he went to sleep?
he really does sleep?
alex_joni... you still around? :)
03jmkasunich 07HEAD * 10emc2/docs/src/hal/vcp.lyx: Restored VCP documentation, added a note that it is deprecated. At some point, we will probably remove VCP completely, but until then its documentation should remain.
jmkasunich_ is now known as jmkasunich
is timeguy around?
[03:02:06] <A-L-P-H-A> http://bash.org/?23601
I was googling "autocad g-code exporter," and I tripped over your website and noticed that nice little plugin you wrote. I do have a few questions, though
for instance, would your plugin convert in three dimensions? I noticed that the name was "2.5," so I'm not entirely sure...
it's just 2.5D
also, since it is written in LISP, I'm worried that speed could be an issue - have you any benchmarks?
that means each cut is at a particular depth, but there can be different depths
so why don't you call it "3D"?
because that's what 2.5D means
it's a common term
3D often means it could contour, or work equally well in all planes, which is not true
what's `diff 2.5 3` ?
you realize it's a program you run once, inside autocad, to generate gcode output, right?
yes, of course, but what if I have a complex drawing?
then wait a bit
say, 5000 lines+arcs+splines?
I'm baffled by why you would care
wait an hour?
why not try it for yourself? it's free software
improve it if it's too slow
maddash, what spindle horsepower do you expect to have, and what kind of material do you expect to cut with your machine?
gene__: I laugh every time I see your ident information, which starts gene@pool
the one from this afternoon?
yes, the lathe or mill or whatever that will be moving a cutting tool around a workpiece
not the PC, the mechanical contraption
cradek, aha, nice catch
SWPadnos, I'll be grazing the surface of soft (unhardened) stainless steel with a cnc miller.
ok. what spindle horsepower do you expect to have?
I'm not sure yet, because I've haven't decided the actual depth of the graze...
the depth isn't the only important thing - a given horsepower has a specific amount of metal that can be removed
but, if you want a range, I'd say about 7-12
ok - one sec
hahaha... bash.org kills me.
haven't had such a good laugh in a while.
well, steel at 100 BHN (whatever that hardness means) takes 1.25 HP per cubic inch/minute
at 400 BHN it's 2 HP / cubic inch/min
[03:16:17] <A-L-P-H-A> http://bash.org/?50891
so at 12 HP, you can remove between 6 and ~9.5 cubic inches of steel per minute
ok. it looks like 18-8 (304) stainless is 250 on the Brinell scale
that takes about 1.85 HP/cubic inch
that gives roughly 6.5 cu in/minute. unless you're planning on using a very small cutter, I think the spindle will likely be your limiting factor, not EMC2
I am looking for a suggestion of how to repurpose M7 and M8. right now they both turn off with an M9 command. I wish I could control them separately. any suggestions?
since that's explicitly in the rs274ngc spec, it's just hardcoded in EMC and not configurable
is there another way to accomplish your bigger goals?
I use a universal stepper controller board, it has 8 relays that can be operated on the board.
most of them are under the control of M3, M4, M7, M8 and M9
that's only true in emc2 if you set it up that way
I would like to be able to control them independently.
are you using emc1 or emc2?
emc2, of course. :-)
yeah that's simply a hal configuration issue.
ok, then change the ppmc_load.hal file to connect the other digital outputs to other signals
*looks at code*
I think there are 8 or 10 from the motion controller that are controlled by M60-ish codes (??)
does the configs/univstep use the ppm_load.hal?
univstep_load maybe ...
I guess I could look
that'd be much appreciated.
perhaps someone could shout out what ppmc is?
M64 Px = set output X; M65 Px = clear output X
are you at your EMC2 machine?
then look in the directory from which you load the emc configuration
in CVS, I see several files, two of which are univstep_io.hal and univstep_load.hal
cradek: what is M64?
M64 and M65 let you control auxiliary outputs
yeah and then I could tie them to the appropriate hardware of the UCS using hal?
you hook the USC digital outputs to the motion controller aux outs (not sure what they're called), and then the SSRs are under the control of M64 and M65 instead of M3, M4, M7, M8, M9
unfortunately I think there are only four of them currently. You can increase that number and recompile though.
oh four would help a lot
* owhite counts aloud.
it's easy to change - just one place
you could easily have 8
I need oxygen on, laser on, ventilation on.
the last 4 sections of univstep_io.hal have those connections
that sounds like the ticket.
laser on should be able to connect to spindle on, I'd think
it is :-)
spindle forward is laser on.
I was doing fine with lube being O2 on too, then I needed ventilation.
ventilation should probably be either mist or flood
since that is cooling
do you always need O2 flow?
02 goes on when the laser does.
but I have them split out.
ok - can they be controlled with the same bit, or do you need them separately controlled?
its true, they could be tied to the same bit. I was just being tricky. thought maybe there would be times when I might supply a different gas, or something.
but since I can do the M64 number, I think I'll leave it.
if you need to use two separate SSRs, you can still drive them from the same Spindle-fwd bit
just linksp spindlefwd <first_dout>
then linksp spindlefwd <other_dout>
oh no kidding.
damn you guys are smart.
you can write to all the outputs with one bit if you want
and you can invert them individually, so you can deal with positive or negative logic, even though they're run from the same bit
well I want ventilation to come on at the beginning of the run, and stay on even if the laser is off.
that would be a mist/flood thing
keeps it similar to conventional milling programs
you can connect O2 to the spindle-brake output, and invert it
hey get this, while I have your attention...
...you know, have to switch on a lot of crap manually.
that should turn the O2 on a little before the laser, and off a little after
...the motor power supplies, the laser controller, which has the UCS, ventilation, and lighting.
sometimes, i'd really like to have control of that even if its in e-stop.
