hi all, anybody got time to help an idiot?
heh... measured everything on the part, and then proceeded to drill a 1/4" hole, where a 3/16 hole should have been. doh... now I gotta buy bigger rivets.
mac, DaveE has a friend with a prob: a Japax 3F uv homing, could you talk to him? ( seems it runs but is at a constant tilt now )
know what I just realized... a center drill 00 is about the perfect size for drilling PCB boards.
stiff, small, readily available.
but, not carbide, so useless for FR4
cradek: you can get them in carbide.
cheaper than pcb drills?
I'm not isolation milling them, just drilling them.
haven't gotten to the point of isolation milling.
is the center on a center drill long enough to go through a .062 board?
seems like they're pretty darn short
[02:28:29] <A-L-P-H-A> http://www.i-logic.com/utilities/CenterDrills.htm
my board media is pretty thin
let me get my calp.
if it's over about .020 I bet you're out of luck
my board media says it's 0.020.
some parts are 0.022.
a center drill of 00 would be good enough as well.
what's up jmk?
these boards are only running data.. and a +5V line... low amps... as it's only a 7805 rated 0.5amp
I've never used that thin stuff, always the .062 FR4
in my head, I thinking copper thicknesses.
let's see.... work on sampler... go to sleep.... decisions decisions
Start new project at 10:30pm? [OK] [Cancel]
I started it yesterday
omg carbide endmills are expensive.
heh use pcb drills
hmm... maybe I will stick with a number #60 drill bit.
come in any diameter and last forever
cradek: I also don't have super highspeeds.
high speed does help for that little stuff
fsckit, sleep wins
cradek: this seem like a good deal? http://cgi.ebay.com/100-Micro-Carbide-Drill-Bits-Bit-PCB-Jewelry-CNC_W0QQitemZ330036081077QQihZ014QQcategoryZ58218QQrdZ1QQcmdZViewItem
good night jmkasunich
A-L-P-H-A: they're all too small for many things
I've never used smaller than #75 which I use for vias (24 ga wire iirc)
let me check what my smallest bit is
I use #69 for most things like dip pads, 1/4w resistors etc
and that's a little on the small side (on purpose)
cradek why's that?
I pretty much used a #60 for everything, except for some bigger stuff.
and I think I lost my #60.
oh well... a ten pack is cheap anyways
cradek: little on the small side (on purpose)
the parts stay on the board better for soldering, plus there's more pad left over
ah, ok. makes sense.
cradek: I just solder the shorter stuff first, and move up till the tallest.
I got a KW/h meter today. Nothing fancy, can handle up to 1800VA (iirc)
Jymmm: what are you measuring?
A-L-P-H-A: Just load of things. Wanted one for years but could never justify buying one. Got this one for free.
Mostly just want to see what things are really sucking the power; crt's, fans, etc
fuck www.busybeetools.ca... $1.49CDN for a #60.
is that for carbide with 1/8 shank? if so it's pretty cheap
straight shank, hss, jobber bit.
I remember them being $0.79
when'd did they double?
the pcb drills I buy are around USD2.35 in small quantity (<10)
[02:58:04] <cradek> http://www.thinktink.com/cgi-bin/carbide.cgi?page=imperial&start=128&step=20
what's a ring? size ring?
a lil plastic ring that indicates the size of the bit.
see... $2.35 I could understand if being carbide, and all.
it's the plastic ring that has the size marked on it, and can also be used to set the length in the collet
it's the 1/8 in shank that's so important
and like I said, they last and last
when drilling by hand you always break them off, but in cnc they drill forever
A-L-P-H-A: I use very few sizes. I keep around: .0625 (carbide end mill is fine), .046, .035, .029, .021
surprisingly, these really do cover most parts quite well
I strongly suggest that before you buy any drills, you go through your parts bin and measure pins on stuff
if you choose carefully like this you can easily get equipped for under $50
well... so far, I've just drilled #60, and shoved stuff in, and bent the pins to hold them in place. I am doing this manually.
what's #60 in inches?
#60 = 0.0400
oh, that's huge then
on the boards I make that would often take the whole pad
you must be doing very low density stuff
I think even wire wrap sockets (the big square pins) will go in a .035 hole
my wire terms are .75mm, while the drill is 1.02mm
my caps, are 2.5mm spacing, 5mm dia... those wires are .48mm
1n4001 = 0.79mm
yeah those are big
led legs, 0.6mm
if they're square, be sure to measure the diagonal, not the flats
yeah, I did. :)
ok, obvious, but an easy mistake to make
header pins are 0.85mm
my other style of wire terms are 0.9mm
I should probably get some #68 drill bits.
which are 0.79mm / .0310"
yeah that sounds like what you might want
like I said, I use #69 for most stuff, and it's just a tad small
#69 = .0292 iirc
#68 = 1/32"
pretty standard sized
.0310 != 1/32
#68 = 0.0310... 1/32 = 0.0313
cradek: like I can measure 0.0003"
(but point taken)
oh hey... this isn't a bad deal. http://www.littlemachineshop.com/products/product_view.php?ProductID=2318
but I already have a height surface gauge.
that must not be something I need very often
holy crap this is expensive.
