Jymmmm: that bottle must be empty?
i'm trying to share homes with limit switches
i've read the user man, int man and the wiki by fred proctor
can anyone point me to an example hal config that works this?
twice2: I don't think any of the sample configs do this
twice2: the nist-lathe and sim/axis configurations do show homing order with a home switch that is shared between axes, but I don't think that's what you're asking abuot.
thank you both, jepler: i have noted your suggestion. but i'm looking for shared home
jepler: yes those configs are good for my setup, maybe i need to somehow ignore limits during the homing
twice2: the docs say what to add to your ini to do that
I forget what it is
sorry, jepler i've just scratched the hal to the point of getting a limit a n.c. switch working
btw, is it ok to tie the pin directly to ground?
any way that's not stuff for this channel so i need to find anything on shared home/limit switches
ejholmgren: I can only hope for =)
twice2: you hook the signal that comes from your switch to both the limit and home inputs in HAL, then set the ini so limits are ignored during homing
twice2: here are some relevant parts of the documentation: http://linuxcnc.org/docs/2.1/html/config/ini_homing/index.html http://linuxcnc.org/docs/2.1/html/config/emc2hal/index.html
for the .ini and .hal stuff respectively
thank you guys!
twice2: welcome, hope you get it going soon
hehe, i prob shouldn't be trying this stuff, makes my dog guide spooky
03cradek 07v2_1_branch * 10emc2/src/Makefile: sim makes .so, not .ko
03cradek 07v2_1_branch * 10emc2/VERSION: 2.1.1 release
03cradek 07v2_1_branch * 10emc2/debian/changelog: 2.1.1 release
ok, i've copied all of your input and i'm 'going in' , thanks again, emc2 is awesome
03cradek 07v2_1_branch * 10emc2/VERSION: bump after release
03cradek 07TRUNK * 10emc2/src/Makefile: sim makes .so, not .ko
EMC 2.1.1 is released
jepler has changed the topic to: Welcome! EMC (Enhanced Machine Controller) is a linux-based CNC control. | Latest release: EMC 2.1.1 | http://www.linuxcnc.org
thanks for actually making the relesae
I got it right in only a few tries
wish alex was still up - he always does the news items on linuxcnc, etc
I think I got most version number mentions on the wiki
ooh, this looks cool: http://www.mindaugas.com/projects/MJoy/
GPL firmware for AVR, acts like a USB joystick with 6 axes and 24 buttons
03jepler 07TRUNK * 10emc2/lib/python/linux_event.py: add GPL license notice
03jepler 07TRUNK * 10emc2/src/hal/user_comps/hal_input.py: add GPL license notice
tip: a chip lathe toolbit makes a big difference in the ability to acheive an acceptable finish!
skullworks-away is now known as skullworks-sore
I just don't understand chi machine builders
they grind all over - except where they really should.
table sides, ground, saddle face, ground, ways - not ground
ways don't really show up on pictures ;)
but there going to see some pics of this
maybe you got it from a surplus batch that was made for a ways scraping class? ;)
they did a roughing pass, then a finish pass, but the rougher must have deflected so much that the finish tool did not clean up.
what kind of machine is this?
Seig X3 via Grizzley
guess you got a bum machine; the mini-lathe.com guy seemed happy with his
I just spent the afternoon tearing it appart into (nearly) manageable chunks to move it home from the shop.
just lots of award lifting
didn't get sand and red grease all over?
red grease - green grease - yup, went thru a bunch of used paper towels doing the first cleanup.
We keep all the used towels people use to dry there hands bagged
used for oil spill cleanup
I think I'm going to set this up like the Tormach CNC
put dual pulleys on motor and spindle - run direct drive for Hi/lo
the gears are sloppy loose
no wonder everyone complains about the noise.
so - bye bye gears
going to put in a 3PH VFD motor?
2hp DC treadmill motor
how much are those going for thesedays?
with additional heatsinking and AC continous cooling fan.
