todays project.... Yay emc2
yup... far too much time on my hands!
a fine example of TTT output, subroutines, g90 to offset strings .... lots learned from this project 'Thanks Guys'
luckily you didn't want to *rotate* any of that text
* jepler hides
oh come on jepler - you should be able to write a python script to rotate TTT text in no time
heh - truetype trace -> gcodetodxf -> rotate -> ace converter.
nah, I'd modify ttt to take an affine matrix in an inconvenient format on the commandline
ooooh - even better
seems like a new gcode is in order ... that works like g92 offset only it does a 3d matrix warping
heh - someone asked about something similar on the userlist.
no no ... ttt -affine ".49 .866 0 -.866 49 0 0 0 1" "Hello Rotated World!"
it has to be user-hostile -- it's a commandline app after all
complete with an example in the documentation that doesn't work as suggested
I did have one idea for axis.... when stopped , it could poll gcode file time/date and reload if it changes.. I used axis to layout dogdish with an editor
if your editor is configurable, you can run this shell command when the gcode file is written: axis-remote --reload
you can also make it load a different file: axis-remote newfile.ngc
since my favorite editor *can* run a command when it saves a file, I prefer to think explicit polling by axis isn't required
open source editors are always configurable!
03jepler 07TRUNK * 10emc2/src/emc/rs274ngc/interp_read.cc: fix sf#1788078: correctly diagnose a missing bracket after 'O- if', 'O- elseif' and 'O- while'
SWPadnos_ is now known as SWPadnos
hmm.. in linux the mounting is easy but the unmounting is not
what's the error
"device or resource busy" .. lsof showed a bunch of stuff was using my dir, so i killed everything and still get the error. anyway then i could unmount it instead of rm'ing it
do you have any shells open with that as the CWD
its fixed now, but, it was inconsistent
step 1: umount /var/tmp: device or resource busy
step 2: delete everything in /var/tmp
step 3: fuser (shows nothing)
step 4: lsof (shows a bunch of kde crap)
of course it didnt complain when i deleted all their stuff
step 5: delete /var/tmp itself: device or resource busy
step 6: umount /var/tmp: success
er, 4.5 killall k* :)
step 7: more tequilia
[06:30:10] <Skullworks-PGAB> http://www.youtube.com/watch?v=KxmHEGy7JUU
I can see the guy missing it - but the whole studio audience?
[06:44:10] <Jymmm> http://www.youtube.com/watch?v=sVPa46W3uMY&NR=1
Heh, Just this one item alone got me hooked http://musiclessonz.com/rebol_tutorial.html#section-6
cradek_ is now known as cradek
JymmmEMC is now known as Jymmmmmmmm
Jymmmmmmmm is now known as JymmmEMC
uh hi, while you think that's funny and/or insulting to us, in actuality it's just misogynistic
cradek: You got a word-of-the-day calendar, didn't ya?
actually no, I'm just well-educated
and, I recently read an article about being honest
I agree with the misogynistic assessment. it would be pretty mean to womankind to insinuate that they're as ugly as this bunch
You have to be political, to be politically correct.
SWPadnos: that's a twist I hadn't considered
SWPadnos: Hey, if the papr/plastic bag fits.
cradek: That's SWPadnos for ya, twisted as can be =)
I think "Morning, Ladies" can only be seen as insulting if it's demeaning to be female, and if it's not insulting, why would it be funny?
SWPadnos: (That's a good thing in my book)
cradek: This is IRC, not the local gym, So the term could very well be true.
cradek, in a sense, it can be demeaning to call anyone something that they aren't. That implies an insult to the conversing party, not necessarily the third party referenced.
if I call you a lump of coal, that's bad for you, but not insulting to coal-kind everywhere
* archivist thinks a lump of coal fits in this instance
SWPadnos: that's interesting
coal is just low-pressure diamond ;)
fwiw: the article about honesty: http://www.esquire.com/print-this/honesty0707
I thought it was entertaining/interesting for a while, but by the end I think it's just some jerk trying to sell his seminars
I agree that the intent is often to say "you're a member of this less desirable/respactable class", but I don't think it's necessary to consider any instance of that type of comment as implicitly demeaning to the third party
maybe you should mention that if there's a comments page
just being honest, you know ;)
I feel asleep after the 3rd sentence
"You appear to be some jerk peddling seminars. How's that for honesty?"
