03alex_joni * 10emc2/configs/stepper/stepper_mm.ini: fixed bug #1406189 , thanks floh for spotting it
SWPadnos is now known as SWP_Away
jmk_dinner is now known as jmkasunich
hellooooo. I really need some help from anybody
what kind of help do you need?
tonyp: what's burning?
I have a z axis problem on a router that wont repeat
what emc are you running?
steppers or servos?
it seems to drift all over the place, I just upgraded to the 4.38 today and it is steppers
and a dozen other questions that we need to understand your situation
4,38 is supported by Paul Corner, we know nothing about it
he has a mailing list, URL in a minute
[01:19:54] <cradek> http://ourproject.org/projects/bdi4emc/
thats the one
When running a z-.010 z0 z-.010 z0 and so on it seems to miss signals up and down and ends up drifting all over the place.
sounds like lost steps, but I really don't know the BDI-4.38 at all
this is not a release problem. I have had this since I built this a year ago
the same problem?
it did the same with rc46
yes the same problem.
why didn't you say so?
I dont type fast enough
have you tried reducing accel?
yes, replaced the gecko, rewired with a 3 foot cat 5 cable and changed to the 4.38 just to make sure the release wasnt the problem
I asked about accel, not about all that
yes was the answer to accel
what was your original accel, what is it now, does it work correctly for any value of accel, even very slow?
are your TRAJ accel and per-AXIS accel all the same?
If you want to use newer versions of emc, you're better off building emc 1.2.0rc1 or emc2 from CVS, not upgrading to bdi 4.xx.
per-AXIS accel is ignored for coordinated moves on those versions!
jepler: that may be so, but he has had the problem for a long time...
bdi 4.xx seems to lack a certain amount of, er, organizational transparency, right now.
no difference even on bdi rc46
tonyp: what are your accel limits, in the TRAJ section and in the AXIS sections of the ini file
I also used smaller and bigger
how much smaller and bigger?
1 to 20
what is max velocity?
is this an inch machine or a mm one?
1.5 or 89 inches per minute inch
X and Y work fine?
When I plug the z into the x or y it works great and the problem moves to the other axis
jmkasunich: right, but if he wants to compile emc1 from source, he's screwed with bdi 4.xx.
sounds like a driver problem
just replaced the gecko yesterday
with a 210 microstep
put on the cat 5 cable per marris's suggestion
tonyp: that isn't clear: when you plug the Z drive into the X motor? when you plug the Z motor into the X drive? Does the problem move with the motor or the drive?
the problem moves to the next axis
the z motor works well when plugged into the x drive
the x drive then starts to drift
huh, you just contradicted yourself
been pullin my hair out
you mean the x axis starts to drift right?
let me try again
you just said: the Z motor works on the X drive, then you said the X drive starts to drift...
did you mean the X motor starts to drift?
* fenn is thinking a faulty electrical connection or electrical interference
yes when plugged into the z drive
so the problem is either the Z drive, or the Z signals from EMC
what does the cat5 cable do?
or the Z signal wiring.
2 sets of twisted pairs that are sure not to be cross with the wrong wire
fenn: reduces interference because it's twisted
cat 5 is probably an attempt to avoid noise
you are using software generates step pulses out of the parport?
(no Pico systems boards or anything like that...)
I thought the z signals were the problem and that is why i got the 4.38
do all axis have the same scaling?
hard wired to the gecko \
no on the scaling
what is the scaling (INPUT_SCALE, OUTPUT_SCALE)
for X, Y, and Z
x axis is 1495
y axis is 2450
* fenn prescribes hal and a square wave signal
z axis is 1350
take two measurements and call me in the morning :)
I dont even know how to do that
tonyp: 1350 steps per inch?
fenn is being difficult
are the Geckos microstepping?
only the y axis as the screw is 20mm lead
1350 microsteps/inch seems low, that means one turn of the stepper shaft (2000 microsteps) is about 1.5 inches travel
The scale is right to an indicator
wow 20mm lead? what diameter?
so you are running X and Z at 200 steps per revolution?
moving 1" =1"
how big are the steppers? nema 34?
nema 34 at 640 oz in
gecko current is set right for the motors?
plenty of power. will bend the y axis if it hits the table
two of the axis work great with the same setup
1350 steps/in at 200 steps/rev = 6.75 turns per inch on the leadscrew
does that match what you have?
