see any car transmission
straight cut are too noisy
dave_e: I'd love to see he's router at work.. my mill wont do faster stuff then 2m/min
it's quite a bit away from it..
*checks datrondynamics.com again*
anonimasu_: you need to see may laser :)
I'm good for about 5 m/min
84m/min rapids :)
some of the more modern lasers hit 200m/min
accel is 20 so nothing is instantanous
ok... i just hate friggin site's that dont have pricing on them!!!
robin_sz: the datrondynamics machines apparently has 600W spindles..
sure...grab you first then lay the price on you ;-(
bait and switch
dave_e... ya.. pisses one off :)
ok... another question.... y would i use servo's instead of stepper motor's (yes... lot's of stupid q's)
well, whats your budget?
how big is the machine?
robin_sz... i would say zero... but i dont think i am getting it for that :)
aslittle as possible
steppers is the way to do
and some geckodrives
the machine is going to be 3m X 2m <--- it's kinda huge
if you had lots of budget, and were more concerned with performance > price
ya.. those geckodrives look quite awsome
servos would be the way, but steppers will do just fine :)
pushing towards servo territory at 3x2
a 2m gantry will be kinda heavy
triple stack NEMA34s I figure ...
1200 ounce/inch stepper motor
whats an ounce?
got any specs for thoose motors
how heavy will the gantry be?
are you american?
no... not american.. but all the sites are
which doesnt help
well, stick to Nm so the rest of us can understand
[00:07:12] <Yuga> http://www.campbelldesigns.com/stepper-motors.php
<--- bottom stepper
u think the bottom botor would be able to do it?
possibly too big
nema42, will have a lot of intertia
the inertia is horrid too..
if you want speed you will have to match the inertia with your gantry weight
intertia matching is essential
with a router like that you would have to dual drive it anyway..
robin_sz: correct me if I am wrong..
* robin_sz hates dual drives
ok... so you would rather recomend 2 640 oz/in steppers? one on each side?
twisted metal, bent gantries :)
or well a single stepper that drives both sides
with a connecting shaft
yeah, corssover axle
anonimasu... the 1200oz.in one that drives both sides?
bit of toothed belt gearing ...
think that would work best?
and a decent motor
anonimasu... not even able to guess?
liek a Sanyo Denki nema 34, 3 stack ... does similar sort of torque with half the intertia
you need to run it through "motioneering" to really find out
I'd run motioneering and calc it
* robin_sz hi fives anonimasu_
not a clue what motioneering is
wait a sec
[00:13:16] <anonimasu_> http://www.motionvillage.com/motioneering/app_engine/
friggin hell... wouldnt have a clue how that works :<
oh, download & play with it
anonimasu... doing so
well, its easier to try and figure it out on the screen
it's pretty self explainatory :)
than it is to rework the router two or three times
calculating it by hand is tough
right ... bedtime
friggin hell.. the link on the site dont work
[00:17:44] <anonimasu_> http://www.danahermotion.com/Service_NA_Frame.asp?target=../../Software/Kollmorgen/Motioneering/Motioneering_App_Engine_page.htm
cool.. that one is working.. the other one wasnt
anonimasu... got any recomendations on what motor's i should look at?
and what board i should use to connect the gecko drives to my pc?
the paralell port..
but I am not sure what requirements thoose steppers have
guess parralell is the easyest?
if the voltage¤t are enough
gecko drives goes up 2 9 amp's i think.. so i'm sure i will be save
was looking at this board. http://www.campbelldesigns.com/breakout-board.php
hm, thoose are large steppers :)
i lied... 7 amp's 80 v
2,3,4,5 PHASE, CLOSED LOOP
stepper motor's... what the hell are they on about?
jeez Yuga don't listen to robin
he loves to spend money on stuff
get a $75 hobbycnc board and little steppers and learn on that
so when you mess up you don't blow your $500 driver electronics box
and a 5 axis router isn't really such a big deal
cheaper than buying lots of weird bits
fenn: how about cam programs?
[00:29:06] <fenn> http://www.rainnea.com/cnc_5axis.htm
CNC toolkit is free 5 axis cam
you have to have 3dsmax i think?
fenn: if you own 3dsmax
fenn: 3dsmax is 5000$
got 3dsmax :)
that's the retail price of it..
well that's dumb
i thought it was just a 3d modeling program like rhino-3d
it's for making movies and stuff..
hence the pricetag
moves cost millions and millions, why would they charge only 5000?
oh, that's for one seat..
maya = ultimate 3d software
*cough overrated cough cough*
if you are going to do 5 axis you need a program
over here atleast
hack from yourself :P
charging $5000 for some numbers is stupid
* Jacky^ agree
fenn: i've yet to see any opensource cam program.. :)
well, i spent all day today blasting my brain with multivariate calculus so that i could write a really good one
anonimasu_: you're not the only one :((
you speak german right anon? try out gCAD3D for me eh?
no I dont :/
alex does :)
fenn: it's open source ?
