col, I don't have experience with the motenc cards, but ifyou only need 3 outputs plus spindle (maybe), then 4 outputs is fine
unless you also need the analog inputs for something, I don't see that they would be an issue (I don't know if the card does integration of tach feedback, but I'm sure EMC2 doesn't)
well teh motenc has encoder inputs
so im assuming that takes care of it all
so the 3 axis and spindle is covered by the 4 outs
and the other stuff like tool change ect would just be digital I/O ?
i have an EMC2.1/motenc index pulse problem
col, I think so. I don't know what the other differences are
HAL: ERROR: signal 'Xindex' slready has output pin
mdynac: did you read the info about changed configs?
[00:10:01] <alex_joni> http://www.ipodlinux.org/Toolchain
[00:10:13] <alex_joni> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?UPDATING
i changed the signal names a while back....
can you put your .hal file(s) on http://pastebin.ca
SWPadnos: thanks iv ordered the lite version :)
hopefully you did more research than talking to me? ;)
yeah a lil
just wondered if there was anything obvious id need the adc inputs for
I don't think so, but they could be useful for stuff like spindle load monitoring
(assuming you have the external hardware for that)
i can always upgrade if needs be
get the other one and keep that for extra axis if i convert to 5 axis
or just get a generic A/D card and write a driver
(or get two, send me one, and have me write the driver :) )
[00:13:55] <mdynac> http://pastebin.ca/333596
my new machines arrive tommorow
[00:14:38] <col> http://img.photobucket.com/albums/v280/rallyslag/02012007171.jpg
[00:14:43] <col> http://img.photobucket.com/albums/v280/rallyslag/28122006158.jpg
mdynac, index isn't the same thing as index-enable
I think the fix is to change the first line for Xindex and Yindex to linksp Xindex <= motenc.0.enc-00-index-enable
the driver has two pins. one tells the state of the index pin, the other is a bidirectional pin that lets the motion controller ask the driver to reset the position when an index is detected
motion does this by setting the bit to 1. The driver then tells the motion controller that an index was found by setting the bit to 0
i changed the names and emc booted
ok. hppefully it still works as you expect :)
i'll check it on thursday when i get back...
gotta be up early for delivery truck :)
heh - see you
Just bought the motor controller as it should work just fine :-) $69 inc shipping not bad value!
col, nice machine ... was that from the GoIndustry auction the other day?
newly added directions and package list for building on Ubuntu 6.10: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Simulator_on_Ubuntu_6_10_Edgy_Eft
did I miss anything good?
but it left
darn, it would have been nice to see it for a little while
cradek: I just added this to the wiki: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Simulator_on_Ubuntu_6_10_Edgy_Eft
after I had declared loudly to someone -- lernean hydra? -- that I'd never have an edgy system :-P
I was installing vmware on my new system
thought "what the heck"
since it's taking me ages to get everything moved, I'm considering just sticking the old drive in the new machine and running the old OS in a virtual machine
that's an approach I hadn't considered
Twingy: thanks for working on that bug I ran into when trying out gcam
I've put my old home directory in a subdir of the new one though
Twingy: I didn't test out the new version yet
I did - and it is fixed
yep, wait a day or two until I release one with additional bug fixes
on tuesdays I mentor at the high school for first robotics league and saturdays all day
so not going to get much done tonight
I can imagine that would be in the "fun but tiring" category
league? It is competitive somehow?
I am also going to start 4th axis milling in Feb
yes, Dean Kamen (segway guy) started it 15 years or so ago
(Dean Kamen went to my school ;-)
aww gee I don't have any names to drop
Twingy: oh, is it a "battle bots" type of thing?
its more goal oriented
the bots perform a task
find objects, bring them back, etc
that sounds better than "battle bots" frankly
I wrote the software for Dean Kamen's Dialysis machine test system
this year it's placing torus's on a 3 tier cage with 8 arms per tier
(how's that for name dropping? :) )
k, I need to relax for a bit
Twingy: see you
so... i finally finished my mill today
works great, 'cept that the x-axis is backwards
crepincdotcom: that should be easy to fix
is was going to just put a hardware NOT on that pin, but then i realized the HAL might be able to do that
just negate the scale
crepincdotcom: just negate INPUT_SCALE for that axis
ok, i was looking at that
thanks a lot, just checking
you could also NOT the pin, but changing the scale is simpler
or if it's a stepper on the parallel port, you can 'setp pin-NN-invert TRUE' in your HAL file
where in the HAL file would that go?
for digital output pins, there is almost always a HAL "invert" parameter
its written wicked well ;-)
crepincdotcom: anywhere after you 'loadrt hal_parport' -- putting it near the line that connects the direction signal to the parport pin is what I would do
thanks a lot guys
have you made chips yet?
no, i figured it'd get it all working properly first
then im going to put a pencil in it
no sense putting a bit through the cmaps
i'm trying to compile emc2 but i'm getting this error when running ./configure --enable-simulator:checking for BWidget using /usr/bin/tclsh8.4... configure: error: not found
i may have the smallest EMC implementation ever.... my whole mill plus CNC setup is 20 lbs
josesito: you need to install a bwidget package
josesito: the bwidget home page is http://tcllib.sourceforge.net/
on debian/ubuntu it's just called "bwidget"
crepincdotcom: I'm having a look at your blog now
no pics yet :-(
haha "Reality check: yes, this is what I do when I get home from work and before bed..."
drat, i need to fix that post about the cnc being "almost" done i guess
crepincdotcom: your stepper driver is unipolar, then?
why did you decide not to go with L298?
i got the motors from canon scanners
the L298 is for bipolar
guys...some language questions: what's the name of a machine that makes only cylindrical pieces??
i mean it does unipolar as well
but the GSS is just as easy with separate transistors
that was easy, ask me a harder one :-)
a machine which can do 3d pieces
ooh ooh i know
actually i guess a lathe can as well
ok, passed .configure script...
I really like the L298. it's fine for driving little servos too
[02:33:19] <jmkasunich> http://en.wikipedia.org/wiki/Milling_machine, http://en.wikipedia.org/wiki/Lathe
handy - the pics clear up any mis-understanding
cradek: i think my servos would certainly be "little" compared to what you likely have
i dont know, i can more easily figure out whats wrong with separate transistors
maybe not, my servo lathe is sherline based
let me link the manual mill i built on top of:
but it takes you 12 sniffs to find the bad part when it only takes me one
my mill: http://www.thisnext.com/media/230x230/9263BDE7.jpg
cradek: point taken, but i can replace one transistor and not the whole array
we definitely need pictures of your conversion
do you have pics of yours?
i should get them up in the next day or so
how long is the table?