I think adding comments to random blogs/articles is just wanking
read like a blog - yuck
SWPadnos: no woman would walk into a room where a group of her (female) friends were, and say "hi, men" or "hi, boys" unless she wanted them to have a laugh about how men suck
I suspect that's true
there must be a reason for calling someone something he/she clearly isn't
cradek: It's called humor
comedy generally implies some classism, but it isn't meant to reinforce the classes, it's meant to (a) provide humor and (b) get people to consider their prejudices
most ly (a)
the word parody comes to mind
JymmmEMC: sorry to make you the butt of my honesty experiment
cradek: Eh, I dont give a shit... it's all good.
the closest thing is irony, but that's when you say the opposite of what you mean, and it's funny just because of that
so maybe it's as simple as "I called you ladies, but you're not, that makes it funny, haha!"
I've generally found total honesty to be a good thing. sometimes I temper it with concern for the feelings of others, but for the most part, I tell you what I really think
SWPadnos: You damn engineer!
JymmmEMC, shut up, I'm tired of your bullshit
err - hahaha
SWPadnos: Liar, no you're not.
SWPadnos: I've always been interested in the question of when it's right to lie, and I don't know the answer yet
so, I just got a catalog from "New Pig", address: One Pork Way
maybe I don't mean "right", maybe "best"
and they have nothing to do with food
cradek: If the lie would do less harm than honesty
what's New Pig?
they make spill prevention, containment, and cleanup products
cradek: I have a hard time at work when something happens. We're told to say "We're looking into the issue", when I'd rather say "There was a HW failure, and it's been looked into", but eh, that's what the boss wants.
or there's a fiber cut, etc.
that isn't one of the times where lying is good. that's strictly a monetary / image thing
ie, it serves the person telling the lie
SWPadnos: It's strictly an image thing in this case.
I guess that's the distinction - if you tell someone something that you know to be untrue, it's not justifiable if it benefits you. It may be justifiable if it benefits the person you're telling
But in a way it's true, We are looking into the issue, but I hate the part of "I'm not sure what the problem is" when I do know.
sure - a company lies to prevent liability issues, and to prevent reduction of stature in their customers eyes
In the end, when someone pushes the issue, the admin's tell them what the problem was anyway, so eh.
either of which is good for the company, not the customer
On the flip side, many custoemr will rant big time if they're down 1200mS
WHAT HAPPENED?! We were down, our business lost $100,000
"It's the internet fscker, chill out already"
if that's true, and their SLA says they shouldn't be down for 1.2 seconds, then they have a point
there is no SLA
otherwise, they canshut up too :)
Sometimes I'd love to be able to say "Don't get your panties in a bunch", but well you know, this is the SF Bay Area.
I wonder if I should be eating this chocolate covered donut just before going to the dentist
SWPadnos: have garlic and a butterfinger bar for good measure
hmmm. I have some pickled garlic
it's oh so tasty
chocolate covered garlic donut?
I think that might be icky
yeah, chocolate ginger is good, but maybe not chocolate garlic
Like buttered microwave popcorn? Read this or die http://seattlepi.nwsource.com/food/330230_popcorn05.html?source=mypi
however, I'd probably try it
JymmmEMC: microwave popcorn is nasty and smells like a chemical factory
cradek: the butter flavor is toxic
just foudn out
couldn't everybody tell by that stench?
I like the smell personally
But, I like the smell of a angle grinder too =)
the "butter topping" used in theaters has some of the same ingredients as Dektol - a film developer
that's probably better for you
or a open camper fire/fireplace
SWPadnos: a common additive to real butter is the same thing used to melt ice on sidewalks
I hear it's conposed of two poisons also
Doritios contain a chem used in anti-freeze, Twinkes contian a chem used in floor wax.
Margarin is one molecule away from plastics
[16:00:43] <cradek> http://www.boingboing.net/200709041918.jpg
"The Post Office. Bringing the technology of the 1890s to you, today."
SWPadnos: Well, they finally did catch up.... internet postage
hey! whatis was actually helpful for a change
hi all, anyone good at debugging stepper motor/drive malfunction ?
I have this microscope stage that just stopped working
the motor has 4 wires, and I can measure two coils around 5 ohms each
that means the motor is OK?
if it's a stepper that could be fine
yes, it's a small stepper 40x40mm square case
the problem is it just jumps back and forth between steps
it's a two axis stage, and the other axis works fine, so I'm going to try to run the strangely behaving motor with the drive I know works
did you check the voltage across the coils as you move one step at a time?
that sounds like a decent idea
no, not if the motor killed the drive
did you check for the coils shorted together
the coils are around 5 ohm on the motor that doesn't work well
I didn't measure the other motor
I mean the coils shorted to one another
nope, coil to coil the multimeter shows something like 10 Mohm which is reasonable I think
I would just replace the steppers and drives, but these genius designers have used the shaft of the motor as the leadscrew, so it's not a standard motor-shaft>coupler>leadscrew thing
sounds like one H bridge is probably dead, you should be able to test that easily
can I measure the output pins of the driver without load? or should the motor be connected
I think either way would be fine
is there a stepping diagram online of what I should see?
