yay, worm gears gashed and hobbed
for future search reference
When trying to configure a linksys wireless G card - good information at http://ubuntuforums.org/showthread.php?t=919964
not actually working yet - at least I can see my card now.
grrr - gotta stop using two keyboards
yay, finished outside milling just as the temp is dropping back below freezing
jmkasunich: you've got a mill sitting outside? :)
Anyone in here familiar with the GA-370 Glentek boards?
what is it?
Board to drive my Glentek 370 servo
too hard to look it up on their site
I have an old Shadow CNC and am trying to convert the system to EMX
Nah, I have all the info on the boards and drives I just wanted to know if any of you got it working with EMC
UncleG, I've never heard of anyone using those (with EMC2 or without)
I figured as much... I guess im an entrapenuer
what does the GA-370 do exactly? (analog output to the drives, resolver feedback, ...)
I think it's 10v analog
one common servo control interface is +- 10 V analog
Im useless with this stuff. I'm a machinist, not a cnc buff :( seems I can't even set up the encoder
My first step was to connect the encoder to the PC
well,at least you have the manuals. start reading :)
ah, the GA-370 is the servo drive board
I wasn't sure if it was a controller to tell the servo drives what to do
well. you see I have a 15 pin D connector from the encoder to the old cnc box with the pinouts. Im assuming I would be able to rewire it into my lpt and I don't even know what wires go where~
I attempted reading the manual for EMC and am so completely lost.
do you have a 15-pin D connector for each motor?
I doubt a parallel port will be sufficient
certainly not a single one
I just need one drive to work
you need a 1-axis CNC?
until I learn enough to make it 2
it's an analog drive, mesa or equivalent required
I suspect that there's an interface that can run several of them
the mesa card eric_unterhausen mentions is probably a good match
At minimum all I want to know is 1. can I just run the 15pin D to the lpt and get a readout from my encoder.
not without some work
well, I don't mind work.
I just need to know where to start.
do you have a pinout of the encoder?
it is possible to make it work, but of course you'll need to rewire things, since you have a 15-pin connector and a parallel port is 25 pins
I have all the wires in front of me ready to wire up
are there A B and Z wires?
do your motors have tachs?
what about ~A /A or some other A (and corresponding B and Z)?
I can give you the existing pinout for the 15-pin D
I think the glentek is going to want a tach
emc doesn't need it
can I scrap the tach then?
EMC isn't a servo drive, so the 70 still needs the tach
sorry I was confusing
you have 3 components: motor, drive, computer
the connection between the motor and drive needs to carry some information (and power), the connection between the PC and the drive carries other info, and the motor also needs to feed position information back to the PC
from motor to drive should be motor power wires and tach feedback (probably)
from motor to PC is encoder feedback
which is where I want to start
from PC to drive is probably analog, which may work somewhat with no hardware other than the parallel port, but it won't be pretty
I should be able to turn the servo by hand and get an encoder feedback correct?
turning by hand is slow though, and there's a speed limit as to how fast the motor shaft can turn before the PC can't keep up
there is no stock "sample" configuration that uses the parallel port for encoder feedback
unless you set it up as spindle feedback
so i could potentially use spindle feedback to get a read
there's a program included with EMC2 that lets you set up some types of machines which use the parallel port. it's called stepconf
using that program, you could select some parallel port pins to be spindle phase A, spindle phase B and spindle index inputs
would you mind looking at my D-pin layout? and getting the titles up to date because I think the old names for the hardware differ from the new
if you connect A B and Z (Z is index) to those parallel port pins, then the spindle feedback should show you something when you spin the motor
that would very good.