I think I saw some of those mills on e-bay while I should have been working
the only photo I have is my X axis motor mount: http://timeguy.com/cradek-files/emc/DSCN6290.JPG
the keyboard is there for, uh, scale
oh wow, that looks great
the Z is not nearly as nice, but it works
my manual machining was not nearly as good when i started this, so a couple of my mounts are astetically off
I know just how that goes
crepincdotcom: how is the backlash on your new mill, or do you know yet?
but it doesn't matter, you can remake a few parts later
heh, i should have used the CNCs here... im at school, we have 11 Hass cnc's, most bigger than my car
jepler: tell you the truth, i dont know what backlash is
<-- machining noob
crepincdotcom: when you reverse direction, it's how far the handwheel turns before the table starts moving again
when you turn the knob one way and then back, there's some part of the turn where the axis doesn't move
and that in turn creates inaccuracies i imagine?
that determines (sort of) the tolerance of the part you can make
cant you just tell EMC to decelerate more slowly as it enters such an area?
cnc is actually worse than manual machining - you can compensate for backlash pretty well when doing manual work
well imagine cutting a circle (something you could not do manually) - at the quadrants, you'll get flat spots
oh i see
so i guess i should cut a circle and see how well it comes out
then i'll let you know
so it's not really true that cnc makes it worse - it's just that you expect to be able to do more!
the etch-a-sketch has a lot of backlash, and the sides of the spiral where an axis changes direction are visibly flattened: http://axis.unpy.net/files/etchcnc/etch-spiral.jpg
that's a great example
it says the spindle goes to 20krpm .. http://www.emachinetool.com/new/catalog/vertical.cfm?ProductID=719
ahhh i get it
my spindle speed is 20k, but my max bit size is about 4mm
wow 20k is a lot
nice for engraving small things
ok it's installed
so you can use 1/8" printed circuit board mills and drills?
now what?? xD
josesito: run the sim/axis sample configuration
cradek: yes, my goal is to be able to make boards
crepincdotcom: that's what I use my small mill for, mostly
so what do you use.... i mean you design your board in some program
crepincdotcom: eliminating backlash is VERY important for that, since the feature size you want is often 6-8 mils
crepincdotcom: I use eagle which has a no-cost version that runs on linux
ahh, i must work on that then
but eagle doesnt go straight to gcode does it?
yes with its scripting language
is that something you wrote?
or avaiable on their site
jepler and I have written some nice exports for EMC
* crepincdotcom wonders where to find such tools
cradek: what "lines and spaces" do you usually use (for example, 12 mil lines and 13 mill spaces means 0.012 wide traces on 0.025 centers)
cradek: is this your site
crepincdotcom: here's a file generated by an older version of the eagle->gcode converter: http://emergent.unpy.net/files/sandbox/elm320.ngc
you can preview it in axis
i'm getting an error!!!
the code looks good though
_tkinter.TclError: couldn't read file "/usr/share/axis/tcl/accel.tcl": no such file or directory
crepincdotcom: maybe this already crossed your mind, but you can easily use your "logic probe" with emc's halscope, for sample rates up to about 50kHz (BASE_PERIOD=20000)
it can read in?
i thought that was just for viewing HAL's output
thanks for letting me know
crepincdotcom: you can configure the parallel port in 'input' mode so that pins 2-9 are read
cradek: sorry if im being thick but i dont seen anything on that site referencing eagle scripts...
crepincdotcom: thinktink.com was for PCB milling bits
ahhh ok sorry
the g-code milling scripts are here (no docs): http://unpy.net/cgi-bin/viewcvs.cgi/eagle/ulp/
josesito: try one of the other front-ends -- select sim/temc instead of sim/axis
crepincdotcom: cradek's site is http://timeguy.com
-- here's an example of a board he milled: http://timeguy.com/cradek/01144545838
oh wow, thats a nice one
well, what if i want to work with a lathe, where can i write the code and run a simulation??
josesito: sim/axis will probably work if you (A) ./configure --prefix=/usr OR (B) ./configure --enable-run-in-place
crepincdotcom: fun with capacitors ;-)
jmkasunich: ha that has the old ones
i got some new ones.... the whole unit weights about 25 pounds
[02:54:19] <jmkasunich> http://jmkasunich.dyndns.org/cgi-bin/blosxom/capbang.html
soounds like a gunshot
jepler, i did ./configure --enable-simulator
josesito: there is an unresolved problem with AXIS when the PREFIX is not /usr: http://sourceforge.net/mailarchive/message.php?msg_id=37495337
crepincdotcom: btdt - except not on purpose, and with much more expensive results
haha jmkasunich nice one. tonight we're going to hit a can of dw40 with mine (from far away)
josesito: then try: ./configure --enable-run-in-place --eanble-simulator
josesito: instead of using "make install", just run scripts/emc from the top emc2 source directory
crepincdotcom: the "plasma jump pattern" and bouncing balls of light were probably molten copper and/or burning aluminum from the can
yeah, we figred that out in the end :-(
i gotta get some pics of the new caps
I work with large cap banks for a living - except we try very very hard to prevent sudden discharges
yeah i bet :-p what do you do?
large motor drives
jepler, same error
the most recent product I worked on: http://literature.rockwellautomation.com/idc/groups/literature/documents/pp/20l-pp001_-en-p.pdf
you're not messing around
thats a serious device there
josesito: find the files in /usr/lib/python2.4/site-packages that were put there by your earlier "make install", and remove them
josesito: then make sure you're running scripts/emc instead of the /usr/local/bin/emc that the earlier "make install" created
but does that mean i'll have to uninstall the one i just installed?? :S
josesito: choose the configuration sim/lathe if you want to work with a simulation of that type of machine
josesito: yes, the files you installed with "make install" are not being used now
cradek: your nixie clocks are awsome
how can i uninsnstall ?? make uninstall?
there is no "uninstall" target
what should i remove then?
alright guys work is over, im heading home to finish those configs on the mill
then fun time :-)
talk to you all later
and thanks again, really
crepincdotcom: see you
crepincdotcom: nice meeting you
you as well
josesito: since almost all users either use "run-in-place" or the debian package, we've never addressed "uninstall"
josesito: if it helps, this is the list of files installed by the package on ubuntu: http://emergent.unpy.net/index.cgi-files/sandbox/files
josesito: I think you will get most of the files if you get rid of /usr/local/lib/emc2 and /usr/local/share/emc and
will there be any solution for my installed emc version?
josesito: it may be fixed in a future release, such as emc 2.1.1, but I don't make any promises.
josesito: there is one other thing you can try, if you want. In the file lib/python/nf.py, change the line that starts 'PREFIX =' to read: PREFIX = "/usr/local"
josesito: then reconfigure without --enable-run-in-place, reinstall, and see if that allows it to work
that file is in emc's source directory??
not under src/
once you've changed that line, the "run-in-place" version will stop working -- so it's not a full solution for everyone
is it safe to install over the existing installation?
