cradek / jepler: what does the roughing function do in the Eagle script? Up to this point I've been perfecting milling without roughing.
I assume it just takes off more material around the traces
but i only have a little bit of PCB left and need to make sure this one comes out correctly :-/
crepincdotcom: it is intended for clearing more space around the traces using a flat end mill. I usually don't use it (set it to 0) too
ahh. so its only astetic
you could say that
usually I use a big GND polygon on the whole board so it doesn't actually do any good
jmkasunich Yo, tooling marks guy!
(topic change) when a servo if rapiding to a location, before it gets to the location does EMC let it coast and watch the encoder to guide it, assume the controller has braking, or reverse the pulses to stop the motor accuractly?
jmkasunich: Shouldn't I be able to have tooling marks less than what I might get from a (lets say) table saw?
i seem to have killed it again :-/
crepincdotcom: I'm pretty sure that emc comensates it's velocity/acceleration for "curves" at the like based upon it's setup.
so "braking"/deccelerating is configurable?
crepincdotcom, if you're using a stepper motor, or a servo drive that accepts step/dir input, then EMC has nothing to do with whether the motor is coasting or braking - that's a driver thing
DEFAULT_VELOCITY = 0.0167
The initial velocity used for axis or coordinated axis motion, in user units per second. The value shown is one inch per minute.
DEFAULT_ACCELERATION = 2.0
[00:28:47] <Jymmmm> http://linuxcnc.org/docs/devel/html/config/ini_config/index.html
but EMC has to know what to do to make it stop at the corect point no?
SWPadnos you sure?
it cant simply allow the motor to coast and assume it will get to point X accuratly
EMC will send a stream of step pulses starting at 0 Hz and going up to some maximum, then for deceleration it will reduce the frequency back to 0 (or whatever the needed end velocity is)
crepincdotcom: are you talking about servo or stepper
so it wont send negative pulses then
are you talking about a driver like Skunworks'
Jymmmm, yes, I'm sure
I'm talking about a generic h-bridge
for servo, there is a PID loop that drives the bridge whichever way it needs to
you have to tune it
ok, with the H-bridge, EMC will tell the driver to push the motor when it needs pushing - there may be coasting from time to time
crepincdotcom: I could be totaly wrong here, but emc won't "brake" ther motor, but you can have it "slow things" (per se)
if it's going too fast, EMC will tell the driver to push the other way a little to slow it down
but if it "slows" the pulses, it is in effect leting the motor coast
it will too ask for braking
SWPadnos: thats what i was asking thank you
braking = reverse current and/or voltage on the motor
Jymmmm, it doesn't do a "braking phase", but the effect of accelerating against the motion is braking ...
SWPadnos what you're talking about is strictly servo and not stepper related, correct?
sort of, but not really
well, yes. there's no feedback with steppers, so it's only servo
so the answer is yes, that's servo only ;)
unless you have steppers with encoders
sorry to complicate things
SWPadnos: isn't part of "look ahead" to accombidate for "um, there's a corner coming up, better slow down a bit"
steppers with encoders are not the same thing as servos
* crepincdotcom hides then
Jymmmm, yes - the motor gets "braked" because that's what we call "decelerating"
SWPadnos: why "braked" instead of just "coast" ?
this isn't driving a car
crepincdotcom, steppers have no reserve torque - if they lag, then there's nothing you can do except slow down the entire move - that's the opposite of what PID usually does
there is no coast, there is no brake
right - what jmkasunich said
there is only applied power vs consumed power
note the "QUOTES"
with steppers, EMC is requesting a position from the motor driver
the drived can apply positive power to the motor, or it can remove power from the motor
and it can do either one in varying amounts
or no power ...
jmkasunich: remove power == coast
"removing power" is pushing the other way in that context, I think
remove power as in exert force against the motion
SWPadnos: right. but if you have steppers with encoders, and the encoder senses that the shaft moves without being told to, EMC will attempt to push it back the other way no?
crepincdotcom, yes, it should
* crepincdotcom wonders who to beleive
it should if you have PID and the position changes
but it won't help when the motor lags behind
steppers with encoders is a very wierd situation, and useless for a discussion like this
they will just confuse things more
crepincdotcom, wait for confirmation from jmkasunich :)
sorry i brought it up then jmkasunich. peope here suggested it before, thought i'd ask
SWPadnos: ok lol
with steppers, EMC has no way to accurately control the amount of "push" so its very limited in what it can do if the encoder says there is an error
03cradek 07TRUNK * 10emc2/configs/max/ (jogwheel.hal max.hal max.ini): add jogwheel support to my well-tested 4 axis configuration
being pushed off position and the motor not having enough power to move are two different situations. they have to be handled diffferently for steppers but they are handled the same for servos.
for steppers, torque is proportional to the displacement between the rotor position and the ideal position (the position the driver is trying to make it go to)
the problem is that as jmkasunich pointed out, there's no proportional drive for steppers - there's only position
if you have very fine microstepping, you could do some closed loop contorl
encoders on steppers might be nice for stopping if you crash into something. to me, it seems not useful for much else.
if you see the motor move a count to the left, you send the driver a few microsteps to the right to correct
you can tell if the part is already screwed up ;)
my steppers lose steps sometimes because they are crap. someone told me to think about encoders.
i guess i won't ;-)
crepincdotcom: that's not good advice :-/
i have some servos, suppose I'll get them running then
are steppers smaller than nema common?
Jymmmm, depends on where you look
there are probably a lot of very tiny steppers and servos in floppy and tape drives
^^ i have a box of those
no, not like that with funky OEM specs, but "standard"
the head motor in a CD-ROM is a tiny motor, and I think they're steppers (could be servo though)
old 5-1/4" disk drives all had NEMA17 unipolar steppers in them
I think I've seen some, but they're too tiny to remember where ;)
nema 17, ok I'll look at that.
I wonder how big motors in 8" floppies were ')
good luck finding 5-1/4" drives these days
* Jymmmm wave his two 5.25" drives at jmkasunich
03cradek 07v2_1_branch * 10emc2/configs/max/ (jogwheel.hal max.hal max.ini): add jogwheel support and tweak jog sliders
[00:43:11] <Jymmmm> http://www.alltronics.com/cgi-bin/item/23M003/55/Minebea%2DAstrosyn%2D17PY%2DQ202%2D03%2DStepper%2DMotor
03cradek 07v2_1_branch * 10emc2/debian/changelog: 'max' jogwheel
note the relatively high voltage and low current - that will be a crappy motor for high speed use
at least its not unipolar like the old floppy drive ones
but how hard would it be to make/find/steal a driver for it?
theyre free ;-)
buy geckos and move the xylotex to these ;)
SWPadnos: mass quantities
That place is just 40 minutes south of me too.
SWPadnos real cheap
Jymmmm: you ask very open ended questions
you mean you need mass quantiteis of steppers and drives?
well realtively cheap compared to motor price of $5.95
do you care about performance?
get L297/298 pairs lol
(top speed, and torque at speed)
the motor is rated at 400mA - you can do that with cheap SOT23 transistors that cost $0.02 each (* 8 or whatever)
jmkasunich max travel is 12" and speed > snail pace
or a snail on a hot date pace (whatever)
a unipolar motor would be easier to drive cheaply, but as jmk is asking, we know nothing about what performance you want and at what speed
Think Sears Craftman's of cnc performance/speed
how many axes?
that kinda rules out the very cheapest approach
well 3/4... could be individual I suppose... I could always machine the PCB's
3 or 4
4 pins per motor, driving 4 transistors, driving a unipolar motor
that would require 12 pins, and a parport doesn't have that many
actually, I bet it does
yes it does
oh hell... this time change shit is messing me up, almost time to go to work.
actually, it's a 5V 400mA motor - you can just about use NPM and PNP transistors from a cheap microcontroller to drive it
NPN, that is
it's not like it's a 200V motor or anything dangerous
still need some passives tho
2n2222 is 800ma
its npn though
phase drive on 12 parport pins is as cheap as you will get, and it'll mostly suck
crepincdotcom is it?
sure - or the SMT version, the MMBT2222
if its a unipolar motor you could do it with those
cradek, that depends on whether you're paying for the computer as well ;)
whats the pnp version od 2n2222 ?
hmmm - 2n2224 or something?
a nice complementary pair is 2n3904/2n3906
* crepincdotcom prefers mosfets
SWPadnos: I don't think so
I wonder if you can get logic level fets in N and P channel cheap
right - 3904/3906 is what I was thinking of - thanks
jmkasunich: cheap is relative ;-)
no drive circuit needed, just tie the gates together, buffer with a cmos chip that has rail-to-rail outputs (if your parport doesn't)
you don't really need a FET fro 5V 400mA ... (though it is "better")
yes, you can - there are $0.11 SMT FETS
MMBT4403 and friends
saves the various resistors and such you'd need with bipolars
03cradek 07v2_1_branch * 10emc2/VERSION: 2.1.2 release
high power CMOS logic ;-)
4401 is NPN, 4403 is PNP
I just remembered something I saw once
600mA or thereabouts, $0.09 in 10 qty at digikey
somebody stacked about 20 4000 series hex inverter packages, then paralleled all 6 sections
ops - that's the 200mA part. $0.15 for the 600mA
haha jmkasunich thats awsome
don't trust any current rating you see on a datasheet
i really want to try that now
I've seen these transistors work at currents that desoldered them. tacked them back on and they still worked
SWPadnos: some of those current ratings are based on case temp of 25C
case, not ambient
not saying they won't work, but they sure as heck won't be guaranteed to work
to know the real current rating you have to figure out the allowable temp rise, then the allowable dissipation, then the current that gives that dissipation
I've had very good experiences with those little parts
most power mosfet vendors spec the current by doing that calc with Tcase = 25C and Tjunction at max
yep. some are rated in case dissipation (350mW or so)
haha i like SWPadnos' line out of context
back in 90
03cradek 07v2_1_branch * 10emc2/VERSION: bump after release
03cradek 07v2_1_branch * 10emc2/debian/emc2.files.in: forgot to add this
holy crap: http://news.nationalgeographic.com/news/2007/02/070226-sinkhole-photo.html
I thought that happened in Cali on a regular basis?
