anyone using bitchx around here?
irc client? Never used it - sorry
any other text mode client?
nope. text - as in dos or terminal?
just what comes with the bdi install and mirc
emc is now known as alex_joni
great help huh
bitchx is no good?
trying to sort it out
there's others.. irssi for instance
just installed it
irssi won't work for me
bitchx is more advanced than most people need
ircii is what i meant http://www.irchelp.org/irchelp/ircii/
irssi does work
heh.. what a fscked up error
I tried connecting before, and it timeout.. or I thought so
hey thats not too hard to learn eh
just to configure auto-channels & such
seems to work here too
irssi I mean
[Global Notice] Good morning all....please take a look at the news page when you get a chance. Happy holidays. :) http://freenode.net/news.shtml
heh i wonder if their webserver crashes every time they say that
wow this is pretty funny.. like 50 people joining #freenode-newyears a second
too much ;)
juve is now known as alex_joni
juve is now known as alex_joni
just came in to check email
if you want noise :)
join that channel for a ride
it's plateau'ed at 150 people
ok i'll join
in for a ride
I can't read that fast haha
looks fun, but I don't like big crowds
alex_joni will be the first in 2006, right?
not the first.. the first from here ;)
I'll greet you from the future
newyear is a madhouse!
les_w: there was another channel set up for debug a while back
when Level 3 was down
that was crazy too
will emc ever get that fast?
sure it will
let's type real fast
all we have to do is say "hey everyone in freenode-newyears come over to #emc"
I can't type at all...as you all know too well
using a lot of clients
fenn: let's not do that
les_w: try the braindead approach
type one word per line
feel a lot
* alex_joni stops now :(
I do miss parties.
I am out so far in the woods...
* alex_joni has an idea..
let's go all to les over for a party ;)
YEAH. Bring girls.
how many gigs?
oh god it's great.. i broke out the megahal bot in there
admins didnt like it
you closed it
I forget what that does
megahal_newyears jbit: That's me in the compile warnings from including math.h anyway, that's really a megahal bot!
other than irritate people?
cradek: that's about it
someone help me with some stupid geometry?
I know that if I'm traveling around a circle the accel vector is toward the center
if I'm accelerating along a circular path, what happens to the accel vector?
* ajoni looks away..
I missed that class ;)
it seems like it would be the sum of the tangential acceleration (the rate at which I'm speeding up along the circular path) and the radial acceleration that's dependant on the instantaneous velocity along the circular path
is my gut right?
I'm sure I can take the derivative of something or other to get the answer...
as I said.. missed that class
les_w where are you??
none of the many alexes know the answer
cradek: sounds good to me
just heard a nice thing in newyears
the limit is up to 32bit of people
that sounds sooo nice ,)
and the next question was.. signed or unsinged :D
wonder how signed people look
I am on skype designing a heat sink for yuga
maybe men are signed, women are unsigned
let me read back
cradek: they want to slashdot the channel ;)
ajoni: it would be interesting to see if they can handle it
they said they probably could
as they have lots of spare bandwith
ok cradek, read back. the instantaneous accel vector is the sum of centripedal (pointing toward the current center of radius of curvature and
so the instantaneous accel vector can point anty direction
so if I'm accelerating past 1ips at 1ips^2 on a radius 1 circle, my accel is sqrt(2) at 45 degrees in front of the center?
I mean if I'm going clockwise, it's 45 degrees to my right
let me think
ok add vectors
yes and no, since it's not in physical space
centripedal is v^2/r pointing toward center
vector sum them
an accel vector isn't made o the same "stuff" as a circle
fenn: I know what a vector is...
but it's still a vector
with a direction
we are talking of a special case where the tangential velocity is changing
sorry.. guess that should have been "no and yes"
les_w: it's not really a special case though
it's a general case
general case is a 6x6 matrix
or some quaternions
come on, we just have some circles here
* cradek draws some circles and arrows
cradek: what are you doing anyway?
fenn: umm, I don't want to say
i'll show you mine if you show me yours
I'm rewriting the TP but I'm not promising anything
mostly I'm in the "gutting it" phase now
but in g code tangential accelerating arcs can't be coded except for helixes
les what are you talking about?
accel is usually constant and pointing toward the center
Geez am I the only one that took Dynamics here?
les_w: of course there's tangential acceleration
any arc with a stop at one or two ends will be accelerating
les_w: i never took dynamics either
didnt know there was such a class
uh, I didn't either
Dynamics is a famous "weedout" course in engineering school
it can be hard.
I was a CE so got to skip it
but I had some nightmarish EE classes
Drop rate at my school was 60%.
well at least it makes sense
That's what makes business school applicants.
the "weed out" class for biology was organic chemistry
but what does industrial chemical manufacturing have to do with biology?
I know my alkenes and alkanes
chris what was your hardest course?
meanwhile, i never got to take biochemistry.. oh well
les_w: ummmm either the relativistic physics one or the field theory one (maxell's equations)
les_w: it may have been the professors as much as the topics
les_w: those are the ones I remember working really hard on.
a bad prof makes things tough yeah
I had two courses in modern physics
the field theory class was probably the worst. He was a complete jerk (and he wrote his own textbook which is NOT a good sign)
no problem...but good prof
I was good at classical field theory though. Aced all the courses.
Green's theorem and all that
in that physics class, the average scores on tests were around 30%
that's not all bad though - after a while you learn you only need to do two or three questions right
maxwell's? that course?
never mind all those other pages
no, the physics
I aced em
much easier thandynamics
I bet the tests were a little different then... nobody aced these (or got 50%)
I never took Dynamics (EE), but my dad has taught it for years and years at UIC ...
I just ask him...
well we had to derive special relativity on a test from no more information than "there is no aether"
anyway thanks on my dumb vector question
As an aero I had to take extra dynamics.
for obvious reasons
took all the EE core too.
les here's a dynamics question
does the horsepower dissipated in the work, tool, and chip decrease or increase as the machining speed increases?
It changes dramatically with increased speed
at high speeds the energy dissipates in the chips
that seems more than a little obvious to me...
so the chips become coolant
no i mean the total energy dissipated
um -"energy dissipated in the work" or "energy used to separate the chips from the block"? :)
at lower speeds the tool dissipated relativiely more power
I think the question is about cutting load (ie, spindle horsepower required to remove X volume of stock)
no score yet
That s why people are all so hyped about HSM (High Speed Machining) .... tools last long, chips fly, productivy soars...but typically one is operating between column resonances..
SWP strangly the hot chips remove most of the energy from the system with HSM
excluding the power dissipated in the spindle due to friction
the whole thing is highly non linear
like most real things
well - it's linear if you neglect everything
On my little MF-70, it ran in HSM mode, even though the slew rate was slow....Ran between the first and second column resonance the best...
the true formula is left as an exercise for the reader ;)
SWPadnos: lol, but if you neglect everything it surely isn't linear
it gets into mechanical impedance matching
at most a dot
first assume a spherical les_w
a really small line
Hey, EM is actually linear...
a spherical les_w ?
but then the concept of impedance is a linear operator
* ajoni pictures..
of uniform density
but we use it for non linear stuff too
but he won't fit in his A6
shame on us
les_w don't turn spherical
anyway - do you know if the amount of energy required to remove a cubic inch of material changes significantly when using HSM vs. "normal" milling speeds?
wb9 EM? electromagnetism? linear?
as long as you live in non-euclidian space, it's linear
EMC - C = EM
Unless you ionize something...
beg to differ wb(. Google "group velocity" and "phase velocity"
classical EM is linear of course.
as long as you don't mind multidimensional lines that aren't straight, it's linear
and if you look at a smith chart, and say "look at all the pretty straight lines", then em is definitely linear
for the rest of us - NOT!
