ray's png of full vs halfstep torque & resonance is amazing. http://imagebin.ca/view/clpgA5sW.html
[01:24:38] <skunkworks> http://www.electronicsam.com/images/house/PICT1275.JPG
[01:24:45] <skunkworks> http://www.electronicsam.com/images/house/PICT1274.JPG
skunkworks: are you *still* not done with that kitchen? geez.
I mean, looks like nice tile
why stand the spacers up instead of laying down at intersections?
they go away and you fill in with grout
yeah, but the same is true if you lay them down at the corners of the tile
pro: keeps the rows in line
con: harder to pull them out
oh, I was thinking underneath (which doesn't work so well I tbelieve)
yeah, I think we planned to re-use outs as well
tile looks nice, but I think I'd rather have good old formica
this was big in se asia, concrete cast in place & ground smooth , http://www.amazon.com/gp/reader/1561584843/ref=sib_dp_pt/104-0426989-5911954#reader-link
or that solid stuff the name of which is on the tip of my tongue
tile/stone have one problem: you have to be careful with ceramic/china dishes or they break
our kitchen's tile (smaller tiles than that, maybe 1/4 the size) and it's always a pain keeping the grout clean-ish
2nd problem is cleaning - don't care how well you seal the grout, it will be harder to keep clean than a seamless surface
SWPadnos: yeah, corian
tile on the floor isn't so bad
that's where we put it :)
but even there... if I drop a plate or glass on my lineoleum floor, I cringe, but most of the time it bounces
between the garage and the downstairs bathroom (the one with the deep sink)
if you drop one on tile, go get the dustpan
use melmac dishes ( from the planet were Alf lived )
I'd rather have a melmac floor and counter
[01:50:55] <tomp2> http://en.wikipedia.org/wiki/Melmac_(planet)
I always wondered how many of the kids who watched that show actually knew what Melmac was
Hi folks. I have developed a problem on my M5I20 servo knee mill that I think is configuration or software breakage and I'm somewhat at a loss how to proceed. Some things still work fine but I'm getting some bizarre GUI behavior and unresponsiveness after loading certain test ngc files. Anybody up for trying to help me?
what exactly is happening?
Hi John. Chris Helgesen here.
Homing and jogging still work fine.
When I run the EMC logo file or tort.ngc everything still seems to work.
jepler: yes -still not done. Started at christmas with a vacation in between. I am slow.
However, when I load 3D_Chips or some of the other example files and try to run, I get no motion and Axis becomes extremely unresponsive.
These worked fine a few weeks ago.
does the preview look OK before you hit run?
Since then, I have done some tuning, added a few new I/Os, and some simple HAL logic. I haven't done anything (knowingly) that I would think would cause severe software breakage. Of course, what I don't know is a lot. :)
Yes. The preview looks fine and can be manipulated before as normal. Once it goes into wacko mode, the graphics window updates are infrequent (big jumps).
if its only certain g-code files that cause the strangeness, its unlikely to be a hal thing
the "wacko mode" is probably something sucking up CPU cycles
see if top will tell you what it is
I'm a linux newbie so I'm not good at diagnosing such things.
start by opening a shell, and run emc from there, not an icon
Chris_sub_1_, did you add any functions to the base thread?
that way you can see errors and warnings
next, open another shell, and type "top" at that prompt - it will give you a list of processes running on your computer, sorted so that the CPU hogs will be on top
the "top" list updates automatically, hit "q" to stop it
Hmmm...I did add 'blocks' to core_servo, but it was working fine (an and function for preventing my mechanical spindle brake from kicking in until the spindle stopped).
that shouldn't have much effect
btw, blocks is deprecated - there are now individual components for everything that was in blocks (and more)
loadrt and count=1 (or however many you need)
yeah - I was just thinking that something extra in the base thread could start to cause issues if BASE_PERIOD is close to the limit
make that "and2", for 2-input and gate
though it should happen for any NC file
That's what I did. Do you mean you just add 'and2' rather than using blocks?
good luck. gotta get some sleep
Is that in the html docs yet?
yeah, blocks was deprecated for 2.1.x I think - I know it was deprecated for 2.2.x
I've been working from a pdf that may be out of date.
Back to the problem...I'll go run top and come back and report.
(Embarassing) newbie question: How do I launch my M5I20 config from a shell?
just type "emc"
the config picker will appear, and go from there
Xorg and Axis are at the top, but they are both under 10%.
top only sees user space stuff, so the problem must be something loading down the RT side
I hit stop and it took about 10 secs for the GUI to respond, same when I hit the MDI tab.
while emc is running, go to the other shell, and type "halcmd show thread"
then paste the output to pastebin.ca, and paste the URL here
[03:05:42] <Chris_sub_1_> http://www.pastebin.ca/937669
your thread times seem OK - 1mS servo thread, that never took longer than 207uS to run
50uS base thread that never took longer than 16uS to run
Chris_sub_1_: what are you using pwmgen for?