when using a gantry on the x axis ... does it matter where I place the leadnut? ie. on either face or in the middle
EMC2.1.2 is released!
cradek has changed the topic to: Welcome! EMC (Enhanced Machine Controller) is a linux-based CNC control. | Latest release: EMC 2.1.2 | http://www.linuxcnc.org
hokay. I'm looking for a CNC backplotter, not a full controller.
has anyone ever tried using EMC for just the backplotting and toolpath stuff?
you can run EMC2 without hardware or realtime
and that will essentially give me the toolpath tracing and that's it?
yes it will be just like running machinery, with no machinery
gah, I apologize
you can build EMC2 on most any modern Linux system, or if you pick Ubuntu 6.06, there are prebuilt packages
but specifically, how is the toolpath trace and backplot stuff
i notice that the EMC g-code isn't exactly Fanuc standard
so I'm concerned testing some production programs on it
so while I'm not so much concerned with what appears to be amazing backplotting
yes I bet there are some differences.
i'm concerned I'm going to crash as 350,000 dollar mill
i guess that's what the e-stop is for
well, I'll give this a shot and see what happens
if you also run the mill with EMC, you'll know they will match
hahah, yeah, i suppose that would be true
does EMC support high speed machining
with nurbs curves or otherwise
we have the beginning of spline and nurb work, but nothing in the releases yet
we have not seen a lot of users asking for it yet, but a few developers are interested and working on it
yeah, not that I know anything about your demographics
but i don't really think most people can afford the inserts to do HSM
if they're hobbyists persuing this
still, it would be cool to see
we have some hobbyists and some pros - a huge variety of machines
do you know offhand if EMC supports being wired to a Fanuc control panel
of course since EMC is free/open source, we can't know exactly what people are doing with it - they're sure not required to tell us
like, for selecting rapid overrides and stuff
EMC2 has a lot of flexibility for wiring to buttons, jogwheels, etc
well, thank you sir for your time
questions, the remix
what's the ubutnu repository for emc
[02:36:40] <cradek> http://linuxcnc.org/index.php?option=com_content&task=view&id=21&Itemid=4&lang=en
some information here - but it's mostly concerned with getting a realtime system
You can get the simulator only:
note those sim packages are ONLY for ubuntu 6.06 LTS release
i'm running 6.10
you will have to build it from source then
no big deal
[02:39:20] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Pure_Simulator
not entirely sure that list of build dependencies is still right - please update it if it's not
[02:40:39] <SWPadnos> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Simulator_on_Ubuntu_6_10_Edgy_Eft
oh hey, look at that
this just keeps getting better
i bet if i just sit here long enough, someone will ssh into my box, install it, then do all my programming for me
now i wonder if this will work
hmmm. I started to do that, but since I'm running edgy for x86_64, it didn't work
SWPadnos: what didn't work, any idea?
there's a problem with asm/bitops.h on x86_64
it isn't there, and the RTAPI bitops says this CPU isn't supported
(at least I think it isn't there)
new question: looking at the list of g-codes in the manual, i don't see support for the lathe g-codes
am i just looking in the wrong place?
there is some lathe support in EMC2 but it's not at all complete yet
still very impressive
and add anything that's missing please :)
we have single point threading with multipass canned cycles (compound feed), and tool shape compensation
and lathe tool offsets
hmmm - that page (and some of the other 20 or so that mention "lathe" may need updating)
SWPadnos: thanks for volunteering!
hahahah - I've almost never touched a lathe :)
you guys do have the really important lathe cycles
that page doesn't say what we have, it's a list of how it's done on various controls
(it may also say what we have, I haven't read the whole thing)
are you wanting to use EMC2 to simulate fanuc lathe programs?
that's probably not going to work well, if so
I assumed you were asking about mill, which is much more complete
a variety of controllers
Fanuc and Haas, primarily
ok you know full well how they're all incompatible then...
my hope was there was some sort of abstraction
between the g-codes themselves and the emc implementation
but, i fully understand the purpose of emc is to be a machine controller, not to be an open source backplot
it's sort of my only hope, as there's no open source backplot
so I'm willing to run with this as far as it goes
we like to keep compatibility with commercial controls when we can, but we're not afraid to head in our own directions when we have to
no doubt, there's no reason to make yourself insane trying to keep up with the million different controllers
it would be cool to have a postprocessor kind of thing though
EMC does not behave the same way as Haas Mill nor does it support the follow idiosyncracies of Lathe mode
not like i'm in any position to write or contribute
yeah, the more commercial controls disagree, the worse our compatibility gets, but that's how it goes
toastydeath: do you have sample code on the Haas that blows up with EMC?
not yet, because i haven't installed EMC fully yet
but I will be putting it through it's paces once i do
I have a perl hack/preprocessor to take my own Haas code and make it EMC palatable
hmm, that's a kind of workable idea
for dumb stuff like just swapping g-codes i have no problem writing a script to swap it out
but it only does Mill; Lathe is too different
EMC with the AXIS gui has the concept of input filters - your script could translate transparently from haas/fanuc to ngc
ds3: your perl script or EMC
there is more than that... some of the can cycles behave slightly differently
toastydeath: my perl script only does Mill; the EMC lathe handling is too different
yes, like the automatic roughing/finishing cycles
i don't have a problem making two seperate scripts
AFAIK, there are no free/cheap controller with G70/G71 support :(
then there is G47 on the mills...
remind us amateurs what g70/g71 are
g70-74 are automatic lathe cycles
that do roughing
so you don't have to sit there and program your toolpath to rough down, and then another one for finishing
G70 and G71 are roughing/finishing cycles; you define the geometry and tell the lathe to leave certain amount of stock in the roughing and tell it the amount to take off per pass
that doesn't sound hard to add. I have a fanuc lathe book that a user sent me just the other day - I'll have a look
it automatically figures out the number of passes; it will also let you over ride the feed rate for roughing; the Fanuc flavor will let you override the tool
basically, without 71-74, the controller is not lathe-ready
I assume the passes are deeper at the beginning?
but it looks like it has amazing mill support
it's not like threading
cradek: the strategy to use is not that obvious as you need to figure out if they should be boring or turning
let me find an example
70, as i recall, is turning
71 is facing
72 is finishing
if you guys would like to help us get the information needed to add those features, it would be great if you would add them to our wiki
and 73 is somthing retarded like 3d roughing
we need lathe experts as we continue adding lathe support
I would be happy to get you the info, samples, and even test them
whose implementation do you want
the wiki is probably the best way
haas or faunc
(haas is what i know better)
i'd be happy to add stuff i run into as i use it as a backplot
I think ngc/mill is more compatible with fanuc, but I know almost nothing about commercial controls
don't know exactly where you want it
haas -G71 is roughing, G70 is finishing
maybe "the one that's most sane" is what we want
G71 P25 Q28 U0.01 W0.02 D0.125 F0.007
ds3: i have it backwards, then
toastydeath: yeah... I am looking at code that I ran
Haas and Faunc are identical on the lathe for most of it
it's only when you get to 5 axis stuff that they deviate
toastydeath: uh... no... threading is quite different
and some other minor areas
(like threading and everything else)
well faunc and haas use different words, yes
hmm, i guess that is kind of important.
the parameters have different meanings and units :(
i don't like the haas controller interface
the editing stuff is cool
G71 - P/Q start and end block numbers that describe the geometry; U/W - amount of material to leave X and Y respectively (U/W are stndard lathe relative mode; there is no G91); D - depth of cut per pass; F feed rate over ride
but the interface itself is kind of dumb
G70 P25 Q28
P/Q are the geometry block numbers
I have only used one Fanuc lathe and that was painful to setup; had to manually type in machine coordinates into the offset tables; got spoiled by just hitting Diameter measure
there are so many damn fanuc controllers it's absurd
some of them are really, really bad
some of them are very good
what i like about them over hass is the knobs for feedrate and rapid override
makes proofing programs very easy
can you single block on Fanuc?
single block, block delete, and opt stop are all switches on the control panel
very clearly marked in case you need to hit them real fast
cradek: is there any EMC lathe person in the bay area with space for a lathe?
I'm really the only one who has worked on the lathe part of EMC2, and I'm in the midwest
and I could use a full size mill more than a lathe :-)
the local community college is thinking of getting rid of a Fanuc control based lathe in the next year or so
would be a good machine to check against; I've used it and it has quirks but it works
don't think I know of any takers, but thanks
I'd offer to house it but Ihave no place
now if you have an old bridgeport cnc, let's talk
a series 1 would make a fine EMC conversion
heheh... with the BOSS5 cabinets? ;)
with or without :-)
wow, emc did NOT like my lathe program
well, i suppose it's not mine
but the school has a Haas lathe already and is going to replace that with another Haas lathe so I can probally convince the instructors to let me check things against it if needed
toastydeath: first fatal problem; EMC does not understand lathe tool change
it also does not like G18
huh. mabye that's not it.
sure, G18 is fine
g18 is fine
it doesn't like G96/G92 either
oops...G96/G97 I mean
nope, no CSS
think it also chokes on G50 if I read my notes right
because i think that's scaling on the mill
or cancel scale, or something
when it says "near line #"
does it mean the line before it or after it
usually it's the line number given, but it may be +- 1
damn, i don't think this can do lathe stuff period
it's not liking my spindle g-codes
so close =(
Almost ANY language that says "near line #" or something like that, usually means " I got to line x, and you mucked up something and I can't continue anymore till you fix it"
hmm, if i strip out all the misc crap
it takes the program
so close =(
pretty nice gui though
i have to admit.
the gui beats the pants off haas
i'm trying to figure out how to setup a tool table so i can do cutter comp, the UG make reference to the 'tool table page' but i can't seem to find that page
i tried google it with no luck
ok, i found in the max sample config dir a max.tbl file, so i guess i just create one with the pocket number of the tool that my cam assigned and try it
I finally resolved to get my bridgeport running and bought a Mesa board
but I'm at a loss as to how to get started.
yes, going analog
They other guys are working on drivers for it, but no ETA
do you have motors / drivers yet?
there is a driver
I have motors and drivers
there is one now, and will be another config "real soon now"
ok - servos with analog inputs?
