jlmjvm: atleast this is a project that could be done in next to no time
your not done yet?
BigJohnT: nice table you got there
thanks, can't wait to get it done
good night all
good night alex
jepler: did you post the .py and .hal for the 5 axis simulation? i was wondering about how you drove it with gcode.
oops, ok, i see browsing CVS (nevermind ;)
Bil1 is now known as Duc05
skinny77cncpuppy:i finally got my homing problem fixed
would somebody please test this link: http://www.anderswallin.net/2007/08/dah-lih-emc2-conversion/
I think anders is having server troubles
Error establishing database connection, but the server responds
thats what I get - wordpress seems to think I want to log in, not just read the page
same on any truncation of that path, even root
wordpress uses a mysql backend
probably something wrong with the isp's mysql server
jepler: where can I find the ngc file that drove the 5 axis simulation? (has XYZBC moves)
oops, maybe cradek, he's the listed author
tomp: it's nc_files/cone.ngc
report and pics of the Wichita trip: http://jmkasunich.com/cgi-bin/blosxom/index.html
and a clean shop too
and well lit
i wish the places i go were that nice, usually i wear a headlamp :)
nice report jmkasunich!
thanks for posting that
EMC in the major leagues
'new sense of scale' haha
that would be a cool write up in Modern Machine Shop Magazine
or digital machinist
hint hint RayH
ray wasn't involved on this one...
i thought Ray published/authored digital machinist
jmkasunich: what kind of bandwidth do you have on your server?
maybe you could copy it over to linuxcnc.org and then we could link to it to show it off?
but SWPadnos generously lets me host the photos on his dreamhost
do a wiki page
anybody who wants to copy is welcome to do so
I'm going to walk the dog ;-)
Ok, but does SWPadnos have the linuxcnc server ? maybe we could put it up there too
linuxcnc.org website is also on his dreamhost
it would probably be nice to have a "how to" for doing that kind of scale + encoder thing
but I'm scared
SWPadnos: what do you think about putting it up on linuxcnc.org?
it is a non-trivial thing, and if we make it too easy to try, people who don't understand what they are doing will dig themselves holes
jmkasunich: you did a good job of explaining it in english
I think that's the best way in this case
note that I didn't explain in detail - if you don't already know PID you shouldn't attempt this
ditto if you aren't already comfortable with HAL
was there a sizable difference between the report from the linear and the rotary? like a not-consistant difference?
is linuxcnc.org up at the moment?
I get a white screen
looks fine to me
tomp: there is a few thou of lash which shows up as difference between the two
also, we didn't know (and really still don't know) the exact scaling of the rotary
you could see the various screw errors in the difference too
it appears to be a 1/2" pitch screw, but we don't know the ratio of the timing belt between motor and screw
so all scaling was pragmatic
errors other than lash were small, less than a thou IIRC
lash was big - about .006
but we tuned it to correct on reversal in a few tens? of msec
yes, pragmatic - scale is 25400 counts per inch, encoder is something like 24848
you think the tuning was difficult because lash or stiction/mass low accel
it wasn't particularly difficult
its just not as dynamic a machine as something like the mazak
we tuned the motor encoder first, then moved the integrator to the scale
120 ipm (2 ips), and IIRC 2.5 inches/second squared accel
by contrast, the mazak does 300 ipm (5ips) and 20 inches/second squared accel
yeah, thats snappier than what i imagined, low top end but good response
we had 0.001" indicators at both ends, and were doing 96" rapids and stopping dead nuts on
at 90ipm (taking one minute each direction)!
oh, thats right, I forgot we slowed it down to 90 ipm from 120
those are hard tests when you have to really pay attention for long periods
we discovered an idiosyncracy of the scales and their electronics
they don't generate a uniform pulse train
every 50uS, the electroncis reads the scale, and bursts out any counts that have happened in that 50uS period
wow, some buffer dump
the pulse length (in the burst) is settable with jumpers, and initially it was too short for the PPMC to detect
but then realtime has a problem if the thread runs faster
so we could reliably count 0, 1, or 2 counts per 50uS period, but if we went fast enough to have 3 counts, we would stop counting totally
the G&L is using Jon Elson's PPMC cards, counting in hardware
so the threads aren't an issue
but jon's boards filter the input digitally, a pulse less than 3uS is ignored
and of course, the scale manual explains this in engrish....