PREFIX = find_prefix(__file__) ---> PREFIX = "/usr/local"
yes, that is what I'd like you to try
(already ran configure)
jmk: nice caps
josesito: is the error exactly the same as the first time? I would expect it to look for /usr/local/share/axis/tcl/accel.tcl which should be created by "make install"
let me see
it doesn't look like the same error
can you put the full error at http://pastebin.ca
[03:20:55] <josesito> http://pastebin.ca/333797
uh-oh, python 2.5
josesito: but sim/axis worked at least once while you were using --enable-run-in-place?
i think so. a small window appeared for some seconds but then dissapeared
josesito: I don't have any guesses about this new error :(
jepler, i ran again configure --enable-run-in-place with the original nf.py and tried sim axis
josesito: what happens now?
the small window appeared (it was just part of a manual or something like that) after a few seconds it dissapeared and now i don't see any window
MOTION: init_module() complete
task 0x8051e60 period = 1000000 ratio=1
*** glibc detected *** /usr/bin/python: free(): invalid pointer: 0x0806c0e0 ***
those are the last 3 output lines of the program
that probably indicates a bug in emc2, but it is not one that I encounter here
it looks like it's still running
if you know how to get a backtrace with gdb, that will help me. otherwise I don't have any suggestions.
i don't know :(
I saw another guy failed to run AXIS with python2.5 but python2.4 worked
but I don't think the error looked anything like that
let me make sure i have this right: +x is to the right, +y is toward the front of the machine, and +z is up?
+y is away from you
right handed coordinate system
but note that means the table moves toward you
right that makes sense
ok thanks, let me make sure i have this all right now
jepler, about the software itself: does it show the shape of the piece i want to build in simulation??
josesito: The AXIS preview shows the path of the tool, not the shape of the remaining material
here's an example, showing a program to mill a printed circuit board: http://axis.unpy.net/files/about/axis11-running.png
for a lathe the display is harder to interpret. here's a lathe program that creates a chess piece: http://emergent.unpy.net/index.cgi-files/sandbox/axis-lathepawn.png
you have to imagine the remaining material "above" the tool path, which is shown in white
the top of the pawn is on the right
jepler, is there any other software like emc2 for linux?
oh btw...i dled coolcnc
i'll try it later
coolcnc is based on emc2 but the version is very old and buggy
some people on this channel are interested in programs that show the remaining material, but they are all at the experimental stage. you can read some about my effort here: http://axis.unpy.net/01169521961 http://axis.unpy.net/01169689661
although my information may be outdated
when you try it, let us know what version of emc2 is on it
last I knew it was an emc2 release from Feb-March '06 and a lot has happened since then :-)
*** glibc detected *** /usr/bin/python: free(): invalid pointer: 0x0806c0e0 ***
jepler, is there any other software like emc2 for linux???
josesito: to control cnc mills and lathes? No, not really.
to preview g-code files for mills or lathes? none that I know of.
ok my scaling is off
1.8 deg/step and 1 rotation per mm
im looking for the config on that now.....
crepincdotcom: are you using full or half stepping? (you should use half-stepping if you can)
3.6 deg/step motors, at half
then I would expect SCALE to be 200 if you are using mm, or 5080 if you're using inches
can i see the math you did there?
360 / 1.8 = 200 steps per revolution
oh, its just asking for steps/rev
steps per "unit"
and unit is a revolution, or mm
unit is mm
yeah - if your screw is one rev per mm, then its steps per rev
that's why it's 5080 if the rest of the ini file is in inches
good night jeff
[22:29] Emily Ross: you asshole!
[22:30] Emily Ross: thanks a lot, seriously
[22:30] Emily Ross: i need tea
sorry sorry sorry, damn putty
friend of yours?
jepler, tkemc works
jeppler went to bed
well, can you tell me what's tkemc for??
whats it for? its an emc front end
so, a gui-like program to help control emc
it can open your gcode files and such, handle jogs, etc
can it simulate a lathe??
it is an alternative to axis (older)
there are 5 of them, each with various features
AXIS is the only one that has specific support for lathe
i personally use axis because its fairly pretty :-) but thats me
AXIS is also the only one that shows a preview of the program before it's run, but others (including tkemc) can show you the program as it runs
hm go figure
i used a bit of ratios to empirically find my SCALE
oh well. it works. night all
lerman_ is now known as lerman
make install is very disruptive :/
ds3: that's why you need to sudo it :D
so it really fscks things up :P
I am trying to install it as a simulator
ds3: ./configure --enable-simulator --enable-run-in-place
sudo make setuid
then scripts/emc and you're donw
problem with run-inplace is I can't use other python scripts that depend on it
so I want a simulator that lets me install entirely in a tree
then install it.. I assume you don't have any other emc installed?
entirely in a tree?
the install follows the LFS .. so it's not going into one place
look at the second half of directory.map
yes, that is what is irritating me for a simulator run
there is no reason a simiulator needs stuff in /etc/init.d for one
well.. patches happily accepted :D
I am this > < close to doing exactly that but you guys would reject it given all I am doing is #'ing it out
you need to do a proper job :P
the emc script needs to be more intelligent
is --enable-simulator exclusive? i.e. does that guarantee I cannot run those binaries with a real time kernel and a machine?
--enable-simulator turns off building of RT components
so there are several things you can't do with such a compile, inlcuding trying to load configs that load hardware drivers
cool... let me see if I can write up a patch to make the emc script not try to load/run /etc/init.d/realtime
(though I know I had parport building under sim at one point, so that may not be completely true)
SWPadnos, you don't sleep do you?
SWPadnos, what does your wife think about you being a night owl?
you still need that - it's a fake RT system
why do I need /etc/init.d/realtime?
A-L-P-H-A, well, she doesn't mind until I sleep late on the weekends (when she'd rather be out doing stuff :) )
ds3, I think it's a fake RT script - it runs a userspace "rtapi" executable
but I'm not sure - I've only used it with --run-in-place
things seem to work w/o it
well in that case - I guess you don't need it :)
something else seems to start rtapi
I think the first loadrt in a hal does
later folks. :)
does anyone other then me want the simulator to not have a tangle of stuff so it can be a quick gcode checker?
yeah - it is getting late :)
ds3: only if you do it right :)
for an installed system, it would probably make sense
ds3: under form of a .deb :D
I like tarballs ;)
but you'd probably want to get rid of a lot of stuff, like all the sample configs (that can't load) ...
none of that AR or pseudo cpio stuff ;)
*nod* but I'd settle for make install working as a normal user w/ only write privs w/in the install location
anyone know where is the place emc.in becomes emc?
that's what i am afraid of... is it written with autoconf?
you probably want to look at configure.in, then run autoconf
Oh no....... grumble...
looks like configure doesn't need to be touched
ds3: after you ". scripts/emc-environment" in your shell, other programs can use the python modules in emc2, even when you are using "run in place"
morning folks. :)
you added me a link to you yet? :)
[13:01:40] <A-L-P-H-A> http://www.lloydleung.com/
you've been added. :P
I just bought a cable crimper :)
120mm^2 max cable section
that's a bit thicker than AWG: 0000 (4/0)
according to this: http://www.sengpielaudio.com/calculator-cross-section.htm
what are you going to crimp?
I normallz use up to 95mm^2
for what though?
alex_joni, what camera do you have? dSLR?
cradek_ is now known as cradek
will a pair of deep-groove ball bearings on each end of my leadscrew be enough?
as long as they're angular contact, sure
might start building the skeleton out of MDF if I feel enterprising tonight
* jepler looks at some of the boards he's done in eagle
looks like making them fit with one dimension under 1.8" would be a big PITA
it doesn't make that inexpensive proxxon mill sound very attractive
1.8 sure is small
I'm not sure whether the table is smaller still, that was the travel specification iirc
the proxxon machines are surprisingly small when you see them in real life
err - wait. I'm thinking of the Unimat machines
SWPadnos: ever seen a 6-axis Unimat?
