Does anyone know if it is possible to control the layout of PyVCP radio buttons? O would prefer a 2x3 layout to a single column
Sure, you can change the layout, but not sure how to do it in such a way that you could save/reapply your changes after the next update.
there are hbox and vbox elements, which arrange their children in horizontal or vertical arrays
the default is more or less a vbox, I guess
(a vbox containing two hboxes, each with 3 buttons, would give you a 2x3 arrangement)
I think it would be simple with checkboxes, but radio buttons are a "family" and all defined in one line of the radio button XML
oh hmmm. then no, I don't know if that's possible
toss in a linebreak maybe
it's the radio button control itself
Are you suggesting a line-break in one of the tag names?
Well, not necessarily a "\n" as much as <br />
It would have to go in the middle of the string array, I can't see that working?
try it in various methods
Though by no means all possible methods, I am sure.
post the file on codepad
post the file on codepad.org
I might have a look through the Widgets code to see if there are undocumented parameters.
I can paste the code here, is is 4 lines....
Or, you could post the ENTIRE file on codepad.org
That's all there is...
If you think that the rest of the file will help, I can do that..
Lines 67 to 73 at http://www.pastebin.ca/1818146
pygtk.org appears to be down, but I'd look there (or somewhere else) for information on making gtk radio button groups have multiple columns
I dun know...
I think I tried that way, and it didn't work.
jepler should know
Though I ought to try again as I might have included another error
Aye, or Anders, looking at the pyvcp_widgets.py fle
I thought that the syntax above crashed Axis, but actually it was that the pins were missing that were referenced in custom-postgui which were the problem. Splitting the choices means that only the ones in the second line get created.
Dumb Question - will emc2 directly control thru parport/optocoupler the ssr's on a hardinge CHNC?
I would imagine so.
I am controlling SSRs, Mechanical relays and VFD button-inputs with the outputs of a Mesa card and a buffer IC
SWPadnos, has anyone successfully loaded mesa firmware with that kernel? It getting the same errors.
hmmm the spindle is moving
* Jymmm sends a priest to JT-Hardinge's to perform an exorcist
across the room? is that good?
clytle374, no idea
skunkworks: Why do you think I'm sending a priest for
SWPadnos, thought maybe you were the owner
nope. I think mozmck built those
not sure though
what mesa card?
gee thanks Jymmm
at least it moves
JT-Hardinge: Hey, no demon possed machinery, m'kay
unless you get good videos
JT-Hardinge: That goes for zombies too
5i20 but need to patch emc for the 7I48 card.
I don't have any mesa cards (yet) so I can't even test that.
what error do you get?
I'm trying to get newer packages made soon now.
loads hostmot2 fine.
there should be an error, either printed on the terminal or in dmesg
soory, i'm trying to get it.
ok, stick a log of some sort on http://pastebin.ca
when you get a chance
I wiped out a bunch of stuff, looking for the firmware now.
I'm loosing my mind obviously.
makes debugging harder...
Danimal-office: my bike frame is on the internet: http://www.flickr.com/photos/mcn7/4398766376/sizes/l/
* JT-Hardinge watches the spindle spin while he tries to finger out what controls it LOL
Okay, wasn't logged in in that terminal, didn't think I deleted it all. This is trying to load the firmware. I have tried lots of firmwares. My own 64bit smp kernel
then 64-bit kernel, then decided it was a 64-bit issue and installed 32bit, now the experimental kernel,
whats a good base mesa card and where do you buy it?
here is dmesg http://pastebin.com/0cK0T1ty
huh. well, I don't know what's wrong by looking at that
I'm assuming that you compiled emc2 yourself, since you need a patched version of hostmot for the 7i48 ...
It looks like you might have the problem I often have there, a failure of RTAI to exit cleanly.
Bah! he left
whats a good base mesa card and where do you buy it?
Oddly enough, I was plotting making a bicycle frame myself this afternoon. Out of 48mm scaffolding tubing.
7i43, straight from mesanet.com
woa usb based?
7i43 is normally used P-Port based
PCI or parallel prot
there is a USB connector on the 7i43, but I don't recall just what you can do with it. EMC2 doesn't use it
will the mesa card work on a second pci parport card with a pico system installed?
sure, as long as they're on separate parallel prots
you can mix and match any drivers as you like, subject only to CPU and I/O processing time
But it seems daft to install a PCI parallel port to connect to a Mesa card, when you could fit a PCI Mesa card (though it might make for simpler cabling in some situations)
not much simpler, as you can't use more than 6' or so of parallel port cable to a 7i43, IIRC
now all I have to do is get the pci card installed. do you normally have to edit 3-4 files to get it working?
it's still nearly $100 cheaper to use a 7i43+parallel port card, but it gives you less I/O and worse performance
at least two, could be as many as you like ;)
actually, you may not have to edit the ini file, so the minimum is only one file
clytle374: Type lsmod, what do you see?
the install intructions on the card are poining to about 4 files in the etc
I think that that pastebin might be a failed HAL boot due to the rtai_sched and rtai_math modules not having quit cleanly.
SWPadnos: Does the pci mesa have a limitation on cabling length?
it was a clean reboot, once it fails you have to reboot. lsmod http://pastebin.com/KXgQeczC
Jymmm, it's likely to be cable and noise limited
crappy time to loose net connection
SWPadnos: So you cant run 25-20' cabls to steppers?
that's an lsmod with emc running?
you can't unload it after the error.
It might be worth removing lp from the load-at-boot modules file. (and just about everything else too). I can't remember the name of the file that lists them though.
lsmod on my machine only shows 5 modules, and 12 with emc running.
clytle374: you need to block parport_pc if you are using the parallel port
I'm just doing a halrun; loadrt hostmot2; and then loadrt hm2_pci config=" blah blah
Which emc version?
I think that requires parport_pc then
What do you have in your /etc/modules file?
Does it require it? I know you don't have to block it anyhow...
for anything greater than 2.3.x that is...