I want to talk to Jon about driving it with one of his servo amps
Hmmmm interesting idea
btw, not sure if you know but there is a SCR based controller for $30
yep puts me that much closer to being able to rigid tap
that's like the holy grail of home built machines
yes There are SCR units - will cost more like $89 to have enough capacity.
where are you looking?
mainly surplus center
oh you are right.. the one I am thinking of is good to about 1HP
need 110V at 18amp
that pushes the upper limit for most units
how much are the servo amps?
but - then I would not need to build the linier power supply :)
amps are $125
but there is a driver issue
Linear PS for a ~1.5KW?!
is that where you cast the PCB out of solid copper? ;)
they are intended to be driven by his controller board
no - not the powersupply - the servo amps
if you can get rigid tapping, I'd be interested in doing a smaller (<1HP) version
the power supply would be a dual coil 80VAC @ 18 amps - a bridge rectifier, and a bank of caps
the x3 comes with a 4000rpm .8hp motor as is
but that's .8 import midget horses ;)
but stock gearing tops out at 2000rpm at the spindle
my target is 5000
is it balanced enough to do 5000?
the minilathes sound scary at 2000RPM
the spindle itself
treadmill motor is 6750 rpm
thats something else
stock spindle hangs out - canterlevered too far
then there is the stock bearings
I want to make a new bearing housing that shortens this up by almost 2 inches
what are you going to do on it to want the 5K?
also my conversion shoul add almost 5" z travel
5K is creeping
I'm used to running at 10-12k
what do you do at 10K+?
Okuma's and Mori-Seiki's
I find myself wanting to go slower usually
high speed machining
but that's stuff stuff with 1/4" or smaller carbide bits, no?
filling 55 gal drums with chips each shift
no thats with 80mm favemills taking .300" depth of cut at 280 IPM
or 3/4" em
the Mori has a 30hp spindle - and I love to run up the company's electric bill
9.2hp for each axis servo
and this is an "old" machine
Now - I have an IH mill on order
I can see why they have doors on those now... just didn't seem justified on those puny 1-2HP HAAS minimills ;)
Haas has had doors since there tupperware days
my other job - I run 3 haas
2 VF0E and one old VF0 tupperware
the 2 VF0E have robots feeding them (I'm lazy)
that IH mill
I may bust my wallet and order ground ballscrewws for it
but when I'm done I want it to be an EMC poster child, with min 400ipm rapids and rigid tapping.
hehe... make a video of it doing #0-80 holes ;)
that puts us back at the level FADAL was in 1986
so - EMC would still only be 10years behind leading edge
consider the price of a leading edve control
and that is something to be proud of because this whole effort runs on virtually no budget.
trust me I know the prices
and how they squeeze you for every little option
its a joke that you can spend $100K+ for a machine and they give you 64K user mem.
there is a full 8mb installed but they use paramiters to disable the rest of the mem.
john - what state you live?
seems most the active folks are farther east
Jon in MO
cradek and jepler are in nebraska
tomp is in illinois I think
rayh in michigan
I'd love to sit in on some of the get togaters like cabin fever, but its easier to get a root canal than more than 2 days off
* skullworks-sore has had the root canals to prove it
did you guys see this stuff? http://5axes.free.fr/
stock removal simul8r, 5 axis code, evevn some apt stuff... pretty dense site
too dense for me - bedtime here
john - nite -
past my bedtime too
Never thought about this before, but is it posisble to jog using a mouse?
Jymmmmmm: jepler just added a component to emc2 that exports HAL pins for _any_ input device
so you can hook up the pins exported from a mouse to halui for example
[08:57:38] <alex_joni> http://linuxcnc.org/docs/devel/html/man/man1/hal_input.1.html
I'll be runnnig my plasma table today
the T arrangement isnt a problem..
got any pics?
too bad :)
I've got the torch holder in the mill right now :)
going to finish it up on the afternoon
and wire up the last gecko..
and hook up my scope
but I doubt EMI will pose a problem
unitronic bus cabling is neat ;)
you need EMC shielding :P
alex_joni: does that involve a cement truck?
jlmjvm: did you update to 2.1.1 ?
not yet,is it ready
yes, already released
jlmjvm: if your Ubuntu is connected to the internet you should get an automatic update
its not showing it yet,im on the internet
jlmjvm: you do have 2.1.0 .. right?
you'll get it
you can activate the check for updates now option somewhere
(forgot the exact place)
im glad i found you guys,its quite a task to learn linux
i tried that already
what all did yall change in 2.1.1
[11:49:03] <alex_joni> https://sourceforge.net/project/shownotes.php?group_id=6744&release_id=489336
i just got it
the "enter" button on my keypad works
the auto updater work flawlessly
that is a big help for new people
what got added in 2.1.1
[12:12:03] <alex_joni> https://sourceforge.net/project/shownotes.php?group_id=6744&release_id=489336
am i reading this right,the hal joystick is now part of emc 2.1.1
where can i get SCARA at
atkbd.c failed to enable keyboardon isa0060/serio1
Alex - just for giigles - how often to you rebuild from head?
ubuntu seems broken on this pc
drop back and punt?