[03:24:55] <A-L-P-H-A> http://www.littlemachineshop.com/products/product_view.php?ProductID=1927
I have the HF version... cost me I think $80
this seems expensive to me for what it is... http://www.littlemachineshop.com/products/product_view.php?ProductID=1943
Jymmm: how's the apt hunting going?
It's goin... WE think we have a 2br 1ba house with garage
Garage is good... :) more storage.
you can put your machines in there... :)
same area? or another area?
willing to move further out?
Not same neighborhood, but nice area about 10 minutes away. Yeah even a sink in the garage... good for wiid finishing work.
Jymmm: got any more photos of some of your work?
not really, I think I've already shown you the jest of what I've done.
when you moving?
when it's all confirmed.
I need a new SD card... so I can use my old video cam...
I don't think it has any onboard.
and I don't think it can hold more than 512.
ah, well 512MB isn't THAT small
no, it ain't....
but I also want a Cannon SD600.
I really shouldn't buy ANYTHING... as I just paid my insurance in full for the year.
$18CDN for 512MB... not bad at all
hah. just realized I gave my sister my media reader as well.
oh... damn. http://canadacomputers.com/index.php?do=ShowProduct&cmd=pd&pid=010969&cid=990.218
I could just get like a few of those... and if I loose them... haha.
I like SD basically becasue my laptop has a builtin SD slot and I can carry phone book, keys, personal data on them easily.
cheap now too
And you can get the throw away (16MB or less) usually for the asking.
128meg is less than a fourbucks coffee.
$10 handling fee per order
it's 20 minutes drive... I can go for chinese around that area too
* alex_joni yawns
it's raining outside.. and I have no urge to do anything :(
same as here
i hope it rained here
to sunny here :(
I have a toolchanger design done :9
the spindle part :9
anonimasu: We'll expect a complete working finished model in 72 hours.
I need to thread a locknut for my gear on the spindle..
anonimasu: No no, you dont understand... I'm not spekaing of JSUT the spindle, I mean THE WHOLE DAMN TOOL CHANGER!
that gives me the oppurtunity to add something to push aginst the drawbar when unscrewing the tool
anonimasu (but I never said it had to work again after the demo)
that kind of sucks :D
just a proof of concept =)
the only hard part about a toolchanger is getting it in place and getting it away afterwards
then you can go back and work out any bugs
mounting or actually changing a tool?
changing the tool..
what kind of machine?
my machine :)
just have it slide in from the back
side or back..
I think side is the best
kinda like a coffee vending machine, but reverse
I have a kind of idea..
the fork slides in from the side/whole changer arm..
you reach in to grab the cup. Instead of the cup being stationary (representing the spindle),
then flips the tool back and mounts the new tool..
the tool changer (representing your arm/hand) slides in from the rear
you need a swap arm too
yeah.. that's what Im pondering about getting there
[11:46:45] <fenn> http://fenn.freeshell.org/retrofest/dcp_0371.jpg
<- on the mazak the swap arm moves up and down
I'd like to have 2 tool positions so I can keep the next tool in the changer
in the changer arm
yeah like that
I need to extend my spindle to do that
as I cant fit a arm thats long enough
seen the haas style changers?
[11:48:53] <anonimasu> http://www.cuttingedgecnc.com/Movies/toolchange.wmv
I think I posted it befoer
yeah url sounds familiar
hard to get a good look at it
looks about like the way the mazak works, just faster
hm, if I settle for one tool in the changer it'll be easier..
that's probably fast enough :D
thats fun to watch on slo-mo
fenn: got any good idea about it?
well, hydraulics since you've already got the pump and stuff
* anonimasu nods
that much I know :)
i dont know when to use hydraulics vs pneumatics
huydralisc when you need force..
how much force to extract a toolholder from the spindle?
pneumatic stuff hardly goes over 8bar ever
fenn: 20 tons
my drawbar will be doing it for me
I'll have the drawbar screw the tool out
just make sure it isnt able to throw the toolholder across the room on accident
the mazak had prox switches for each position of the arm
actually a small robot would be the best things
yep a small robot could be really useful
load and unload parts and tools
though getting a decent payload is hard
not if you're making it
then you can make it slow and strong
they dont make too many slow robots
I saw some with a payload of 500kg :) in a magazine I got at work
not slow though
used for amusement park rides
seen the motoman robots+
[11:59:33] <anonimasu> http://www.motoman.com/
bleh large robot with 3kg payload
[12:00:28] <anonimasu> http://www.motoman.com/products/datasheets/HP3XF.pdf
what does "indirect cooling" mean?
