in the last week nothing..
I think he's in holidays
states again !! :)
maybe.. i've no idea
Jacky> so you have a toy machine
cutting or milling jobs !
I would like to made some 3D relief on wood
the result is not bad at all, for now..
any pics around the net !
lemma se :)
[00:05:34] <Jacky^> http://www.roboitalia.com/modules.php?name=coppermine&file=thumbnails&album=75
My budget is small, and for now, i'm just interesting in learn emc..
for example, I was looking at this project : http://vinvin.dyndns.org/projects/scanner.html
but can't find any software on the website :\
maybe, should not so difficult write it from myself, i will try..
pics looks nice a goot start you made
hehe.. not really
but work fine for me
for soft wood it wil work for sure
the controllers made yourself aswel !
yeah, I solved a difficult issue with opto-limit switch here, tnx to Phydbleep
the controllers are very simple, L297-L298
thatz were frends are for hey
ow the L types chips works great
yes, I suppose
the parts more expensive has been the 3 joint: http://img-europe.electrocomponents.com/largeimages/C748342-01.jpg
Euro 19 everyone
* Jacky^ sigh
seriously, stuff are ridiculously cheap everywhere but here :/
always dat way
I'd be glad if I could get thoose < 45 euro
in a metal junk yard i almost got hit by some serius steppers
they fall of a truck includet the control case all from berger lahrr
jacky you dont need those couplers for your machine
take a piece of vinyl or rubber tubing
hose clamp it to the leadscrew on one side, and to your motor shaft on the other
hi feen, I tried..
motor was dancing :\
and big noise
Hmm, I threw my mail to picosystems now
and if they got hot !!it slips ore burn off
hopefully I'll be getting reply during tomorrow
couplers are better though..
I was have too much alignment error :(
the screw are also inexpensive.. about 30 cent
it could not work well without couplers
alignment is easier with real couplers
I can say, now there's no noise but for the rest it work as before..
was too bad heard the noise and see the motor moving
just for this, no improvement for the rest
what's the problem with the rest?
"no improvement for the rest"
I machined a part with and withou couplers, no difference
oh, peace of mind is a improvement..
a great one
I would not have shown my machine to nobody in that way
i have to take some rest - lots of bruses on me boddy
roel01: ok do taht
yeah but $19 euro is probably more than your motors cost, right?
right, the motors are 10 euro
it's so easy to spend huge amounts of money
not for me, really :(
heh that's what i mean
* anonimasu yawns
* Jacky^ agree with anonimasu
robin_sz: how are things going?
laser ran most of today, so i guess thats a good thing
* anonimasu ordered a USC from picosystems a bit ago
although I've yet to receive a reply
one customer had asked us to quote
and then only ordered some of the bits
other bits he ordered from another supplier ...
today he came in a panic .. he'd got the drawing worng ;) .. could we make them in 4 days?
so i made them in 2 hours, just to show him :)
hopefully the USC will make my emc box behave better
I wont be getting much more speed, but ~3m/min isnt too bad
I've got the period set high to be able to crank out enough pulses
I wonder if I run 5x at the geckos, if so I'll be getting 5m/min
which is acceptable rapids
I'd want to do 500
but it'd throw my table into the wall
and your point is?
trouble is ...
emc you need to set period MUCH faster
because the pulses become rough as you approache the maximum freq
robin_sz: yep, hence why I want/ordered a usc
it'll be way smoother..
it will give you more torque
* robin_sz nods
steppers are fine if you elt them spin
well, it'll probably be smoother anyway..
get them over the corner freq and they are stunning
the software generated pulsetrains is nothing to admire
being able to use the UI is fun too
yeah, using the same UI is neat
I hate it when I get delays in the UI
it scares me every time
it looks like I might have a customer for the EMC interp :)
someone contacted me last week because they needed a Gcode interp .. why me I don;t know
did I have a clue how to write one ...
take the PD one, scrub it up a bit ...
they just paid a guy to write one ...
it does G0,1,2,3 ... errm thats about it
and it keeps forgetting its current Z position
and cant do full circles
has to do two arcs
no G4 ...
infact not much really
I wish you could offset x,y in polar..
for g2 centres poitns you mean?
hm to mill \ and stuff,
you can do it with a g1 x1 y2
but still, offsetting everything around a point would be much better :)
wtf all these quadrature decoder chips only go up to 20khz
and they're $10 each to boot!
* fenn needs 100khz
* Jymmm lol @ fenn
10MHz $9.38 for < 10 pcs
1k pcs $6.02/ea
all i want is something that counts pulses @100khz and gives me 4 bits out
and the computer can take care of the rest
for DRO ?
to keep track of servos
ah, well these are 22bit
with serial interface
hey even better
how do you get it out of serial and into parallel?
why would you need paralllel?
so i can bit-bang the parallel port
no no no you silly bitch... not parallel port, parelel I/O
to got from serial ---> to pin on parallel port I think you can use a MAX222
or it's equiv (which are cheaper)
so when a chip says it talks serial, it's using rs232?
