hey people. I put a new PCI card in my machine to add an extra parallel port. any idea how I could check if the operating system sees it?
lsmod or something?
hey peeps. anybody home?
what should I look for in the output of lspci?
it should show a parallel port identified and an address lix 0xC000 or such
try it, and if you dont see what you want, pastebin the result, maybe i can help
drat. I dont see nuffin. I always forget where to post to pastebin. is it pastebin.ca?
[00:40:19] <tomp_> http://pastebin.ca/
[00:40:45] <owhite> http://pastebin.ca/326325
all I did was open the computer, drop in the card, fired ubuntu back up. didnt know if I had to install anything.
yep, the system doesnt recognize it, got any docs? who made it?
i like china!
jumpers. hm. lemme look, I have to open the lid.
dual boot sys? see if windows can smell it.
(jumpers unlikely sorry if you already opened it up )
no jumpers. sorry its not dual boot.
hmm, manufacturer, model # before you close it?
or, it this the system you're speaking from?
jeepers I should really go back to school. I thought it would be recognized by the OS. do you think its okay if I yank the card to look at manufacturer?
NO NO NON NO NO
oh. just to be clear, is that a no?
it certainly was olly
hm. well, I'm hangstrung because I only have one machine available to me today -- the harddrive on my laptop died.
well, this time write down and keep it someplace, maybe pasted outside the box
oh, here's a model number on the box.
no manufacturer, just a model number.
and the winning # is ???
it appears also as A-9835-1P-2S
what do I win?!
ao sound intl io card
1 par & 2 ser?
would you believe the company is called "Generic"
yes. 1 parallel and 2 serial.
should come with a 2nd rear plate to hold the 2 extra dongle db9s & AND a cd!
made by asound international
who ripped you?
huh. well it just was a little PCI card with a DB-25, and it has a dongle going to another 2 serial ports.
figure I got to find a driver for it?
driver found here but win only http://www.asound.com/ProductDetails_cn.asp?ProductClass=36&ID=245
for linux, try to find the chipset :( = pull card again :(
okie doke. I'll power down, pull the card.
argh, the zip file they have is munged too
owhite: likely a NetMos 9835
should be supported in the kernel
try removing some of the other pci cards
owhite: the info on the NNM8235 for use as printer port (1284) http://pastebin.ca/326354
fenn's got a good idea, may be a clash, and defeat any on-the-motherboard parports too
I found docs!
[01:23:06] <owhite> http://pastebin.ca/326367
owhite: (mr! woohoo!) i pastebin'd the same thing! no problem if you get it working :)
tomp_ wave hello if you're available.
owhite: sorry, whats up?
hey. I found some documentation for my card....
yep, any help?
[01:49:57] <owhite> http://pastebin.ca/326367
look up, i pasted same one :)
it says I should be able to more /proc/pci
but there aint no /proc/pci
can you view /proc at all? (huge)
I just checked my dapper system
no /proc/pci here either
use the "lspci" command to view what's detected on the PCI bus
did that, nada parport
the port itself might not be identified as such
he pasted http://pastebin.ca/326325
"lspci" has the information that /proc/pci would have contained
lspci might just show the PCI interface chip that connects the port to the bus
the docs also say I should be able to lsmod parport.o and that reports no such file
owhite: I'd try doing lspci >foo
then remove the card, and do lspci >bar
and diff the two
if something appears when you add the card, that is the parport (even if it don't look like a parport)
right. by the way I said lsmod parport.o, its insmod parport.o
jmkasunich: right. but shouldnt it also show an address like 0xC00 or something?
owhite: is emc2 installed on this system? If so, we've prohibited the Linux kernel driver for parallel port cards from loading by default
jepler: yes. emc2 is installed.
owhite's lspci output is 10 lines long http://pastebin.ca/326325,
the chip is a NetMos 9835
try: 'sudo modprobe -i parport_pc' to load the linux kernel parport driver
then look at dmesg to see if it detected anything
sorry, where is dmesg located?
run this in a shell: dmesg | tail
dmesg is a command
just type "dmesg" on the command line
sorry, I forgot its a command.
I got these lines when I ran that command (I just have the onboard parport): http://pastebin.ca/326402
parport: PnPBIOS parport detected.
parport0: PC-style at 0x378, irq 7 [PCSPP,TRISTATE,EPP]
lp0: using parport0 (interrupt-driven).
sorry I hope the spam is okay.
three lines is fine
10 isn't ;-)
that's likely to be the onboard port unless you don't have one -- I'd say nothing in linux is detecting the presence of this card
so I am interpretting that to mean I just have one parport, 0x378
is that the mobo parport, not pci
some channels/servers wull auto-kick you if you paste too many lines too fast
did I hear you say this was a multi-function card?
jmk: 2 ser 1 par NM9835 chipset
owhite: did you disable the motherboard parport in bios? or see if there was one in bios?
in 2005 this person was getting 'lspic' entries for that chipset: http://lists.infradead.org/pipermail/linux-parport/2005-February/000251.html
I am really surprised there's not a pci card entry for it
is the card known to be functional?
the card is identical in appearance to this: http://www.newegg.com/Product/Product.asp?Item=N82E16815166006
tomp_ I tried disabling the mobo parallel port in bios. rebooted, and I dont think anything was loaded based on using lspci
jmkasunich: the docs for the card are here: http://pastebin.ca/326367
it gave me the impression I could use this under linux. I dunno.
you can also try lspci -vvv, possibly with sudo (though I think the extra information from sudo isn't useful)
owhite: yeah, I already read that page
ok, it must still be disabled ( at least mobo pport not seen by any linux tools ), jmk asked... did it >ever< work?
and thats why I asked if the card is known to work
that page says it should show up on lspci, but it doesn't...
Try to load the parport driver again?
so maybe its busted
oh. ever work? I just bought it and threw it in.
wrong direction :)
throw it out :)
do you have another PC you could try it in (maybe a doze box?)
right. well it came from china, so it must be under superior quality control.
omg - windows? ;)
sorry, no I cant check with another box.
I think you lack imagination to be blaming the -hardware itself- :-)
then take it back and tell them it doesn't work
you don't have to tell them it doesn't work with linux
anything, including hardware, can be busted
lots of low end consumer electronics is never functionally tested
trust him - he designs hardware ;)
well I feel like I get a fairly small budget of using everyone's time, and I'm running out very quickly. :-)
the customer is the quality control
I can try testing the card somehow.
it came incomplete, no cd as it should have according to maker 'asound', so it was an opened box ???
owhite: if they have not banned me yet - you have no problem. ;)
yeah, don't worry about time
no actually it had a CD.
its just that, when you've checked the basics, its time to go back to troubleshooting 101
what works, what doesn't
if you can't find _anything_ that recognizes the existance of the board, then maybe its dead
this is getting philosophical. "I am detected, therefore I am".
heh, thats about it
its funny - a parport is about 100-200 gates
owhite, have you tried lspci -vv ?
but the interface to the PCI bus, with detection and all that crap, is probably 20,000 gates
that's what the PLX chip is for ;)
wow. what should I look for in that output?
the PLX chip that seems dead
(plx on the 5i20 - owhite probably has a single chip board)
look for "parallel" or "1284"
then look for parallel or 1284 ;-)
in series - because once its pastebin'ed the browser can do the search for you
[02:13:47] <owhite> http://pastebin.ca/326418
but we can all look in parallel
check pastebin for NM9835
i'm playing new orleans jazz band funeral music
magic eightball says future uncertain.
parallel, 1284, epp, ecp, 9835, nada
nope - I don't see it either
robin_sz: did you see anything?!
survey says NADA!
you can try a different PCI slot, but other tahn that, it looks like crapola
looks like it, smells, like it, tastes like it, might be ^%!!#
hey, the _slot_ could be screwy?
did you get it from a local store, or mail-order?
anything is possible
I even have the receipt.
the slot or the IRQs / resources allocated to that slot by the BIOS
I just cant believe the lack of faith in our consumer electronics industry.
