main emphasis on cheap and easy, so i guess i will just pick the cheapest and hope it's also the easiest
and what is this "forward current transfer ratio" thing all about. why would anybody buy a 0.5 when a 1.5 is the same cost
i think i've got it narrowed down to 13 pages of results
I think we have a winner: http://www.lumex.com/pdf/OCP-PCT114+C.pdf
at $2 for 10 of them I guess I can afford to be wrong
Dmess: if it works I'll send you a pcb
looking at how others interface stuff is handy http://www.anderswallin.net/2006/08/optoisolator-cards-for-mesa-5i20-servocard/
hmm wonder why he chose that one
hcpl2531 is rated to 20v output with only 8ma/channel
recommended by another whiz, he talks about it
and 5x more expensive too
wonder if they have better speed characteristics or some other desirable feature
gotta shut down, finished b/u this hd, later!
ahh that would be it, they are good into the mhz range whereas my cheapies are only good to 85khz or so
tomp: thanks for the link
i think for my application 85khz should be plenty good
it's so annoying that GTK2 accelerators don't work
I've got a 6.06 install problem. Live CD only see's my video at 640 x 480. I can't navigate through the manual disk install when I get to partition selection to continue. Is there key strokes that will let me pan the screen or a way to change the desktop size? Admin>screen resolution only gives 640 x 480 resolution
I've seen that problem, but it was so long ago I don't remember the solution
I hate graphical installers
im trying to get a driver to work and all that happens is stepper clicks and config crashes
i bought one of those mechatronics drivers off ebay
when you say the config crashes, what do you mean?
wonder if he's using stepconf
I wasn't sure if "driver" meant software or hardware
have you configured it propoerly...its config seemed a little odd to me and i have the same 1 not wired up yet
the pin outs are backwards it seems
backwards or inverted
anybody know the pulse rate for pdm output?
like a ballpark figure?
|dareposte|: up to one output transition per BASE_PERIOD
would that be normal? the developers manual suggested 10khz i think
as an upper limit
two transitions to go from 0 to 1 back to 0, so for BASE_PERIOD=25000 you get 2*25000ns = 50000ns = 50us = 20kHz
so for generic spindle speed control would it be wise to attempt to run a pdm that fast? it would definitely help the filter circuit to have a higher pulse rate, but what effect would it have on the cpu load
very little, I doubt you could measure it
choose you base period to provide the step rates you need for your other axes, and take whatever you get for the PDM frequency
is there any way to set it up to allow for decel / accel times?
if my drive takes 0.2s to accelerate then it would always be 0.2s behind for constant surface speed work
or is that not generally a problem
it's not a problem
you don't change radius that fast, and if you also use feed-per-rev you will never notice it in your part finish
i was thinking about a facing operation on a large part, but it was only a thought
yes facing is especially a time to use fpr+css
at cutting speeds you don't move far in 0.2s anyway
that is true enough
it is more of a problem if you rapid out to a large diameter and then start cutting right away
since the speed can still be too high
easy enough to program around i guess
usually, if you set a VFD for 0.2 (or even 2) second ramp time, that means "the ramp rate is such that it will go form zero to max speed in 0.2 (or 2.0) seconds
yes you just have to think about it when programming
if you are going from 1000 to 1100 RPM as part of css, the time will be much less
you'll see it on the first run anyway...
since you piqued my curiousity, here are two plots of function execution time for 10kHz pwm vs pdm: http://emergent.unpy.net/index.cgi-files/sandbox/pwmgen-pdm.png http://emergent.unpy.net/index.cgi-files/sandbox/pwmgen-pwm-10khz.png
-- the variation due to cache effects is much much greater than the variation due to pwm vs pdm
jmkasunich: i was assuming the default accel time of 5s for my drive, and maybe going only a 5% change in speed
adding a braking resistor will help a lot
(I need to get one sometime)
well, I'd set the accel time to as fast as I feel comfortable with, rather than using the default
jepler: did you just make those??
I have my vfd set for "don't fault no matter how long it takes to decel"
PWM and PDM do the same amount of math, the only differences are that they use different comparison thresholds and PDM does a lot more OUT instructions
PDM does NOT do more out instructions
by the way, that number is in CPU cycles; the CPU is about 2.13GHZ so when all the stars (and caches) are aligned it takes about 46*nano*seconds to do the pwm/pdm calculations.
pwmgen only sets the HAL pin, and the parport driver updates every time whether the output changed state or not
I thought that was essentially the same as "interlaced" PWM??
does the parport driver optimize away redundant writes (if new == old don't bother)?