I dont have the actual numbers/ the machine is in the garage and I dont remember exactly. these are within 100 though
1.5" per sec max velocity = 2025 steps per second at max speed
2025 is certainly not excessive
5 inch/sec^2 accel means about 1/3 second to go from zero to max speed
also seems reasonable
depends on how much iron you're slinging around
I get no following errors. I have been using the machine for a long time now for 2 axis stuff but seems to be getting worse
what is getting worse? I thought Z was always bad?
running a trim router on a ball screw with smaller than a 1/4 cutter
usually not very noticable on signs
should be able to get way higher accels and velocities
I can on the x and y
you said: When running a z-.010 z0 z-.010 z0 and so on it seems to miss signals up and down and ends up drifting all over the place.
do you have another computer to test with?
and now you say "usualy not noticable"
which is it?
he meant it loses steps i think
if it drifts both ways 10 thou in a program that is cut .040 deep with a 1/32 cutter
I'm asking HIM to clarify
doesnt seem to lose steps as much as get a bad signal. I can see it happen on very slow feed rates and hear the motor move too far
too far? thats differnet
in both directions
now I'm starting to agree with fenn
that is what I mean by drifting
electrical noise causing extra steps
it is not consistant
i was starting to disagree with myself right then actually :)
noise rarely is
does it do it with the spindle off?
it's not overshooting and backing up is it?
I just rewired it/ Where would the noise come from
noise has its ways...
does it do it with the spindle off?
cradek: no spindle on
Had something similar when I was over-driving the stepper - try lowering the current. Had a stepper that we thought would handle high current - cut the current in half and no more problem. (gaining steps)
tonyp: are you saying it happens ONLY when the spindle is on?
spindle doesn't matter
have you put a scope on the step/dir lines right at the gecko?
on or off
how do I cut the current in half. I am not very smart
cradek: I dont have a scope
you followed the gecko instructions for parallel, series, half-winding connections, and you selected a current setting resistor according to their instructions?
I got the resistor spec from Marris. But I can try a smaller current resistor if it will help
I have no idea if it will help or not
if Mariss thinks you have the right current, you probably do
that is what i thought. besides it works on the other drives x and y
they are set up the same way
are all three motors the same?
someone mentioned a different computer
ok, we can probably rule out motor current
don't go there yet
still on motor current, are you sure the all three resistors are the same value?
(measure with an ohmeter is best, verify color codes or labels is next best)
I will double check
if that is OK, then we can pretty much rule out current issues
* alex_joni is off to bed
are all three drives mounted next to each other? (in the same box or panel)
I bought them at the same time and they came from the same pak
yes they are mounted in line
the resistors? thats good
do you have a breakout board? or does the cat5 wire run straight to the PC?
straight to the pc
so you have a bare naked DB-25 connector plugged into the PC?
one wire from each cat5 pair soldered to the connector?
a db 25 plug in the panel to plug into
the control panel
you just said it went straight to the PC?!?!?!
one end of the wire is screwed to the gecko
where is the other end of THAT PIECE of wire?
geckos are wired to a male plug on the panel the cat 5 plugs into that and then to the computer
ok, see if I have this right....
only a plug between the gecko and the computer
a homemade printer cable from the back of the PC to the connector on the control panel (made from cat 5 cable)
what kind of wire from the connector on the control panel to the gecko?
stranded copper soldered to the db25
twisted or anything?
24 gauge not twisted
how long is that part? (the connector to gecko part)
8 inches max
3 feet to the computer on the cat 5
what are you using for +5V?
computer power supply
you have two pairs per axis in the cat 5, right?
step/+5, and dir/+5 for each axis
total of 6 pairs? (which means 2 pieces of cat 5)
no only one piece of cat 5 with four twisted pair
one pair per axis step and dir
how can you send 6 signals (X, Y, Z, step and dir for each) thru 4 pairs?
x step and x dir is only two wires and so on
so you aren't actually using pairs
that is not good for noise, and is probably not what Mariss had in mind
I really dont think it is noise because it is only on one axis
so your pairs are: (Xstep, Xdir) (Ystep, Ydir) (Zstep, Zdir) (+5v, +5v) ?
and when you get to the control panel DB-25, how is that pinned out?
same as the wire pairing?
then inside the control panel, you have three pieces of 24ga carrying +5 to the three drives?
and 6 more carrying the signals, 2 to each drive?
only on on the +5 and jumped to the other drives, yes to the 6 more carrying the signals
what is the order of the jumped +5? connector --> X --> Y --> Z, or something else?
z jumped to y jumped to x
z is first in line
there are two +5 wires in the cat5, right?
and two +5 pins on the connector?
uh, what about GND?
I am using a seperate power supply the +5 doesnt come from the cat 5
fenn: no ground, the geckos use common +5v, and pull low to turn on the opto leds
tonyp: I thought you said you were using the computer power supply!!?!?
a seperate computer power supply from a junk computer sorry for that
that is located inside the panel
[21:04:49] <jmkasunich> so your pairs are: (Xstep, Xdir) (Ystep, Ydir) (Zstep, Zdir) (+5v, +5v) ?