I am a bit fuzzy on it, but seem to recall that it had to be a loadable module, not an installed kernel function, when I built my custom kernel
/etc/modules contains lp
And that is all?
Probably not the problem, but you could try commenting it out.
Does it matter since I'm olny trying to load the mesa firmware?
I'll try it.
I wonder why all the file systems are loaded?
Hang on, is this a PCI card or a p-port card
Ignore me, I have the wrong end of the stick then.
But it does seem a very extensive modules list.
Not sure why, there wasn't that many with my kernel
2am, need to sleep.
This is new. The the kernel is loading com20020_pci for the mesa card. I'd think that was the problem, but it wasn't happening before
Kernel driver in use: hm2_pci Kernel modules: com20020-pci
That's interesting, I was going to ask what that was.
I didn't even have that built in my kernel so that wasn't the problem
The card was getting an irq of 255 on the 64bit kernel
I've installed ubuntu 9.10 64bit twice, 32bit twice and about 40 kernels between them.
Full time job for 5 days now. Starting to want to throw things
I get real good latency test
what kind of latency are you getting?
I posted it here earlier, don't remember off the top of my head
I get about 15,000
BUG: unable to handle kernel paging request at f8e32046 Bug in what?
I have a major linux/emc problem. I was on earlier in the day asking about a step_conf wizard session that resulted in a system that would hang the computer when EMC2 booted - freezing up everything. I got some good ideas to try BUT since then I have discovered that even booting off the CD and running the sim "fake" machine causes the whole PC to hang. The same CD loads on another computer and runs OK. The PC in question has a 2.
Guest404: running sim shouldnt hang the computer. Does it pass memtest (also on the EMC2 LiveCD, i think)?
it's a sim config, not running in sim (non-RT) mode
so it still loads RT, but no hardware drivers
SWPadnos: Thoth Industries? ;-)
seb_kuzminsky, did you see the pastebin from clytle374 ?
no, Thoth Systems :)
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-03-02.txt
huh. I wonder why my backscroll buffer is suddenly only about 100 lines long
but I don't wonder enough to stay awake trying to find out. night all
ah that sucks
clytle374: hi i just saw your pastebin with the hostmot2 bug
is the bug in hostmot2?
i'm not sure yet, but it does look suspicious
is this on a 32-bit machine or 64-bit?
Thanks for any help, I'll do what I can to help.
It is a 32-bit machine. Ubuntu 9.10.
i've never tried it on 9.10
i also dont recognize the bitfile, where did you get it?
I had the same error a few days ago on 64-bit smp, tried non smp, then 32 bit
I built it for the muxed card. I have tried 'factory' bit files with the same results. It was the only on I could get quick for the post
I'm using the experimental 9.10 packages now, trying to drill the problem down
I can load the firmware with the live 8.04. Not my firmware, but I kinda expect that since it emc requires a patch.
The patch is not applied to the current machine
clytle374: Do you have a particular reason for al this customization you keep applying unsuccessfully?
It is needed for the 7I48 daughter card. emc is straight from git master.
rtai was magma CVS, not sure what is in the experimental packages. kernel is 126.96.36.199 or 188.8.131.52 for the experimental packages
I thought it was to get a little better latency (reason for the custom kernel)
skunkworks, While you are compiling a kernel, why not remove unnecessary items? That was the game at first, but I didn't have my mesa cards yet either.
Now I'm just trying to get something to work
clytle374: try this: http://pastebin.com/0cK0T1ty
because the git version of emc will run fine on the stock kernel... ;) (I do stuff like this all the time) Ask me about my 'simple' h-bridge I spent a lot of time on that I probably am not going to use any time soon.
i noticed a missing sanity check on the incoming bitfile, this patch adds it (plus some belts & suspenders stuff)
seb_kuzminsky, I don't see a patch, looks like my dmesg
[03:04:25] <seb_kuzminsky> http://pastebin.ca/1818342
i forgot cut-n-paste buffers dont carry over from virtual machines to the host machine :-/
I see a download link.. wget?
there's a "download raw" link on the left, click it and save the document you get, apply it with "patch -p1 < FILE" in the root of your up-to-date master checkout
then rebuild and try it
i'm holding my breath and crossing all my fingers
I'm working on it through ssh, too cold where the machine is at
fingers and toes
ugh, sounds like my garage...
this is vanilla master? or do you have local changes?
or a 200 year old part of the house
no changes. trying to make your life easier;)
thx, i appreciate it :-)
to make my life even easier, would you try loading one of the standard firmwares first? then if that works we can try your custom one
bad news, same thing.
SV12.BIT. easiest to type
is the dmesg exactly the same? can you pastebin it?
here it is. http://pastebin.com/6f5UrdPH
I will be back in 10 min, battery is dead
Back home and the netbook is reacquainted with its charger
horray for netbook
Best $250 I ever spent.
(applies to clean master, not on top of the previous patch)
bunch more debugging output, no real fix tho
okay, can the patch be backed out of? I know I should have a back up, but got pissed and was taking shortcuts.
"git checkout FILE" will revert any local changes to that file
or if you're feeling masochistic, "patch -R" removes a patch
git checkout src/hal/drivers/mesa-hostmot2/bitfile.c was that it?
should I have make clean?
just checkout bitfile.c, apply the new patch, and run make
good takes longer. My desktop is 64bits so I had distcc on it to speed up the endless rebuilds, but doesn't work so easy/well between 32/64 machines
I guess that isn't needed
are you using distcc spread across 64-bit & 32-bit machine to build emc?
no need for configure
same thing + more I remembered to clear dmesg this time, http://pastebin.ca/1818390
No distcc was not installed on the 32 machine, only when I was building 64 bits. Causes trouble if you mix them up
i have trouble calling a sub file. the examples i wrote years ago on the wiki dont work. heres a simple test. http://pastebin.ca/1818392
clytle374: well that looks good
did request_firmware() change between 2.6.24 and 2.6.31? is the firmware that the kernel fetches read-only now?