I guess the four wires switch between -V and +V in some sequence (or maybe 0V and +V)
the coils will alternately toggle polarity
this might be in the emc manual? type-XXX stepping?
there are pretty diagrams that document all the various stepping outputs stepgen can generate
ok, I'll try to open the box and scope the output pins, thanks for the help.
I worry more about Dihydrous Monoxide
its in everything and is unregulated
skunkworks: Teflon frying pans is another biggy
we lover or thick bottom stainless pans. Love love love
cast iron baby
I got my #8 and my #10
If I was a big bacon and egg eater - I would do cast iron.. :)
great for campsite too
good for pork chops, stews, etc
My pasta pot is SS with a thick bottom.
These are MINE, not my gf's.
cradek, jepler: one more thing about my stepper: when I command full speed, and give the motor a push by hand it does rotate at a good speed. would that be consistent with one part of the H-bridge being broken? i.e. at full speed inertia takes it over the 'bad spot' in the cycle and the motor can rotate with only half a H-bridge functioning
awallin: is there a heavy load on your machine?
JymmmEMC: no, this problem occurs at zero load also. one motor just jumps back and forth instead of taking steps in one direction
one specific motor, so if you swap motors does the issue follow the motor or the driver?
one specific motor? so if you swap motors does the issue follow the motor or the driver?
right. I should swap the motor with a known good driver and test. it's just an akward place to work and the short small wires will have to be cut and soldered etc.
why soldered? Maybe try using a crimp connectors
or even a screw terminal
yeah, I should do that...
cheap easy solution, especially for testing purposes
awallin: I like using these type of terminal strips http://www.allelectronics.com/cgi-bin/item/TB-24/755/24_AMP_FEED-THROUGH_TERMINAL_STRIP_.html
yes, but this is a microscope stage with about 25x25mm of travel, 40x40mm square steppers, so the wires are _tiny_
smaller than 22ga?
maybe 1mm diam
I'll go and see what I can do now... bbl
well, maybe solder the wires to some CAT5 cable if you have to.
i tested the trajectory planner and try to modify it
do you have some interesting results?
cncjunior: actually what did you change?
i am working on a optimization project and emc2 is a very good platform
optimization of what
at first was feed rate and spindle speed versus time/cost/productivity
now is interesting the precision
i want to try not only a simple trapezoidal profile but a multiple trapezoidal
jepler: i read your explanations on this topic and i found also some examples in matlab
the link from http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Simple_Tp_Notes
to this examples don't work
stepper problem: I now made a cross-cable to connect motor1 with with drive2 and vice versa. it seems the drive is at fault, both motors work flawlessly with a good driver
cncjunior: that wiki page is mostly by me
awallin: At least you've narrowed down where the problem is.
awallin: what drivers are you using?
cncjunior: you mean this? http://wiki.linuxcnc.org/uploads/test.m
and this http://wiki.linuxcnc.org/uploads/runcycle.m
JymmmEMC: yeah, only problem is that this is a commercial microscope stage and production of it was stopped some 5 years ago
awallin: ok, but what drivers does it use?
skunkworks : yes, these are the examples
the wiki needs to be fixed
skunkworks: thanks for the links
cncjunior: I remember that working through simple_tp was not _that_ hard.
it's the blending that's hard. cradek probably understands it, but it's not documented anywhere
JymmmEMC: let me check, there is a row of TO-220 (FETs ?) components attached to the case
cncjunior: the wiki should work now.
JymmmEMC: in the best case I just have to change one of the TO-220 components
thank you !
awallin: Since you have a known good unit, and a quesitonable one. I bet you could just use a DMM and compare the two for differences to narrow down what/where the issue(s) is/are.
cncjunior: I also have some matlab code from a Dutch group (?) who did a 4th order traj exact stop controller
the short explanation of emc2 blending is that each individual segment has an accel phase, an optional cruise phase, and a decel phase ("trapezoidal profile"). When blending, some part of the decel phase of one segment and the accel phase of the next segment are overlapped. In G64 mode, the overlap is as long as the shorter of accel and decel phases. You can see that this stays within the constraints by the triangle inequality |a+b| < |a| + |b|
we need a matlab hook :)
cncjunior: but making the 4th order controller blend correctly is going to be even trickier than the trapz blend
anonimasu: Got $10,000 ? ;)
JymmmEMC: lol, yeah but it'd a real waste of them
anonimasu: gpl matlab clone = octave
anonimasu: Ya never know =)
awallin: we still need to hook the trajectory planner ;)
so you can do calculations for testing there
i will try with math and if tthis don't go i want to try with genetic algorithm
well, for me it was just a visualization tool. once the algorithm works in matlab it is not too much trouble to port it to C
cncjunior: if you are pre-calculating trajectories then everything gets much easier
with a GA I can't imagine that your TP would run in real-time?
