand that
i dont get why people are so reluctant to release a schematic
it's not like i can't just desolder everything and scan the board in
I know, the Chinese would reverse engineer this for fun, twice
a la tormach
yeah, and that actually had engineering involved in the original design
unlike this board
actually, this board comes without any components on it
it's a kit, no desoldering needed
you could scan it and generate production files and have a board house mail a quote to hobbycnc for giggles
its only 9.8k Ohms
so that's right
where are you getting the 5 v from?
is there more than one 5v supply?
do i need to power the switch with the +5v
I would think that would have some amount of isolation
have you measured the voltage across the switch?
24mv when connected
that's the switch?
I thought that was at the port
yes. Its the same. I checked it to ground.
there should be a pullup voltage across the switch when it is open
24mv is not it
what is the function of j1, j2, j3, j4, etc?
OK slow down a bit. The switch is only wired from pin 10 to ground. I have on external power.
pin 10 of the hobbycnc board, or pin 10 of the parport?
The jumpers are for stepping 1, 1/2, 1/4 etc and idle current reduction.
is the switch connected to pin 10 of the hobby cnc board?
Pin 10 of the hobby board whitch should be pin 10 of the parport.
I think I was operating under the mistaken impression that the "Pro" board would have buffering
There doesn't appear to be any isolation. Continuuity check shows it goes right thru.
what do you think Q4 does?
oh i bet that's the l297
it's a secreet
carefully sanded so you can't read the part number
Something to do with syncronous idle current stuff. Dave was offended when I asked about this.
are you sure the 5v regulator is putting out 5v?
I guess it is though
* fenn goes to bed
good luck divining the answer :)
yep. I checked it at 5.17V. Is this suppose to drive the 5V on the limit switch.
is the ground at the board at the same potential as the ground at the computer?
I'll need to verify that. Youre talking about the parpot ground?
I would power up the board disconnected from the computer and see what the voltage from pin 10 to ground is
those two grounds need to be the same
The voltage from 10 to ground is 5.14V
so it is being pulled low by the parport
OK now I am lost.
me too, but that's because I'm somewhere on the internet
Why doesn't emc like this ?
was the 5.14v from 10 to ground with it connected to the computer?
No I disconnected the limit switch and the parport.
connect one back up
your choice which
connecting the parport 4.2mv and if I just connect the limit switch N.C. I get 0.
I guess it was worth a try, but it looks like your conclusion an hour ago may have been correct
OK I bought a parport and I'll install it tomorrow. Really appreciate your help. Let fenn know as well.
let us know
You bet. Till then.
cradek: it seems like that spindle-on bug is still there in some capacity
oh? what details can you give?
i haven't put my finger on exactly what sequence of events causes it yet
i will write it down
the spindle has to be on, like in low speed
I have not seen it and I've used trunk a LOT
then execute a g43 command, and it will speed up a lot
sure, the radius changes when you issue g43
if it's in CSS mode, it will change speed
yeah for css i guess it would make sense
but my two tools i was using today were very close to the same centerpoint
definitely let us know if you think there is a problem and you can demonstrate it somehow
sim/lathe is good for that because it has a spindle rpm readout
yeah i will try to take better notes
i was just in a rush to get parts made today
I've sure been in that position
got 42 made, but 8 left for tomorrow
it sure doesn't help one make a good bug report
there also seemed to be some weirdness with touch-off in X to tool table, but that might have been me
several times i could have sworn i only touched off to tool table, like for correcting an oversized part, and it wound up wtih a g92 offset
the only way i know how to fix that is close down emc, edit the .var file, and restart it
touch off definitely doesn't make g92 offsets
probably an easier way but i haven't found it yet
okay maybe it is just me going too fast through the motions
you can clear g92 offsets with mdi g92.1 or on the machine menu
is there a way you can view what the offset is?
