sigh ... bought a new phone
shoot me now.
probably the worst bit of technology I have ever EVER had the misfortune to own
this is perverse - I hacked emc to have the same quirks my bridgeport does, so I can check and preview my gcode
instead of locking up, though, emc gives an error message!
Hello guys i´m back
I was dinning
well, i restored the config files, it solved my problem
Thanks a lot!
OK - here is an off the wall question.
Could a rotary axis be hidden/embedded in classic ladder - such that it would not show up in a gui EMC2 front end.
sure, you can have extra pid loops/stepgens/etc in hal that emc itself doesn't know about
imagine a rotary table used as a tool changer turret
the application is to put a servo on a rotary table - mount the rotary table on a lathe and bolt a turret front on it
hehe java batik now running squiggle & the hektor simulation, i can get file writes with the data now for halstreamer.. but cpu is at 100% and it's taking forever to trace 1 pass ( well it can run like a rendering farm, while i sleep ;)
so - use a 4" on a 7x12 lathe - might be able to fit 8 tool positions - but might only work with 6
Need to model it up in Solidworks
hell I should do this, I think I have a rotary table that could work
first tool changer on a sherline lathe, I bet
nice part is you could make both turrets - and swap them off the rotary table.
Seig - cast iron
7 x 12 from Homier is $299
well I already have this one and it's already cnc
I can buy one of those Rubbermade carts - put all the PC & Powersupplies etc on the lower shelf
lathe on top
cradek - we have a monster Hwacheon lathe with an oversived turret - the design allows the tool Centerline to be below the turret centerline
Mmmmm minilathe with tool changer
some lathes take CAT-40
and have an 80-100 slot toolchanger
i think this is a brilliant idea, ablit slower
on the lathe tool part or for live tooling?
SWPadnos: you awake?
almost asleep. what's up?
anyone up for a RS274X -> G-Code demo video?
reading about sony's psychic phenomena research
[12:10:33] <fenn> http://www.remoteviewer.nu/index.php?name=News&file=article&sid=723
yoichiro sako has 172 patents
im going to machine some stuff..
a housing for a sensor
skunkworks_: Estimated delivery: Aug 28, 2007
horse drawn UPS
it seems to take forever for stuff to come from CA
its horrible waiting for a new toy
just be glad it's not USPS
USPS to the UK is by boat
actually I have had really good luck wht usps priority. I always get it within 2 days after purchasing it.
archivist: you should move over here to what we like to call the "first world"
usps international is always > 1 month and can take up to 3 months, I suspect they ship the stuff around the world before it gets delivered
jepler er this is the first world
archivist: american rednecks disagree
to bad the second world does not exist anymore
was that france? ;)
1.st world == Western Europe and USA
2nd world == USSR and allies
most of which are now just considered to be a part of "Europe"
don't worry, I am saying it tongue-in-cheek
1st world UK, 2nd world ... the rest
in fact I'd love to get back to the UK for another vacation
archivist: and USA is just a colony right ?
reality: first world = middle east, China, and India
yah - who do you get when you call for tech support..
well, I'm speaking historically, but I guess everything goes full circle at some point ...
Egypt = middle east, but you forgot that Indian and China both have recorded history going back around 5000-6000 years as well
(as do the Israelis and Egyptians)
there was a neat show on how they had ways to drill for oil, gas and salt thousands of years ago.
03jepler 07TRUNK * 10emc2/src/hal/drivers/pluto_step.comp:
add "speed range" to pluto-step, which allows a trade-off between resolution
at low velocity and highest speed. recompile firmware.
03jepler 07TRUNK * 10emc2/src/hal/drivers/pluto_step_firmware/ (5 files):
add "speed range" to pluto-step, which allows a trade-off between resolution
at low velocity and highest speed. recompile firmware.