if its local, I'd just march back in there and say "this piece of crap doesn't work, I want my money back"
and a pound of flesh
also check the BIOS settings for "Plug-n-Play OS", and set to "no" if you have the option
oh. do you imagine there is something I could change in bios?
what is plug-n-play?
yes, set the Plug-n-Play option to No :)
okay before I log off, other things to check in bios?
it may not be there at all, but it has to do with whether the BIOS sets up peripherals or leaves it to the OS
nothing offhand, but let us know what BIOS/chipset the motherboard is
and pastebin it!? :-)
okay. *mental notetaking* plug and play, and the BIOS/chipset of motherboard.
paper and pencil notetaking works better
and try another slot while your there
I know this from experience - I can't remember shit
actually, change one thing at a time - fix the BIOS if you can, and if you made changes, boot into Linux and save the output of lspci -vv
does the basic m5i20 setup still include enc-00-cnt-latch enc-00-count enc-00-idx-latch enc-00index enc-00-latch-index enc-00-pos-latch enc-00-position enc-00-resetcount ?
i making a pyvcp test panel
that's an excellent question :)
to make the pyvcp panel, i need to know the data types and the data direction
I take it you don't have a 5i20 card?
you can take it i do
start the m5i20 config, open a shell, and do halcmd show pin
halcmd loadrt hal_m5i20
yes, thanks, halcmd more powerful than halmeter (uh i been told before ) now i know direction and type, thanks
what widget is useful for bit i/o ?
you mean pyvcp widgets?
checkbox/button and LED
leds for inputs
what he said
checkbox for things you want to keep at some value, button for "temporary" things
heh, that reminds me of something
before vcp came along, during the first round of the mazak
yes, but halcmd show says the pin is bit and is i or o enc-00-latch-index
ray was testing some IO using "newsig foo; linksp foo pin-to-test; sets foo 1; sets foo 0"
of course each command was typed individually
when he tested the coolant pump bit, it started squirting stinky old coolant at the ceiling because the pipes were disconnected
and of course its hard to type "setp foo 0" fast when there is stinky coolant shooting at the ceiling
today a pyvcp button would be a good choice
tomp_: sorry, got distracted with my story
you are talking about a pin that is listed as "IO"
neither in nor out, but both
ok, I misunderstood
to be honest, I'd just stick an LED on it
ok, thats easy
if you really want to test the index functionality, you can setp it to 1, the led will light up
hmmm - can you connect an IO to an O?
or just other Is
then turn the shaft, when you hit the index the light will go out and the counter should reset to zero
SWPadnos: not anymore
an IO-mode "set only" pin would be useful in VCP
sets to 1 when you push
but doesn't update
but somethng else can turn it off
in theory, "O" means "I'm gonna write to the pin every time I run", so that would cause contention with somebody else who (might) write it
.. pops back out when the pin goes back to 0?
tomp_, yep - that would be ideal
write 1 when pressed, detect reset to 0 before popping button out
that sounds like fun
jepler: did you see vpython?
how about "displayed state of button matches the signal, when clicked it writes (only once) the opposite state
tomp_: i've heard about it, never used it
jmkasunich: race condition between the eye and the hand
oh, because it might change just as you are clicking it?
well, if you are gonna have one that writes a 1 when clicked, then you ought to have another that will write a zero when clicked
for test purposes ( no motors allowed), just checking functionality, the btn could begin a timer for 1 sec & the led reflect the state ??
I was trying to kill two birds with one stone
but maybe two stones is better
tomp_ no clue what you are talking about
if you are testing the index function of an encoder (the only reason you would access that pin) then you want the widget that we are talking about
you could make a compound widget, with a "set" button, a "reset" button, and an LED
well, the "write 1" button is the set, the "write 0" button is the reset, and we already have the LED
I wouldn't build a widget
just make a box and stick the widgets in it
(of course, that means more hal and xml code)
ok, write1 write0 and led in a frame with a label
actually, you can leave it as a button that stays down until the signal goes back to 0
2 radiobuttons, 1 led 1 label 1 box
not radiobuttons - those are "permanent"
they should be called but-ons
and a butt-off
there is a momentary button in pyvcp, right?
yes, butt-on and butt-off
butt dup on off ;)
ok, but-on but-off ( thought i could write to the radiobuttons ( control thier state))
there is a momentary button in pyvcp, right?
yes, button' is momentary
then all you need is one that doesn't write a 0 when the button is released
it writes a zero (repeatedly) when off, and a 1 (repeatedly I hope) when held down?
it may continue to write a 1 while held down
it should (both when down and up)
not foran IO signal ...
"halpin is 1 when button pressed, 0 otherwise "
anyway, I'd suggest an attribute for "button" that says "export an IO pin instead of an out pin, and only write it when clicked"
is there a toggle button?
* jmkasunich should RTFM
toggle button is actually a better candidate for this
checkbutton is close
display the state of the pin, when you click it, read the state, and write the opposite state one time
oh, checkbutton is togglebutton
just different names
it's a checkbox, not a checkbutton ;)
anyway, the "normal" behavior for checkbutton/checkbox/togglebutton/funnywidget should be to export an OUT pin and repeatedly write it
and an XML tag should be able to say "export an IO pin and write it only when clicked"
is out pycomp.newpin(halpin, HAL_BIT, HAL_OUT)
I would guess so, but I don't know
that was a quote from pyvcp_widgets.py
probably. change the HAL_OUT to HAL_IO for an IO pin
oh thats how it'd be done ( to get the IO pintype) but the rule 'only when checked' thats not to be done inside pyvcp_widgtes, maybe in vcpparse.py , it may need an xml token added
the button widget will need an internal variable, and the update function will need to check that variable to see if it's OK to write to the pin
the variable will be set with an additional XML token
I'm not looking at the code - is there a way of doing "internal options"
like <option>io<value>true</value></option> or something
yes,there's a list of tokens for allowed options, and in the widget __init__ you collect any name/value pairs that make it past the parser
ok, that should make it easy then
(he says, without knowing anything ;) )
tomp is collecting which of many ideas was the final suggestion...
I think it was an option to the standard button, which when set causes the exported pin to be IO, and the update function to only update (toggle) when the user clicks the button
SWPadnos, it's arrived.
well, it's actually at the post office, where I have to pick it up, and pay something like $26 for duties and shit.
they came by, but didn't ring my doorbell.
didn't hear the other one.
well, at least you know it's around, and available for a finite amount of cash
wonder if customs would have been a bitch to have marked it for $20.
so far http://imagebin.org/6994
for a while, I thought customs was opening my mail...
I had to fill out a postal declaration that tells what's inside
as anything from a tech company was being opened, mysteriously, if it was addressed to me.
SWPadnos, always use it first, then mark it used parts, cheaper
hmmm. it should be sealed when you get it
tomp_, oh hey! good idea
tomp_, heh - I could have done that. though I included a copy of the receipt for warranty use
SWPadnos, mail that separately next time!
customs can open everythng, i have to put a hole in crates, cover with tin, for them
tomp_, where you from?
tomp_, otehr than the input numbers being the same (I00-I07 three times), that looks cool
or whatever I put there
what are those things for the outputs?
are those the checkbuttons?
looks like a copy/paste error to me :)
funky looking things
not screamingly obvious whether they are on or off
what's the word on step/dir functionality?
the word for today is "wait"
same as the last time somebody asked
jmkasunich. :) the man of the hour
I think it was a day or so ago, on one of the lsits
make sure it has the good packing nuts in it.. I heard of others getting sub-par packing material
jmkasunich, I'm on the list, but don't read it... too many posts
and the answer was "in a couple months"
the phrase for today is "if you want it faster, do it yourself!" :)
* skunkworks waits longingly for the next dev or user list posts.
Jon Elson just posted one ...
jmkasunich, I don't see that when I searched for mesa, or 5i20... :(
I don't see the post that is.
this is the firstname.lastname@example.org list right?
look for step, in the body ;)
I see something, but nothing in regards to mesa step/dir project completion date. :) Just jmkasunich saying he's using the mesa, cause it's better suited to his needs.
