jmkasunich_ is now known as jmkasunich
anybody here ever used the home to index
HAL: ERROR: signal 'Xindex' already has output pin
HAL:66: link failed
im getting this error
[03:29:08] <jlmjvm> http://pastebin.ca/709973
if i comment the x index out,it wont find the index when homing
what about line 62?
says it already has an output pin,but i cant find it
You might want to compare your file to a known good one
this was the 1 Jon told me to download and use
do you have to have the latch velocity set really slow?
I'm the wrong one to ask about USC stuff
jlmjvm, comment out those few lines, load EMC, and look at the pins and signals that are there
if you see XIndex, then you'll know what's connected to it, and you can find which HAL file does that
(you can also go to the directory where the files are and type `grep -il xindex *` to see a list of files with the word "xindex" in them
the -il means ignore case and only list file names (rather than showing all lines with the word), xindex is what it's looking for, and * is the list of files to look at (you could use *.hal instead)
gawd he's good ;)
i cant find it
oh - right
I'm not sure of the specifics, but there have been changes in how index is handled. I know it works now, but I'm not sure if those changes are in 2.1.x or if they're only in CVS at the moment
would be my luck they arent
that would explain why they were commented out
but i was told it would home to the index
one thing you can do - run emc with those lines commented, then run the following command:
(from a terminal)
halcmd show pin | grep -i index
do that and stick the output on pastebin
[04:01:48] <jlmjvm> http://pastebin.ca/710000
ok, I think maybe you're supposed to connect axis.n.index-enable to ppmc.0.encoder.nn.index-enable
lemme check a CVS config
well, try that. it can't work any worse than now :)
I can explain what I'm seeing as the problem though
notice the column in that pastebin that says "I/O or "OUT" for each of the bits?
I/O means the component/driver may read it or write it
OUT means that the component/driver will always write
(IN, not shown, means the copmponent/driver reads only)
you can only connect one pin that's capable of writing to a signal
if you allowed multiple writes, then the order of the functions would cause changes in the data (you never know the value if multiple things can write to it)
the one exception is when all of the pins are I/O or IN pins - I/O pins know how to work with others, and IN pins don't affect the signal data
so the error you're getting is because ppmc.0.encoder.00.index is an OUT, and that HAL file attempts to connect it to an I/O pin
got it? :)
so it should connect it to an IN ?
none of those pins are ins
but there are teo I/O pins - I'm just not sure they're the right ones.
I think they are - I think index-enable turns on the index finding feature, and index is just the state of the index input when the scan was done
its not finding the index
ok. this is a fairly involved setup, and I'm not sure I can help much more tonight
you may want to ask about tomorrow, or email JonE about it
im sure i will be in touch
already gave up on the closed loop mode,was supposed to install in the morning,the index should be working like i was told it would when i
I'm sure it'll work (Jon did the final touches to the FPGA code just before Fest) - I just don't know the specifics of how to configure it correctly
i shoulda went with something that i could have made work,like a breakout board
is this the stepper machine or one with servos?
the servo machine
I'd email Jon (or the user list) with a description of the trouble you're having. Closed loop operation is the default with EMC, and I know it works with the USC
index is a bit harder to set up, but it also works if configured properly
there are things that must be set up on the actual machine, such as PID tuning
shouldnt you just have to enable index,and have an encoder with a z channel
so it's not possible to ship a fully-configured setup to an end user, unless the machine is part of it
no, there are other setup options
look in the user manual about homing, I think there are about 8 options for the homing process to use
4 that i saw
ok, 4 then :)
im using #3
ok - I don't know the homing setups too well, that's why I'm saying that I don't think I can help
there are others that know them much better
im gonna take the usc off the panel and put a breakout board in its place,then i will know if it will home to the index
I don't think that would be constructive
if 1 works,it gets the job
if neither works then its something else
you can't use the USC for step generation and a breakout for index
you'd be limited to the step rate that software can generate if you just use a parport
this will only be a 100 ipm machine,software can easily run that,im doing that on my own mill
the speed of the usc would never be utilized
machine has 45v servos,at 1200 rpm,with the 30v factory ps you might get about 750rpm or 150 ipm,but would never be able to move all simultaniously at that speed
jlmjvm: removing the usc is stupid..
did anyone see tomp around ?