"floating point" step generation ;)
cradek: I haven't fully tested the "trade-off" yet but I wanted to get it out of my local tree
is the divider in powers-of-2, or divide-by-n?
ok, that should wok nicely (either would, but 2^n should give better range)
FWIW, the G-Rex does the same thing
they didn't choose a big enough FPGA that there was room for "big enough" accumulators?
that would sure have been my first choice
hmmm. actually, the G2002 did that, I'm not sure the G-Rex still does
well, the FPGA is 50kgates, and Mariss says it's totally full
which I believe, because it does have a lightweight microprocessor embedded in it :)
that could do it
the 6 step generators and motor encoders, plus SPI for the digital I/O also contribute
the updated docs including an explanation of "speedrange": http://linuxcnc.org/docs/devel/html/man/man9/pluto_step.9.html
hmmm. 76 Hz minimum may be a little high. if there's a spare bit or two for the divider, I'd use it
when the speed is very low, emc will switch between commanding zero speed or commanding minimum speed
oh - this is a load-time parameter, not changed on the fly (?)
ah - then it isn't like the G-Rex :)
it also applies to all the step generators
doesn't anything slower than one step per servo cycle become unnecessary, since emc can generate one step at a time in that case?
there's this weird interplay between the servo rate, the minimum step rate, and fractional step feedback
you still want to be able to generate lots of steps, but a G1 X10 Y0.1 F10 needs very low rates on the Y axis
you can get as low a step rate as you want on average, but the steps are issued at a somewhat irregular rate
jepler: are you planning on getting one of the parallel port mesa cards?
skunkworks_: whether I'll actually put it to use is another thing altogether
huh I wonder what I changed
what are you going to do with all those gates? ;)
you will be swimming in them
I was able to enlarge the velocity registers from 11 to 15 bits and it still fits
I should say, I wonder what *else* I changed
I guess I mean from 12 to 16, the new size is reg[15:0] in vhdl
increasing it beyond 16 bits requires more changes..
but if I can enlarge it 4 bits and still fit, that gives the same benefits as the speed range selection
yeah, it almost worked OK before...
if the irregularity was 1/16 as big (4 bits more resoultion) it might not be noticed
howdy Ray - long time no see
argh - time to restart Mozilla. I've got a stuck window on screen - it loads pages fine, but doesn't respond to mouse/keyboard events :(
how do I change my cloak here.
unbutton it at the top ... :)
* archivist covers his eyes
it's just a cloak
that is an odd question
I think you have to ask a freenode admin, but I don't know how to find them
or read the FreeNode faq but I never worked it out
the FAQ says "/quote stats p" but that lists 0 people
Hallo , I still have a problem witth running my robot on EMC. in manuel mode I can run it as fast as i want, in Auto or MDI i could take F.. as high as i want, but my Robot is not nearly as fast as in Manuel. It work but to slow. Is it possible that the Cycletimes or periods had to be Change ??
you mean jogging the machine is fast - but mdi or auto is slow?
could you pastebin.ca your ini file? - I don't know if I can help..
could it be the limits in the traj section?
[traj] section of the ini - I would think jogging may only use the limits for each axis/joint but coordinated motion would first limit by the traj and individual axiss.
(I am not an expert)
same says alex, but it doesnt change anything , move them up or down makes no sens
skunkworks_: that sounds right to me
plattschnauze: could you pastebin.ca your ini?
fenn: thanks - Sometimes I make stuff up that sounds good to me ;)
Is there a setting in EMC for a spindle speed multiplier? My spindle has a few belt options and I wish to just leave it on a high ratio.
I will be putting an encoder on the motor side aswell, so the encoder will need the mulitplier aswell to display the correct spindle speed.
i think you would have to cook something up with halvcp and hal to multiply the spindle speed by the belt ratio
the scale component should do what you want
or if you didnt want to use halvcp you could just set the ratio with halcmd (through a shell script probably)
ok I will look into that stuff, I havent started any software yet, just hardware.