I don't know about the change. It is owned by root now, I'm pretty sure it wasn't in the past. Wan't me to change it?
jt-plasma: the oword file call here doesnt work http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/catid,20/id,958/lang,english/
i got a good feeling about this
(Stripping trailing CRs from patch.)??? Okay?
why? doesnt it apply cleanly to top of master?
git checkout src/hal/drivers/mesa-hostmot2/bitfile.c
patch -p1 < 1818413
why add all the unnecessary parentheses in that patch?
because i'm not very good at c operator precedence...
I didn't get an error, just several of the (Stripping trailing CRs from patch.)
+ and - are higher precedence than deref
clytle374: should be fine
no they're not...
oh, that's unary + and unary -
see, i told you i'm not very good ;-)
- chunk->data = &fw->data[*i];
+ chunk->data = &fw->data[(*i)];
no offense, but this one's extra silly :-)
ok ok i'll take it out before i commit
are you guys finding the 64 bit problem (if there is one)?
i think (and clytle374 is checking) that in 2.6.31, the buffer you get back from request_firmware is mapped in read-only
oh so you can't byte swap in-place?
new to me message make: warning: Clock skew detected. Your build may be incomplete.
so now i byte-swap as i write it out instead
the kernel guys sure change stuff around
they like to keep us on our toes
if that's even what it is... we'll see
they have no concept of not breaking APIs except in major version number changes
shall a undo that patch?
clytle374: something went wrong with your build, the function that's erroring in that dmesg is not called any more
cradek: they're talking about dropping the "2.6" prefix, so each new kernel is a major version number change
I was going to say they ought to just use increasing integers, because they have no major/minor organization - but I figured it would just be whiny
so, it's funny you should say that.
um, what now?
cradek: remember that control i showed you i said i could get from Fadal?
they said i can come get it for $400
comes with the pendant and everything, including a bunch of fanuc cards that i can ebay
thinking about it... more than i want to spend, but it is brand new
looks to be the perfect control since it has the keyboard and all the high quality buttons and pendant, including the big swivel arm
clytle374: how did you apply the patch?
yeah, seems like plenty of $ for 'parts'
cory@cory-cnc:~/emc/emc2-dev$ patch -p1 < 1818413
well that looks reasonable...
[04:18:26] <Danimal-office> http://kzekkani.com/Projects/104.gif
try "grep reverse ~/emc/emc2-dev/src/hal/drivers/mesa-hostmot2/hostmot2.c"
think it's a little steep cradek, or would you jump on it?
i know, you prefer touchy
re: file call this o-word file call facility just doesnt work, the code is ignored/skipped over, no errs occur. i have no idea how far back i'd need to go to get it to work again. others have reported similar in last few days. zramaro 2010/02/24 08:01 is this feature abandoned?
I'm still doing halrun; loadrt hostmot2; loadrt hm2_pci config="firmware=5i20/SV12.BIT"
i think somehow you're getting the wrong hostmot2 module
heh, doesn't matter whether I'd buy it...
how did you compile after your applied the patch?
it is run in place
I guess you have to compare it to buying a bunch of steel, welding it up, and painting it... that's not very quick/easy/cheap either
tom4p: is there a full bug report somewhere I can see?
what can i get for you?
example was posted at http://pastebin.ca/1818392
sample program that shows incorrect behavior. it'd be best if you'd put it in the sf bug tracker because it might be something ken needs to handle, and he doesn't monitor here
cradek: i figure the pendant is at least worth 1/4 of that, plus each button with the contact block is at least $50 new, and the industrial keyboard isnt cheap either
uh, i didnt know sf was involved after move to git, will look at sf
tom4p: what's the behavior? I see a M2 in xxx.ngc which surprises me
M2 is in the dcos on wiki and .org, even tho it doesnt make sense
m2 to end a program?
the behaiviour is that the called file doesnt get exec'd, just skipped over
the emc post for mastercam uses m2 for the end of programs
(m2 doesnt make sense because it doesnt get interp'd becuz its after the return to caller)
tom4p: yeah we still use the trackers on sourceforge
and code in main DOES get execd after return, just squat doen in middle
cradek: the lathe throws a ferror in x, but only in one direction... it also sounds a little rough in that direction. i think it's a brush maybe.
tom4p: this isn't my area of expertise but I think Ken L will hear about it if you make a bug report.
Danimal-office: could be motor or tach - scope will tell you which
clytle374: ls -al hostmot2.ko hal/drivers/mesa-hostmot2/hostmot2.c
in the emc2-dev/src dir pls
how will scope tell me? what should i look at/look for?
put the scope across the tach (don't ground one side - use invert + add) and see if it's smooth
ls: cannot access hostmot2.ko: No such file or directory
-rw-r--r-- 1 cory cory 35302 2010-03-01 22:58 hal/drivers/mesa-hostmot2/hostmot2.c
I jsut did make clean, and am making during that command
oh, that's why
you should just about never have to make clean
sf has sedonc 'open cnc system base on emc2' and emc^2 'EMC² - Eclipse Mail Client' when emc2 is searched
but it shouldn't hurt
I really, did something change with that in the last few years?
with make clean? i dont think so
cradek: you mean a physical scope, or halscope, using a 5i20 input?
okay, good to know
crappers, i dont know anyone with one
nor do i have one
I've got a spare you can have, it only needs one hard-to-find part :-(
cradek, like the tube?
i guess i should buy one maybe
but i'm sure they cost as much as another HNC
no, if I remember right, it's the vertical amp IC
Danimal-office: you can get plenty of scope for $100
never really looked at one since it was all greek to me
yeah 80s technology is what you want :-)
it'll just match everything else in my shop
i swapped out the pc in my mill, seems to work better
seems to have fixed the flakey jogging
and it has a real agp card in it now
it was getting 10k for the servo thread, and 14k for the base thread... not perfect good enough, i would think, huh?
doesnt seem to lock up when i run that long program anymore
not perfect, but good enough*
ack, goodnight folks
gnite artifact successfully created
almost done, sorry I got anxious
clytle374: still working on that atom?
giving you problems?
make: Leaving directory `/home/cory/emc/emc2-dev/src'
make: Entering directory `/home/cory/emc/emc2-dev/src'
it's running in a freaking loop
did you configure with --disable-build-documentation?
i'm not finding anything about request_firmware having changed between 2.6.24 and .31, so maybe this is a red herring
it's Depending the same 20-30 files over and over.
well that sounds wrong
do you have any goofy symlinks in your working dir?