Re: [Emc-developers] Generating step and PWM control
Sun, 21 Jan 2007 19:11:00
ooooooooh... I'm no the devel list.
oh, then you are excused for asking ;-)
wow look at this edm'ed racecar part http://www.moldmakingtechnology.com/articles/040508.html
bah - I can do that with my 6" long 1/64 end mill :)
woot - an LED dimmer!
what, it wont turn it off all the way? or (more likely) pwm to an led?
I'm messing with the DAC outputs (which are actually PWM)
and I have the 7i31 LED board connected
i found the stuff on vpython when i found stuff on how to capture desktop to swf/flv/ogg, and in that i found some python tutors, http://showmedo.com/videos/python
we need more visual communication
in then end i found i could just use ffmpeg to do screen capture ( not istanbul or vnc ... )
videos to learn programming?
thats just... wrong
thats from a good programmer, who doesnt click much, but for the 'rest of us' (woooooo i heard that before) :)
I can see the value in pictures, especially when learning GUI software where you have to show people what to click, or when talking about toolkits and you need to show them what the widgets look like
video (or audio) forces you to proceed at the pace of the media
you're right, it was handy to see how some ide's worked...
its inconvenient to slow down over the parts you need to think about and speed up over the simple parts
its hard to index so you can go back and review a specific topic
it just plain sucks as a teaching medium, IMHO
this is what video is for http://www.i-bar.ch/en/video/
the technology is tfir? an acrylic sheet, some leds, an ir camera and a lcd projector
Frustrated Total Internal Reflection
I didn't realize the 5i20 dacs wire pwm+dir, I thought they were up/down (two PWM pins)
ftir means that when you touch the acrylic it changes the index of refraction and so your finger "glows" and the camera can see that from the other side
its lit from the side by led's
I wonder how long it takes before the top gets scratched
the web shows more greasy fingerprint pix than scratches, and scratches are bad for the detection, jef han at TED http://www.youtube.com/watch?v=UcKqyn-gUbY
jmkasunich what are you describing about the dacs, it's a single pin, and the signal is above/below 0?? so pwm can cause polarity and magnitude?
no, its two pins (logic signals can't go below zero)
one pin has PWM on it, based on the absolute value of the desired analog
the other is either on or off based on the sign
that is pwm/dir (similar to step/dir)
the alternative is up/down, where one or the other pin PWMs, based on the sign of the analog value
up/down is more suited to driving H bridges - send up to one side of the bridge and down to the other
ok, i eventually need to turn that into standard analog bipolar... this was discussed before & i better find it ,some integration as I recall ( no h bridge, stock analog amps )
you can convert from one to the other with a couple of gates, but its a pain to have to do that in external hardware
if you are using analog amps, the easy choice is the mesa 7i33 card
oh, thats why i bought one =-O
they do that for you - convert the PWM into analog (including the biploar part), filter it, and gain it up to +/- 10V
(off to get those docs)
no, off to finish the dang mesa test panel
03jmkasunich 07TRUNK * 10emc2/docs/man/man9/ (siggen.9 sim_encoder.9 supply.9): more new manpages
03jmkasunich 07v2_1_branch * 10emc2/debian/emc2.files.in: backport: new manpages
03jmkasunich 07v2_1_branch * 10emc2/docs/man/man9/ (siggen.9 sim_encoder.9 supply.9): backport: new manpages
[06:10:07] <A-L-P-H-A> http://www.youtube.com/watch?v=7IbV7ad2xgY
any opinions on which is the move advance linux 3D system?
3D cad system that is
s/move advance/most mature/
varicad looks slick
let me look at them
okay varicad is out, commercial
hmmm brlcad looks nice but no support for engineering analysis...
brlcad is hard to use and lacking important features like opengl shading
but its very mature :)
okay, is there a useable 3D cad that will support de facto file formats?
Good night :-)
ds3: no :(
ds3: yes, yes there is!
i think o'reilly is running out of animals to put on their book covers
ds3, its called "Solidworks"
fenn, do they really put animals on their books?
How to you get dhcp to renew under ubuntu/debian?
Jymmm, do an if up/down
robin_sz tried, didn't work
sudo pump ?
did it search for a DHCP response and didnt get one?
you mean 'if etho down' ???
sudo pump --renew
pump? new one on me .. I try
nah, thats the server side
ubuntu/debian dont use dhcpcd i think
oops, I thought he meant dhcpd
I use dhcpcd -R eth0 to bring up the interface on my firewall/router
I just use ifup ifdown ... the dhcp happens as part fo network initialisation
a of course, I should add I am a Debian guy, its possible the the ubuntu fairies have fucked it up
all good things to know
dhclient seems to work also
here is a jumping whatever for your pleasure http://www.oreilly.com/catalog/pythoncook2/cover.html
ok, 'sudo ifdown eth0' 'sudo ifup eth0' worked
oh, PICTURES of animals .. that makes more sense
weird that ifconfig eth0 down didn't renew
no its not
wtf is that ...
some sort of rodent
not a fucking python thats for sure ;)
the product of rabbit/squirrel procreation?
The animal on the cover of Python Cookbook, Second Edition is a springhaas (Pedetes capensis)
[08:24:55] <Jymmm> http://www.oreilly.com/catalog/pythoncook2/colophon.html
so there's a huge python on the back cover about ready to swallow its ass?
Jymmm, ifconfig brings interfaces up and down ... it does just the "interface up" part of it .. not all the "and run this script, do all these other things" that are in /etc/network
only ifup ifdown do that
ifconfig is lower level
The python cookbook. Take one medium python. Skin and wash. Slice into1/2" thick slices, fry lightly in oil with a little garlic, serve in a bread roll with a few onions and a little mustard relish
never heard of it
alex_joni you awake?
robin_sz's up early
alex_joni installed ubuntu, any issues uninstalling gnome and installing kde?
Jymmm: only that some stuff will stay behind
that means diskspace
other than that, not that I know of
(disclaimer)> I never tried it myself *grin*
I have ried installing kubuntu, has issues installing where ubuntu doesn't. weird, but
all I know is that sudo apt-get install kubuntu-desktop should switch you over
eeesh, I'm running firefox and a terminal, yet 89 processes are running, almost as bad as M$
and kde is less resource intensive than gnome?
just a different flavour
if you want less resource intensive then try xubuntu with XFCE4
I liked it a lot
and you like yugo's too,
so, that aint saying much =)
ok, it aint too bad
I don't like yugo's :P
but I'd like a smart
(the roadster though)
You should me a pic once
I might have.. but I don't like it
[09:43:07] <alex_joni> http://www.dieselstation.com/wallpapers/Smart-Roadster/Smart-Roadster-016.jpg
well the pic had you in a 2pc bikini on the hood of one!
who makes smart?
some european consortium incl. Mercedes
the normal smart, and forfour (with 4 places) is butt ugly imo
I'd rather my gf get smart instead of a mini
it's somehow cheaper :D
but the roadster is still 20k$
[09:45:25] <alex_joni> http://www.autointell.de/europaeische-hersteller/smart/smart-forfour-diesel/smart-forfour-cdi-125.jpg
so is a toyota =)
Jymmm: yeah, but a toyota is a bit bigger than 3.5m length
the standard smart is 2.5m
we were watching a movie once, and there was a car smaller than a mini in it and she wanted it *sigh*
"The Roadster is powered by 45 or 60 kW (61 or 82 PS) versions of the turbocharged 3-cylinder engine in the rear, which is engineered by Mercedes-Benz."
that is a motor you can take out of the car by yourself
with your bare hands
does that mean I don't have to worry about breaking my 4 ton floor jack?
what's a nice smb browser client?
* alex_joni uses smbclient
some don't like textmode though
yeah, want gui for this
Jymmm: in kde you just type smb://hostname to browse
what about in gnome?