[21:04:58] <tonyp> yes
so is there a (+5, +5) pair in the cat 5 or not????
no, I am sorry for that confusion. I am not very good at this. only step and dir and limit and ground in the cat 5
* fenn hides
+5 coming from the seperate computer supply
so your pairs are: (Xstep, Xdir) (Ystep, Ydir) (Zstep, Zdir) (ground, limitsw) ?
fenn: I have been trying that for awhile
and in the panel, the ground from the extra power supply is tied to the ground pin of the connector (the ground wire in the cat5)?
is that extra power supply running anything else?
that is not the best setup...
the extra power supply's ground is internally connected to its power cord ground
so you have a ground loop, from the PC power cord ground, thru its supply into the PC case, thru your cat5 ground lead, to the spare supply ground, thru its case to its power cord ground, and back to the wall socket
are the PC and the extra power supply plugged into the same wall socket or power strip? if not, that will make it even worse
yes all is run from the same powwer source.
one power cable to the machine and divided from there
the amps and spindle are on the same power cord tho, right?
the whole machine is run from one power cord
the only thing that isn't is the monitor for the computor
so noise from the main gecko power supply, and/or from the spindle motor, could couple into the ground of that extra power supply
so getting back to the wiring from the connector to the geckos
wouldn't that effect three axis's
there are six signal wires, each one goes from one connector pin to one gecko, right?
then there is a wire that goes from the ground connector pin to the extra power supply ground
and one from the extra power supply +5v output to Z, then jumps to Y, then jumps to X
what is the total length from connector to power supply to Z?
so the supply is right next to the connector
below and to the left
but very close
inside the same box
within 3 inches
if the total length is 12", the connector to PS must be about 6, and PS to gecko another 6
you cut the power supply wires short I guess (normally they're a foot or more long)
yes I did
trying to keep things neater
this must be an older supply, the new ones won't even start up unless they get a signal from the motherboard
what's the preferred way to do twisted pair from a parallel port? each signal twisted with +5v?
fenn: depends on the load, for a gecko thats what I'd try to do
some of the very old ones (XT, AT) are really noisy and need a good load to work well
cradek: good point
tonyp: is your auxiliary power supply super old (like it has the big red toggle switch on the side?)
from a 486 i think
so its switch is one of those four-wire jobs on a pigtail from the supply?
those aren't so bad
tonyp: no switch at all then?
the 5 volt I get from it is solid
a solid 5 on a voltmeter doesn't mean it's clean
no switch. plugs into the 110 and comes out 5 and 12 volt
just like that, no startup signal needed?
does the fan in the PS run as soon as you plug it in?
cradek: had an electrician check it for me
runs as soon as it is plugged in.
did I hear right that it's only ever the Z drive that screws up
only the z
then the z runs great on the x or the y drive
why not get out the soldering iron and switch the X and Z wires right on the back of the computer
the x and the y drive screw up when hooked to the z
you'll know once and for all whether it's software
tony is being unclear again
* cradek ships jmkasunich a beer
not software. I just tried two different ones today
the Z MOTOR works when hooked to the X DRIVE
the X MOTOR screws up when hooked to the Z DRIVE
yes the same drifting as seen with the z
the X DRIVE never screws up, I don't think?
the x or the y drive never miss a step
this is why I said "wrong": <tonyp> the x and the y drive screw up when hooked to the z
ok so what I said still holds
you could rewire the connection to the parport
* jmkasunich is sorry to be such a nit-picker, but getting the facts absolutely straight is critical for this kind of stuff
and if the problem stays on Z, you know it's not emc
on the mill's Z
the z drive is the problem. when I unplug the z motor from the z drive and plug the z motor to the y drive it works perfectly
tony: its not neccessarily the Z drive
the z drive OR the power supply or the z wiring or your parallel port or emc
when I plug the y motor into the z drive the y axis starts acting erratically
it could be the Z signals from the PC, or the Z wiring in the cat5, or the Z wiring from the panel connector to the gecko, OR the drive itself
my suggestion will narrow it down quite a bit
switch the Z and X wires at the parport connector at the back of the PC
so keep track of what you know it's NOT
so far: the motor
is the problem stays on the Z drive and motor, it must be wiring or drive
it it moves, then it must be the signals
I replaced the gecko. I replaced the par port cable with the cat 5 and I replaced emc with a different version since yesterday and still the same problem
I get that!
ok, probably not the gecko
ok, what haven't you replaced? the power supply and the internal wiring
and since that is neither shielded nor twisted, it is vulnerable to noise
isnt there still a ground loop in there somewhere?
instead of just switching the motor, switch the wiring at the connector. if the problem moves it isthe par port
or the computer
would it be easier to just use a different computer. I can put emc on my laptop in 30 minutes and try that rather than the soldering?
laptop parport have their own issues
laptops are bad at running emc for several reasons
3.3 volt logic makes for a crappy parport, lots of people have had problems with that
I have used it before on a smithy mill with out any problems
also, the power management crap in laptops wreaks havoc on realtime code
sometimes you can get lucky, I know some people use laptops for emc, but mine didn't work
it's just luck
tonyp: laptoms sometimes work fine, but when you are troubleshooting, you don't want to add more variables
rc46 power management didn't work
* fenn would like to see a list of laptops that work
all power management is always disabled with realtime
you don't want your computer to suspend when the mill is running!
so please give me the best next step. I will do that in the morning and check back
swap signals at the parport end of the cable
ok it will be done!
if that doesn't work, draw a schematic diagram and send it to us
and check all your ground connections once more
(how do you have two power supplies connected together and no ground loops?)
yeah a picture would be great
On another note. I was running a smithy mill in the middle of a program and the laptop froze up. it finished the program before shutting down. pretty amazing.
fenn: he has a ground loop, no getting away from that
can you put toroids on it or something?