I can take a quick look
it's past my bedtime :)
ty, I will get you the link
[06:20:07] <UncleG> http://www.shadowcnc.com/shadowmanual.htm
then find amplifier and the layout will be on the left
ctrl + f amplifier
glentek 40 series motors?
well, it's not simple. you need a 5V supply for the encoder
you can get that from a USB connector, but the parallel port has no good power output
I can handle that
the GA-370 does need a tach on the motor
index is called "mark not" ?
or at least it has a tach input
I think so anyway
I was curous about mark not
it confused me
the "not" means that it's an active low signal, so check the "invert" checkbox for the pin you select as spindle index input
most older brushed motor controls require a tach
ok, im in stepconf right now setting things up as we speak.
and yes, the drive takes +- 10V control signals
(9 pin shell, pin 1 "SERVO COMMAND -10V to +10V)
I told you that after dloading the manual
had to register and everything
I thought you were complaining that you didn't want to wade through the site :)
I didn't want to do that
invert "spindle index" and Spindle Phase A/B and I have 2 more open pins for input will that be enough?
I registered as Joe Mader
you don't need to invert A and B
I didnt. I wrote that poorly
next time I'm going to register as terry mader, which means your mother in Hindi
it doesn't hurt if you do, generally each inversion will reverse the effective direction of the encoder
which may need to be done after that motor runs away
so if you see the numbers going up where you think they should go down, you can either swap the pins in the connector, or invert one (but not both) inputs
I'm thinking there won't be any power to anything except the encoder at first (hint hint)
where's the fun in that?
so, not all hope is lost with this servo setup :)
in the immortal words of Monty Python
I did manage to connect the servo drive board to minicom and make it run with that :p
UncleG, you won't be able to make a useful system by going down this path, but it should be a learning experience for you
in one direction at full blast :)
20% over full blast even
there are several boards that can drive your system
the Mesa boards start at $200 fora PCI card + $70 or so for a 4-servo interface (analog out and encoder in for each of the 4 channels)
there's a parallel port connected card also, which is $90
almost doubles the cost
more, for the 7i43
since you need I/O and analog/encoder cards
man, with all these companies going out of business in the US I should be able to pick up a full blown cnc system for about the same price at auction
if you can pick it up and bring it home, that may be true
but not with controls
Yeah, I have the means but I really would like to learn this
it is possible to use parallel ports to do this, but at some point you'll have to buy more ports, build or buy interface cards and whatnot, and you'll still have a system that's limited by the PC
on that page I sent you
and you'll be on your own, since nobody really does that :)
cw limit and ccw limit home and comm
I have no idea what those are for, could you explain?
limit switches and ???
home position and
I wont need the limit switches right?
if the motor isn't moving machine parts around, probably not
Im running this in a cam system anyhow
ok, bedtime. good luck and enjoy
thank you very much.
UncleG: Which CAM system? (goodnight SWP)
the servos drive cams
like a mechanical cam
old index machine
what did it do?
like a turret lathe with barloader
I just wired up the encoder.
I used a/b/index and home and added the usb 5vdc for the power to the encoder. now where will it give me a readout of the encoder position?
the hal scope didnt give me any readout of encoder.0.position or phase A
it's possible that it isn't electrically compatible with your parport
Would you happen to know how one might check for such a scenario?
are you sure you are powering your encoder? multimeter looking at voltages?
I would like to be able to see changes on the a/b/i lines
well, I did wire into it with a 5vt line from the usb
before I paniced
and it checked out as 5vt
with the encoder attached?
but you should be able to see changes on the a/b/i lines
where on the line should I check it, right at the encoder?
by moving very, very slowly
don't bother, check the outputs of the encoder
what pins did you hook up on the parport?
10 11 um....lemme see what else
10 11 12 13
spindle phase A B Index and home-x
wiki is being served by hamsters typing
maybe I am using scope wrong
did you check your a/b lines for voltage changes?
you probably are using scope wrong, it isn't trivial stuff
I see a flat line.
What am I looking for?
I get a voltage output on phase A with the voltmeter nothing for phase b
does phase a change when you turn the encoder?
on the scope I get a flat line and nothing changed when the encoder moves
I'm talking about voltmeter changes
it just stays less than 5 v
if the voltages going into the computer aren't changing, the scope inside the computer isn't going to see any changes?
some encoders take a reasonable amount of power to actually get working
for example, gecko drives can't power up all known encoders
reasonable amount of power?