I don't think so
I only saw them at Photokina
it is kinda cool that the whole thing fits in a briefcase though
[17:23:29] <alex_joni> http://dsplabs.cs.utt.ro/~juve/thecooltool/pics/
crepincdotcom has the proxxon mill and he said "my whole mill plus CNC setup is 20 lbs" -- sounds pretty small
heh - cool
I'm trying to understand what the 6th (?) axis gives you (the linear axis that's sometimes parallel to Z) that the others don't
do you mean rotary?
ie, C instead of W
no I actually meant what I typed for once
I guess it depends on the machine. I can imagine motorizing the knee on my BP, which would be a W axis
[17:37:03] <cradek> http://dsplabs.cs.utt.ro/~juve/thecooltool/pics/DSC00031.jpg
it looks like the motor at 11:00 controls a linear axis that's parallel to Z right now
it seems that a combination of X and Z movement could always give you the same motion
so what's the purpose of that joint?
maybe the "C" motion is easier to write gcode for?
I don't doubt that
actually, you need to be able to move the whole head up/down, and extend/retract the tool along its axis
easier to implemnt who knows what instead of kins
since A and B are there, the tool isn't necessarily parallel to Z
I just looked at that image and I can't make any sense of it
SWPadnos: as chris said.. if you couple Z&X then you get that
ah - right :)
jepler: there is anothe one :)
[17:40:17] <SWPadnos> http://dsplabs.cs.utt.ro/~juve/thecooltool/pics/DSC00030.jpg
that one may be better
[17:40:24] <alex_joni> http://dsplabs.cs.utt.ro/~juve/thecooltool/pics/DSC00030.jpg
or that one
yeah 30 is easier to understand
I think SWP is right
orient the whole head thing with that extra rotary
then move along the linear for drilling
easier than writing XZ codes
sure, I agree it's easier to write by hand
I doubt the people buying these buy 6-axes CAM
it depends on the ararngemet of the head. it's possible to have the tool oriented such that a simple "peck" motion is a coordinated XYZ move (though not on that head)
a servode BP head would have that problem, since hte two axes are "nod" and "rotate"
argh. typos abound
yes this doesn't have "nod"
Results 1 - 10 of about 411 for servode </google>
like a geode
alex_joni: how do they control backlash on this machine?
cradek: I think there are some screws to adjust backlash
damn - here'sa lot of perfect transformers for mid-size stepper systems (and maybe even servos): http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=290078294981
just like on the sherline
SWPadnos: those don't look like 10 lbs in the pic
I don't understand the markings
no - they don't seem that big
cradek, yeah - it's not clear if there are dual primaries or something
does it also have a multitap 1:1 coil?
alex_joni, hmmm - if it has ha;f the density of steel, then the 10.83 pounds is pretty close
half, that was
SWPadnos: I'm more used to something like this: http://pictures.mastervolt.com/fotofolder/medfoto/afb4vS_ky_.jpg
SWPadnos: did I tell you I got a wire crimper today?
alex_joni, looks nice. maybe a bit big though :)
yes, for that big-ass wire you were talking about before ;)
yeah.. and it was quite cheap
70$ or so
(actually, you didn't tell me - I read back)
oh, and a nice adjustable wrench :)
is that metric? ;)
it's 4 inches :P
for disassembling coating pumps
there was a kid in advanced training (military) at the same time as me, who was apparently annoying. the guys finally asked him to go get a metric Crescent wrench
he was gone for about an hour ;)
i had a brit ask for a "torch" on site in michigan, he jumped when the oxy went 'pop' behind him, he was expecting a flashlight
SWPadnos: thoose are super rare and expensive :D
I think 500eur expensive
I have a new encoder today
bad day today
the laser decided not to play ball ... or more correctly, the controller decdied]
I just wired up my unregulated linear power supply, except I inadvertently wire one of the caps up backwards. It didn't like that. Is there much likelihood I damaged anything else?
[19:33:31] <jepler> http://blog.modernmechanix.com/2007/01/31/giant-voltage-may-blast-atom/
owad: any load on it?
it may be okay, stick a scope on it after talking out the cap
should see 60 or 120Hz ripples
transformer, rectifier, and 3 caps
Is it ok to test a unreg linear without load?
just test w/scope or meter only...
I was getting about -16v with the one cap backwards
and while I was wondering about that, it popped
the cap may be shorted
the backwards one?
or the others?
but if you still see voltage, its probally open ;)
the backwards one
if you remove it and don't see ripples, check the fuses
the backwards one stank up the whole garage
blew any paper on the ceiling?
i assume it is a electrolytic?
didn't check. I just unplugged it, opened the garage door, and figured it was a good time for a snack
the other stuff is probally ok
check with a scope to see if the wave form is ok...it'd depend onhow you have the rectifier setup
ds3: did you see my earlier remark about the emc-environment script to se PYTHONPATH?
I used this wiring: http://www.cnczone.com/forums/attachment.php?attachmentid=29355&d=1168695167
to set PYTHONPATH
(except I only have 3 caps)
jepler: yes... but there is still no reason for a simulator only build to run stuff in /etc/rc.d
(marconi confetti generator = elecrolytic cap & ac ) remove all the caps & see if any ac out of bridge before hooking caps back ( worry about ac not just ripple )
owad: disconnect it at the terminals marked 'DC' and probe with w/a scope...should see 120Hz riples on there
if you see that, then fix the caps and scope the output of that...should be pretty smooth
oops... disconnect at the rectifier DC terminals
owad: switchers are what you can't reliably testw/o a load in all cases
ds3: I don't care to debate that point.
thanks. (I was certain of that until this one blew up in my face! :-) )
ds3: that script still performs a function even in the simulator system, and it is desirable for the simulator to be largely the same as the real version in terms of installed file locations.
jepler: yes I agree... I am making up patches to check if it is a simulator, don't bother running /etc/init.d/realtime..that's all
the only other thing is to not bother installing things there
ds3: but 'realtime stop', 'realtime status' and 'realtime restart' all do useful things with a simulator system
yes. For instance, "realtime stop" makes sure rtapi_app exits and removes a shared memory segment.
rtapi_app seems to be managed elsewhere
I'll let you get to the bottom of it, then.
good point about the SHM :/
jepler: I really want a nonintrusive setup on my normal workstation for G code checking
then why not use run-in-place?
don't have much of an argument for that right now
i suppose I can go back now that you mention the script... it all started with that gdepth.py stuff not running the other day
I'm sure the existence and purpose of emc-environment could be better documented
of course, I am still wondering why make install was trying to install stff in /usr/share/locale/de/LC_MESSAGES
the impressing I am getting is, the entire simulator side is not well documented
as opposed to /usr/local/share/locale? That might be a bug
the 'de' part
what do you believe the proper path would be for the german translation files?