Jacky^: i think so, yes
its in english but i had no way to verify the gcode output cause my EMC is borked right now
and some of the functions are still in german
reads iges and dxf and lwo and more
mm, i cannot help :(
anonimasu_: how proceed your jobs ?
got a day off tomorrow
going to try to get some stuff machines
the toolchange is more on hold since I got my new lathe
drastic...you are going to produce something. ;-)
nice.. got some pictures ?
not yet ?
just the same as I posted
got it stripped in molecules now
going to replace/clean everything
and paint it
make: *** [videohook] Error 2
ouch :\ stop
* Jacky^ tired
I spent a lot of time :(��
well, apt-get install ffmpeg !
without mp3lame support, doh !
take care... I'm gone.
Night? it's only 1900
alright.. calling all 8th graders
I bet we're all ex-eighth graders
with 13 watts of power, can we reach .6 m/s in 1.5 m travel with a 100kg load
this is so easy i can't figure out how to do it :)
this is a case for UNITS man!
constant power or constant torque?
well it's really constant torque, but let's try to solve the simple problem first
why? just do the one you care about
const torq means max power at the end
P = 0.6m/s * F
so you know F
A = F/M, so you know A
t = 0.6m/s / A, so you know t
d = 1/2 at^2, so you know d
is d < 1.5m?
thanks jmk mybrain isnt working right now
I know the feeling... I was flailing around with kbuild for nearly 5 hours last night, finally gave up
today I figured it out in 15 mins
now to make that accel problem really fun, do it in imperial units :-(
how many ergs in an oersted-foot-pound?
and what's with this oz-in stuff anyway
they should average the torque curve and give you average power output from 0-800 rpm
(for steppers at least)
hehe i'm still stuck on the first line :)
i'm going to bed
not a bad idea... I was up till 5am this morning, gotta work tomorrow
nope. Good night :)
I ate him
he was yummy
hope you won't get an indigestion
yesterday was aweful
seriously... I'm lactose intolerant.
and I had a glass of milk with chocolate cake.
I was gassy for 8 hrs.
it was like a system flush... except with gas in my stomach.
nope, not at all.
that's up today?
catching up on work
just woke up
started to write some code
have a brake
have a smoke.
hire a concubine.
heh.. thought those weren't for hire
drink on the job.
flirt with the secretary
concubines are like just a masters mistress/slave women.
that sounds like a plan
used specifically for sex.
I think they're illegal now... but I haven't had the oppurtunity to test it out.
could you have a concubine, if you're single male? Wouldn't that be called your GF?
that being the girl/concubine in question.
if you had multiple... maybe it would be...
like you have to feed them (dinner, and you pay).
I'll get off this train of though, because it's rationalized quiet easily in my head to equate my exGF to a concubine.
hey! look! over there... something shiney
* anonimasu_ yawns
7am... I should go to sleep
anonimasu_: machining ?
currently drinking tea :)
it seem to me that roughing toolpath is useful if there are big empty areas on the piece, otherwise not, right ?
roughing is useful for roughing
you usually rough then do a finish pass
to get within spec
in some case, it seem tale more time
because toolpath become more complex :\
so, maybe is better to do all in one with ballnose, without roughing..
roughing means more stepover..
hence faster cutting times
and a bit different strategy..
when you finish you usually have a very fine stepover to get a good surface finish
I found useful to calculate finish toolpath using a small ballnose(3 mm) in 3D relief
this remove some material exceed
Well roughing is faster then finishing
when first toolpath was with 6 mm ballnose
* Jacky^ confused
i'm missing something
maybe I should select the vector before calculate latest toolpath..
what are you using?
tool do not pass again in the same points that are finished
have a nice features
but it's a bit difficult to find the right strategy
anonimasu_: have you seen the inkpot I made ?
[13:58:30] <Jacky^> http://www.roboitalia.com/modules.php?name=coppermine&file=displayimage&pos=-1273
which tool would you use to machining that ?
are there very much detail?
should be a gift for my brother..
a ball endmill..