I just do not understand I put 5vdc to the encoder and spin it , phase A should cause fluxuations on voltmeter correct?
you may have to spin slowly to see it on a voltmeter
very, very slowly
ch a/b are 50% on
and there is the possibility that they are open collector, which means they need pullup resistors to work at all
yeah, I read about that.
Wish I knew wth I was doing.
Seeing anything more than a flat line would be big improvement on this scope.
Looks like I have a lot of learning to do.
so, this page can be changed; http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Supported_Hardware
I have the linistepper working with EMC
You should look into making a GA370 servo work with it :)
I'll send you kudos.
JustinXJS2_ is now known as JustinXJS2
Lerman_______ is now known as Lerman
anyone know if BigjohnT is out of electricity again ?
he is on dial up
last week he was out because of ice
piasdom: he's probably asleep now
I have a question..
in : addf funct_name servo-thread -1
what is the number -1 standing for?
Vq^_ is now known as Vq^
alex_joni, hi hi
robin_z: what's new?
we got snow!
and a fantastic boomin economy
not enough snow
not here in eastmids
im in kidderminster
oh ... hmmm
dry ski slope IIRC
with added real snow today
yeah, totally awesome
150mm of booming economy?
sounds like .. a lot? .. not really
we haz no ekonomee
sure we have ..
its just that this months orders are probably stuck in the snow ...
could be worse
archivist we have had about 40-50mm down in poole
still burger all
wimps on London moaning about 8 inches
they have closed the skools here!!! unbelieveable
I dont remember schools being closed as a kid
I used to have to take a dog sled to school
BigJohnT: So, how does whale/seal blubber taste like??
salted salmon is much better
pjm__, part the kit (but a little over priced these days) fleabay 110345883898
archivist did u see that cnc spindle + VFD 180324271889
hmm looks like he is making them or something
on friday, the forecast was for "crushing" amounts of snow today
now they say 1"
neutron enriched snow
that's not snow, it's titanium dioxide!
hello everybody, i need help with the user defined M codes, if someone could help me i would really appreciate it, thanks
what are you trying to do?
i need to read a parport input
ok, but what are you trying to do? :)
the parallel port input will already be read by the HAL parallel port driver, and its status will be available
i need the g code program wait for the M code and when de M code reads that the parport pin is TRUE then the program continues with the next line
ok, you don't need custom M codes to do that
sorry for my english i'm from argentina and i'm not used to talk in english
if you have a recent version of
no problem, you're doing well
so what can i do?
there are M codes to wait for an input to be true (or false)
the m66 ?
I think they are in the M66 range, but I'm not sure
yes, that would be it :)
yes i tried that but nothing happens, i mean when the program reaches that line (the m66 line) the program stops until the Q time reaches 0 and then continues the program
and when the m66 is waiting i changed the port from false to true state but nothing happened
did you connect the port pin to one if the motion controller digital inputs?
and i checked that in the hal show panel
it seems to be ok but when i change the port to true nothig happens.. the Q time reaches 0 and then the program continues
ok. well, if I knew more about the motion controller I/O M codes, I might be able to help
but I don't
i need to do this because i'm using an external measurement system with and encoder.. so when the encoder reaches the value that i need sends a signal to the port
ok, got it
well, thanks for your help and i'll keep trying to make it work
what version of emc are you using?
i downloaded the linux hardy heron from the linux cnc web
it's the emc 2.X i don't know exactly what version is
emc 2.2.3 includes the following change: * fix #1853953 M66 doesn't work with parport inputs
determine your version of emc and upgrade if it's older than 2.2.3
If you are using the AXIS user interface, the version should appear in the title bar and in the window shown by Help > About
otherwise, you can look in the packagemanager
ok, thank you very much im gonna look that and if it's older i'll upgrade the emc
[15:23:58] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?InstallingUpdates
thank you very much you are very helpfull
heh. I figured someone would be :)
i'm gonna take a look and then i tell you if i solved the problem
btw - I think I all but liquified one of the mosfets.. (I kept turning the current up while testing) needless to say - it is the first 2 mosfets I have destroyed so far.
come back when you've gassified them :)
oh - and then I took the bridge out. (it was way too small)
heh - It was cool - the metal on the back of the mosfet looked 'rough' and it had 'leaked' around the edges
skunkworks: this thread is useless without pics
with smello vision
the bridge really smoked.. The mosfets really didn't smell all that much.