Was more thinking along the lines of stronger currents, stronger fields linear...
oh - that stuff ;)
like "more mass = heavier"
Well in school we usually conformal mapped field stuff so things were straight
Z transforms and all that
Unlike mechanical stuff, stronger forces , stronger strain nonlinearly...
Typical real world EM stuff is linear over 130 dB or so...
Wb9, in classical electrodynamics, yeah.
I look at mechanical things and see quantum effects all the time though.
Otherwise your cell phone would have to be completly different...
remember the yodelling example I gave?
As most cell phones are full duplex and tx and rx on the same antenna...
but they're using different frequencies for tx and rx
non linear stress/strain on vocal chords. resonant frequency suddenly "pops" from one energy level to another.
machine tools do that too. A lot!
If the antenna was nonlinear the way a beam is, you would not be able to hear the RX signal at the same time you transmit...
well - I may not get how a cell phone demonstrates linearity in EM
And the Beam nonlinearity is the key to milling machine problem...
that's superposition, not linearity
how come we have to talk one at a time on walkie talkie's then?
you also wouldn't be able to see without superposition
fenn: if you cary 2 you don't
on different channels (like a cell phone ;) )
My best example of the group velocity concept in nonlinear wave propagation is throwing a rock in a still pond
watch the waves
Walkie talkies are on the same freq....Cell phone has a rx and tx on different frequencies, with fancy EM stuff to allow the tx and rx to operate independantly...
Try to listen to a third walkie talkie on a different channel while the second is transmitting the next to the first ? Wont work...
depends on the walkie-talkie
the newer digital talkies that people are so fond of aren't actually on difrerent channels
I gave a moto walkie talkie to my 98 yo grandma down the street but she can't figure it out
they just use a different code for digital transmissions
a CB, on the other hand, is a set of discrete carrier frequencies
That s called CDMA (Code Division Multiple Access) ... At a distance, all the different transmitters can be heard independantly on their different codes...
(I'm still not sure how this relates to linearity though - I may just be brain-dead from too much accounting work)
But even CDMA has duplexors in the cell phone...
The only radio stuff I deal with is aircraft voice, VOR, and radar transponder when I am flying.
pretty simple stuff I guess
And the tx and rx are on seperate bands....within the bandwidth of the antenna...
and NPR, of course
sure - they use a 45MHz separation, so the receive local oscillator is the same frequency as the transmitter
oh - and it's "Hi Hi Hi" - Paul McCartney and Wings :)
Well, I have to do a lot of acoustics in my consulting work. It's kind of like RF but a much more complicated non linear propagation mode.
subharmonics, shock waves, and the like
yeah - you probably have exponentials in sound propagation
You could never make an ultrasonic cell phone with the same range and performance as an EM cell phone...even if the ultrasound propagated as well...
similar to thermals
Have a lot of Bessel functions right now.
Having to design a PZT resonator.
I have a DVD player - does that count? ;)
The pressure wave heats the air distorting the speed ...
if I could remember what a Bessel function was, I'd probably explode
but my mother would love to hear about it
(the Bessel functions, not the explosion)
Different temps, different speeds....That s why exhaust systems on cars cannot be tuned well (as well as a cell phone duplexor) ...
Well, can't say much, but my acoustics research has sound that will burn skin badly.
what frequency are you talking about?
ok - low freq ;)
at 10^5 pascal overpressures
Bessel functions are sins of sins if I remeber right...in EE they relate a FM modulation to its RF spectrum...
or are they gaussians superimposed on sines or the like?
They also define how circular discs move in mechanics
Are there not several different types ?
they have order and kind
are you an IEEE member?
That s the analogy to FM...the disks...
like bessel function of first order of the second kind
acoustics and vibration of machines use a lot of classical DE solutions like Legendre polynomials, Bessel functions, Chebychev, etc.
In SAW class the professor told us to be careful if we ever played with PZT ....There are stories of sailors getting killed doing maintanance on them, where they had to
Ah yes - I think my mother mentioned Legendre and acoustics in the same sentence
Also used of course for active filters and such
heat them up to reallign the molecules,,,then did not discharge after cooling...
if you can download IEEE documents, do a search for TAST and Sachs (the Thermo Acoustic Sensing Technique, Tom Sachs)
Yes that is poling...done in hot silicone oil
I have hired Morgan electro ceramics to build devices for me
as I recall, the perturbation beams were around 1 MHz
They invented PZT.
I am getting good funding support.
cool - send a few boards my way - help me pay for Altium and the new scope ;)
Above --- "them" = sonar emitters ...
first can I buy an audi?
yes - go ahead
When I parked at J and L today, there was a Prius next to me, then 3 others on the way home...they are calling to me...hi...
really...I must start driving a whole bunch. I need a comfy old man car.
I'll be back and forth to Morgan I'll bet a lot
Where is Morgan ?
I am having them mix up a special "snake snot" hard low permittivity PZT
What you doing with PZT , making a active milling machine column ?
Aw saying too much. Law dawgs say I cannot talk.
but yes I mess with active structures. Some.
Well,,,its going to happen...with all those active mirrors on big telescopes...same sorta thing...
just get freenode to sign an NDA ;)
yeah. We had a big NDA fight.
I assume you're allowed to get subcontractors as long as they sign the appropriate papers (probably provided by the higher-ups and their law-dogs)
SWPadnos: we're just trying to convince lilo to log all channels :D
on freenode and publish them
Finally had to agree to disagree and turn the lawyers off so we could get some work done.
lawyer prob monitor IRC logs haha
I'm sure NSA does!
Well, I know the spamers do.....got about 10 spams about a word I said on here the other day...
Having to do with chinese time devices...
Oops...another 5 spams...
you shouldn't talk about Cialis and Viagra here, I guess
or Fake Rolexes
I jst get rolex, viagra, nigerian inheritance.
or a Hot New Penny Stock
Oh boy...now you ve done it...one year of spams for SWPadnos...
maybe we can screw them up with email addresses like firstname.lastname@example.org
oh...that's my next door neighbor
well - does he know how to use a computer? ;)
Did you get any Audi ads les ?
remember that winders southern edition?
les_w: can you tell me what a quaternion with s=1, x,y,z=0 represents? Is that just an "identity quaternion"?
Enhance you r perf0rmance with a fake rolex Audi v1agra!
paint icon was "colorin book"
ah now I can get the car spams
suprised I have not...since I posted on audi and infiniti forums
infinty crosses audi
Oh oh somebody is working on the TP...
les_w: can you tell me what a quaternion with s=1, x,y,z=0 represents? Is that just an "identity quaternion"?
are you gonna make me hit the books at this hour?
it does look like an identity
nah, if you don't know don't worry
I'm asking google too
[02:46:20] <wb9mjn> http://mathworld.wolfram.com/Quaternion.html
I have to hunt it up. I don't use quaternions regularly. The 3-d graphics guys do I guess
I've always used transformation matrixes
quaternions can be more computationally efficient
but the big feature is...
no gimbal lock.