I think that works too
the normal explorer understands smb afaik
where is the normal explorer in gnome?
that sounds like it works :)
argh... Totem Movie player is whining that it can't read from the source.
alex_joni oh xubuntu is what you were talking about by chance?
11:39 < alex_joni> if you want less resource intensive then try xubuntu with
11:39 < Jymmm> icky
read after that =)
I can't get movie player to see the smb share
same as mount
alex_joni, the roadster isn't too bad... http://www.dieselstation.com/wallpapers/Smart-Roadster/Smart-Roadster-016.jpg
I wouldn't buy it... but if a chick drove it, I think it's be okay.
if you drove, I'd probably laugh. :)
I saw a smart roadster with 420hp
a motorcycle engine with turbo
no, that was a smart motorcycle car..
having issues trying to install emc
says theres a package missing on apt
do you have universe enabled in the repositories?
Fetched 4B in 6s (1B/s)
Failed to fetch http://www.linuxcnc.org/emc2/dists/dapper/Release
Unable to find expected entry emc2/binary-amd64/Packages in Meta-index file (malformed Release file?)
Reading package lists... Done
E: Some index files failed to download, they have been ignored, or old ones used instead.
Reading package lists... Done
Building dependency tree... Done
E: Couldn't find package emc2-axis
hm, shit, I really need a faster Z
ah well, 11 minutes per part isnt bad at all
but it's be 5 with a faster z
and I could do paralell contouring
or plunge for the edges
sorry got cut off
col: there are no 64 bit emc packages
why is the script trying to install em then
the scripts installs based on your subsystem I think
did you install the 64 bit ubuntu?
i installed the livecd
from the emc site
will there ever be a 64 bit version of EMC?
the LiveCD already has EMC on it
so somehow you must have ended up with something else
it didnt install with it i dont think
Martin_Lundstrom: there is sim, not sure about RT
maybe i isntalled from a different disk
col: can you run "uname -a" please?
Linux colin-desktop 2.6.15-27-amd64-generic #1 SMP PREEMPT Fri Dec 8 17:50:54 UTC 2006 x86_64 GNU/Linux
that wasn't the emc livecd :-)
definately not :D
that would explain why its not there lol
amd64 & SMP ..
it'll be a while till emc2 runs on that (if ever..)
maybe the 2.2 series of releases - but we're just starting 2.1
so emc wont install on the 64 bit
the problem is the realtime system, not so much emc, but in short - no
ill have to reinstall
I hate 3d contouring
getting my machines next week
it takes so damn long time
should be fun trying to work out everything
think im gonna go wih the montenc controller boards
why not the mesa cards or something?
I _think_ they are better priced
are they better?
you should as SWPadnos or someone about it
[13:31:12] <anonimasu> http://www.mesanet.com/motioncardinfo.html
you should ask SWPadnos or someone about them
[13:32:41] <anonimasu> http://www.mesanet.com/
im just hoping that the servo control boards in the mill already are usable with teh controller
col: what are you getting?
motenc have real DAC's on them, the mesa boards use PWM + filtering for that
both are supported by emc2
yeah the built-in real DAC and ADC seem to be the motenc's main advantages
* jepler was just reading that page
the non-lite card has a lot more total I/O too
yeah, but it's way more expensive
it's also 8-axes afaik
more expensive than two -lites?
don't think so..
but still more expensive than 2 5i20's
getting a hurco milling machine and churchill lathe
bed on the mill has about 1m travel
lathe has 12" chuck hydrostatic bearings 10 tool changer and a C axis
the lathes using siemens controls at the moment
hm, what's wrong with the stock control?
other then that it's siemens
on the lathe nothing
on the mill the memory board has gone
I wouldnt replace a working controller
if it's not old as hell
its pretty ancient
but if it works for now ill be happy
the mill is the main one
are the servo controllers usualy reusable from old mills ?
or am i probably looking at new controllers
yes they are
if you are lucky ofcourse
ill have to take a load of pics of em when i get it
if not is somthing like the gecko servo drive what id need?
yeah, but that depends on your motors
i had a look at the motors
didnt really have a lot of info on em
how do you tell?
look on the back of the motors and see if there's a marking at the thing at the end
on that thing :)
fjungclaus__: hi Frank
col: sometimes it helps to count wires :)
tach is 2 wires usually, works together with velocity controlled servo (usually on DC motors)
for position loop you usually get an encoder (can be 6, 8 wires)
AC motors usually have an resolver on them (6 wires, pretty neatly shielded)
i belive its a dc motor with encoder then
you can either keep the old amps (if they work)
the most probable interface from the control is +/- 10V
so you can hook them up directly to the motenc
thats what i was hoping for
or if they are busted you can a). get new ones b). get something like the G340 form gecko
b). means you don't need the motenc
you use step/dir (like a stepper) from the control to the drive
and keep the PID inside teh drive
but that has several limitation (and it's also cheaper :)
limitations: 1. possibly smaller max vel (because of software generation of steps)
2. you don't have the position fed back to emc2
id rather get teh montenc then
3. you need to tune the PID inside the drive
col: if you can afford it, then go for it :)
well to me then montenc isnt that expensive with exchange rates ect
col: where from?
oh cool.. :)
there are some other people from there on this channel once in a while
robin_sz is from Worcestershire
im also considering getting the 8 axis controller
coz ill want 5 axis at some point i reckon
you can always get a second lite :)
the driver supports multiple cards
but it's 2 x 550$ vs. 1 x 765$
hmm.. I see the Motenc Lite has 8 channels of DAC/ADC
only the encoder counters are limited to 4 channels
maybe it's cheaper to get a Motenc + a 5i20 from Mesa
use the Motenc for 8 channel DAC/ADC, and 4 channel encoder read + 4 channel encoder read on the Mesa
that means 550$ + 200$ pretty close to the 8 channel :)
but you'll have shitloads of IO on the 5i20 :D
for the price difference i might aswell just get the 8 chan
$765 and its all there
i might just rip out all the old amps from the mill anyway i can probably sell them for more than the costs of some new geckos
yeah, but as I said.. geckos are inferior to DC amps
although they can be crippled to run as DC amps
not the product is inferior.. the solution
the geckos are great drives
what would be a good DC amp then?
or the galil ones..
i thought the geckos were brush DC amps
col: yes, but driven as steppers
so you need Dc motor + encoder => gecko
ahh i get you
the gecko closes the loop itself
ill try keep the existing amps then
[14:25:59] <alex_joni> http://www.copleycontrols.com/motion/downloads/pdf/4122z.pdf
col: if they work.. it's best
do the amps just connect to the montenc with 10v out ?
and you need the encoders connected to the motenc too
i should be ok then
and a bit of tuning :)
aparnetly the machine was running fine its just teh memory board on the old controll that gave up
alex_joni: you make it sound so simple
especially the tuning part
* alex_joni doesn't want to go there :D
ill be happy if i dont have to refit the lathe
the hydraulics pipe are just as much of a spagetti junction as the wiring is
right must dash
work to do
thanks for the help :)
rayh: I added some commented out configuration and instructions in configs/stepper/standard_pinout.hal. Would you check it and let me know if there are other common questions you hear that we could answer with an example in there?
cradek: without looking at the file: home/limit switches, amp enable/disable, (power save for amps - less common), spindle on, estop..
I did not touch estop or power save (actually I don't think we have a good approach for power save)
the rest I put in there to some extent
I had a user ask once for that..
I made him a hal file which used ddt and a window comparator
that means the amps are still in power save at the beginning of every move, since the enable signal will be delayed a bit
it would need to be in stepgen to do right
the steps need to be delayed (just like they are currently when there's a direction change)
Hi cradek. Let me look in just a bit. Phone.
cradek: yes, I agree with you that's the proper fix
but this particular user reported it's ok like this
had another one ask for the exact opposite :)
output a pulse once every 2 seconds so the drive doesn't go into powersaving
what an evil hack
it was easy to get constant pulses out with emc1 :-)
I had that often when my pseudopid and deadband were wrong
rayh: right, siggen + sum2 + ...
siggen one step worth of distance?