i.e. to check if it's non-zero
hmm, all I can think of is to do g92.2, g92.3, see if the dro changes
but props again on the touch-off to tool table, it makes minor adjustments for wear so much easier
if a part comes out 0.05 over, simple to touch it up for the next one
i love it
cool, glad you like it, me too
I just used it tonight for adjusting a boring bar
bore came out perfectly on size and with some nice chatter marks that only happened on the last pass :-/
i get that sometimes
if i try to baby a finishing pass
yeah that's a sure failure
gotta make a cut i guess, i'm learning
they have to all be the same, you have to prepare and then cross your fingers on the last one
(with carbide anyway)
with sharp HSS you can baby it
i learned that with my boring head on the mill
take a couple, measure, extrapolate, set it, and let it rip
sharp hss with a big rake angle i hear works well, but i never have bothered trying it
I use it a lot actually
I have a little boring bar I made from 1/8 square hss
for small holes in aluminum
i was impressed with my lathe today... it was turning out parts repeatable to 0.01mm for hours
i could watch the OD's slowly creep up over 5-8 parts or so as the tool edge wore in a bit
then a minor adjustment and right back on target
very cool, i have never experienced that before
what is different?
speeds, feeds, better coding
slowed the rapids down, i think they may have been skipping a step here and there before
ah, not good
i had them up at 2800mm/min, too fast really
slowed them down to 1800 and they seem to be more repeatable now
also learned to program around backlash in my screws, it seems to help too
do you know if anybody has ever considered adding the basic roughing cycles?
I was just thinking about it today
it would be somewhat challenging to do
i would be willing to help, but i'm not sure i would want to lead something like that yet
don't know my way around the source yet
I think it would be about as hard as the new cutter compensation I did recently
it seems like it should be fairly straight forward... just a projection into a revolved plane really
as far as implementing it, maybe not so easy though
that took several weeks (?) of most of my spare time
yeah coding eats up time
yeah it's not conceptually hard, I agree
it's the details
i bet it is
and the weird profiles
like an undercut or something? i wonder how those are handled by a commercial controller
for example making a chess pawn
you can quickly get into the realm I consider CAM
is the source only patchable in c?
our gcode interpreter is written in C/C++ and that's where this would go
maybe i can take a crack at it and see what breaks
if it fails miserably then at least i might learn something
you could also help by collecting and distilling the data required to write a real meaningful spec saying what it should do and how it should work at the gcode level
oh great, i have a bunch of reference i was reading over already
the intersection of "expert concerning how lots of commercial controls work" and "expert at improving the emc2 interpreter" is exactly zero
so someone who can do the go-between would be very useful
i can definitely do that
hm, I should get to bed
glad to hear of your successes
i'm thoroughly enjoying this software
so.. is this the right place to ask questions about cnc control in general?
izua: no problem in doing so :)
but usually, around this time of day there is little going on in here
heh, guess so. around here is 10 am
anyway, i'm trying to understand the basic principle behind controlled acc/dec - so successive smaller lines are 'blended' together
it's relatively simple on a single axis. consider that axis is part of a two axis cnc (foam cutter or so). i know my angle, max speed and max accel, so I can compute my axis speed (vector) out of the max speed (vector sum). I also know my deplacement (or time that speed needs to be applied - synced with the other axis).
IN theory all is wonderful for a single axis
But for two of 'em, do you simply double everything?
why couldn't you still do the vector sum?
no, i mean, i don't see how blending will be achieved there
i do the vector sum there (well it's technically vector.. decomposition or something) so I can obtain individual axis speeds, which summed up give my max speed
you can see how it's done in EMC by downloading the source :)
..that's the way it should work, right?
i've tried actually. i even burned the whole distro on a cd (man it took me a long time to find a blank cd :P) but the truth is, i'm not that good with software involving more than two-three files..
izua: think about it a bit easier
you have 2 moves
the first move is decelerating to a halt
the next move is accelerating from a stop to move vel
now if you combine the two (at the moment when you start decelerating move1, you start accelerating move2)
and you add the two outputs, you get blending
yes.. but how exactly do I blend the two?
I need a parameter, at least, that says how much can be blended together.. so in some instances (like very sharp angles) it would deccelerate rather than blend
my long-term purpose here would be to make a stand-alone board that does what emc does. released under o/s of course.
the electronics and implementing an algo wouldn't be such a hassle i guess, but I still need to understand the principle before hand, hehe
if that's what you are doing, you should download the EMC source and look through the kinematics section
sure. i'll check it
as another side question, how does error correction fits with acc/dec
izua, its already in the trajectory planner
he wants to port emc to an 8 pin Pic or something like that
and pigs might fly
* archivist has seen a pig fly
it's been ported to a z80
it would probably be a avr with a fpga around, or an avr32 (which would be waaay easier)
but I intend on rewriting the whole thing, which would prove I understood the concept :)
you should port it to an ARM
and run linux
[08:49:22] <archivist> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Trapezoidal_Velocity_Profile_Trajectory_Planner
and probably don't require much of a port
yes, that would simply be a port
yeah! that's exactly what I was saying earlier
i understand it perfectly for one axis
i even made a c# sim somewhere (yeah, bash me, .net)
[08:50:17] <archivist> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Simple_Tp_Notes
but do you simply duplicate everything for 2 axis?