no point for the most part, since you can't have a machine that will crash if it ever has to output the bytes :)
|dareposte|: hal has some nice bits built in for performance monitoring -- you can track in halscope the execution time of each HAL function (such as pwmgen.0.make-pulses), and you can view in halcmd (but unfortunately not graph in halscope) the total execution time of each thread
I suppose it could, but that wouldn't help the worst case (would hurt it a tiny bit) and it would increase jitter
it would only increase jitter for the second write, and of course anything that comes after that function
in any case, PWM and PDM do the same amount of work ;)
jepler: I have not gotten so far as to actually get the electricals hooked up yet, I was just making my circuit for PDM to analog 10v signal and was trying to get the design right
I'm sure I will be getting intimately familiar with halscope very soon though
|dareposte|: it's a very useful tool, you'll be glad once you've learned it
thank you for the help and advice everybody who jumped in, very informative
on the accel time wouldn't maxing it out be harder on the gear train though
or belt in my case
dareposte, did you see the $28 PWM->analog spindle converter?
a bit, but nothing compared to a normal "across-the-line" start of the motor
yeah I saw it
ok. it may not be quite what you need
(if you discover that PDM is too fast for your optos or whatever, you can simply select an appropriate pwm frequency and turn on dither -- that'll be essentially as good as you can get. "minimum on time" (or off time) would be another interesting mode to have in pwmgen, hmm)
you probably need to use the analog version, which means you'd have to filter anyway
my optos are supposedly have a cutoff frequency of 85khz
swpadnos: I did like the $27 circuit, but I'm trying to incorporate the digital IO and analog speed control all onto the same board
there were those too, for $65 or so I think
which $27 or $28 circuit is this?
may have been a different site
homann designs DigiSpeed GX or XL or something
maybe, i ordered enough optos and opamps to make 5 boards for about $20 though
isoalted analog output from 0-5V, 0-10V, or PWM,
and maybe serial or something also
(different versions though)
is it common practice to actually use the double-redundant e-stop lines?
on these premade boards that is
I don't think so, but it depends on what you mean by that :)
since e-stop should always be designed as fail-safe (ie, if a wire breaks, you can't get the machine out of e-stop), having two chains just adds extra stuff that can break
i don't know i guess its some safety standard to have two e-stop lines in parallel
yes, there are some situations where that's done. I think the semiconductor industry is one
they make us do it at work that way, and i just assumed it was some safety law or rule they had adopted
it is, for some industries
whether it's necessary / advisable really depends on your situation I guess
is cnc machine tools one of those industries? :)
I don't think it's required, but I don't know for sure
the main things are that you can't have a single point of failure that could make it so the machine can harm someone (at least for Semi-S2)
and you have to remove alll hazardous energy from the machine as fast as possible when E-stop is hit
(and also poisonous gases etc, in the semi industry)
no poison gases here fortunately
there is also a rule (again, I don't know if it applies to CNC specifically) that e-stop can't depend on silicon at all
that rules out my optoisolator e-stop circuti
I don't know for sure whether that means no programmable silicon, no microcontroller/FPGA, or no transistors
I think all of the above
in my work its required to be hard-wired with dual channel "safety" relays
as long as you can guarantee that a failure of the opto will cause an estop, then it might be OK
relays only is my understanding (I am not an expert)
cradek, I think so too, but never confirmed it
I designed that in my recent project - just seems smart.
yeah an e-stop button is one thing that i definitely want to have work
the trouble with transistors is that they can fail shorted, which isn't acceptable in an estop chain
estop should be isolated through a mechanical relay, and EMC should only have an output to that chain
couldn't a relay stick though
dry contacts are unlikely to stick
but you could use a contactor also
a tiny one :)
mine's a decent size contactor
those safety relays sort of look like contactors now that you mention it
switches off the AC that runs all the other contactors
is estop in series with servo power or something?
so for the e-stop chain it should probably be hard wired with EMC having a single e-stop out into a relay which breaks the NC chain
no, to a relay that makes the estop chain while EMC is ready to run
you don't have to do that, but you can
EMC needs to have an E-stop input too though right, so it can stop the steppers
I use emc to trigger the event to come out of estop. all the buttons have to be out for this to happen
cradek: my vfd says not to cut the AC to it in event of estop, as it will stop faster if it decels
right, and the contactor is in a latching configuration
yes, that's a yough question
|dareposte|: that's a "pick your poison" case. I chose to send 'external fault' to mine and 'unplug' it
I guess I'd rather everything go limp than try to decel under power when something is wrong
but it's only me who has to use it. I don't pretend to know what anybody's "best practice" is
the documentation for mine says it is likely to hurt the vfd as well (??)