[15:09:01] <alex_joni> http://dsplabs.cs.utt.ro/~juve/blog/index.cgi-files/sandbox/core_stepper.hal
I appreciate your cleverness, but that really should have been solved at the hardware level!
cradek: I agree.. but this was a custom driver which probably didn't have an option to switch the powersaving off
probably = microcontroller based, with little info :)
[15:11:13] <cradek> http://cvs.linuxcnc.org/cvs/emc2/configs/stepper/standard_pinout.hal?rev=1.4
anyways.. definately not something for standard_pinout
^^ these are the changes
cradek: although I like the last part (net command) I think it might scare newbies
I think 'net' is easier than newsig+linksp/linkps
and better than linkpp
for 2.2 we should change the standard configurations to use 'net' everywhere
standard_pinout.hal changes Look good to me.
I tend to feel that this level of documentation ought rather to be in the integrator manual.
rayh: it's not a replacement
rayh: it's rather a quick info to get users going, instead of sending them in here to ask about it :)
the Integrator Manual is rather incomplete too.. :(
Sure. There are lots of things that ought to be more readily available.
anyways.. running home.. bbl
rayh: I wanted to put only the very basic example stuff in there - now that I go look at it again, I may have overdone it
it's just another attempt to avoid the first common stumbling blocks that people hit
Sure I can understand that.\
oops (\) typoing
hmmm - I like the information there, but it definitely looks "complicated"
I'm thinking that the approaches you use here could as easily be applied in many other configurations.
but the stepper stuff is the most common entry point.
yeah it's not supposed to be documentation, just a way to reduce questions -- many people seem to make it to that file, but don't quite understand what to put there
SWPadnos: any ideas for how to simplify it?
only one, and it doesn't simplify, it just makes things easier to find in the editor
the idea is to stick "big text" section names in the file: http://pastebin.ca/326904
"HOMING" "ENABLES" "LIMITS" ...
not really. it gives a visual cue to someone who doesn't know what they're looking at
it's stupid, but it works
I agree it's stupid, but don't necessarily agree that it works
proportional fonts are common in editors now
You guys know where yelp gets it's toplevel directory list?
that's true - in which case it ouwl look like the "html representation" there :)
right, like someone puked some letters on the screen
rayh: do you mean the contents of the Help Topics screen?
well that's what's under the line I guess, hmm
I looked at that but didn't see anything but i18
<description>Guides to using the desktop</description>
oh but what does it link to, hmm
I guess the answer to your question is "no" :-)
thanks for trying.
in the source's README there's something about "scrollkeeper" and "docbook"
<long shot> maybe someone #ubuntu could tell you how to add stuff to it </long shot>
it'll turn out our documentation is in the wrong format
fjungclaus__ is now known as fjungclaus
not to be negative or anything..
jepler: I think it can show man pages prettied-up
yeah, it does a nice job at that
someone should get rid of this remark in stepper_mm.ini: # axis is an external interface, check http://axis.unpy.net/
I'd do it but I have local modifications in that file right now
I thought alex cleaned those up
maybe I don't have the latest version, or he did it only on the branch?
I'm working in HEAD
now that I've been reminded of the INCREMENTS feature I got around to adding something else I wanted
INCREMENTS = 1 cm, 1/8 in, .1 in, 1/16 in, 1 mm, 1/32 in, .01 in, 1 mil, 10 um, .1 mil
03cradek 07TRUNK * 10emc2/configs/stepper/stepper_mm.ini: remove outdated info
jepler: do they show up just like that in the listbox?
cradek: they appear in the order listed
a user could choose to sort them by length (as I have) or split them into metric and imperial, or list them with the smallest increment first
I agree that's probably smarter than automatic sorting, which is what I was thinking originally
you know - what may be more generally useful would be a halcmd syntax hilighting file
I'm pretty sure there's a way to get all of the standard text viewing/editing tools to use it (less, nano, gedit, kate)
fjungclaus is now known as fjungclaus-away
oops.. forgot about TRUNK
wanted to do that last night
alex_joni: Hi, Alex (with a 2.5 hours delay :-) I just left my irc-client running for monitoring the traffic here. And let me say, I'm astonished about the amount of traffic on #emc!
fjungclaus-away is now known as fjungclaus
fjungclaus: you should have been around this weekend :D
how much of it was on topic?
SWPadnos: ha ha that is a funny joke
03alex_joni 07TRUNK * 10emc2/configs/stg/stg.ini: remove references to AXIS beeing an external app
03alex_joni 07TRUNK * 10emc2/configs/vti/vti.ini: remove references to AXIS beeing an external app
03alex_joni 07TRUNK * 10emc2/configs/hexapod-sim/minitetra.ini: remove references to AXIS beeing an external app
jepler, which one? (I never know what people think is a joke ;) )
SWPadnos: syntax highlighting that is used by less, nano, gedit, and kate
SWPadnos: calling those programs "all the standard text viewing/editing tools" is also rpetty funny
well, right-click in a file browser, and you get gedit, sudoedit gets you nano ...
we're lucky they can all properly view and edit the encoding we chose for hal files (7 bit ascii)
less is just - well, less
jepler: you didn't like the MAIN LOOP joke?
I'm quite sure I've seen color syntax hilighting in less and nano
(if the romans cared about the date)
they did.. but used a slightly different calendar
if only we could specify jog increments in roman numerals :)
SWPadnos: anyway, fixing "sudoedit gives nano" is one of the first things to do after a fresh ubuntu install
jepler: it does give mc here :D
heh - vi has syntax hilighting also, doesn't it?
and I like it that way
the editor "6"
vii of ix
bunch of goof balls.
stfu! u're just jealous :D
skunkworks: kidding.. btw, there was something I was meaning to ask you :P
03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: support units for jog distances: um, mm, cm, in, inch, mil
alex_joni: whats up?
no time yet :(
heh.. you seem to know the drill :P
new documentation for jog increments: http://pastebin.ca/326984
"machine's native unit" -> is that the TRAJ setting?
or the corresponding AXIS/JOINT setting ?
I think so, anyway
Oh, "units for jog distances"! Great! Somebody was able to read my thoughts (wishes) :)
alex_joni: if there are bugs in axis support for machines with different units for each axis, that's a separate issue
I think the doc you wrote there makes it clear for the user
(beyond having different units for XYZ and ABC)
it doesn't make it quite clear for me though, but I bet/hope users don't try all possible combos
it's not likely that a 1um jog will be useful on very many machines, I guess
can someone tell me what a good value for INCREMENTS on a metric machine really would be?
fjungclaus: you must have some idea
My metric setup is: INCREMENTS= 0.0125 0.1 0.5 1.0 10.0
Don't wonder about the 0.0125. That's the minimum stepwidth my machine is able to handle
ah I was just going to ask
I suspect that a lot of people have near 1 uM resolution - that's only 25400 steps/inch
whether it's useful to jog that little is another question
0.5 but not 5.0?
I was thinking of changing it to this: INCREMENTS = 10mm, 1 mm, 0.5 mm, 0.1 mm, 0.5mm, 0.01 mm
er, something's not right there
one "step" on the pluto lathe is 1/6000 mm
INCREMENTS = 10 mm, 5 mm, 1 mm, 0.5 mm, 0.1 mm, 0.05mm, 0.01 mm
that seems a bit nutty
that's how the encoder worked out
and JonE has his 120000 step/inch machine as well
of course it can't position like that
no I bet not
10TPI screw * 2000 steps/rev (for 10x microstepping steppers) = 20ksteps/inch, which is ~1.25 um /step
(1.27, to be exact)
jepler: Your metric suggestion sounds ok for me
cradek: you want to stake out a position on useful imperial step distances?
the smallest two should be .001 and .0001, the largest depends on the machine
for me the largest useful would probably be 1"
you wouldn't rather do that with continuous jog?
depends on the jog speed - on the lathe it's a little fast for that
I cringe using incrimental jog. for me I really don't see a use.