so much better to read the code
ello people :)
in the ubuntu emc cd ... why cant i update it ? :)
MrSunshine: can you be a bit more specific ?
update the cd, or update the resultant install
man this is huge..
the installed version
izua: basicly the trajectory planner doesn't work with individual joints
only very late in the process
ive installed it and its 357 updates available :)
until then it works with vectors
which contain all moving joints
MrSunshine_: so what's the error?
alex_joni: the emc2 page told me not to update :)
cause then emc stops working
it told you not to update to another Ubuntu version
err.. let me be more specific
it told you not to _upgrade_ to another Ubuntu version
so i can do the updates if i want to then ? :)
without emc stopping to work :9
k il trust you .. :)
the only possible problem I know is if you installed ubuntu first
then installed emc2 (not by using the LiveCD)
then if you install regular ubuntu kernel updates it will redo the order of kernels in grub
and you need to be carefull to chose the right one (2.6.24-16-rtai)
but if you started from the LiveCD there should be nothing that interferes with emc2
* MrSunshine_ got his first cnc machine up and running today .. tho still missing 4 axial ballbearings
small cute pcb milling machine :)
what work area?
about a eurocard :P
dunno what they are, 10x16cm ?
it became ALOT smaller then i thought it would be when i started building :)
that's pretty cool
but its made for pcb milling so :)
dont need much bigger for hobby stuff
also good for robotics and stuff..
no need for a lot of big parts
got a real mill in the cellar im going to cnc convert :)
got started on it but then the money ran out :)
tho its very slow, bad motors and controllers :)
only 48 steps/rev and no microstepping :P
i realy should get stuff to make x, y, z, max switches tho
settng the config to "both limit X" that will be just one input and the program figures it out by the direction its going right ?
and the home thingie then it goes to the home position == when the switch turns on ? :)
adding a couple of microswitches shouldnt be to hard, have to change the cable from the cnc to the pc tho :)
just use the hardware
add some good motors and some a3986 drivers and you should be fine
heck, even 3977 should work if they are less powerfull motors
real hardware may well need beefier drivers
3977 blows up over 2 amps
Im running 2.5 to 3 amps at 30v per stepper on a small mill
and speaking of drivers, has anyone made an o/s gecko yet? what magic do geckos have in them, anyway?
[10:05:12] <archivist> http://www.arceurotrade.co.uk/Catalogue/Stepper-Motors/Stepper-Drivers
have you opened them up yet? :)
I looked in one
izua: there was a post of an gecko clone somewhere on the web
izua, I know from a previous day job that I may as well get pre-built drivers, too much work and not enough money saved when you do it yourself
archivist: that is probably the sane way to think about it
but otoh we only get to hear from the people who think otherwise in the beginning ;)
and I want it to work
I have no problems spending 4-5k EUR on a servo drive
even if I could probably build one in 3-4 years maybe?
time to implement a machine is a factor
anyone got any ideas how i can increase the torque of the motors? anything i should think about while doing the settings in the wizard?
step times etc
i could increase them to 7V but then they will get mighty hot fast :P
isnt it high voltage low apare when moving and high ampare low voltage while holding ?
i was thinking mostley the step hold etc times
if they have something to do with it ?
MrSunshine_: not really
different times need to be correct so you don't lose steps
but they can't add more torque
for that you probably need smarter drivers, or better motors
(smarter drivers = current limited drivers with high voltage input)
usually it's best to drive a stepper around 20-30 times the rated voltage
acceleraton then ... what is that? :)
I run mine (small steppers) at around 40V
acceleration is how fast the axis spins up to speed
if you get lost steps then you can set lower accel
but that will cause blending to have a much bigger effect
e.g. higher path deviation to keep the speed up
just turn blending off... that is always an option
OoBIGeye: sure it is, but then you get lousy performance :)
as it needs to stop at each end of every move
well, if accuracy is nothing to do with performance :)
when blend everything to the max...