(err - analog inputs on the drives)
yes, +/- 10v
more specifically, I'm trying to figure out how to enable the drives and hook up limit switches
have you looked at the m5i20 sample config?
I should look at it again
ok. that sets up the mesa hardware and many connections to the motion controller
I'm not sure it deals with all the other I/O, but I think it does have some spindle outputs and limit/home switch inputs in it
probably a coolant output as well, but I'm not sure (I'm not looking at it at the moment)
yep - it looks to be all there
I guess I need to put the mill on the network, can't look at it from here
look in the file m5i20_io.hal for spindle, coolant, lube, mist, spindle brake ...
heh - it's a lot easier if it's on the net
It could be, I only have one network cable on my bench hooked to another computer
you have to click on the revision number to see the file
I guess ubuntu isn't configured to take remote ssh connections?
oh snap, it looks like you are right
I can ping it but it refused to connect
what kind of madness is this
back to the basement
[04:58:08] <Jymmm> https://help.ubuntu.com/community/SSHHowto
what linux distro doesn't come with ssh-server
this is pure madness
03jmkasunich 07TRUNK * 10emc2/src/hal/components/stepgen.c: stepgen reworked - new core step generator with high resolution feedback, and tweaks to the position control loop to take advantage of it
so sudo apt-get openssh-server should install it?
yep, then you might have to enable it
I think it enables when you install
we'll know in a few minutes, apparently I need the excercise
ubuntu is targeted at desktops, no ssh by default is more secure
I had little trouble getting it to work tho - apt-get install was about it if I recall correctly
its been some months ago
didn't know ssh wasn't secure
it's an extra service for portscanning
and most people (those who want a desktop OS that's easy to use) don't need ssh
any running service is a _potential_ hole
true, and most people have no idea what it is and have no need to use it
is there a better pdf display program on linux, evince is driving me nuts
I like kpdf
if you install it, it'll grab lots of the KDE libs, so I hope you have a fast connection ;)
SWPadnos I do =)
it was already installed, but evince is the default
SWPadnos Yeah, I'll have you foaming at the mouth soon enough
I still have a faster computer and a higher resolution monitor than you :P
SWPadnos who cares
SWPadnos and I got multiple 10GigE pipes baby!!!!!
unterh, right-click a pdf file, and select "open with other ...
i was thinking about purchasing 3 19" monitors
the bill comes out to 500 bucks
I had a computer with 2 24" lcd monitors on it, but things kept getting lost
then select kpdf, and I think it'll be default from then on
heh - you guys just desxcribed two of the computers in my office ;)
the only issue i ever have on large displays is the mouse
SWPadnos 800MB in 5.7 minutes =)
03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc: fix zero-radius tool comp
at one time, i used 6 laptops for a unified monitor
or was that 8000MB =)
I have 3x19" CRT on this computer, 2x24" LCD on the other one logged in here, and a 22" 3840x2400 LCD on another one :)
a 3072x1536 display is hard on finding your mouse
try it on a 200 PPI LCD
isn't there a beryl module
for like, mouse radar
hit a button and your mouse gives off ripples
probably, but I don't have beryl installed there (or at least not enabled)
i haven't found it in the config yet though, i'm sure it's there
03cradek 07v2_1_branch * 10emc2/src/emc/rs274ngc/interp_convert.cc: fix zero-radius tool comp
03compile-farm 07Ubuntu 5.10 (breezy) non-realtime (2.6.12-10-386) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot1_log.txt
toastyde1th, under System->Preferences->MOuse, there's an option on the Pointers tab called "Hilight the pointer when you press Ctrl"
that is not awesome water ripples
i demand awesome water ripples.
* Jymmm hits toastyde1th with a fire hose
also i demand sshd enabled by default
but apparently i'm not getting that
you get nothing!
dear ubuntu: this is not a kerberos server
please enable sshd
Dear toastyde1th, RTFM <3 ubuntu
what does reading the manual have to do with this
toastyde1th This aint Burger King, you can't have it your way!
so i am noticing!
yes you can, you just have to get in the right line
the "any distribution other than ubuntu" line
no - I'm sure Linspire doens't have sshd enabled by default either
I could almost get used to ubuntu, but emc2 in /etc?
never would have found it
if OpenBSD enables sshd out of the box
it's probably okay for a desktop install of ubuntu
most people would never use it though, not a horrible decision
not a horrible decision, no
it does make it more difficult when you need to ssh into somebody's box to fix something
and it's impossible
sudo apt-get install openssh-server made it run
indeed, it does
nobody can pick a set of default packages that will make everyone happy...
i wasn't aware anyone was upset with sshd being on until a half hour ago
I would be upset if telnet was on
i would be upset if telenet was on, too
unless it was kerberized
then i'd be okay with it.
it's best to have no unneeded ports listening - you can't argue with that
i can, and i will
exit, stage left
i am suggesting it's absurd to turn ssh off and remove a standard posix troubleshooting tool
certainly, turn off as many unused demons as you can
but the only place I've ever seen a security administrator turn off sshd was on a kerberos server
people who run linux should be able to read the source and help themselves, right?
not at all
or at least search a wiki and help themselves
search a wiki, absolutely.
actually, there is so much trash out there about linux, it's amazing people can help themselves at all
though the first time i installed linux it took me 3 weeks to get x11 running on slackware
I was looking up something the other day, and there were a batch of pages that talked about 2.2 kernels
a lot of places are still on 2.2 kernels
but 2.2 is pretty irrelevent from a desktop standpoint
I was trying to get my wireless to work
yeah, the people who care about 2.2 probably not too concerned with wireless
who uses 2.2?
datacenter applications, cluster computing
people who care about a uptime a whole damn lot
very willing to sacrifice performance for reliablity
I have realtime stuff at work running on 2.4, but it is behind a firewall. University finds it, and I'm in trouble
(why would you be in trouble)
no security updates, I'd have to unplug it from the network
we had a computer rooted within 2 hours one time before they added firewalls
i'm still trying to man up enough to install selinux/ldap/kerberos
and play with it
but the last time i tried i was so confused.
* cradek uses NIS and rsh behind a firewall
talk about bravery
living on the edge
now i have that scarface song stuck in my head
i hate you all
push it to the limit.
PAST THE POINT OF NO RETURN
you got me singing that old Cartman favorite, "I hate you guys"
c'mon guys, sing along, you know the words
come sail away, come sail away, come sail away with me - you guys
03compile-farm 07Ubuntu 5.10 (breezy) non-realtime (2.6.12-10-386) * 10emc2head/: build FAILED ; see http://linuxcnc.org/compile_farm/emc2head_slot1_log.txt
apparently the beryl rain effects require a better graphics card than i have
what is the ssh command line option to display X?
that was my guess, can't get emc to start
are you trying to forward X
I use emc that way often
really, I thought it would do it
i used to do that and forgot how
sez "Application initialization failed: couldn't connect to display "myipaddress:0"
also do an xhost +
although that is insecure as all get out
after you use ssh -X, don't mess with DISPLAY
ssh will set it to tunnel for you
ok, that's it then
nevermind xhost is for running dmx
completely ignore me
I'd ignore toastyde1th
strangely, I was, but not on purpose
at work, I ssh into a linux cluster from a windows pc, and so the ssh client has me all spoiled
back in my day i ssh'ed into clusters in the rain
AND LIKED IT
with windows ssh server, you have to click with your toes
i broke two endmills the other day
milling tool steel on a bridgeport kind of sucks
that's no fun
one was a 1" 6 flute dealy
wow, that's really no fun
the other was a 3/4" 3-flute, solid carbide
03jmkasunich 07TRUNK * 10emc2/src/hal/components/stepgen.c: non-realtime simulator doesnt have 'u32'... grumble grumble
but the toolmakers at work convined with me, and gave me a new cutter that is supposedly a+ for this application
five flute roughing endmill, plus some bridgeport tricks to get it to cut properly
the roughing mills are much better
i am really hoping this one works
I was really shocked how much movement you get out of a bridgeport head when milling steel
yeah, and our mills aren't level
they rock on the base
that's exciting too
so one good jolt starts the whole mill vibrating
I hope mine doesn't move too much, I have a series II, over 5000lbs of cast iron
that's what killed the carbide
that, and the powerfeed is jerky at low speed, adding to the problem
yeah, series II shouldn't be a problem
i wanted to take it in to work and run it with a 10" slab mill on the big horizontals
if this doesn't work i'm going to ask if i can do that
what are you making?
03compile-farm 07Ubuntu 5.10 (breezy) non-realtime (2.6.12-10-386) * 10emc2head/: build PASSED
I can go to sleep now
<-- just broke an engraving cutter :(
concrete floors and sharp carbide cutter are a bad combo
do they make ceramic endmill inserts
I'm trying to decide if ds3 dropped the cutter, or is engraving his floor
i'm trying to remember what they use for high speed machining on steel and i can't remember if it's a specific coating, or they somehow got ceramic to survive endmill cuts
oh, it looks like they can use ceramic in endmills
the floor could use some engravings, but alas, it is nothing that reasonable... damn collet released at a bad moment
sorry to hear that
I'm thinking about skipping the home switches on my mill, anyone have any thoughts?
makes for a good re-zeroing of the machine axes
no home switches might have problems if you plan to have multiple machining sessions
if you are concerned about wander
or use fixtures
I did want to use linear encoders originally, which would solve the problem
but it's too much work
abosolute encoders, i assume
no, almost all linear encoders have an index
did not know that
on the Heidenhains, you can move it, it seems to be magnetic
i have not looked at components to build a cnc
so i am unaware as to the options out there
I learned too much shopping on Ebay
"that looks like a good deal, let's go find the manual"
5 hours later...
let's just say I have some drives that need to go back on Ebay from whence they came
SWPadnos wake up!
anyone remember the name of that surplus place that has motors among other things?