Ive been scared for a year
tonyp: don't give up
you've entered the spooky realm of analog electronics
you're in a twisty maze of ground wires, all alike
(you have been playing too many online rpg's)
it's from adventure
probably means I haven't been playing enough
ok, just put the kids to bed. even with the problem i am blessed. thanks for all of your help and I will check back tomorrow.
me too. gnight folks
still early here!
man what a mess
"Configuring bicyclerepair" (ubuntu install)
he needs a nice 3 watt 5 volt wall wart
yeah, his PS is probably no good for that.
and his pairs are configured all wrong
there should be grounds in them or something
(gnd, step) (gnd, dir) x 3
then at the control panel, each gnd goes to the gecko +5v pin thru a 0.1uF cap
yeah couldn't hurt
bypass caps are good superstition
tie all the +5V pins together thru 10 ohm resistors, tie all gnds together thru 10 ohm resistors, and connect the tie points to +5 and gnd of an isolated 5V supply
that way for AC signals, each pair goes directly to the geckos, and the supply is decoupled by the 10 ohm resistors
the resistors would protect you from shorts somewhat too
SNIFF SNIFF what's that smell??
ah, the aroma of toasted resistor
a debian install sure has to screw with the packages for a long time after reboot
I have only used redhat and freebsd recently and they reboot and are done
SNIFF, SNIFF.... smells like a SEI... 1/4 watt... thin film.... 10K, no 6.8K...
no 6.40K! oh no!!
nah, thats a 6.8K, the 6.4K has a touch more burnt plastic in the bouquet
hints of nutmeg and bat guano?
don't give away my eggnog recipe
I've become more accustomed to the smell of fried IGBT
burnt elephant snot
glad you stayed out of that wiring discussion?
I like the spell of soldering circuit boards (fiberglass not bakelite). But I have developed a twitch
my eyes glazed over
smell I mean
I get so frustrated with folks who don't get the importance of accuratly reporting things
yeah things like that require a lot of precision
jmk - I wasn't getting that from you at all ;)
precision in language takes practice
* jepler knows how to tweak cradek
its difficult you mean?
Im anal about apostrophe's and other puctuashion.
I got email from a user the other day, he was having problems with emc2 PID stuff
sounded like an absolute idiot
spelling, grammar, etc
you sure he was a native english speaker?
so I go into "rank beginner mode"
the idiot filter is only fair for native speakers
and he comes back (still with awful grammer) that he has an Masters in control theory, 35 yrs exp, and is working on his PHD
wonder who will write his dissertation for him
I eventually spoke to him on the phone
he has a russian accent, but communicates very well verbally, no grammer problems that I noticed (of course, you notice more in writing)
I'm just sitting here, laughing.
he's in the US, has written papers in english but must have had an editor/proofreader to check his grammEr
grammer, grammer, grammer
which one means the mother of your father or mother?
dictionary.com says grammer is a city in Indiana
well, isn't Indiana the state that says pi is three? so what do they know about it?
of course, now that I've intentionally mis-spelled it so many times, I'll probably screw it up forever
grammar, grammar, grammar
she sells grammar by the grammar shore
we like you anyway
dammit... I hate unscheduled perforations...
* jmkasunich looks for a band-aid
huh. I thought at first this new bug was an axis bug, but I guess he's saying it's an ini-parsing bug?
[03:13:50] <jepler> https://sourceforge.net/tracker/?func=detail&atid=106744&aid=1406189&group_id=6744
yeah alex "fixed" it by moving the comment
the value of an ini file variable is everything after the = sign
probably shouldn't be, but thats the way it works right now
ok, it's a feature then
is there documentation on the .ini format as used by emc?
its .. surprising .. to me that it works that way, but if^Wonce it's documented I'm happy to shut up
* jmkasunich looks
the only "documentation" is a comment at line 582 of src/libnml/inifile/inifile.cc
/* it matches-- return string after =, or NULL */
I'm pretty sure I don't care which way it works
although typically you can put comments there I think
comments are detected in "stripwhite()" which is called to strip leading whitespace only
if the text starts with (or ;) the entire line is dropped, but a after that is treated as just part of the text
[insert tcl insult here]
I assume there were some #s in what you just said? I got some reverse video text, between where two of them should have appeared
is that happening in my client? I thought I'd turned off all the stupid
I got that too, but I was embarassed to say so
cradek: invalid command name insert
cradek: invalid command name "insert"
cradek: when you type that into tclsh
 means that whatever is inside is interpretted as a command
so you tried to execute the command "insert" with args "tcl" "insult" "here"
oh forget it
for pete's sake
it was a joke
tcl is also too stupid to parse comments like that, I was being funny
* jepler considers earnestly telling cradek that he can (sometimes) write ;# and it'll be treated as a comment
sorry for the failed humor
[Global Notice] Hi all. We just experienced routing problems with a main rotation server. We're pulling it and checking with the sponsor. Apologies for the inconvenience.