I didn't put any there on purpose.
It still had hal hung up in the background, reboot and seems to be notlooping
I mean it is looping :(
make: Warning: File `config.h' has modification time 2.3e+03 s in the future
dude your machine ain't right
I'm getting that feeling too, wouldn't be the first barnd new something that was broke
Just having a hard time believing that I can have the same problem in 2 different OS, but not the live CD
cant find my windows cd :(
and it be hardware
Danimal-office, don't even look for it.
i need it
cant use mastercam or solidworks on linux
and more importantly, cant watch netflix on linux
Danimal-office: You can in a VM
This is also the first 'weirdness' I've had with this thing. Never had a problem with outher things I've built.
huh, look at this: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=8d8fe198c6a756ae96617548af4776e7a86c0d3e
Having hal going crazy in the background can't help.
Jymmm: wouldnt that be worse than running a windows machine?
and wouldnt i still need to find my windows cd?
Danimal-office: Not at all. But you either have to have a pre-built VM, or installer for hte OS you want to use, then just backup the VM you jsut created, and save for future installs
seb_kuzminsky, In my non expert, and usually flawed opinion.... I think you found something
I have a generic XP VM I built, and I just copy that when I need different XP VM's
so you're basically running an OS with an OS?
I've got a new tree, should I apply the last patch?
and the machine wouldnt be slower than if it just had windows on it?
clytle374: sure, try it
Danimal-office: Well, dont try to use/setup a VM on a machine from 1942
Danimal-office: No cheap bastardness allowed
i have to be cheap
it's my MO
Then no NetFlix for you!
well the problem would probably be solidworks
that's a bit of a resource hog, as it mastercam
the computer i have is just a 2.8ghz p4 HT with a decent video card
working on a laser etcher hopefully controlled by emc :)
a really really small one :P
i want one
with cd burner infrared diodes
can it burn metal?
probably will start with dvd diodes
seb_kuzminsky, so if that is it, how hard is the fix?
as they are less easy to destroy
Danimal-office: no, sadly :(
cutting metal with a laser is pretty insane
I guess it will have to be fixed anyway, unless it's going to be 8.04 forever
there are far better ways to cut metal :P
well i mean engrave metal
* TD-Linux is still on 8.04 on his laptop
Danimal-office: no, only plastic and wood
unless I have really soft metal
i need to build something that can etch anodized aluminum
clytle374: it'll have to be fixed in emc 2.4, which will come out in April or May or so
this might do it if you anodize it black :P
so thanks for testing it and letting me know
the last patch i sent you has the fix if the firmware gets delivered to the driver read-only
I try to contribute, only qualified for testing and some docs..
The firmware only needs to be written anyway, right?
famous last words
the firmware comes out of the Xilinx VHDL compiler in a different format than what the FPGA wants, so the driver has to transform it
it's slightly cleaner to transform it in place, but it can be done as in the latest patch: as it's written out to the FPGA
is that the bitfile_reverse_bits_of_chunk message?
yeah, that's the function that tries to transform the firmware to the FPGA's preferred format
well, we got a different message insmod: error inserting '/home/cory/emc/emc-test/rtlib/hm2_pci.ko': -1 Unknown symbol in module
and dmesg http://pastebin.ca/1818519
here's the commit that made fw->data readonly: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;f=drivers/base/firmware_class.c;h=8a267c4276291bb775e2590a6935825ae35e9305;hp=d3a59c688fe4fcaf0f7ebf514f44f59a08e90693;hb=6e03a201bbe8137487f340d26aa662110e324b20;hpb=091bf7624d1c90cec9e578a18529f615213ff847
I don't understand, git pull?
Now I understand A kernel git not a emc git.
ok try this one: http://pastebin.ca/1818530
it adds the missing symbols to hm2_pci (and hm2_7i43)
i should have seen the error before sending you the previous patch
wich rt lib is --with-realtime looking for ?
I shoulda done lots of things, that well didn't happen... No big deal
patch a clean tree? or on top of?
Tecan: it's in /usr/realtime-*
clytle374: clean tree
"git checkout" in the top dir resets the whole tree
im on linux mint and i installed the realtime kernel
linux mint is close to ubuntu though
uses all the same debs
whoa, motor controller in fgpa?
TD-Linux: best place for it ;-)
as of 5 minutes ago I got my apple II in my fpga working
by "working" I mean the only external I/O being the speaker
seb_kuzminsky: for simple stepper control I can see it, but for servo motions and stuff I don't think fpga would be optimal
TD-Linux: cool :-)
seb_kuzminsky, doesn't work in a copy. I copied the tree this time.
TD-Linux: only part of the servo control that's done on the fpga is the pwm output and the encoder counting
emc2 closes the loop between those two up on the PC
which is the latest rtai ?
4 or 24
[05:41:07] <Tecan> https://www.rtai.org/RTAI/
seb_kuzminsky: okay that makes sense... FRC robots do the same thing
seb_kuzminsky: what kind of interface goes between that and the PC?
oops i see 3 in there too
TD-Linux: EPP, PCI, PCIe, PC/104, or PC/104-Plus
that's quite a variety :P
it's a pretty flexible interface :-)
3.8 would be the latest i guess
I have no clue how PCIe works
TD-Linux: it's like PCI, biggest difference (i think) is that it has good support for variable width busses
I should add a PCMCIA interface for my laptop
... actually, my laptop's docking port has PCIe... that could work :P
PCMCIA is being replaced with ExpressCard, which is PCIe in a different form factor
clytle374: how's the build coming?
jsut started it. I think the harddrive is failing.