I guess my other option is to just find the same pully for the motor and run a 1:1 ratio
Adam-alt: different belt ratios is a good thing, you should keep it
adam-alt: is it a toothed belt? if not - you cannot put the encoder on the motor side of the belt. There will be slipage
(like a timing belt.)
as long as he's not using it for position it should be fine
if the belt is tightened properly
right - I figured it was for rigid tapping..
No tapping just for speed display
its just a V-Belt
my motor gives me around 3400rpm, on a bridgeport copy - I think that will be enough for the machine... I dont want it to vibrate too much.
i dont really get why they still make machines with v belts
Yeah I like a geared head myself
i'm a big fan of round-tooth belts
its so simple and works so well
I've been spoiled with Maho-Deckel's now I have a bridgeport copy that I need to retrofit from an 1980's control.
the dialog control?
there was a brief flash of interest in making a dialog emulator for emc
lets see if it went anywhere
I am upgrading from an anilam crusader
[17:56:18] <skunkworks_> http://www.electronicsam.com/images/me.JPG
anyone know if you can extend the shaft of an induction motor out of its rear-end?
skunkworks_: is that from image-to-gcode?
mm I could use a beer
I like how you can read the guinnes
jepler: do you see the terminal in the lower left? Is that error normal? I get it every time I run image to gcode. It seems to work ok.
jeez is this ass-backwards or what: http://www.practicalmachinist.com/ubb/ultimatebb.php/topic/6/2104/4.html#000132
skunkworks_: huh I don't know
jepler: trunk. 3 different computers.
like I say - it doesn't seem to effect anything
that I can see
with files that size - the plot and preview need to be shut off.
on a 1.7ghz
how big are the files?
I don't know - how can you tell? I did image>gcode without it making and intermediate file.
hmmm. I wonder if file | properties will tell you
or file|info or whatever it's called :)
I wonder too
yes - well it is odd
56K file size -- 74k lines of gcode.
isn't that odd?
oh, and are you saying that the AXIS plot/preview is slow, or the image-to-gcode process is slow?
that is odd. I'd expect more lines of code for that file size
err - wait a minute. I'd expect more coffee to help me analyze that :)
I think 56K is the size of the image file, not the size of the generated gcode
out of curiosity, where is the G-code in this case?
SWPadnos: somewhere in /tmp
ok, so there is a real file then - not just a pipe
that's good (and necessary, AFAIK)
heh - it says it should take 12min to run.. It has been running for 15min and maybe 1/4 done.
that is at 300% feedrated
skunkworks_: yeah I've had complaints that the estimate is not too accurate
what's the programmed feedrate vs. the [TRAJ] max?
I think I had the feedrate set to 100ipm - the max feed per axis is 200ipm
it dips down to 12ipm as it is scanning with an accel set at 25in/s/s
the estimate assumes that the F requested is actually achived, among other limitations
I set the taj section to 5.777 in/sec
which is 346in/min
that is sqr(axisa^2+axisa^2+axisa^2)
I guess with lots of small segments, it would end up much slower
sqrt(), hopefully :)
skunkworks So, you still gonna sell it?
[19:21:38] <skunkworks_> http://www.electronicsam.com/images/eme.png
[19:21:58] <skunkworks_> http://www.electronicsam.com/images/me.JPG
and the final result?
what is a good material to cut that in? to backlight it. white styrine?
dont forget to lightly sand and stain it after
I was planning to cut it in plastic
and backlight it. what do they call it - lithograph?
I've never full sheet carved plastic like that
usually I use carved / uncarved areas to contrast
whats the max Z ?
that picture is .1 depth
poor lighting made this pic unfocused.... http://farm1.static.flickr.com/84/206158222_cb5c34978c_m.jpg
that's clear acrylic
but like I mentioned, I use the carved / un carved areas to contrast
[19:29:49] <skunkworks_> http://www.cnczone.com/forums/showthread.php?t=41845&highlight=litho
but isn't lithograph 2D line art type?
something like that
corian is dense
got any delrin?