OoBIGeye: ideally you have infinite accel
so you don't need to deviate from the path while blending
alex_joni: ye these drivers are constant current/voltage
but I haven't seen machines that can do more than 3-4g
5v with constant what the motor sucks in it
so its qute bad stuff :)
5v seems awfully low
what current limit do you have on these motors?
3-4g is alot!
OoBIGeye: I _know_
especially on a 100kg gantry :D
it's really scary
the new linear motors are rated at 80g when they start using them it'll be fun...
I am having a VERY frustrating time here. My calculation results look correct as long as I bypass the kinematics equation in my kinematics file. Using gdb the values are what I want. When I connect the kinematics calculations the values are not the same. I cannot see why this is different. I have had the equations working for 3 or 4 months. I cannot get them to work in the control.
good morning jepler
are you ready for another fun day?
oh they're keeping me busy at work here
THAT is a good thing
yeah it is
and good to hear
the part where my manager's manager gives me new priorities every 2 days is starting to chafe a bit
I want this thing done NOW and then I want that done before you do the first thing
the consolation is - he is getting the same instructions from either his boss or the customers
heh after a while I bite back in that situation, job swapping costs time!
stustev: so you have two PCs, one where you used --enable-simulator so you can run gdb and one where you actually want to run hardware
stustev: and the two of them behave differently?
I hear people say I am going in business for myself because I am tired of having a boss - heh - now every customer is your boss :)
free software projects are the same, except your customers don't pay you either :-P
jepler: how many :) :) :) are payment enough?
I am using only one PC - everything is in simulator
I recompile and restart to change the configuration of the kinematics
when I get this figured out I will then go to the machine with it
can you be specific about what you change?
you modify the source code? a hal file?
I modify the source code of my kinematics
I use pos->a = joints or pos->a = (a number)
when I use the former (bypassing the kinematics calculation) the variable I calculate and view using gdb are correct
oh -- so in your first kind of testing you are using a specific hardcoded number instead of the actual joint position .. and in the other kind you are using a changing number that comes from emc
when I try to use pos->a = (my calculated variable) the viewed calculations are incorrect
pos->a = joints and joints = pos->a bypass my calculation. In this configuration I view my calculated variables and they look correct
when I try to use my calculated variable ie pos->a = (calculated position) and joints = (calculated joint position) my calculations are not correct
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-03-05.txt
thanks for the insight
it's the best insight I have at the moment :(
WEELLLLL - I already had that much figured out :)
can you easily let me take a look at the source? not that I can take much time to look at it right now
done - the source is www.mpm1.com:8080/cinci/5axhydrotelkins_work.c
you can see where is comment/uncomment things
can I just change axis_9axis to use this kins file, or do I need more "set up" than that?
you would need some hal file changes also
the ini hal var tbl file are there now
the 5axhydrotelgui.py is there now - I have it in src/hal/user_comps/vismach - I have the config files in config/vismach
I had hacked it out already
you are too fast :) (or maybe I am just too slow) ;)
BJT-Work_ is now known as BJT-Work
serious WTF - has steel gone up 2000% since I last ordered any?
quick poll - how much should a 1/2" round 36" long rod cost?
stustev: here's all I've learned so far: I print facomp on the terminal and look at "A" in the DRO, and they match..
check B also - and don't even try to tell me it just works I will scream - I may be bald the next time you see me
welll I wouldn't say it works -- when I was jogging only B, A was changing as well..
A joint should change with a B only change if the error variables have anything but zero in them
B joint should change with A joint motion
the axis positions shouldn't change though
I was looking at axis positions, not joints
it started at B=360 and said it was on a limit switch (which I'm sure it was) -- is this what it does for you too?
that is the same as what I see
you can set the axis limit at 10000 and you will still get the axis limit switch warning
$ double fbcomp = (r2d(atan2(fva22,fva17))) + 180;
well I can tell you one thing -- this is the line that gives you 360 when you want 0
r2d(atan2) returns a value from -180 to 180 .. when you add 180, it gives you a value from 0 to 360
there's a discontinuity between 0 and 360, and you end up on the wrong side of it
when I have the calculation bypassed that line gives me zero when I view it with gdb and 360 when I don't have the calculation bypassed
isn't it -360?