haha japanese scales? oh yeah sony magnetron
these tiny tiny tools in huge machines are funny: http://willepadnos.net/jmkasunich/cinci-spindle-1850.jpg
that machine has about a 10 foot long table
over on one corner is a Kurt vise, and in the vise is a 3 x 5" plate about 1/16" thick
seems like they could really use something the size of a bridgeport but with a 6-10krpm spindle
they're drilling 1/8" (or smaller) holes in it
well, in that case, they were spinning that drill bit at a few hundred RPM
the part is inconel ;-)
well, its 30 mins since I said I was going to walk the dog, and he's whining at me
me too, goodnight
i got some math err... 90ipm = 1.5ips = 38.1mm/s = 1um/26.247uS. the linear scale was 1um. oh, ok it broke at 3 counts in 50uS ( almost broke at 2 )
yeah 2 counts per 50uS was on the edge
i find it interesting that the g&l has a tiny integrated computer/monitor
[04:56:50] <fenn> http://www.mpm1.com:8080/machines/GandL/pictures/S1030251.JPG
what brand is that?
fenn: can you get to linuxcnc.org at the moment?
probably a gateway profile 3
I can't seem to reach it from here
works for me
i can get to http://www.linuxcnc.org
but cant get that last pix posted
I restarted firefox- now it works
JPG is coming down very slowly here
anderswallin.net is in the first page of google results for 'gateway profile'
DAMN... eta 7minutes
714Bytes per second... somebody needs to get some bandwidth
fenn: what website is that?
and it's only a 308K file too
well, it worked for me
I even connected to my server at the datacenter, so it aint my dsl that's for sure =)
fenn: this better be some damn good p0rn =)
[05:19:38] <fenn> http://fennetic.net/pub/irc/S1030251.JPG
fenn: Um, where's the boobies?! There's no boobies in that!
What computer is that? Never seen one quite like that before.
fenn: THOSE are boobies fyi.
google sees all
even your 'hidden' wallpaper collection
hidden my ass =)
still waiting on the coffee maker
i hear ya,on my second cup
mmmm, first sip
I found a way via fenn to get the config path so it won't have to be hard coded into tlo
I haven't tried it yet, I was reassembling my plasma cutter after a round with my mill modding parts...
thats gonna be a good looking machine
thanks, it's taking long enough...
i sure liked that idea yesterday,cant get it out of my head
a bolt on plasma attachment
just keep in the back of your head that your mill might not reach the speeds you need for the best cuts on thinner material
however it would work for a small part every now and then
do you need more than 100 ipm?
on thin material like 22 gauge I cut at 415IPM
but it will cut at slower speeds
good to know
just has more droos
to clean off
how fast for something like 1/4 inch
for my hypertherm 1250 with the 80 amp tip it's 105 IPM with the 40 amp tip its 48 IPM for mild steel
that will depend on your plasma cutter
do you own one now?
no,but im seriously considering 1
tractor supply had 1 for 700 bucks that caught my eye
ok, some advice on them. the thickness rating is usually for severing. severing is a french word for fxxking ugly cut
you can start from an edge on thicker material than you can blowing through
do you have an oxy/ac torch?
that would cut thicker material...
just depends on what you want to cut
i will get with ya before i purchase 1 to make sure its correct for the application
ok, what type of materials do you think you might cut
alum and mild steel probably
they are the same until you get above 3/4" for speeds and amps
looks like fenn's fix only returns the hardcoded path to the nml file not the one that is in use...