Spindle speed control using Steve Stallings' board.
why not use one of the 5i20 dac outputs?
Saving the 4th axis for a 4th axis. :)
are you sending the PWM out thru a 5i20 output, or thru a parport to steve's breakout?
(the reason I ask is because you aren't updating parport at all, and you are only updating the 5i20 outputs in the servo thread - too slow if you are creating PWM in the base thread)
All my I/O is through opto22 modules, two of which feed a pwm and dir signal to the spindle board. The pwm is only 10Hz to give me decent resolution.
Steve's board takes it across an opto barrier and simulates a pot input to the vfd.
it still doesn't make sense to generate PWM in a 50uS thread if you are only gonna update the actual output every 1mS
Now that you mention it...yep.
I don't see any connection between that and your problem though
Is there anything persistent related to work offsets, G92 and so on that could be hosing things up? I was messing with them recently.
BTW - I tried a re-install using the package manager and that didn't fix things.
I believe you can remove your .var file to get rid of any residual offsets, etc - emc will recreate a blank one when you start the next time
(remove while emc is not running)
I don't think that's true
[03:17:06] <cradek> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CoordinateSystems
oops - I should avoid commenting on interpreter stuff
^ this url explains offsets and how to clear them
Another observation: When I'm in the broken state, on the MDI tab no gcode lines are highlighted, even though the run button is on.
run button is on while you're in mdi?
I can switch to the MDI tab with the run button on. Some of the controls are very unresponsive.
that seems strange to me
oh, I can too, forget it
you have to do that if you want to see the active gcodes
ok, I lied a little
oops,. wrong channel
If homing and jogging both work fine, what might be different about even starting to run ngc files (but only certain ones)?
And they all preview just fine.
I agree that's very mysterious
ngc files run in a completely different mode than jogging and homing
coordinated mode, rather than individual axis mode
but the odd thing here is that only some g-code files mess it up
I ran the EMC logo file with no trouble though.
chips is bad but tort is ok?
That's why I was thinking it might be something persistent to do with offsets. Maybe I discovered a non-feature.
I don't think this would have anything to do with offsets.
try turning off various things in the View menu
In the broken state, even loading a file takes a long time.
what's at the top of top when it's doing this?
Just a thought: Could I have fat-fingered something that would cause it to be waiting for my non-existent tool changer to do something?
Xorg and axis are at the top, but both under 10%.
waiting for a toolchanger wouldn't suck up CPU and slow stuff down
what version are you running?
In MDI I can enter g0 commands and I don't get either an error or motion.
I got this in the shell window when I reloaded 3D_Chips: http://www.pastebin.ca/937713
you said something about a non-existant tool changer? what does that mean?
I don't have a tool changer. I was wondering if something might have been made active somehow that was waiting for a hardware event that would never happen related to a tool change.
try tn / m6 in mdi and watch the logs/terminal/responsiveness
It takes the commands without the misbehavior that I see with the problem programs or new error messages to the console.
got a spindle that would need to decel before a tool change? ( anything prior to and necc for a tool change ready )
EMC_TOOL_PREPARE timed out ... dunno, sorry
no, that didnt time out, the EMC_TOOL_ABORT did
I'm not familiar with the tool changer interface at all.
Q - If the program calls for a tool change, and I mucked up something with the coordinate systems so that the machine thinks it should be elsewhere for the change, could that be a problem because there is no command to send it there?
dunno, but you could try the same program sans tool change to test
( sorry running in & out of hotel room doing laundry )
Hey, there's someone who is awake. I'll give it a try.
tomp2: That worked! I commented out the tool change line at the beginning (N50T1M6) and the universe started to make sense again. Thanks! There must be something modal that is on but shouldn't be, or maybe I broke something in the config files.
It really freaked me out when it started running because this is the first time I ran it since I wired up my new spindle control. It's much louder now. :)
cool, hope you can find why the tool change is a problem, right now you just have a work-around.
please talk to the devs about what you saw and what changed that. (domani)
Thanks (all). I'll have to work on the 'why' tomorrow night.