I was going to steal some styrene from work here or what ever else i could find. Just to try
I have some abs I think also
Eh, too soft I think
you could try a small sample though
I might have some colored lexan or plexiglass
oh, you could try counter top material
name slips my mind
yeah, that's it
or something like that
Be sure to use a dust maks BIG TIME
and not those $0.99 ones either
those are not dust masks ;)
the dust is just like asbestos
you could try those cutter boards too
hmm - I think I have an old one. :)
looks like that picture should take about an hour to cut.
how would somebody go about calibrating a 3 phase motor with a VFD to give a correct spindle rotation?
rotational speed i should say
you mean CWvsCCW?
all you have to do is swap any 2 of the 3 wires.
no I mean correct rpm
if I put 2000 rpm into EMC and the VFD gives me 1500, what is a good way of measuring that?
if you attach the encoder - you can look at the velecity in hal.
I am not sure if I can attach an encoder to my motor, I would have to do a swap...
I dont have any physical room to attachit
the vfd should give you rpm in some mode
I would think.
in its display
maybe there is a handheld tach that I can shoot the spindle with and then adjust the VFD accordingly.
technically - the rpm of the motor is a direct relationship the the frequency of the 3phase. with the normal 60hz 3phase the motor spins say - 1800rpm. at 30hz it will run at 900rpm. take away a little slip.
at 60hz I am at 3400 rpm then.
[20:05:26] <cradek> http://www.sherline.com/rpmgage.pdf
works best with a neon bulb in a darkened room
I used this to calibrate my lathe spindle drive
SWPadnos: bona sera
not quite.. I'm in italy right now :)
ah - mea culpa :)
c'est pas d'probleme
not much - still trying to get the best cross-section of cost, performance, and availability for this analog board design :)
[20:20:09] <cradek> http://earz-mag.com/2007/March/Why_Ban_Comic_Sans.gif
* alex_joni has free wifi for tonights hotel
[20:23:45] <alex_joni> http://www.hvenice.it/homeenglish.htm
that is always nice.. Normally around here - wifi is free in hotels. Not over there?
it even has "reduced electric fields"
skunkworks_: not always
and "Air-conditioning with 2 Totems functioning with methane without in house electric fields."
sorry alex, it says the hotel is only for me
it's a shame though, because I'll probably never go there
I'll give it back tomorrow
seems like a bad business plan.
is it a big problem to mix servo drives and servos? I have stock X & Y drives and servos. I wish to add a Z on my quill and its going to be a big pain to match components up.
it's not a problem
allright, figure I alraedy have a glass scale on it might aswell attach a servo.
though im not sure how well using glass scales togther with emc works
a position loop with any backlash is trouble, it will oscillate
I wonder how commercial machines do it..
a position loop around a velocity loop is ok, but not a typical emc configuration
they do have backslash compensation(atleast the heidenhain ones)
yeah it has a tiny bit of backlash
of course emc can do whatever you can manage to tune
probably all others too
so what could I do for a quill feed if it has backlash?
put an encoder on the motor
cradek: would it be hard to use a linear encoder to correct for positioning errors?
anonimasu: well you could definitely make compensation tables by reading from your scales
cradek: feed it back into the loop, to correct for pitch errors and such..
so have the encoder on the motor and the glass scale aswell?
but that's not something anyone has done before..
what I know of..
then I am toast with that
that won't be easy to tune, especially because you'd be the first EMC2 user to do it (as far as I'm aware, anyway)
there's a reason cnc motors typically have encoders on them...
cradek: well, most higher end machines have linear scales too :9
and that would be fine in emc, you could have two pid loops per axis
but you have to tune it. it would take patience + talent I think
couldnt you feed the current_pos - actual_pos
using the scales to generate compensation tables should be easy. if they have a DRO, it's *very* easy
cradek: im just speculating..