[12:19:41] <jlmjvm> http://cgi.ebay.com/40-Amp-Plasma-Cutter-Inverter-Dual-Voltage-110-220-Volt_W0QQitemZ180216268431QQihZ008QQcategoryZ113743QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
do you think that 1 would work
more coffee brb
it is 40 amp so it "should" cut 1/4" ok. Don't expect too much from a cheap plasma cutter and you won't be disapointed
he has enough bad feedback to scare me away from buying from him even if he has 8k transactions...
didnt even check that
jlmjvm: also make sure you can purchase consumables for the plasma torch. <Very Important
[13:19:36] <jlmjvm> http://www.tractorsupply.com/webapp/wcs/stores/servlet/ProductDisplay_10551_10001_36265_-1______?rFlag=true&cFlag=1
they dont have the 700 dollar 1 any more
for that price range look for a Hypertherm...
the $700 one might have been a local special
prolly was,it was around christmas
[13:25:50] <jlmjvm> http://www.youtube.com/watch?v=EHBiuCbqQXY&feature=related
i'm on dialup so it will take me a while to see it
k,its just a plasma cutting a mud flap lady
notice the bandaged finger LOL
i see that
that's plasma cutting at it's finest
[13:36:09] <BigJohnT_> http://951753.pastebin.ca/916232
what's on my mind is to "read" the list of tools from the tool table and then you pick the one you want to update, what do you think?
if you could make this 1 not force you to enter a dia and not overwrite an existing dia it would be awesome
that's why I'm thinking that you pick the tool you want and it pops the existing values in there and you change what you want
do you have an email?
jlmjvm: I opened another dialog window
BigJohnT:how do you do that,never done it before
I just right click on your name
did you see it?
try it again
did you see it pop up under #EMC?
nope.i think i may have to register my nick for that to work
seems like i saw something about that before
brb,breakfast time,will register and reboot chatzilla after i eat
ok anyhow at home its jet1024 at semo dot net
Duc05 is now known as Duc05_
BigJohnT:i sent ya 1,gonna turn this off and try to register
Does anyone know how to setup syntax highlighting for .ngc files in gedit? I tried to follow the guide at linuxcnc but that seems to be for an older version of gedit. There isn't the option to change the colors in the menu anymore
I might be able to help you
let me look at my notes
I'm looking for the particular directories....
well the guide I followed had me put a ngc.lang file in /usr/share/gtksourceview-1.0/language-specs
but now I think it uses gtksourceview-2.0, and the formatting of the file is different
yes, I think formatting is different
you can look at another language file to see how
give me your email over in #cam and I will send what I have
maybe /usr/share/gedit-2/plugins/snippets/ seems to be xml desciptions of different file types, then you select one from menu bar View | highlight mode | others
I think ngc.xml would be for the snippets actions
anyone running a mesa 5i20 card .. i have a couple of questions
tzak: just ask, maybe someone knows
well i had my system up and runnig, fully servoed and running programs... then i started up my machihne yesterday and all the limits were on and i couldnt start the machine. Then i had one of my amps blow ,and im not sure if i damaged the mesa at all
i turned it back on and tried a working amp , i am getting counts on the encoder and limt switches are working however i cant get a servo???
"all limits on and couldnt start" sounds undesirable to you but a correct action for the control. maybe power loss thru limit swx and e-stop circuit? "get counts , limits work, cant get servo" sounds like e-stop really didnt release or no enable to amps.. check with hal meter or voltmeter.
is there anyway to check the card to see if the pwm and dir are working
i looked at the enable and they seem to be getting the correct voltages .. nothing has changed since the original setup which was working ??
scope ( or a meter knowing low velocity pwm is read as a lower voltage and hi vel is read as a higher voltage. ) exercise the mesa using hal ( not emc2 )
oh and something has changed, you dont know where right now
yeah i guess you right on that
how should i test using hal ??
this is real crude.. the basic idea.. 1) you run a halconfig that starts the realtime system and asks that a 5i20 be recognized for use 2) you get a list of the pins by asking hal 3) you force some values and observe the outputs
the specifics... i'm looking for an example in the wiki now...
this may be handy, it's a gui that shows you what hal thinks is happening ( despite possibly blown ioutputs ) http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?M5i20_Halvcp_Test_Panel
yes, i think his shell script sets it all up without using emc2
how do you run the test panel ... from terminal?
wait, there's bug in script, i made it run, will pastebin.. put all 3 files in /home/yourdirectory/youremc2root/ ok?
ok i'll give it a try
tzak: also chg permissions on the shell script so it's executable, then change to the directory where it is, and type './iotest.sh'
can you see the panel?
i have to try this on my linux machine right now im on my windows box
i will have to get back to you on the results... what do i need to look for one i get the test panel up and running?
