are you talking about lines 48,49?
it is 47 and 52 here
f = n = [y1 + self.offset_y, z1 + self.offset_z, x1 + self.offset_x, a, b, c, u, v, w]
it looks right to me
I changed it
I suspect you broke it then
have you found any bad behavior before or since you changed it?
f = n = [y1+self.offset_x,z1+self.offset_y,x1+self.offset_z, a, b, c, u, v, w] -- befoe
you should test previews of arcs in different planes
can you tell me how to reproduce the problem you saw?
I just think it is wrong to add offset_x to anything, but x. I may be wrong :)
did you understand my question? I was asking if you saw an incorrect preview and it led you here, or did you only see the code and think it looks odd?
no. just looking into the code.
if you load 3dtest.ngc and set some offsets, I think you will see that your change breaks arc previews
arc_feed is a strange canon call. it is plane-agnostic, meaning the order of the arguments changes as the active plane changes
I am nearly 100% sure that code is correct
I'd say it is broken now. Did you looked at it?
look at what?
the code, or the preview plot?
yes, and it looks right
xy plot is circle, both others are broken
with your changed code, or with trunk as-is?
send me the gcode that shows the breakage please
So anybody up for some Hostmot2 SPI talk?
this is 3dtest.ngc you've just mentioned -- backplot looks ok, BTW
3dtest.ngc looks fine here
make sure you have unmodified trunk and try again
doing what right now
ok, this is my change, that broke it. Sorry for the noise.
SWPadnos: I have an external PWM generator now. LM628 same chip used on the mesa 4127 However I have some uncertainty's about how I can get the commands from EMC to the chip, everything else should be already set up. Where might I find an example or manual for emc to lm628
UncleG_, I don't really have any time tonight, but I will say that the LM628 won't work with EMC2
the LM628 is a motion controller chip which uses a parallel bus to talk to a host CPU. EMC2 can't really use an external controller, and there's no parallel bus to connect the chip to
yep. sorry about that
thanks tho :)
If anybody in interested, or has some time later, here is a better idea for hostmot2 SPI... http://geo01005-ideas.blogspot.com/2009/02/basic-hm2-spi-driver.html
better, than my previous ideas... I hope.
looks pretty good to me
one thing though - there is always data being transmitted and received
how do you mean?
it's a question of whether it's relevant or not
well every time you clock a bit, one goes in and one goes out
in other words, the devices will all be reading some data whenever they receive a clock. they may ignore it (and only shift it through to their SDO pin)
similarly, the receive register on the FPGA card will be clocking data in, though it may not be relevant
so the RECEIVE HAL pins can just be left unconnected if they have useless data
Yes, that is true, but you would have to specify that on the loadrt line, right?
any thing that needs to send a specific value (like "start ADC conversion") can have the SEND pin set to a constant
I just think the READ (BIT) may not be necessary
and maybe ENABLE also
the number of bits you need to send/receive is determined by the attached devices
so disabling transmission of one frame would cause the bits after that frame to end up in the wrong chip
(and before actually - you wouldn't send enough clocks to get the data all the way to the end of the chain)
so it makes it easier :)
The reason for the read bit is so that say you have a ADC you want to read and act on within the thread cycle, and perhaps you have a PWM that is calculated based on the ADC, both on SPI.
you only need the CD (some flags and the number of bits), and the send and receive data for each word
You would want to read the ADC on the HM2 read function and write the PWM on the write function.
I wasn't sure about the enable...
they may be useful then
anyway, got to go. like I said though, looks pretty good to me
I'll try to put it on the Wiki later.
Have good one.
see you UncleG_
thanks. headed to Honolulu tomorrow morning - gotta finish packing :)
I'll hold off hounding you with questions till you get back home
cradek: how about renaming arguments to something like this, to avoid future confusion ? def arc_feed(self, e1, e2, c1, c2, rot, ea, a, b, c, u, v, w)
[01:33:52] <maddash> http://126.96.36.199/mp/rmi/_dlp/MBIO/Lecture%2001.wmv?advc=rmi&advv=1.104.1&advx=1235352488&advt=6m0nZbtlaG5IA7W8C62hflXi85u2nV2dhivD66EqatCwZFR8BWiLnf+hoUcHLzDKHX3cMVolAtpFfbm1keom4Q==&MSWMExt=.asf
can someone try streaming that? it won't work for me
what do you see?