03jmkasunich * 10emc2/src/rtapi/rtai_rtapi.c: fixed an ifdef that was disabling time measurements even on systems that are capable of making them...
jmkasunich is now known as jmk_sleep
[Global Notice] Hi all. We just experienced loss of bandwidth on a main rotation server. We've rearranged things accordingly. Apologies for the inconvenience, and thank you for using freenode.
SWPadnos is now known as SWP_Away
TorbaX_bed is now known as TorbaX
I made emc2
I have an error
with emc scriptù
after I installed axis
I did only the step suggest by Axis site
to install Axis on emc2
I am on emc2 path
when I execute: ./scripts/emc
./scripts/emc: line 581: [: too many arguments
Traceback (most recent call last):
File "/root/emc2/bin/axis", line 1636, in ?
s = emc.stst(); s.poll()
emc.error: emcStatusBuffer invalid
TorbaX... uh... the people you want, are probably not here.
and emc shutting down
cradek is the guy you want to talk to about axis.
he's the author.
same with oh... who was the other guy...
I post the question to cradek later
cradek: when you are active in channel read those messages, please:)
The axis version is latest.tar.bzip2
see you soon
I make snack, I'm hungry
the emc2 cvs version cant start if i do not manually add the adeos module, it unloads it, but it doesnt load it. this should be changed
hi Lucio65 :)
cncuser: wich distro are you using to run emc2 ?
Jacky^: puppy 1.0.7 = puppysit = coolcnc
hmm, jacky, youre right i should configure the modules correctly then they should get loaded
I'm not familiar with that :(
alex_joni: could help you
Jacky^: your distroquestn reminded me :)
Jacky^: i am familiar with that, i just didnt think about fixing it from that side :)
puppy its cool, I have to try it !
Jacky^: hehe, maybe i get a beta online this week
found some bugs in the howto / scripts
good .. youre a very good cncuser :)
Lucio65 is now known as Lucio_away
anybody know about how to wire geckos
where does a fuse go if you want to protect the gecko against a motor shorting out. On the high voltage going in?
the motor is a 5.5 amp. what size fuse would be best. trying to fix a problem and I have another motor to try but am scared of it
tonyps: read the geko manual, it's in there
gecko even ;)
they say about 67% of the motors amp, should come from the PSU
great, I'll look that up right now. be back later
so in your case the fuse might be 5-6A
I'll look that up and put a little smaller in it. I was told the motor was the cause of blowing a gecko before but I am doubting that now
I think the electrition dropped a screwdriver when adjusting the pot and blamed it on the motor.
None of the leads seeem to be shorted to the case or to each other. Just the two phases
I am going to go try this out. I'll be back (arnold accent)
What is on the agenda for today?
* alex_joni will do the stg configs
We should take a bit of time to discuss retroFest and codeFest for 2006.
* alex_joni knows little about that ..
good morning guys
is chris up too?
ok.. you usually come in pairs ..
that's why I asked :)
how've you been?
Pretty good. Many changes are a-happening.
that's good to hear, hope they are for the best
I'm now officially engaged, my girlfriend is taking 21 credits this semester so she can graduate in may. Looks like we'll be marrying in either June or July, and we're also trying to figure out how to be first time homeowners.
sounds great,I'm happy for you..
And she's a totally too-energetic nut, trying to cram all planning in before school starts, last week of Jan.
sounds like something I should do too, pretty soon ;)
Today, however, she's off elsewhere, so I actually have some few hours to myself, am here to check on the latest and greatest in the EMC arena, perhaps there's something new I can build to exhibit at Cabin Fever Expo next weekend?...
well, emc2 is pretty advanced
we're working towards a release soon
suggestion for a location for EMCcodefest: Stuttgart or Aalen :-)
Imperator_: don't think the amis will travel over the pond for it :D
I traveld last jear :-)
are there some details about Codefest ?