wait a 2nd path dependancy in script... need to pastebin again
i should just copy this script ???
hmm, i'm still having trouble with the script, it closes the halcmd session communications to the terminal, so some debug isnt available, like 'show pins'...
so i have a m5i20 and am trying to manually type the commands so it doesnt do that...
ok .. well i'll give it my best this type of testing is new to me
another question is there a way to switch the enable signal .. to reverse it ?
yes, (if reverse means change its logic level ) use a 'comp' or route the output thru an inverted output
how do you route the output through a inverted output?
i'm in the shell scripts now, bu ti think if you look at the hal docs for m5i20, the outputs may have normal and inverted pins ( or i'm thinking of the parport...) this is a good reason to get the halcmd running with the panel, then you can see the resources available :)
yeah i understand.. well i'll try getting the panel running and see if that works
jmkasunich; that's quite a mill! (the giddings & lewis)
tzak: i can exercise the dac enables and dac outputs values, but havent gotten the simple digital i/o to work yet
thats through the test panel
click enable, slide the thumb, see the output
ok i'll try that.. like i said i have to go try it on my linux box but i'll let you know how it turns out
for the digital i/o i thought the output was warpped to the input so clicking the output btn should change the led color
but not on mine !?!? not on panel, not in halcmd show pin
dang i had built a breakout to breadboard for it, and its at work, had leds and dip swx and test points
ok, try it, and i'm off to openGL machine design land :-)
yeah i have to go and test everything at my work to .. ill let you know the results
nap time for /me
I have a question about comments in .ngc files. How can you do a block comment, and how can you comment out just a line or portion of code
no block comments as far as I know
() are comments
there is block delete, iirc
can ( and ) be used to span multiple lines?
* alex_joni looks it up
re: block delete..
you can prefix lines with '/'
yeah so it can be deleted at runtime
there is a hack using O-words to "comment out" a large (or small) block of code - just put "O400 sub" at the start, and "O400 endsub" at the end
those lines get interpreted, executed only if the block delete is on
400 could be any number not used for any other O-word in your program
A left parenthesis always starts a comment. The comment ends at the first right parenthesis found thereafter. Once a left parenthesis is placed on a line, a matching right parenthesis must appear before the end of the line.
ok, so () comments are one line only
[17:51:19] <alex_joni> http://www.linuxcnc.org/docview/html//gcode_main.html#sec:Comments
I think I'll just stick with () in case I run this on a different machine
also, you can't nest () comments
alex_joni, thanks for the link
tzak: (ok u not here) it all 'works', but i have better luck using the terminal and the panel than using the script, and i just saw the whole section "# link digital outputs to button" is missing (may be why the button didnt cause anything to happen ;)
will pastebin new scripts and a recording of the shell session
jmkasunich: nice blog entry
will you write a small description on the encoder + scales thingy?
I did ;-)
it might be good to include in the integrators manual, and send to the user list
you did? didn't see the commit :P
its in the blog posting
u didnt see it cuz it was very small
I gain for the scale PID, P+D+FF for the encoder PID, sum them in HAL
tomp: oh, I saw it.. but was hoping jmk could bring it more into a manual form :)
anyone who isn't sufficiently PID and HAL proficient to start with that shouldn't be doing this, IMO
(call me an elitist if you want....)
jmkasunich: you elitist you =)
if you dont know what diddywadiddy mean, then dont mess with it lady (R Crumb) ==> PIDdywadiddy
jmkasunich: The sub, endsub trick will work in some places, but not within another sub.
Instead, use if  ... endif
That should work anywhere and to disable it, change the zero to one.
yeah, dunno what I was thinking
(Or, I should say, to reenable the code.)
guess I can claim its early
oh, since you are here....
or to value of variable at top of pgm? so you can edit the block at top of pgm easily
tomp: good idea.
maybe I already asked about this? how hard would it be to error out when a second Oword with the same number is encountered?
You mean in the case of two subs with the same o-word?
or two while loops, etc
I have a habit of writing say a roughing loop, then copying it farther down and modifying it for finish passes, or roughing another feature
If and when I get a machine to work on EMC with, I'll be implementing named o-words.
the endwhile of the 2nd loop goes back to the while of the first one
Yup. It will do that. I should be able to generate an error in that case.