E-stop / stop / power design is a really big question :)
like, one of the things I'd like to flip on is the laser controller itself, which has the UCS board, when its not on, the machine is in e-stop.
do you mean USC?
that board should be powered ewhenever EMC2 is running, I'd think
so like its a catch-22.
what'd be cool is to fire up EMC2, and use it to start the laser controller, which would power up the board.
you can do all sorts of sequencing with ladder logic
do it all in software. I turn the thing on and off a lot. it'd be nice if was mouse driven.
well, you need to think about power sequencing a lot, and what states you want to have which pieces of the machine powered in
but can I do _any_ ladder logic, if its in e-stop? It will be in e-stop until the USC board is on, and it is recieving good signals from things.
* owhite fears this is making no sense, per usual.
you probably need 3 or 4 separate power zones. The PC is in one, the USC is in another, then the servos and laser (possibly separate), ventilation and lighting could probably be enabled in the same zone as the USC
yah, that's it.
since the USC will control those things, it needs to be on before you can do anything with them
since the PC controls the USC, it must be powered before you can do anything with it ...
and the way I wired the USC, it wont come out of e-stop until the laser power supply is on.
E-Stop doesn't need to kill ventilation and lighting, so those can be "before" the motors and laser and support equipment, like the O2 controls)
the laser spits out logic that its getting good pwm signals from the controller, and the USC reads that.
I assume that the laser power supply isn't your means of controlling the laser output? (ie, there's an enable input on the laser)
how is the laser controller controlled?
(you mentioned PWM, but the univstep is step+dir, not PWM)
so. I have a custum board that sends it PWM, and the USC controls the enable.
ah, so you only have enable/disable, not power control
the laser also sends out states, indicating its not overheating, those are read by custom electronics, and if they are in good condition, they allow e-stop to be off.
power control can be toggle by emc2, to two different states.
I figured that would work so I could either cut, or engrave.
I thought some day I could have finer control with M100 codes.
how about with spindle speed?
yeah, could do.
course, spindle is tied to laser on! :-)
not spindle speed
but that could be changed. i'd still have to make my own electronics convert the spindle speed game to pwm.
spindle speed is a float
oh oh oh.
but its output usually goes where?
to a VFD or a dialog box asking the user to set the speed ;)
*awaits the inevitable "hal" answer*
oh sorry - "wherever you want with HAL"
what are my options for hardware that could receive the spindle speed?
the way I set my pwm, is that its controlled by 5 bits.
well, with the right hardware, you could just output PWM from EMC
that would be nice.
it may be doable in software, depending
I've been think ing of re-doing all my electronics. but havent had a good enough reason.
you'd need another parallel port or other I/O device though
500 yards of wire wrap cant be all wrong, or something.
yes it can
* owhite is proud of his wirewrap, dont knock it.
I think 500 yards of wire wrap is ALWAYS wrong :-)
I'm only a lowly biologist, learned that stuff myself.
I have a friend who built an entire Z-80 based computer with wire wrap
dont be cruel guys.
actually, I think he did two or three that way
that's what the smiley is for
TRS-80 compatible, too
okay, so I think the answer is to look into M64.
for the aux digital controls, I think so
control m'little SSRs, turn on my ventilation.
you probably only have one though - lighting
I put 1800 cfm system into my basement, made a plexiglas enclosure, ran hosing, buste holes through the house foundation -- I'm ready to fire this thing up.
O2 is analogous to inverted spindle-brake, ventilation is analogous to mist or flood coolant, and laser enable is analogous to spindlefwd
but I swear the way the USC is configured if you hit spindlefwd, then spindle brake will go low.
I think that's rs274ngc?
I think that behavior is hardcoded too.
*takes a mad shot at saying something intelligent, awaits retrobution*
yes, but there should be a delay
I WAS RIGHT!
but there's no reason to hook up brake to something if you don't want it
you invert the spindle-brake signal to make it work for O2
sounds confusing, but I'll look into it.
it becomes a "spindle is going to be on/was just on" output
I don't think there's a delay, but I could be wrong
something like setp ppmc.0.dout.02.out-invert true
but what the problem with M64?
oh, g'night cradek.
see you cradek
it's not normally used
if you use the controls that are already there for a lot of this stuff, then your code will be easier to interchange with milling machines
yeah but the same arguement could be made that I am shoe-horning my laser to work like a milling machine, when it aint.
I never use other people's g-code, its all dxf-->perl-->g-code.
SWPadnos: well I pledge to have a careful look at both options.
looking at the code, spindle brake is set/cleared at the same time as spindlefwd, so there's no need to use it
that's the behavior I've been seeing.
you can use classic ladder to delay that, but it isn't there by default
the other thing to note is that the USC guy has recently implemented another 8 bits on his board to work.
yep, the P2 header, I think
and I assume all that will be driven by M64 or M101.
depends on how you wire it up in HAL ;)
right. I made a relay box to plug into that. that's why I was asking those questions about lighting and stuff.
that header was actually intended as a spindle speed control, and Jon is making a DAC for it
but you can of course use it for what you want, if you do it right ;)
hey that'd be pretty good for the PWM of the laser power.
well, I'd still (for the last time) take your recommendation on the better way to go -- would that be a M64 for a M101 thing?
I think it'll be DAC only, using all the bits there (plus a couple of the other outputs)
M64, if it's in the version you have
I'm not sure if that's a v2_1_branch or HEAD thing, or if it's in the 2.0.5 package
ah, I may have to do another check out? hm, that'll be a bunch of patches.
could you shout out the cvs command to check?
you're running HEAD?
* owhite looks puzzled.
(ie, no specified version when you check out)
I think so, but I'm not that bright.
heh, when it comes to CVS, I'm not too bright myself
yeah but head is still a snap shot of what was checked in at the time right?
from the directory where you checked out (the one above src/), run cvs update -dP
that wont over write anything, will it?
it will overwrite anything that's changed in your CVS directory
no it won't
err - that was ambiguous
whatya trying to do, kill me? I've done a lot of code changes.