I know the general idea I think
another thing to google.
just don't know the math
just a complex number.
plenty of web stuff on them
i imagine a "quaternion gimbal" looks like this: http://www.anthrobot.com/omni_III/
although i still cant figure out how it works
* ajoni just ate a gimbal for dinner
aha, wikipedia says it's equivalent to the identity matrix
I'm a bit bloated though
good enough for me
cradek: I thought it was but did not remember for sure
I forget things.
Really the computational efficiency is a non issue these days. Not blowing up at gimbal lock is the big thing with quaternions. It's about the only reason to use them.
I don't plan on using them - I just need to fool pmInitLine/pmInitCircle into not puking
Typically they would only be used in non trivial inverse kinematics
my plan is the TP will be able to follow a path in 6-space
you just need to provide it a representation of the path, an evaluation function that gives the 6-space point a certain distance along its path, and the path length
haha on # newyear someone wants an open sourced gpl'ed human genome
6 space is good
of course there's just support for lines and circles (helixes) now
les_w: crazy stuff on #newyears
but very good to do more than just xyxuvw if you can
simple_tp is coming along - I think everything is done but run_cycle() (the periodic function)
which of them?
gpl genome...they are getting very drunk!
tp_run_cycle() or tc_run_cycle() ?
well that's the part I'm going to ditch and rewrite the simplest possible
for every segment it will accel, cruise (if possible), decel, exact stop
what is a segment?
start with the basics and document it?
right now, a line or arc
les_w: right. I want it to be clear and functional
les_w: it should be obvious how it all works
once that's done, we can do path generation in non-rt (task)
right if you use a time scaling scheme for feed override
which seems ok to me
ajoni: yeah the task controller could feed it with "whatever"
les_w: I haven't figured how to do feed override correctly yet (you have to accel/decel when it changes)
any particular reasons for task?
les_w: baby steps
and not canon ?
ajoni: yeah I mean canon, which is part of task?
ajoni: whatever calls the tp api
well.. canon gets run by task
and then resultes get passed to motion controller which runs TP
Well, feed override by simple time scaling in the RT thread will not be time optimal....but so what? commanded (g-code) will be...
les_w: do you mean that accel would also scale down?
it just means that if you slow down with feed override it will not get tighter that's all
les_w: yeah I don't care about that :-)
les_w: my goal is just the simplest-possible proof-of-concept TP
or fred. We talked about it a bit.
les_w: mainly for documentation purposes (but also maybe as a reference implementation)
and as a starting point for people that want to do greater things
* ajoni starts pointing around
* fenn hides
so I won't have to hear how confusing and hard it is to interface anymore!
cradek: next we'll have to do the same for nml ..
I see overriding feed in real time as a correction to a g code programming mistake. Time optimization is not needed. You just don't want the machine to break!
and I did NOT want to talk about this until it's more done... :-/
les_w: I see feed override as a step down
At least lawyers aren't gaging you...
I guess I'm far enough that I would commit it anyway even if I can't make it work
always plan with the highest possible speed, then reduce it if needed
there is another usefull app for it, and that's feedstop
I turn feed override to keep the machine from estopping...if I need to do that then I need to improve my g code program so I don't have to anymore
ok.. night all
* ajoni gets the last sleep this year
happy newe year
what's left of it..
it's 5 am already .)
ah... #newyear is finally slowing down...they have all passed out
are they not a bit early?
burn bright, burn out
I would think
"I've got something to say. It's better to burn out than fade away" - the Kurgan
what's up fenn? posting just stopped dead
oh I know...
cia mind control waves
you stopped posting! haha
03cradek 07simple_tp * 10emc2/src/emc/kinematics/ (tc.c tc.h tp.c): checkpoint
(for you guys to look at if you like)
* fenn takes a break from irc-land
I'm trying not to change the API in tp.c but I might later
so it's somewhat simpler but not great yet
but tc.c is "done"
so starting from simple exact stop?
if I get that done, it should be easy to add colinear blends, for instance (just don't decel)
I want it to be simple enough that we can make branches and play with it
sounds like a plan
I notice nobody ran over to cvs and looked at the code :-)
so help me out again
hey - I was reading the email :)
vel is (new position - old position) / cycle time
accel is (new vel - old vel) / cycle time
is this right?
basically difference equations
so vel needs two position points
accel needs three
hmm, another case I didn't consider
for each long segment, there will be three phases, accel - cruise - decel
for a shorter segment there will be no cruise phase - the vel will just a triangular profile /\
for a segment that's shorter still, you may jump from the start to end point in one traj cycle
I'm having trouble visualizing what will happen for very short segments
that may be too small
or are these 10x he servo rate? ()for example)
cradek: you are doing exact stop mode so that will never happen
I don't know
fenn: ahh you're right
fenn: every segment will take at least one traj cycle for accel and one traj cycle for decel (to 0)
with short segments with no cruise phase the averasge velociy will be lower...due to the lower height of the triangle
so it will velocity adapt
note emc does not do this!
that's not really what i think of as velocity adaptation
well I see it as this...
if it can't do the commanded path it slows down until it can
again, emc does not do this and the algo breaks down causing queue starvation
Had a long talk with Fred about that
the whole queue idea is bunk if you ask me
If you request a certain path with certain accel and vel and the programmed values don't allowit...
you have two choices...
slow down until the path can be done...
for a while
fenn: there has to be a queue of upcoming motions, since only part of the program is realtime
i like the spline segment idea better
you only pass one spline segment, and stop at the end of it
you'd still have a queue of spline segments
yes but you have to have lookahead to make a spline
the idea there would be to make a series of cplines that share all but one control point
that means a queue
well, move the queue into realtime where it doesnt matter
you just rotate the points through the low-level interpolator
er, userspace not realtime
I think the queue has to be in realtime
if it's going to have to stop and think, it should do that somewhere it won't hurt the part
it certainly has to be read in real time
the queue is serviced in realtime
well fwiw, the framework I'm building will work as well with splines as circles/lines
as long as the splines can be represented in a fixed space
I doubt we would want to malloc each one but I guess we could.
I guess I don't even know if we have malloc in realtime
... I bet we don't
the spline must have a bounded complexity.. in other words, it cant take forever to calculate
Are you doing a TP queue in RT or user?
the queue is in rt
I don't see how you can get away from that
yes - there's an RT malloc
I wouldn't know...I don't know much about Shared memory and such in Linux
at least in HAL, so it must also exist in general
ok, so we have malloc
(god forbid we need it)
some RT version of that
what situation would you use malloc in?
and it's 4k at a time
fenn: to queue splines
queues with pointers, I'd bet
the queue should just be a fixed size array, with head and tail pointers
array of whats? pointers to kmalloced splines maybe.
well - a spline control point (or coefficient set) should be a fixed data type
ok, then it's easy
I would hope so, else it would get really nasty really quickly
doing this in trivial kinematics?