except Z - you really want a brake on a head that could fall
SWPadnos: very true
you shouldn't cut power in normal opeeration
|dareposte|: docs with mine are foggy about that
but by definition, estop is an emergency
and ideally those wouldn't happen too often
i would envision using the estop button when i realize my code is no good
the docs say not to do it as a way to normally control it, use it just for emergency
it says it "may" "shorten" the "life" of the drive or some nonsense
though I must say, estop was a different thing on a particular machine I saw
better than shortening the life of the machine operator
i was thinking maybe use the red mushroom estop button to command a rapid spindle stop, and decel and hold all axes
that's one option
and if things ever really got out of hand have a manual disconnect
that cuts power to the whole rig
usually you cut power to anything that might get a command from the PC
because the PC could be crashed ...
decel and hold all axes how?
hadn't figured out the how yet
maybe force step and direction high for each drive?
oh steppers? they're easy to stop :-)
decelerating may work on a DC or AC servo, but it doesn't work easily on steppers
better to turn them off
you get the best decel when you slow the pulse stream
yeah - I wonder if drives look like a short when they're depowered
i would like to not lose steps when i hit the e-stop button
yeah, those are incompatible goals
if you use abort, not estop, you get a controlled stop with position preserved
maybe a yellow abort button, with a red estop button then
even if you use encoders (which I'm not suggesting), you would have a full electrical cycle (4 full steps) of uncertainty when you restart, since you don't know where the stepper was beforehand, and the drive will come on in some state that probably isn't the same
yes, abort is different
that's "stop pleast"
rather than "STOP THIS BEFORE IT EATS MY ARM!!!"
is abort a HAL input?
err rather can it become one
when you think about it in terms of loss of life or limb, little things like reducing the life of your VFD become less significant
you cannot turn off steppers and keep position
(or they should)
I bet halui has an abort input
if it doesn't it could
i like my life and limbs
so i should probably just deal with having to rezero the machine after an estop
the only good way to avoid that is to have a well-designed feedback system, with careful consideration of how it's powered and when (in addition to whatever it connects to, such as a breakout board or whatnot)
my stepper machines weigh (much) less than me. my servo machines weigh (much) more than me. this makes me generalize that stepper machines are not dangerous. I suspect this is wrong.
maybe that they are less dangerous
at low speeds, or with a running spindle (for mills), they'll hurt you pretty well
you don't even need a motor to be moving to have a mill cut you
the spindle doesn't even need to be on
if the machine can cut metal, it can cut bone even easier
that's why we take the tool out before we peel the pcb off the table...
unles syou have titanium limbs ;)
i have bad news guys
there are rumors that EA is making system shock 3
I'm not sure that's going to matter much
should EA and 'system shock' mean something to me?
because most of the people in here don't bother playing FPS games
but system shock was barely an FPS
(I do from time to time, but I like stuff like Wolf:ET :) )
see how much I know? ;)
system shock 2 is one of those games where if you say "this is the greatest game of all time" you do not get much arguement
ah, so it's TRON, only better :)
EA is a horrible, horrible game developer but they own the license to SS
the guys who did it the first time around moved to a different development house
hmmm. did Portal ever get released?
and put out bioshock
portal did get released, and it was awesome
bummer. I missed it
it's on steam if you want to try it, it's worth the money
I'd prefer a Linux version
wow after like 2 years i have a proper setup
pathetic for even a hobo
anyone want to whip me with a chain
um. no thanks
i was tricking, usually i have to pay
you can try disabling estop and performing dangerous feats of milling instead
um. no thanks. i still have to connect everything
the story of my life..
a truckload of parts, a big spool of wire, and a crimp tool
SWPadnos: would just taking a blowtorch to the estop button , then dropping the chargepump in LN2 be much EASIER?????
depends on which hand got caught in the spindle
then the answer is no
SWPadnos: Fine, see how you are... we'll just wire the estop button to shunt the mains instead
jmkasunich: alt left btn drag will move the window so you can get at the buttons
i've determined that emc only works for me when booted from live cd
that way it wont conflict any drivers/settings to my system
is that an ok and recommended way of using emc?
the PC I use was installed from a live CD , it has a plug in tray to replace the hard disk to go back to windaz
wish i had fast pluggable hd's myself
hmm, i know i have 60 gig unpartitioned space
can i install another linux partition from the live cd without modifying the existing system?
is it smart to install the system on my unpartitioned space
probably but Ive never done that so dunno
gonna try it
hope i wont screw the remaining partitions
do you think 30 gig is enough?
mine is on 8gig
so i can pat myself on the back for having 60 gig free unpartitioned space
I had a partition size problem with an old bios
anyway, lets roll
brave and foolish aren't that different
i've chosen a new "xyz" configuration
i wanna play around with jog shuttle and simple g commands
wander what "joint 0 on limit switch error" means
You have a config that looks for limit switches.
Are you setting this up with stepconf?
yeah i did
Run it again and where it shows limits select unused.