does anyone use insert tooling
hmm, it is near 5 am est
chances are i will not get a very wide set of answers
do you measure your wear
on the insert?
or use like, a cut log
I dont bother too much :p all I do is one off stuf
apparently there's some philosophy that says with inserts you can measure your wear
and like, when you decide an insert is dead
and chart it
and be able to decide before you ever start cutting how hard you can go
based on your datalogging
i.e. how much difference an increase in RPM will make in the life of your cutting edge
it seems like a really neat and simple process
we have a microscope at work, i think i'm going to try it out next time
start logging my cutting data, cutting edge, etc
just to see what comes of it
The_Ball_ is now known as The_Ball
anyone want to play "guess the feed rate"?
let's start easy: F2
darn, nobody wants to play
2 microns per year
it's a regular inch machine
you think I would try to trick you?
well F2 gives a feed of zero but that's not the punchline yet
how about F3.
3 ipm! it's obvious
how about F3.00
F3.00 causes it to cut 30 ipm IN THE WRONG PLACE
the French would expect 300 ipm
(this is not emc of course)
I had some cuts with plunge of F2.00 and cut of F4.00... it actually did cut at 40 for a while before I stopped it
not really, mostly it melted the half inch plate to the cutter
use a sharper cutter next time
didn't ruin the work though, since it was cutting in the wrong place!
cradek: going to start the retrofit soon?
what a insane control :)
jepler:does 2.1.7 have the same homing to index as the cvs
jlmjvm: there are some driver differences between 2.1.7 which may affect homing. remind me which board you are using?
I think 2.1.7 has all the known index fixes in the various drivers
does that use 'loadrt hal_ppmc'?
I think it can also depend on the version of the firmware chip on the board
dont see 'loadrt hal_ppmc'?
I think that must be the driver name
[14:00:43] <jlmjvm> http://pastebin.ca/710373
'ppmc' is in the name of a lot of things in your hal files, right?
heres my ini file
yeah this would be a hal file thing
dont think so
it looks like hal_ppmc is the same in 2.1.7 and in the cvs development version
'grep loadrt *.hal' in a terminal in the directory with your inifile and check whether hal_ppmc is there; I'm pretty sure that's the driver for the usc board.
[14:03:51] <jlmjvm> http://pastebin.ca/710376
[14:10:19] <jlmjvm> http://pastebin.ca/710380
this is what it shows when i do the grep in terminal
univstep_load.hal:loadrt hal_ppmc extradac=0x00
OK, I was right that the driver you're using is hal_ppmc
find out from jon elson whether home-to-index is supposed to work with the firmware that is on the board -- I think there's a small chip on it which shows the firmware version number.
i was told it would home to index
it's new, so it "should work"
only delivered last week :)
my first pastebin shows how im tring to use index
[14:14:58] <jlmjvm> http://pastebin.ca/710373
my ini file
what's the symptom?
ive tried homing option # and #4
never finds the index,just keeps moving
have you hooked the ppmc's index-enable to the axis index-enable?
no the index connections are commented out
I notice in one of the hal files you showed on pastebin the lines were commented out and probably refer to the incorrect ppmc pin names
... and wrong
yeah, what he says
they were commented out when i downloaded
net Xindex ppmc.0.encoder.00.index-enable => axis.0.index-enable
same for Y,Z
ok - I wasn't positive that index-enable was the right one to use, but I thought it was
jlmjvm: complain to your vendor then
* SWPadnos tried to help a little last night
the files in the 2.1.7 release are right
a look at 'show pin ppmc*index*' while emc is running will verify the names that need to be used
if you got these from elson's site they are outdated and need to be fixed, or preferably, deleted
they are what i was told to use yesterday
* alex_joni just started up a new robot :)
"Drop Your Weapon"
just wondered if i was doing it right on my end
</end robocop references>
there are only 2 homing options that use index,correct?
SWPadnos: this is 9-axes
#3 and #4
6-robot, one longitudinal to move the robot around (8m long)
but no machine guns :)
and 2 rotating tables (500kg max each)
oh, a big robo-DJ ;)
play "misty" for me
I really love it when I'm up until 4:00 AM working, and the neighbors have a construction crew building them a nice new back deck
and it's even better when my wife has left the bedroom window open
next time you'l think twice about working so hard
alex_joni:have you ever used home to index with emc?
does anyone here have a machine that homes to index,if so i would like to see your ini file
jlmjvm: we said above what the problem is in your hal file...