it shouldn't matter, really
kinematics and emc don't mix, after all
an acceleration limit doesn't really mean anything to a robot
if you use splines of sufficient order, then each axis can have a curved or straight path between points
les_w: like the old TP, three coordinated axes that are lines/circles and the other three that tag along
well just need 6 degrees of freedom fo rthe most generalcase
les_w: I think that's generalized 6-space with some special cases (the circles)
the RT code should only be evaluating an equation for each axis, the coefficient choice is a hard(er) problem, in userspace
trivial knematic plans can always be converterd to joint space by inverse kinematics if possible
les_w: trivial kinematics only works for linear axes
I'm no expert but it seems like kinematics can be done way before the tp
one thing I considered was generalizing arcs to allow us to do the racecar-algorithm tp we talked about before
so precalculate a path were all endpoints are tangent (blend with arcs)
I would consider xyzuvw or whatever trivial if orthagonal
i.e. machine space
les_w: I don't follow
les_w: you have to take into account whether the tool is mounted to the rotary table, or the part is, and where the rotary table is on the bed, etc etc etc
cradek: right, the RT code should only be following a prescribed path, not dealing with joint mapping
most people just program that in, but you're not doing kinematics in the computer
the curves used to describe the joint motion would be calculated in userspace
SWPadnos: yes I think that's right
any plan created in that coordinate system can be transformed to realizable joint space by Jacobean opereators
after all, G code is simple kinematics
I need to not get bogged down in all this, guys
indeed - I should be doing accounting, but too much counting makes me sick ;)
i'm just pointing out the limitations of the way emc was written
do any of you have a XYZA machine?
just keep it trivial kinematics...the other stuff can be don eelsewhere
if you write it over again the same way, but simpler, you'll have the same problems
once I put the motors on the Bridgeport and the rotary table, yes ;)
fenn: no, the problem is everyone thinks it's intractable
fenn: that's what I'm trying to solve
but a lot of the stuff in there that is confusing was caused by fuzzy thinking
fenn: I'm trying to make it possible for others who are smarter in other ways to work on it
improvement isn't a simple task - you have to think about what you want, but also how to get there without completely breaking the software in the process
a complete replacement would be grerat, but you might as well just start a new project
all you have to do is know how to use cvs
* fenn coughs
SWPadnos: that's just fatalistic thinking. There's no reason why we can't break the tp on a branch
look at me, I'm doing it now
this software is not that complex. I can build it in five minutes.
I can understand all of it in only ... a few years
I think fenn is looking toward a total replacement of the communications, the UI, the kinematics, the motion controller, etc
i think "looking toward" is the wrong word
ie, a replacement of emc
even if so, you can do that in stages
that's my point - you *have* to do it in stages, when people already use the software
jeff and I already "replaced" the gui with something (we think is) better
i'm "replacing" the glue that holds everything together
but then you have to replace everything
SWPadnos: maybe I missed your point, I'm not sure
fenn: even that's not true. You just patch up the APIs as you go along
I was pointing out that your approach is a good one in this situation
the high level stuff is really pretty hard to get perfect on the first try
fenn: the APIs are pretty clean here - they have to be since there are clear partitions
but its easier to think ahead and fix problems before you have to deal with them
SWPadnos: oh, ok then, I must have totally missed it!
I'm not sure I completed it, actually ;)
fenn: but I'm not smart enough to plan all the goals of the whole project at once...
that's too bad
i thought you said that it was pretty simple
it's true the program is "not that complex"
there's a simple goal "make a machine controller that can be used for arbitrary machines"
but everyone's goals at once? not so easy.
it's not complex compared to, oh, the Linux kernel
Me either for anything. Still, the MBAs gant chart "when I will innovate"
but compared to a mailreader client? yeah, it's complex. :)
so, I've got this pile of 460V VFDs
anyone need one?
i can see how to do it with mathematical concepts, but it's when the bits hit the fan that stuff gets hard
nope, but what do you think of this:
wish I had 460
if emc ran in userspace it would be easy
[04:14:16] <SWPadnos> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=7575290551
to meet everyone's goals
les_w: me too
but the realtime requirement blows everything up
one way to realize a lot of (possibly conflicting) goals is to have "pluggable" components
but that's not so easy at all the levels at which emc works
* djb_rh wonders what a 1Ghz oscope has to do with anything
(RT, kernel, userspace, GUI, remote)
about as much as a 460V VFD :)
hey, you can run a mill off one!
SWPadnos: if you're offering to give me a scope like that, by all means speak up
people here run mills!
djb_rh: thy do? i thought we just talked about it
I'm just not sure if I should buy that one - I can't tell if it's legitimate
okay, maybe so
feedback of 7
SWPadnos: when in doubt, use your phone, talk to the seller
and the userid has "golf" in it
yeah, and a $10,000 purchase
SWPadnos: you can read a lot more then
and they are selling a $10k oscope
seems not right
I was just wiring up the rest of the new 240v 4 Kw vfd/spindle. That fat 20mm wire is not fun to route.
we emailed back and forth a couple of times, but I never got any real information
SWPadnos: escrow escrow escrow
SWPadnos: walk away then! trust your gut
I may just go for the "real thing" from Agilent, but it's $5k more
yeah - I've got an OK gut
SWPadnos: tell him you have a buddy an hour away who will come and make sure it exists and see what he says
$15k << $10k + $15k
I wonder how he expects to acecpt $10k via PayPal
I need better equipment for the electroncs lab
SWPadnos: hell, for that kind of money, fly out and get it
only one feedback in the last 12 months probably means stolen account
using old analog tektronics storage scopes
hey - he is in florida - that could be a good thing ;)
les_w: I love my tek 466
well they work
I love my Fluke digital
heh - maybe I shuold donate my old 54622D to the cause ;)
forgot the model number
I've never met a digital scope I liked - but I've also not met any modern ones
they take some getting used to
the 6000's are pretty darned amazing
the controls just aren't "right"
I like to turn that knob until it looks like I want
I had a pretty good one in Chicago
until you get the hang of them, anyway
it cost $16K
they operate like an analog scope - they have a 100k wavefornm/second update rate
cradek: I was willing to get used to one given the size
but no way I will spend that much with my own business
I used one in school and I would have traded it in a minute for a vacuum-tube tek
ha, which is what I had at home at the time!
I do some pinball repairs in the field from time to time and it's nice to have one that's *really* portable
the 8 bit ones are no good
the nice thing about these is the deep memory - you can capture 1M samples, and zoom in after the event
12 or 16 is better
almost all digital scopes are 8 bit
yeah, pre-triggering like halscope would be nice and it's impossible with an analog storage scope
these are also nice, since they have 16 channels of logic analyzer on them as well
I think we're accelerating out along tangents here...
I should go do something else for a while before bed
first the VFDs, now this ;)
SWP I know...and not good for me...If I get anything else it will be a laptop scope module
oh, it's all MY fault!
hey, unlike your $10k "deal", mine really exists!
cradek: go watch tv for a bit
don't bother, unless you're willing to spend around $8k on it
my first CNC control box shows up next week
the low and midrange laptop (or PCI) modules generally have crappy triggering, and poor "live" performance
unfortunately the parts to convert the mill won't be here for another month
pre-built, or a cabinet waiting for homemade goodness?
or listen to XM radio elevator music or something
SWPadnos: it's just a pico box for a mini-mill
got my PC running that's gonna control it (assuming it's fast enough)
SWP: I really need signal analyxer functions
man, it's a dog to surf with running X
shouldn't be a problem - that's waht I've got
FFT, cross correlation, coherence, etc
load average goes above 1 just logging in on FC4
these do FFT, but you'd need a higher end model (like the Infiniium) for the communications testing functions
did BDI go from 4.29 to 4.30 in the last three days? or did I just find a mirror that isn't updating? The latest I saw was 4.29. :(
actually, there's a 4.38 now
Paul just uploaded it 2 days ago, I think
[04:26:48] <SWPadnos> http://www.cncgear.com/EMC/BDI/
well that's one coaster burnt
I was considering building a house with all the AOL discs I got - maybe unused BDIs should go in the pile
they're sending controller, motors, and power supply
that means I can actually test EMC
I'll be able to make motors spin!
that's always fun
won't that be big fun
did you get a PPMC / USC, or just the motor drivers?