K`zan nite pooky
the devel channel shouldn't be flooded with crap :)
SWPadnos: NEC WT610 DLP beamer
40cm image size, at 5.5cm distance
SWPadnos_ is now known as SWPadnos
SWPadnos: NEC WT610 DLP beamer
40cm image size, at 5.5cm distance
err - what?
an interesting beamer
[13:07:07] <alex_joni> http://www.projectorcentral.com/NEC-WT610.htm
not that small, but has some lenses/mirrors inside
to shorten the distance to the projecting surface
hmmm. 40 inches image size at 0.2 feet? that seems impossible
20" at 2" distance seems more impossible
less impossible - 0.2 feet is 24 inches :)
argh. 2.4 inches
40" at 2.5 inches, 100" at 25.9"
seems nice for tight spaces :)
they must have some very good anti-keystoning optics/image processing
I'll bet the throw distance is being represented as the distance from the projector to the surface, and that you need a very high vertical pitch to the screen
wihch makes sense from the photo of the unit
they say something about 55° projection angle
well, it looks pretty cool (though it's only 1080i, and the native res is 1024x768)
do you have one?
no, I just noticed it in an online store
out of production ...
they seemingly stopped manufacturing them
yeah, probably there's another model now
[13:17:00] <alex_joni> http://www.projectorcentral.com/NEC-WT610E.htm
but it has wireless :P
I don't think I'd recommend anything non-1080p to anyone right now
but then again, I'm like that ;)
What's the diff between 1080i and p?
1080i is like normal broadcast, interlaced half-frames at 60 Hz to give you full frames at 30 Hz
1080p is 30 (or 60) full frames per second, displayed prograssively (like a computer display)
interlacing often causes odd artifacts with scenes that have a lot of motion
(half the frame is from 1/60 second ago, and the other half, intertwined with the first half, is from "now")
can you actually get TV where you are? ;)
SWPadnos: ever seen http://www.vitalsystem.com/web/motion/dspmc.php
sounds like an eth-grex
odd. 2 MHz encoder counter, 10 KHz step generator??
that's what that page says ...
wonder what DSP they used :)
sounds like a more crappy one
no - it's probably a DSP rather than an FPGA
that's what I said..
and 10 Khz (if that's no t a typo) is probably being done in software
that's still way too low for a DSP
you said "more crappy one" - I assumed you meant more crappy than the G-Rex, which doesn't have a DSP
hmmm. have you looked at their instructions for using EMC with the motenc-100?
[13:45:51] <SWPadnos> http://www.vitalsystem.com/web/motion/motion100.php
use BDI, get this zip file, copy source code, compile ...
it's so broken
get emc form CVS but use generic.run
oh - that's not as expensive as I thought - $795.50 for the 8-axis board
or maybe I'm used to high proces now
imo.. this is a DSP: http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MSC8144&nodeId=0127958594
heh, or a TI 32000 (?)
4 1GHz cores :)
10 MB 128-bit memory
yep - that's a DSP
I should make a cell-based motion controller
separate core per axis, 6 axis ;)
haha, and send SMS's for g-code
the PS3 motion controller
hmmm - you know, that might actually work
a $300-$500 computer that has excessive performance and multiple processor cores, and can run Linux
yeah, but no RT for it
unless you want to spend a lot of time :D
well, there's RT and then there's RT ;)
the real problem would be getting any signal out of it other than video
you mean for driving motors?
step/dir, analog ...
use the video signal then :D
heh - not enough pins
sure there are..
but update rate is a bit slow
a DVI breakout board
I was thinking analog
use a couple lines for one motor
ok - for analog the output is way fast enough, but not high enough resolution
you need 6 motors lets say
divide one full image in 6 parts
with DVI, if you can get to the TMDS chip, then you can get a 165 MHz update rate
and use that for resolution :)
SC140 Image Processing Library (StarCore) <- NICE
does it have object tracking and stuff?
robin_sz: I changed the PSU, and now it's great :D
Dallur: fix your panel problem? i just found some filename length limit and solved it by shortening the filename. this was in a tabbed (notebook) panel that included other files. very repeatable but i use a custom pyvcp_widgets.py.
So I followed the direction here. https://wiki.ubuntu.com/LiveUsbPendrivePersistent
and it would not see it as a boot device. At the bottom of the page I tried the install-mbr (didn't work) so then I tried the lilo. It sees it as a boot device now but I get an error cannot find kernel vmlinux.
skunkworks: X versus Z check that you told it to use vmlinuz (notice the Z) when you typed in the lilo/grub info if you said vmlinux (notice the X). it may never find a vmlinux , becasue the compressed image is usually called vmlinuz, and searching the web page you posted shows vmlinuz not vmlinux
hmm - ok thanks
i did that once too many times ;)
I thought it was vml<tab> ;)
I just did an apt-get lilo or whatever then ran the command on the page.. I take it I need to do more?
i'm more familiar with grub becuz over the many years of using lilo, i found for the last few years that grub was more flexible for me. at boot time you can ASK grub whats going on...
the instructions on that page look very complex... and i've no experience with booting usb drives (old computers here)
thats ok - I will keep looking.
i'd suggest before you try booting with the usb, that you mount it and look at it
look for the files you told lilo to use and where they are, and look at what device the usb is ( likely sda1 )
then make sure the lilo config agrees
tomp: brilliant, I will test when I get home
attack the error you recieved first, you may be very close already I get an error cannot find kernel vmlinux.
Dallur: heres the ugly bugger... this works t1234567890123456789012345.xml this doesnt t12345678901234567890123456.xml urff!!!
some buffer too small
tomp: hmm In my case the filename was probably around 10 chars
hmm, can you post/paste anything (simplified )
tomp: once I get home I can copy/paste with a screenshot
Dallur: that includes the path!
Dallur: i take that back... does not include path
tomp: python should have no trouble handling paths with 100s of characters, and neither should any common linux filesystem
ok, it gives an error on rendering acheckbutton in one case and renders to screen and works fine in the other
the difference is the file is copied to a differnt length name... just copied, no edits
oooohhhhh -- the HAL pin name is based on the filename, and those have a fairly small size limit
k, i can use a smaller name when i know the rules of the game :)
>>> h.newpin("a" * 50, hal.HAL_BIT, hal.HAL_OUT)
>>> h.newpin("a" * 90, hal.HAL_BIT, hal.HAL_OUT)
HAL: ERROR: duplicate variable 'x.aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa'
I would have expected an error message printed to the terminal, though
righto, i did see that it was the pin name and didnt make that known to you
is alex_joni on line ?
in emc2.1.1, - steppers-mm I can't move spindle speed override with mouse
you need to set the MIN and MAX for spindle override to be able to do so
alex_joni hi, this is dumitru pop
alex_joni ahere to put min and max ?
03cradek 07TRUNK * 10emc2/src/emc/rs274ngc/interp_convert.cc: fix another sticky-word problem
03cradek 07v2_1_branch * 10emc2/src/emc/rs274ngc/interp_convert.cc: fix another sticky-word problem
Guest354: in your .ini file ( eg: ~/emc2-head/configs/m5i20/m5i20.ini ) you will find MAX_FEED_OVERRIDE = somevalue i see no MIN anywhere
tomp: thank you, I fix it.
don't know how to control spindle speed
which configuration do you use?
I use steppers mm for a milling machine
does the widget move now?
yes, it's ok
does the spindle speed change?
don't have all hardware setup
ok, you can do some checking with halcmd ( show ) and with halmeter to view what the system thinks is happening
i suppose that all is ok but i don't know how to implement spindle speed control in HAL
03cradek 07TRUNK * 10emc2/tests/ccomp/mill-zchanges/ (expected test.ngc test.sh test.tbl test.var): test of sticky Z changes for both entry types
the ini file you chose(and associated hal), has on/off control for spindle, but not speed control ( as far as i can see )
tomp: yes, but i tried some examples grom hal manuals.
so you can use computer controlled on/off with manual velocity for now, and later write code thru hal to allow spindle speed control ( examples... understood, will look )
in EMC user manual, chapter 4.3.5
tomp: i want to understand that example
i have different emc user manual, i see you are in right place, in analog output... but my pdf goes 4.3.3 -> 4.4 so i cannot follow exactly
[16:38:21] <SWPadnos> http://www.linuxcnc.org/docs/EMC2_User_Manual.pdf
the heading is "Adding PWM Spindle Speed Control"
emc2 user manual version from February, 19, 2007
Guest354: they just pointed us to a March 7 2007 version, lets both get that
I think that section is unchanged, but it's a good idea to use the correct one for the installed version of EMC2
in a moment
Guest354: in the AXIS user interface, the spindle override slider will not appear unless the HAL pin motion.spindle-speed-out is connected to something
in tkemc it probably appears all the time
in tkemc the slider became active after I put values for MIN_SPINDEL_OVERRIDE and MAX_SPINDLE_OVERRIDE
Guest354: in what file?
in the ini file
tomp: the .ini file, section [DISPLAY] (I am guessing)
stepper mm, I think
the section 4.3.5 is the same in emc2 user guide February and March versions
ok - that's good :)
Guest354: what kind of speed input does your spindle have? The example in section 4.3.5 is for a spindle which has a PWM speed input or a 0-5V analog speed input
my cvs checkout of 20070211 has stepper_mm.ini and the word override only appears in [DISPLAY], and the only as MAX_FEED_OVERRIDE not for spindle the only reference to spindle in the file is ... 'spindle' never occurs
ok - there's no information in the manual avout those INI settings
tomp: right, that's why Guest354 had to add it
because there is no reference, that's why I asked here
i want to use pwm speed input
input for controller, output from parralel port
(output) the example outputs pwm on pin 9 of the loaded parallel port. did you look at that pin with a meter or scope?
no, i want to use a real scope.
did you add the code to the .hal file ( yes real meter real scope, np)
yes but i don't know if it works
you're using tkemc?
ok - you need to hit F1 (estop reset) then F2 (machine on) to make emc output anything to the spinde
my experimental setup works at that level, i can move all 3 axes
connect real scope to pin 9
ok, so if you issue M3S100 in MDI mode, you get no output?
ok, but i can't do it now
many thanks to the emc2 team for help&time
no problem. come back any time
best of luck
how can I get a user name ?
send a message to nickserv for help: /msg nickserv help
are you using the web client from Linuxcnc.org?
ok. if you have Mozilla or FireFox installed on your machine, you can install the extension "Chatzilla"
it's a little easier than going through the java headache every time, and you can set it up to automatically connect to freenode (and this channel), and to log conversations
ok, i will try that
well - I got emc to boot off of a keychain drive. (was my mistake initially as I forgot to copy a few files that where needed)
took 3 computers before it worked - the first two I tried must have been too new.