(bad football analogy)
jlmjvm: the scara configuration with the simulated robot is in emc2 cvs trunk
still nursing that 1st cup of coffee, reading with one eye open
cradek: got any idea why this stuff happens to me?
evil past life?
that may be it
alex: I've read the new page in the wiki about joint axes etc.
in the scara robot example maybe AXIS number should be added
I was just wondering how often the Developer crew rebuild there test box from head?
things seem to change so fast, on many fronts at the same time.
fsck, some checksum failed.
xemet: didn't understand that
in the YA example, in the imaginary ini file you specify Joint_0 type linear AXIS = 1
yeah, that's a special case for trivkins
alex: in the YA example in the ini imaginary file you specify Joint_0 Type linear, AXIS = 1
because you want that the joint_0 control the Y axis
so when you specify Joint_1 you say type angular, Axis = 3
yeah, that's a special case for trivkins
because you want it to control the A axis
so, in the scare example
scara doesn't have a 1-1 mapping from joints to axes
should axes numbers be specified?
well I'm off to push the green button. - Have fun folks.
but scara moves in the XYZ space...so when you say joint_0 angular...the gui interface should know what to move when joint 0 moves...
xemet: the GUI doesn't care
xemet: it's teh motion controller that does the conversion from carthesian to joint space
ok, and GUI shows it taking values from motion controller
ok, but using the GUI, for example axis
if I'm with the scara
I will have 4 axes, as described in the ini
no, you will have 6 axes
you will have 4 joints
I would like to understand this
what will I control from gui?
the cartesian coordinates?
first you control joints
or the joint position?
you can move each joint, and home them
once all have been homed you can switch to world view (carthesian)
now you control the position of the tool tip in carthesian space
so if I want to move X 10
(e.g. joggin in X, Y, Z)
just I type G1 X10
the program you run is G1 X10, right
and the motion controller take care to convert it in the joint moves
but you _must_ home all joints before you do that
ok, of course
xemet: yes, that already works
try the scara example
I've to run
see you later
cradek: good morning
thanks for doing the linuxcnc.org news
hope it's ok..
I also sent an email to linuxcnc.org registered users
yeah I saw - I have mixed feelings about that
bbl, getting ready for work
alex_joni: that change I made didn't work for the smi.
I will do it again post the error
good morning all
I wonder if the timing required by USB (time spent in interrupts) makes an AVR USB jogwheel infeasible
jepler: good morning.
jepler: did you end up getting your joystick yet?
skunkworks: yes, I got the "spaceball" friday and played with it this weekend
it's quite sensitive, and very hard to move on just the desired axis
I hacked a version of quake (zquake) to use it as an input device but I mostly spent time bumping into walls, staring at the ceiling or floor, and falling into pools of slime
maybe I need to reduce the overall sensitivity or the "dead band" around the device's center
now someone needs to send me one of these usb jogwheels so I can verify it works with the new hal_input
cradek: yeah, I know .. me too.. it was more of a test and see what kind of reactions I get
so far I only got one reply: "Hi Alex, Thanks for the Update! Mark"
alex_joni: never mind. It seems to work.. I wonder if loading the rtai_smi manually screwed up the loading of emc last night. Seems to be working today.
alex_joni: yep - seems to be working great. No unexpected realtime delays.
alex_joni: I just noticed that my name is not on the list of EMC developers. I'm not sure if that is a good thing or a bad thing. :-) Before adding my name to the list please show a name change to Aaron :-) That would place me first in the list.
alex: might be a good idea to put the list on the wiki, make a contributor page ?
lerman: that list is referring to emc2 only, maybe you forgot to add yourself there when you commited stuff to emc2 ?
Correct. I never did add myself. But I blame you anyway :-) signed: Aaron lerman Aardvark.
I'm making progress on my new version of GWiz. See: https://www.se-ltd.com/~lerman/gallery/G-Wiz/Pocket2
for the latest.
03alex_joni 07TRUNK * 10emc2/docs/AUTHORS: add Aaron lerman Aardvark.
alex_joni: my emc2 wont install
for some reason
lerman: is any of your gwiz code available for download ?
does ubuntu work?
trying safemode now
that's what imtrno idea, I dont have ubuntu(other then the livecd)
anonimasu: you mean the liveCD ?
that's not really emc2
and I'll have to point you to #ubuntu for specifics (I only installed it a couple of times without problems)
alex_joni: gee thanks..
hm, gdm firing up now..
works with safemode.
awallin: it is not ready for prime time, yet, but I'll but a tar file up on may machine. http://www.se-ltd.com/~lerman/files/gwiz.tar.gz
Note that the directory contains a lot of junk -- also tested mainly on windows with Python 2.5. Use at your own risk. It really isn't ready to cut metal, yet.