#newyear woke up again...what an innovative nick I see..."lilo"
I'm too much of a noob to understand that question
what's in the box you're getting? :)
I told you: motors, controller, power supply!
it's the whole package from http://www.stirlingsteele.com
ah - you said Pico - I assumed you meant Pico Systems (Jon Elson's company)
stirling steele sounds like a romance novel author
yeah - and his "pico" controller ;)
a very small whip
can you trademark an SI prefix?
you can trademark a piece of glass in a frame, as used for about 1200 years
it is Jon's controller
Stirling is a reseller
fenn products introduces the Femto (tm)
I'll go with the Zepto
djb_rh: i think it's a different product
I think you're right
yeah - I'm pretty sure it isn't Jon's
[04:33:50] <fenn> http://pico-systems.com/motion.html
<- jon e's
yeah, I'm going to be getting Jon's for my Bridgeport
gotta get cracking on that
that's gonna be a bitch of a project
nah - it's fun
(or it better be, that's what I'm doing)
it's a big dirty mess of a machine right now that I have no idea what works and what don't
though I've got to say, I'm not as happy with Jon's USC after helping write the HAL driver for it
buy lots of Oil Eater
what bugs me is it's already CNC
so I'm gonna try to use a lot of the existing stuff
NC or CNC?
motors, encoders, power supply, amps
that should be good
had some old PC thing with it
dual 8" floppies
and a tape drive!
ok - you'll probably want a newer PC
I think servo motenc-lite is the way to go
and one of the servo cards, like Les says
djb_rh: hey did you figure out what kind of machine it is?
the Mesa card is pretty cool too, though I don't know anyone using it on a machine
cradek: the computer? nope
djb_rh: I think I bugged you about it last time
SWPadnos: instead of Jon's controller?
cradek: yeah, you did and I forgot to get some pics of it
djb_rh: practice reaching for your wallet, and remember to stretch and warm up so you dont hurt youself
it's a PCI card, so communications are a lot faster
SWPadnos: too late, I'm definitely going with Jon's...I'm going to need a lot of help and he's been very willing to give it so far, including helping me find spare parts if I need it
it's also changeable, whereas Jon's would need a reoplacement chip from him
fenn: oh, I'm already over $3k invested in my stupid mini-mill!
well, with tooling and work holding and CNC kits and upgrades and ...
Servo update is all important with the old emc, and the pci motemc has demontrated 8 kHz.
yeah - 3700 reps of the "WalletReach"is a tough workout
no reason why the Mesa couldn't be that fast as well
the FPGA code is included with the driver
(or downloadable from the website)
I suppose I should see if mine works ;)
I can't wait to make my first PCB
well, with a mill, I mean
I've made PCBs before
ah - etchant in the nostrils
servo update is important because trajectory update is an integer fraction...and you want to stay away from the queue starvation area
djb_rh: it's really fun (compared to etching)
SWPadnos: I started etching them when I was 13 or so...I still remember well the time I was searching for a pan the right size to use for a particular board so as not to waste a bunch of etchant and didn't notice the pan that most correctly fit the board happened to be, err, METAL
cradek: I'm sure
heh - oops
actually etching isn't bad, it's drilling that I hate
that pan started smoking something fierce...all I could think was run outside and toss it in the grass
etching double sided boards was always fun
and hose it down
never tackled any double sided stuff
Yes I absorbed enough ferric chloride to have a lifetime dose of geritol
and yeah, drilling blows
we did a memory upgrade card for a Vic-20 when I was in High School
les_w: you're gonna need it :)
I had one of the drill press attachments for my Dremel
made it almost bearable
I think that one was DS
I had a summer job etching IC leadframes with ferric in school. Had enough for a lifetime.
I assume the process would be to use an endmill to cut away the copper you don't need and then you replace the endmill with your tiny bit and let the mill drill the holes, right?
all under CNC control, obviously
not an endmill - a v tool
and pottasium ferricyanide.
a v tool?
damned noob here again
crap, I can't get to my web site
there is some stuff on my site timeguy.com
a tool that's tapered to a point
"the internet" is broken again though
[04:46:08] <fenn> http://www.pilotltd.net/engraving.htm
okay, i get the tapered to a point (and figured that's what was meant)
but don't you typically need to remove a lot of surface area of copper?
nah, just isolate the traces
why are digital boards always completely etched away?
oh, you leave everything else?
because they use etchant
wouldnt the ground plane help prevent emi?
8because people don't use big ground planes
also, in large quantity, the extra copper may add a few cents to the board cost
but you have to start with a whole board anyway
yes, but the board houses do their own plating, and they can reclaim the etched copper
I had my last proto boards milled
i thought it maybe had something to do with lower capactiance
cost about 10x as much as etching
not usually on FR4 board - there are controlled dielectric boards (including ceramics) available for those applications
les_w, how long did it take?
.005 end mills
from the time you sent the files until you had boards?
I did it because ITW tech center bought a machine. I had to spend hours getting it straightened out.
I charged the usual fee
etching would have been much quicker
ah - one of those "get the machine that goes 'Bing'" moments
hey I made a bunch of money
could have sent the art to a proto house and have it done for $60
"I try to please the client"
client bought the little machine...needed to use it.
had to screw around for hours getting a layout with no plated through holes
and no soldermask
well - that works
leeft shift key is sticking every time
hey what's the green stuff do anyway?
damn proto keyboard
I ve used that printable mask paper...works well for simple stuff....
it lets you take food from stores without getting chased
oh - that green stuff ;)
why is it green?
it isn't always
heh on "modded' pcb boards it's hot pink
But , you have to choose and tune the foil application for the trace widths...very sensative to that...
you can get green, yellow, red, blue, black, silver, purple, etc
it's there so that solder doesn't go all over the place when you put the parts on the board
it also protects the tinning on the copper traces
G10 is green, so they mask green as well...
it is usually blue or green to have the right emisivity for infrared or uv curing
yellow is fairly common as well
Kinda wierd when you have white circut board...
I Used "delco blue" for car circuit boards
or absense of tinning...
good IR curing properties
actually, we got a run of boards that looked all gold one time
Probably why G10 is green too ?
they had yellow soldermask, and the traces weren't tinned
what is G10?
Motorola likes red...
is this stuff epoxy or silicone or what?
G10 is a standard epoxy/fiberglass laminate used for PCB s...
the glass raises the dielectric or something?
ok - I always see FR42 unless I ask for something else
Just keeps it stiff...