3rd times a charm. ran emc2 stepper_inch config no problem
[Global Notice] Hi all. Services are going down for a quick re-start, hold on to your ops and we'll see you on the flipside. Thank you for flying freenode and have a great day!
skunkworks, that's off a USB stick?
great - you dod the work so I don't have to ;)
it's basically like that Ubuntu wiki page?
I had to use lilo before it would actually boot - but the direction on the page are what I used
what fixed it?
hmmm. I didn't read that page - do they mention syslinux or grub?
or did those not work correctly
I had forgotten to coppie a few of the files that needed to be in the root of the usb drive
a bit weird it used casper
syslinux is what I tried initially - but none of my computers would see it as a boot device untul I did lilo
the wiki showed 3 different ways to make it bootable - but only lilo worked.
odd. I'll have to try messing with that soon (in the next 3 weeks or so at least)
this should be persistant also - but I have not tried it yet.
hmmm. I guess I should get another >600M USB stick
but it's still booting the "liveCD" system right? so it can detect video and all that jazz
I installed it on a 1gb dive. I made the primary partion 750mb and the ext3 partion what was left over
ok - that's what I'll need
yes it is,
like I say - I have only gotten 1 computer so far to boot from it. usb/bios seems to be a bit pickyer
hi, thnak you for your help with spindle speed override
sorry I left in the middle of it :)
glad you figured it out
* alex_joni goes to put the stuff into the documentation
can we talk in romanian here ?
better in /msg
SWPadnos: I should say 4 machines have booted off of it - only one has started the gui. (the rest didn't find the video card)
ah - ok
the joys of making 25 of something, repeatively. :(
horray for CNC...
otherwise I would never ever ever ever do it.
03jepler 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: begin documenting items that configure the GUIs
03jepler 07TRUNK * 10emc2/docs/src/gui/axis.lyx: label used from ini_config.lyx
Which would be better for emc2: 600MHz w/512 cache -OR- 1000MHz w/256 cache (both SLOT1) ?
whichever one works better
either is likely to be fine
does that mean PIII? both are fine
I am running on a 600 pIII right now. works good.
So L2 cache means squat?
it's possible that the larger cache would help if you use an exceedingly short BASE_PERIOD
but otherwise it won't change much
use the 600/512
Jymmmm: if you can, run the latency test on both (from the CD) and see which gives the lower latency number
that's 18 uS (nominal), which is 90000 cycles. that gives a better chance of base thread functions remaining in cache
is the cd bootable?
Jymmmm: have you seen jmk's page on stepper setup, including the spreadsheet? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TweakingSoftwareStepGeneration
err - 10800 cycles
Jymmmm, yes, that's why it's called a LiveCD
Jymmmm: yes, there's a bootable/installable emc2 live cd on the linuxcnc.org website
Last I knew, it was based in ubuntu, never knew there was a live cdi after puppy
I have a question - I was running some text yesterday and I thought it was a bit jerkyier than I would hav liked. It would be smooth then would star to jerk in some locations. Could it be starved for segments?
skunkworks: listening to/watching the machine, or just watching the plot?
feeling/watching the machine
skunkworks: does a tight base period make your machine pretty slow while emc is running? that might cause starvation
you're using 2.1.x I assume?
I am running 50us - it will run lower
yes - latest installed emc2
I was going to hook it up to my laptop and see if it ran better - just have not gotten around to it.
this was using g64 strait up.
FWIW... under DOWNLOAD section it doesn't list the LIVECD, except for the sidebar.
err NEWS Sidebar
skunkworks: can you post the gcode somewhere? maybe your ini too?
Jymmmm: You're *supposed* to find it by clicking Download, then Basic Installation
ok - I will tonight.
Jymmmm: maybe you didn't have any idea what Basic Installation was, since it is not explained on the Download page?
(I don't blame you)
jepler: maybe add a "(also known as LiveCD)"
cradek: I will tonight.
ok - jerkiness usually does mean starvation, but it's hard to diagnose without more info, maybe one of us can reproduce it
wait - I think I have it on my keychain drive
where is wget usually?
Jymmmm: /usr/bin/wget on ubuntu
this was mostly x/y movement so 50us base period and input scale of x and y of 2540 and z of 10000
do you remember if there's a certain part of the program that was jerky?
it was on every letter that was curved.
I reamed the 'download' page pretty heavily: http://linuxcnc.org/index.php?option=com_content&task=view&id=2&Itemid=4&lang=en
cradek: it could be me.
what do you see?
I'm watching "sim" cut the Ds and halscope shows the tangential velocity staying at about .333 (F20)
were you cutting at F20?
acc set to 10
for x and y
sim is 20 - I will change it to 10
anyone ever deal with loctite 609?
skunkworks: it looks quite good here - tangential velocity varies between .333 and .32 going around those curves
skunkworks: maybe your TRAJ or SERVO thread is set badly
ok - could just be my slow computer. Will try my portable.
a-l-p-h-a, I have a bottle of it
hmmm - no, mine's 680
cradek: I will check that.
I doubt it's a slow computer - around those curves it's only using 10-15 segments a second, which is not at all hard to keep up with
if you can't spot the trouble, I'd like to see your ini next
Ok - I really don't have that here ;)
cradek: the sing turned out really good though
what did you use to generate the font?
Not your program.. It was some font to g-code utility I had found on the web
SWPadnos, looking for something that's like super glue, but for metal.
I then offset it to do the pocketing. sort of a pain ;)
yeah, mine can't fill like that
oh - I did that manually
609 has like a gap fill of 0.005". But I'm not sure if that's good enough
ahh, that does sound like a pain
it went pretty quick - using acad is like riding a bike.....
I used a #30drill bit, for a 0.125 +/- 0.002 stainless steel 304 shaft... doing a test to see if they fit with each other or not.
yeah, me too
and its offset works well
autocad... :) crayons + ruler for the computer.
wish I bought their stock when it was $5, now it's $50, after 3-4 years
although acad 2000 offsets much better than acad 12. acad 12 would get lost on some of the letters - ended up finishing it in 2000
03petev 07TRUNK * 10emc2/src/libnml/inifile/ (inifile.cc inifile.hh): -Added table driven Find() method to handle enums, bool.
skunkworks, you should try 2006+, it's really nice, especially if you have a relatively new computer and vid card. it's a really nice GUI now
a-l-p-h-a: cool - have not looked at it
autocad's really nice gui = dos, 1996
cradek, I've been using autocad since v10.
in 1994. :(
11 here, maybe 95
never had anything newer than 12 though
I was teaching autocad in highschool, as a peer tutor...
yah - I think 11 here also
hm, now it's time to find out how to do preheat and stuff with sheetcam
in 11 it was a pain to edit polylines because there were no grips
03petev 07TRUNK * 10emc2/src/emc/ini/ (emcIniFile.cc emcIniFile.hh): -New module for handling EMC specific types in INI file.
that's the big improvement I remember
need to wire up the cutting valve tomorrow
got 130 parts to run :)
anonimasu, did you make the plasma cutter?? :)
anonimasu, any hawt CNC porn... or real porn if you must?
d a m n y o u
plasma is overrated for thick stuff..
unless you have a big(bigger then my plasma)
im cutting 20mm at ~400mm/min now..
and no videos. nor photos.
cradek: good afternoon... on thepwmgen module in hal, what determines the total length of the pwm cycle?
15-300mm is the rating of the torch I have :)
uh, so how'd you 'overclock' it?
all hail jlmjvm
I dont, I run oxyfuel..
anonimasu, :) okay... how clean are the edges/
LawrenceG: did you read 'man pwmgen'?
no kerf at all
cradek: I would like to try making a coarse spindle spped control for a trim router using a "hockey puck" SSR
really? cool... did you use the T or H gantry design?
cool. I'm hungry... bbiab.
the trouble is the belts(length)
jepler: yes.... I see the function that is called from a high speed thread,
jepler: that would determine the minimum pulse width
but as I drive both rails it works really well..
LawrenceG: did you read man pwmgen?
LawrenceG: look for 'pwm-freq' in the manual page
LawrenceG: there's a whole mess of options
oh look, jepler gave the same advice above
it must be good advice
a-l-p-h-a: the only problem is the preheat and pierce stuff
cradek: np.... using a triac based SSR means any pulse will hold the SSR on for up to 8ms
a-l-p-h-a: I have no idea how to make sheetcam do it
with no feedback it's not really going to work very well, is it?
cradek: but I really only need 3 or 4 speeds .... trim router has very little load with pcb engraving bits
is there a dwell g-code?
cradek: will give it a try on the bench
anonimasu: yes. G4. quick reference: http://linuxcnc.org/docs/2.1/html/gcode.html
LawrenceG: definitely sounds like it's worth a try
what does 2.1.2 fix?
duerz: are you on the users list? I sent the changelog with the announcement
duerz: there's a list here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Released
no - i dont think im on the users list
can I get an update with a .sh file?
just wish it would fix this freakin vti card
Cool, I can upload iamges to my gps (splash screen, waypoint symbols, etc)
duerz: did you try contacting eric johnson? he is the last developer of hal_vti and from his e-mail messages it sounds like he actually owned one of the cards.
yes - he e-mailed back - but i still cant get it to work.
its closer now
but no cigar
duerz: it takes a developer with access to the hardware, and nobody but him seems to fit that bill.