(jogging 4 to be just a tiny bit positive before switching to coord mode gives you B=small, not B=360)
other than that I don't see obvious problems -- I think you only have subtle problems
yey, priliminary table done, Z axis in place :)
sorry I'm not able to immediately solve it for you
here too - with A also - if my logic was complete the A joint would move when I move B and vice versa
[14:25:56] <skunkworks_> http://cnczone.com/forums/showthread.php?t=74908
thanks though - your insight has been a tremendous relief - now on the the SUBTLE things :)
I just may have to take a trip to Lincoln to find the subtle things :)
I'm not sure what good that'd do you
bbl, coffee time here at the office
skunkworks, I think that user on cnczone should look at all the youtube wire bending vids
a trip to lincoln could possible clear some of the fog behind my eyes
does atan2(-y,-x) give you the same as your atan2(y,x)+180 but without the range problem?
"The new generation PLC-UART motherboard also support DIP by optional pins that can be used to directly plug to user's board without doing screwing"
that sounds like a bad translationl....
cradek: good morning - I will try that
I removed the '+ 180' from the line - the cone is upside down now - I will try the (-y,-x) now :)
cradek: how do you do that? this is working now - I see the numbers I expected. Now the subtle things will come into play. My results are not perfect but now at least EMC2 matches gdb. bbiam
jepler spotted the problem - I just figured out the easiest solution
you can be inclusive (more than one person)
check out this machine: http://blog.makezine.com/archive/2009/03/cnc_car_form_prototyping.html?CMP=OTC-0D6B48984890
i dont think it's emc2, unfortunately :-/
seb_kuzminsky: that sure is a big-ass machine
hiya guys :-)
for machines that size, the control software cost is negligable
a few k$ sure don't count ;)
seeing k$ reminded me..
why do people call microsoft M$ and not u$ :P
seb_kuzminsky: although I must say .. the CAM they use is pretty dumb :)
you think they could optimize their toolpaths a bit?
it's not really 5-axis toolpaths
it is 5-axis when they apply that glue/compound
you see how the tool stays perpendicular on the surface
I just got to the part where they mill after applying the gue, that looks more like it :)
* JymmmEMC shoves alex_joni into the machine and turns him into a YUGO!
JymmmEMC: unlike you I am actually familiar with the Yugo ;)
even drove one once
alex_joni: I thought you actually modified one (or two) ?
and survived !
er.. nope ;)
hmm.. didn't know they used to sell them to US too
i burned out the clutch while offroading in a yugo once in tajikistan
Now I guess it's the "SmartCar"
Diknky lil eurotrash thing
thats a Merc
you mean the Smart
archivist: sure is priced like one
[17:21:02] <JymmmEMC> http://www.smartusa.com/
so embarrassed they renamed it
although I must say I like how the Smart Roadster Brabus looks
don't think that one is available on smartusa.com
alex_joni: DUDE! If I walked to fast and bumped into one, it be totaled!
video removed due to copyright
or this: http://stb.msn.com/i/25/A9BCAC5FB576EA6C534C887CC97F6E.jpg
JymmmEMC: found another one: http://www.youtube.com/watch?v=NKzUcL4g60E
the back end looks fugly
there are 2 versions
the roadster and the roadster coupe
[17:30:38] <alex_joni> http://www.carenthusiast.com/smart/smart_roadsterbrabus2004_56.jpg
still looks fugly =)
the CRX has returned!
and smaller than ever, apparently: http://www.sparkplug.org.uk/gallery.php?gallery=images/smart/roadster/gallery/&index=3
gaaah, that thing is hideous.
the car is ok though.
this looks much better http://www.sparkplug.org.uk/gallery.php?gallery=images/smart/roadster/gallery/&index=5
strange looking car
that's uglier imo :)
Good evening.one council.what is the more stable version of debian linux and rtai ? thanks
i have problem with 5.0 lenny and rtai 3.7.1
motioncontrol: what kind of problems?
good evening alex.i have install this configuration on 2 maschine for prove
the sistem when you open the windown or simple operation system , some times stop the mouse and the sistem is not operate.the keiboard is kill and i reset the pc for init the syetem.
sounds like a problem with the kernel
how did you install emc2?
i don't have install emc2 , because the sistem is no stable
i have disableb the pc speacker
oh, you mean that is with the default lenny kernel?
no i have install 18.104.22.168 kernel
well.. if that doesn't work, why not use the default kernel?
anyways, probably #debian is more appropriate than here..
which patch the defaul kernel?
surely they will say "use our kernel"
i don't have the source
just like we say use the packages we built
the rtai3.7.i / i386 not have the 22.214.171.124 patch
your package run on ubuntu or debian?