most other packages default to using --prefix .... then append share/locale
I was surprised it was installing german files...that's all
but I do believe it should go in --prefix ... share/....
why would 'make install' not install all the available translations?
donno... guess I am not used to internationalization :/
per ./configure --help:
--datadir=DIR read-only architecture-independent data [PREFIX/share]
I can look at it more later I suppose
if I had to guess why /usr/share is hardcoded, it looks like python's gettext looks for translations in /usr/share/locale by default
and would not look at /usr/local/share/locale
(that is, assuming python's --prefix was /usr/share)
Hmmmm I see... except python's --prefix is /opt/emc32 is my case ;)
now, that might be a 1-liner fix too
sounds like your system is pretty unique
unique, yes. but I am just using features provided by the packages
like I said, I want a standalone simulator and I was hoping to make a tar ball of /opt/emc32 and move it between my machines
Hmmmm this is an interesting option:
--with-locale-dir=DIR Location of the locale file(s)
I see where /usr/share/locale is coming from
the --help is just wrong ;)
guess I'll install autoconf later and muck with configure.in
looks like emc-environment is documented in two places: the AXIS manual section "Python Modules" and the "Compiling EMC2 from source" chapter, section "Run-in-place"
from the AXIS manual: "To use these modules in your own scripts, you must ensure that the directory where they reside is on Python's module path. When running an installed version of emc2, this should happen automatically. When running "in-place", this can be done by using scripts/emc-environment."
where is the axis manual?
[20:07:41] <jepler> http://linuxcnc.org/docs/2.1/EMC2_User_Manual.pdf
apparently that's Chapter 9, page 91
hey jepler hows it going
(while watching live firefight in port city of Umm Qasr on TV)
<eva> fucking campersmy god
i am so sorry
* crepincdotcom kills right click on his mouse
ah... that is new to 2.1 docs
(axis was separate from emc2.0 and really didn't have any docs of its own)
03jepler 07v2_1_branch * 10emc2/docs/src/install/compiling_emc2.lyx:
* talk about TRUNK, not HEAD, for newest CVS development versions
* talk about --enable-simulator
* mention putting emc-environment in ~/.bash_profile
03jepler 07TRUNK * 10emc2/docs/src/install/compiling_emc2.lyx: merge rev 220.127.116.11: HEAD vs TESTING, emc-environment, --enable-simulator
03jepler 07v2_1_branch * 10emc2/docs/src/gui/axis-limit.png: the image came out as black rectangle so I made the background white instead of transparent
03jepler 07TRUNK * 10emc2/docs/src/gui/axis-limit.png: the image came out as black rectangle so I made the background white instead of transparent
anyone around to clarify some things with G2/G3 for me by chance?
do you know howto imatate a potentiometer electronically?
wait a sex
just a quick clarification, youre not trying to read a potentiometer from the parallel port are you? ie, your goal is to use it as an encoder?
sorry to just ahead, just had to ask
No I wish to imatate a pot digitaly for example I have a speed controller that uses a 5k pot to set speed but I wish to do it digitaly.
how's it connected, like a voltage divider?
yeah, it depends entirely on the circuit it's in
crepincdotcom: what's your G2/G3 question?
I think so
if it is, and the rail-to-rail voltages are sane then PWM and a cap that gives a short enough time-constant would be simple
otherwise I think there are circuits that do that same thing if the rail-rail voltage doesn't happen to be that nice
I was looking at a digipot ic but the denford I have is was made in circa 1988 and these where not available then.
how does the pwm / cap setup work?
just have an output PWM pin that has an average voltage which is the voltage you want and a cap, you may need a diode if the output doesn't like being held high/low when the set state is low/high respectively
the cap smooths out the voltage
pulse width modulation: depending on the % of time the signal is high versus low determines the capacitor voltage
larger cap gives a smoother voltage, but a slower delta-v
and vice versa
oh, you'll want a resistor too if the cap is isolated with a diode
keep R*C (time constant) something sane for your application
[21:10:10] <jepler> http://emergent.unpy.net/files/sandbox/pwm-to-5v-analog.png
OK I think I understand I take the 12v output from the controller chop it and feed this back to the input depending on the pwm % will determine the speed of the machine right?
jepler; that assumes that the output can be held in the opposite state it's set to
mostly that should be safe though
Comming back to the digipot ic they have up / down pin (high low) could this be easily interfaced with emc2 ... Its probably a lot easier to plug one of thease in. example > http://www.analog.com/en/prod/0,,761_797_AD5228%2C00.html
The previous one is a push button this one uses a clk / up/down configuration. http://www.analog.com/en/prod/0,,761_797_AD5227%2C00.html
you could use stepgen or freqgen to generate the pulses
according to the PDF on the first page, the voltage allowed on the A and B terminals is only up to Vdd
Jepler I think the first one would be easier to interface ie 1 pin for and another for down that should do the trick.
to avoid any complexities with isolation etc., I'm tempted to put a tiny stepper on the speed control potentiometer on my lathe.
err, surely you jest
Theres another solution but so elegant ;-)
that's like taking a sledgehammer to kill a fly sitting on a precious glass table
lerneaen_hydra: I guess you have some pretty strong feelings about that
These IC's are like 39 cents each not bad value ... lol
jepler; well, I for one would definetly not go for a mechanical solution when a solid-state one is available
This one does'nt look to sad as it comes in a DIP format > http://www.analog.com/en/prodDesc/0,2895,AD5220%255F0,00.html
I think it would be a clever hack actually
you need a driver for the motor, you have to home it, etc etc
homing is really the issue
IMO a rather clumsy solution
and I can't really see the benefits
(I'll stop whining now)
Also you will need pid to get constant speed for threading an so on... stepper > pot may not be accurate enough.
you don't need closely controlled speed for threading - that's why there's an encoder on the spindle
you probably dont even need feedback to EMC if the speed control is any good
Sure but will you get the fine control by ajusting a pot physicaly i mean?
paragon36: yeah, plenty good enough
a pot will probably spin 270 degrees, and a 1.8 degree stepper should give at least 100 steps for that range
so more than enough
20-30 steps would probably be enough
I agree with that but the pot track is not always perfect is it?
paragon36: on mine it's pretty linear
good enough probably
then again, go for PWM and it's 100% linear ;)
only in the 10% to 90% range :)
PWM tends to have nonlinearities inear the edges
I don't dought it the pot I have been using to manualy control my mill is a bit eratic in places probalbly due to its age and finding it in my bit's and bob's pot ;-)
depends on the pwm implementation
if you have an error-integrating PWM it should be good
lerneaen_hydra, somewhat, but none of them is perfect - remember that the transistors aren't perfect switches
jepler: i just noticed pyvcp doesnt return the command prompt (eg: 'bin/pyvcp -c fred m5i20-test2-di.xml') . i wanted to open multiple panels. (in, out, dac, encoder, misc ), any ideas on how to get more than 1 panel up?