* alex_joni goes away for a while
03rayhenry * 10emc2/tcl/bin/halconfig.tcl: watch mode works a bit better.
kevin: is it true you cannot access wikipedia in china?
sorry, I cann't catch your meaning
03alex_joni * 10emc2/configs/stg/ (stg.ini stg.tbl stg.var README stg_io.hal stg_motion.hal): added initial STG configs
03alex_joni * 10emc2/configs/motenc/ (motenc.ini motenc_motion.hal): unified usage of P gains naming inside the ini's and core_servo
kevin: try looking at the emc2 code notes.pdf if you're interested in libnml
03alex_joni * 10emc2/configs/common/core_servo.hal: unified usage of P gains naming inside the ini's and core_servo
[16:35:30] <fenn> http://www.linuxcnc.org/EMC2_Code_Notes.pdf
what are you trying to do?
03alex_joni * 10emc2/configs/common/core_stepper.hal: added headroom for stepgen maxvel
03alex_joni * 10emc2/configs/stepper/ (stepper_inch.ini stepper_mm.ini): added headroom for stepgen maxvel
emc2_code_notes.pdf say little about it
kevin: what do you want to do?
is it possible to parted it to emc?
i dont understand what you mean
The RCSlib (Real-time Control System) was replaced with libNML in EMC2, And now I plan to use libNML in EMC1 directly
why not just switch to emc2?
Is it possible?
i am the wrong person to ask, but the person who knows doesn't come around here very much anymore
kevin: again, what do you want to do?
what does emc1 do that you cannot do in emc2?
* fenn sighs
fenn: probably poor choice of an already registered nick
hi fenn alex_joni
we have done much work in modifying emc1 to fit our environment. and now we plan to use libnml in our emc1
i would suggest reading the CVS logs to see what changes happened between emc1 and emc2
but it sounds like a lot of work
kevin__: the emc on the BDI is mostlyemc1 with libnml
but also some other things are different
what kernel/RT are you using?
the linux kernel is 2.4.16 and RT is 3.1
the BDI emc is in CVS
I can not see some character in IRC
under the bdi-4 branch in emc2
I need to go away, I'll be back later
cradek: are you online now?
alex_joni: I made emc2 and it works
... my screen only accept 640x480 resolution
i've a problem with tkemc and other GUIs
it doesnt' fit all windows..
but xemc it does.
see u later
there is something wrong with my debian
axis doesn't fit?
we've designed axis for an 800x600 minimum screen size
axis seems to fit correct....when it worked...
jepler: because I installed emc2 in root folder and Axis installed on its....
I've strange problem
emc2 with DISPLAY=axis
I follow the guide on axis site
emc 2 shows an error
what is the error?
just a moment
[17:11:36] <TorbaX> http://tinypic.com/karc5u.png
do you see it on image?
and your DISPLAY = line doesn't have anything after "axis"?
This doesn't work: DISPLAY = axis # I like this one best
because emc doesn't actually treat it as a comment
jepler: hehe, yes, this is allmost a classic ;)
there was a default row commented
default this row are commented (with a comment after 'axis' word)
but I suggest to move the row comment before the "# DISPLAY = axis" row
jmk_sleep is now known as jmkasunich
with a future diff patch to correct this simple mistake
SWP_Away is now known as SWPadnos
TorbaX: I'm here for a bit
TorbaX: that problem is fixed in cvs already (yesterday)
another problem now appears
[17:22:53] <TorbaX> http://tinypic.com/kas2a1.png
cradek: is there a changelog to latest cvs fix?
I will find it in webcvs, just a minute
because I made ok
this latest error may be because of your remote desktop connection
I don't understand what that is
that happens every time?
[17:25:29] <cradek> http://cvs.sourceforge.net/viewcvs.py/emc/emc2/configs/stepper/stepper_mm.ini?r1=1.1&r2=1.2
now i check if this error happens on my PC and not by remote connection
TorbaX, can you run other X apps from that shell? (like kate or something)
the remote conection is available by vncserver from BDI
did you say you can run xemc but not tkemc or axis?
on desktop axis ran
by vnc it doesnt run
it's incredible. ...kate and other programs are ok
does "glxgears" work in vnc?
hmm, maybe because of gl
it may be that OpenGL programs don't work properly in vnc, but usually that gives an error about the "GLX extension"
yeah that's a funny error
cradek: and a funny window ID
cradek: i'd swear I've seen that on install2, intermittently
good, I thank that theproblem was depend by emc2 or axis,
but I thank bad...:-)
TorbaX: you get the error consistently when running in vnc? Or just once?
jepler: no just this error i see
other I didnt
obviously by vnc.....
ok axis are ok
Is it possible to run DISPLAY = xemc ?
damit, what a set back. compiling huge software on a 500mhz pIII compared to my moboless 2.6ghz powered lamehorse
but xemc ...nothing
I don't believe that the xemc or yemc got moved over to emc2
sorry ...xemc is the only one that runs on 640x480 resolution
tkemc may work if you change some of the fonts
ok i tried to copy old /usr/local/emc/plat/linux_rtai/bin/xemc to "bin" folder of emc2
and it works good
Mini should work on 640
There may have been a few changes in NML since last xemc was looked at.