I will see if I still have the mosfet.
which component(s) do you mean when you say "the bridge"?
rectifier I was using in the power supply
yah - sorry - that was confusing
[15:55:51] <alex_joni> http://business.timesonline.co.uk/tol/business/industry_sectors/technology/article5639463.ece
"Boffins at Stanford University say they have managed to write "SU" in letters smaller than atoms."
I'll believe it when I see it
"Manoharan and his nanohologram collaborator Chris Moon say it's difficult to properly express the size of their stacked S and U, but the equivalent would be 0.3 nanometres."
sounds about right for a new "big" DRO :)
SWPadnos: the 10$ laptop you mean?
no, the subatomic letters :)
I find the 10$ laptop harder to believe ;)
RS500 could be a lot in India
IBM claimed something a few years back as well regarding moving atoms
that's just a rendering. I want to *see* it :)
SWPadnos: guess your sight is not as good
The researchers encoded the letters "S" and "U" (as in Stanford University) within the interference patterns formed by quantum electron waves on the surface of a sliver of copper. The wave patterns even project a tiny hologram of the data, which can be viewed with a powerful microscope
didn't the ibm one actually show a picture. I vaugly remember that.
[16:30:43] <skunkworks> http://news-service.stanford.edu/news/2009/january28/videos/427_flash.html
skunkworks, I saw a picture but cannot remember where it was (could have been in a magazine)
$10 laptop is a piece of cake, with proper fudging of the definition of "laptop"
think monochrome e-ink segment display, arm processor, solar powered
what amazes me is that TI graphing calculators are still $100
they should probably be a lot better than they are for that much
they have to suck in order to prevent students from shortcutting around the problem with the advanced computing device in their hands
* fenn wonders is that sentence made any sense
I used to write programs during tests
"everything has changed but our thinking"
I think calculators should be banned from tests
and then they should ban tests after that
so much cheating going on
cheating is the least of the problem
my advisor gave one problem that was a total non sequitur -- if you got it right you were cheating, I'm almost positive about that
ugh, I'm so glad school is behind me
what did you go to school for?
for about 7 years
I studied comp sci and french
but quit without a degree
first they should ban students
can't do that, who would do my work for me?
yep somebody needs to automate those students out of a job
or maybe we could just outsource it
outsourcing part is done, try finding an American grad student
does anyone still have xemet/manfredi's nurbs patch?
i thought i saved it but i guess not
trying to talk myself into riding two bikes into town today, racing bike needs work
remember to put some padding on your shoulder
I just push the second one
guess i dont see how that works
one time I was taking my tandem to the shop and the cable to the only brake I was using broke
that was fun
[Global Notice] Hi all, It would appear one of our client servers just dropped off the face of the planet. We're looking into the issue and should hopefully have it back soon. Affected users just over 2,000. Apologies for the inconvenience and have a good day.
good evening.some people can help me for problem in rigid tapping with 5i20?thanks
they can surely try if you state what the problem is..
alex i have configure my spindle whith analog */_10volt and when i command in mdi m3s1000 and m4 s2000 the velocity is ok
but i create one program : g0 z0;m3 s200;g33.1 z-5 k1;g0 x10 y10;m2. the z axix down in z-15.00 and stop not move at z=0 z position .