...But if you want to look at it and comment (read the code, even) it sure would be appreciated. I'm a novice at python and wxPython is far from well documented (or I haven't found it yet).
how long does loading the livecd usually take?
lerman: thanks, I'll look at the code. Mainly I'm interested in the functions that actually create G-code (I might use those for some of my own stuff). Can I still reach you by email at lerman-awallin "at" se-ltd.com I might be able to help you out with the wxpython
Yup. That email wil work. There are no functions that create gcode. More precisely, the only function that creates gcode is one that generates a subroutine call.
anonimasu: depends on CPU
but usually 5-10 minutes max
anonimasu: gotta remember that 256<B is absolute minimum
anonimasu: gotta remember that 256MB RAM is absolute minimum
this doesnt wor.
x starts and it locks up
anonimasu: did you test the CD?
I mean after burning
it failed, but im sure it's the cd reader..
which I replaced..
this is the 2nd cd that's faulty in that case..
and I burned it slower then usual..
as the first one were broken
I did md5 the cd before recording also
people must be busy :)
skunkworks: I was having a look at http://elm-chan.org/works/smc/report_e.html
pier_gar: cool little package. (emc will do the pid on my setup)
yes, as a matter of fact pwm can be performed by emc
and pwm ;)
this morning instead I had a look at the UHU project
always step/dir but pid on board
pier_gar: have you seen this? one of my projects http://www.cnczone.com/forums/showthread.php?t=25929
* pier_gar reading
although now I am moving towards using hardware.. pluto and mesa.
[17:51:20] <pier_gar> http://emergent.unpy.net/files/projects/01142347802/servo-sch.png
is this yours
nope - that is jepler
pier_gar: he's amongst the big boys now :D
this is the schematic in eagle for mine. But the parts where picked strictly for size - not acutal componants.
[17:54:06] <skunkworks> http://www.electronicsam.com/images/KandT/servostart/schem/
I have to make up a parts list
* pier_gar already looking at it
did you see that the change to the rtai.conf worked?
skunkworks: yes I had already had a look at it... I remember the picture of the lab
lab - heh heh heh
heh, nice name for that
skunkworks: I'll have a deep look at it as I am not as much with electronics
I have a degree in it but have not done anything except tinker. ;)
does emc do the control part and your interfece the power?
* pier_gar opened it in eagle
The h-bridge has 2 inputs - pwm-cw and pwm-ccw
so emc or pluto outputs and up/down pwm signal to runit.
its optically isolated but no current limiting yet.
is the h bridge with power transistor mosfet?
I read l298 somewhere..
perhaps in the cnc forum
that is what jepler and cradek used for thier stuff
* pier_gar too
it is good for 2 amps but I think it can be paralleled.
until I get hold of some good servo
yes max 2A
but as I love rummaging in the scrap heap I don't think I'll be able to get such nice thing done
pier_gar: where are you located?
in the garage
cold here too.. 4C today
no heater as wife's going to grill salmon...
we just got almost 2ft of snow this last weekend.
pier_gar: that's yummy
pier_gar: what's on the side?
none here... not once to ski this season
* alex_joni likes penne al salmone
no I don't eat pasta
but at weekend
once a week is enough
* alex_joni goes home..
too mauch carbo and waist goes out of control
any idea on where to buy dc (not much expensive) motors and encoders?
skunkworks: I would like to try your interface
pier_gar: ebay? I really don't know other wise.
I just use junk ;)
skunkworks: have you tried wipers' motors?
no - someone on here said toyota heater fan motors are really good. - that and a hacked mouse encoder :)
I read somewhere that starters and fans motor are not suitable
I have plenty on unused printers encoders
with on-board circuit... I'll have to count the steps with a lent
I think the main thing is - how many poles. There is a cogging effect when you use non servo dc motors. I had asked the toyota heater motor guy and he said cogging was minimal.