* fenn blinks confusedly and looks at the channel topic
ah, so it's not #diybio after all
aystarik: what are you using to view it?
wtf, doesn't work in mine
aystarik: do you mind: 'mplayer --version'
Crouching Tiger Hidden Dragon, Desperado, and 5th Element all three sealed for $5.99 - not bad at all
with EMC it seems possible to use servo motors, but what I can't figure out is if the encoders are read via parallel port or by a counter chip
there are several options
the best way to run servos is with special interface hardware - you are right that the quadrature encoder signal gets counted in hardware and emc reads that
so the hardware sends via serial interface the encoder position
mesa and pico make hardware that does this
mesa is parport or PCI, pico is parport
not serial - serial is too slow
well the only quadrature counter chips I"ve used use serial interfaces (not serial port)
there are other options too, but mesa and pico are the best choices
googling mesa and pico isn't very specific
pretty sure they both use FPGA
do you have a more specific title or a link?
ic thank you
smart to put emc2 in there lol
my google skills are second to none
my big lathe uses the mesa 5i20 interface and I am perfectly happy with it
it also uses resolver-to-quadrature converter boards from pico systems
I kind of wanted to create my own electronics but w/minimal software adapation
I've created measuring systems with microcontrollers, QEI counter chips, and optical encoders
cradek: find how to find unique ID on Movie DVD's =)
I'm building a PCB mill first with steppers
so after that I'll probably have a better understanding of how to make my own electronics
I appreciate the help
You know, they spent more money in the '3 pack' packaging itself than they did on the media. All brushed silver and a raised bubble badge.
any body here have an idea about how to get halscope to load? It normally does but it seems I have somehow managed to screw it up. It gives an error about too many active channels, and quits
i think i was using it, then ran stepconf to invert a signal, and now it won't load
probably got some setting screwy where it's confused now I guess, but reloading emc and hal didn't seem to help any
remove .scope.cfg from either the current directory, your home directory, or the directory that holds the config you're loading
or all 3
i was thinking there might be a trick along those lines
tricky hidden files :)
file not found
find / -name .scope.cfg
ah find, i always forget about that
it wouldn't be any different in the pre-2.3 head version would it
I don't think so
actually, do you get that error immediately, or after you select which thread to sample in (and how many channels/samples per channeL)?
i did find an autosave.halscope file in my configs directory
and now halscope loads
there does not appear to be any .scope.cfg anywehre in my system
but problem solved
how i love the smell of solder flux...
ok, that's fine
thank you for your help
i think i owe you more than a few brews by now
hey great the spindle index pulse really works now... thanks SWPadnos
if I could just remember the difference between the emitter and collector on a transistor it would have been working about 2 circuits ago, but third time's a charm
so anybody tried cutting npt threads using emc?
I haven't done it, but I probably will one of these days
there's no canned cycle so you'd have to construct it with a g33 for each pass
steve_stallings is now known as steves_logging
the G76 won't do it?
what is the "entry" and "exit" tapers thing then, I was hoping that might be able to do it for me
something for easier starting i guess?
hm, maybe you could do it with an "exit" taper
make the exit taper most of the length
I didn't think of that.
it says its an error if E is > 1/2 the drive length
(in the manual)
what if E=0.5 P, and you use both entry and exit tapers with opposing signs
I don't think that's actually an error
looking at the code...
let me know if you try it!
it would be convenient if it wasn't an error
i'm not exactly sure I'm understanding what the "entry" and "exit" tapers are
if I'm reading it right, making the "entry" taper being equal to the NPT required taper should give the proper taper on the threads
you'd use the exit taper I think
yes if threading right to left like normal, moving left and X outward is the "exit"
that's the one you want to build your thread with
oh i see
entry taper is like when the tailstock is in the way or something - rarely used I bet
i was actually going to be threading this part from the left to right, with the small end of the thread on the left (near the spindle)
i have a relief groove cut in the part
I don't think I've ever threaded left to right with g76
now that I think about it that would cut LH threads
well, depends which way the spindle is turning
you could put the tool upside-down
yeah my spindle runs forward only though
heh, better thread right to left then :-)
it runs backwards slowly but doesn't have very good speed control
or any at all
(i'm not sure why, in theory it should run fine in either direction)
some problem with the vfd i guess
but i guess i could start at the right with a relief cut, and thread the taper to be smaller
i think i will take your advice and make a G33 cycle for it... seems far safer
yeah you could still do the small side on the left
G33 can specify X values right
yes G33 can go any direction
you would not need a relief on the left. just g33 over and g0 out. it will not pause there
oh that's nice
i already have the relief for my LH turning tool though, so no biggie
it won't fit without a big honkin relief cut
what's the part?
its a npt to 10mm adapter
trying to do it all at once, without rechucking it
1/8 npt external to M10x1.0 internal
are you single-pointing the inside thread too?