:( Z works
jepler: could you please post a filelisting of the files your script found out to be necessary to run axis into the wiki ?
but X and Y doesn't go
I don't think an xemc binary copied from emc1 to emc2 will work right.
cradek: on my PC it works
"right" is the operative word
IMO you should be able to do all the "usual" machine operation things thought.
typ0ing 1o1 again
fwiw, it should not be hard to get it to compile. I think emcsh and xemc are similar, so one could look at the changes to emcsh
also axis's emcmodule.c is very much like xemc
cu in a minute
is there a program that allows to export a HPGL file....???
I think hp2xx converts from hpgl to other formats
but I'm searching a program to generate an HPGL file
after I try to convert into a GC file
after I try to convert into a gcode file
because it's difficult to obtain a gc file immediately by programs
qcad doesnt export to hpgl or gcode
03jmkasunich * 10emc2/src/hal/drivers/hal_skeleton.c: fixed a bug in the 'skeleton' HAL driver - the HAL component name and the linux kernel module name were not the same, which causes problems for the 'unloadrt' command.
what do you need g-code for???
* fenn smacks dmessier with a wet trout
dmessier: i need a gcode generator from hpgl or other format because emc support gcode file...Did I think right?
yes... but ARE you a cad /cam jockey... could you handle a real cad/cam application
cncuser: the list may be different for other versions of python, or when it is compiled with different options.. but the files my script finds are here: http://emergent.unpy.net/index.cgi-files/sandbox/python-for-axis.files
I know for a fact that some of the "bwidget" files can be skipped, but I'm not sure which ones.
jepler: ill play around with it later :)
* jepler disappears again
woow, python is only a shadow of itself :)
hm, that list doesn't help, because you don't see the contents of python23.zip
[18:12:35] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/python23.zip.txt
jepler: it helps, for i just make it into a bashscript and copy the files needed, then i do some unionfstricks and et voilla i can test if it works :)
(BaseHTTPServer? I wonder why that's pulled in!)
ok, more files ;)
that would have been to good to be true ;)
fixed one thing .. check python23.zip.txt again
ok gpot it
some of encodings/ can be trimmed, certainly the microsoft-style codepage ones
hehe, palmos :)
you allays learn something new
hmm, the ssl modules possibly also isnt necessary
perect start :)
ok, im of to dubdevelopments
oh my god
the gcode examples file of kcam4 Win program don't load into emc
why am i not surprised?
various file .gc .ngc etc are not loaded
what's the error?
at the first rows
did you fix the line endings?
there are lots of different flavo(u)rs of g-code too
SWPadnos: I try nowe
emc load the file
if I press Run
the parse of gcode shows an error at the first gcode rows
N001 .......etc etc
can you paste in the g-code line and the error?
only the 28th rows
because the file STEPR_297_V7.GC is too long
the post processor is using square brackets for comments "", emc uses parentheses "()"
(this file is a part of kcam4)
also I think % must be the first line of the file
and the % is the end of file, I'm not sure if it's also used to mark the beginning
the error point to first line
change the  to ()
on both line 1 and line 2
no, just remove those lines
because the % has to be the first line
i test the edit function
and imark the 3rd line
to inizialize the run of gcode
rayh is now known as rayh-lunch
the error doenst appear
but another ....appears
TorbaX: probably the program creating this file is not RS274NGC compatible
Cannot do G1 with zero feed rate
you need a speed setting for any G1
check the file for an F word, before the G1
if it's not there add it, G1X...F10
or what speed you want
sorry........it doenst exist
ok, add it (you probably need to know how to edit a file for that)
I've added F1 1.0
before the G01 ...codes
and Axis now load correctly the file!!!
F1 1.0 means F11.0
rayh-lunch is now known as rayh
F=speed, 1 1.0 is a number, 11.0
there is no F1, F2, etc.. just F
I really suggest you start reading on G-Code, and what the symbols mean
alex_joni: excuse me but I don't have any practice in Gcode ...
TorbaX: nothing to be excused from, we all started knowing nothing ;)
I printed yesterday the handbook of emc that it shows principal Gcodes command
but there is a lot to learn before you can start using it like you should
the last one problem apparently I have is that only Axis Z move correctly anche other axis X..Y doenst move
TorbaX: what setup are you using?
i set on standard_pinout hal file
the corret pins
and the correct direction with setp parameter
did you test motors / drives?
I wrote a simple C program wich allow to test the axis X Y Z by parallel port
it ran on my desktop
and on BDI
in this program simply I set #define to map the parallel port
and check it
the motors sounds good
on EMC only axis Z move
X and Y nothing
seems something might be wrong
did you try to jog manually?
did you look at the pins to see if it's stepping?
alex_joni: I tried manually to jog them
and only Z moves
no, meter or scope on the actual parport pins
no I don't check it
I check by IO_Show if the Pin are correct
what are the pins you are using?
mmm it shows wrong pins....
not at all
only these pins I set
ok, you can use halcmd to test if these work
limit switch I don't set
emc2/bin/halcmd show pins parport
should display the values
try executing the command a few times during emc runs, and see if the values change as they should
or maybe run emc2/tcl/bin/halconfig.tcl
it should show you the values
the script shows an error
you probably don't have bwidget
"I didn't find the bwigdet library in the usual place
you're a magician!!!!