I will let them know once I figure out the cause.
i mean you could ask them/ report the observation and the change
lerman_ is now known as lerman
[06:10:20] <renesis> http://www.nasa.gov/multimedia/nasatv/index.html
[06:23:24] <renesis> http://www.nasa.gov/55644main_NASATV_Windows.asx
10min was the hold, my bad
Anyone in the south awake? Shuttle launch @ 2:28 AM
Oh boy a python update
cradek: wondering if you seen the pic I posted on photobucket for you? 2 days ago
yes, thank you
but, the generator is at the mechanic's today :-)
[14:45:12] <alex_joni> http://www.derailer.org/clientquotes/?qid=509
RS274D => RS274/NGC conversion tool somewhere?
I started writing one a while ago, got side tracked....
RS274D as described in my old machinery's handbook seemed fairly compatible, at least for simple kinds of programs
how much changing does your D code need?
Don't know yet, code generated from pads layout (mentor)
oh maybe RS274D can mean more than one thing
RS274D is also gerber format for PC boards
so it has aperture and tool handling that's totally different from RS274NGC
oh, that will be much harder
and depends what fills it has as well /me been reading spec
don't forget flash vs. (whatever the other thing was) for pad/hole tools
or should I say "tools"
the flash part is easy, except for arbitary shapes
there are other options.. convert between programs.. Like - if you can get it into eagle - there are many tools to create emc compatable g-code
I think there are also some cad software that will take gerber files..
I'm pretty sure there's no conversion between Pads and Eagle
SWPadnos: #2 pencil.... DUH!
you can load gerbers into various PCB CAD/CAM programs, dunno if Eagle can import those
that reminds me - I was going to request that Altium add dEagle export/import to Designer
later gerber has some negative options as well
I didn't know that was part of the file spec. I know most PCB programs let you decide whether a layer is positive or negative image though
there is a layer polarity which is separate from the image polarity
thats RS274X the later spec
I thought the only difference between RS274Xand RS274X was that the apertures are included in an -X file
err - between D and X
no loads of enhancements
GCAM also supports RS274X (Gerber) and Excellon drill files for circuit board milling.
[15:27:48] <skunkworks_> http://gcam.js.cx/index.php/Main_Page
www.archivist.info/cnc/pcb/rs274xrevd_e.pdf should be the pdf
converting from a format of traces/pads to the necessary tool paths to cut isolation around those traces/pads seems like a very hard problem. it's not a simple conversion of file formats.
I agree it isnt
it needs a db of connecting lands building to isolate a land/track/area, then generating a path around
but you have to analyze all the tracks at once, since cutting around one might cut into another
as it doesnt have to be drawn in any order in the original
yes hence a storing joined in a db to chect/test joins seems the best to me
never converting to gerber in the first place seems much better
that's just the wrong representation for milling
true, pcb programs dont expect milling machines though
pardon my ignorance, but what is the 'right' representation for milling?
isolation tool paths and where to drill what size of hole
fenn_: STEP :P
cradek: If I wanted to convert, I'd first convert to a (very fine) raster.
Converting from faster to isolation shouldn't be that hard.
i think gerber (a list of lines) would be easier
anyway it's a moot point since gcam does it already
But gerber isn't a list of lines. It's a list of exposures. (for pads, rectangles, etc)
I saw gcam, not busting a gut to write mine
You have to merge the exposures.
archivist: works for me.
and it has area fills and polygons
it's a list of "shape trajectories" and "shaped points"
Think of it as two D constructive geometry.
since you can have an ellipsoid "tool" go along a path with RS274D
oh, that sucks
I've written code to do that. Take an arbitrary shaped brush and move it along a path. (For the purpose of *raster* display.
I don't know how far the shape definitions go, but I know you can have round or rect tools, and those tools travel the paths
sure, but there's no easy analytical way to go from a swept shape to the outline
there are also flash tools, which I'm pretty sure can have complex shapes (like a star)
fenn_ is now known as fenn
that should be easy enough
Or like a pad.
yes - or a pad with thermal (cross) connection to a plane
[Did someone say star?] :-)
anything involving 'flash' is now useless as a google search
I got as far as drawing the plotted image, in mine
lerman, yeah - the early photoplotters actually had physical apertures that would get switched in front of the light source - they could have any shape you wanted
I was noticing that the pad with a thermal is pretty much a star.
lerman, you forgot telnet the protocol :)
I just bought the laser for my laser interferometer. I now have the light source and the receiver. Still need the optics and some cables.
Yup. I did forget that.
why would anyone want to run emc over telnet?