the internal threads are easy enough and obviously have to be accessible on the non-chucked end, so that means the 1/8-npt has to be on the chucked end
(that's a bit on the small side)
yeah attempting to
I'd rigid tap it with a spiral point tap, assuming it's drilled through
it works in plastic, but stainless may not be so easy on the tooling
303 at least
it's not so awful as 304 is
oh you don't have a good spindle reverse - forget it
may wind up hand tapping the lot if the internal threading doesn't work out
is part of it hex, or just the threads?
part of it is hex
sounds cool - good luck and make us a video :-)
I have 80% of the program done, just gotta work out the npt part now
i will make a video when it's successful
so far it's bloated to 7 tool changes
I know how that goes
if I figure out a good way to do the npt threads, how hard would it be to make a G76.1 to canned cycle it?
or maybe some other code for npt, not sure what the normal method is
the hard part is figuring out how to specify the necessary things in the gcode. the math isn't hard. I'd happily review a patch that does that.
although if g76 can already be coerced into doing it (and I think it can) it would probably be wrong to add another code
I'll have to experiment one of these days, or if you try it, please let me know
at the moment I need to get these parts made, so I'm thinking G33 is the fastest way to do that. but I will definitely play with G76 to see what it can do with the exit taper when i'm done
i am concerned that it is explicitly stated in the manual about the error for being over half the drive length
guess there's one way to find out though
now G33 says the K value is for the total distance travelled in XYZ, so I guess you have to trig out the taper to get an accurate tpi in the Z when you're cutting a tapered threaed
so a 1/10 thread would normally have a K value of 0.10, but with the taper it would have to be slightly less
assuming NPT pitch is along Z, not along the taper (I have no idea without looking)
yeah it is
you are right then, you have to tweak K
I think it will be more, not less
ah yeah that's right
its like a 3.6 degree taper, it might not be so much
.015 over a .25 inch long thread
.1 / cos(3.6) = .10019
yeah that's right
not much difference, like you said
almost a thousandth over my 10mm length
maybe not so significant
thanks for your help cradek
[04:48:38] <maddash> http://www.youtube.com/watch?v=EKL6elkbFy0
i think he could've done without the steps prior to the taper
that machine looks like it has the torque to manage
very cool though
hooray emc users list re-issued their notice of dropping me, now i can reply (did)
haha 'invalid confirmation string' after 'leaving the Subject: Line intact" useless bureaucracy!
if the system knows the string has expired, then why does it keep sending messages! typical
this is handier, https://lists.sourceforge.net/lists/listinfo/emc-users
just invent a bozo name and join again
argh, after i shake off the aggravation,i realize there was absolutely no attempt to find the cause. which means its likely to happen again
wasn't there a "visit the confirmation page" link?
yes, and i did, and i think it will happen again unless the problem is understood
well, that's probably true
it could be related to the internal mail isues SF is having - there have been a few times where they've blacklisted some of their own servers for sending too many emails ;)
not your problem, i'm always happier with a cloud on the horizon anyway
hah! i bet that was noisy
excellent. I'll just go shower and head for Honolulu then :)
hey have fun beach boy!
I'll be there in about 15 hours :)
see you later
i grew up in chicago, on the south side is 'Blue Island' a suitor of one sister always showed up in hawian shirts and a big packard with hula girl bobble heads in rear window. till i was 16 i thought blue island was tropical ;)
the suitor was from 'blue island' (boy i can f up a good strory)
this gets better , just got a 'mail privacy alert' ( i subscribed with my old name not bozo name) "it is possible a bad person was trying to probe the list for memberships'... 'you are already subscribed' .... in the end it says it didnt work & it did work !! clever!
this is jerry lewis logic "he's a nice man' 'shes a bad lady'
[Server Notice] Hi all, the server you are currently connected to (lindbohm.freenode.net) will go down for scheduled maintenance tomorrow (February 24th 2009) at 0900GMT. You may wish to re-connect to a server in main rotation by changing to irc.freenode.net prior to this time. Thank you for using freenode and apologies for any inconvenience this may cause.