I've learned that when a program takes a long time to load/preview, its time to hit escape and check for that particular mistake
<tomp> question for ya
I'll also make o-words be local, so that two subs can have the same (internal) o-words. That should eliminate a lot of clashes.
im trying to load the test panel and not sure where the files need to go in the linux file system
tzak: i making sure the m5i20 stuff is reproducable ;)
tzak: ok, the files go...
jmkasunich: If you generate a bug report for that with low priority and assign it to me, I might remember to do that. :-)
ok well i can test thing now so let give it a try
is it a bug or just a feature request?
tzak: where is emc2 installed on your system? find out. we'll call that the 'root' of emc2
You can call it a bug. It was probably caused by my being lazy.
well im logged in as 'root' as the user
no, dont do that and that's a different meaning of the word
log in as you ( a plain ol user )
this meaning of root is like root of a tree ( a computer file directory tree )
did you get my private message? got that facility?
no i didnt
never mind, we can talk here
are you a normal user now?
ok yeah normal user
then open a terminal
cd to your emc2 installation directory (maybe /home/tzak/emc2.2somethingsomething )
copy these files to that directory 'iotest.sh' 'iotest.xml' 'pyiotest.py'
sorry i can't see the emc2.2 directory???
ok, what is the linux you are running? ( for example, ubuntu )
yeah ubuntu with emc2.2
oh crap, it's been so long since i used the instlled version, i always use a run-in-place.... sorry i'm a bit lost where to put your files... we need input
tomp: what exactly are you trying to do? what files?
jmkasunich: thx, he wants to test/exercise the m5i20, thinks maybe hdwr failure, using awallinns pyvcp test panel
jmkasunich: where would the files go http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?M5i20_Halvcp_Test_Panel
why not just use halcmd show?
the original project was "test m5i20", but now the project has become "get pyvcp scripts working"
yes, i thought the gui was a good starting place :(
im having a hard time trying to figure out what went wrong with my system and tomp suggested the test panel which look like it could help
tzak: have you read the hal tutorial? do you know how to use halcmd?
i have used the halcmd
tzak: i'm working on a recording of a terminal session where i look at and examine the m5i20 pins, it aint ready yet
then I would suggest simply using halcmd to see what the 5i20 is doing
im just not sure what to check ??
well, what do you think is broken?
I seem to recall that encoders count, so that part must be OK
encoders are counting my amps are working i think something with the enable maybe ..
so put a meter (real one) on the enable output, and use setp or sets to toggle it from halcmd
see if the physical pin toggles when you tell it to
i have done that when i turn on/off the power in emc and it seems to change.. i will try with halcmd
hey guys, if i do a cvs update on trunk, am i getting the 'all tags / default branch' ???
tzak: a halcmd session in a terminal http://pastebin.ca/916503
beware you make one error in the terminal and halcmd drops out to prompt leaving you with a mess to clean up, if so, rerun halcmd -f, type 'unload all', then 'quit', then, back in terminal, type 'scripts/realtime stop', then close that terminal, open a NEW terminal, and start again.
"halcmd -kf" means "take input from stdin, and keep going in the event of an error"
it won't drop out until you say halcmd quit
also, even if halcmd does exit, that doesn't change the running HAL configuration, you can simply restart halcmd and keep doing stuff
didnt know you could re-enter either, was always trying to clean up and start anew
halcmd is kind of like a shell for HAL
if you do "mv <somefile> <somewherelese>" in bash, then close the shell, the file is still where you put it
likewise, if you do "setp <someparam> <somevalue>" then exist halcmd, that param still has the value you gave it
ok, got a typo and dropped out now, will re-enter with halcmd -kf
cool, i was in midst of adding all the functions and could just continue
tzak: if you look at the pastebin of the terminal session, just change 'halcmd -f' to 'halcmd -kf'
i tried what is in the paste bin and it seem to be working
tzak: ok, try the enable , begin by looking "show pin m5i20.0.dac*"
tzak: then try to enable dac0 with 'setp m5i20.0.dac.00-enable 1'
tzak: and see if the real meter shows the amp is enabled
is that going to be on pin 1?