I see 4 patches for 2.6.26 in RTAI magma
rtai magma i don't see on rtai site ...? what dowload it?
[19:54:16] <alex_joni> http://cvs.gna.org/cvsweb/magma/base/arch/x86/patches/?cvsroot=rtai
one question on ubuntu i can compile emc2 ? i don't have all library or package the debian
on ubuntu you don't have to compile, you can simply install the packages
but only on ubuntu 6.06 or 8.04
if i use the live cd ubuntu i can't compile emc2 .is possible install all package for compile emc2 ?
I'm about to order electronics. How is EMC2 with the HobbyCNC Pro controller?
[19:57:45] <BJT-Work> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Hardware_Problems
* BJT-Work goes to the screw store
ok alex i want compile emc2 .what is the linux version for compile and use the emc2 stable ?
any version should be ok
_but_ if you go with ubuntu, which we support, then you know at least the kernel & rtai packages are ok
so you don't waste time building those until they work
Thak you, BJT-Work. I appreciate it.
alex can i compile on debian maschine and use on ubuntu maschine work ?
BJT-Work: wow, what a writeup
tombrown_: if I recall, kim was not the only one to have trouble with the hobbycnc board.
the xylotex is twice the price but has no known bugs and runs bipolar which is also superior
something on the mailing list iirc
Thank you, Cradek.
can i compile on debian maschine and use on ubuntu maschine work ?
I see Zylotex has a bundle with a 4 axis drive box and 4 x 425 oz in steppers. http://www.xylotex.com/Econo4Ax425.htm
cradek when you compile use debian or ununtu ?
It looks like what I need.
it doesn't say what mounting the motors are
Others have told me to go with the HobbyCNC 425 oz in motors, regardless of choice of controller.
motioncontrol: I use only ubuntu
tombrown_: what size of machine?
It's a JoesCNC 4x4 Hybrid. (48" x 48" x 5")
I'd like to have 200 ipm rapids, if possible.
what kind of leadscrews?
1/2-10 5 start
yeah that's a lot of travel - fast rapids would be nice
ahh that means two turns per inch right?
ah, I thought you were in dreamland, but maybe not
Another builder has achieved 300 ipm with 570 oz in steppers.
you definitely want the big motors because those screws will back-drive the motors with any cutting force
i want use one maschine for compile and one maschine for use emc in stable version.please one council for me.
motioncontrol: it's easy - use the same OS on both
tombrown_: maybe you should try to get the same steppers and drivers he has then
Hey motioncontrol, if you have to ask Debian versus Ubuntu, go Ubuntu. There's nothing wrong with Debian but it's aimed at an audience who likes to tinker with details more.
the xylotex is way too wimpy for this speed I bet. it is only 24 volt.
i want use ubuntu , but on ubuntu i can't compile emc2
That means 4 x G201 controllers, 4 x 570 oz in steppers, and a 600 Watt x 70V PSU.
my advice is keep your $500 until you've figured out the best way to get what you need
Thank you, Cradek.
you will want geckos for that speed
motioncontrol: then you are doing something wrong. instructions here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2
tombrown_: (or servos...)
but that's another world of complexity
I think motioncontrol is trying to install packages from a LiveCD boot. On Kubuntu 8.10, there is no direct Apt support. It's just meant to install and kick around in a bit if you want to play or fix a system. I assume Ubuntu is the same.
it is possible to use apt on ubuntu while booted from the live cd
Fair enough. I will not comment further on Ubuntu since I'm a KDE snob.
tombrown_: if you're in that deep already, maybe talk to pico systems and see if he can set you up with a turnkey servo setup
do you already have the leadscrews?
following the instructions and using ubuntu just works TM
with servos you wouldn't need those extremely coarse screws
No. I do not have the leadscrews.
ok i have the live cd , but on live cd run emc2 2.2.5.i know the possibili the upgrade emc2 whit apt, but i want conpile emc2 source, and on live cd not is the package for compile emc2
motioncontrol: READ THE INSTRUCTIONS at the link I pasted earlier
install from livecd, then get needs for compliation
its all on the wiki I followed it, it works
I have the option of 1/2" 2 start lead screws (5 turns per inch) but then I would be concerned about whipping.
yeah, 1/2 seems pretty small for a 2' screw
you don't have one end free do you? both ends in bearings?