I meant anything that's changed in CVS will be downloaded, unless you changed it locally
when I grep -R M64 I dont get any hits to anything.
in which case it may be merged or ignored
when running emc, do "bin/halcmd show pin motion" from the same directory where you did your checkout
actuall, I found this with grep: src/emc/rs274ngc/interp_convert.cc: M64 sets a DIO imediately
if the output includes "motion.digital-out-00" thru "-03", then you have the M64 code, if not, you don't
its in there :-)
then M64 should work
do a M64 from MDI, and then repeat the hal show command, see if any of those pins changes state
M64 P1 on my system sets motion.digital-out-01 true
but its late and I cant really run the system, everyone is sleeping.
oh, should I have run M64 P1?
I just ran M64
I didn't know the syntax myself, when I tried M64 I got an error message, so I muddled around
M64 P1 turns on output 1
dunno how to turn it off yet ;-)
well I could run sim....hang on.
M64Px turns outputs on, M65Px turns outputs off
(cradek mentioned that eariler, or I wouldn't know :) )
yeah there we go, worked in sim mode.
you know, whoever came up with the hal was pretty smart.
that's darn nice.
that'd be jmkasunich
jmkasunich: good on you man.
and happy new year too.
glad you like it
how to save a log of the irc session in xchat?
hmmm. in chatzilla, you have to turn on logging, at which point it starts the log ...
you can probably copy/paste the text into an editor, or there may be a "save log" menu option /command somewhere
there is a save in an odd part of the menu.
alright guys. take it easy. thanks again.
see you. I was thnking the same thing
20cm of snow in the last 24 hours, and I thought we didn't get snow anymore
we've got, no snow at all
+8 degrees C
hell must be nearing the surface of the earth ;)
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2007-01-12.txt
LH: the real question with hell though is if it is endothermic or exothermic -> http://www.pinetree.net/humor/thermodynamics.html
that means that hell indeed is nearing the surface
Dallur: that's a nice one
alex_joni: I laughed my pants off when I first saw it :D
I can understand why :)
good morning folks
eugenics... have you been chemically castrated? is that why you like eugenics? Preservation of human gene pool?
A-L-P-H-A: go to sleep
I did... from about 11pm till 5am
it's now 5:51am
good morning then >(
so, did you follow my advice and not go to sleep yesterday?
nah... I just took a 1mL of melatonin...
alex_joni, made anything cool lately?
lerneaen_hydra, where's photos of your bike? And I don't get your photos of how you machined the waterblocks on your side.
A-L-P-H-A, I think you might been confused, check it out http://en.wikipedia.org/wiki/Eugenics
A-L-P-H-A: im waiting for parts
3 weeks *nnngh*
A-L-P-H-A, and apperently me too, anyway eugenics also means that the human beeing is searching to improve its way, term used in psycology
A-L-P-H-A: been working on a magazine commercial
ok, I should change my nick again...
eugenics is also some sientific rase enhansing filosify
I did not know
^eugenics is now known as kamekaze
eugenics can also be a study of "improving human genetic qualities"
[11:26:34] <alex_joni> http://18.104.22.168:8026/cms/acct/vietweekly/english/english_vw3n09/lastKamekaze.html
but its is anyway a sensetive topic, and I dont want to affend anyone
I imagine kamikaze beeing a sensitive topic to US guys
how about politically_correct_nick_01 ?
well.. it's in english
only that is sensisitive to a dozen middle-eastern countries
I like that one :)
kamekaze is now known as politically_corr
politically_corr is now known as nonafensivenick
nonafensivenick is now known as nonofffensivenic
nonofffensivenic is now known as nonoffensivenick
well its a hard job choosing a nick
how about name?
or is that too obvious?
btw my name is martin, but the nick is taken
first letter of last name?
ok, you convinced me
nonoffensivenick is now known as martin_lundstrom
on phone with slide reseller
I didn't realize I caused so much problems.
someone type my nick please.
nick highlighting, that's all.
A-L-P-H-A: what problems?
anonimasu: any luck with the slides?
alex_joni: 3 weeks of shipping :/
just hadn't set it up that's all.
alex_joni: waiting for the guy to call me
anonimasu? what'd you order?
alex_joni, did you get your mesa?
A-L-P-H-A: some slides off a company here..
SWPadnos owes me a mesa, and a few other parts. Once the check clears....
anonimasu, why do you need slides?
A-L-P-H-A: plasma/oxyfuel table..
X and Y rails..
what are you gonna make in the mean time.
go make stainless steel 316 ring for yourself. :)
A-L-P-H-A: I'm going to draw up all the parts we are going to cut..
functional cutter? or art cutter?
production stuff for work
I got the square beam for the base today
who was the guy that worked as a volunteer for a racing team? was it you anonimasu?
that was me :9
still talking with them?
A-L-P-H-A: I've got neither time or will to be involved anymore :)
I get the real delivery date on monday
and we changed to a roller bearing carrige..
instead of a linear bearing one..
he thought that it would be the best thing to cope with the dust
how bit are the rails/what are they charging ?
anonimasu: not to bad
that's prefab ones with motor and all.. :)
wow, that's good then :D
belt drive also..
alex_joni: plasma and ballscrews is a great combo
* alex_joni doesn't trust any nonmetalic parts
but good low-backlash gears are really really expensive so ....
anonimasu? WITH motors??
oh, drive belt... not screws.
but you need the speed I suppose.
the more important issue is dust..
ballscrews and plasma dust..
you're not making sense to me.
A-L-P-H-A: 2.5 meter ballscrew..
A-L-P-H-A: how do you get it totally dust proof?