I usually use incremental jog for only the shortest jogs -- the final steps of a "touch off", basically
I might choose .005in, .001 in, <machine step> actually
I could see that - I guess I do touch off differently. I get the axis close - then measure the tool hight off the part.
cradek had a good system with a feeler gague and a .0001 final step size, for doing circuit boards
geez I'm spelling like skunkworks today
fjungclaus is now known as fjungclaus-away
03jepler 07TRUNK * 10emc2/docs/src/gui/axis.lyx:
document new jog increment units feature
in axisrc example, show how to turn DTG on by default
we use this a lot for finding material.. http://www.goantiques.com/scripts/images,id,990648.html
fjungclaus-away is now known as fjungclaus
fjungclaus: since you're still messing with your IRC client.. make it join #emc-devel aswell
jepler: "show how to turn DTG on by default": This is specially for me, isn't it :-) :-)
fjungclaus: mmmm could be
subconcious actions ;)
on both of you :P
Anybody any suggestions for a nice irc-client? By now I'm using (aka fighting) with Chatzilla ...
fjungclaus: I like the terminal-based "irssi" program
xchat seems also nice
* alex_joni uses irssi too
TNX. I'll give them both a short look
fjungclaus is now known as fjungclaus-away
alex_joni: do you have any idea why my Z tris to keep up with all other axes
even though I've limited it in my config
what do you mean?
on this line
there's the right one..
and when I jog X or Y and Z
Z tries to keep up and starts mising steps..
skunkworks: do you understand?
anonimasu, why are you using G91 on those lines?
G91Z0 should result in no motion ...
in incremental mode
G28 Automatic return to reference point G28 Automatic Zero Return
g28 is go home. I read it is - go home through xyz point.
G28 homes all axes, not those specified
right - if you specify a position, then the sequence is "go to position, then go to home"
if you try G0 <whatever the home coordinates are>, do you still have problems with Z?
if so, then I'd suspect a configuration issue. if not, then the interp is busted, since G28 is equivalent to G0
it does the same when jogging Z and other axes at the same time
g0 for just Z works.
ko, so there's wither a config problem, or there's a TP problem (I'm betting on config ;) )
I can post the config in a bit..
can you pastebin it?
just need to go out and shop some
will you be here a while?
in and out, but yes
you're welcome (in case I help any :) )
I'm planning on setting up a dual stepper gantry. Would it be possible to "replicate" the axis by connecting dir/step to both motors drivers, and their individual homing switches in series?
Or should I go the "software" way?
software is probably better, since homing can be a problem if the two motors are driven by identical signals
ie, you would have no easy way to "un-rack" the motors if they ever got out of synch
You mean if they are so out of sync that the gantry jams.
slundell: Even if you want to use the same step/dir signals to drive both motors I recommend multiplexing the signal in HAL by hooking up multiple output pins to the same signal generator, much easier than doing hardware and much better (this is what I did btw)
slundell: I plan to test out the new gantry kins soon though :D
Dallur: oki, servos or steppers?
slundell: you ?
slundell: it also saves you from reversing the phases on the motors to make one motor run backwards, you can just reverse the signal in software
How about homing, were there any problems?
slundell: dallur-thc config in 2.1 is my config which you can look at
slundell: I do manual homing at the moment, but gantry kins will solve homing by enabling me to sync the sides during the homing cycle
thanks, will look at that.
Do you have any kind of feedback from the steppers?
slundell: I don't have any encoders or optics no, the only feedback I have is the limit/home switches
I was thinking that one could use some sort of adjustable switches to be able to tune the alignment.
Huston, i heve a problem.... a big problem
hi all, what's going on?
in mi system
.... we have a mains B undervolt...
slundell: I use inductive NC switches for the home/limit switches, those work just fine and your table really should have those for safety
sorry, that is from apollo 13
slundell: This way the homing sequence that is already in EMC with gantry kins can automatically sync up your gantry sides during the homing cycle
marley: what is the problem?
this problem is, joint 2 following error
dallur: oki. Do you mean that standard micro switches won't do?
can you pastebin.ca your ini file?
slundell: standard microswitches are just fine if they work for you, depends a bit on what you are doing with your table though, interference from plasma HF throws many switches
no, because i´m in windowsssss
slundell: I prefer solid state switches myself but that's just me :D
dallur: this is router that i build for fun. I have no plans on adding a plasma or such.
haw to configure ferror
marley: could be a few things. running the axis too fast for your base period, You don't have any headroom in the stepgen max velocity,max accelleration.
marley: steppers I assume?
slundell: then your microswitches are probably all you need, might want to check the repeatability though depending on how you mount them and what kind of switches they are
dallur: K, will have solid state switches on Z, because the actuator was shipped with them. microswitches on both Ys and X.
marley: it's impossible to help without seeing your ini file. You have a misconfiguration.
dallur: that was my next question, the repetability of microswitches...
in ini file ferror, is 20.
marley: it's impossible to help without seeing your ini file. You have a misconfiguration.
slundell: I don't have enough experience to help you with that question (or the info), probably just need to try it a couple of times and measure
friends watt is ferror?
marley: proper setup of the ini file will most likely never have issues with the default following error settings.
We really need to see your ini file.
with steppers, it just means you have configured the ini file incorrectly
i see it
please one moment
and i paste it in pastebin ok?
my paste bin is:
[19:54:05] <marley> http://www.pastebin.ca/327187
[19:54:03] <anonimasu> http://www.almaskin.se/univstep.ini
that's the config
my pastebin is http://www.pastebin.ca/327187
anonimasu, two things jump out at me: first, Z is the only one with nonzero backlash, second, Z has ~4x the resolution
anonimasu, can you also like to the .hal file(s) (or are they unmodified?)
you see mi ini file?
SWPadnos: the Z axis cant have backslash.. as it's damn heavy..
I see [AXIS_2]BACKLASH = 0.011
maybe I did set it.
I suspect it's a accel issue
MAX_VELOCITY = 5.0
MAX_ACCELERATION = 20.0
well, it's possible that the TP is mishandling widely disparate accels, but I thought that was all fixed
this is with head.
I don't think that's been broken lately ...
i do maker?
marley - I'm looking, please wait a little
is when I grabbed it..
marley: DEFAULT_VELOCITY = 2000.0
that is not correct
argh - is output_scale still being written to the ini?
SWPadnos: surely not
I'm led to believe that by the single spaces around the = sign
I think you could line up several people who are sure we took out all the "ini writing" behavior long ago
* alex_joni is amongst those people
I agree - that's why I ask incredulously ...
I took it out
and I'm someho pretty sure OUTPUT_SCALE is useless for steppers
yeah, which is a whole other topic ;)
SWPadnos: what do you think about the accel?
the motor's pretty small for that axis..
so that may be the issue..
but it occurs when jogging also
dunno - it's strange that it works fine on Z-only G0 moves
or does it?
anonimasu: can you step back a bit?
I wasn't around :)
my Z axis ( AXIS_2 )
misses when jogging X or Y togther with it
alex_joni, look around 20:42 (your time, assuming it's ~22:08 there now :) )
what is "misses"?
cradek: it goes "whirr"
anonimasu, are you using the USC in "internal feedback" mode?
or do you actually have encoders?
I do on X and Y but they are not connected..
they are servo axes..
whirr sounds like too much accel
but shouldnt a g0 be at the fastest rate possible?
ok, so you're running geckos with the USC, and are using the encoder counters to count the steps output, not encoder feedback ...
maybe MAX_ACCEL = 20 is too much
and you get ferrors, not just weird noises and a machine in the wrong position?
the stepper goes too fast..