Look at your Rockwell water meter remotely interogatable board in your house if you have one. I did that. It is blue, green, and black.
i think we're talking about different things
or a Soyo Dragon Ultra Platinum motherboard
i was talking about the green "stuff" that they put on circuit boards
yes - that's the Soldermask - it keeps solder from briddging between nearby pads/traces
G10 - FR4 do not know the difference...same stuff I guess...
could be - are you in Canada?
Great Grandparents on one side were...
I wasn't sure if it was a national border thing
here's what PCBFabExpress says about G-10:
again...the green promotes IR absorbing rather than visible light for rapid curing
G-10 General-purpose epoxy/fiberglass woven fabric PCB material. Most vendors do not stock this material; instead, they use high-grade material like FR-4 as a substitution if you specify G-10.
I don't think so, les
Red would reflect IR...
no - IR isn't visible - you can have visibly clear goggles that absorb all IR
green reflects green light.
Also had some elemetary schooling in the UP, and they spell like Canadians up there...and it leaks out in my typing sometime...
exactly - that's why I say the color doesn't matter for things we can't see anyway
SWPadnos: you can see IR if you wear goggles that absorb all other visible light
the sky is black, and you can't see car taillights that use LED's
fenn - that may be so, but the sensitivity of the eye to IR is so low that you do have to block out *all* other light to get any usable S/N ratio
We use quartz elements to cure pcb coatings. They have broadband emission.
Ir is more penetrating
we don't want visible
sure - the skin depth is higher for longer wavelengths
[05:06:04] <fenn> http://www.amasci.com/amateur/irgoggl.html
green dyes reflect visible
that's why we use them!
Welp ROYGBIV...Yellow Green are in the middle...
we want to absorb at the approx 3 micron methyl resonance.
Anything north of Green should probably absorb IR...
so the dyes are not just for looks!
the visible color is totally independent of IR absorption
though the formulation of the dye makes a big difrerence, and the "better" ones may tend toward certain colors
(ie, the ones with a good IR absorbing molecule may also reflect e.g. green)
Well, one has to start an investigation somewhere....I heard once that Aluminum is actually a really poor IR refelctor....
Is that true ?
depends on the wavelength
SWP: have to disagree there. It can be strongly related.
IR is a whole lotta spectrum
it can be, but only if you restrict yourself to certain dye formulations
So, if one were trying to make a solar kiln, one should stay away from Aluminum reflectors ?
I recall in the military that people were afraid to use the "clear" goggles around the laser rangefinders
I used them, because I realized that one could filter out all of the IR, and still be left with the full visible spectrum
Yea, probably because one could not tell if the coatings were really there or not...
and I didn't explode any blood vessels ;)
yeah - that was the problem
Even if the coatings did block IR...
well we go for dyes that maximize curing...and trhat usually means rejecting visible since it skins over films due to lack of penetration
you couldn't readily telll them from normal goggles
ok - it may just be that there are newer dyes that work enarly as well as good ol' green
I think the other colors only became available in the last 10-15 years
There was a company here in town that made laser diodes....the rumor was they would party on top of there building and shoot the 4 watt lasers at each others houses when they got drunk...
You know those crazy engineer types...
dyes are very powerful...they lase and fluoresce
yes - dye lasers come to mind ;)
like day glow pigments
or white LEDs, for that matter
(and fluorescent bulbs)
starch - glows in blacklight
haha #newyears has become a moderated channel
they actually absorb shorter wavelengths like blue
and re radiate
yeah starch is powerfully flourescent
what color does starch fluoresce?
I worked with narrow spectrum flourescents a lot
starch looks blue white tothe eye
what I did was this....
mixed narrow bandwidth compounds together to make a spectrogram that had a simple didital code
spray the stuff anywhere...
and it's like a barcode
heh - and read the code with a spectrograph
it is used to markand catalog trees to be cut
it is used on currency too
reminds me of dna sequencing
it is used on check paper as well
I just put it in spray cans for tree trunks.
BTW, it is invisible to the naked eye if desired
dont want those pesky humans to see that you're gonna cut down their trees
Imagine...a few hundred ppm of this stuff and materials can be identified from great distances just by shining a light on them
well. you can already do that
and they are.
without the stuff
but not things like...
vin numbers of cars
how many bits can you squeeze into a drop of paint?
10 or so
the stuff looks gray mostly.
argh - good night guys
happy new year if I don't see you before then ;)
it is expensive. Rare earth minerals are used. Small concentrations though.
SWPadnos is now known as SWP_Away
Jymmm is now known as Red70sShow
Red70sShow is now known as Jymmm
Jymmm is now known as Red70sShow
Red70sShow is now known as Jymmm
[Global Notice] Hi all. In about 45 minutes, it will be 2006 on Kiritimati (Christmas Island).... don't forget to go get your invitation to freenode's New Year's celebration.... http://freenode.net/news.shtml
[Global Notice] Okay, folks, this is it. We are approaching New Year's on Kiritimati, which is pretty much the first moment it's 2006 on Planet Earth. Please come by #freenode-newyears and help us celebrate! :)
search for 0917504
hey rayh ! :)
How are you today?
rayh: just ready for the party :)
alex_joni: I tried my first photomerge: http://digilander.libero.it/jackydgl0/panoramica.jpg
I think, I missed all :(
4 photos ..
Ah. Party is good. May the new year grant all you wish and need.
rayh: hope so, for all ;)
Interesting image, Jacky^
rayh: thats naples gulf
could be nice , are 4 images merged ..
but as you can see the merge is bad :(
as first experiment should be enough :(
hehe yeah, really beautiful
automatic shutter speed and fstop?
the bad weather do not help to get nice photos in these days
rayh: I bought a digital camera Sony, im playng with it attempting to get best photos I can
for now IM just using auto-mode features
manual setting should be much better ..
You'll get the hang of it.
You run Linux?
rayh: can you look a bit in the board channel?
not sure my messages got through
Jacky^: I have a dcc for you
It did Alex. I'm a bit slow today.
ouch .. looking how to accept it in chatzilla ..
rayh: no problem, it's just my net connection is crappy..
not sure how much I'll be around :(
may is disabled from settings
say when Jacky^
dcc enabled.. alex_joni try again
rayh: the modem is USB powered, and it seems it's draining more power than it receives from the USB
so the battery runs dry :(
ok, dont worry :)
Jacky^: still there?
alex_joni_laptop is now known as alex_joni_
trying again now
Jacky^: that's taken in bad weather..
doesn't look like one on a sunny day
but I like it
alex_joni_: realy nice :))
not that nice..
I like ssh ;)
Fetched 5053kB in 6s (765kB/s)
nice, can be used as wallpaper :D
want a bigger one?
send me ty :)
oh gawd.. how do I tell debian NOT to update packages?
I rarely used that option, but its possible to lock a package
ok, alex_joni_ this is much better ;)
I just want to install irssi..
not update the kernel..
fsck it ;)
Hi Ray...gotta question about the TCL GUI...
How does one know which coordinate system a MDI command will act on ? I set a relative coordinate system for a part...Then did a little program in Auto...
After M2, and returning the MDI, even though the relative coordinates were on the screen, the MDI commands acted on the machine coordinates...
I figured out that if i did a reset, the MDI commands would then act on the relative coordinates....
somebody not afk?
Sorry wb9mjn, was off reading
Was off trying to do IRC from the milling machine...
KSIRC is working, but ugly...KVIRC was working before, but wont connect now for some reason...