You can mouse over those boxes and use the center mouse button or click and drag the highlight down to the bottom.
i opened another config "stepper xyza" now and now it doesn't bother with limit switches
The stepper configs don't either.
its a program example with the emc logo
as a matter of fact, i'm only using two motors at the moment
its a wire winder
Do you write gcode for the number of layers of wire?
thats what i plan to do
the spindle is a 200/steps revolution motor driven at 1/8 microstepping
How fast do you want to spin that?
tbh, i'm not sure of the capabilities of it
would be nice if i could try that out somehow
You could think of it as a linear with 1600 steps per unit.
Or an angular with the same.
the graphical representation on screen wont be relevant in my case
how do i flush the existing g-code
In fact I'd use a non plotting gui
Just load a new one.
but i can write my own right
I've not tried loading an empty file but one with an m2 on a single line would work.
sure you can.. but I don't think you want to
probably running tkemc or xemc should be good enough
i've figured out i cant edit the lines at the bottom
It is easiest if you put them in ~/emc2/ncfiles.
mr_boo: no, you can't.. AXIS doesn't allow editing of programs
it only opens/runs programs
so the emc2 app is not an editor?
nope, it's a machine controller
If you start the file browser and click to emc2/ncfiles
there are tons of editors out there, you can pick your favourite ont
I keep an editor open and just reload
right, that's how it's supposed to work
then right click over an empty part of the display and select new file it will create an empty.
I'd use the mini interface cause it has an editor built in.
But that's just me.
i'm a total newbie but for what i've seen so far its the coolest software on the planet
Yes it is.
methinks would be nice if axis could be extended one day to edit when not running
wish i had a cnc mill now
You drag the wire back and forth over the spool?
as it spins?
thats the plan yes
maybe i should post a picture of my machine
That should work fine.
looks a bit messy
the real things look messy
we have 3 here
the spindle part of the system is very simple
I see that.
the threaded bar is intended as wire guide and i need to figure out how many steps required for traveling a certain distance
What is the thread pitch?
1.2mm i think
but its not direct drived from stepper
What is that ratio?
unless the back tension is good enough it will jump off the guide
im not entirely confident this is actually gonna work
Is that a 1.8 degree per step motor or 7.5?
wheee.. they discovered water on mars
[12:19:31] <alex_joni> http://www.phys.ncku.edu.tw/~astrolab/mirrors/apod/image/0504/WaterOnMars2_gcc.jpg
this will give me a clue of the travel distance versus number of steps
for mm units you'd use 84.444444444
Then you'd use the diameter of the wire times the number of turns for the total travel for a layer.
Do you use more than one wire diameter?
that will happen later on but its easy to recalculate
first step is to be aware of traveling distance as function of number of step pulses
Sure. I was thinking of a general program where wire diameter was one variable and number of layers another.
And number of turns a third.
But that will come as you get into gcode.
some calcs and a subroutine in gcode
If the wire does hop over the thread you could make a simple nut that the wire threads through.
That's the idea archivist_ub
but leave about a wire space per layer
i've calculated that 768 steps should travel 1.25 mm
now i wanna put this to the test
good thinking. the wire will drop into the groove of the layer below.
* archivist_ub chalenges mr_boo to do wave winding
can i write g codes to the "MDI command" field for testing?
Yes you can. It will only make the one set of moves (block of gcode)
* mr_boo googles for some g codes
i know the first set of g codes are for positioning and traveling along axes
What is the diameter of your wire?
Which axis is the spindle?
i haven't even rigged that up yet but it will be around 0.1mm
Okay. so 10 turns per mm
spindle is on x and guide on y
hmm google fails for real wave winding, seems the term has been stolen for as different connection method
How many turns
haven't decided yet but i know i want something to yield around 300 ohm
Ah. What width will the spool be?
This first sample run?
inner diameter is around 6mm and width is 7.4mm
so 74 turns with a 0.1mm wire?
but im gonna get the hang of emc first before i rig stuff
Sure. I'm just trying to imagine gcode to do an imaginary spool.
so start with g0 x0 y0
maybe i should edit a file now
are they *.ngc
then g1 fxx x73 y7.4
i'm in ~/emc2/nc_files/examples/ now
maybe i should copy and modify an existing file was my thought
If you right click and select create document and empty file it will make a blank one.
Then name it myfirstspool.ngc
Then left click on and that will open it in gedit.
you may not be able to create or edit files in the "examples" folder (it's actually a link into a system directory). use ~/emc2/nc_files instead.
ah thanks jeff. I didn't see the examples added there.
back up one to nc_files
* jepler slaps himself for mixing "folder" and "directory" in the same sentence.
it wants an end command or "percent sign" at the end of file as it appears
my myfirstspool.ngc only had the line "g0 x0 y0" atm
Yes. I'd use m2
yes; I always write "M2" at the end of all my gcode files, though there are other things you can write instead
Automatic spool changer? m60
worked just fine
If you've set your unit scale so that one unit is a rotation and one linear unit is a mm
you can program using real winding and length numbers.
yeah, i know there are g codes to configure such things
so x74 y7.4 would give 74 wraps
i need to tell emc that 768 steps corresponds to 1.25 mm travel
turns excuse me.