[14:29:53] <BJT-Work> http://www.cnczone.com/forums/showthread.php?t=73943
Good evening.I have one curiosity.The emc is possible compile on linux for powerpc processor ?
in simulation mode yes
becouse only in simulation ?
not exist realtime for ppc
the hardware driver not exist for ppc ?
usually ppc doesn't have a parport
maybe one could try to compile a pci driver on ppc.. but I don't know anything about ppc to comment
I saw talks on the rtai list lately about ppc, so maybe that works too
ok thanks alex.i thing one version for amigaos ppc , but it is only love for amiga platform.
i was wondering if someone could give me a little help on an parallel port issue
just ask your question, and if someone is around to answer it, they will
the problem is, i installed the package based on ubuntu 8.04, configured it but it does not work propperly
the pins 1-4 of the parallel port are working perfectly (2 stepx, 3 dirx, 4 stept, 5 diry,...etc)
but the pin 5, 6, 7 do not work at all
i tried all possible configurations from bios
but no succes, i even tried installing the whole package on an other computer but same results
does anybody have a clue about what it may be?
how did you decide they don't work?
i meassured the voltage on the diry pin (5) and dirz pin(6) and while "moving" the machine in both directions no change was there
the stepy works, the drive moves allways in the same direction
i also tried to do it by the hal configuration
just start HAL (no emc2)
halrun -i (but read the manual to be sure)
loadrt hal_parport cfg=0x378
i i tried it to
loadrt threads (check man threads how to specify the rate)
addf parport.0.read threadname
addf parport.0.write threadname
without a thread, loading the hal_parport won't do anything
but still, the x axis works fine, the y axis moves but only one direction, the z axis does not move at all
what could be the problem?
the port works physikaly fine
how do you connect things?
parport -> ... ?
the connection is ok
I mean, do you have an opto board?
does it go directly to the drives?
the cables are ok, i measured them
what kind of drives
there are general numeric drives
justt like the geko
I don't understand "general numeric drives"
but they work like the stepper drives
i just need dir and step
I suspect wither the timing is wrong
or there is some electrical problem
either not wither
all drives work well, timing is ok, i connected them to the pins 2 and 3 and worked perfectly
electrically it should be ok
what voltages do you get for on/off ?
it's a compaq
(without drives attached, and with drives attached)
i get around 4 V for on and 0 for off
without drives 5/0
on all of them?
no clue what else to check..
i put a resistor between the drives and pc to make the amps smaller
do these drives have optocouplers?
(for example the gecko's have some)
the drives are very professional
industrial ones work at -20 degrees perfectly
then I would check (with an oscilloscope) before and after the optos
i don't think the problem is the drives
the problem must be the software somehow
I didn't say the drives are the problem
I'm fairly sure it's not a software problem
because every drive connected to pin 2 and 3 works perfectly
what configuration did you run?
what do you mean by configuration
i configured my own system
with stepconf ?
try one of the sample configs
when you start emc2 from the menu, you get the config picker
alex_joni, are you sure your getting 0v on paralel port? 0,6 might give wrong information to drivers( this means the paralel port is not consuming the necessary current
can you show me your drivers schematic?
sorry, i don't have the schematic of the drives
alexfloc: try the stepper/stepper_mm or stepper_inch config
but right after the optocoupler it goes into atmega 128
ok then i sugest getting a paralel port tester
where you can set lpt bits by the press of a mouse button
step/dir going to microcontroller: big red warning light
maybe your timings are wrong. step pulses are too short.
no, the drives work fine on pin 2 and 3
because microcontrollers poll their inputs and are slow
pin 4 and 1 also work fine, the problem is after pin 5
pin 1 is input
yes i know
then maybe your parport is bad
i have the machine emergency stop on it
the only way for a parport to stop working after a certain pin is if it is defective
i tried an other computer too
the same problem :(
but since you tested another computer, then I bet the problem is in timing or electrical
then maybe your wiring is wrong
but you need to troubleshoot. use a logic probe or scope and see what's happening.
i tried it, i get nothing on pin 5
we can't guess (and when we do, you just say we're wrong)
alexfloc: try a sample configuration for stepper (not one generated by stepconf), if the drives (for which you didn't say you have timings) have odd timing restrictions, then you can sometimes get results, sometimes not
with emc2 running try this:
open a terminal
type "halcmd -kf"
then "unlinkp parport.0.pin-*-out"
where * is the pin number you want
(you can use tab complete for this)
then you can use commands like this:
i did that allready
setp parport.0.pin-05-out 1
i get no voltage on that pins
and check the voltage
are you sure those pins are output?
but an old printer works
that is the problem
pin 2-9 are output
databits fot printing
i will try again tomorrow
try this http://www.beyondlogic.org/pardebug/pdebug.htm
then you cab make sure if problem is in pc or drivers
or you can also inject 5 and 0 volts to those pins without pc
to see if drivers are working fine
put a small resistence on 5 volts
for the test
lol he left
[Global Notice] Hi all, as you probably noticed our services took a short break there. They're back online now and we're investigating the cause of the crash. Thanks for using freenode and sorry for the inconvenience!
excess speed crash or driver was drunk
good night all