SWPadnos; no, that's true, but good enough, definetly in this case ;)
& is your friend
lerneaen_hydra, yep, for sure :)
& doesnt work, if you try maybe you can tell me how
oh. then & isn't your friend ;)
tomp: to run any unix command in the background, put "&" at the end of the commandline. Compare the behavior of 'sleep 5' and 'sleep 5 &'
FYI I purchased a new spindle controller off of ebay for 35ukp including shipping it has the same breakout as the original denford. >> http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&ih=006&sspagename=STRK%3AMEWN%3AIT&viewitem=&item=160079098176&rd=1&rd=1
I had tried bin/pyvcp -c fred m5i20-test2-di.xml & and the command prompt didnt return, it waits till the panel is closed.
that is seriously messed up
did you try pressing enter?
sometimes the prompt doesn't get printed but control has returned to the shell
tomp for completeness you can also run the command then ctrl z then type bg to put it in the background (fg brings it back to the foreground)
[21:34:20] <cradek> http://surpluscenter.com/item.asp?UID=2007013115303699&item=11-3167&catname=
no - ctrl-Z should suspend pyvcp and return control to the sheel
^Z == sleep
SWPadnos: would something like this ^^^ give isolation? i.e. hook right to the parport and an outlet?
fjungclaus__ is now known as fjungclaus
no, return was the trick, needed to be nudged
I mean, np
OH the command did not return to the prompt?
cradek, it looks like it should work, since the 15ma is within the parallel port spec (I think)
right, and a return got the prompt, using kterm on gnome :/
anybody with more experience than me want to weigh in on exactly how wrongheaded this circuit is? http://emergent.unpy.net/files/sandbox/pwm-to-isolated-analog.png
tomp: yep it does that sometimes.
no, even better, the prompt does return, then pyvpc takes it back , pasting it now ...
SWPadnos: I was asking mostly about isolation -- the datasheet says Dielectric Strength 4000 Vrms
does that mean what I think it does?
jepler; why not have one optocoupler, and connect the output to an RC object?
I think it does
what 4000V before failure of issolation>
[21:39:19] <tomp> http://pastebin.ca/334782
lerneaen_hydra: because you need to both charge and discharge C1 (I think)
tomp: that page is not loading! it hangs...
hm depending on turn on/off characteristics I suppose that circuit may short V+ and GNDA
jepler; uh, let me just sketch something
tomp: it does not take it back it is just telling you that it has finished.
tomp: try executing the program with && echo "Finished" added to the end of the command. This will print the Finished if the program executes correctly.
this assumes current_limit << resistor_RC_loop
jepler, you probably want a single opto, and a push-pull transistor pair on the isolated side
also note (as you probably already did) that V+ needs to be isolated from Vcc
lerneaen_hydra: if current_limit << resistor_RC_loop then isn't the largest output voltage possible much lower than V+?
if they were reversed that would be true ;)
tomp, you already got the prompt - it's before the status print from pyvcp
if current_limit << resistor_RC_loop then doesn't a small duty cycle give an output much closer to V+ than a linear response curve would?
the max voltage is Rbleed/(Rclimit + Rbleed), but I think it's just as linear otherwise
oh, hmm, that may be true
Jepler: regarding your lathe speed control take a look at this doc fig1 (looks like step/dir) should be very easy to implement. http://www.ionaphysics.org/lobby/robotics/pdf/Ad5220.pdf
doesn't time_constant(charging) = Rclimit * C, time_constant(discharging) = Rbleed * C?
paragon: you're right, it doesnt print 'finished' till the app closes(releases?), meanwhile the app runs and controls i/o. I gotta run some stock stuff to see if this always happened http://imagebin.org/7123
yeah, that's true
assuming that the optocoupler can deliver current
which means that r_charge must be = r_bleed
which gives a max output voltage that is half the optocoupler voltage (?)
sounds like the push-pull approach is best then
it avoids all these worries
more components though :
yeah - it's basically a bridge driver on the isolated side
the only issue would be if both conduct at the same time
well, half bridge
I'm not sure I'd worry about the discharge time being longer. that should be set for the "update rate", and the current limit set for some reasonable charge rate
maybe the best solution would be something like a schmitt inverter that delays the on(off signal?
tomp: I think it is the code not returning a CR to the stdout. Could you run strace <command> then copy the last few lines into the pastebin
ie adding a schmitt inverter
heh - just stick a 74hc buffer on the other side :)
that should be a push/pull output with a little bit of current behind it
paragon36, tomp: no - there's no error there. the sequence of events in the paste
... pastebin is run program, shell tells you that it's backgrounded, shell prints prompt, pyvcp prints status message (with CR)
SWPadnos: i dont have the prompt, the terminal is not useable. is that to be expected?
you can't re-focus the terminal window and get a prompt if you press <enter>?
SWPadnos I know the program is exiting correctly but the shell is not returning the prompt unless a manual CR is hit. This happens alot but can cause issues if one is using the output to pipe to another script etc etc.
SWPadnos, yes, i can re-focus, but I guess I'm concerned about automating opening panels that way
i found my panels getting too complex, so i factored them, and got stuck
03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd_completion.h: seems that readline.h depends on stdio.h being included, at least on some old readline versions
paragon36, it *is* returning the prompt! it's just done before pyvcp has parsed the input file and printed its status line
you can plainly see it on line 3 of the pastebin
tomp, in a script or .hal file, it won't be a problem
you would end each line in the script with &
note that halcmd loadusr starts the program in the background by default
or, with -W or -Wn, waits only for the new HAL component to call hal_ready()
ok, trying that now thanks ( and trying the strace too, i had to revert to original pyvcp_widgets, only to see same behaviour ).
or the new waitforcomp (or whatever) command
you could also modify pyvcp to not print anything by default, then you won't see this problem with its output competing with the shell prompt
I agree with you SWPadnos... it's the shell that is not handling the return correctly (ie it needs to be prompted to return)
The shell is handling it the way UNIX has always handled it
bad unix shame shame
it is handling it correctly. the program is just printing crap after the prompt
or you could redirect the output: pyvcp ... > /dev/null (and/or 2>/dev/null as required)
line 3 of the pastebin shows me "shell prompt" "exraneous text"
extraneous, that is
Some programs return the command prompt after exiting other don't. I have never realy looked into the reason why this is I have just accepted it.
you could type a command instead of just <return>, and it would work
it's a scheduling thing
when you run a program in the background, there's nno guarantee that bash will get to run before the program
the only thing that's (probably) certain is that the  8307 (or whatever bash process number / PID) will print before any program output
sleep 1 && echo asdf &
Really? I thought the program retuns to it's parent. In this case the shell?
the program isn't returning - it's being run in the background
it finally returns when you close the pyvcp window
at that point (once you press enter), the shell will tell you that the program has finished
good night all
see you Alex
SWPadnos: Im still a little mithed why some programs return with pid and done and others need prompting to do so?
if you specify that the program should be run in the background (with & appended to the command), then you get the pid and done messages
if you just run a program, you don't (and you don't get the prompt back until the program finishes)
Sorry I mean background jobs!
then I don't understand the question
some background jobs return on exit with pid program done. While other background jobs just sit there until one hits the keyboard or the return key then they indicate that they have completed ie pid program name done even though the background job could have finnished sometime in the past.