I'd make that 1mm and the number of steps required for that.
is that a g code for that or should i tweak the stepper setup for that?
tweak the stepper setup.
The rotary isn't quite going to conform to the stepconf X axis page.
You could put 1 in the leadscrew pitch and it should come out the way I'm thinking of it.
that makes sense
I got 1013.3 for axis scale on y
other way :)
768/1.25, not 768*1.25 I think
I just filled in the blanks in stepconf.
it seems wrong though
And it's early here;)
Steps per rev = 48
it's early for me here ;)
microstepping = 8
oh, could be a different axis then
pulley teeth 9:19
I was looking at the 768 steps=1.25mm thing
which is 960 if you do it wrong
see - its early ;)
wow, you sure can get cheap flights to Houston for this weekend
i'm playing around with the axis test feature
having lots of fun :)
SWPadnos: how is your sister?
she and the kids (and the ex) got out before the storm hit, but she;s not sure there's a house to go back to
yikes. Glad she is taking her time. sounds pretty rough right now.
apparently there's damage to the causeway linking Galveston island to the mainland, and they don't expect to have power for weeks
she wants to go back to see what's left, but she'll need a boat to get there
SWPadnos: perfect excuse to buy a yacht
and lend it to her for a couple days/weeks
well, my wife did bring up the fact that houseboats are mobile, and somewhat less susceptible to flooding ;)
not that we can afford one
Maybe a fly over would tell the tale.
Would certainly be a good use for sat images on the web this week.
I just saw a video shot on Saturday, and the couple of buildings I recognized as being near her house were pretty unscathed. there wasn't even a lot of water in the street there
rayh: i've found out my spindle motor only can do around 200 rpm
rayh: does this make sense?
faster than that and it stalled
and this is at 8th microstepping mode
steppers are not good at speed
What voltage have you got to it's driver?
but higher voltages didn't help
not higher current setting
gonna take ages to wind these fellas
I know nothing about that driver but with a gecko higher voltage will encourage a faster spin.
remember that my motors are a joke
as well as driver
How many layers do you guess you'll need.
the driver is an A3977 based one
but i've got some prodrive 2000's that i'm gonna use when i get real stepper motors
hopefully all three of them can be used then
Well you've got a start with what you have and 3 wraps a minute is better than hand winding.
ah, finally done with the stepper setup
maybe i should rig stuff up now before writing code
must install flash first to this partition
wish i had gear like that
he sold it :P
to buy something better i believe
na - to build a garage...
rayh: have you figured out the g code for a wire winder?
I'd set up the rotary as 1 unit per turn, because it's natural to talk in turns instead of degrees
so if the coil is 1" long and you want 100 turns, you'd program G1 X1 A100
about g codes, would G1 X3 Y4 make a straight line between (0,0) and (3,4)?
between "wherever you are" and (3,4)
i guess G2 is for relative motion
G2 is an arc
have a look at the gcode quick reference
g2 is for arcs. look up the G-code quick reference on the linucxnx website
[13:47:35] <cradek> http://www.linuxcnc.org/docs/html//gcode.html
err - linuxcnc
you know it's good advice if two people give it at the same time
even if one mis-spells it
(<1/2 cup of coffee so far, so I have an excuse)
we have spelling on IRC?
pulled out your IV again?
it fell out overnight
[13:50:57] <alex_joni> http://findarticles.com/p/articles/mi_m1272/is_n2629_v126/ai_19841026
[14:38:36] <skunkworks_> http://www.cnczone.com/forums/showthread.php?t=64683
I have 4 axis in one port what is he doing wrong
he doesn't say whether he's out of inputs or outputs
does hal_input work both ways? IIRC it does..
archivist_ub: winding activation outputs for three axes takes up 12 outputs...
skunkworks_: yes but don't recommend that unless you know what he needs the IO for
I think the only output type is an LED or something
yes well not using proper stepper drivers is doing something wrong :)
Pokeys emulates keyboard and joystick
that's just probably not an appropriate device
it would not be realtime for sure :)P
unless he's got a lot of non-RT inputs to move there, and can use the parallel ports as output
the poKeys is 49.99EUR for the terminal version
$75-ish, might as well get a 7i43
skunkworks_: uh, maybe use the amp-enable output? jeez
or machine on
skunkworks_: if you do anything else you'll get pid buildup while the amps are off, which is bad
ah, well then :)
I guess I'd make it so it powers up the drives in estop reset, but enables them in machine on
Magellan or Copernicus, I think
one of those anyway
I think the accessibility tools include a magnifier, but I'm not sure about a screen reader
those are so irritating
I hate them
ah - gnopernicus :)
if your posting everyone assumes you need help
well it's not just me then (I hate web bbses)
I think Benny did that :)
SWPadnos: the screen reader is unlikely to work for apps based on tk such as tkemc or axis.
what are you guys talking about?