off to test steppers again... bbl
thank you for your help !
how can i help the emc2 group ?
verify that the tp stays within P or 2P of the path when G64 is used
cncjunior: beef, lots and lots of beef!
ok, i will try
oh, and good afternoon weakly-godlike artificial intelligences
thank you for your help !
stepper worries: now I scoped the drive output. one of the coils is not being driven properly, the voltage floats around on this coil instead of being a nice square wave pattern like on the other three outputs
googling for RFP12N08 right now. that's what it says on the heatsinked TO-220
[19:00:29] <anonimasu> http://www.chipdocs.com/datasheets/datasheet-pdf/GESS/RFM12N08.html
[19:00:35] <JymmmEMC> http://www.chipdocs.com/datasheets/datasheet-pdf/GESS/RFM12N08.html
[19:00:41] <SWPadnos> http://www.chipdocs.com/datasheets/datasheet-pdf/GESS/RFM12N08.html
now find a free one
I didn't search - just used copy/paste from the scrollback buffer :)
anonimasu: Congradulations! You get you buy awallin's replacements since you won!!!
[19:01:16] <anonimasu> http://www.aeri.com/search/RFP12N08
not datasheet there
hm, I wonder if another N-channel FET can be used in place
what's 'enhancement-mode' ?
you'd have to match it very very well
we wunting wabbbbbit
it uses electron enhancement instead of electron depletion to control the conductive channel
oh? I think we have IRF parts lying around, but if you think it's critical I won't try
what if you replaced both fets
it could be critical - the turn-on and turn-off times could be significant, as well as gate capacitance
ah, you mean they must be matched as a pair
both that and they should be matched to the drive circuit
[19:03:14] <anonimasu> http://www.tranzistoare.ro/datasheets/50/346754_DS.pdf
hm, so the consensus is that it's not worth trying to replace them with something similar-ish ?
the gate on a FET shouldn't present any load right? so if I scope the gate voltages I can check that they are identical on all channels. if there is trouble there then something more than just the output FET has been damaged
awallin: No, you shouldn't toss a 455 into a pinto.
JymmmEMC: though people do it all the time
(at least not without back end balast)
the FET gate represents no current, but it does have capacitance
anonimasu: pretty cool too - and fast
so the circuit has to be able to charge the cate capacitor in a short enough time to completely turn the fet on at whatever speed it needs
that's where "speedup caps" are used - they have enough juice to quickly dump into a FET gate
ok, I'll scope the gate voltages next. if they look OK on all channels but only this one pair has the wrong output voltages then I can be pretty sure that the FETs on the faulty channel are bad
awallin: what if you check on the good driver..
this is also why CMOS devices use more power linearly as the clock rate increases - you're charging/discharging lots of capacitors at each clock, but in betwee, there's almost no energy loss
scope voltages part per part..
and write the diff down
going scoping now... bbl
anyone have any leads for finding a service manual, troubleshooting, or at least wiring diagram for the CSR Contraves NC400 servo amps?
uh oh, problems with one of your amps?
I think it's outside the amp itself, since I swapped two and the problem stayed on Z
it faults sometimes, but when just sitting there or at startup, never when in use
I have no idea what all the "fault" might mean (overvoltage? undervoltage? overcurrent? open brushes? bad tach?)
so, I'm not real sure where to start to figure it out
back. I scoped the gate voltages and they are nice square waves on the FETs that work, but totally dead at 0 V flat-line for the FETs that don't work. Does that indicate that the FET is wrong or the drive-circuit is wrong? (or both?)
there's a RFP12P08 for each of the heat-sinked N-versions. Can the P-channel FET be the driver for the N-channel one?
the signals on the P-channel FETs look slightly better. The gate voltage is a perfect square wave, and the output is ok-ish
maybe the H-bridge consists of two N-channel FETs and two P-channel FETs.