for the controller
no idea at all :9
500 degrees/sec rotation
i guess none of that means much without acceleration ratings
[12:04:15] <anonimasu> http://www.motoman.com/products/datasheets/HP6.pdf
that design is really nice
eekk .. motoman ;)
anonimasu: competitors :D
wouldn't say they are bad
anonimasu: but you have ABB closer
at least for a while..
alex_joni: no $ for a robot..
would emc run a robot like that?
anonimasu: after a while
fenn: how so?
you'd probably want another interpreter, as g-code sux for robots
no collision avoidance, no dynamic acceleration ratings
then you'd want something to talk CAN to the drives
collision avoidance is easy
yeah just keep stuff out of the way..
fenn: usually it's done by ESTOP
the robot arm completes the estop circuit?
the only "smart" collision detection is current limitation in the drives :D
emc doesn't do JOINT acceleration/velocity limits
well.. not very well
or at least that was the idea i got
but that's the same for comercial controls :P
so even if I'd build a robot i'd ve fscked.. :)
stepgen does joint limits
anonimasu: the robot you'll build is easy to drive with emc :D
yeah usually the hard part is deciphering the weird serial protocols right?
fenn: CAN nowadays
what about old surplus robots?
anonimasu: there's also orocos which was designed for 6dof robots from the ground up
i dont know how complete it is though
seems like orcos would be a shitload of work
unimates are so ugly though
yeah, but you have kins written for them :D
put some G340's on this, and it should work quite well
ROFL "We are in the Desert Southwest so items may need dusting/cleaning when they arrive."
I'd rather build a robot.
the design part isnt that fancy, but the motors are pricey..
anonimasu: like I said.. a recent robot is up to 10k$ / joint
that's really horrid unless you do production..
anonimasu: have you been looking on ebay for motors an such?
im in sweden
nobody ships here
hey maybe you could use hydraulics :P
anonimasu: just ask.
dont you have piles of hydraulics actuators lying around?
I know I have shipped over seas if I am asked
they are all oversize
believe me - don't go hydrolics.
you can run tendons to the robot
skunkworks: I know :D
skunkworks: positioning huydralics are a hell
that is what is on our big nc mill. yeck.
(granted it did work)
if you have large stuff it works well..
I tried to do positioning with a 11cm cylinder ;)
this was k&t precision version - it actually did .0001 (scary huh)
that's very good :)
hydrolic motors - not cylinders.
with cool proportional valves that had little vibrators in them to keep them from sticking :)
that's neat :D
well - if thats all you have....
I wonder if mariss will start doing brushed drives
I thought he had a brushed drive - step/dir in
do you mean brushless?
for brushless motors
alex_joni: know of any ro to en translaters?
* anonimasu has passed on several motors because he's had no clue about drives
alex_joni : http://www.radioamator.ro/articole/view.php?id=248
alex_joni: You're human (most of the time), so start translating =)
I'm logging. I don't understand '!thislog', Bo^Dick. Try /msg logger_aj help
alex_joni been there, done that.
nothing on wikipedia?
[12:42:31] <alex_joni> http://en.wikipedia.org/wiki/Maidenhead_Locator_System
[12:42:42] <fenn> http://www.tranexp.com:2000/Translate/result.shtml
alex_joni: Yes, that's how I got the url I just gave you.
Jymmm: sorry.. a bit busy to translate all that :)
alex_joni thats whay I asked if you have a ro to en trans, google doens't have one.
i saw a link to a allegro stepper driver chip. can someone remember who gave it to me
[12:45:19] <skunkworks> http://www.cnczone.com/forums/showthread.php?t=24699&highlight=stepper+motor+chip
Bo^Dick it's the same one that xylotex uses btw
i'm working on a design that has auto adjustable mixed decay
Bo^Dick: Yeah, but does it have a built in coffee maker?!
well, wth?!?!?! lol
i'll meet any demands
fine, have it ready in two minutes.
Jymmm: that thread is about a new chip, not the one in the xylotex.
jymmm you're a tough boss. i'd like to invest in your company
Jymmm: among other advantages, it allows higher voltages (up to 50V) and currents (up to the limit of the external MOSFETs chosen)
does the A3986 support the anti-resonance feature?
i'm not sure how that works but i'd like to invoke that feature to my design
since i'm using a microcontroller i could maybe make some adjustments to meet that
maybe the resonance-compensation involves a change between the two phases?
if that would be the case it would be pretty simple to implement with the microcontroller
jepler: ah, very cool! Didn't realize the MOSFETS were external.
jepler: Please insert credit card here --> [ ]
1234 5678 9012 3456, expiration 13/08
I'll take two
lol, morning SWPadnos
hmmm. morning + SWPadnos don't mix well
SWPadnos + prime rib
mmmmm - prime rib
I'll take that as a yes =)
well, it does depend on the prime rib, but yes
nothing really beats a good fish fry though
working my way around eagle again.. Maybe today I will try to actually draw a componant.
except for really good prime rib
hmmm - I guess the Friday Night Fish Fry is an acquired taste ;)
yah. its mainly anything breaded and deep fried.
heh - well that I can agree with :)
its foolish to design my own stepper friver isnt it
if your measure of foolishness is economic, then yes
Bo^Dick: Not really... if you have the time
if your measure of foolishness relates to learning, then no
btw, what do they cost these A3986?