I don't see any indication that the question has anything to do with using EMC
jepler, right. but though I think accessibility software is good in general, blind people shouldn't be running milling machines
I met a blind wood turner!
SWPadnos: that's racist
emc2 - not for the blind?
he felt the work
there was a Linux distribution for the blind a few years ago. the entire install had speech synthesis available. the only thing you had to do was get the CD in the drive right side up
I can do that in only two tries
my grandfather fixed cars after he went blind from diabetes by feel
you don't want to check tool position by feel ...
cradek: some people take more than two tries
i wanna thank for all help
have fun , see you
be back later
I have just intalled version 2.2.5
but I have some problems with latency
the rtai latency test gives me several overruns
which disappear after switchin from nvidia to vesa driver
but axis won't start with vesa driver
yup, nvidia driver is one of the known culprits for bad RT performance
yes I know
installing the nvidia driver will surely bork a good working vesa driver
that's why I switched to vesa
not sure if you can fix that without reinstalling
but as I said axis don't want to start
nvidia (the binary driver) does some odd things to the software Mesa install
even if you uninstall the nvidia driver, things stay messed up
alex_joni: even if I load vesa module?
I don't know if (or how) you can fix it, I just know people have complained about this before
can you try uninstalling (with --purge) both nvidia and mesa, then reinstalling mesa?
I don't know if it will work, but if it does that would be good to know
I remember poking it until it worked.. but this was at least a couple years ago
I have official nvidia driver installed (legacy-drivers)
I wonder if removing then installing the vesa driver might also be necessary or helpful
and mesa libs 6.5.2
what does official mean?
the one that's not open source..
I think it's called 'nvidia' vs 'nv' the open source one
anyways.. I would try to uninstall the 'nvidia' driver
I'll try and unmerge the driver
also .. AXIS is probably failing because of missing GLX extensions
it's known to happen that you simply need to add the option to Xorg.conf
ah... I'll have a look at xorg.conf
so you're on Gentoo then?
how much of a pain was it to compile emc2?
not a pain at all
got some info what kernel/rtai/gcc/python you used?
gcc version 4.1.2 (Gentoo 4.1.2 p1.1)
got an ebuild for us? :)
if you have >= 1G of ram, you should update to 3.6.1 rtai
SWPadnos: that was an idea but I think it is beyond me
the ebuild is easy
yeah, me too. jepler could do it, it's python :)
supporting the millions of users is painfull
* alex_joni is kidding..
pierp: glxinfo give you faster testing if it works
tried to restart X with vesa and glx but axis sent a BadWindow (invalid Window parameter) error
glxinfo | grep GLX I think
yup.. mesa is messed up
that's why you remove mesa and nvidia with purge - they both fiddle with some GL library. mesa wants its own library there, nvidia symlinks it somewhere else
getting rid of both might remove the file, which might enable a reinstall to work
but it's just a theory
ok I'll have a go with it
Jymm, get back to work. slashdot seems to be down
or at least ask.slashdot.com is
uh - .org - you know what I mean
hmmm. actually, everything but the front page seems dead. I wonder if the database server got slashdotted
SWPadnos: works here :)
yep, it's back here too
I bought this stack of smashed lathes on Saturday: <http://sandbox.owad.org/lathes.jpg>.
I'm hoping to retrofit one for CNC, and turn another into a rotary table for the router.
for the rotary table, I was thinking about just replacing the motor with a stepper, and putting another chuck in the tailstock. Is that a reasonable way to go about it?
depends what kind of rotary you want/need
What kinds of rotary are there?
various, platform and on the end of a spindle
I use rotary for gear cutting
this'll be on the wood router, so I'll probably be using it mostly for carving/models/that sort of stuff
theres a goos vid on youtube of someone engraging wood objects on a rotary
[17:12:08] <owad> http://www.youtube.com/watch?v=C0bh2SUeGHI
archivist_ub has changed the topic to: Welcome! EMC (Enhanced Machine Controller) is a linux-based opensource CNC control. | Latest release: EMC 2.2.6 | http://www.linuxcnc.org
cant see it I need to reload firefox, been up too many days
ChanServ has changed the topic to: Welcome! EMC (Enhanced Machine Controller) is a linux-based opensource CNC control. | Latest release: EMC 2.2.6 | http://www.linuxcnc.org
heh what happened there
I was just saying bleh cant see it I need to reload firefox, been up too many days
archivist_ub: somehow you had a /topic in front :)
but I never type /topic!