"whirr" just like when thre's too much accel..
but I dont know where it comes from
ah - ok. I thought you were getting errors in EMC2
anonimasu: I suggest you scope the velocity/accel for that axis
you need to tune the accel/vel parameters
set the limits tighter to ensure EMC2 does what it needs
or what limits
alex_joni: I can scope it right now
if you hold on
SWPadnos: the thing is that a G0 should be at max_accel right?
then something is wrong
actually, any move should be at max_accel, I think
SWPadnos: when I do g0 it works..
at least any independent move
maybe 20 is just too fast..
that's 20 units/s
[20:12:45] <Jymmmm> http://news.yahoo.com/s/ap/20070122/ap_on_fe_st/tiny_apartment
anonimasu: is this in mm?
so it would acc to 20mm/s or 1200mm/min in one second
I cant remember
anonimasu: err.. 20mm/s/s
that means you get 20 mm/sec in one sec
so, for your max vel of 5 mm/sec, you need about 1/4 seconds
sounds like quite an accel
depends on the motors and the weight to move ofcourse, and ballscrew/vs normal thread etc.
I'm eager to see what ACC/VEL the 1100ozin motors and 2.5mm ballscrews will give me
once I get the servo electronics done, and the mechanics... so that means 1-2months at least
awallin: sounds like you will be able to form metal with a stationary spindle :P
I think the table weighs about 40kg, so from the torque and ballscrew pitch it should be possible to calculate some upper bound
it happens with 10 as accel..
Dallur: 1100ozin is the peak torque
anonimasu: I see lots of PID settings there
do those get used by the UxC ?
alex_joni: this works fine, but when I move another axis _whirr_
want me to plot them?
yes more data than "whirr" would be great :-)
that's the bug report :-)
I mean: does it go whirr by itself and whirrrr with another axis?
it goes whirr when I move another axis..
servos never go whirr.
they just move smoothly..
cradek: the stepper over accelerates.
ah, I already calculated it earlier, 1.6Nm torque with a 2.5mm pitch ballscrew, 80% efficiency should give about 3000N of force. with a table weighing 50kg I get: a = F/m = 3000N/50kg= 60 m/s^2 ! wow
I have emc2 running by remote now
awallin; don't forget about the force the cutter places
that's with the rated continuous torque... maybe I made a mistake, 6g sounds like a lot for this system
1G is damn fast
lerneaen_hydra: yeah, this would be for rapids only
uh, 1g isn't exactly fast is it?
yes 1g is fats..
anonimasu: well, I haven't taken into account the inertia of the rotating parts at all, I wonder how much that is
hmm, IMO it isn't :p
not super fast
about good enough though
try taking a corner, at 1g with a machien..
while machining is a different thing ;)
with or w/o it being anchored 5' into the ground?
just doing stuff in air, no cutting force seems easy
anonimasu: with or w/o it being anchored 5' into the ground?
Jymmmm: without :D
I mean, just take an object and drop it in your hand
doesn't seem to be unreasonably fast accel
anonimasu Awwwwww, that's no fun.... I'd liek to see the foundation shift ever so slightly
lerneaen_hydra: a 50kg object
the accel is the same regardless of weight
lerneaen_hydra: most hobby machines isnt capable of 1g
not hobby machines no
oh, was this a hobby machine awallin was going to add it to?
Heh, I can make my machine jump sideways
we have 2 feet of concrete under our big machine. but it is also what connects x/z to y
lerneaen_hydra: do you have any idea how fast 6g is..?
it was bit scary when I ran a OMFG-big lathe that wasn't anchored and did G1 moves
Heh, I can make my machine jump sideways if I crank it up, and that's NO WHERE NEAR a 1g
anonimasu; 6G is fast as hell, 1G is more "ok"
hm, so nobody has any idea about my issue?
I seem to remember the ballscrews have some limit on dynamic loading... maybe that will limit what acc I can get
I thought you were going to get us some plots of the problem
cradek: what plots do you want?
accel/vel of that particular axis=
a plot that demonstrates whatever you think is the wrong behavior
cradek: well, I'd think it would be when Z ignores the acceleration limit set
im getting a plot of the accel now
probably the commanded velocity and the actual velocity would be something to look at
what's thoose parameters called?
I can find accel but not velocity
err vel but not accel
vel should be enough
you probably would need to calculate acc yourself with a ddt hal comp
well, it should be visible..
it's a big difference
anonimasu: add another ddt
link the velocity to it's input, output is accel
a very basic test here (g0z1 vs g0z1x.1) shows sane accelerations
g0/g1 works well..
cradek: on the same config?
no, on sim
I mean settings
anonimasu: what doesn't work?
cradek: When I move X And Z the stepper accelerates too fast
faster then the limits should allow it to
anonimasu: read again what cradek said
(g0z1 vs g0z1x.1)
I also tried g0z1x1
trying to find what pin to check
anonimasu: check for signals, those are usually labeled better :D
I suck with halscope..
you may not have vel or accel signals - you have to generate those with ddt (see sim/axis)
hm, I have.. a vel signal
core_sim.hal or something like that?
do you just have to load it?
anonimasu: what motion causes this? you said not g0 or g1?
cradek: no, jogging
cradek: and that particular G28 command..
G28 is the same as one or two G0
but it makes it occur..
as it moves seveal axes fast..
cradek: select units?
g91 g28 is an odd combination
I was just pondering that
what is that supposed to do?
go to zero..
G91 Z0 isn't even a move is it?
G91 only selects
it can be on a line by itself, or concated with other moves
03alex_joni 07v2_1_branch * 10emc2/configs/ppmc/ (ppmc.ini ppmc_motion.hal): get rid of PWM_OUTPUT_SCALE in favour of OUTPUT_SCALE. emc2 doesn't write anything on shutdown anymore, so it's safe
03alex_joni 07v2_1_branch * 10emc2/configs/univpwm/ (univpwm.ini univpwm_motion.hal): get rid of PWM_OUTPUT_SCALE in favour of OUTPUT_SCALE. emc2 doesn't write anything on shutdown anymore, so it's safe
03alex_joni 07v2_1_branch * 10emc2/configs/univstep/univstep_motion.hal: get rid of PWM_OUTPUT_SCALE in favour of OUTPUT_SCALE. emc2 doesn't write anything on shutdown anymore, so it's safe
03alex_joni 07TRUNK * 10emc2/configs/ppmc/ (ppmc.ini ppmc_motion.hal): get rid of PWM_OUTPUT_SCALE in favour of OUTPUT_SCALE. emc2 doesn't write anything on shutdown anymore, so it's safe
03alex_joni 07TRUNK * 10emc2/configs/univpwm/ (univpwm.ini univpwm_motion.hal): get rid of PWM_OUTPUT_SCALE in favour of OUTPUT_SCALE. emc2 doesn't write anything on shutdown anymore, so it's safe
03alex_joni 07TRUNK * 10emc2/configs/univstep/univstep_motion.hal: get rid of PWM_OUTPUT_SCALE in favour of OUTPUT_SCALE. emc2 doesn't write anything on shutdown anymore, so it's safe
03alex_joni 07TRUNK * 10emc2/configs/etch-servo/.cvsignore: ignore copied emc.nml
sorry I had to run out a bit
Yeah, I think this error occurs when doing g0 also..
03alex_joni 07TRUNK * 10emc2/configs/dallur-thc/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/demo_mazak/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/demo_step_cl/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/halui_halvcp/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/m5i20/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/hexapod-sim/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/max/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/motenc/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/nist-lathe/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/sim/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/stg/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/stepper/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/vti/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/univstep/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/stepper-xyza/.cvsignore: remove core_axis.hal
03alex_joni 07TRUNK * 10emc2/configs/univpwm/.cvsignore: remove core_axis.hal
it's kind of disturbing, loosing your Z axis in the middle of a job isnt fun :)
03alex_joni 07v2_1_branch * 10emc2/configs/dallur-thc/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/demo_mazak/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/demo_step_cl/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/halui_halvcp/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/hexapod-sim/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/m5i20/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/max/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/motenc/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/nist-lathe/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/sim/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/stepper/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/stepper-xyza/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/stg/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/univpwm/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/vti/.cvsignore: backport: core_axis.hal doesn't exist anymore
03alex_joni 07v2_1_branch * 10emc2/configs/univstep/.cvsignore: backport: core_axis.hal doesn't exist anymore
[21:09:03] <alex_joni> http://cia.navi.cx/stats/project/EMC
hm, there's this show where there are some porn actresses, that want to be real actors. in a play
... and they'll do anything to get the role?
oh then forget it
they have to convince some judges..
and they suck :D
"uh, but then you have to like read books, and that takes time"
yeah, that's kinda what I was suggesting
some english reality show :D
did you get the time in the future on the cia page?