* fenn googles
it's a single output on one pin, instead of 4 pins (as an example)
only two wires are needed.... GND and DATA
so technically one
ok, what does the protocol look like?
or, more importantly, can I multiplex 6 of these babies while still in serial
to connect to six input lines of a parallel port you mean?
well, if i use your chip, then I've got six serial data lines.. how do i get that into the computer?
i've got an idea for how to multiplex stuff on the parallel port using an addressing scheme
but i dont know much about interfacing multiple serial lines to a serial port
I guess you could use some I/O board.
wait... MAX222 is a line level converter, nm.
CMOS <-> TTL
[02:13:41] <Jymmm> http://www.genapta.com/quadrature%20decoder.html
i wonder if it might be easier to just use a microcontroller and have it talk to the quadrature signals
i mean, no decoder chips
fenn: I am afraid that'll be too slow
Basic Stmap has 16 I/O pins.
i was hoping to get by without having to use a micro so i wouldnt have to learn how to program the buggers
why too slow? dont they go up to like hundreds of MHz?
well, still a lot more than 100khz, which is my target rate
here's one at 50MHz
if each micro cycle can read all the pins, figure out which way it's going, and increment/decrement the position variable, and tell the pc what happened, then i'm fine
unless it has to read 22 bits of serial data on each pin
[02:20:32] <Jymmm> http://www.parallax.com/detail.asp?product_id=BS2PX-IC
here's what i was thinking when i originally thought of this: make a small circuit to figure out step and direction based on the quadrature outputs
hook that circuit up to a binary counter, with step on the clock pin, and direction on the increment/dec pin
then choose 4 pins on the counter to determine the final resolution, based on the speed of the computer and desired motor speeds
(since i already have 2000 ppr encoders)
which is way more than enough
the parallel port multiplexes through 6 of these modules by outputting which device number it wants to read
then it reads it on the next cycle
fenn gawd you whine too much sometimes ---> http://pegasus.me.jhu.edu/~allisono/courses/530.420/labs/lab5.pdf
it's my job :)
fenn : http://126.96.36.199/search?q=cache:VhIupwN1ilYJ:www.cnczone.com/forums/archive/index.php/t-4424.html+basic+stamp+quadrature+decoder+&hl=en
based on the "19,000 instructions per second" it looks like its way too slow
fenn LEARN TO GOOGLE ! =)
hey i didn't want to use a basic stamp in the first place
there are other uC's I only own the BasicStamp =)
basic is useless for anything fast
BASIC and AVR Mega
[02:32:40] <Jymmm> http://mcglothin.us/RobotScrapbook/CommunityPcbMill2005/BasicStampGCodeInterpreter/
I think it's kinda neat for a proof of concept (which he says it is)
heh.. Unfortunately, I ran out of 9-volt batteries while starting the testing, but I can assure you that the program does compile without error.
fenn: you will need a counter chip
you need some overhead for sending out the data also.. :)
but a atmel is like 10$
horrid isnt it?! ^_^
what time is it in the us?
9 pm here
ok, just curious about when I might get a reply from picosystems
anonimasu: why would I need a microcontroller if i have a counter chip?
fenn: to pass your data back to your pc..
wait a minute.. if i somehow got step/dir out of the encoder and fed that to a counter, I could go straight from the counter chip to the parallel port right?
you get 2 pulsetrains from the encoder..
there will be a (hopefully small) logic circuit to decode the quadrature output
you can go straight to it, but that's as easy..
that's what a counter chip does..
i mean a binary counter.. you give it a pulse, it adds 1 to the number
i think you are referring to what I call a quadrature decoder
hm, how would you determinate which direction it's going..
the problem is the quadrature decoder chips i see are all too slow
maybe i just need to dig more
dig more :)
that and they are all serial i/o which makes it impossible to interface to without a micro
fenn : Seiously... what makes you think that? A parallel port is nothing more than 8 serial pins
would USB be easier for you to interface to than parallel port?
no, there's no USB stuff in HAL yet
besides how would i multiplex serial signals?
anyway, about using serial i/o through the parallel port... i'd have to write a driver to talk to the decoder chips, which is doable
i think that's a good solution
* anonimasu yawns
* fenn grumbles
well cheap one..