well.. good night guys, been a long drive till here
still have 1000 miles to go
bona sera Alex :)
hope the rest of the trip is enjoyable
So I could skip the encoders and stick with the scales and generate compensation tables for them?
yeah, it's fun
you could skip using the scales and use encoders and generate compensation tables for your axis with the scales
no - you can use the encoders, use the scales during setup (and periodically later on) to generate compensation tables, then forget the scales for the most part
see you cradek
SWPadnos: though that limits the usability of scales..
well, they're limited by (a) slow readout rates and (b) potential backlash between the motor and the feedback
I think we said the same thing anyway :)
SWPadnos: last time I checked the cheap scales would do very high speeds..
if they have quadrature output then that may be true, but I haven't really looked into scales enough to know
my scales are quadrature
some have quad out some have vpp ~
ancient but quad
you could always try using it..
yah - you really have nothing to lose
if your machine is tight - it may not even be a problems
im kind of wondering how the commerical machines do it..
because it works well :p
if they dont have any backslash or if they are very good at compensating it
Yeah well the quill is just a gear right so there is going to be backlash and it will only ge tworse over time....
or it is just geared...
cradek: changing the g64 from p.001 to p.002 changes run time from 60min to <45 min :)
i think machines with linear scales usually have tachos on the motor
gives a better signal than encoders anyway
I think that is how Adam-alt's machine is setup right now with its original controller
you should be able to just feed the tach output into the motion planner velocity feedback? but emc doesnt seem to have that anymore
yeah the X&Y servos have tachos...
the tachos are delt with on the drive boards.
I can get a z-board and a suitable servo for it right now.
i am worried about that oscillation that somebody was talking about earlier though. Damn quill has a tiny bit of play.
I like this guy's attitude! http://www.freesteel.co.uk/wpblog/2007/08/a-message-to-lawyers-contemplating-acts-of-legal-harassment/
It would be nice to have more time for mill building:)
yeah, same thing here. haven't got the DC servos installed yet... will take another 4-weeks at least before its done
I have now replaced all my steppers with servos
the cnc-fi googlegroup has been quite dead for a while. do you know if the finnish cnc-crowd hangs out somewhere else on the net?
ok... although I don't have much time for irc...
Martzi: are you good at programming? maybe you saw that I made some tests with a CAM-algorithm
hmm, I am not sure if I have seen
[21:46:21] <awallin> http://www.anderswallin.net/2007/08/drop-cutter-in-c-v2/
I have some programming knowledge
hopefully I'll have time to add some more algorithms during the winter
when the program does something useful I'm hoping more developers will get interested and help out
I can help
At least try
good. are you working under windows? the idea was to code it in .NET so that it would run under Linux/mono, but there are some issues with TAO- the OpenGL bindings which make it not work under linux right now
I can use both Windows and Linux
Perhaps it was my clone
yeah, that was your alter ego
What is needed to put the building environment working?
the free Visual Studio Express (C#) from microsoft, and the Tao (http://www.taoframework.com/).
that's under windows
for linux there is mono, monodevelop, and tao. but there are issues with tao as I said
and you need to have .NET 2.0 on yuor windows machine, but most modern machines should have that
i have it
Is express 2008 Beta 2 Ok?
or should I use older?
.Net 3.0 isnt compatible with mono
too much work and traveling... not much CAM getting done.
same with me at the moment
I feel a bit bad about it
no worries:) ... as we planned this is a 6-month thing to get to v1.0
dont ask... two weeks ago in Austin and Boston. Now I'm flying helsinki->JFK on saturday. 7-8h on a plane is "fun"
what a pain
Ok, everything is installed and I can start coding;)
good. nice. if you do something significant and it doesn't brake everything let me know your gmail account and I can add you to the googlecode project. that way you can commit your changes via svn
I need to go to bed
do VFD's generally protect from back emf?
some have terminals for connecting a dump resistor. I guess it's used with really big motors.