A-L-P-H-A: the dust is from the vaporized metal, very high carbon steel particles everywhere :P
Dallur: how are you running your machine?
rack & pinion?
dual rack & pinion system with a belt reduction drive
I have two pinions for each drive point, then I put tension between them to eliminate all backlash
A-L-P-H-A: ballscrews get filled with dust and well, your machine locks up..
same with linear rails..
connection problems ?
seens to be a lot of that going around
no, just his nick is getting longer
no word - know by the end of the day
anonimasu, okay... use teflon style linear slides... not much of a dust problem.
what about HDPE and anodised alu?
programming an upload section for gallery for a client.
for fun or code-monkeying?
ah, that's nice too
good morning folks
03jepler 07v2_1_branch * 10emc2/docs/src/gcode/main.lyx: merge rev 1.17
awallin: where's the sharks ?
alex_joni: hi. I understand that was a slashdot thing from a while back... don't have time to read /. myself...
Dallur mentioned on the list that he had done some work in pythno for DXF-> G-code. interesting
awallin: there was a TV movie (sharks with lasers)
Hey, Anders, yup
alex_joni: oh, ok...
awallin: When I get home tonight I will send you a copy of what I have been working on :D
Dallur: great, I've been thinking about something like this myself.
If you've solved the DXF reading part then that will save me a lot of time
awallin: I got the dxf reading mostly done but only for dxf v.1.2
"You know, I have one simple request. And that is to have sharks with frickin' laser beams attached to their heads! Now evidently my cycloptic colleague informs me that that cannot be done."
awallin: So it does not handle splines, 3dfaces or solids but everything else works fine
alex_joni: I can supply the lasers, if you get the sharks
awallin: Austin Powers
[14:02:11] <alex_joni> http://www.imdb.com/title/tt0118655/quotes
Dallur: lines and points would be enough for a lot of interesting stuff
awallin: I'll get the sharks if you mount the laser :D
Dallur: and maybe circles?
awallin: circles are working fine
I thought I saw something about sharks!
awallin: so are polylines and lines and points
awallin: and curves
Dallur: good. Your G-code is an exact copy of the imported geometry? no offsetting or anything?
I mean the G-code output from the program
awallin: yup, that was one of my requirements
Dallur: for a closed curve, how does it decide where to start
awallin: it's stupid at the moment, starts all curves at the western most part
that's an interesting idea :D
to put it in a better way, the lowest X value
I was thinking about OpenGL for the 3D view...
awallin: vpython uses opengl :D
Dallur: ok, I didn't know that... does vpython include stuff for rotating the view and selecting geometry etc
awallin: Yes, currently I render a 3d viewpoint which you can rotate around, zoom in and work with
awallin: might be nice to split to multiple viewpoints but initially I just added the view to assist in debugging the code
sounds great, can't wait to see the code! how long have you worked on this?
awallin: I worked on it for a week or two early in november/december but then I got swamped with work and have not had time to look at it, hoping to get some focus back on it soon
awallin: I was really surprised with how fast you can do things like this in python
're not jarl on the mailing list by any chance?
gah, ' and enter are too close
Jarl is my first name
Rugludallur=Dallur=Jarl to be more precise, Rugludallur@home, Dallur@work and Jarl@RealLife
oh, I see
Dallur: what kind of work do you do?
how well-developed is the stl-to-gcode app you're making?
*curiuous about iceland*
is it based on jepler's image-to-gcode?
it's dxf to g-code
isn't that 2.5d?
anonimasu: Software/Hardware Consultant, mostly enterprise systems for large companies (airlines, telcos) but sometimes smaller stuff to
btw the converter is called dxf2nc, and at the moment it is good enough so you can generate any 2d toolpaths, it's missing things like lead-ins and spindle control so most likely you will have to edit the g-code somewhat, but all the toolpaths will be done for you
Dallur: sounds good, I was thinking about trying out some CAM algorithms... your work with dxf2nc will be an ideal starting point
awallin: How about I just create the sourceforge project real quick, and once I get home I can upload and add you to the project ?
Dallur: that would work too, but I'm not sure we have exactly the same goal in mind
I'd be interested in toolpaths for a mill, for your plasma you are probably satisfied with directly following the dxf line
[14:30:34] <anonimasu> http://sourceforge.net/projects/verot/
awallin: the plasma is not quite that simplistic but when I first sat down I figured there were two sets of separate requirements, 2d and 3d, they should still share the same parser/writer for the dxf and gcode imho
to bad that project is dead though
eh, can you find _any_ files in the cvs for verot ?
just a few..
looks like c++
I think python would be ideal for rapid prototyping something like this.
cpu-intensive stuff can be ported to C if required later
you can implement the modules/toolpahting in c..
and just use python anwyay..
anonimasu: start with even the toolpath generation in python. it can't be that slow...
[14:35:12] <anonimasu> http://repositories.cdlib.org/cgi/viewcontent.cgi?article=2493&context=postprints
The most intensive stuff is probably the toolpath optimization, finding the shortest closed path between multiple vectors is a classic NP problem
Dallur: are you using a gui toolkit (Tkinter, wx ?) or just what vpython gives you ?
* anonimasu would love to help out
though my python isnt that advanced
awallin: at the moment just what vpython gives me
anonimasu: that paper looks more like RT-control of a machine, not toolpath generation
anonimasu: neither is mine :P
awallin: yeah I just found it
Dallur: in the long run a raw OpenGL viewport could be more versatile, but for rapidly getting things show I guess vpython is ideal
awallin: very true on both points :D
[14:38:58] <anonimasu> http://mathforum.org/library/view/66099.html
start by reading all references in this paper: http://www.cgl.uwaterloo.ca/~smann/Papers/survey2.pdf
dunno if anything is useable..
anonimasu: if you read freesteel, boundary offset is not what they recommend, if I understand correctly
jepler: you haven't done anything new in image-to-gcode by any chance?
awallin: well, for constant overlap contouring.. that's a good thing to do..
capslock or pissed? ;)
sticky shift lock
er, shift key
awallin: though paralell pocketing ends up faster in most apps..