this is my hal file for spindle:
#Load a ddt and scale to get spindle-speed from spindle-position#
loadrt ddt count=1
loadrt scale count=1
net Spos hm2_5i20.0.encoder.03.position => ddt.0.in
net Svel ddt.0.out => scale.0.in
setp scale.0.gain 0.00445
net Srpm scale.0.out
#link together the spindle encoder, motion module and PWM generator#
net Spos => motion.spindle-revs
net Srpm => motion.spindle-speed-in
net Scmd motion.spindle-speed-out => hm2_5i20.0.pwmgen.03.value
setp hm2_5i20.0.pwmgen.03.scale 3000.0
setp hm2_5i20.0.pwmgen.03.output-type 1
net Sidx hm2_5i20.0.encoder.03.index-enable => motion.spindle-index-enable
#enable the PWM gnerator too#
net Enable => hm2_5i20.0.pwmgen.03.enable
#calculate spindle speed
addf ddt.0 servo-thread
addf scale.0 servo-thread
does your spindle encoder have an index, and if so is it connected to the 5i20?
yes my spindle have index and is connect at m5i20 the encoder is 2048 cout/rev(quadrature encoder)
what version of emc2?
please read my hal file and if you can control is it ok
where is the hm2_5i20.0.encoder.03.* connected to motion?
the index-enable or whatever it's called
net Sidx hm2_5i20.0.encoder.03.index-enable => motion.spindle-index-enable
Spos and Sidx both seem right
cradek, 2.2.8 doesn't have the at-speed thing, does it?
motioncontrol: do m3/m4 cause spindle forward and reverse?
yes the spindle move ok in m3 and m4
did you put halmeter on motion.spindle-revs and see that it increases 1.0 for each turn in the M3 direction and decreases 1.0 for each turn in the M4 direction?
SWPadnos: right, no at-speed
if i command m3 s2000 the motion.spindle-speed-in is 2000 value
motioncontrol: cradek asked specifically about motion.spindle-revs
ok i control in my office after this variable.this variable motion.spindle-rev whath is?
motion.spindle-revs should increase 1.0 for each turn of the spindle in the positive (M3) direction
it's called motion.spindle-revs (spindle revolutions)
ok i control after 10 minut thanks for now.
archivist are u here?
ok i return i'm motioncontrol, but now stay in office
i ceck the motion.spindle-rev and this variable is = couter encoder not revolution increment
fasting: then you need to set the encoder scaling
i don't have scaled encoder but pwm.03.scale at 3000
that is not relevant. you need to set the encoder scale.
please control my hal file i thing is no correct
i don't have scale encode .03 signal
yes i add scale encoder.03 and now the splindle-rev increment at 1.0 at one turn the spindle.i prove wait please 5 mints
ok the rigid tapping is ok, only one problem when revert the motor the z axis move is fast and for one istand vibration , after finish to z=0 pos and now z axis fast revert for one istand.The axis position is ok.i ceck my max velocity is 300mm/min.because?
fasting: that didn't make sense, try expressing it otherwise
excuse for my english i reppet
fasting: no need to excuse, just try to say it differently
it is normal for Z to move up past 0 at the end of the tap cycle
it then moves back down to the original height (0)
here is the explanation of the rigid tap cycle: http://www.linuxcnc.org/docs/html/gcode_main.html#sec:G33,-G33.1:-Spindle-Synchronized
my command is :g0 z0;m3s200;g33.1 z-5 k1;m3 s1000;m2.the move when the z axis revet at -5 to 0 is very immediately(very fast) because.the axis z receive the meccanical shock when z revert the movement
fasting: it starts moving from Z0 down to -5 .. right?
then just before -5, it starts stopping the spindle, and changing direction
at Z-5, spindle speed is 0 and direction changes
yes , the direction change
is that when Z makes a jump?
the problem is when the direction change
yes is jump
I think it depends how fast your spindle stops/changes direction
ok i decreese the accelereation spindle
is Z/K in mm?
it seems like a very short tap
yes i use mm
ok thanks for all.very important for me your help.at after.
is it better with slower spindle accel?
I am not sure I understand the problem
but if the axis accel cannot keep up with the spindle accel, it will break the tap
I think he's air-cutting for now .. maybe
[20:28:45] <toastatwork> http://www.onesentence.org/stories/2283/
[Global Notice] Hi folks! Earlier one of our leaf (client) servers had a few connection issues. This was due to latency issues on the sponsor's network. Fortunately, the sponsor now reports that the problems have been resolved and so everything is back to normal. We're very sorry for the inconvenience. Thanks for using freenode and have a nice day!