$3.04 in quantities of 5000
i'm NOT intrested in such quantities
heh - me either ;)
so what do they go for each
5000$ in quantities of 3.04
dunno. Mouser doesn't seem to have them, and digikey seems to be offline
they're pretty intresting parts
but that also depends on the price
I'd think that the single price would be 2-3x the quantity price, but it could be more
actually, there's a CNCzone comment that says they're about $5.00 at DigiKey
that's pretty cheap
then add external MOSFETs, I think
still fairly cheap though
if someone could spy on the zylotex
btw, would the mosfets be more expensive?
that depends on the voltage and current you'd want
one allegro part would require 8 mosfets right?
yay - digikey seems to be (almost) back online
there is still one particular advantage with my design
5.12 in quantities of 1 at digikey.
since i'm programming a µc in assembler i could implement code to deal with the anti-resonance feature (assuming allegro hasn't already done this)
a microcontroller is a blessing and a curse in this type of situation
the resonance-phenomenon is quite undesirable in CNC apps right?
in any app
I would think
you can do anything you want in the code (more or less), but everything you do takes more time and slows down the max step rate
why would it be a curse?
that's true although the ATmega8515 has 16Mips
well, if you add a resistor and acapacitor to the board, you get a filter that runs in realtime, and doesn't affect (for the most part) anything else on the board (except for the filtering function)
if you want to do a filter in a microcontroller, you add code to do the filtering, and it executes sequentially
that affects the timing of other code in the system
a microcontroller can do just about anything, given enough time
hardware just runs all the time
this is why i use asm rather than C
i consider this application time-critical
I always use asm, it doesn't matter
that only pushes the line between success and failure, it doesn't eliminate it
but i can't say in advance that i know for sure that the µc will provide a desirable step-rate. i just hope this is gonna be the case
well i'll have to read up on the resonance subject regarding stepper motors
so far it has been tricky to find detailed information about it
i've heard that this subject is the critical point between good and bad stepper drivers
so it's really funny there is so little talking about it
it'sa pain in the ass to implement, I gather
still it's a guess
prolly a good guess regarding the circumstances
it's got to be something special about it since allegro has neglected it
I think basically everyone has neglected it except for Mariss
that's really strange
it's probably there in more expensive drives, but I don't know
it's not so strange if you think about the market for stepper drives
if it makes the drivers better and more reliable the subject really would deserve more attention
for the most part, large motors haven't been steppers for years. they changed to DC servo, then AC servo (at least in the CNC world)
hobby-grade drives need to be inexpensive, and the compensation isn't as critical
the geckodrives sit on the edge between hobby and pro drives, so some hobbyists or light industrial users may get them, and need the compensation
so however the anti-resonance is implemented it's likely to affect the price of the drive?
add components = add cost = increase price
now i'm talking about large changes in price not marginally differences
it either changes the price due to cost, or it cahnges the price due to marketing (ie, better feature = you can charge more)
oh - as a cost, it may not change much - I think the circuit is only a dozen or so components
it doesn't feel 100% sensible to me
to make a drive as good as possible to the lowest possible price, wouldn't that be most attractive product to buy for the consumer?
good question. why haven't you bought a Geckodrive then?
it has only 1/10 stepping
fixed that is
no, it has 1/10, 1/5, 1/2 and 1/1
if you get the PLL board (the G210)
not the cheapest gecko
ok, so there's an amount you're willing to pay for a certain set of features
you just happen to want more features and/or a lower price than anything currently on the market
if some gecko would ever be intresting to me it would be their cheapest namely the one that goes for the "bargain" price of 114$ each
if you count your time at $1/hour, and the time of people here at the same rate, then you could have bought several geckos for the time spent designing your drive
observe that this is not the price for a three axis system, oh no, it's the price for only one driver
and $1/hour is significantly lower than my consulting rate ;)
so a three axis gecko system is almost 350$
so as i said, if you want to learn something, then designing your own drive may be worthwhile, but if you're doing it for monetary reasons only, it's a loss
is that a bargain?