AFAIK, no background job "done" messages are printed until you hit return
the program itself may print something, which will happen any time, but the bash "this pricess has finished" message is only printed when the user presses enter
run from script with & on tail of cmd, same result , going to try pyvcp ... >/dev/null
It could be a bash thing SWPadnos ... I use KSH on Solaris much more then Linux and I am sure it returns with intervention...
that could be
with out intervention
tomp, try adding "sleep 5" to the end of the script (with no &, of course :) )
ah thebitbucket trick works, shell is available to me after running the script
Ill try it tommorow... runing windows here right now... tut tut..
that'll give you a prompt after 5 seconds, which is probably enough time for hte messages to get printed
The shell learns immediately whenever a job changes state. Normally,
bash waits until it is about to print a prompt before reporting changes
in a job’s status so as to not interrupt any other output. If the -b
option to the set builtin command is enabled, bash reports such changes
immediately. Any trap on SIGCHLD is executed for each child that
bitbucket aka > /dev/null & will try the wait now
it's man bash
sorry if what I've pasted doesn't look pretty
so if you type "set -b"
then type "sleep 1 &"
That's the answer! ... ksh can be annoying on Solaris as when one runs a bg job and it returns it sometime screws up the command prompt usuarly when entering a huge perl oneliner ;-)
then it will tell you immediately when it's finished
sleep 5 ok
sleep 1 ok
tomp, np. it's probably better to have errors printed to the terminal if there are any
yes, but i wanted to get 5 panel up , in, out, dac, encoder & misc ( avoiding notebook widget in tkinter )
next i want to hide the unwanted panels by radiobutton ( imitate notebook)
I was just in the garage working on the power supply for the stepper drives. I get the appropriate 24vdc out of the bridge rectifier, but when I attach the two 6300uF caps and the bleeder diode, I just get a trickle of voltage at the caps. Then I noticed it started to smell like it did before the third cap blew last time (the one I hooked up backwards), so I shut it of. Are there any potential causes, besides bad caps? The bleeder r
Should it not be a bleeder resister?
what is a trickle of voltage?
less than 1v
so was the current nearly short-circuit current?
oops - resistor
incorrect resistors? shorted resistors? shorted caps?
the resistor is only 570 ohm which I think is a bit low, but wouldn't that just cause it to bleed slower?
owad: try it without the bleeder!
bridge putting out ac ?(trickle read on dc range?)
without the caps attached, the bridge puts out the appropriate 24vdc
with the 570 ohm bleeder you should get about 1A of bleed
24VDC or 24VAC?
chek for any ac from bridge, ac kills electrolytics
without caps, I strongly question that measurement
24VDC from the bridge
it should be a 120Hz AC signal, with an average DC of 24V
(assuming it's a 120 V -> 24 V transformer)
I'll go check again, with both AC and DC, at the bridge, with caps off
* robin_sz questions that assumption
I think the clue is in the word "average"
ok. It is in fact a 120v->24v transformer. This one here: http://www.mpja.com/category/Transformers/24V_CT_(12-0-12)10A_TRANSFORMER_7846_TR.asp
on the outputs of the bridge rectifier, the multimeter gets readings of:
without capacitors presumably
how nasty can you let the power supply be for the rest of the power net?
if you have a transformer and some caps it'l pull current from the net unevenly
depending on the cap size and power dray
the drive I'm using (EAS Microstep) recommends an unregulated linear power supply.
and 2 of those is all I plan to hook up to it
umm, sorta. you tend to get a fairly even sinewave of current on the primary side as the bandwidth of the txfmr is fairly low, sow the primary side doesnt see the peaks
03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd_commands.c: fix getting the count of arguments in loadusr
robin_sz; doesn't that depend on the transformer?
size, inductance and so on
to an extent yes
Should I be getting 9.7 vac on the rectifier outputs? I think somebody said earlier that could be a problem.
but txfmr are nothing like as peaky as switch mode supplies that rectify the mains in
owad, yes, that sounds about right
you'd be better off looking with an oscilloscope ... strangely, i always find troubleshooting DC supplies is easier with a scope than a meter
owad, it's AC because it's varying with time, not because it's going "negative"
if that's all good, my inclination then is to try a different cap, since I have no idea where these came from (I either pulled them out of used power supplies or got them with some other equipment)
owad: did you try it without the bleeder resister?
I can do that
and I can take it down to just one cap, too
I don't think the resistor would cause a cap to fail
but if it's putting to much drain / short that would indicate why he is getting such a low voltage right?
03jepler 07TRUNK * 10emc2/tcl/emc.tcl.in: we have IMAGEDIR, why aren't we using it?
wheee! 1 script, 5 panels, terminal free, ... onto hiding all but the desired one http://imagebin.org/7124
paragon36, the resistor would have to be much lower resistance than the transformer to cuase a voltage drop like that
tomp, cool! :)
I'm sure your right SWPadnos. I'm no expert with linear power supplies but I just thought take it out to minimilise the equation. :-)
good idea, always :)
owad: the ac seen after the rectifier is called 'ripple', and 9 vac for 24v ac seems large, a scope will show you more. you look for any swings below ground ( real ac ) if none, then the bridge is 'rectifying' (reversing all negative swings to above the 'ground' ... you wanna see bumps, not snakes :)
it seems kind of small to me, considering that the measurement was with no filtering caps
oh, no caps, then it can swing to gnd potential
but... no lower ( eats caps _
It's not going negative we dont think.
right - the low measurement may be due to the frequency not being what the meter expects, or averaging
use an adc and hal :)
I think I am definately going to invest in one of those AD5220 digipots and hook it up to my new spindle controller. :-)
just remember isolation when you do it ...
Isolation from para port > digipot, digipot > controller or both?
Looking at http://www.ionaphysics.org/lobby/robotics/pdf/Ad5220.pdf
fig1 I dont think it can be isolated or its not need as I understand the pot side is direct replacement for a traditional pot.
the pot side is internally linked to the logi side
so you likely need isolation somewhere between the speed controller and the computer
Agreed. I use TL94 ( from memory) opto isolaters from paraport to driver control logic.
s\isolaters\isolators\ not sure why I am bothering correcting my spelling at this late stage ;-)
Hope owad is OK he's been a while! Hope he hasn't shorted out!