Both ends in bearings.
if it is like the original joes - they are bearinged on both ends
I see 1/2" acme shoudl be good for around 800 RPM. That would put me at 160 IPM. That's not bad.
I don't know the math but I bet the screws have a max speed for a certain length...
I do know the math. The math shows about 800 RPM for 48" but the lead scews are a few inches longer than that.
can you go up a diameter?
I hessitate to because it would add complexity to the build. Bearings get a little bigger and maybe don't mount in the same delrin inserts I had Joe machine for me.
or maybe your 5 start and large steppers are what you want, just brainstorming
Maybe there is a 1/2" 4 start?
I will check with McMaster-Carr
with steppers direct-driving a 2tpi screw, each step is .0025 inches
that seems awful coarse to me
I found a 1/2" - 8 tpi - 2 start. That might be interesting.
There's my 200 ipm.
plastic nuts or bronze?
Plastc - Dumpster ABL nuts.
ah, those are nice. I have them on one of my desktop machines. you'll give them a workout with this long travel.
I assume they'll need replacing every few hundred hours.
... but ball nuts aren't nearly as affordable.
yeah, and they have to be kept scrupulously clean and oiled
* BJT-Work just got a box with some 50 pin ribbon cables :)
Thanks again for the help, Cradek. I really appreciate it.
you're welcome, my advice is worth at least every penny you paid :-)
mine is worth less than that sometimes :)
* BJT-Work heads out ...
I pay people to give them my advise. Very therapeutic
skunkworks_: I could use some of your advise :P
hi to all
good night all
yey got my first run done today also, with all axis working flawlessly :)
is there support for "probing" in emc ? .. that is if i have a metal place and take my mill down to it it stops or records the position when it gets a current running in it? :)
did you try looking in the gcode reference for it?
im totaly new to the cnc scene =)
so no ? :)
[21:59:44] <jepler> http://www.linuxcnc.org/docview/html/gcode_main.html#sub:G38.2:-Straight-Probe
[21:59:52] <fenn> http://linuxcnc.org/docs/html/gcode.html
is a good one
G38.2 looks like something
but i have to put that myself into the code then ?
how else would a computer know that you wanted to probe, except that you directed it to?
jepler, i was thinking if there was a button in emc or something .. "go find depth" :P
or something .)
actually there is a "touch off" button
yes but it doesn't make the machine move
latest version off emc what requirements for pc do i neet to have
x86, 512MB RAM, not a laptop
[22:46:50] <cradek> http://www.linuxcnc.org/docs/devel/html/common_System_Requirements.html
oh wow an official answer
ow thanks cradek
second link on the docs page...
weird, I thought G28.1/G30.1 were in the docs but I see they are missing
i broke my first machine
the doodad that holds both the ballscrews in one of the bridgeports, it's this weird cast part, i snapped a good chunk out of that
is that what it's called?
under the table where they intersect?
I think so
I think people break them all the time :-/
it stuck me as an awfully retarded way to mount the screws.
what did you do?
one of the non-machinist guys in R&D brought the machine back up wrong after a power loss and left it
so when i went to go use it, i pressed one of the "rapid feed" buttons, unaware
it rapid fed, all right
and didn't stop
and didn't have any of the software limits in the machine anymore
and no hardware limits !
so it rammed up against the front of the knee and snapped out
well there was a "hardware" limit
pull the plug !
too far away
about all i had time to think was, "i don't remember the rapid feed working this way before"
on the positive side now i know how to disassemble most of a bridgeport
bizarre - doesn't it home?
im using only a small router for drilling pwb so emc should accept excellon file's
roel01: emc uses gcode
i know cradec im not the first time in this room he'
i'm trying to find out iff emc accept encoder inputs on lpt port
yes it will but not very fast rate due to limits
im working with stepper motors so not that fast i think
what are you using encoder for?
position off x an y ax
if ur encoders are not high count, 100ppr or so u will probly be ok, but i cant say for sure sorry, as i have not used or looked at limits of LPT encoder ins
should be in the docs i belive, in the intergrator manual