I have piercing and stuff working
does he got a phone number?
wrong place :D
lots of people looking for people
if he emailed you back, surely that's the way to get his phone number if he wants to give it to you
anonimasu, I don't know sheetcam.
a-l-p-h-a: I just changed the post, a bit
a-l-p-h-a: I'll wire up the valve for preheat/cut tomorrow
anonimasu, photos. :)
a-l-p-h-a: no :D
photos are your friend, as well as ours. :)
a-l-p-h-a: I've got one but I cant get it off my phone
you don't share any of the good cnc porn.
03petev 07TRUNK * 10emc2/src/emc/nml_intf/emc.hh: -Changed EMC axis type defines to an enum.
yeah I found a cable.
I've got a picture
03petev 07TRUNK * 10emc2/src/emc/nml_intf/Submakefile: -Added emc/ini/emcIniFile to libemc sources.
03petev 07TRUNK * 10emc2/src/emc/ini/iniaxis.cc: -Cleaned up access to INI file.
a-l-p-h-a: a action shot :)
but I need to get my mouse living agian
CODE BLUE... Rodent Ward!
just need to grab 162 images off my cellphone
has anybody seen my camel?
No, we haven't seen your wife. <rimshot>
fuck it's already 1300.... this time change always throws me off
[19:46:38] <anonimasu> http://imagebin.org/7613
dust protection is another issue I\ve yet to tackle.
anonimasu wrap the whole thing in plastic sheet =)
the outer edge looks bad because it's re-cut..
how thick is the stock?
ah, I was gonna say 1/2"
got some parts in 40mm comming up soon
but I badly need the preheat stuff working
it's a bad one :/
it's blurry because I trued to hurry to get one in the cut :)
816x612 is that the resolution your camera takes?
I scaled it in paint ;)
what phone is this?
never scale by anything ither than an integer factor, unless you're using professional tools (and not even then) :)
nice cutting though - very cool (or hot)
I need height control to go faster :/
thick metal there
~500mm/min is the speed it should be cutting 20mm at..
how fast are you going now ?
still not bad.. but the closer you get the nozzle the faster you can go
how far off you running now ?, 3mm ?
Dallur: also, it may be a setup issue(how much gas)
"oxygen" sorry :)
need to set up home swithces tomorrow
lerman: around? I have a program that reportedly runs on EMC2.1 but not on cvs trunk, can I send it on to you?
anonimasu: Yeah, somehow they always seem to be the last thing on the agenda although they probably should be the first :D
email to lerman-emc at se-ltd.com
Dallur: I'm never building a cnc without adding them first..
anonimasu, so waht you going to do for height control?
plate rider, or someting electronic?
robin_sz: I've yet to decide..
cradek: where did the program come from?
mmm .. all very novel for oxy
I'm having some issues finding sensors..
anonimasu: inductive will give you about 2% error in repeatability
Dallur: depends on what kind of sensor..
capacative is probably the most common
Dallur: there are very repeatable inductive..
that an ionisation
indictive is unlikely to work
I have analog inductive sensors ;)
that I do positioning with at the machines we build..
lerman: a user sent it to me
but where will you put it?
robin_sz: that's the issue ;)
ultrasonic would work..
capacative is usally doen with a ring, with a 25mm hole or so in it
ultrasonic would fail for the same reasons
yeah but finding one seems hard..
robin_sz: ultrasonic are long range..
mounting them on top of the slide would work..
you going to sense what ? just ahead of the cut?
yeah, or besides..
besdies excellent choice
anonimasu, besides is great ... assuming your machine knows where "besdies" is .. or you only cut straingt lines on one axis
robin_sz: changing the PSU worked wonders
robin_sz: you have no clue about what kind of logic I'd use to control it..
no more hum, not even while on IR
robin_sz: it's not like sheets go +4mm in a few decimeters..
anonimasu, I have seen more cutting machines and more attempts to try and sense height than I have had hot dinners
anonimasu, err, wait till you cut multiple parts and one sticks up a bit ...
anonimasu, or you go to the edge of the sheet
robin_sz: also ultrasonic have fairly wide area..
capacative ring ...
robin_sz: where do you get that kind of sensor?
for a sane price?
alex_joni, you have seen " a few" oxy fuel CNC cutters?
To get the best edge finish, feed slow, rom's high?
anonimasu, less than $10 .. build it yourself
robin_sz: a few, and also a bit of plasma's
alex_joni, explain to anonimasu why sensing is done the way it is and how laser, ultrasonic etc are just a painful way to have no fun
anonimasu, simple oscillator from a couple of TTL invertors at about 10mhz
robin_sz: well, I have no idea how to build that.
well then you are screwed
anonimasu: all those robin_sz mentioned migh work in 80% of the cases
it's the other 20% that will bite you
unless you use more than one sensor
and do some min/max math on the feedbacks
and the 20% wil be the ones that bust the tip off your torch
robin_sz: it's replaceable :D
robin_sz: I dont see why my torch would dip that low, if I cap it..
anonimasu: if you're willing to mount 4-5 ultrasonics around the head, it might just work
alex_joni, spoken liek a true torch salesman :)
robin_sz: _anything_ is replaceable
alex_joni: I'd bet you a ultrasonic would work.
OK, if you cant build electronics ... then the Rutex height sensor is good
while we can talk about it forever..
well, ive seen em tried ...
the sheets are very flat..
anonimasu: indeed.. prove us wrong :D
alex_joni: I'd like a capacitive ring sensor, but finding one seems hard.
[20:11:15] <robin_sz> http://www.rutex.com/us/home/Oxy_Torch_&_Plasma_Control/R982kit.htm
use the rutex one
or build one for $10 .. your call
robin_sz: well, if I can find a schematic it wouldnt be a problem
robin_sz: filtering out HF might be an issue
for someone that has never tried it before
anonimasu, on oxy?
alex_joni, on oxy
sorry, that was for alex
alex_joni, no HF on oxy :)
cradek: I'm having trouble just starting emc2 -- what did I screw up?
err - TRUNK?
anonimasu, just buy this: http://www.rutex.com/us/home/Oxy_Torch_&_Plasma_Control/R982pcb.htm
I getr Application initialization failed ... this isn't a TK application...
robin_sz: yeah, that's a good idea..
robin_sz: or maybe I can zoom more and figure out what components they use ;)
its not exactly difficult to brew something up on your own ... we built one for the CNC laser project
ours was a bit OTT though
over the top
I could cook one togther with enough time..
basically, you make an oscilator
RC oscillator froma TTL invertor?
connect the ring sensor to the input ...
as it goes nearer/further from sheet, the frequency changes
convert freq to voltage through some means, job done.
I think there are ttl chips that do it directly
ah well, at 99$ I can afford to buy one, time spent building one costs more then buying the damn thing..
that's the spirit
alex_joni: well, it's true, normally I'd look more at it..
I've got about 2 months with 6 hours of overtime a day.. comming up now
I dont have time :)
it looks like that torch height control would work with the gates of an h bridge instead of 2 relays
cradek: are you around?
lerman: I thought you had that problem before, and found it had something to do with your rgb.txt
seems familiar, but I don't remember the cause or solution.
using vnc or something else unusual?
robin_sz: what's inside the thc ring?
I'm trying to run over vnc. It used to run. Oh well -- I'll go upstairs to that computer and see if it runs there.
anonimasu: small people looking at the material
I bet your vnc has bitrotted
I'll give you a pic..
[20:38:19] <anonimasu> http://www.rutex.com/pdf/R982.pdf
or is that between the shield and the sensor?
lerman: lots of similar reports on google
more that just vnc. My xterm is 640x480 and doesn't seem to want to let me set other preferences.
I believe I rebooted since the last time I ran emc (power failure).
for the heck of it, I'm rebooting again.
(not that it should matter, but what the heck)
rebooting solved the problem (I can now run emc).
hmm, that's always nice
rebooting to fix random problems
well.. running emc is nice.. but rebooting to fix stuff.. is .. ^&^*&^*^*#
gosh - that sounded like a microsoft product.. ;)
oh man - I am a very happy camper right now
lerman: that file also does not load for me here, I just tried it
even though I'm not camping
03alex_joni 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: there is a MIN_SPINDLE_OVERRIDE too
some (free) update to Altium fixed import of very old Protel files - I can now directly import all my old designs
yeah. I didn't even see it in the changelogs, but I tried loading a file today, and it worked :)
a file from 2001
that's an old one?
it's the oldest of mine, in my current consulting business ;)
the software predates Protel 99SE though, it's from Advanced PCB 2.75 or so
some say I am
I'm not so sure though
* alex_joni waits for the red pill
do commerical oxyfuel machine pierce?
and if so, does it kill the torch?
anonimasu: they usually do some fancy dance before the pierce
helix move around the spot to heat the material up
then jumping up and down during pierce
is there temperature feedback?
tomp: not on the ones I saw
tomp: just preprogrammed.. based on material &thickness I guess
i like 'fancy dance' :)
alex_joni: I guess the fancy dance isnt that essential if not cutting very thick..
alex_joni: I pierced 20mm manually in less then 0.5 sec with a bit of preheat..
anonimasu, the ring is just .. a ring
anonimasu: yeah.. it's up to 30-50mm for piercing
anonimasu, its noting more than a capacative plate ...
yeah but the resistor in the drawing?
the rutex pdf..
although I've seen thicker getting pierced.. but that usually kills the torch slowly
alex_joni: yeah figures
the rutex pdf resistor? unit, 2 relays , motor...
I'm just going to add a spray cap to it to keep molten metal from flying all around..
the resistor in the part with the ring..
robin_sz: did you read my question about piercing too?
ahh, looks like from the shielding to the ring frame
i was looking at the 'schematic'
that's to avoid the shielding from being detected?
anonimasu, the resistor is not so critical .. it would work OK without one I suspect
robin_sz: just making sure, that I can just order the kit and build the ring..
as for piercing ...