It was used. it came out of my grandma's computer. But free ofset the fact the computer died when grandma house was hit by lightening.
I shoulda known better.
It's so hot I'm surprised it's working, hang in there.
If this works it will probably work in 64 bit too.
well i'm off to bed, clytle374 please tell me if it works (here in irc is fine, i'll see it in the morning)
thanks again for the bug report & all your testing
;/ I got this agian BUG: unable to handle kernel paging request at f8e3204d
wait what? where?
just like before
I double checked that installed the right patch
argh, i sent you the wrong patch, sorry
5th time's the charm, i always say
[06:07:07] <seb_kuzminsky> http://pastebin.ca/1818570
ok now i'm off to bed, see you tomorrow
Thanks for helping. I was starting to go mad thinking I was screwing it up
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2010-03-02.txt
making any progress?
I think. building now
git checkout seems to list the changes but not revert them
I'm sure that was the hard way. But patch went in this time.
Danimal_office, did you end up buying one?
clytle374: no, i didnt yet
i think i found one of my mobos that i thought was dead actually works
the video card was bad, and a stick of ram was dead as well
once i find my windows disk i'll know for sure
I really hope it works. Cheap and real quick. Not fast at building software. But would make a real nice desktop if you didn't do games, cad, or stuff like that/
oh, I thought you were building a control
yea, it looks like a nice little home theater pc
clytle374: yes, well, i was going to put the atom in my lathe, and take the dell that's in my lathe and use it as my windows box. My current windows box i just put into my mill, because the pc in there was kinda crappy
kinda playing musical chairs
probably be good for that. I might get one as a file server, to replace the desktop that's on 24/7.
doesn't use much power.. The fan is on the chipset, not the cpu
yea probably more efficient
i'm just trying to use what i have right now so i can allocate some funds for other stuff
nothing like new socks
I know that feeling.
testing now, I can't bare to look
Now to see if the program light went on the card, it didn't crash.
so far so good?
After 6 18-20 hour days trying to get something to work.... I just lost my purpose in life;)
nice job, congrats
seb_kuzminsky, IT WORKS. Thank you.
Danimal_office, tell seb_kuzminsky that. I just ruled out operator error.
Not to try 64bits..................In the morning.
Now to try
well i'm off to bed, adios
good night all. :):):)
if i were trying to build rtai it keeps saying my kernel tree is not configured
how do i configure a kbuild ?
Tecan: did you read the RTAI howto?
Tecan: here's the ubuntu RATI howto http://neo-technical.wikispaces.com/emc2-ubuntu
* BlackMoon looks to random new burns on his hand and arm
now I remember why I hate turning steel.
get a little too agressive and your lathe spits out a shower of firey doom
I always wear long sleeves
and you gotta stick around to turn the fucker off.
archivist: good *ouch* recommendation *ouch*
* BlackMoon ices a burn
What about the burns on my hands though?
I mean I guess I could go back to HSS tooling :)
Its only the blue and black chips that bother me
silver/straw colored just bounce right off me like bullets off superman!
* Tecan waves to BlackMoon
only slightly less dramatic.
thats strange its asking for qt3 when compiling the kernel with make xconfig
they should make coolant for your hands.
make oldconfig is working better
or make menuconfig
ah yes much nicer ;)
micges_work1 is now known as micges
i was in kazan', russia for two weeks. my wife lives there
and during this time i was robbed
here in minsk
you don't live with her?
ewww that's not good
i do live, she studies there
i will maybe move to kazan too
but i dislike russia
I'm afraid doing bisness with them:)
)) yeahm they are much too barbarous
but they are clever too, but well :( very erhhh naaasty
izua_ is now known as izua
hi to all
i have an emc2 question, there is anyone?
just ask the question
in axis there is a signal " axis.0.joint-pos-fb" that it gives back the position of axis 0 (x in my case)
how I make to read the variable from a program python?
helich is now known as bassogigas
is it a difficult question ?
ah...sorry for my bad english.......
eng is enough. no one knows it seems. try to look through different examples
Guest103: you can use the position from emc2 status
maybe there you'll find the answer. in particular, cradek's site
I don't think you can access the joint-pos-fb directly
the only way to do that is export a pin, and link joint-pos-fb to that pin
simply , i need to know the axix 0 position into a python program variable
Guest103: use emc.stat
what is emc.stat?
a class for the current status
so, when i exported pin , how i read from python?
s = emc.stat()
err.. 'import emc' first
you select which axis you want there, 0 in your case
you use either s.position or s.joint_actual_position
alex_joni: emc.nmlfile path must be set
if you have a trivial kinematics then the 2 will be equal
what micges_work1 said
Guest103: here's a working example that uses position to print it out: http://git.linuxcnc.org/gitweb?p=emc2.git;a=blob;f=src/emc/usr_intf/axis/scripts/teach-in.py;h=bc1c52e9e7291091cfb99ae2ad10329d39ff0bcb;hb=HEAD
well alex_joni, now i go to lunch,here is 13,00
in the afternoon i'try it, thank's
hmm why last G0 line desn't show on preview? http://www.pastebin.ca/1819108
micges_work1 is now known as micges
* Jymmm lol @ JT-QuadCore
damm gateway died again
* JT-QuadCore said well I'll just get another mobo for it
WRONG! the card slots on the gateway are bassackwards from normal
* Jymmm waiting for JT-286, JT-386, JT-386SX25, JT-386DX50, JT-486, JT-Pentium, JT-P2, JT-P3, JT-P4, etc
and JT-XT, JT-AT
JT-CP/M, JT-Atari2600, JT-Atari9600, JT-C64, JT-Coleco, JT-PET, JT-Apple2e, JT-TRS80, JT-Z80, and finally.... JT-UNIVAC
Why are you using a PC for a gateway?
a gateway PC
should I did out my 286 mobo?