well, you'd get a system that has mid-band resonance compensation, plus things like a warranty
i'm not even sure the motors themselves are that expensive
well you're right at one point, i'll get the midband resonance
but my driver can step 1/1, 1/2, 1/4, 1/8, 1/16 and 1/32
plus documentation, support, warranty, and the knowledge that if something is wrong, it's probably in your wiring, not the drive
that is definetly more than what the cheapest gecko can
btw, has the cheapest gecko mixed decay?
one thing is for sure, if i manage to get through the stuff with resonance, my driver will without doubt be cheaper and better than the gecko
but now it's all up to that detail
I hope you manage it :)
besides this, since all this has gotten me into the world of µc:s i could also start developing a frequency conformer for brushless motors and they are really expensive on the market
what microcontroller are you using?
very nice device
hey SWP how come your bridgeport is still in pieces?
if it's mounts that are bugging you, just make some crappy temporary ones out of hollow rectangular section
This nickname is owned by someone else
If this is your nickname, type /msg NickServ IDENTIFY <password>
this may be interesting: http://www.youtube.com/watch?v=NZNTgglPbUA
my perl skills are weak :(
There is no log URI
now I borked it for good :D
This nickname is owned by someone else
If this is your nickname, type /msg NickServ IDENTIFY <password>
fenn, I'm lazy ;)
sorry bout the toggling :)
I'm feeling lazy .. but here's the log anyways: http://126.96.36.199/irc/irc.freenode.net:6667/emc/2006-10-05#T14-27-39
Try this address for searching the logs: http://188.8.131.52/cgi-bin/search.cgi
logger_aj: chump a
weird - I just got Alex's 3 CVS commit messages
from October 3
heh.. me too
might be the lists were down?
so, I've seen a fair amount of discussion about "alter" - where is that?
didn't see anything on emc-devel or emc-users in the last days
SWPadnos: it's an alter ego
kidding.. not quite sure
me either. though I thought there might be a problem with my mail server, since I had zero emails from a 6-hour period last night
well, one of your commit messages says that you didn't change alter, needs discussion ...
double alter;/* additive dynamic compensation */
err - command.c
no.. in emc/motion/motion.h
int usrmotAlter(int axis, double alter)
/*! \todo FIXME - the comp stuff is temporarily disabled, so "alter" must
be disabled too. BTW, what the heck is "alter" anyway?
that's in usrmotintf.cc
bet jmk wrote that before he found out what it is
it would be cool to have lxr available for the emc repository
but I still don't quite know.. it's supposed to be an offset which gets added to the motor pos
so you can dynamically change it while it runs
SWPadnos: here's the most usefull I've found about alter
[14:33:38] <alex_joni> http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc/src/emctask/alter.c
sounds like a poor man's HAL
well, that's noly 5 years, 11 months old ;)
alex_joni: didn't someone just do that with emc2?
here's a usefull diff: http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc/src/emctask/alter.c.diff?r1=1.2;r2=1.3
skunkworks: no, I only said it used to be in emc1
hmm - I thought a few weeks ago.. they guy using kins to map the lead screw error also. Or am I confused.
skunkworks: different thing
I don't think alter is part of leadscrew compensation, except that the comp struct seemed to be a good place to put the "external offset"
SWPadnos: yeah, but in emc2 there's a motor offset anyways
right, so alter is probably redundant in emc2
so alter might/should go away/different place
that's why I wanted to talk about this..
maybe jmk will be around later.. he did most of this stuff anyways (emc2)
note that the compensation array is fixed at 256 elements as well
SWPadnos: noted.. but at least it reports an error :)
in emc1 it just silenced away I think
maybe I'll set up a Linux machine so I can join the discussion a bit more intelligently
doesn't help me any. :)
it may be possible to have the program that fills the array do some interpolating to cut down the number of coefficients down
SWPadnos: why would you want that?
alex_joni: WRT the discussion on the dev list yesterday and the strange stepover in axis, is that resolved yet?
Lerneaen_Hydra: don't think so
was that running sim or real iron?
alex_joni, so you could still run with a small-ish array. not the best solution, but possibly better than just turning off compensation if a larger set of points is provided
SWPadnos: if the user needs a larger array, it's one #define away
that's one recompile away though, so it's too complex for many users ;)
sudo apt-get build-dep emc2
apt-get source emc2
then they should find someone to do it for them :P
im going to come whine here about stuff like that :D
anonimasu: you I have no hard feelings to slap around
im looking forward to moving the mill with the usc :)
Yes for a few hours.
hours doesn't sound right :)
don't I know.
rayh: keeping busy http://www.electronicsam.com/images/KandT/servostart/DSC_2105.JPG
suprisingly not distructive testing yet.