anyway, I don't have any plans for image-to-gcode these days
awallin: then just do a finish pass around the boundry :)
jepler: why not? IIRC work was progressing nicely, then just stopped?
anonimasu: parallell pocketing would probably be a simple case to start with.
awallin: that's a very standard function..
awallin: all cam programs use it as finishing.. when pocketing and stuff..
but I bet it's not dead-simple to do right
awallin: I think they use it to make, a offset..
then they face the area paralell -----
then they sweep around the outer boundry/multiple passes for the finish
tell me if im off track..
lerneaen_hydra: as far as I can tell, it's "done" for the case of items that can be machined in a single pass
awallin: the other page I looked at talked about fp errors..
lerneaen_hydra: and since it's GPL and in the emc2 CVS tree, someone else can add new features such as multiple passes if they like
jepler: oh, were lead-ins implemented?
anonimasu: I don't quite follow, drawings are much better than text for describing toolpaths
lerneaen_hydra: I don't remember 100% what was implemented. I think arc-in was added, but I'm not sure.
oh, I see
awallin: floating point..
if so then that would be pretty much everything
do you recall if you could tell it to do two passes very crudely (essentially two programs after each other, 90 degrees ofset to each other)
anonimasu: right. that may become an issue, but python should contain som high precision number format?
jepler: well then I don't have to bug you about that any more :D
pymath(whatever pyplot) depends on
what was it you coded it in? python?
lerneaen_hydra: yes, python
of course python
neato. does it "come with" a standard cvs checkout?
anonimasu: here's something on blend/morph pocketing: http://www.ima.umn.edu/industrial/2000-2001/bieterman/bieterman.pdf
*saves this stuff*
anonimasu: when the open source CAM discussion on cnczone was active, I started on a wiki page about open source CAM: http://www.editthis.info/opencam/Main_Page
especially, see the "oo ideas" section and see if it makes any sense
looking in a bit
awallin: Thanks for the list (it's what I used to find most of the apps I evaluated)
anyone want to help me with the logic of handling a graphic file to creat uniform thumbnails?
sometimes I'll pass it landscape images, sometimes I'll pass it portrait images... but the generation of the thumbs need to be a given size. Looking to creat a minimum thumbsize.
A-L-P-H-A: imagemagick ?
I can do the grunt, just the logic... I'm looking at it now.
using the GD2 library from php 5+
Dallur: you're welcome!
lerneaen_hydra: you'll get image-to-gcode with a checkout of the 2.1 branch or HEAD
jepler: oh, ok, nice
is there even a premade link to it in the axis GUI?
lerneaen_hydra: you probably have to add some additional python modules: python2.4-imaging-tk and python2.4-numarray
lerneaen_hydra: yes, if you choose a .png file it kicks off image-to-gcode automatically
I take it 2.1 testing adds those packages when using apt?
awallin: it makes sense
lerman_______ is now known as lerman
lerneaen_hydra: yes, once there is a package installable with apt, those requirements will automatically be pulled in
anonimasu: good. that's a start :)
jepler: sounds perfect :)
awallin: let's continue this after I've gotten home
anonimasu: sure, I'll be online later today also
awallin: later or late?
well I'll be here later or late.. but I dont know how active I'll be
pertty tired :9
anonimasu: in my part of the world it's now about 17. I'll be at work until maybe 21, then at home
it's 16:04 here :)
that sounds like when I'll be around
anonimasu: you're part of the swedish emc-gang? ;)
need go get the dog outside
let's hope Dallur digs out the python source before that
I will dig out the python stuff in 3 hours 11 minutes and 7 seconds
03jepler 07HEAD * 10emc2/debian/control.in: groff is required to build html version of manpages
oh... awallin, finished.
finished like probably 30 minutes ago... but didn't say anything
admiring it myself. :) http://22.214.171.124/bc3/gallery.php
nice. know of any easy image plugin for wordpress? I'd need that.
awallin, yeah... I hacked one together.
[15:46:40] <A-L-P-H-A> http://126.96.36.199/mn/index.php/archives/commercial-retail/6
I hacked a single.php file, to select all related attachments that are image mime types.
and to display them, with a lightbox2 script.
what would it take to get that working on a standard wordpress install ?
really, not much.
and will it all go to hell the next time I upgrade wordpress? I don't want that
no. it won't go through hell.
as it's part of your theme, not wordpress itself.
I'll share the code in a minute or two.
bbl, heading home but stopping in a couple of places
just editing it myself to make it look nicer.
A-L-P-H-A: you do web-design for a living?
sending you teh file.
you can look at the code in there.
03jepler 07HEAD * 10emc2/debian/ (emc2-dev.files emc2.files.in): put new files in .debs
where does it go? in the theme dir
well... see that theme I have...
yep, the link you sent
it's really customized theme... it's not a direct plugin... you also need the lightbox wordpress plugin.
and a hack I put together... starts on line 27 to line 42.
so you need lightbox wordpress plugin, phpthumb, and lines 27 to line 42 of that file I just dcc'd you.
then you can just attach files, like you normally do, to the post, and it'll have the images automatically come up. all the time.
you could setup a tag system, to have conditions for the gallery to come up.
but I didn't bother.
anonimasu: you there?