maybe clicked the topic?
dunno I was watching the keyboard
why would you want to do that?
hmm I did copy the wiki url a few secs ago
if you're using chatzilla, that may have left the topic focused, and it is editable
anyway owad, now firefox, restarted it wasnt that one but you got the idea
at least it is sometimes
usually one gets a telling off, you dont have ops etc
as " #php-support :You need to be a channel operator to do that"
thats where I can op mysqlf
if +t isn't set anyone can change the topic
+t was not set
now it is
toastatwork: you can change it all you want.. ChanServ will put it back
alex_joni: just because chanserv puts it back doesn't mean you can't chagne the topic, which is what i was explaining regarding not having ops yet still being able to change said topic
*change is now a meme
Yesterday you talk me about inversed deadband
I use the last livecd and i can't find comp ?
how did you look for it?
you should be able to type comp --install at a command line
in the home
comp --install something
if you're trying to find it in a menu, you will fail :)
yes but i have command not found , sorry
that's a little strange. you may need to install emc2-dev
ok i need to connect to internet, then
actually, I think you need to install a couple of things so comp can do its job. you need at least build-essential (for the compiler)
yes, I think so
it will not be easy
thanks a lot, bye
there is a section of the manual on using comp, but I'm not sure it tells you what needs to be installed before using it
well, I guess you do need emc2-dev
SWPadnos: the wiki page ContributedComponents does say to install emc2-dev and build-essential. I don't know what the pdf manuals say, I didn't check
right - I got the same info from the wiki page and also didn't check the manual
hi all, I was trying to run the emc latency test
but perhaps I do something wrong
as I get an error
realtime RTAI latency test works with no overrun
pierp: if you are waiting for an answer, be aware that everyone is waiting for you to say what the error is
File "/home/pier/Programmi/emc2/bin/pyvcp", line 38, in ?
File "/home/pier/Programmi/emc2/lib/python/vcpparse.py", line 24, in ?
ImportError: No module named ext
lat.hal:5: pyvcp exited without becoming ready
I believe that should come from the package python-xml
on my system it does anyway
is this an os other than ubuntu 6.06 / ubuntu 8.04?
ah, you are missing some python xml stuff then, you will have to figure out how to get it
perhaps some missing flag in python
totally unrelated to EMC: why don't GTK2 button accelerators work? i.e., a button with the text, "OK" where the "O" is underlined does not activated when the O-key is pressed
ok thanks, I'll check the flags in python
thanks, cradek, I'll try that.
crap late for class
he sure is aptly named
argh too late
apparently it's part of the gnome human interface guidelines that access keys always are paired with alt even when the active control is not one that uses unmodified keys.
that's a good thing
anyone using a taig closed system?
do you mean closed loop?
i'm wondering if there are any hal files for the taig
I don't think there are any. What sits between the PC and the motors?
a black box!
heh. well, if you can get specs for what's in that black box, someone can help you :)
tahnks, what sort of info would be handy?
well emerging pyxml made latency test to work
nibbs, if it's parallel port connected, there's hope that pinouts would help
if it's a serial or USB-connected box, then you'll need some hardware changes
if it's a stepper system that also has encoder (shouldn't be, but I just found such a description on the web), then you'll need to think about which parts you want to keep :)
yep its parallel
yep its got encoders
servos or steppers
[20:12:50] <nibbs> http://www.microproto.com/MMDSLS.htm
its runnig fine on mach3
it probably just has a missed step pin that runs an estop signal or something.
hmm broken website on my firefox
yep mine too ff3?
yes , ive looked at source and its there
ff3 seems a bit buggy. no more drag n drop of images into apps as well
hit ctrl-A and you'll be able to see the text
its built in frontpage so.......
with FF3 you can set the style to "no style"
heh good cheat
you'll lose the background but the text will be visible
i hope his electrical engineering skills are better than his web dev skills, frontpage? pah!
well, it looks like there isn't much technical info on that site
I wonder if he uses the encoders to fix microstep errors
he does say that the DSLS hardware does the work of making sure the motors follow the command position
so it seems like it's still a step/dir system as far as the PC is concerned
yes, i upgraded from the previous system to this one to cure my lost steps
there aren't enough I/Os on a single parallel port to have step/dir out and encoder in for 4 axes
the cure for lost steps is to run slower or get bigger motors (or in some cases, to change drivers or increase the power supply voltage)
Includes: DSLS Electronic Control Driver Unit
the encoders cured my lost steps and helped me run rapids a little quicker
cool. glad it worked out :)
me too, it wasnt cheap!