* pier_gar looking up for cogging
assuming I spelled it right
not regular angular velocity
he also has a servo controller
[18:42:44] <skunkworks> http://members.shaw.ca/swstuff/dspic-servo.html
skunkworks: I dismantled an old mouse on that purpose but the segments where few
this is lawrenceG's site - he is on this channel
yes - but if you belt it down - it maybe enough resolution that you want. maybe. remember your counting all the edges - 60*4
just throwing stuff out there ;)
jepler: u around ?
awallin: I'm back now
awallin: thanks for the link to that book. I'm reading it now.
jepler: I was wondering about updating the contents of a tkinter frame. Is there a way to get a list of all the objects packed into the frame, so I can delete the old widgets and create new ones
lerman: hope it helps, someone else was already asking about gwiz on the list so you better release something soon ;)
awallin: in raw tk there's "winfo children", "grid slaves" and "pack slaves". Each returns a list of widgets. I assume these have been wrapped for Tkinter as winfo_children, grid_slaves and pack_slaves, but I'm not sure
yeah, where's gwiz ?
>>> t = Tkinter.Tk()
>>> b = Tkinter.Button(t)
[<Tkinter.Button instance at 0xb7e0a30c>]
[<Tkinter.Button instance at 0xb7e0a30c>]
Now that I'm using python and wxPython, I'm making good progress. It would go faster if I were familiar with either of them
ah, thanks. I hacked together something for editing the tool table, but then got distracted by this CAM stuff which I think is more fun...
I need some idons for push buttons. Approx 32x32. One for insert line (before). The other for insert wizard (before).
I'm not an artist (or graphic designer). idons -> icons.
jepler: in opengl, is it possible to draw some graphics with lighting on, and some with lighting off ?
awallin: yes. anytime outside of glBegin()/glEnd(), you can glEnable(GL_LIGHTING) or glDisable(GL_LIGHTING)
right. I'll try that
you can also turn individual lights on or off with glEnable(GL_LIGHTi)
Hows do you spell gcode? As in GWiz is a gcode Wizard. A GCode Wizard? A Gcode Wizard? A GCODE Wizard? Which is correct? Which looks good?
I haven't tried changing any lights yet, I've just used the default that glu gives me
lerman: I think it's G-Code Wizard
alex_joni: works for me.
did you get my e-mail about the spaceball?
I got an email you wrote also to vojtech
did you send other?
yes, that one
but did you succeed in using it with inputattach?
yes, after modifying the kernel driver
but I think maybe if you configure it for "right hand" use it will work without the changed kernel driver
what should I modify?
and...how to configure it?
on my spaceball there is a hand-rest that can be moved from one side of the spaceball to the other
also in the mine
in the right-hand position it closes a switch, which allows the computer to sense left-handed or right-handed
so if I change it...it change the configuration?
damn!! I had not thought it could activate a switch!
am I correct that your spaceball is configured for the left hand?
now that I see it...I think so
you might also be able to insert a piece of thick paper or a coin to make the switch be closed
now I try
I'm starting the other computer
jepler: about the G5 g-code you was adding to EMC2
is it possible to have the patch source?, I would like to study it
today I've carefully read your work in python with biarcs
I have a patch, but unfortunately it includes a bunch of unrelated stuff. http://emergent.unpy.net/index.cgi-files/sandbox/spline-and-unwanted-junk.patch
I haven't worked any further on this since the last time we spoke
now I've understood it, I've written may code and I'm going to apply it to approximate the nurbs
ok...also if it is incompleted, I would like to understand How to add a G-code
jepler: did you commit the velocity patch for the pluto?
skunkworks: yes, it should be in 2.1.1 and TRUNK
thanks - If you didn't I was going to say it works ;)
xemet: in the diff, I think these are the files that are important: src/emc/nml_intf/canon.hh src/emc/rs274ngc/gcodemodule.cc src/emc/rs274ngc/interp_array.cc src/emc/rs274ngc/interp_check.cc src/emc/rs274ngc/interp_convert.cc src/emc/rs274ngc/interp_internal.hh src/emc/rs274ngc/interp_write.cc src/emc/rs274ngc/rs274ngc.hh src/emc/task/emccanon.cc
xemet: the most cryptic change is probably in src/emc/rs274ngc/interp_array.cc where I make codes G5 and G5.1 (50th and 51st entries in the array) belong to modal group 1. Before they were -1 which meant invalid
xemet: as you can see I have to check in a bunch of places for conditions relating to the new codes, which are called G5 and G5_1 in the source
so If I download these with CVS I will get your modifications?