52 minutes ago
A-L-P-H-A: hmm, it's not so easy afterall.. I'll have a go at lightbox sometime
this is sad... I want pizza.
but I'm having pizza tonight already.
so I can';t have it for lunch...
pizza dinner date?
tonight's pizza I'm buying, and movies at a friends house.
tomorrow is the real date... dinner and clubing with another girl.
not the same one as monday... another girl.
carribean food... :)
haven't had that in a little while.
now now A-L-P-H-A, what's this, you, two-timing?
you naughty boy
first one is leaving at the end of the month to Aussieland... so I might as well get a head start with another girl.
lerneaen_hydra, you would too, if you could. :)
aha -- in jmk's schematic the things labeled "K" that look like capactors aren't.
yeah it took me awhile to figure that out
heh. I guess the coffee is working - I managed it before the PDF came up :)
* jepler wonders where his coffee mug is
oh - that's a bad situation
so K1 closes and lets the caps charge at a limited current. then K2 and K3 close in order, taking the resistors out of the circuit and also completing the "closed when ready to run" circuit
after power off, the big caps discharge with a 27-second time constant? that seems pretty long.
you mean the 2700 uF ones?
it's not bad actually. that's the DC supply for the VFD, so it isn't directly connected to motor loads
A-L-P-H-A: hmm, I wouldn't be too sure
The spec on the time delay relay is ambiguous. It should say time delay on close, NO delay on open. If there is a delay on open, that will delay the estop.
there will be another supply connected to the bottom set of AC lines for the servos, and that will likely have a faster time constant
I wouldn't be able to do that with a clear concionce
(that spelling is so bad I can barely stand it)
Did you consciously spell it that way?
I did, that's the thing
he subconsciously spelled it incorrectly, so his conscience would be alerted to the unconscionable lack of spelling prowess :)
hmm, in sweden K is very often used as an abbrieviation for kontaktor, which is a large relay (say, >10A, usually three-phase)
well, C was already taken ;)
so, K isn't a standard letter for relay-thing?
(with a nod to my Spanish speaking colleagues.)
oh, and not SI as in metric convention :p
K IS a standard symbol for relay contacts.
it is a SI standard, I think he's trying to say ;)
(or an SI standard, depending)
In case no one noticed that I was gone, I'm back from San Francisco.
I mean, welcome back :)
* lerneaen_hydra is reminded of the famous bash quote:
I did get some work done. Changed the named parameter stuff to the alternative (that seemed to be what most people wanted). And implemented comments with embedded parameters to help in debugging.
cool. I just saw the commits
okay. lunch time.
[16:52:40] <lerneaen_hydra> http://www.bash.org/?5300
though CIA seems to be shirking
And I was thinking bash-- Bourne Again SHell
very dissimilar things ;)
Does verfiy report the line number it finds an error on?
doesn't seem to in tkemc
I just get a window that says "Unknown word starting with a" and an OK button
(I wrote 'ha ha ha' part way through 3D_Chips.ngc)
okay. Thanks Jeff.
Does axis report the offending line number.
Yes. It says: Near line 44 of 3D_Chips.ngc: Unknown word starting with a
"ha ha ha" is actually on line 43
glad I could help
Now if we made "ha ha ha" a named variable....
ha ha ha
03jepler 07v2_1_branch * 10emc2/debian/ (emc2-docs.files emc2-dev.files emc2.files.in): some files were unpackaged -- package them
03jepler 07HEAD * 10emc2/debian/ (emc2-dev.files emc2.files.in): move section 3 manpages from the main package to -dev
anyone happen to have the URL of the video with the lathe that cuts flats?
the one with the really terrible narration
why does this circuit work only with the C1458 op-amp and not with M5218? http://members.home.nl/guidomennen/temp/voeding2.GIF
SWPadnos: re the emc-users mail, you don't happen to know of an open parametric modeling format?
is IGES open ?
K`zan, bought geckos? :D
A-L-P-H-A: I've pestered you about the book "the game", right?
kinda interested in reading it
it's damn good
but I don't really have a problem... but would be fun to read already
I mainly read it for the comic effect
so you pick up many swedish girls with that kind of technique?
not so much the poor sods that need that help
I'm surprised they don't hate your guts... cause you're an americant
haha, people don't notice untill I tell them
then they're all "ey wtf, you don't seem like a stereotypical american"
and then I say that "no, I can't really say I fulfil the stereotypical obese-religous-patriot-braindead american feeling"
and then all is good
seriously though, most people here wouldn't hate americans, unless they show very stereotypical behavior
it's americant. :)
or I send the US on an invade-canada campaign again
(south park the movie)
Dallur: I'm waiting for that python code real soon now... ;)
you know... I saw the south park movie with this girl... she was sweet... but weird.
made out with her, and went drinking with her too... her bro was a good friend of mine actually... and I was younger then her by like 3-4 years if I remember correctly.
their parents gave all all the siblings names that started with A... kinda weird...
going out with a friends sister
it was Alexis, Andrew, Adam, and Adam's twin... I can't remember... She was the oldest.
she was older than me, but I think 3 years...
Andrew was older than me, and we were friends from highschool.
Andrew had anger problems.
didn't you get... strange situations?
especially with that?
<shrug> I think Alexis abused him pychologically when he was growing up. She was fun that way. :)
Andrew didn't know... until later.
ah, heh, interesting
:) He was actually cool with it...
not a lot you can do about it
As he can't tell her what to do... and I'm not a huge asshole in real life... just on irc. :)
like GAAH I HAVE ANGER PROBLEMS!!!111oneonoen argh!
will only make it self-fulfilling
so you say
nono... I played him squash on day... and he was doing terrible... I was killing him.
so he takes his racquet and smashes against the wall... at the time he had an alu racquet... so it got all deformed.
never saw him spaz out before... kinda made me not want to be friends with him.
friendship ended soon after that... cause he was spazstic...
he had an old racquet.
eholmgren, you don't know what squash is?
[18:45:39] <A-L-P-H-A> http://en.wikipedia.org/wiki/Squash_%28sport%29
balls are softer I see
no. my balls are tender.
the american raquet ones are hard as hell and hurt real nice when you get nailed
getting hit by a squash ball isn't much fun either.
I get hit in the back of my head, cause I was in the way... that hurt.
and another one in a butt cheek... that stung bad.
my only experience is from playing raquetball with a tennis raquet as a 10 y/o
the extra leverage probably didn't make it any safer
whel are you?