Oh, Gateway (tm)
you remember it the one that won't boot up if the power is dropped
Yeah, paint a bullseye on it and grab the rifle
I'm thinking 44 magnum
ammo is too expensive, use a .22lr instead
* JT-QuadCore wanders out to the shop
we reload them
* JT-Hardinge tries to figure out what is missing from the spindle part of my config
toy for the USAians who want a little accuracy http://cgi.ebay.co.uk/HP-Agilent-5518A-Laser-Head-Machine-Tool-Interferometer_W0QQitemZ300402310940QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item45f15f7f1c
I have Spindle: and the +- buttons in Axis but the start button is not there, what did I miss?
hi JT-Hardinge, did u have spindle-cw and spindle-ccw linkes etc?
ha! motion.spindle-on does the trick
getting closer, I can make the spindle spin but can't make it completely stop
* JT-Hardinge off to work
work on it tonight
re: "George Bulliss... has been working with WCC to get EMC a nice place to meet during the CNC Workshop", will this be free or do the emc workshops have the 100+$ cost attached?
archivist: i imagine the interferometer is for sale w/o the retroreflector due to some ugly crash while setting up comp tables
tom3p, I'm pretty sure we all have to pay the attendee fee ($125 for the next couple of weeks, then $150)
clytle374: glad it worked! thanks again for finding the bug & testing the fix :-)
can't get much more binary than that http://www.electronicsam.com/images/KandT/conversion/shiftlogic.JPG
i'll commit it to 2.4 & master today some time
seb_kuzminsky: what was it?
cradek: the linux guys made the firmware data readonly in this commit: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;f=drivers/base/firmware_class.c;h=8a267c4276291bb775e2590a6935825ae35e9305;hp=d3a59c688fe4fcaf0f7ebf514f44f59a08e90693;hb=6e03a201bbe8137487f340d26aa662110e324b20;hpb=091bf7624d1c90cec9e578a18529f615213ff847
the fpgas want the bits reversed from how the xilinx tools spit them out (mirror each byte left to right)
lovely! so it wasn't 64 bit at all.
right, all this testing was done on 32-bit ubuntu 9.10
so we'll need that fix in emc 2.4
i'll commit it later today, without the mockery-inducing parentheses ;-)
is there such a thing as friendly mockery?
cradek: You so funny!
(that sound pretty friendly to me =)
i took it as friendly
See, there ya go =)
skunkworks_: top speed is 2400? ouch.
I posted a link to a project that allowed you to document a electronic project, draw a schematic, and even produced a PCB. would anyone happen to remember the link?
[14:54:02] <SWPadnos> http://www.altium.com
It was open source
oh. definitely not Altium then
More along the lines of arduino type project.
I think Olimex has some PCB design software now
and they have (had?) a deal where they'll make a couple of boards for free, if you let them sell the design on their site
Sell your Soul is more like it
No IP issues there =)
or you can just pay them (or someone else) to make boards for you
* Jymmm hands cradek $20, I'll take 20 boards pleasse.
cradek: yes - it use to have a hydraulic spindle motor... Now it has a vfd. Although I don;t know how much more we want to run it. (if we need higher we would probably put a motor in the spindle)
*how much faster we want to run it.
[15:02:43] <Jymmm> http://fritzing.org/
Here it is
cradek: it runs a 12inch shell mill just fine... :)
hi everyone. Question about ballscrews. What is a typical price per foot ?
elmo40: google would better answer that
since ballscrews come in diameters from 3/8" or so to about 6", and there are a dozen other variables as well, there's no "typical" price per foot
Jymmm, OK, affiliated with that Processing language as well - interesting
SWPadnos: It's alpha, but lots of potential
wow, 6" ballscrews. would take descent power to just turn those ones :P
elmo40: Nah, just a couple of AA batteries
anyone knows how I could get some from of timestamp in HALSCOPE
I need the timestamps of the signals that I sample
i see 20mm screws for $2.40 / in. Just sounds expensive. Plus the $130 flange nut...
malem-cnc: all you could do is generate another signal representing the time, and plot it
I'm not very familiar with hal, but I will look for a timestamp kind signal in HAL doc
I don't believe there is any timestamp component at the moment
I have scope.sample.time
I'm not sure what it is, it looks like the delta between sample
if it is, I could deduce the timestamp from it
but system time would be better
yeah. I think time of day is a little harder than "time since starting", but that may not be true these days
FYI... Google Bokmarks 'labels' are case sensative! ARGH
[15:26:55] <Jymmm> http://www.web-tronics.com/liti124.html
hmm comp's script language is quite nice
I'd only have to find what rtapi linbrary would give system time
Holy BaJeeBus.... NEMA23, 380oz-in, sgl shaft$ 299.00
Jymmm: that a good price?
hi to all
are you here?
Guest188: just ask
hi, i have a problem with emc2
i try to read the axis position with python program
I know, how it's going?
emcstatusbuffer invalid error=3
when i start the program
where is the mistake?
are-you attached to an emc process?
i start it from an emc session
Guest188: paste it program at www.pastebin.ca
loaded the .ini file? in the right working directory?
i have sent the program to pastebin
then take the address of it and link it here
[15:54:37] <Guest188> http://ewww.pastebin.ca/1819405
it stops at line 21?
line 88 and line 65
there is no lines 88 and 65 in there :P
sorry line 66 and 42
it looks like you are not correctly attached
is there documentation for the emc python api somewhere
just to make sure : you did 1) emc& 2) started the gui of your choice 3) ran your script
i'd like to have a quick look but can't find such a thing
no all source code
all in emcmodule.cc iirc
yes i start emc2 then open the file
I think you have to run you script in your .ini directory
it should work
but i'd start the script into emc2
is it possible?
you would have to patch the makefile
much simpler to attach to the process using emc&
in a console
>emc& (then choose the right .ini file)
> python yourscript.py yourargs
you put no args!