xemet: eventually it calls a function Interp::convert_spline() which looks at the supplied values such as block->i_number
xemet: no, these changes are not yet in CVS -- you would use "patch" to apply the changes from 'spline-and-unwanted-junk.patch' to your CVS checkout
damn, the spaceball now works :)
xemet: Inter::convert_spline() calls SPLINE_FEED(). SPLINE_FEED() is one of these so-called "canon calls", newly created for G5/G5.1. In the case of emc, emccanon.cc performs the biarc interpolation and calls ARC_FEED with the results. In the case of AXIS, gcodemodule.cc converts them into a number of lines and calls STRAIGHT_FEED with the results.
xemet: the task would have been much more complicated, going through several additional layers, if I had added splines all the way down into the realtime trajectory planner (aside from the difficulties of performing motion directly from spline curves)
the spaceball works, but maybe now I should make some kind of filter component in hal to filter axes
because If I move Y...X and Z moves too
xemet: before I will include the patch in the CVS version, I want to fix several things. First, I want to add support for splines in the other planes (YZ and XZ). Second, I want to allow movement in the axis perpendicular to the plane (linear interpolation like helical arcs). Third I want to use an adaptive method instead of fixed subdivision for the arcs.
so it is difficult to move only one axis
xemet: yes, I found it extremely hard to move just one axis on the spaceball as well
the original driver had a filter feature
xemet: anyway, once I can achieve all three of those things, I will feel that splines are ready to be included in emc.
well, others plane shouldn't be too difficult
I agree -- that's the simplest of the items
linear motion in the third axis is not too hard either, once I have found the arcs I can use the arc length as my distance function and easily find the Z coordinate for the end of each arc
I was thinking about that
the adaptive subdivision is the hardest, I'll need to study the papers in greather depth though
I've downloaded the paper you used
optimal biarc fitting etc. etc.
I mentioned one paper in my blog entry but I didn't actually use their method for finding the optimal t-hat
but you used the biarc formulation
I've still not read the part of your python relative to the spline
only the arc and biarc functions
and the giarc
I'm reading it now... so you calculate the cubic spline function,
the spline part just consists of dividing the spline into N+1 equal parts t=(0, 1/n, 2/n, ..., (n-1)/n, 1) and using the position and slope at each pair of successive locations to define a biarc
you increment the u parameter
in equal parts
this divides the spline in equal parts?
yes and no
in the nurbs, I cal points incrementing the parameter in equal parts
it's an equal division along the parameter 't' but each part might have a different length
infact I get shorter segments whe curvature is >
and you would like to subdivide it in order to obtain?
same lenght segments?
xemet: you need those to plan the trajectory (I think)
There are two cases: First, with no programmed tolerance, the arcs should be the minimum length that allows emc to deliver the requested speed
with programmed tolerance, the subdivision should be such that the tolerance parameter is met, even if it means lower speeds
and how do you know be the minimum length that allows emc to deliver the requested speed ?
chris and I have talked about that, and I'm not sure I have the necessary formula yet
ah, I've talked to joni two days ago about the possibility to have only tow axes in AXIS
is it possible?
lathe mode does that
lathe mode does X and Z only
no...in the mill mode
run sim/lathe.ini to see how that works
something similar could be done to allow XY only but nobody's asked for it or offered to write it
a friend of mine asked to me if it was possible to set two axes, one X linear and other A, angular
no, you have to have dummy Y and Z axes defined to do that
what kind of machine is this, with X and A axes only?
easter egg plotter?
I don't know
I think he has to work with stepper motor assembly
he assembles stepper motors
If I've understood well...but I've to investigate better
he asked it to me...and I asked it to you...:)
jepler: I am fairly confident only YA can be made to work, but there are certain parts of emc which need fixing
that's why I started http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?IniChanges
maybe a highly tuned armature wrapper :)
if axis has a nice way to get knowledge about the used axes and joints from the .ini I'll make it use it
that will "fix" the DRO and joint/axis buttons
maybe it shouldn't come from the ini file
I'm still trying to think it through
supporting angular axes for preview and backplot is much harder though
yes, I know
but I saw that axis A works for the xyza config
alex_joni: whereever the information comes from, I'll work on making axis use it
"works" is such a strong term
works == moves :D
IIRC the preview plot and backplot don't match
which is the opposite of everything AXIS stands for
but the cone rotates...:)
jepler: got any ideas about that page?
alex_joni: I haven't taken the time to read it :(
it's fairly short so far, but there's no pressure
oh. my spaceball moves my cnc machine. Thanks to jeff!
it is quite impossible to move it right now...filter is absolutely needed
I will think about it tomorrow
xemet: When I am at my laptop I can send you some files that will let you rebuild the spaceball kernel driver so that the left-handed mode works. I don't have access to those files right now, though
but now my real job is calling me
oh, but I don't need the left handed...