Now I'm just hoping that the N-channel ones have failed and can be replaced
yes it definitely takes 4 switches to make an H bridge
although I'm a bit worried that there is no signal on the gate-pin of the N-channel FETs
I was just about to ask how many of each type
how easy to remove the N-channels from the circuit and see if the gate voltage is OK then?
but maybe the N-channel ones have failed in such a mode so that they ground the gate?
sometimes FETs fail "melted" (everything shorted together)
jepler: yes, that's a good idea, the fast way is ofcourse to use cutters...
I have failed to ask the internet whether gate-source short is a common way for a fet to fail ..
* cradek mumbles about horizontal output transistors
is there a resistor on the gate? if so, you could try measuring on the opposite side of that resistor
it's a two-sided pcb so de-soldering might be a bit tricky. I'll try that next
that usually works fairly well
can't see a resistor on the gate
heat + solder sucker
snip first so you can do one lead at a time - very easy then
yes! cutting the faulty FETs I recover the gate signal! nice square wave
so, 100eur to the first person that delivers two RFP12N08 FETs to my office tomorrow morning at 8 ! anyone?
I need to get home, not healty being at work until 22:30...
good night awallin
good luck finding your replacement parts too
gnight everyone, thanks for the help. I'll take a pic of the damn thing tomorrow so you can all see what you've helped fix :)
[19:51:54] <skunkworks> http://cgi.ebay.com/Contraves-Cinetheodolite-Missile-Tracking-Telescope_W0QQitemZ230168085532QQihZ013QQcategoryZ74930QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
skunkworks: now that's something that's handy to have if you live in Iran
I think quite a few places
skunkworks: it would also make for a great platform for star gazing, to bad the price tag is a bit out of my reach
the unit will operate on 110VAC, 30 amps (household current)
uh, not my household...
yes it would - I bet there is no backlash in that mechenism
Best Offer: $200
that would be fun to take to bit to see how it worked
[20:18:58] <JymmmEMC> http://i7.ebayimg.com/01/i/05/e1/e5/e8_1.JPG
I do collect theodolites,....but
will a pc do
a 1973 mainframe at that anyway so a modern pc is probably overkill
How you guys checked out REBOL ?
meh. REBOL has a license that makes it not worth my interest. "You cannot modify the software" -- http://www.rebol.com/license.html
hmmm. that could just be a statement of fact (for most people)
I assume they mean "may not", smartass
never wanted to be a lawyer ;)
license aside... it aint bad
SWPadnos: VIEW.EXE http://www.rebol.com/view-platforms.html
SWPadnos: Tell it not to install, then click it again
um. no, thanks. not at the moment
[20:42:42] <JymmmEMC> http://www.rebol.com/oneliners.html
OD NOT WANT
DO NOT WANT
Leapyear?: function[Date[date!]][Year][Year: Date/year any[all[0 = remainder Year 4 0 <> remainder Year 100]0 = remainder Year 400]]
(defun date-leap-year-p (year) "Return t if YEAR is a leap year." (or (and (zerop (% year 4)) (not (zerop (% year 100)))) (zerop (% year 400))))
: LEAP ( year -- flag ) DUP 4 MOD 0= OVER 100 MOD AND SWAP 400 MOD 0= OR ; \ nonzero flag means a leap year
I think we can leave forth out of this
'0 = remainder Year 4 0 <>' sure made me think if forth
except I guess it's still some weird midfix notation
oh I guess I grouped that wrong
"0 = remaider Year 4" and "0 <> remainder Year 100" are the two args to all
your favorite language sucks, particularly if it's REBOL
You guys realize these are all GUI, and not just command line
though it does CLI too
I suspect jepler has his reasons, but it reminds of the common feel ppl have of flash... annoying banner ads. Flash is mouch more that that.
JymmmEMC: when I have dozens if not hundreds of decent languages that are Free or standards-based, why should I dick with a language which is neither?
jepler: If it serves a specific purpose, I feel the right tool for the job w/o (reasonable) prejudice is justifiable.
lawn mow fence paint post fix (ancient forth joke )
do whatever you like; I'm just explaining to you why I am uninterested in REBOL.
jepler: (ok, it's M$) But for GUI, this seemed impressive to me.... http://musiclessonz.com/rebol_tutorial.html#section-6
jepler: Not to worry, I'm not trying to convert you to the "Dark Side", just expressing my enthusiasm for what it is, specifically GUI
[21:58:30] <jepler> http://shootout.alioth.debian.org/gp4sandbox/benchmark.php?test=all&lang=rebol&lang2=python
anyone still have the plans for the "small home-built THERMO-nuclear device" that i found posted a while back???
sorry.. gov confiscated
check the logs
have most ALL materials and proceses in place... just wante to re-check some minor things...
like the plutonium
nahh... that comes from Uranium.... we've mine that....