this page http://www.microproto.com/DSLS.htm
makes it sound like it is encoder counting + PID in an external box, similar to gecko g320
* archivist_ub contemplates what that actually is
but .. who knows .. it's "proprietary"
I imagine we could erm do something similar
one thing I'm currently thinking would be cool is an open source CPLD-based stepper driver aiming for a gecko-like feature set (whatever exactly that means :-P) and a wide range of max voltage/current capacities by allowing choice of transistors and high-side drivers.
with over current for short term
oh wow microproto.com's software (i.e., mach3) has "Complete G-code set of ALL G-codes."
heh for what value of ALL
it even has G1001
jepler: and the CPLD would be on a pci card, and only the power stage outside of the PC?
that might be a bit noise prone
alex_joni: no, I'm not talking about step generation and sarcastically reinventing mesa
I mean the driver itself, current sense, etc
why not have it all in the same fpga?
stepgen + part of the driver?
one of mesa's add-on boards has that but I don't think hostmot has the support for it. and I think the current control is open-loop only, not closed-loop.
you can put the stepgen , but bether if it control directlt the mictrosteps
so depending of the speed , smaller or biger microstep
with the feedback it can easily move to the correct place, no need to suffer microstep error any more
and outside just a single H bridge
if you want to have fast moves , forget the microsteps , but at small speed no other way
The only driver I know to do that is the gecko 203
may respond faster any way as its under control
the board I'm thinking of is the mesa 7i32
maybe it is closed-loop current control inside the 7i32, I'm not sure
what speed are you targeting
"the 7i32 sets the stepper motor phase currents based on the PWM inputs. These inputs are filtered on the 7i32 card to provide a reference voltage for setting the current." earlier, the manual talks about working with softdmc.
issy: well, this isn't a product I need (my machine has tiny 2A bipolar motors which even the simplest boards can drive)
The 7I32 is closed loop but fairly wimpy...
i've got a question about the so called 'feedrate'
you can use simple allegro chips , they are 3 A ,single chip drivers , accepts steps and dir
is that distance units per seconds?
F word in gcode is units/minute
We tried the allegro Step&dir chips -- no good!
i'm using allegro chips atm
yes right A3977 , for 2A they work fine
mr_boo: units / second
in the INI I presume you are asking..
and you can switch the step
The problm we had was with the new 80V chip...
i'm using two motors, one of them is a 90 ohm/phase and the other 4.6 ohm/phase
pcw: heh, wishfull thinking?
poor combination, yes i know
for them , steppers from www.motionking.com
up to 70V
<alex> I mixed it up with a step controller we are working on (which is 80V or >)
ooh.. sounds interesting enough :D
issy: how is the gui progressing?
stuste1: there's a binary you can try out.. but no source yet to be inspected
only if we can get the price low enough (uses $2.00 DSP)
pcw: I sent you a private msg
finish the cycle time , fiew bugs in the coordinate system. now Iworking on the hal visualaizer
my inspection of the source is suspect anyway :)
no licence statement on it .. so it might eat your computer
in the ini file can be set which hal component where and under what name to be visualized
alex_joni: thanks - as this is a zip file is this for windows only?
I think both
I ran the windows executable, but saw an ELF in there too
I have it. I will look and see what I can do with it. Thanks
ust one advice.
reed the ini file carefuly
blah, "linear move would exceed limits on line 2"
mr_boo: need to adjust your travel
how do i bypass all that stuff
i've got a wire winder
aint got any travel limits in that sense
does that have anything to do with the stepper setup?
you can put in some really long limits
you probably want a couple thousand rotations, not more.. right?
and the rest
G0 X0 Y0 will probably take a while though :D
uhm, where do i adjust limits?
did you use stepconf
the .ini file is almost as simple as the normal file - wow :)
so it does indeed have to do with stepconf
it sets a default
I believe this is the first time I have seen a copyright on a gcode file
the file '3D_Chips.nc' in the library has a copyright notice - no problem - just different
stuste1: it's been in the file for a long time
is there another way to reset the program other than "home axis" on all of them?
every emc had that bundled in the last couple of years I think
it was taken from the emc sources , no changes made by me
not a problem at all - just hadn't realized there would be a need to copyright a gcode file - I have created thousands of gcode files and never dreamed to copyright them
ask the owner :)
duh - thanks - I probably won't
any sugestion or questions on the gui are wellcome
naw is the time
tried to wind my first coil
the guide did not catch up thats for sure
gotta check my calculations
It's a start
how do i outcommend g code statements?
put () around them
could be that the guide moved in the wrong direction?
well, lets mount thread at other end
That would do it.
how do i reset the program so it starts from scratch?
just reload it with ctrl+r?
i must've done a poor job in setup, when i instruct it to move 40 it moves 35 on guide axis
the easiest to check is instruct it to move one motor rotation
dangit, I missed pcw
anyway, thanks for all help, gotta go