however it would be interesting
if you have the time, send them to me
I'll try to remember
aha -- look like this got fixed in the kernel last summer, but the ubuntu dapper kernel is much older than that
[20:35:13] <jepler> http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=2c1dd69d28cba5a51e838897e0335f82a292f366
jepler: only a question... HOW DO YOU DISCOVERED IT???
xemet: I added printk() (print to kernel log) in the spaceball driver until I discovered which line it failed on
xemet: once I knew which line it was, it was "obvious": it was comparing SPACEBALL_MAX_ID (8) to SPACEBALL_4000FLX_L (9)
xemet: after I changed SPACEBALL_MAX_ID to 9, just as Nick Martin did, it worked
dapper didn't reconize my touch pad as a synoptic either
update asked me if I wanted to overwrite the rtapi.conf. I said yes then had to re-edit it again ;) updating to 2.1.1
2.2 is down the road a bit
couldn't get to linuxcnc.org quick enough to check ;)
how is the stepper hadware generator going?
(I had some distractions this weekend)
skunkworks: next time it asks.. don't let it overwrite it
now you tell me ;) nothing in there that would change most likely?
skunkworks: I'll let you know if it ever will
(probably when rtai-4.x appears (if we decide we want that))
or when we switch to xenomai (no real reason for that)
are we well?
had a bit of a printer bargain this week
a while back I bought a Xerox Phaser, hot wax printer ... great quality
anyway .. it had a broken plastic bit ... and I saw on eBay ... 4 of them for spares, dirt cheap
bought them ... and to my suprise, they came with the additional lower paper trays
those alone worth more than I paid
even better deal
and .. they were all full with ink sticks
better still ... two of them work perfectly :)
so now I have 3 working printers and enough ink for a year and loads of spares should I need them
so ... apart from that ... I collected the conveyor system for our powder coating plat this weekend
robin_sz: care to give one away? :P
alex_joni, the shipping would cost a fortune .. they are very heavy
anyone know how linux behaves with multiple keyboards?
attached to one computer
I have a feeling that by default they all type letters
and idea of how many simultaneous keypresses the system can register?
lerneaen_hydra: I somehow suspect there is a very small probability that you'll ever get 2 keys at the same time
well, I need 8 ;)
I think that depends more on the keyboard than on Linux
four it can do without problems
if the keyboard can do it
[21:31:52] <jepler> http://www.braille-pokadot.com/keyboard.html
<-- this page indicates that it is hard to find keyboards with 6-key rollover
damn, I suppose I could make do with 4-key input
lerneaen_hydra: what'cha doing?
building a second DDR pad and trying to connect it to a slaughtered keyboard
what you want to do is build an "mjoy" and let me know how it works. http://www.mindaugas.com/projects/MJoy/
a bit overkill for my purposes ;)
still, it sounds like a neat project
to bad implementing usb is such a PITA
besides not fully obeying the USB standard, the AVR software-only USB works pretty well
for low data rates anyway
oh, I see
anyway, 'night all
see you lerneaen_hydra
if the avr can read some quadrature inputs often enough while tending to usb, a variation of mjoy with some jogwheels and LEDs added and a few buttons taken away would really be nice for an emc physical control panel
jepler: did you see I was running my 1024 line encoder at 1500rpm.. No issues there. want to spin it faster :)
my first guess is that the USB interrupts last less than 500uS (750 bit times) but probably not less than 250uS (375 bit times), which might be OK for jogwheels
skunkworks: that's about one transition per 10ns? faster than you could do with software counting, but the pluto should be able to go much faster than that.
Yes - I could test the limit ;) but at 1500rpm it is faster than we will need for the k&t
what linear speed will 1500rpm give you?
3tpi lead screw - direct drive
would allow for even higher line encoders.
you have to look at the rise and fall times specified for the encoder, not the speed of the pluto.
it looks like the theoretical max speed for a 1024-line encoder is over 500krpm (units 40MHz/4096 1/minute)
yah- that would be screaming.