6 mm ?
maybe a bit larger for roughing
then a toolchange and finishing with a smaller
I dont think you can machine the smallest detail with a 6mm ballnose
3 mm ballnose should be ok
but it take a lot of time :(
* Jacky^ sigh
I am still deciding if I should go out and machine this part of mine
need a mount for the die grinder
going to use it for engraving
* anonimasu_ is trying to find out how to mount it
try mount -a
how are you ? :)
tired, bored, doing too much PCB layout work lately
no holidays yet ?
is Fenn here?
I used my holiday time to go to the CNC Workshop, then came back with a head cold that lated 3 weeks
looking forward to finishing current layout job so I can go play some 8-)
news update on website ?
news, what news?
form latest meeting..
gee, I have seen a couple of reports, guess none made into the Wiki
ah, well ok :)
Jacky, there are a batch of pics named "mazak*" at: http://linuxcnc.org/dropbox/
nice, thanks :)
That is the machine that the EMC crew was retrofitting
gee i'm popular this morning
sought after and popular are two different things 8-)
last night I ran across an interesting web site about 5 axis CAM and thought you might find it interesting
is it rainnea.com? :)
the page starts in French but includes English version below
[16:13:16] <steve_stallings> http://olivier.coma.free.fr/opencascade/opencascade.html
quite interesting application with code available, runs doze or Linux supposedly
hmmm i havent seen this page yet
i wonder if i have to download the opencascade libraries now
you have to jump through some hoops
thought it was a non-issue for personal use
yes, but i still don't have them on my machine.. i remember there being some annoying stuff you had to do
* anonimasu_ just machined a impeller
what was the problem?
it came out great but I made it 80mm instead of 40
I am re-machining it
You were having issues the other day
hm, remind me with what :)
no clue, lack of sleep?
that might be it
cool beans =)
Ill bring you a pic of it as soon as I've machined it again
if it turns out ok
otherwise I'll put it in the not so good idea bin
lol, fair enough
but from the other one I made I think it'll come out great
so, where's mine then?
Yep, the one you're making for me
I sell them at 20$ + s&h
you said free
argh... I feel stranded! Waiting for replacement parts to be made for the router, and an email back from Xylotex about the steppers and driver board (which I think is blown).
I know the feeling
I had a dead gecko way back..
and I didnt use it before the warranty expired..
so I have a stepper/crap axis
Did he deny the warranty?
I dont think he would have either
but shipping it back costs more then a new drive
I'll be going with real serovamps soon anyway..
oh man, that stinks
and a vital
what do you mean "REAL" ?
geckos are step servo amps..
you wont see the kinds of speed I want with them..
unless you got a hardware pulse generator
g200x.. or something
ok, what kindas of speed can you get from geckos?
2m/min right now at most
I dont want to sacrifice the stepsize for speed..
about 78 IPM ?
shit... with what I have I can only get 28IPM w/o it stalling
and thats just jogging
I want 5m/min when cutting
what voltage are your motors?
and even more for rapids
(hi guys :) )
right now, I have 48V
and several 100A
mine are 150V, 40A (max), but that's not important right now :)
I think the servos are 100V per axis
I'll be buying new servos soon
the Geckos will limit me to 80V, but that's still enough for 1700 RPM or so
plenty fast with a 2:1 ballscrew drive
[18:17:48] <anonimasu_> http://www.galilmc.com/products/motors/servomotors.html
the first ones are the ones I have
belt drive to the ballscrew, that is
am I on crack to think jogging at 28TPI is wrong (wiht the steppers that is)?
I'll be buying the larger ones later
just need somone to shop via cant afford them in sweden
acceleration may be too high if they stall during jogs
NEMA 34 brush servo motor, 1000 ppr encoder with round cable
thoose should be a good match
SWPadnos tell me more
10nm peak torque
Jymmm, those steppers are rated at 2.8A, 1.13 ohm. That's a 3.16V motor.
I would need to calc it
the inertia though
the Gecko recommendation is for 5x to 25x the voltage, so you're in the low end of that
the ones I have now are a bit too small..
for the speeds I want
but I need to upgrade other stuff first :)
also, the current from the supply is limited - they explicitly mention 2.8A/phase, not per motor.
SWPadnos Ok, so since it's underpowered, why would it stall ?
SWPadnos or are we talking apples/oranges
I think you have 2 phases on all the time, so you actually need 2/3 (5.6*3) or around 11A, not 4.5
because you have too little power to accelerate it to the speed..
does it stall when not connected to anything?
you cant push the rotor in place before the next step..
so you end up going nowhere..