09:17:06 < cradek> net Xindex ppmc.0.encoder.00.index-enable => axis.0.index-enable
09:17:09 < cradek> same for Y,Z
you are missing the index-enable hookups because your hal file is old
the correct connections are in the file that came with emc 2.1.7, you can use that OR add those three lines to what you have
k,lemmee add those in
they're the same in the univstep TRUNK config - commented out
03jepler 07TRUNK * 10emc2/src/hal/drivers/hal_parport.c: add reset to the 4 output pins on the 'control port' -- not yet tested
and the wrong names
index instead of index-enable
and they're the same on Jon's site
well, that's very discouraging.
oh hey - it looks like it's right in the 2.1.7 release (at least for ppmc), but not in TRUNK
yes I nagged him before 2.1.7 release and he fixed it
yep. ppmc for 2.1.7 has index-enable connected, but univstep and univpwm don'r
by golly it worked
many thanks guys
03jepler 07TRUNK * 10emc2/src/hal/drivers/hal_parport.c: it helps to write to the correct port
im gonna celebrate with a fresh cup o coffee
have a dnout too
err - donut
bring enough to share
I would never do that!
reminds me of holiday in the states, dunkin donuts oooooooooooooooooooooooooo
(I'm still eating the granola bars)
heh - much better than Road Chef or that other one you guys have
those are awesome donuts
seeing this work was like a shot of adrenaline
I used to live in a development right behind a shopping center. After a while I remembered the item codes for their store-made donuts
Road Chef and a Little Chef over here, but they dont have the donuts
ah - Little Chef - that's the one (they have the green logos, right?)
or are they the red one
hmmm. there's a green one I remember - do you think that's just the BP stations?
BP is rebranding away from green
ah. I haven't driven the motorways since last December (only took the train this last trip)
SWPadnos:where can i order some ssr,s that i can use for a vfd,they need to be able to switch dc
I think the series is D2W or something
note that standard box-shaped opto-22's don't fit on the USC
those are for ac
the catalog page says AC/DC I think
Load Current @ Voltage3.5A @ 280V AC
you need to select those for yourself - I've never gotten any that fit the board
(I can sell you several that are too big though ;) )
also note that you must populate the fuse sockets - those are in series with the relay output
i saw that
This is from an email Jon sent me a couple of years back:
The relays I use are Crydom D2W202F, available from Digi-Key or Mouser. There is also a
5 A version. The fuses are to protect the SSRs and the board traces from overcurrents.
Use "Picofuses", such as the WK1088CT from Digi-Key for 2 A.
so i need to get 2 more like i have that fit,2 more external relays to switch the 24v from the vfd,and 2 more pic fuses
could be. I don't know what you have or what you need, so I can't really comment
i gotta give a vfd the fwd and rev commands,and it is 24v
what is delvin?
[17:56:54] <SWPadnos> http://en.wikipedia.org/wiki/Delvin
SWPadnos: very funny
never heard (or saw) the term before now
some sort of split between debian and vindows?
Guest576 is now known as skunkworks_
or typoed "delving"
skunkworks_: whatever happened to the dude trying to jog his tripod?
did he succeed? (there was a thread in the forum you "moderate")
he came on IRC and got it workinh
did he now?
I only was here ther first? time, when he was set on track
he came back.
[18:09:09] <skunkworks_> http://www.cnczone.com/forums/showthread.php?t=43872
did you hear what he was going to use it for?
did he ever post his kinematics file?
I bet I know why his Y and Z are swapped
ok ok - I will tell.. - he wants to fly a ghost around his yard for halloween.