tzak: for pinout of m5i20 see Hal Users Manual section 6.5
pin 23 connector P2 is dac-00-enable output
(theres differences between mesa's naming and hal naming )
so i am going off the 5i20 pinout
yes, other breakouts or amps we wouldnt know about unless you tell us :)
tzak: going to put you on audible now, the irc client should read me your messages ;)
ok so i tried turning the enable on/off true/false and the voltage stays at 3.00v
'setp m5i20.0.dac.00-enable 1' not true not on but 1 worked for me
maybe try it unloaded ( not connected to device )
make sure that the appropriate hal functions are in a thread and that threads are running
jmkasunich: i hope the pastebin did that http://pastebin.ca/916503
looks like it does
any reason thread tt2 is a 1 second period? seems very slow
just what awallin had
yeh a billiion
actually, you don't seem to be adding anything to that thread at all.... why have it?
unloaded?? i have it into a breakout board and the only thing runnign is the computer everything else is off
tzak, then its unloaded unless the breakout is suckin it down/nailing it, try another output
jmkasunich: yes, its useless, just what was in the orig scripts
ok i will try m5i20.0.dac.01-enable
use a meter to measure FROM your pin under test TO connector P2 on any even pin ( all those are GND)
i checked all the enables 00-03 and they are all at 3v when i switch them on and off
ok, check P2 pin49 to GND to see if your supply voltage is ok
and when you say "the only thing runnign is the computer everything else is off " I'd like everything else disconnected, at least the one wire under test
i checked all the enables 00-03 and they are all at 3v when i switch them on and off same result
i took the ribbon cable straight and measure pin 49 to gnd and had 3v then did the enable and its still 3v
ok, the outputs dont do what your ask... is your m5i20 working on 5V or 3.3V? ( what do you measure on pin 49 to 50)
(49 = 'POWER' 50 = 'GND')
soory my mistake pin 49 to gnd is 5v and pin 47 is 3v
good night all
anything else to check ?
tzak: what does hal think about this? type 'setp m5i20.0.out-00 1", then ask by typing 'show pin m5i20.0.out-00'... does hal say it's 1?
then dbl check with type 'setp m5i20.0.out-00 0", then ask by typing 'show pin m5i20.0.out-00'... does hal say it's 0?
ok i will check
if hal says its what hal set it to, and the meter sez no, then the hardware is broken
or not connected
? did the 'other end' come off?
ok so setp m5i20.0.out-00 1 = true
setp m5i20.0.out-00 0 = false
yes, do that , each time checking what hal thinks and what the meter sez
which pin is out-00
you'll need this for reference, the Hal Users manual, in section 6.5 it shows out-00 is connector P3 pin 33
on your system the manuals should be on the menu bar Applications | CNC | HAL_User_Manual
yeah i found it was pin 33
still no change in voltage
and does hal say it changes? this is important
yeah it says its false =0 true =1 but the voltage stays put
last idea: can you check the end of the cable going into the breakout , without connecting it to the breakout?
i tried that on the enable signal but i'll try on the out-00 just to make sure
still no luck the voltage doesn't change
sorry, i think the board is not working
well at least thats then end of that goose chase..
would mesa be able to repair it?
dunno, i cant see the board so dont know if drivers are socketed or what
they made it in the first place, so they can certainly repair it - however, the cost to repair might be high
there are no drivers, so if it is really fried then the FPGA probably needs replaced
from what I have seen scroll by, I don't trust the testing that tzak has done
what dont you trust ? is there something else i should check
it sounds like you are following instructions without really knowing what is going on
whenever that happens I don't trust the results
well im not sure what else i can do .. i had this thing working it was running programs a week ago without any issues
"working" is the big picture - that means many little things were all working correctly
when the big thing stops working, the thing to do is to track it down to which little thing has stopped
tzak: do any of the onboard leds turn on? ( a gross indicator )
yeah the three of the led are on
which 3 ( please refer to the pdf on your desktop )
yeah the big thing that happend was a amp cooked
the 5i20 board has many functions - digital inputs, digital outputs, encoder counting, and ADC/PWM outputs
<tomp> with the current halcmd setup you recommended only CR8 is on
Good Day gents..