still gotta get data out of the computer
six serial lines eats up 12 pins
[03:04:21] <Jymmm> http://www.beyondlogic.org/spp/parallel.htm
no six serial lines eat 7 pins
the 7th being GND
well 6 data linese
fenn read the url I just posted entirely
hmm i guess i dont have to transmit any data to the counter chip
90% of this is already taken care of in HAL
but 100% is things you aren't aware of
well, if you have lines to waste that's not a big deal
anonimasu thats what SPI is for =)
or I2C =)
Jymmm: yeah but how do you do i2c with thoose counter chips..
anonimasu no clue, fenn said they were too slow
yeah thoose but there are fast enough counter chips
I gave him a url for 10MHz ones =)
20 to 32bit too =)
sampels are always free =)
fenn : http://www.beyondlogic.org/spp/parallel.htm#7
the parallel port is really not a very good I/O system huh? :)
it's good :)
just depends on what you are going to do with it
not the best thing for intensive realtime simultaneous input and output
are there any good way to do thoose on a pc?
what bout those boards like vital systems and such
there's just weird stuff about parallel ports that doesn't need to be there
* Jymmm smacks fenn with a branding iron!
there's plenty of stuff you should read
whats the rul to the datasheet for the chips you have fenn ?
i dont have any chips
i'm trying to get a datasheet for a binary counter right now
then i'm going to try to figure out how to make a circuit to decode quadrature and feed it to the counter
* fenn wonders if he's getting in over his head
well, if you feel like it's hopeless dig for counter chips :)
that's how the servocards stg/vital/stuff does it
do they have a micro on the board? how do they store 32-odd bits of position data
maybe thats a dumb question
you might read them directly..
how's the pc talk to the chips? i dont know anything about isa bus or pci bus
fenn dont go there...
a micro does not store 32 bit's of position data..
but, there are probably counter chips that are less..
a uC stores NOTHING
you wont need that kind of range
since you will be polling the counters frequently
yeah i only want 4 bits
well, it doesnt really matter
4bit? that aint shit
it is when you're updating it at 5 khz
fenn without soundsing liek a smartass here.... what do YOU think it's "4bit" of?
4 bits of position data
4 least significant bits
which is what range deimally?
the computer tags it onto the end of the position variable
you mean 16 decimal?
that means there's only 16 possible positions
2000 ppr encoders
it only reads the 4 LSB's
it knows which direction it's headed, and the direction won't change between read cycles
basically i have too much resolution and need to trim it down a bit
i can choose which 4 bits to use as my LSB's
no such thing as too much resolution
yeah, there is
if you dont have the capability to keep up with that much data
i can't handle 6 axes at 100khz
not through a parallel port at least
what is 100KHz?
2000 ppr * 2700 rpm = about 90khz
2700 rpm is the max rpm of my motors
yeah but you can poll them frequently thus reducing the need to keep that much data in memory
poll what.. the decoder chip?
the micro handling the chips..
or you could buy a FPGA and do it in hardware..
i understand digital circuits, in theory at least.. but micro's and fpga's are another ball game
also, i just dont see how a micro that can do "19000 instructions per second" will help me
you can poll the counters 10000 times per sec..
and pass the data to the pc every x cycles..
or well let the pc poll them
how's that better than letting the pc talk to the decoder chips directly?
EMC can do 50khz i/o
so that's like 25000 times you can request/receive data
or is there some circuit in the micro that automatically gets serial data and stores it in a buffer? like dma
depends on what micro you use, and how you do it..
you should have a talk with SWP I think he's been doing more work about it
he'd probably pummel me into the ground with his crushing intellect :)
but you wont be doing much more then polling your counters
I really dont know how you transfer 32 bit by the paralell port with decent speed
i wonder how fast is good enough
for servo control
i'm used to thinking about steppers
well, you would probably need a vital or something
but it seems you dont need to update servo velocity anywhere near as quickly
fenn: a 16MHz AVR can do millions of instructions a second
nearly one per clock cycle
ASM instructions or a line of C?
depends on the line of C
one line of c is usually multiple assembler ops..
if p then q++
is that 2 instrs?
that's 2 or 3
fenn : You relaize that a parallel port does 115K BYTES/seconds right?
Jymmm: where do you get that number?
cradek (old standard bi-dir)
that's for a stream of data
cradek it wasn't the number i was reffering to... it was bits -vs- bytes
you have to know what you want to send before you send it
the CPU can't keep up with that rate, is what i mean
er, i guess it's the motherboard can't get data from the PP to the CPU and back to the PP fast enough
and a 400MB/ firewire can?