TorbaX: apt-get install bwidget
i see many number
I suppose tha's a pin port
the pins are correct
the limit switch pin11,12,13 AND the enable pin3 aren't setted
maybe IO_Show is easier for this problem, has labels for pins and for function
because the enable pin is harwired to all motors but the enable pin is not set by me(I couldn't find its)
tomp: the IO_Show doesn't work
yes it do
works may be wrong word
it appears and displays pins
TorbaX: you need to connect the enable pin to an emc enable pin
now i try it
newsig axis.0.enable-amp bit
linsp axis.0.enable-amp parport.0.pin-03-out
you need to set a name
newsig enablesignal bit
linksp enablesignal axis.0.enable-amp
linksp enablesignal parport.0.pin-03-out
probably the naming is wrong, I used memory for the names..
in the HAL Handbook
are there a list of names?
halcmd should show you what you need
sudo bin/halcmd show sig axis.0
halcmd doesnt work...
emclinux:~/emc2/bin# ./halcmd show pin axis.0
RTAPI: ERROR: could not open shared memory
HAL: ERROR: rtapi init failed
halcmd: hal_init() failed
NOTE: 'rtapi' kernel module must be loaded
TorbaX: of course you need emc2 running for halcmd to work
and for that you need to comment out the lines which don't work now ;)
now halcmd work
it's the correct name
and the limit switch are:
TorbaX: you should be able to manage from here
good job guys.
now I start the new pin assignment
now all pin are ok
now the last tips
setp parport.0.pin-11-in-invert 1
because i tried to invert the logic state of limit switch
emc shows an error to this
what is the errpr text?
in other words ...the limit switch are set to high state....and I check always override limits
HAL:54: ERROR: parameter 'parport.0.pin-11-in-invert' not found
before this, I setted the inverted pin-out successfully.
but these dosent work
is the name ov Input pin Inverted?
only output pins have an invert parameters
input pins have 2 pins, pin-11-in and pin-11-in-not
you can use either one
sudo bin/halcmd show pin parport.0 will list the names according to Hal
the "not" are logic inverted
i saw it
If you add the bwidget tool set, you can use halconfig.tcl to see all the pins, params, and such as well as make changes and watch values.
is the logic reversed
newsig limZ bit
linksp limZ axis.2.neg-lim-sw-in
linksp limZ parport.0.pin-11-in-not
In most of the existing systems the signal would be named Zlim
now tkemc shows only two axis over the limits
error: joint 2 error
no, its typical :o
tomp: not really
slow down the feed or override ( and typical for me )
alex_joni: joint 1 following error
alex_joni: taskinf.cc 787: Error on Axis1, command number 122
right, probably speed is not right
in the ini
is line 122 a G0 as opposed to others being G1/2/3?
the errors are
the errors appear when I had set the limit pins to the .pin-11-in-not
instead of .pin-15-in
now with pin-xx-in setted
the limits are active....
idem problem......with pin-xx-in setted
TorbaX: can you please take a conclusion before you post here? it's hard to follow..
we are trying to help, but only can do that if you have a specific question to ask..
yes, you're right
and please try to write connected things on one line.. it helps us keeping an overview..
I try to disable the limit switch and the enable pin assignment
use a standard stepper config first
make sure your motors work as they should, and afterwards start by adding enable pin and limit switches
after I ran emc, and I see that the axis Z is ok but X and Y don't work...
I clean my configuration
and I restart new template
thanks guys to incrediible support which you have got me, tomorrow I'll try to check entire configration,
see you later soon....
TorbaX: no problem
You bet TorbaX Glad you got the understanding to get it going.
03jmkasunich * 10emc2/src/hal/utils/halcmd.c: fixed a simple typo that resulted in u8 params printing the decimal and hex values even in script mode (where it should only print the decimal value)
ray, watcha doing with the parsing of the types? is this for halconfig?
Hi tom sorry was off reading logger.
Yes halconfig needs to find type to convert value from sci note to decimal
03rayhenry * 10emc2/tcl/bin/halconfig.tcl: switch scientific notation to decimal for floats
03jmkasunich * 10emc2/src/hal/ (7 files in 3 dirs): Added HAL Virtual Control Panel. Still very much a work in progress, will be finished after the release of emc-2.0.0, committed only to allow others to look at it. Do not use.
03jmkasunich * 10emc2/configs/sim/test.vcp:
Added test.vcp to the sim config directory. Eventually the sim configuration
will be used to demonstrate the use of virtual control panels, so the .vcp file
should live here. For not it is only used for testing during the development of