look at your line 14
basically, you did not load any .ini file
from your script
umm... if it doesn't work with no args, it'd be a good idea to return with an error message, maybe? 8)
(in the case no arguments are given)
ok thank's, tomorrow i try all now i stop work and .......go home
iniFile = emc.ini(iniFilePath)
nmlFile = self.iniFile.find("EMC", "NML_FILE")
if (nmlFile is not None):
emc.nmlfile = nmlFile
that a sample from my code
ok thank you, tomorrow i try it, bye
Jymmm: i can find NEMA32 600oz/in for $110.
anyone know if p4 and p4 ht processors are generally interchangable on motherboards?
they should be, if they use the same socket type
i discovered some good parts that i thought were bad, so i'm pretty happy
got a couple extra much-needed pc's out of them
SWPadnos : http://pastebin.ca/1819459
HAL: ERROR: function 'timestamp' not found
univpwm_load.hal:24: addf failed
maybe add a reset input
it was installed using
sudo comp --install timestamp.comp
you don't need "timestamp_" for the function, only _
... etc ...
yeah that is what I first tried
well it was basically a cut & paste from http://linuxcnc.org/docs/2.4/html/hal_comp.html
the function name will be timestamp.0.timestamp
if you had more than one, or you gave them names, they'd be named differently
the problem is that I cant even attach it to my thread
leave the signal attaching lines out of your hal files, and use halcmd to see what the actual name is
ok, as that example shows, you can leave out the function/FUNCTION declarations
I removed the addf
it shows up in hal config
but it has no user
ok, so use that name in the addf
thanks a lot
I see it in halscope, love this
JT-Work: hows it going
ran the spindle this morning
Danimal_office, did you see the link to my frame?
[16:56:42] <Eric_K1> http://www.flickr.com/photos/mcn7/4398766376/sizes/l/
the spindle would not stop turning LOL I could raise and lower the speed
Jymmm: did you see where the Fritzing stuff prints out an overlay for a proto-board, and the user just punches the components thru the paper into the grid? http://fritzing.org/news/paper-templates-for-your-breadboard-prototypes/
some configuration isssue
JT-Work, how you controlling the spindle?
do you know why it wouldn't stop?
not at the moment
I assume the drive was getting a little bit of a velocity signal due to my faulty programming
Eric_K1, whos the cute girl on pic?
How can I communicate with components pins from python code
but it only works for creating new
JT-Hardinge: what are you using to enable the spindle drive?
Eric_K1: yes, very nice!
JT-Hardinge: you're using the 7i33 to enable it? what changes direction? -10v?
I thought most spindles had voltage/direction control
if you're using the 7i37 turn it on and off, try reversing the polarity of your output wires going to the 7i37
Eric_K1: you set up a booth at the nahbs?
yeah, in the new builder section
did you see a couple Boo bikes there? they're bamboo frames
good way to wast $1k anyway :)
I saw a batch of bamboo, but I don't know if I saw those
I was in a fog for most of the show
did you get some good response?
hard to say, a couple of people seemed interested
I need a better on-line presence
[17:16:06] <Danimal_office> http://velonews.competitor.com/2010/02/bikes-tech/boo-bikes-headed-for-nahbs_106217
this guy ran my chain tensioners
on both those bikes
[17:16:49] <Danimal_office> http://velonews.competitor.com/2010/02/photos/boo-bikes-headed-for-nahbs-a-photo-gallery_106219?pid=2600
that's nice, I think I saw them
my first magazine appearance!
those are cool
the ever-present bottle opener
you can start hanging around the web forums... i pretty much owe my business to MTBR.com
yeah, that's my plan
and give my friends bikes at cost, that sort of thing
my crappy blog usually gets 100-200 hits a day, and mostly from MTBR.com
that's a good point
Eric_K1, i'll be your friend!
I have an account there, haven't been posting
make a single speed and post it in the singlespeed section of MTBR
does it have to be a mtb?
tons of roadie SS's on there
replying to myself (for future references) http://firstname.lastname@example.org/msg16029.html
(read hal pin from python code)
what's your user account at mtbr?
wouldn't have guessed that :) I'm unterhausen
that is a really nice frame
i love the color
Danimal_office: when the hardinge gets up to speed I might have some ideas for you
it's up and running
i do have a little work to do to it, but she runs
Danimal_office: made you a buddy on mtbr and added a link to your blog on my blog
cool, what's your blog? i'll do the same
from image-to-gcode, what are "I" and "K" coordinates it keeps putting out? :)
? g2 and g3? those are arcs. http://linuxcnc.org/docs/html/gcode_main.html#sub:G2,-G3:-Arc
it's the radius
yea what skunkworks said
how's it going skunkworks?
Good. Keeping out of trouble for the most part.
how's the maching coming
wondering what people use for controlling RPM on their spindles.
Danimal_office: working on it consistanly.. figuring out I/O needs. we know we have to control 21 hydraulic solinoids atleast ;)
so how many 5i20's do you need?
or are you going with a 5i23?
probably more than 1. We have 2 5120's just in case. I think though I will need 1 more servo interface card.
they have a 6 channel that works with emc now i believe, right?
do you need to control every one of them, or could a small plc device get a start cmd and hand over a fin signal? (just in case a cheap plc was < a few cards and slots )
that will give me what I need to run the 4th axis and spindle control (10v). plus a few encoder inputs.
tom3p: I think it would be easyer to control everything within emc.. (through hal and ladder logic)
tool change sequences and table rotation and...
you still need all the I/O points, whether they're in a PLC or attached to a Mesa card
right so its just money, simplification is always good
luckly my SSR's work with my 7i42ta so i get quite a bit of outputs with that
5v is enough for them
we are going to be using a lot of opto22's
like jmk's? (3-32vdc)
very similar - just opto22 brand. (his seem to be a bit better quality.. but they are the larger size. (we have a bunch of g4 stlyle we are using)
think that is it..
gordos opto-22 crydon kyoto all similar wait for a price was ;)
seems to be coming togather so far. No major suprisses yet...
the 230 found in homes (eg: for elec dryer) theres 110 on each side to gnd, how far out of phase are they?
it is just a center tapped transformer (ground is the centertap)
you got hot/hot to get the 240
ah good, else id need more cap to smooth (gimme your left your right yer left...)
uh i think i got that wrong, at 180 deg, they're simultaneous (overlapped but delayed current) ?
if each were half wave rectified and summed, there'd be way huge ripple
lemmy find an image, sec
any two wires give you one phase
it's just like rectifying 120V, only twice the voltage
it is appropriate to call this power distribution system a 3-wire, single-phase, mid-point neutral system.
so hot+ to hot- gives 240, either hot to neutral gives 120
which means they much be 180-deg out
(also ive hooked the scope up to the wall ;-)
[19:12:45] <jackc> http://www.beananimal.com/media/3246/split-phase-sine-wave.gif
green and red are the hots
x axis is neutral
[19:16:06] <skunkworks_> http://groups.yahoo.com/group/DIY-CNC/message/29824
skunkworks_: it's behind a login - if it's imporant, pastebin it for us?