SWPadnos, 9.7v seems about right to me .. its probably a true RMS meter, and if you tink abou thte shape of a recified sinewave... 24VAC .turn one halfwave the other way up .. at best 12Vac, then the formfactor .. 9.7 seems about right
it could be, especially with the 1.414x factor taken into account
12 x 0.717
that would be ~8.5
* robin_sz surugs
so 9.7 is not that low then
I think you just can't measure that waveform with a meter and get something meaningful - it's the wrong tool
anyway, it's within an order of pi :)
1W/1.414VA = 0.707
cradek, right .. scopes are the tool for DC supplies
and spectrum analyzers for AC supplies :)
meters measure sine waves and DC only
I know you all know that :-)
oscopes ftw :-)
even worse, they generally measure sine waves at 60 Hz (or 50 in europe)
crepincdotcom: hey did you figure out your arc question?
cradek: no no yet
to be fair, modern meters of qulaity (eg fluke) are pretty good upt to 20khz or so
nothing's "wrong" per se, im just wondering about the syntax
so you supply an endpooint and a point to pass through on the way?
ok, then I'm unclear
endpoint and center or endpoint and radius
[23:14:01] <cradek> http://www.linuxcnc.org/handbook/RS274NGC_3/RS274NGC_33a.html#1002442
SWPadnos, when I was with the BBC, we checked our Fluke "true rms" pocket meters agaisnt a calibrated audio source from 10hz to 20 khz and they where within 1db of where they should be
i guess i'll have to fool with it to make it more clear
to find that, I opened the gcode quick ref, and clicked on G2
haha, you could read the docs instead :-)
i've read them, but it just wasn't clicking so I thought I'd ask here
robin_sz, sure - with a $300-$1000 meter, I'd expect correct results. fora Radio Shack $12 meter, maybe not :)
center format is preferred because R format is ambiguous in some cases
It works without the resistor - now I'm getting 34.4vdc, and no smell.
owad, what value resistor did you have?
hmm. that shouldn't have caused any problems
hm... whats an example of an arc starting at x1 y0 z0 and ending at x0 y0 z0 with a center at x0 y0?
SWPadnos, well, we had Fluke 77s and 79s :)
the endpoint and center can't be the same
sorry, end at x0 y1 z0
I did think of a posible way for the resistor to kill the caps - heating due to charge/discharge current
after I unplugged the power supply, I used that same resistor to discharge the caps, and nothing complained
ok go to the starting point: G0 x1 y0 z0
arc to X0 Y1 with the center -1,0 from the start: G2 X0 Y1 I-1 J0
IJK is relative to the start point
ohhhhhhh that was the problem
thank you :-)
to get the other half circle, just switch to G3
to get a full circle, make the start and end points the same
whats the difference again between g2 and g3?
cw vs ccw
thanks again, ill work on that now
nice pix on g02 g02, cw, ccw, xyij xyr http://www.etec.wwu.edu/faculty/McKell/nccnc_prog3.pdf
owad: I had a similiar problem on make shift powersupply for an induction heater I was building. In the end I just shorted the caps whith a resister when I was done with them. Not the best solution.
Any harm in leaving them charge (assuming I don't touch them)?
not the safest method! I would'nt personaly.
hmmm ... now, thats what I call a mill!!
Marin’s 5 axis equipment, which has the following specification:
Working Range X x Y x Z 13000 x 4000 x 2000 mm
Feed Speed up to 150 m/min
Precision +/- 0.10mm
Controller Bosch CC 220 M
Spindle 12 KW, 500-20,000 rpm.
I'd use a bleeder resistor if you can. are you sure that one was 570 ohms? (could it have been 57 * 10^0?)
it's down in the garage, so I can't check it right now, but I sure thought it was.
was it a standard resistor with color bands, or one with a number printed on it?
How much resistance would you recommend, ideally? I was looking for a 1k, but the closest I had was the 570 ohm and a 2.6K
you need V^2/R = wattage of resistor
or less, ideally :)
24 x 24 / 1000 = ?
ok. that should be (thinks) - blue violet brown. if it's blue violet black, then it's a 57 ohm
570 ohm is 1 watt at 24V
it's DC 34V, after rectification ...
so 34 ^2 / 570 = 2W
(makes sense, since that's root 2 that 24V gets multiplied by :) )
so a resistor about 6mm dia and 12mm long ... or bigger :)
Is owad using a .25w resister do you think?
not for long ...
That could be the smell he's smelling ;-)
2W dissaption from the armpits = Gotta Hurt. lol
How do I know if it's a .25w resistor?
[23:30:35] <tomp> http://www.bright.net/~agarb/STMD/Power%20Supply.html
(tho i use 3 pages from a '60s article in popular electronics )
owad, how big is it?
,25w is tiny , fits under your fingernail
owad, i gave you approx sizes for a 2w
03jepler 07TRUNK * 10emc2/lib/python/bwidget.py: fix several notebook-related methods so they work
it's the same size as all the resistors I have. The ones you'd typically use on a circuit board
then it's 1/4 or 1/8 watt
That's your answer right there!
jepler: in what scope? pyvcp or in axis ( in c or python )
owad, it needs to be a big fat one :)
tomp: my intent is to make the bwidget tabs usable from pyvcp but that requires some more extensive changes
size of the filter on a cigarette
I've got some slightly larger ones, but they're all exceedingly old
You could uses larger values in parallel right?
old is good for power resistors
tomp: if I commit some changes to vcpparse.py and pyvcp_widgets.py, will I be stepping on your toes?
jepler: yes, i got tabs to work, but couldnt add stuff to thier frames... no problem on m end, i can ewrite pretty quick now
ie 2 1k in paralell = 500ohm at .5 watt
thers that spelling thing again ;-)
jepler: you gonna add your pybwidget module?
assuming .25w resistors yes
so ... 8 to be safe
tomp: that's what bwidget.py is -- it's already in emc2
8 4K7s in //
how about 10, for a safety margin? :)
or some old carbon rsistors from the junk box
Yep That should do it 10 * 5k6
or my preference ..
those old ones are still only 1/2 w
dont bother :)
the electronics will sag it down, eventually :)
an average human puts out xxx watts of heat ... guesses?
when chatting on IRC, or actually doing work>?
What about a neon switch? or is the voltage too low < 55V (cant remember)
when typing furiously?
olympic sportspeople can output around 200W for a while. normal people are probably <100W
though that's work, not waste heat
100W consumption average 500W hard labour ... thats "energy used" though, not output
My Wife is on about 3-5KW when she gets talking ;-)
the brain uses 20 to 40W of that
jepler: i been messing with that for days :-P, but now we got it , thanks ( was that python or pyvcp ? got code / )
What wattage does a peanut contain?
tomp: the XML is here: http://emergent.unpy.net/index.cgi-files/sandbox/tabs.xml
:) i keep trying to click on it , lol
tomp: do you want me to commit the code to make it work?
it? notebooks/tabs? sure, i'llpport it with the m5i20 xmpl
CIA-8: where are you?
and what have you done with Scooby Doo?
Good Talking with you all, I'm off to bed... Night!
err, anyone know how to read a message?
sending them is easy enough
try aasking memoserv: /msg memoserv help
ah, got it to work
msn memoserv help only told me how to send a message :p
strange - it should have told you how to read them as well
apparently you use the syntax msg memoserv read [n] where n is the number of the message
hmm, now it gave me the whole list
before it seemed to truncate it to one row
jepler: i 'spose i need to checkout 2.1 cuz 2.1-alpha0 gives me a blank wndo :(, while all my panels still work
/msg memoserv list
that may help
tomp: you'll need TRUNK, that's where I put my changes
I got it to work
meh, I'm off to bed now
see you LH