* alex_joni kinda likes irssi
a bit hard to get used to it, but it's ok
Here I am back on the milling machine...
wb9mjn-servo: you knowit's not the best idea to log on as root
Oops...you can see that on here ?
Oh...there it is...
in some channel you,ll get banned as root :(
Never really setup anything but root on this machine yet...
Can't you start ksirc as a different user?
Question - what version of Lyx is being used now for the docs?
something like "su rayh" in a terminal and then start ksirc
good question jtr
Let me try...
Ok - developer.pdf mentioned version incompatibility. I'm trying to get my head around EMC2, want to look at the latest docs.
jtr: Any recent LyX will handle the task 1.3.xx should do the job.
The 1.3.4 does a good job with graphics.
I've not tried the 1.4 pre stuff yet.
Good - my rendering speed on the pdfs is slow, plus I don't know how well they keep up with the masters.
We need some help to get the masters ahead of the pdfs.
I did a bit of work on Hal_Introduction.lyx the other day.
We have a couple days work to get that into shape before the release.
Well, I think I've gota lot of learning to do before I can be of a lot of help; just getting back into linux after a long absence.
first exposure was patching kernal sources for RH5.2 - never got the RT tests to pass on that 486 machine.
work took over and I haven't been back til recently.
jtr: seems your start was too steppe :)
darn, don't like sluggish ssh's, can't see what I'm typing :(
03alex_joni * 10emc/ (12 files): reverted RUM changes on inifiles, preparing for release
whats up ?
laex is up :D
[15:32:16] <alex_joni> http://solaris.cs.utt.ro/~emc/
<- works again, didn't notice that the make tgz was failing :)
Actually, didn't mind the patching and compiling - enjoyed tracking down the failed hunks.
morning - now did you know samco?
glutton for punishment.
skunkworks: sorry, what?
<alex_joni> morning samco
oh, I meant sam :D
but samco is correct, right?
but I didnt know samco
< skunkworks> morning - now did you know samco?
or should that have been how?
I am on a lot of boards as samco. (when I was younger I used it as my "company" name)
I might tell you someday :P
that sounds scary. What am I wearing right now?
I don't dare looking
Did I see cradek working on a new trajectory planer last night? <- acting like I know what I am talking about
skunkworks: you did
and I just looked today at the results
looks very promising, half is gone already .)))
sweet. How did he do?
next half the next few days
cradek seems like a nice guy. ;) He has a thing with clocks but I don't hold it against him.
not sure.. a bad timing.. is crucial for a new TP
new camera photos - no pressure
I started a place for them at least
[15:46:42] <alex_joni> http://dsplabs.cs.utt.ro/~juve/blog/index.cgi/photography
didn't really add all I wanted
dad has been running emc/axis for a few weeks now. He is getting used to it and liking it. (he wants to get a rotory axis on the machine to play with)
skunkworks: ever solve the issue with the stepper backing up a few degrees after a move?
I think that disapeared after adding headroom in the stepper hal for acceleration
skunkworks: nice to hear that
cool - trying to get my head around EMC2 sources, then see if I can find any bugs...
it was either that or the issue with the default acceleration overriding the max_acceleration of each axis. Cradek patched that.
skunkworks: the backing up was because of headroom
(It might have been when I had the default acceleration set higher than the max_acceleration of an axis.
jtr: you're very welcomed if you do
Thanks; currently I write equipement interfaces for video conferencing systems - wouldn't quite call them device drivers
the language is simplistic and I feel my skills have slipped - this would help me sharpen up and
give me a new toy to play with.
I swear there are not enough hours in the day
skunkworks: sure there are
24 is plenty
you just waste too much of it
the bad thing is I like sleeping
then don't complain
thinking back - I think you are correct - it was the hal acceleration headroom that fixed it.
I know I am correct :D
and modest too about it
I know - I know
I feel bad that all I can really do is break emc. ;)
that's called beta testing - and is a required function...
there, isin't it nice to know you provide a needed function to the borg? ;)
It is nice being part of the colective
alex_joni: the webpage is looking nice. You might want to add clear:right to the css of the floating images, though. On my wide screen (1920x1200) I get a stairstep effect.
thanks for pointing it out, looking now
Hi - cradek
alex_joni_ is now known as aj_laptop
djb_rh_ is now known as djb_rh
Jymmm is now known as Red70sShow
Red70sShow is now known as Jymmm
aj_laptop is now known as alex_joni_
Praktica B100 - is that a rebranded camera or its own company?
Praktica? one very big brand
could have googled I suppose.
ajoni is now known as alex_joni_
Welp...still trying to get another user setup on the system.....got the user there, but EMC bombs when it starts...no permisitions to write to the usr/local/emc directory....
wb9mjn: try sudo in front of it
and make sure your user is properly set up in /etc/sudoers.conf
User is setup...just does not have permissions to do anything outside of /home/<username> ...
as I said.. use 'sudo'
well - off to work on the basement. Have fun.
sudo <username> ?
no, sudo command
sudo executes a command as root
just sudo ?
I m executing from the menu....
then probably you are not added in the /etc/sudoers file
What does sudo do ?
<alex_joni_> sudo executes a command as root
What good is that , if the purpose is to avoid all the dangerous things a root user can do , when one just wants to run EMC ?
I could just login as root ?
EMC needs to get run as root
there is a difference in beeing logged on as root, and running a command as root
EMC needs to insert kernel modules, etc, which all need root access rights
Let me go look at it and see if the menu will let me set a startup command, not just a file...
the file executed already has sudo in it
in the proper places
but because you added the user yourself, it has not been added to /etc/sudoers
heh, that reminded me....
Ok...put in "execute command as another user" ....still bombed...
wb9mjn: open a console, and try it by hand
see what happens
What is wierd is that all the other preinstalled EMC startups work...just not mine...as a user....but it works just fine from a root login....
just not as root from the user login...
Well, in effect its doing that...but will try...
Some friends had bought some clams, saw one and thought it might be bad, then decided to cook it anyway. When it was opened, inside was a rare purple pearl. Was told only about 13 in the world.
Could you imagine tossing it out then to fins someone else had rummaged the trash and found it?!
It's my story, I can embelish it if I want to =)
it could have ended in trash :D
that's what I said, but ppl do go dumpster diving ya know
not for clamps
who would want that? not even a desperate homeless
alex_joni ok, your in the dumpster, you stop on an unopened clam. you crack the shell, you see somethign purple. it's a peral
I'm not in the dumpster
you'ld make a lousy dumpster diver!
I don't dive for pearls
and you'ld mever find a rare perl
especially not inside dumpsters
we've found brand new dishwashers, lawn mowers, weight sets, computers, and a whole lot more.
ok, but no pearls.. right?
Jymmm has kicked alex_joni_ from #emc
Knew there had to be something different about the menu editor setup of the preinstalled EMC's...
see.. told you there are no pearls in dumpsters
I started by copying and editing one of those to make a new menu entry...
But then when I selected the wb9mjn-servo.run , instead of generic.run, it just put in a file path...
In the preinstalled entries, sudo preceedes these, even though you cannot see it in the selection window...
alex_joni_ has kicked Jymmm from #emc
alex_joni that was lame
not lame, but there weren't any pearls there
So, by scrolling to the far left in the selection window, one can then enter sudo manually, which was left out by the "browse" selection tool...