I personally wouldnt recommend it
anonimasu: the us rutex dealer is tom ???? nice guy, should be able to help ( oh you in the states? )
tomp: I'd had all stuff already..
if I were
anonimasu, avoid piercings if you can
ears are OK I guess ...
but noses, lips, genitalia etc are not attractive
robin_sz: piercing workpices ;)
oh THAT ;)
pierced workpieces aren't attractive either
like alex said, on thin sheet, like 6mm its easy, lower, wait, pierce , wait, cut.
robin_sz: so how do you do it? always cut from a edge..
anonimasu: you can't always do that
alex_joni: I know
e.g. cut out some circles from a part
alex_joni: that's why im asking :)
on 20mm its more tricky .. circular preheat path, then raise to a safe height ... begin lowerign at a feed rate ...
all sorts of strategies
anonimasu, this is part of the start of the story where me and EMC fell out :)
sheetcam will do that..
well.. now at least
IMHO it shouldnt
not circular preheat..
I guess I could do that as a macro..
IMHO it should be something like :
03alex_joni 07TRUNK * 10emc2/docs/src/config/ini_config.lyx: chris says it's usefull to write useful
Mxx P 22.0
that's hard to do :)
(set material thickness to 22mm)
G1 xblah y blah
alex_joni: can you assign g code a M code?
the "M3" should handle the pierce, including the torch raise, the height set, the circular pre-heat et etc etc
but, thats just my opinion :)
alex_joni: assign macros "G" functions..
anonimasu: not atm
although... maybe :)
robin_sz, wouldn't the procedure depend on the material, not the machine?
in *some* controllers, you can assign a load of stuff to a Mxx macro
SWPadnos, it would depend on both material and machine
in which case, a single M3 code wouldn't really be possible
alex_joni: acutally I'd be happy to call a file that does the circular pierce..
anonimasu: you can call a procedure in emc2
SWPadnos, well, are you sure about that?
alex_joni: is it in the wiki?
i was thinking the torch height control could provide a way to follow a surface and polish it. (who want a robo-back massage?? )
anonimasu: not the pierce
but the call-sub
SWPadnos, because I can tell you thats how almost every single plasma and oxy fuel control I have ever seen does it
alex_joni: oh, I dont need the pierce.. just the stuff to do it..
o-words or smthg like that
alex_joni: can you call files with parameters like height?
or set parameters before you call one?
robin_sz: and your problem was?
though it'll be a bit of a post hack :)
SWPadnos, of course you usually have to select a material thickness/type from a database of paremeters pre-set for each material thickness
technology data base
anonimasu, its not easily acheivable in emc, no wait, it was nto easiliy acheivabel in emc at the time I tried it
trying to do that now with edm tables
robin_sz: I guess I'll have it working tomorrow..
I tried to hijack the tool table ... using diamter, length etc as params for the piercing ..
it soon became a mess
yeah I can see why
now how to get a subroutine call into my code..
anonimasu, this is when I discovered my hatrred for the way emc passed messages :)
alex_joni: does g04 do millisecond pauses?
anonimasu, its a lot better now though ...loads better
anonimasu: I think so
one problem with doing piercing like this is getting the delay time right..
better slightly too long than too short :)
I figured I'd do a quarter arc.. then pause for 0.1s, and another one..
feedback from an ir thermometer gun strapped to head, reading surface temp, fed into hal?
mayber smaller ones..
tomp: let's not go there ;)
anonimasu: just use a slower feed
I have PLC>hal stuff..
alex_joni: hm I guess I can calc feed and timing..
straight delays are fine ...
it just needs experience to determine the correct delays
im not sure I need circular preheat.
not at what im cutting :)
20mm .. borderline
I'll be doing 40mm later..
but I can avoid piercing..
hm, the nozzle is from 15-300
robin_sz: I did 80
alex_joni, stand well back!
here's the spec..
300mm at 100-70mm/min
I've seen photos of thicker than that being cut with a "standard" torch ...
still that's not a bad speed..
and what speed in 10mm?
im not concerned about speed, if I can keep the machine automated..
bout the same as my laser then
that's for 10-15..
dont compare apples and pears.
apples and rotted pears
still pretty impressive for fairly crude technology
though if we talk about the $ put into it..
what tolerance you reckon you get?
the parts I've made several off are identical..
right .. good enough then
I didnt use a caliper though but when I lay them on top of eachother they match
and with the drawing too.
good enough :)
we used to cut stuff by hand before
so way better than sending it out to be done then?
as in cheaper, quicker
I dont know..
at least you know tha machine isn't booked
also, it's a work issue..
the welder at work has better things to do then cut out 40 of something and 40 or something other
like play spider solitaire
heh like weld stuff
or well cut one off stuff..
is anyone awake with a axis working?
as in XYZ_A_ ?
as in the gui..
I have one around
now for the fun stuff how to make a full arc..
wot's the question
I dont know I and J :/
alex_joni: I need someone that can try this if it works..
/is legal g-code
"/" is not legal G-code
it is legal, however, at the beginning of a lilne of legal G-code, and indicates that the line (block) is to be ignored, if the "optional block delete switch" is active
yeah.. I know
but that werent what I were wondering about
I'm still waiting on what to try or inspect ...
one thing to note: I and J are always relative to the current position
SWPadnos: I need to make a arc first :/
that's what I want..
* alex_joni gave up waiting
do you have to use IJ, or is R usable?
I can use either, as long as it's from where I'm at
but I need to specify my coordinates right?
of course, or EMC doesn't know where to go to
that's harder then..
anonimasu: the start is where you are.. you specify the end, and either radius or center
so if I do "g01 x-10" I can do "g02 r10" ;)
no, because there's no endpoint
that's like G1 <nothing>
hm, beats me how to do a circle wherever you are at..
a full circle, or parts of a circle
maybe I should just go to incremental and wiggle around a bit..
Tell me where you are?
i can be anywhere ;)
that's the problem with the subroutine..
yea right always a smart ass around.
as for location im in sweden..
I think an IJk arc can do full circle, you just respecify the current point as the endpoint
you are trying to make a circle in a subroutine.
I'd give the subroutine the center and radius of the circle
#1=x #2=y #3=R
the subroutine is then something like this:
I and j can cause a full circle if you use current position for endpoint.
G1 X[#1-#2] Y#2
G2 X[#1-#3] J#3
oops - typo on the first line
G1 X[#1-#3] Y#2
G2 X[#1-#3] J#3
I guess I can make the post do that..
G1 X[#1-#3] Y#2
G2 X[#1-#3] I#3
that should be all you need
I think this conversation will definitely make it into the bash.org top 200 in no time
cradek: are you here?
it moves to X-R,Y does a circle back to the same spot, and uses R as the center offset ofthe circle
cradek: I have a fix. :-)
[22:16:27] <jepler> http://pastebin.ca/392544
unter1: which one?
very slightly more readable, IMO: http://pastebin.ca/392549
hm, now lets see..
I always fancied doing it directly from an M3 ...
robin_sz: well, you can always hack that into the interpreter yourself.
you could code it in a HAL thing, but, apparently, HAL can't do motion
lerman: cool, what was it?
anonimasu, well, I ended up using Mach2 :(
the only thing left is to calc the feedrate based on the delay..
robin_sz, HAL does all motion, it's just a question of where the commands come from ;)
SWPadnos, ah, ok
The first time I hit a DO, I add the o-word to the symbol table (of o-words). When I loop back, I shouldn't add it again.
SWPadnos, I think the problem was with the interpreter not knowingthta the position had changed
robin_sz, sure - that's one problem. there are ways around it, but they're not very easy to use
recall the discussion about "weaving" some time ago
03lerman 07TRUNK * 10emc2/src/emc/rs274ngc/interp_o_word.cc:
Add an o-word to the symbol table on ONLY THE FIRST time it is hit; not each
time through the loop.
cradek: did you see my ini file?
hm, it'd be nice if you could just call up a file as macero with a parameter..
SWPadnos, the easy way would be for the M3 thing to inject G1 xxxx intot the interpreter directly
The trick to debugging this stuff is to turn on debugging. It took me a while to igure out that I had put that stuff into the .ini file.
but that's probably doable..
SWPadnos, so long as it was careful not to inject an M3 it should be fine :)
lerman: do you want to let him know it's fixed, or do you want me to?
lerman: thanks for fixing it
skunkworks: not yet, I'm away from all my emc machines now
You can let him know (please). But it now gives an error about starting and ending radii being different.
cradek: no problem
anonimasu, wait until you try it with Plasma ... much trickier
It goes with the job description... Fixes all bugs in code that he wrote.