* A-L-P-H-A looks at rafa weird.
eholmgren, different tensions in the strings/gut...
when go emc 21?
when go emc 2.1?
* eholmgren hits A-L-P-H-A in the butt cheek with a well-aimed shot
rafa, when it's ready.
voccÊs são fogo
traceroute gets me to brazil
too bad unicode is messed up somewhere
I'm contemplating porting my gwiz code to python. Is there a good tutorial on writing gui code in pyton? I have the TKinter document, but that doesn't seem to show the stuff that Axis uses: menus, tabbed panes, etc.
Looked tat axis code, found the nf library source, but that seems pretty opaque.
I know no python
Python (an interpreted language) seems to call tcl (another interpreted language) and uses a library (Tk) which is written in tcl.
lerman: for simple widgets you can look at pyvcp
I at least need to have menus (like at the top of the Axis screen).
lerman: axis uses some strange and low-level things to interface to Tk
lerman: it's not very good to learn Tkinter from
Strange -- even to you :-)
[19:08:36] <lerman> http://infohost.nmt.edu/tcc/help/pubs/tkinter/tkinter.pdf
should tell me what I need.
But I don't see the type of menus at the top of Axis.
It could be that it is there. The document shows examples of code, but not examples of the output from the code.
lerman: here's a simple example with a menubar: http://pastebin.ca/314422
a toolbar is simply a Frame with a bunch of Button widgets inside it
meh, starwars over telnet is slow today :(
lerneaen_hydra: that's so old
if you go over ipv6 it's in colour
still ASCII, right?
thanks, jepler. That's easy. I just ran it and it looks like what I need to start.
lerman: you're welcome
Can anyone explain the watch dog reset used on the vital motenc boards?where this signal originates. how i get it to a pin?
I know what a shear and brake do, but what does a slip roll do to sheet metal?
mtedad: afaik a watchdog expects a periodic signal to a certain pin
I would check the docs for the frequency needed, then set up a freqgen in axis, and link the square wave output to that pin on the motenc
i was hoping the "EMCMOT_ENABLE_WATCHDOG" in the file motion.h was the origin.
mtedad: if it's enabled I suspect a proper pin gets exported by the driver
but you still need to do the rest
afaik there was a pin belonging to the motion controller
let me look
mtedad: nope, I was wrong
no watchdog pin in the motion controller
I'm going from the comments in hal_motenc.c here, but it looks like you probably ant to set the watchdog-control parameter to 0x14 or 0x15 so that DAC writes reset the watchdog
otherwise you can arrange for a component to always write TRUE to a bit output and link that to watchdog-reset
jepler: always true doesn't really sound like a watchdog
maybe something at the beginning of the servo thread to write 0, shortly afterwards 1
but the 0x14 sounds better
mtedad: setp motenc.0.watchdog-control 0x14
mtedad: that should be enough
alex_joni: OK I read the code a little bit on that -- the watchdog-reset pin is HAL_IO, and it resets it to 0 after each read and reset of the hardware
jepler: sorry, the watchdog reset seems to be self clearing.. so writing 1 to it is enough
where do i check to make sure it got to the 0x14 parm?
mtedad: setp motenc.0.watchdog-control 0x14
you need to put that in your .hal file
you can check with halcmd show param motenc.0.watchdog-control
03jepler 07v2_1_branch * 10emc2/docs/src/hal/drivers.lyx: improve documentation of motenc watchdog based on source code comments
03jepler 07HEAD * 10emc2/docs/src/hal/drivers.lyx: merge rev 188.8.131.52: improve documentation of motenc watchdog based on source code comments
alex what is afaik.
"as far as I know"
also you may run into "afaict", "as far as I can tell"
np= no problem
iirc - if I recall correctly
irc = nice waste of time
got the setp in the hal motion file. now how do I get it to a pin.
if you used 0x14 as the value for watchdog-control then there's no need to do anything with watchdog-reset -- just make sure the dac-write function is being called by hal.
(at least, that's how I read the source -- I don't own one of these cards)
it's a parameter,setp = set parameter
mtedad: same here
I'll comment out the wdresets in the io file.
mtedad: you don't need to
that setp is enough
the dac-write is coming from a halfile. Emc won't load uncommented its looking for a pin that does not exist.
mtedad: as we said.. we don't own a board, might elaborate a bit?
what pin doesn't exist?
mtedad: as I said before all you need is the "setp motenc.0.watchdog-control 0x14" line
you can leave the rest commented out
now i hope this cleans up the problems im haviing calibrating the axis.
watchdog is not related to calibration
is it not related to the count period.
no, watchdog only resets the card if the computer locks, as a safety feature
a watchdog is a safety feature which needs a periodic action
if that periodic action doesn't happen, the DAC's on the card get zeroed
sometimes a watchdog also resets the computer
well give me an idea for the calibration . i can move + ten inches in actual and on readout, but - ten goes 10.024.
subsequent moves sum the error
mtedad: so you command a 10" move, and it goes 10.024
if you commandit to go to 0.. what happens?
I command +10 it goes +10. I command -10 it goes -10.024
how about 10,0,10,0 ?
haven't tried it.
Rugludallur: found the python thing we talked about yet?
awallin: about to hit the send button :D
awallin: there's news for you :P
awallin: who else wanted it ?
alex_joni: oh, tell me!
awallin: LH ?
awallin: is your nick registered?
alex_joni: yes I think so
ok.. plan to use any other nick, or will this one stick?
this will probably work ok
ok, care for a cloak?
what's that? :)
something like that :)
awallin: Your name is damn similiar to mine
alex_joni: how is that going to improve my life?
awallin: not sure.. people recognize you as an emc devel?