I baulked at teh login too :)
jackc you're right, they're 180 out, when 1 is high other is low, i do need the correct amt of caps times 2 http://www.allaboutcircuits.com/vol_2/chpt_10/1.html
sorry. It really isn't that important.
some sort of argument about if linux runs under rtai or rtai runs under the linux kernel.. and then that rtai sucks because it is it 'may not work for every proccessor' and it is trying to do everything for everyone. (I am paraphrasing) ;)
it seemed more of a rant against open source projects and documentation in general
"Computers are hard!"
[19:22:33] <SWPadnos> http://pastebin.ca/1819716
I hear that to be convincing, lies should have one specific detail, not more. I wonder if rants are the same.
hmmm. I guess Tom's Hardware must be really convincing, since they have one detail per page
ok, I read it, thanks for that enrichment...
sorry. nothing personal
I meant the pastebin
me too :)
* cradek shrugs
(ie, I didn't mean to torture you or anything)
GUYS MY XOLOTEX Z CHAN DIED
* renesis hands out the tiny violins
(its okay because i have an A chan i never use)
(but wtf same config for like 4 years just like decides to die)
hey this violin broke I wanna refund
skree - eeeee -eeeeeeee - ooooooooo - ooooooooook
not bad, considering I haven't played in 35 years
SWPadnos: how old are you>
me 15.4 > you :)
me 35.973, yay
hmm I think I should use hex
robotito: of what the xylotex?
no drama, i didnt pull the sink tho
i just disable that channel and switch wires/config for using A
i think the bridge is fucked tho
I think it is a right of passage to destroy a xylotex board..
the sink was still warm but the stepper wasnt at all
and i think it was loading up the psu a ton
because when i switched channels, the X and Y sounded alot crisper too
ough, 3d milling goes slow :P
skunkworks_: yeah i kinda felt lucky up until now
i guess im just normal person after all
yeah looks totoally fine except for a bit of dust
MrSunshine: it all has to do with accelleration... (do you use g64px.xxx - x.xxx is how close you want emc to follow the programmed path)
when i get home and back to the machine ill take the heatsink off
and then right after that
the wires on the push-down terminal MTA-100 on my Y stepper broke
that g64p### works wonders
i thought my xylotex Y died but i redid the connector and its fine (yay)
skunkworks_, i just generate it with image-to-gcode and mill it =)
and i guess i could add up some acceleration on Z
its quite slow :/
and i need a FASTER cnc computer!
nah, P4 but bad graphics card etc
i think its a p4 :P
celeron55_, im not at that computer atm
and no network on it =)
i want my quadcore running now
that should be enough to run this stuff =)
tom3p: No I didn't, that's pretty cool.
a mesa board should be cheaper than a faster pc
Danimal_office: did the new computer fix your issues?
I seem to have been siginifcantly too tardy grabbing the pyvcp spinbox patch that allows direct entry from the pastebin where it was put. Does anyone know what I am on about, and where I might find it again>
skunkworks: seems to have
that computer was a bit old and it had some weird proprietary stuff on it
it was an old server
so i threw in an asus mobo with an amd athlon xp2600 processor
and a gig of ram
so far so good
still not the newest, but who cares as long as it works
i just put together a couple other computers from spare parts and everything is working out pretty good, so i'm happy. i didnt have to spend a dime
Tecan is now known as COMMA
COMMA is now known as COMMAD
Danimal_office, p4, ram?
[21:15:35] <COMMAD> http://img651.imageshack.us/img651/9467/screenshot1uv.png
<< the tor network hehe
COMMAD is now known as Tecan
robotito: it's an amd, not a p4
You mean "It's an amd, not intel"
is there a wiki page about encoders, types, setting them up etc?
i have noob Qs about quadrature etc
[21:20:32] <Jymmm> http://www.google.com/search?q=encoder+site%3Alinuxcnc.org&client=firefox-a
haha well played
not official, will get stomped on by any upgrade, cvs or git-ification
Sorry for the late reply, I was in the workshop cutting metal. I think I might have damaged my hearing with that metal-geared mini-mill churning away for 2 hours
* frallzor is pretty sure irc is more based on the ability to see :P
digitized voice 'frallzor is pretty sure irc is more based on the ability to see. grin'
wow compiling a kernel takes forever
It is a Rule of Linux that no matter what time you start compiling a kernel, it will finish at 4am.
half the reason to gut them
clytle374: would you try the v2.4_branch without any patches and make sure it works on your Klingy Koala?
i just committed the patch we worked on last night
seb_kuzminsky, yep, starting on that now.
Still doing a drive copy to a newer-ish drive. should be done soon.
overheated building a kernel with make -j2 lol
Tecan, I prefer -j8
yo its all about the cores kids
n+1 is how we roll
2xn unless is breaks
distcc likes extra
seb_kuzminsky, don't worry, I didn't thread your test.
seb_kuzminsky, odd, I got the original error and did a fresh pull for the test.
And the patch cleanly applies
patch -p1 --dry-run < 1818570 gives no errors
wait, is the emc2-dev the same as the v2.4_branch?
clytle374: no... emc2-dev is the repo, it contains many branches. "git branch -a" lists them