I will probably be gone until later
lerman: I won't tell him it's ok yet then :-/
I've got to head off to class in a little while. I'm just glad I got it fixed before I left for class.
yes thank you
lerman: I'll try to figure out the arc thing (since it runs in 2.1)
It's probably a user error.
ooops. Did you say that the error doesn't occur in 2.1?
robin_sz: well, it's not like it's a problem still
It could still be my error, then. Code that uses parameters could be broken by me.
anonimasu, piercing with plasma is tricker, because you have to wait for a "arc detected" before you start the timer, if it fials, then retract the head
i think that would still be a problem
lerman: maybe I can determine that if so
robin_sz: here's how I'd do it, I'd throw a command to my plc to strike a arc.. and have it pause emc, until there's an arc..
lets say you didnt get an arc ...
within hte 5seconds allowed
as I can do both hal and nml from it I can back up to the line before the pierce and retry it..
use optional block delete on the rest of the "macro", and the PLC sets the optional block delete output on error
this is where 2.5d systems meet 3d systems
robin_sz: so tell me where the problem is.. :)
Darn that is a pretty limited subroutine for running circles.
hm, I wonder how to time the delays..
anonimasu, the desired bhaviour is: raise head to safe height, stop machine, inform operator
robin_sz: well, I dont care what the desired behaviour is, I could do either..
if the operator presse cycle start again .. then try again
skunkworks: maybe try TRAJ_PERIOD of 3000000, every ten servo cycles seems not very often
ok, then thats great .. emc has indeed come a long way
it was a pita when i tried to do it
also reduce [TASK]CYCLE_TIME to 0.001 (not sure if that's still important but it can't hurt)
so, that wonderful .. things have come alonmg a long way
sure it'd be a bit of work, but running hal and nml from the plc took a few days of work(about 3 hours once I learned there was python modules for hal and nml) ;)
the last time I checked, hal couldnt command things like g0 z 100 to raise the head
then 10 minutes for the plc code..
robin_sz: hal still can't .. but nml can easily
yeah, but you can do mdi via nml ;)
and you can hack it in hal too
skunkworks: don't thank me unless it helps :-)
cradek: what's the dpi stuff in xorg.conf ? to make fonts bigger?
if it doesn't, maybe you can make some scope plots to help debug it
right ... and you can fir off any nml message you want from a bit of hal code?
hm, im very confused about this..
no, from the stuff that talks to my plc..
robin_sz, no, you fire off NML code from a python script
but I guess you could build a python program that takes a nml input..
and fires off a nml message..
in 5 lines :)
alex_joni: I don't know about those interactions, but setting everything to 75 seems to help legacy programs
so hal calls a python script,
python script fies off nml ...
python script is running, and connected to hal..
robin_sz, the scripting available in EMC is waaaaay more powerful than in Mach, but integration isn't as good (from a "normal user" point of view)
when stuff happens or dont, you fire off the pin..
from cl or so..
cradek: I meant the actual entry.. is it called DPI ?
I know you told me this once..
SWPadnos, right ...
* alex_joni has issues with a large screens, and small fonts
I think you have to specify the screen dimensions, and it calculates the dpi
I wonder if this works at all..
in gnome at least, you can change all the truetype sizes under preferences/font/advanced or whatever it is
try that first...
some drivers have a DPI setting (NVidia is one of them, don't know about nv)
option "DPI" "200x200"
SWPadnos, so, with a bit of gui goodness we are in a winning position then, the core work is done, now its just config tools? right?
it's set at 96
robin_sz, for the most part, yes - I think so
cradek: thanks, that worked wonders
SWPadnos, certianly the vcp stuff looks very promising
there are other options for customization and configuration, which would be more like some commercial controls rather than Mach
SWPadnos, its good to know the scriptign side is equally powerful
cradek: changing the number under advanced
to that it's tools that missing :)
still, I lack much understanding how to do stuff.
cradek: hmm.. not for all windows though :(
SWPadnos, excellent .. the point is .. it is doable as a config thing rahter than a recompilingthing, so thats a big step forward
[22:44:20] <anonimasu> http://www.pastebin.ca/392588
missed a \n at the subroutine
the interfaces that Axis uses to communicate with emc are available to other python scripts, including NML, interpreter, and HAL interfaces
though that's userspace, not RT (like any stupid VB stuff would be)
new realtime parts can also be written with a minimum of fuss
jepler: what language?
though you do need to compile those
SWPadnos, so, how far away are we from being able to have gui tool that lets you drag and drop widgets onto a canvas, save the config and then run an axis-like display with those widgets?
anonimasu: similar to C, but you only write a couple of lines for a module
robin_sz: as soon as you write it
robin_sz, dunno - how are you coming along on the GUI? ;)
alex_joni: can you look at that file?
anonimasu: 'comp' is a preprocessor that helps you write HAL realtime modules. It is basically C with extra declarations at the start. documentation: http://linuxcnc.org/docs/2.1/html/hal/comp/
SWPadnos, so the rest of it is all there, the running from an XML config?
jepler: it looks like im being force fed with emc-documentation today and it's good im learning lots of new stuff :D
robin_sz, no, is XML somehow easier than plain text?
robin_sz: pyvcp uses xml to store the layout of the control panel, but right now you have to hand-code it
SWPadnos, not really, I dont really care what format the config is in, xml would be a logical choice, but it doenst matter so long as it works
it positions things based on "pack" and "grid", which are very different from a drag&drop user interface editor
jepler, thats where i thought the xml came in ...
so pyvcp and axis use different config formats?
right now the config file is a directory, with files in several formats (unfortnately)
hm pyvcp is the panel stuff that runs besides axis..
axis doesn't have a config format, other than some ini settings
but you can hang a pyvcp panel on the side of the axis window: http://axis.unpy.net/files/01167748606/lathe_sim_pyvcp.png
so the position of buttons and widgets in axis is determined by?
the layout of the axis window is fixed, and will remain so.
robin_sz: jepler and cradek
and pyvcp and axis use the same widgets?
can you change the Mach3 screen layout, or just create new layouts that look largely similar?
SWPadnos, create any one you want
the core code will crash in just the same way though
I know you can create any one you want, I'm asking if you can *change* the one(s) you get with Mach
ok - thanks
robin_sz: not the same widgets
is that truly important, as a separate function from making yuor own layouts?
but there are plenty already for pyvcp
SWPadnos, is what truly important?
is the ability to change the actual file they provide you with important, or is the important part that you can create new screen files?
I'm thinking that with enough man-hours, you could use the Mach screens on emc. I just say this for the amusement value
unter1, go home :)
if someone submits a patch that turns the AXIS backplot and program listing into pyvcp widgets, it would be accepted. but it is a waste of my time to give the user the choice of where to put the "clear backplot" toolbar button.
SWPadnos, say you want to have a screen almost exactly liek the one you have now, but say, less on control, or with an extra control .. you should be able to start with the standard one and modify it ... or build from scratch ..
you can do the adding with axis right now, but no moving or deleting of controls
jepler, sure, but say anonimasu wants a "piercing wobble diameter" control instead of spindle overide?
robin_sz: he can put it on the right-hand pyvcp
that'd go into the pyvcp pane..
he can add one with pyvcp, and live with the fact that spindle override is still there
robin_sz: if he doesn't connect the 'spindle-speed-override' pin, the spindle override slider will disappear
or I could just change axis to that..
jepler, and to get rid of the spindle control?
remove the control in axis..
jepler, ahh, right
robin_sz: most of the stuff on the left-hand side disappears when the associated hal pin is unconnected
anonimasu: yes, that was a new feature in emc 2.1
heh - auto-cleaning UI ...
robin_sz: I think you are too concerned about customization..
robin_sz: integrators are different from end users anyway..
anonimasu, not really - that is one of the big things people tout for Mach
robin_sz: I'm not that concerned with ever changing my ui if it does what I want it to..
SWPadnos: x86_64 edgy download completed, I'll see if I can figure out your edgy-sim compile problems
shrug, customisation is the most important thing,
people do seem to love the fact they can make their own front panel on Mach
robin_sz: so you mean you change your mach setup once you have a working system?
I care about making parts..
anonimasu, well, mach doesnt actually work well enough to be useable for me .. but apart from that ...
some people are happy enough to have "spindle on" meaning plasma on, other integrators would not be
robin_sz: so, tell me about this customization stuff.. do you change your mach frontend just for kicks all the time?
i dont use mach
no but if you were..
I have no working isntallation of mach
I have noi intenttion of installing it ever again
but ... on a cnc control, as an integrator
I would expect to be able to configure the display in some way
for example, I would not expect to ship a mill with a "torch height" button on the display
neither would I expect to ship a plasma with a "coolant on" button
robin_sz: well you could do either..
removing the widget in axis isnt that hard either..
but as they dissapear automagically it's even better..
you can even change the text on the spindle speed to height and connect it to another pin..
an integrator is also in the position to write a new GUI from scratch, or edit any of the existing GUIs
it would be nice if the integrator edited a config for the GUI, like in pyvcp, rather than the GUI itself, so was able to keep up to date with new releases without having to start from sctratch
the integrator had better be smarter than "click and drool" -- otherwise I'd be scared to use his machine
jepler: I agree wholeheartedly
since these files are just source code, CVS will do a lot to help the integrator get updates and make sure they don't stomp on the same parts of the source as his customizations
there's nothing to say that it has to be dead esay.. to be a integrator.
not that I would turn down more "click and drool" stuff, because most of our users seem to be end-users, not integrators -- but I also don't have the spare time to write them, nor the vision for how they should work
I guess in the end it's how good the parts end up now how it looks while making them and the cost/functionality
users/integrators dont like hearing that stuff isnt possible.
dont get me wrong ... emc has come along light-years recently ...
robin_sz: well, something that dosent make parts is scrap, nice gui or no gui.. :)
not that long ago, most of this stuff was on the "never happen" list ...
and, here it is already
I wouldnt like to bet on what we will have in 12 months time ...
actually, i really shoudl investigate pyvcp more when I get a moment
yeah - stop wasting time on IRC ;)
its relaxing :)
you'll find that it's not as well-documented as you might like :-P
while Im doing this, I aint worrying about work
sad, but true
robin_sz: I know exactly how you feel
although I guess I still try to avoid that much responsability throughout the day :)
what was 'sim.var' 's intended use?
lucky you :)
(or any .var )
tomp: hold variables
alex_joni, the great thing about having your own business is you can work the hours you want ....
and how did we access those variables?
some interpreter internal variables that need to get remembered
alex_joni, the bad news is its 24x7
well.. I came home at 19:something today
in g code? like G1 X#5273 ?
alex_joni, I try hard to do !work int he evenings .. for my sanity
alex_joni, annoying you lot is a hobby for me :)
tomp: yeah, there too
hmmm ... speaking of which, where si Jymmmm ?
tomp: but there are other uses for that
tomp: most are described in the docs
thanks, i wonder if they can be written to by gcode of an m func ( Mxxx Pindex Qvalue )
tomp: no, the .var file is not re-read during execution of a program
could they be written to a hal comp?
i dont need to write to the var file, just random access reads . i can write when not running
tomp: the file is also not updated during the run of a program