I'm a Anders too :D
LH: nothing :D
anonimasu: you can grab it from there to
Rugludallur: what's this magic happening here?
awallin: something else :)
lerneaen_hydra: you're not in the loop
boo hoo hoo :p
doing htat now
ok, im going to be afk for the rest of the evening, working on some hardware :P
Rugludallur: have fun
does anyone know which library the db25 connector is in eagle?
did anyone get Rugludallur's python code to run? I installed vpython but get some errors...
I'm trying dxf2nc.py Colors.dxf
03jepler 07HEAD * 10emc2/src/emc/motion/motion.c:
in emc_message_handler, calling emcErrorPut directly (instead of through
reportError) avoids a level of printf formatting, which could have caused
problems for messages containing percent characters.
any way to create a simple DXF file in ubuntu?
qcad will write a dxf
I'm not sure if the Colors.dxf that Dallur included is a sample file, or something else... it's causing an error when I try to run dxf2nc on it
wtf wth ? http://blog.modernmechanix.com/2007/01/12/no-wires-on-radio-knife/
hrrmmnn... eagle seems nice
what's so special about python variables beginning with underscore? _myvar = 1
awallin: _ as a prefix often indicates that the author intended it to be "private"
ok, but it's not a language construct
I've seen _ as a suffix when a name conflicts with a reserved word: from_ = ...
there is a rule about names in classes which begin *but do not end* with __
I don't exactly know what the rule is, but the consequence is that I never use such names
how many of the last pins on a db25 are ground?
[21:20:51] <jepler> http://www.diyha.co.uk/electronics/parallel.html
why do I get invalid token when I split lines with '\' ?
is mystring.partition() something included in the standard library ?
awallin: partition is new in python2.5, and is not available in 2.4.
not sure about your problem with backslashes
jepler: right, so I need to upgrade to 2.5 is that available for ubuntu via the normal packet manager
at least, I don't think so
I see 2.2 2.3 and 2.4 in the packet list
the program Rugludallur sent uses partition(), and I'd like to try it
is it hard to compile from source, I only see source downloads on www.python.org
python itself is not hard to compile: ./configure; make; sudo make altinstall
but if Rugludallur's program requires other libraries that build on python (I think he mentioned vpython) then you have to compile all of those for the new version of python
and some are not easy to build
I'm downloading the 10Mb 2.5 source now
(I don't know about vpython specifically, whether it's tough or not)
you probably want to 'apt-get build-dep python2.4' before trying to build that package
there's an install-sh
looks like install-sh is a script for copying one file to its installed location, not for building and installing all of python
jepler: now I have the 'python' executable, you suggested sudo make altinstall ? that will move it to the correct place ?
awallin: altinstall should copy it to /usr/local/bin/python2.5, meaning that it will be a separate python installation from the system version of python
(you don't want to replace the system version because many of the programs installed depend on that version and the modules available for it)
now it's compiling a whole lot of .py files... maybe vpython is among those?
now in 2.4 import visual works, but in 2.5 I get 'no module named visual'
yep that's what I was saying about the "extensions"
"if Rugludallur's program requires other libraries that build on python (I think he mentioned vpython) then you have to compile all of those for the new version of python"
yes, I remember.
I'm now looking at how vpython is built
def partition(s, sep):
parts = s.split(sep, 1)
if len(parts) == 1: return s, "", ""
else: return parts, sep, parts
this may be an easier way
here's an implementation of the 'partition' method ^^^
thanks, I'll try that
change each s.partition(sep) to partition(s, sep)
and put the def partition in the file near the top
now it seems to run, but I get warnings because of unsupported objects in my DXF file
yay, now it worked! with the supplied sample file
I wish I'd suggested doing that at the outset
ok, so for those of you who want to try Dallur's dxf to g-code program he posted earlier, here is a version that works with python 2.4 http://www.pastebin.ca/314604
additionally you will need dxf.py and the sample file colors.dxf from dallur's zip file.
the g-code output from colors.dxf is not that interesting, but it's a start...
what's the glxgears option to see the framerate?
is 2700 ok for a standard glxgears window size?
it's outstanding for software rendering, or about 1/4 -1/5 the rate for a 7800GT
this is a 5200
I get about 700 fps for full screen
my "big machine" gets 13500 or 14000, I think
this is my home (oldish) machine
that goes way down at full screen, since that's 1920x1200
this is only 1024x768 :)
I'm sitting about 2m away from my 21"
so roughly 1/3 the number of pixels ...
hmm, for a small window I get around 500. but this is under vmware...
vmware actually does have openGL in their guest drivers
you have to disable it if you don't want it
drops to about 20 for full screen :)
yep - 13830-ish at original size, around 800 for full screen (on teh monitor with the menu bars)
and only 740 on the empty monitor :)
alex_joni: they accepted the offer :)
glxgears is *not* a benchmark
is that on the quadro 3500 fx ?
nope, the 7800GT
guess you'll be moving now?
SWPadnos: ah.. ok
:) closing in march
skunkworks: bought a house?
you get very varied performance on equivalent hardware
jepler: yes - on the missisippi - (going to have learn how to spell it now)
the Quadro FX 3500 was about 1/3 the speed, last time I checked
in trempeauleau - going to be able to have walnut burgers when ever we want now ;)
how far is that from where you're living now?
about 20 minutes
at 15 miles away you can't have walnut burgers whenever you want?
ooh time to go home
have a good weekend everybody!
oh - it is a block away (walnut burgers now)
same to you jepler
see you later jepler
skunkworks: still spelled it wrong
skunkworks, now you'll be able to grill the walnut burgers yourself ;)
skunkworks: how about all the work you put into this house?
and throw them in the river if you burn them
you gonna do that again now?
alex_joni: Hope it will pay for it's self ;)
this house is pretty much done - just needs a garage.. (don't have one now anyways)
I should be able to surf the internet sitting by the river ;)
(this house is on second street - overlooking the river.
* skunkworks might be a bit excited
so there won't be much tinkering time for a few months
* awallin is going to sleep
well -time to run and have dinner. bbl