I don't think an apple would help him at that point
How do I set up my udev devices on Ubuntu 8.04?
I found some help here - http://www.britishideas.com/2007/11/07/how-to-install-emc2-on-unbuntu-710-gutsy-from-scratch/
Would they be the same for 8.04?
Read this too but don't know what to do - http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Debian_Lenny_Compile_RTAI
did you install from the EMC2 liveCD, or stock Ubuntu then install EMC2/RTAI?
8.04 is the currently supported version so you could just install our packages.
otherwise, please describe what exactly your goal is
I have a smp box
Had to compile from scratch
do you already have a rtai patched smp kernel that works?
The kernel is running now, and I have compiled the rtai modules
I have not installed them yet
about the only special thing emc2 packages do to is disable the loading of the parport_pc kernel module: http://cvs.linuxcnc.org/cvs/emc2/debian/extras-Ubuntu-8.04/etc/modprobe.d/emc2?rev=1.1
I disabled parport in the kernel when I compiled it.
Is that enough?
I think that would work, but I have not done it that way
I will find out I guess. :)
Just need to get past the udev thing.
what is your specific question about udev? maybe I don't understand what you're asking.
I have to make a file called /etc/init.d/create-rtai-devices.sh
Don't know if that is right for 8.04
I don't know about any such file
recent rtai versions have udev support. I suggest you just try it without any trickery first.
install rtai and run their test suite. if they work, it's probably fine
I guess it is for the rtai_shm, and rtf[0-9] devices?
UbunUbun: yeah, but those should get created by default by rtai
if you use a fairly recent rtai package
(newer then 3.6. or so)
I used rtai 3.6 cv
I'll give it a whirl
one more question. do I have to sudo to install the rtai modules, or just my user account?
in general, you need to use sudo to install things. I'm sure this will be the case for kernel modules
3.6-cv would have been my choice too
Ok, they are installed. going to reboot and run the test suite.
well - good thing you guys piped up - I was going to point him to the wiki on setting up a joystick
cradek: on that board I recently milled, I realized after almost all the trace isolation milling was done that I'd done it all at 200% FO. It didn't seem to hurt the finish, and I figure with my tendency to break tools due to dumb mistakes it doesn't matter so much if I'm shortening my tool life
I think that means I was going at F40..
zip zip zip
have you never yet replaced one because it was dull?
well, sort of -- I I junked all the tools I had tried in the dremel, their tips were not very sharp anymore
which was 2 of the nice ones, plus several those cheaper ebay ones that I never got satisfactory results on
there must be a use for dremels, but I don't know what it is
the ones with one flat face don't work as well?
I am trying to remember if I ever tried one in the wolfgang spindle..
now that I'm using all tools with depth rings, I should test how consistent my tool lengths are and find out whether I can skip using the tool length probe
the ring setting is said by the manufacturer to be +-.002" which is fine if I touch off and start with the fine trace isolation tool
they give a recommended Z-Axis-Offset for each tool: http://www.precisebits.com/Reference/drillfeedspeed.asp http://www.precisebits.com/Reference/diamondcutfeedspeed.asp
I assume this is the amount you should change the depth by when cutting with this tool
instead of cutting to Z-.063, cut to Z[-.063-.020]
or whatever it happens to be
bizarre - I wonder why
because of the vee at the tip of the drill? I dunno.
do you remember the link fenn posted a couple days ago?
regarding double belt drive?
there was a movie and all
I don't remember the link, but I remember the subject matter
ah, too bad
Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-02-02.txt
was it devel maybe?
Alex - I have a new bare install of ubuntu 8.04, no packages other than the default packages are loaded at this point. Which specific source tar balls should I download? linux_2.6.24-16.30.linuxcnc.4.tar.gz and rtai_3.6.1-linuxcnc.4.tar.gz?
[22:49:27] <alex_joni> http://www.designnews.com/article/160365-A_Better_Belt_Drive.php?nid=2337&rid=1696394
ehj: lets move to devel
but yes, those seem right
oops, I thought I was in devel.