I think encoders are limited to 1-2MHz for pulse generation
some of them at least
the first usdigital encoder whose datasheet I looked at ("E6") specifies 'Tracks from from 0 to 100,000 cycles/sec"
or 6000RPM for a 1000-cycle encoder
the cheap ones from automationdirect say 200khz response frequency
what ever that means ;)
is this spinning the encoder with the drill? Have you done closed-loop with the pluto yet?
ooh time to go home
just spinning it with a drill
I need to build a powersupply with enough punch to run the big servos
stupid term froze up
know anything about electro craft dc motors ?
got some dc motors on my mill
they have encoders on
and i have the dc feed wires identifyed
but there were some other wires that come out
it says on the side permenant magnet servo motor-tach
the encoder wires come out the top
and but there are 2 sets of 2 wires im not sure about
if it says tach then that's different from encoder
a tach is a thing that gives an output voltage proportional to speed of rotation
it would be useful for a spindle motor (controlling speed) but not for moving an axis (controlling position)
so i dont really need the tach?
jepler: what's better DLT or DAT tapes?
col: depends on the motor drive you have
there not likley to be for brakes or anything?
some drives work in velocity mode, and use a tach to close a speed loop
alex_joni: beats me. I have a 35/70 DLT for my backups, and it works OK -- except I haven't done a backup lately
how many wires is a tach normally?
what could the other 2 be ?
you can measure them with an ohmmeter
what would i expect to see from a brake/tach
or could i run the motors with a 12 v feed and see if i get any voltage readings from the wires?
ie for the tach
tach would be very low AC voltages
you should even see it when spinning by hand
alex_joni: blank DAT tapes were a lot less expensive than blank DLT tapes last time I looked.
if the motor turns by hand, then there is no brake
lerman: yeah, I read that DLT is a bit better
Faster is better if you have lots of data and not much time. I consulted for EMC...(the storage people)... They had a system that ran something like twenty drives at a time to get the backup transfer rates they needed.
DLT is still not fast
I have a DLT-III (I think) drive, 35/70GB capacity. One tape takes hours to read or write.
(yes, they have several newer generations of drive by now)
I think AIT is pretty fast
I have an old DEC version of DLT sitting around here someplace that I've never used. I forget where I got it from. My guess is that it's a DLT-II or III.
the motors do turn by hand
cradek pointed me to these drives because you can get them inexpensively on e-bay, and you can even get "removed from service" tapes pretty cheap
Yes. EMC was using AIT, I think.
well i can turn the worm drive with a spanner fairly easy
would the brake work on voltage to release ?
col: usually yes
this is nice: http://cgi.ebay.de/DELL-15-TFT-Rack-Console-15FP-LCD-Monitor_W0QQitemZ160089896441QQihZ006QQcategoryZ19546QQrdZ1QQcmdZViewItem#ebayphotohosting
any ideas on what i need to do to find out what the wires do ?
also is it safe to apply say 12 v to the motors to move them about ?
col: it would be best if you have a lab PSU
one that limits current too
was just gonna try using a small psu or battery
know how disgusting your keyboard is? give it a shake... eeew.
a-l-p-h-a: I'd rather not
good night all
i almost have shared home and limits working, but when i link more than one axis in hal in stops working.
[23:31:55] <twice2> http://linux.pastebin.ca/373716
when you home axis 0, the HOME_IGNORE_LIMITS means the limit switch for axis 0 will be ignored during that time
but the limit switch for axis 1 won't be
that's a bug in emc
ah, kick the can, it's all wired up ;)
but all five switches are wired in series and only go to one pin
well, we all told you it would work
too bad that wasn't true
hmm - I thought someone just set that up a few weeks ago..
iows, with just axis 2 linked in hal i can send it to search for home and make it think its home by opening axis 0 limit
yuck, I'm not sure how to solve this -- it involves making the code that checks for limit switches know about the homing process
jepler: why would it matter - if your homing axis 0 all the other ones should be off the limits.
skunkworks: twice2 is using one pin for X home, X limit, Y home and Y limit
ah when you hit the axis 0 limit - it trips axis 1 duh
wasn't thinking in reverswe
twice2: if you set accurate soft limits then you can have a fair degree of confidence that emc will stay within them, issuing an error and aborting if a movement would go outside those limits
twice2: if you're comfortable with that, you could leave the limit switches unused for now, and just hook up (in HAL) the hom-sw-in
twice2: now that I'm aware of this bug I'd like to see it fixed, but that'll need some brainstorming with my fellow developers, none of whom seem to be chiming in right now
let's see then that may work, dunno
ok, well it's a feature then :/
my best idea so far is to add a pin in HAL that is TRUE when an axis is homing. In your case, you would use some 'and's and 'not's to create a 'limit-switch' signal which is TRUE when parport.0.pin-10-in is TRUE and all of the axis.N.homing are FALSE