SWPadnos : Sitting on the table - no problem. Pick it up off the table - stalls. Tight grip in my hand - np. loosen grip - it stalls.
weird - I'd expect the opposite (stall when held tightly)
the thing about steppers is that they need power to get out of the detents
MotorA stall more at low speeds, MotorB stall at high speed (same conditions above)
the problem follows the motor when you swap axes?
SWPadnos : Jeff (from Xylotex) was saying this was resonance causing the issue and that being on the machien would relive this.
uh - yeah
right - I'd expect it to work worse on the machine than when not mounted
and he says the opposite.
you work with RF, right (so I can use appropriate analogies :) )
I don't know what to think atm.
brb machining time
when you energize a coil, it doesn't try to turn the rotor - it makes the relationship between the rotor and the case change
if the case is bolted down, then the rotor will move, but if the rotor is held, the case will move
the reason I think it should work better when helo loosely or not bolted down, is that there are two places to put the torque - the case and the rotor
since Ke = 1/2 m(v^2), it takes less energy to accelerate the case/rotor relative to each other
more mass, but less velocity, so a square law reduction in power, but a linear increase due to mass
yielding a linear-ish decrease in power
Ok, I don't think it's torque (directly) but resonance. But WHY/WHAT is causing it to resonant so much (it is loud and very widespread)? Could it be the way the pulse are coming?
the pulse train could well be the problem.
do you have test equipment or the ability to make a 555-based pulse generator?
I have a breadboard around here somewhere
I also hace a Basic Stamp
ok - that would work
there's a 555 timer circuit in the CCED files section
(sidenote... right now the driver board is blown) why/how not sure.
"manual pulse generator.pdf" - or I can dcc it to you (I think)
that could also have been part of the problem.
actually, the power supply may not be too small - I think I was remembering the servo formula - not sure it's the same for steppers
no clue what the casue is at this point. Which is why I'm taking the stepper diag crash course.
have you read mariss' whitepaper?
this driver/PS is underpower for it's rating yes.
I have it pritned out, but it doens't mention resonance
also, it looks like a regulated supply - that's not an advantage for motor drive
it's switching PS
right - not the best thing. you'd be better off with an AC wall wart from DigiKey
(plus rectifier, of course)
well, the funny thing is that 'Jeff' said it makes no differnce (though I read it does), but that's sorta why I bought the kit. if it doesn't work/have issues it's difficult to blame it on something else.
unfortunately, they don't have any AC wall-warts that are high enough current
the switching supply won't have the same overcurrent capacity that a bulk power supply has
curious... do servos have the same resoance issues that steppers do?
nope - silky smooth, from what I hear (once properly tuned)
well I NEVER heard of stepper resonance issures till now.
and I tried reading up quite a bit first.
never heard of mid-band resonance? you haven't looked enough :)
my impeller works :9
servos are silent
they whirr a little bit..
Maybe it was information overload at the time, I've heard the term SWPadnos, but maybe I missed the point.
heh - shit happens
I am upploading the images..
[18:46:21] <anonimasu_> http://www.bojn.net/~an0n/img01.jpg
SWPadnos but I dont think it should happen at this low a speed (>28IPM)
1 minute of cutting
cool. What material is that?
looks like UHMW
I made it too small but :)
going to machine a large one and turn the base instead of mill it..
the round part..
* anonimasu_ isnt happy with it
visualmill did something freaky
but the blades are sweet
they did end up too thin but it dosent matter it was a tes
Jymmm, what's the drive like? (belts, screw pitch, etc)
I was looking for an impeller for the coolant pump on my bandsaw - pretty expensive little buggers
1/2"-10 acme leadscrew. Direct coupled
ok, so 2000 steps / inch
SWPadnos the coupler has slots cut in it too
so it's falling over at about 1 KHz step rate
ah - 10x?
ok - 8k pulses / sec - what did you end up with for PERIOD on that machine?
SWPadnos : I was never able to get emc running properly. So I dl MAch2 and used the 'minimill' profile they had. I also tried two different computers as well.
I also tried TurboCNC too
OK, do you know the maximum pulse rate that those other programs were set to?
par set to bidirectional on both machiens
SWPadnos what that be like 8K Hz ?