I could ask him to post it.
that has to be a first for emc ;)
"unfortunately teleop mode jogging is busted in the CVS version of AXIS right now."
it might help. a simple error would have been to forget that Z+ means retracting the wires, not loosening them
(ie, the math would be worked out above the XY plane, but the hanging thing is below hte XY plane)
so you end up with a left-hand coordinate system
in that case, if you get it so one axis is aligned, then the other two are swapped
cradek: I wanted to bounce an axis feature idea off you... I would like to add cursor co-ordinate display in the non 3d windows
cradek: I have been hand editing gcode and using axis as the display... some idea of mouse pointer positioning would help
either on the display or the status bar would work
teleop jogging in AXIS works fine for me
fenn: I'd appreciate a bug report here, if you think there's something wrong
not sure what fenn was smoking :)
I think he was smoking PUMA or SCARA configs
try jogging C in the scara config
for some reason when i give an m3 command it turns my power enable contactor off,m5 turns it back on
oh btw that comment was before i fixed axis to only send the first 6 coordinates to teleop
I can't run the scara config
it works now for other configs
* fenn is getting various different bugs mixed up now
AXIS needs work in the case where the "axis numbers" aren't the same as the "joint numbers" -- for instance if you have an XYUV machine with joints 0..3 through kinematics, you won't hava "jog joint 3" button when in jointspace
i think i remember seeing 1 2 3 5
1,2,3,5 is really odd.. maybe 0,1,2,4
0 1 2 5 sorry
the scara config
aren't the joints numbered 0..3 though?
Instead of having one set of buttons whose labels are changed, it should probably be changed to have two sets of buttons in one frame, which are swapped in and out depending on joint vs cartesian mode
SWPadnos:jon had some ssr,s that will switch dc,sending me a couple
'lo, anyone know if fsck generates a more verbose log than the one in /var/log/fsck/checkfs?
according to the manpage, fsck.ext3 can be run with the "-v" flag to be more verbose
hmm, I was thinking for a run it had already finished
it printed some text to the terminal when starting up, though I can't seem to find that anywhere
if checking the root fs, the root must be read-only, so almost by definition fsck can't log anything except in dmesg
unless you log to a different fs
oh, I see, so it dumps it in dmesg
hmm "exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0" doesn't look to healthy, fsck tried to recover some log too :/ , later on "Buffer I/O error on device sdb1, logical block 4704"
dying disk maybe?
it seems to have lowered the ATA speed between lots of failures
lerneaen_hydra: how old is the drive?
not that old, a couple years
in the middle of the bathtub curve
lerneaen_hydra: what brand?
seagate, 250gb , sata
seagate has a utility you can test the drive with
do you have SMART enabled?
yeah gonna run that after the backup is finished
probably, haven't looked at the logs
where are smart logs placed in linux?
lerneaen_hydra: 'apt-cache -n search smart' has a package called smartmontools - control and monitor storage systems using S.M.A.R.T.
of course if your disk is dying, trying to install a new package may not be the smartest thing to do
and wont do much good AFTER the fact.
this is not the root disk fortunately
on my system there is a file /proc/ide/hda/smart_values but as I don't have root access on this machine I can't see what's inside
sudo su =)
hmm, apparently I only get smart value for my PATA disks, not my SATA disks
here's what I get from smartmontools on a different machine with ubuntu gutsy beta on it: http://emergent.unpy.net/index.cgi-files/sandbox/smart-log.txt
smartctl --smart=on --all > smart-log.txt
jepler; had a couple errors then?
[19:44:39] <JymmmEMC> http://www.cyberciti.biz/tips/linux-find-out-if-harddisk-failing.html
ll'll be damn... you can use SMART behind a 3ware card.
well, it says SCSI.
^ the small drill chuck I use on max
maybe could be useful for yours too
cradek: I'll look...
good night all
does anyone know if emc has s curve acceleration?
I believe that accel is trapezoidal, and position is an s-curve
there's no jerk limiting at the moment
velocity is trapezoidal
err - right
that's what I meant, really :)
so a position plot looks like an s curve
k,someone in the mach group said it had it
btw,everything is working on the usc,will install in the morning,thanks for the help
good to hear it
i was under an unreasonable timeline,but everything is groovy now
should just have to hook up motor,encoder,limit switch,and ac power when i get there
heh - that'll be amazing if it works out ;)
murphy has a habit of creeping in at the most annoying times :)
its gonna work,ive done a hurco before,and im pretty determined
that's the right attitude at least
gung ho joe
give her shit... whats the worst that can happen.... you let the smoke out and it has to go back to the factory to get recharged...
boy - Nortons is a bloated pos
uninstalling it has taken atleast 15 min so far. It has to tell you everthing it is uninstalling. right now it is removing the internet worm protection.