* rayh looking
Okay skunkworks. What you doing with that.
that's quite a heatsink.
rayh: just testing some mosfets I got. jmk was right - they are 44a but I will be lucky to run them at 20a contiuous
That whole heat sink would glow.
without a fan - it boils water (spit) at 20a
how about circulating water through the heatsink?
anonimasu: maybe. still goofing around
holy crap. dreamhost just increased the storage and bandwidth limits a whole bunch
like 400G storage, 4TB transfer / month
for how much $$$?
(if you prepay, $19.95 if you don't)
I may need that someday
plus, they add 2GB/week of storage, and 32GB/mo transfer every week
I don't think I can generate enough data or use enough bandwidth to catch up with the increases
* bill20r3 can
well, if I start taking zillions of photos/videos, then yeah ;)
Server uptime: 2 days 22 hours 22 minutes 6 seconds
Total accesses: 1339903 - Total Traffic: 38.7 GB
although 4TB is a ton of traffic in any case.
yeah, and with the increases I've got, I'm up to 422 GB storage, and 4.75 TB transfer
hmmm - that's roughly the equivalent of a 15 MBit/sec internet connection, at 100% usage
so if I had FIOS (and they allowed servers), I could almost serve more data than this plan ;)
fancy input/output system
skunkworks: if you really need that kind of power out of one
Fiber something something Service
it's Verizon's high speed ISP service, but even the most expensive plan is only 30 MBits down/5 MBits up
and that's $180/month
SWPadnos: you kidding, right?
the fastest fios is 30/5
nope - not kidding
almost as fast as what you can get for $20 or so in Korea and Japan
well, actually it's not :(
anonimasu: I would like to. We have some servos that are 220v 17a continuous and 87a peak
not saying I can get something to work with them
skunkworks: perhaps you need bigger MOSFET's
or whatever they use in apps like that
perhaps. but I am just playing to see what I can do. the price was decent on these
hm, I should go outside and start working on the spindle
going to mount the motors there now
I need it, I need to modify my spindle a bit
going to add some thread and make provisions for a toolchanger :)
tool changers are nice :)
it's too bad I cant get a retainer in there easily
I'll have to use the drawbar
but it's ok..
anonimasu: 20 amps http://www.electronicsam.com/images/KandT/servostart/DSC_2110.JPG
is that 42.2
not hot at all
is that at 220v
no - I could live with that. -
with a big heatsink that shouldnt be a problem at all
nope just what it took to put 20amps thru it.
pretty much juction voltage :)
yeah but at what voltage?
I'd like to see that at 220v
that is the plan :) should really not make a diffence if I am thinking right. 20 amps is 20 amp.
the power the disapated is still going to be the juction voltage
hm, that's kind of odd
but im not a wiz at electrics/electronics
20a at 220 is still going to be 2.4v on juction voltage if that made sense
just a bit scaryer if something goes wrong :)
ok. I might be getting the hang of eagle.
what's a graphical FTP program for gnome?
does that upload?
need sftp as well.
just open a console and type "sftp"
xterm is graphical enough
yeah, it runs on X
you can pick what font and color
and transparency on modern ones :)
* ValarQ prefers rxvt-unicode
one of the few that supports true transparency btw
good night all
in a shell, how can I move an entire folder? say I have /home/ABC/asdf, but I want all of asdf to be in /home/asdf (removing the ABC)
mv only moves files...
no, it moves dirs as well
also, I have a folder called /home/asdf
so should I kill /home/asdf first before moving the other one?
so you want to move files into an existing dir, or you want to move the dir to a new spot?
it doesn't matter
-r is illegal
if you move one directory to another directory it will put it in the directory (not overwrite the old one)
mv /home/ABC/asdf/* /home/asdf/
SWPadnos: that just moved the files, not he dirs.
doing it manually
you did type * not *.* right?
there's lots of sub folders.
okay... move things, just empty folders.
mv /home/ABC/asdf /home
yeah - that too ;)
* cradek hands the #emc channel a linux newbie book
will that work if /home/asdf already exists?
cradek: anything new with the sherline?
no.... okay... I'm missing a few words in my last statement
it moved stuff, just not folders that already existed.