i gotta lookup cr8, but am slicing chicken now
wait no chgs should be needed to configs or jmprs
if anyone is interested I have 6 pairs of Yaskawa 200w driver / motors pure servo speed driven..
what is max speed? how large (physically) are the motors?
im going to check what leds are on when i start emc
they are new in the box.. the are the CACR series.. let me get the specs...
good idea, look for diff b4 + after
so with emc running cr2 cr3 and cr8 are lit
tomp, tzak: unless you have some documentation that says what the LEDs mean you are wasting time
I don't think any hal pins are connected to the LEDs
well here is something else .. emc is telling me that the xlimit is tripped and its not even connected
ok.. the motors are ( face plate ) 3" x 3" and 6" deep.. max speed is 3000 rpm, 4000rpm max, these are ac servos ( brushless , if needed all specs can be posted.. )
that in itself doesn't mean much - a properly designed limit switch setip is failsafe, so it will report tripped if disconnected or broken wire
ottos: with encoders? with brakes?
you (tzak) are the one who designed the limit wiring, right? you should know how it is connected and what it should report if not connected
ottos: thanks - I'm looking for something in the 100-200 watt range but fast, to drive an engraving or grinding spindle
how much are you looking to get for one motor and drive?
<jmkasunich> yeah when i fired it up yesterday all the limits were tripped which never happened before. Then about five minutes later my amp burnt so i have no idea whats going on
who designed the system?
but you don't know how it works?
I bought them at 200$ a pair.. and newer used tham.. they are still the original packaging.. with manuals.. I can offer you the same.. how many ware you interested in..?
my limits were working fine
ottos: one, maybe
im just learning how all this goes together so far i have managed to get thing running so when unexplained errors start happening im not sure what to do
the main issue is speed - I will be using belts to increase, but my target is 15000 RPM or more, so a 3000 RPM motor needs more of an increase than say a 5000 RPM motor
Bah, I have way too many PS's (type, not qty)
ottos: where are you? usa or elsewhere?
canada ( toronto )
tzak: unless you can get one input or output to work, the board is not useful
if NO inputs or outputs work, either the board is totally fried (maybe) or you are not testing it properly (more likely)
I thought you said encoders still work
yeah they were just working until this limit switch thing just started happening like 5mins ago
if things are changing then get power off these expensive things untill you know its safe to power up
I thought you said the limits stopped working yesterday, then a drive smoked, but the encoders were still working
you've been talking to tomp for a LOT more than five minutes - if things that were working before you started talking are not working now, that means you probably broke them poking around with the meter or something
ottos: I pm'ed you, did you get it?
here is what happend .. i started the system yesterday.. i go emc thinking that all the limits were tripped then the amp cooked i shut everything off check it all took out the bad amp and fired it up again. Now the encoders were counting but no servo. So between the suggestions that tomp has been getting me to do this same strange limit switch problem is happening . Hopefully that make some sense
chicken schnitzels done!
tzak: it makes some sense, i'd look at the m5i20 with nothing connected to it, not even the breakout. just one flatband , the one on P2 and see if i cloud get output 8 (for instance) to do what i ask.
if not, i'd say the m5i20 was bad
ok tomp thanks for all the help i have to run now enjoy the chicken
best of luck, we're all learning this stuff :0
yeah i hope i didnt sound like a complete moron .. i made it this far didnt i
yep, i make a practice of being a moron 1st thing every day, then it doesnt bother me later ;)
ok cheers have a good one
are the 200w's a good deal?
I dunno - I really need something faster than 3000 RPM, so I decided to pass
so many projects, so little time - I can keep busy for quite a while without a motor
and now for something completely different: tomorrow, BBC7 will start the 2nd series of "the Spaceship" i find it very funny, it's about her majesty's ship, the 'Really Invincible 3'
is that anything like the unsinkable titanic?
yeh, if you dont the joke in the name, dont listen to the series
"dial.com", guessing not
rayh: did you see http://jmkasunich.dyndns.org/cgi-bin/blosxom/index.html
Wow. That is great stuff.
it was a fun trip
I got a call recently about fitting a K&T 5 axis.