SWP was explaining it to me
well, the cpu is multitasking
well, even without worrying about multitasking
wonder if i should butcher it and try to explain it to you or just try to find the log
[03:50:34] <Jymmm> http://www.atmel.com/dyn/resources/prod_documents/DOC2506.PDF
well that doesnt mean you can do 30,000 read/writes a second from the cpu
er, wait n/m
what i meant was the PP can read/write 1MB/sec but you can't do 500,000 read-writes a second
which is why it's 60 kb
since you _can_ do that much
I wish I found find a small, inexpensive digital/analog scope
for measuring mains 60Hz cycle
can't you use your PC soundcard to do that?
you want to check for RF noise or something?
no not a PC based scope, a REAL scope
actually, I could use a specturm analyizer far more than a scope.
talk to phydbleep about that.. he built one I think
some kind of PIC
[04:07:41] <fenn> http://www.eio.com/g321skel.htm
This one will do (for now).... http://www.home.agilent.com/USeng/nav/-12093.536881892/pd.html
50 GHz huh
oh it's only $80k, where's my credit card
you dont know anyone with a scope?
heh. I found $300 bucks in a parking lot. 3x $100 bills. :)
might want to wash the blood/cocaine off of it
yea then burn it :)
the bleach will take care of that.
the money's earmarked... $30 to charity. and $25 for that stupid parking ticket I got today.
The rest goes towards my debt
I've got a reply on the USC
hey there anonimasu
how's things goind?
but that's normal when one's doing research
never seem to find what u want when u are wantinf it
i just finnished downloading bdi... want to install it on one of my old pc's to c how it work's
it works good but it's quite a mess to set up
that's the part i am scared of... not good with linux
paypal is www.paypal.com
well, they have my card number so they better be
well it's paid now
paid in eur instead of $
what u buy?
a hardware pulsegenerator
some one recomended that i use rhino 3d to do all my cad work... what a complex program
lol... looking for easy cad program :) is there such a thing?
not really :)
autocad is also one of those impossible program's to figure out
any cad program will be hard to learn
cradek: axis does an entirely wrong thing for a file like this one (slashes separate lines): G91 / G1X1 / X1 / M2. The preview plot is always shown anchored at (0,0,0) even though the machine will do some motions that are all relative to the current point
I just think nobody should write a program like that
hm, jepler do you know if you can do a full circle?
g2 x0 y0 r10
or do you have to split it in arcs?
I dont have the machine fired up so I can test it right now
03yabosukz * 10emc2/src/ (10 files in 7 dirs): fix make
anonimasu: not with the 'r' format, but with the 'i j' format.
anonimasu: you can do a full circle
oh! I only use ij
jepler: is G91 relative?
jepler: I guess I don't care about that then
jepler: I agree that that's a stupid program
jepler: with my latest fixes I should try a program that uses tool offsets
jepler: when you change tools, I think it should display the path correctly
here's a program that draws a full circle with only one g2: g0 x0 y0 / g2 x0 y0 i1 j0 f10 / m2
the similar program with r1 instead of i1 j0 is ambigious, because the center could be *anywhere* that is distance 1 from x0y0
gosh this # has been so dead the last 2 days
hm, my USC should be on it's way soon
wonder if logger_aj think's we care ?
do androids dream of electric sheep?
fenn_ is now known as fenn
[21:51:15] <Yuga> http://www.owned.com/Owned_Pictures/Funny_Owned_Pics/Match-making_gone_wrong_doggystyle/OWNED.html
that is the dumbest website i've ever seen
who knows bitbake !!
is that japanese? :)
hehe nope the specialists with linux knows were i talking about
sou desu ka
roel01: Will you explain what you mean with bitbaking
anonimasu: it's some kind of auto-compilation scheme like portage
fenn: yeah, but I've never ever seen it around
we used bitkeeper until they get hungry $$
you mean a RCS tool...
i'm reading the history of bitkeeper.. the drama never seems to end, does it?
[22:25:22] <roel01> http://svn.berlios.de/bitbake/
this is the site we use untill yesterday
svn: Can't connect to host 'svn.berlios.de': Connection refused
ah.. it's not there because you aren't using it anymore, right?
is there any other site for this
a site that hosts bitbake repositories?
are there multiple software products called bitbake?
maybe a miiror site
[22:31:31] <fenn> http://svn.berlios.de/viewcvs/bitbake/
you should ask the bitbake developers if you really need cvs access to bitbake
they probably know more about it than we do :)
* anonimasu is looking forward for the USC to arrive
how's things gound?
welcome robin :)
oops was afk
going OK, another busy day of cutting
robin stil playing with lasergun :)
robin_sz I wish I was so lucky!
roel01: running the big laser for commercial cutting at the moment, seems to be profitable
thinking about getting that Co2 laser running as a second machine
and building an auto sheet loader
i work a long wile ago with two punch machines with in the middle a sheet loader
we got rid of our punch
a Rhoades Pearsall 12 station CNC punch ... hateful device
we used salvagnini s4 back in 1990 or so
like windows (error da cap - manny times ) :)