I suspect the issue may be granularity of pulse rates
or the pulse width?
the max rate may be 20 KHz, but you'd still have a problem.
not necessarily width - any pulse that a PC can do will be longer than the minimum necessary (for a Gecko - not sure about the Xylotex)
it's granularity - take a 25 KHz interrupt
count the set number of "blank" interrupts
when you get to the right count, turn on the output and exit the interrupt
the next time, see that the output is on, and turn it off, plus reload counters
then repeat the process.
so, each pulse takes at minimum two interrupts - 12.5KHz pulses, in this example.
but the next slower rate is 25KHz / 3 (one period on, two off instead of one) = 8.333 KHz
the next slower is 6.25 KHz, etc.
so what happens when you want to go at a rate between 8.33 KHz and 12.5 KHz?
Steps per unit: 2000, Minimum pulse width: 1uS
velocity: 7.025 units/sec
you output a few fast pulses, then a few slow ones, then more fast ones, then more slow ones, etc.
it's high speed acceleration, due to granularity
hm, inconsistency in the pulsetrain
SWPadnos: I'm sorry, I'd be lying if I said I udnerstand what you said completely.
like alternately hitting the gas and the brake - you go an average speed of 30 MPH, but it's a lot harder on you (and the car)
now THAT I understand =)
* anonimasu_ yawns
Anyone know if paul keeps a mirror for things he puts on the BDI?
all i know of is homepage.ntlworld.com/bdi-emc or whatever
should do ... the source code should be made available somewhere
My problem is that I did an apt-get update and install of the dev tools and now my .deb file is incompatible with the current BDI
ah well ...
you should be able to get the current release as well as the old BDI ..
I rpesume tis a current dev tools newer than the bdi version thing
mayeb you just need to apt-get emc too
w00t! .. I just bought a compressor :)
I start with the BDI and then add the dev tools so I can compile, prob is that it updates libc6 and then my binary won't install on a raw BDU install
so update the emc binary too?
[20:00:02] <robin_sz> http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=7529721040
at last .. a nice quiet compressor!
Well I ahve to make a binary for endusers that run a plain BDI ( they can't update via web because they run at 56K)
people use this?
would take them forever to update ( not to mention they lack the knowledge )
* anonimasu_ nods
SWPadnos : I'm in the CCEM file area /files/circuits/ which one were you talking about?
It may be on the Gecko list
it's called "MANUAL PULSE GENERATOR.pdf"
I can DCC it to you if you like
firewall, no DCC available
well then - that won't work.
funny - it said the file was sent
I think I got it... damn yaoo and theri session BS
any one know where i can find some pricing on servo motor's?
cool... will try surpluscenter.com dont like ebay
what was the verdict on using treadmill motors for cnc servos?
total waste of time?
Les loves em
he drives his fancy ass router with treadmill motors?
no he uses fancy stuff
the surplus center: one of the only benefits of living in Lincoln
cradek : chasmea who?
fenn: he does not run he's large router on thoose..
I hate having to ghost in a timeframe
ciao Jymmmm :D
Damn... that's the 3rd Maxtor drive I've had to take apart this month... two 30GB and now a 540MB
I've always said HDD mfg's cycle about every 8 years or so
you took apart a 30GB drive to fix it?
or just to play with it
two 30GB were dying
you dont take apart HDD's and fix them
and if you do, you're a better woman than me!
* Jymmmm dogpiles alex_joni
don't bug a tired man
it's 1am here
and I still gotta do a presentation :(
quit your whining! Unless you shoveled concrete for 16 hours, I don't wanna hear it!
is fenn around?
i looked in /dev/shm
tried what I said?
it's an empty directory
any of them there?
make dev doesnt do anything
there should be a /dev/shm file too
a file and directory with the same name?
smthg like that
but the files are shm0 etc
does make dev error out?
or simply not do anything?
doesn't say anything
is there a /dev/rtai_shm ?
do you think udev might have deleted them?
oh there is RTAI_SHM
or /dev/rtfx (where x is a number from 0 to 9 ) ?
yes rtf[0-9] and rtc
Jymmmm is now known as Jymmm
and /dev/rtai_shm ?
ok.. then it's ok
at least from this point of view
Jymmm is now known as ImInNoMoodForBS
ImInNoMoodForBS is now known as Jymmm
fenn: what's still failing?
* fenn pokes Jymmm with a pointy stick.
fenn : Jes?
forgot to check that part :)
talking to alex
talking to alex works?
alex_joni: well it dies on axis, but that's normal :)
probably those damned /dev's missing
and make dev restored them
alex_joni: i think udev is deleting them on reboot
probably udev has deleted them
I remember paul_c saying he fixed that in the BDI branch
for the load script to check on that
or to handle udev
yeah you just copy them to /etc/udev/devices or some such
didn't have the time to check that
maybe ./configure should do that
anyways.. going back to my presentation now
have fun on rtai-3.2 ;)