Fantastic. We are proving that we can make real machine tools work.
seems you'll be joining the real internet soon ;)
(actually internet-experience ;)
Yep I was a member once back in about 95 but got lost someplace.
I think modems where the thing up to ~99..2000
I've got a pile of dead ones.
I guess I fell behind about 8 years ago then.
It's what I get for living in the wilderness.
I bet it's nice there..
does it mean you plan to be more often online?
You bet. Someday we need to have a mini fest here.
Always on with my own dotted address.
Hence the server.
cool.. then I guess you'll want to look at irssi & screen ;)
Don't plan to put much on there.
irssi & screen??
irssi is a text mode IRC client
screen is a virtual terminal, which you can detach from, then attach later on
I think a couple of us use the technique, to keep the client always running
then one can connect from virtually anywhere through ssh
Okay. I plan to run my current LAN behind firestarter on the athalon box.
and see what's happening..
Sure. Right know I port knock to get on to one customer's network.
now not know
port knocking is fun
I see folk trying it all the time on my dialup.
Question about a mini HAL.
I need to write a script to check the status of a parport pin.
This is an EMC 2.1.7 based system.
So I've created the docs from source.
Are we good to go with the way that the HAL doc talks about starting a real time system, adding a thread, and such?
might be overkill to use HAL for that.. do you need RT?
No. Is there a simpler way?
small C application
or even tkio, or what that was called in emc1
direct access to the port.. needs sudo or setuid though
The parport is connected to a PMDX-131 board. But I think input IO is passed through anyway.
Without charge pump and all.
hmm.. maybe steve knows better
Do I have sudo problems if I start realtime on a EMC2 machine?
if run-in-place: scripts/realtime start
I got halrun to go
so I can just enter commands.
That looks like a pretty direct path.
boing... I am actually here, instead of just failing to adjust my user name since last reboot
so Ray, who did you have to threaten to get your link installed?
on PMDX-131 pin 10 is dedicated to E-Stop, all others are direct real time inputs but somewhat bandwidth limited by the optocouplers
He steve. If the 131 is powered but no charge pump, will I still see input signals from switches?
Okay. So I should be able to see the polarity of the limit switches.
wow $100 install + $35/mo is about what we pay for DSL
charge pump status has no effect on reading input bits of PMDX-131
Cradek - reading back, new office is almost there, hope to move in about 2 weeks, have electric, heat, water, etc. just need final inspect and occupancy permit
oh, and a little detail like getting Asterisk Voip PABX running since we installed ethernet but no phone wiring 8-)
hey can't you run phone signals over the unused ethernet pairs?
eh voip is better anyway
of course, if there are enough unused drops, but we did not plan dual drops in each outlet
I noticed that even though there is a place to specify the home location in the stepconf wizzard it does not put that value into your ini file
BigJohnT_: sounds like a bug to me.. please file a bugreport
* alex_joni heads to bed
rayh: glad to see you around again ;)
Glad to be around. Hope to be more often
* alex_joni sticks around 5 more minutes
rayh: I pm'ed you
it's just not close to the other HOME entries... hmm, now why do I get some experiance and not what I expected...
BigJohnT_: you're subscribed to the emc-users mailing list.. right?
follow jmkasunich's steps for the bugreport
no bug alex, just it was not where I expected it to be
in the ini file
BigJohnT_: he had a very nice way of putting things
1. I did this
2. I expected this
3. instead this happened
(would fit here too)
I saw that, I say if I don't get what I expected I just got some experiance
now I have to learn from it
I need to put that step by step on wiki it was very good
well, sometimes your expectations may be wrong, but other times your expectations are quite reasonable and the program is wrong
the task is to figure out which is which...
its the great existential question of life
at least the answer is simple
if you can get ahold of the author
hmm, I'm missing some piece of information... and thus drawing a wrong conclusion I must assume... I think it might be the min/max limit
my Y axis is 0 to -50 inches should the min_limit be 0 or -50?
that's what I have
hmm, I have home=0, home_offset=0.25 but after homing it does not move off the limit switch by .25
both search and latch vel are positive as in example 2 of the inegrator manual
what did I miss?
hang on, let me reread that
but I think you want offset to be negative aswell
and the home not on 0