Lerneaen_Hydra: nope, I haven't messed with it much
so I'll just rename the old folders
cradek: oh, ok
SWPadnos: depends whether the existing asdf is a dir or file
well also depends what you mean by "work", heh
heh - "do the expected and hoped for thing" ;)
it does the sane thing - if it exists and is a dir, it puts it in the dir; if it exists and is a file, it gives an error
mv: cannot overwrite non-directory `b' with directory `a'
I think there's -f
K`zan is now known as wv9k
wv9k is now known as K`zan
on a completely separate note, this cooler works *very* well: http://www.newegg.com/Product/Product.asp?Item=N82E16835128009
I did lap the bottom surface, so mine is probably doing a bit better than normal, but it's keeping the CPU 6 C cooler than the AMD stock heatsink/fan at idle
cooling from a high temperature (stupid motherboard - don't ask), it was closer to 15 C lower, and was 2C cooler with the fan not running
I'd prefer if the lights weren't there, but the case will be closed anyway ;)
I don't think I could have made a better choice. This thing is amazing!! Looks super rad too in the dark. </reviewer>
Pros: Can anyone help me walk through installation
Cons: I tried googling for directions, got nothing
this reviewer has little to no experience withthis type of product" ;)
man these are great reviews
I generally glance at the reviews, then pretty much ignore what they say
bah, a real geek machines his own cooling (prefferably water-based) ;)
I generally use the AMD retail coolers (I like the 3-year warranty on the CPUs), but I'm trying to quiet things down a bit
water would help that, of course
undervolting/underclocking really helps a lot too
well, I prefer to have it cool while actually doing large smounts of useful work, so the underclock thing isn't really helpful for me
as heat scales linearly with clockspeed and squared with voltage
amounts, not smounts
water is for wenies
a real geek uses hydrazine
real geeks use a mineral oil bath cooled by liquid nitrogen
* fenn wonders how he managed to remember the specific heat of hydrazine
real geeks buy a fast enough machine in the first place...
SWPadnos: -80 degrees on a bad day
I've actually seen a overclocked machine cooled with liquid nitrogen
and dry ice
and 2 cascade coolers..
yep - I've seen it on the web, but not in real life
it was cool
what was it, an atomic clock?
we expected the cascade coolers do blow :)
nah, just a p4 they overclocked..
I've seen photos of a machine submerged in mineral oil, and cooled by liquid nitrogen or dry ice
there was a ball of frost around the cpu..
the guys had -100c at the processor ;)
[20:44:31] <anonimasu> http://www.nordichardware.com/Articles/?skrivelse=482
then they weren't pushing it hard enough1
they poured liquid nitrogen down at it..
oo lithium hydride has an even higher specific heat capacity
specific heat is probably less important than thermal conductivity
fenn: lithium hydride?! sounds rather non-inert ;)
it's a rocket fuel :)
oh, that's where I heard it from
* anonimasu pours oxidizer at it
or more like fwoom
hmm - then my servo amp would be 44a....
We have a laser that the output tubes are submersed in oil. the oil is pumped and cooled thru a radiator.
a cutting laser?
at work - not home ;)
it'd be handy to keep away solicitors.
bah, 'night folks
bill20r3: how is the mill coming?
slowly, but it's still coming.
I keep getting frustrated with my inability to get precise measurements.
it's probablly 1/2 me and 1/2 cheap chinese tools.
does anyone know the bandwidth of the geckodrive?
ah, found it
you mean the maximum step rate? That's what I compute from the datasheet, .5uS "0" time + 4.5uS "1" time = 5uS, 200kHz
does that yield a maximum rpm of only 100 when using a 200 step/rev motor?
yes, 2000 microsteps/rev = 100 RPS = 6000 RPM
that's pretty fast
however, there isn't a stepper motor in existence that will be able to do any work at that speed
and you won't attain that rate with PC software step generation either
in fact, i've seen my computer output 200kHz at the parport output but that was only two lines of code
no CNC software could possibly work that quick and efficient
you son't have a good pulsetrain at that rate, especially when doing anything else (like motion planning, for instance)
actually there it is
[21:23:13] <anonimasu> http://jelinux.pico-systems.com/univstep.html
of course - I'm talking about software generated steps
silly commercial made me laugh... this yogo (yogurt drink tube thing), bunch of random people in pink outfits are popping in and out of holes. Then dracula comes in... one of the pink guys goes to him 'seriously you can't be here' "why" 'cause you make people nervous' "fine..." dracular walks off... heh
software pulserates are crap in general
hence devices like the USC and G-Rex
to get the resolution I wanted I had to really push it..
are they better than geckos?
that's just a pulse generator for emc
SWPadnos: wheren't you making one at one time?
a pulse gen...
ah well, really it's a i/o board.. you can run steppers with encoders with a usc + the daughterboard
I have a "manual pulse generator", that makes a stepper motor/driver act like a power feed
I was going to make a step generator, but the project went away
it looked like more than jsut a pulse gen
or just backburner?
poof / backburner
Bo^Dick: well i thas some relays for i/o too
stepper driver==stepper controller?
the controller dosent do anything drive-ish
it just outputs smooth pulsetrains
to me i thought stepper driver=(stepper controller+drive stage)
oh wow.. $250
250$ for what?
you can have a single thing that does both, like the pic-servo chip (I think)
[21:29:13] <A-L-P-H-A> http://jelinux.pico-systems.com/pricelist.html
oops - nevermind me
it's a pretty good deal for what it does
you're not possibly saying the pulsegen was 250$
yes, it is
4 axis, 100-ns resolution, 4 encoder inputs, 16 digital inputs and 8 outputs, all on one parallel port - not a bad thing actually
some people charge $100 for a good breakout board
this isn't the first time i've encountered external timing-hardware for CNC purposes
but hey 250$
the reason for this is poor timing accuracy through an operating system
CS-4A00-0 Software and Signal Generator (Board Only) $1045
what a bargain then
Bo^Dick: build one
well, i'm slowly getting into the world of microcontrollers
yeah - it's sad that a $2 microcontroller can have a pulsetrain that's about 100x as consistent as a 2GHz+ PC
the flashcut signal gen is $1.1K!?!?!
that is a freak'n PROFIT!