I'm getting xwin32 so I can run emc and get that damn plot..
I cant remember what I plottes last time..
but then I werent running head
57s in the future EMC
alex_joni: says "3 minutes ago" now
fjungclaus-away is now known as fjungclaus
well.. it was 4 minutes ago :)
now I have it running again..
later, im heading home
* anonimasu sighs
* lerneaen_hydra farts in your general direction!
your mother was a hamster and your father smelt of elderberries!
now how do you link that stuff so I can see velocitites?
anonimasu; you didn't get the reference?
lerneaen_hydra: the what?
yeah, it looks like my scope's broken today :/
no, the "your mother was a hamster and your father smelt of elderberries!" bit
anonimasu: loadrt ddt count=3
alex_joni: it's =6 right now
in the config
addf ddt.0 servo-thread
.. (for the others)
linksp Xpos-cmd => ddt.0.in
position into ddt0
output is velocity
linkps ddt.0.out => Xvel
linksp Xvel => ddt.1.in
linkps ddt.1.out => Xacc
if you look at sim/servo_sim.hal you'll see what I just wrote
doesn't the canonical encoder interface have a delts / velocity output?
but for all 6 axes
now lets see..
SWPadnos: wanna bet that the univ* doesn't have it ?
dunno - it's been a while since I worked on it :)
I never touched it.. except backporting it
yep. there's a delta output
ha, I was wrong :D
I thought I put that in there :)
(then again, someone may have added it after me :) )
you still need to get that into a ddt
not the delta
to get accel?
oh yeah - for accel you do
this dosent work
anonimasu: what doesn't ?
the addf stuff..
for getting accel..
what's the error?
why not just try to reproduce the error in sim/axis which has all this stuff already
cradek: maybe because the error is only visible on the motor..
it might not even be an emc2 error
simply a ppmc/UxC/univ* misconfiguration
hm, it's with the almost stock config..
I can't keep those boards apart :)
anonimasu: it's univstep.. right?
I see that has PID in it
I got it working
bet you need to tune it
alex_joni: it should be tuned..
i've done that before already..
when steppers go whirrr.. I think it's not tuned :D
why dosent it do that on normal motion
just when interpolated?
when you're using step-to-servo drives, and you don't attach the servo encoders to the USC, you shouldn't use PID (I think)
HAL:109: ERROR: systemv failed, returned 1
HAL:109: ERROR: insmod failed, returned -1
huh? how did you get that?
did you sudo make setuid ?
I added the stuff from sim into the file..
actaully, I take that back. you still need an I term to make sure the endpoint is reached
anonimasu: what emc2 are you running?
you need to change the ddt num=6 to num=9 (or whatever 6+the number you need for scoping is)
oh, and you can't have more than one loadrt ddt line in the hal file
* alex_joni goes to bed
nighty night alex
this stuff makes me insane.
saying good night?
col: really funny
having a axis that dosent work as supposed..
and not a damn clue how to even plot anything
have you checked the realtime latency?
maybe you need more gigabytes
this is with a usc
try it with the power to x/y disconnected
fenn: surely that'll just ferror x and y right away
well, turn up ferror too
jepler: as I have no feedback i twont..
also I have a 200A power supply..
or was it 300..
could be noise on the lines or something
unlikely, since the encoders aren't being fed back to EMC
no, noise on the step/dir lines
I doubt that very much.
got any other ideas?
fenn: I'd get gross errors on the servos.. also..
fenn: they stay still when jogging other axes..
if you turn down z accel and top speed,does it still happen?
top speed, I dont know
max_velocity is 5 btw..
fenn: _if_ you want I can put up the scope on the step/dir lines..
but I've ran a scope on other stuff and I have no noise whatsoever
wait -where is your servo loop? is it just in the gecko's?
well, ruling out a possibility never hurts
it's just in the geckos.
so how are you getting following errors? are you not just sending out step and directions
skunkworks: 2 g340'
it's not following errors..
it's physical errors on my Z axis..
the only stepper on my machine..
it overspeeds when jogging X/Y togther with it..
or over accelerates
did this just start?
but I havent made lots of parts and I've seldom experienced it..
I could take it at the end of a part..
but, now that I need to make some parts it's a bother..
Sorry - was not up to speed. are you also running a gecko for the z?
but a stepper gecko
friends please explain me what he is ferror and as I can decide my problem
joint 2 folinwinf error
joint 2 folinwing error
it goes away with accel set to 2.5
err vel to 2.5
watt is accel
st my acceleration in 2.5?
its i slow
ser my accelaration in 2.5?
set my accelaration in 2.5?
we are troubleshooting mine :D
dont set it to that..
how to function off ferror in file ini?
but you can increase it
ultimately you want to take a look a P I D tuning also
i have one machine for client
i need yor help
I can try to in a bit. im out by my machine
and I have shitty kbd.. so chatting is a pain
friend you say me watt function off ferror in file ini?
You cant turn off ferror, but you can increase it.
ferror and min_ferror
You cant turn off ferror?
friend i have one motor for axix Z
what kind of machine is this?
and this motor turn 200 pulse for one revolve
what a ferror fot this?
I have no idea what your settings are
look in the INI file and increase it
im my file INI
ferror is 20
is that mm?
that's pretty large
watt i make?
i dont ferror?
Look at pid tuning in the wiki
I do not know what he is ferror
it would like to understand the concept
IÃ can explain it when I get inside
thus it could decide my problem
it could teach to me?
ok, it describes me for me.
Hi bob. This is a stepper driven machine?
rayh: as he said 200 steps per revolution I would assume it is :)
bob: ferror is how much the machine is off where it should be
off from where it should be
it is a stepping machine.
bob: internally in the EMC step generator
yes... steps motors ...
skunkworks: I found my error
too much accel..
I can go to 500mm/min with the steppers..
but with a low accel..
so im going to gear it 2 times more then now..
I run around under 10 inches per second^2 on my machine.
I need 200mm/min and nice accel..
accel is everything speed is not
bob: What configuration are you using?
atleast for a usable machine
ok... one moment please... i'm go my configuration...
I take a beer for my credit, until to see everything.
1 beer for my account takes you in www.virtualbartender.com... :)
sorry in www.virtualbartender.beer.com :)
I guess everyone went there and typed in "catfight" ...
I can double my speed/accels
of the Z axis
joint 0 on limit swith error?
exactly being with the set in motion keys or not.
hello huston... other problem...
bob: copy/paste your .ini to pastebin.ca
oh, limit switch error? hmm.
paste the .hal too :)
my pastebin is http://www.pastebin.ca/327468
what is the voltage on parport pin 10?
some wrong thing?
x home is on the same pin as all of the limit switches?
please one moment
fenn: two parports
this file ini is backup
on machine x home is pin 13
bob: what does 'halcmd show sig limit-reached' say?
type halcmd show sig limit-reached
in terminal or axix?
in a terminal
fjungclaus__ is now known as fjungclaus-away
i go make it
terminal returns this:
bit TRUE limit-reached
ok i get the idea
if you press the limit switch, does it close the circuit or break it?
put a halmeter on "limit-reached", and watch it while you manually trip the switch
ok then you want to change parport.1.pin-10-in-not to parport.1.pin-10-in (in the .hal file)
my brain just died again
* fenn weeps
you want to put a halmeter on the signal, and see if it does the right thing (true when you hit the limit, false if not)
if it is backwards, then change from parport.1.pin-10-in-not to parport.1.pin-10-in
USE the meter, thats why we have it
halmeter and halcommand give you visibility into what is going on - not using that visibility is stupid
i not use swhitch, i use a sensor
can you trip the sensor by moving the axis manually? or by waving a piece of steel in front of it?