In Root, this omission did not matter ...
Now, you do not even need to type in the root password, and it starts up just fine...
After lunch/errands, I wll try the KVIRC/SIRC thing again...
* Jymmm liked kvirc
I linke KVIRC too, but it wont connect anymore for some reason....
Maybe with the reboot it ll be fixed....
I tried a few months ago and it was fine...
But then tried KSIRC today, and afterward then tried KVIRC, and nogo...
a lot of peoples here around :P
hi alex_joni :)
get ready for restaurant :D
and dance all night !!
if that suits you :D
18:30 right now here
19:30 here ;)
gone to watch a movie ..
ready to go .. we (me and k4ts) wish you a Very Happy 2006 !
have a fun ! youre Great :D
all the best to you too
* jtr is back
* jtr is back
well that's wierd.
jtr is now known as jtr_away
test received ;)
Just setting up KVIRC here on my milling machine computer...
that's cool; got emc running?
hey djb, you around?
Oh, for some years....this is the latest machine though...
Its a bench top class machine with servos....
ah, I gotcha. I'm relatively new to emc, but I'm designing AC servo hardware
This machine has brushed DC servos, with Gallil amps and a Motenc-100 interface board...
nice.. have it working/tuned?
Yep...its a 2 year old project...holds x and y to better than .0002...
I bought a bunch of 1kW sanyo denki AC servos from surpluscenter for $40 apiece
so now it's just about figuring out how to drive them
I bough a real junky piece of iron off the internet,,,stripped out all the 1970's electronics, and rebuilt...
1 KW would be a good spindle for this machine...hi..
yah.. i don't know what kind of machine i'm actually going to build.. i've got a mini-mill with steppers now
It has a DC spindle drive now...would like to upgrade it eventually...
I started with a really small mill with steppers...a Proxon MF-70....
Rebuilt the table, and added motor mounts....Got Gecko drivers for it...they are really good...
I had this computer on it , but then took it off for the servo mill, and put the old 100 MHz P1 back on...what a dif !
i've thought about going to a servo spindle for the mini mill. it'd be a fair bit of work. i have a sherline mill with the CNC conversion
If I was starting again...I think I would go the Sherline route....save time with the infrastructure and do more...
Although learned allot about steppers setting that all up...
well, i had just gotten the mini-mill at first, and spent enough time retrofitting etc ;)
i got the pico systems amplifier board, built a chassis for that, set up the motor mounts. still, it was a lot closer to turnkey than it could have been
About time, since I had a control prof who was one of the pioneers in microstepping...many years ago...hi..
I do like the sherline machine, but, it's not the most rigid/robust thing in the world.
I got the MF-70 because it had the 15 K rpm 1/8 inch (Dremel style) spindle...figured it could cut anything with the 1/8 inch cutters
that are available...
yah.. more rpm would be nice. It'd make up for a lot of the lack of rigidity
that i'm dealing with. i can't take very much off at all per pass
My new machine is allot bigger and rigid, but its only about 2 K rpm...I replaced the original table with two linear tables off Ebay...
Not as rigid, but I mostly do aluminum and brass...and it goes through them like butter...
The original X-Y table on this thing was really poor....some poorly done amateur class hack of a cheap china table....
Was really disappointed when I got it....The machine was originally a tech school thing...
The column is nice though...and the sheet metal enclosure is ideal for the situation/rebuild...
* icee nods
I need to decide what these big motors are going to drive. have been thinking about either a bridgeport conversion or rigging up a big plasma cutter
Some learning experience with this too...squaring the X-Y table...tuning the servos...
i'm just about ready with the amplifier design etc to be able to spin them
Probably better for a router/plasma cutter class machine...
Might be too big for a Bridgeport X/Y ...
Although, might work for direct drive on a BP...
Most BP implementations use a belt drive...
yah, that's what i was thinking.. direct drive. they're probably a little marginal for the knee
I have a 2 HP AC (Brushless DC) servo on my lathe belt drive...works well...
Oh...what I have seen is people drive the quill, not the knee...
Although what I ve heard about BP's is the quil should be locked during milling....never figured out that incongruety...
Its an 8 inch swing lathe....
yah, if you can lock the quill you get more Z axis travel and a lot more rigidity in the spindle
A 1 KW would be a good drive for that size of lathe, with various drive ratios...
* alex_joni is about to enter a new year
15 more minutes :D
Alex is over in Romania I think...
hehe. 10 hours here
Hi Alex, this is the KVIRC now on the servo mill...
wb9mjn-servo: : nice
will greet you from the future soon :D
Did this so I can play with the EMC while asking questions on here, rather than run between rooms...
wb9mjn-servo: that's the proper way to do it ::)
Just made it through the post office after Ieft, they closed down at noon, and I got there with minutes to spare...
Have to get those bills out by the end of the month/year deci-decade...hi!
Hard to believe its already been 5 years since 2000...
Trying to get rid of the times in front of the text...
Happy New Romania Year ?
ooh happy new year alex!
* fenn spills champagne in your lap! "oh damn my bad"
* cradek celebrates the new year by making backups
* fenn too
my colocation crashed yesterday and I found my last backups were from October
shame on me
fortunately a reboot fixed it (it ran out of swap, and crashed after logging a polite warning message)
any one know root password for live bdi 4.18??
it's probably randomized at install
then what?? install what i want??
when you installed the system, I think a random root password was chosen
you can sure set it if you want, but using sudo is just as good (better?)
im running from cd live boot
bdi4 is not a live cd
4.18 or 4.2... the morphix based one
are you logged in as emc or whatever?
it logs in a morph
if so, just use sudo
with what options
sudo [command] runs that command as root
so if you want a root shell, sudo sh
or 'sudo -s'
i wanna install a deb package
I didn't have any luck doing that until after doing a hard-drive install
dmess: sudo apt-get install ...
leap second in an hour!
you are such a geek
your clock's supposed to show it correctly?
no, I have no idea how to observe it
I doubt I have any clock will do it right
have it ready by the next leap second
that might be many years!
make fake wwvb signals
I don't think wwvb has a provision for it!
I'm pretty sure it doesn't
greetings from 2006
hello alex from the future
who know the command to apt-get install xmms ???
hi guys, how's the past?
dmess: you just said it
[23:05:32] <jepler> http://www.leapsecond.com/notes/ls-wwvb-97.htm
i thiught so... so why does it puke??
dmess: the Live is pretty old
last year I think..
so xmms wont support it any longer??
so there's no warning like there is for DST switch, just an extra marker second
or does this thing even have sound
dmess: sure it will, but you need to find a place where you find a compatible package
that might be a problem
cradek: no, one of those bits is called "leap second pending", at least that's what the text implies
maybe it's the -3th digit before the final P?
[23:09:27] <cradek> http://tf.nist.gov/timefreq/stations/wwvbtimecode.htm
Day changed to 01 Jan 2006 <- irssi sayd a while ago
see you guys next year
yup, bye jeff
it's been a great year
well simple_tp moves!
it actually follows the programmed path
never mind the coordinates flash "-inf -inf nan"
where is that? .. must be outside the machine area
yeah, must be
can't do G1 with 0 feedrate
you use that for machining complex parts
nah, that's my home position
haha complex parts
fenn: to obtain perfect angles