Well it does come with the software to talk with it.
SWPadnos what would it take to commission you to write an Atmel signal generator?
$100000 + a week in Hawai'i
and a pony.
how about $1000 + weekend in Toronto + a meal at Rodney's Oyster bar?
err - nope, that's not enough ;)
how about $1000 + long weekend in Toronto + a meal at Rodney's Oyster bar?
what kind of signal, just something to use as a clock reference?
it depends on the specs. I can make an AVR-based step generator woth a serial interface without too much trouble, but making it into a product takes time
I've been playing with avr's lately, they're pretty cool.
avrs are great
SWPadnos: I don't even know what's up with that.
yep - very nice chips. I've been using them since they were first released
maybe I should hack g-rex, and convert it to an atmel..
I don't know why they went with a rabbit
I've considered making an ARM or AVR32 CPU for the G-Rex
it was easy - the ydidn't have to develop it
but like develop what?
all they have right now is... is making a pulse train, and turn have some IO ports.
relay it all back to via USB or Ethernet.
they didn't have to design a CPU board with ethernet and memory and stuff
actually, the software in the rabbit wil do coordinated motion, with acceleration and velocity limiting
gcode in -> buffer stack -> pulse train out.
so... we could hack the shit outta it... and port it to AVR.
making that sucked CHEAP.
the rabbits are expensive mofos.
$100/chip + $150 programmer or something
not really that expensive for what they have, actually
$60 or so for the module, I think
wait... I'm dead wrong.
the main issue with it is that you have to use their C compiler, which only runs on Windows
I'm angry atm... I don't know why.
I need some tea to calm me down.
trying to figure out the speeds isn't as easy as I would have expected.
pulse gen speeds that is.
acceleration, constant velocity, deceleration. Transitional movements.
continuous velocity as well
it can be a bit of a pain, but it's not too bad
the real issue becomes quantization error, since you'd probably use integer math in a microcontroller
SWPadnos: you taught me a trick...
multiply it by whatever decimal place I want...
I'm sure ;)
sure, and you can use difference equations for vel and accel as well
I think AVR C can handle some of the math for us, couldn't we?
in fact, you can do constant contouring that way (or close to it) also
toss some ram if we need it
SWPadnos: any motivation for you to do so?
the AVR isn't all that fast for that kind of thing, so it depends on the algorithm
well, I'd love to make either an open-source compatible G-Rex CPU, or a completely differnet EMC-compatible ethernet-connected unit
the latter being my preference
but it's a big job, and I still have to pay the bills and stuff
I say EMC compatible devices.
I agree. the G-Rex FPGA isn't really geared toward feedback control over ethernet/USB (which makes some sense)
make it better than the g-rex, and sell it for the same price!
so even an ARM/AVR32 CPU wouldn't really make it great for EMC
right - that's the idea, and Eric Johnson recently did some RTNet/HAL work that I need to look at as well
woo hoo. finally made my first part in eagle. (actully not... just put a package to a symbol. but hey)
skunkworks: now you get to make the board and find out if you were right
working on it :)
I made a USB connector symbol and package, and got the pins the wrong way 'round
bill20r3: yeah basically
skunkworks: yeah... it's sometimes a pain. :) but I've done it for several parts I ordered from digikey.ca
This is a large diode - so I have a 50/50 chance that it is right :)
I walked into the shop... and just looked... i didn't even touch a tool.. just turned on the light... :(
too much to clean.
03jepler 07HEAD * 10emc2/configs/common/emc.nml:
with this change, emc2 --enable-simulator seems to run in 64-bit mode
$ file bin/emcsvr
bin/emcsvr: ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for GNU/Linux 2.4.0
how does one read what amp rating a wire is good for?
first hit for "wire gauge current", maybe it helps? http://www.powerstream.com/Wire_Size.htm
CSA CMG 50C 22 AWG LL83251 (UL) CM E129988 .... CSA AWM II A/B 05C 600V FT4 LL67097 WO# 102459
What's chassis wiring?
not knowing what it means, I'd use the lower (power transmission) figure for anything
but that may be for short lengths (within a chassis) not long lengths (power transmission)
yeah, I just gathered that after looking it up in google