It makes for a cheap way to give commands.
but ssh already works
locally or remotely.
ssh using what interface to emc?
emcsh or an X11 gui
the EMC2 telnet "server" is another UI running on the machine
i think the telnet hack only does emcsh commands
I believe so - it's emcsh+telnet
most the GUI commands are in emcsh too
but that's what tkemc runs with (emcsh), so it can do anything tkemc can do (for the most part)
and it's quite close to halui too
oh crap. I have to get to the dentist
too bad the short way has a flooded road
Sometimes just getting there can be like pulling teeth. :-)
bbl -- I'm going upstairs to see what I broke trying to add named o-words.
hopefully no vase
anyone using GCAM?
Enco has 7x10 lathe stuff on sale. Thought I'd pass it along.
[18:03:23] <SkinnYPuppY> http://www.use-enco.com/CGI/INSRIT?DCMP=EMC-1408851&PMAKA=110-0800
BigJohnT: I had installed it on the windows side... But about all I did was open it - cause the screen to go black and close it.
skunkworks, I opened it in windoz and it seemed to work
I was disapointed that it would not open a dxf file...
sure are a lot of words
just finished testing my prototype vaccum wrench and it really sucks, my customer will be REAL happy when I tell them...
* archivist wants to see pics of the sucker
hmmm, it's top secret but maybe...
[19:04:09] <skunkworks_> http://www.youtube.com/watch?v=zKZYj51Exn0
(selling it on ebay)
ok just this once http://i47.photobucket.com/albums/f163/johnplctech/HPIM0629.jpg
fsdafsd is now known as Unit41
[20:29:47] <Unit41> http://www.youtube.com/v/Y58Ab192ggQ
my half finished lathe
fyi that z axis not for that lathe
thats my gate
it was easy getting to that point, im stuck for ideas on the z axis though
to me it looks like a big gantry router..
it'll take the weight
right link? (I am not listening to the audio)
double check the link you posted... It is not what I would concider a lathe.
unless I just don't get it :)
<skunkworks_> lathe? :)
* archivist_win sees no lathe, just wobbly vision
oops - wrong user..
I just don't understand I guess.. http://www.youtube.com/user/unit411
unit411 is the user name - The quote is 'start of lathe'
Im convinced i did not see a lathe
great - now we scared him away.
you guys are so mean
dont want to to scare then but....
what was that
I don't know where unit41 resides.. maybe it is a communication issue..
was playing with my router
ah. So - why do you call it a lathe?
because it is
looks like a router
its going to be for milling steel eventually
oh, i bet
far too flimsy
you might even work your way up to plastic with the right cutters
ya but I have more acme rod and another motor for a 2 axis drive system
a lathe rotates the metal
dsafsd_: might be smarter to make something smaller so it's not so wobbly
its not wobbly
yes it is, it's made out of 1 inch pipe
just the top rail
might be good for a plasma table
or oxyacetylene if you cant afford plasma
its for plasma and dremmel
i allready have 60 amp plasma
ready for 110
eventually i'll do wood burnings with hho
oh please shut up
are you jelous ?
i would be
fenn are you by chance part of the ass family ?
smart ass, dumb ass , rude ass take yer pic
dsafsd_ is now known as Unit41
so Unit41, how do you plan to make that into a lathe? (just curious)
at the moment, it looks a lot like the base of a plasma cutter or something to me
i'll probably make a lathe out of something else
spin the discs i need into annular form
know any good diy lathe plans ?
I think fenn had the gingery lathe plans (that is a self-build thing, isn't it?)
cutting and bending forces are quite high in lathes and milling machines Unit41
im going to switch to 3 inch steel pipe eventually
that thick stuff for drilling
with my plasma cutter should I even hope to cut rack out with it ?
i need rack and pinion bad
from what I've seen of plasma, I wouldn't expect to cut rack accurately or cleanly with it
but I've seen precious little
the old plasma table i ran could do it
it would still require alot of buffering
yeah - I was thinking the edge wouldn't be too clean, but if you cut a little oversize and grind back to length, that would give both accuracy and a clean finish
actually cut bicycle teeth out and just make a divinci style pinion
Speaking of gingy lathes, what kind of rpm would a smallish ground spindle be able to run in a babbet/copper/steel bearing like an automotive main bearing with thrust surface built in fore and aft
With just cup feed oiling not requiring a gear/lube box
depends how well made
should run a few thousand rpm no problem
I suppose ballance and surface finish are main concern.
oil retention helps, as high rpm will fling it out
eg small motors run 10-20k in oilite bushes
too much oil will cause drag
my gingery lathe is quite messy since for some reason the oil turns black before it gets flung everywhere
hmm black means wear
What are you using for spindle oil ?
i think its some electrochemical reaction between the sulfur rust inhibitor in the oil and the aluminum/steel to form iron sulfide
i'm using mobil vactra 2 way oil
it does it on the ways too (but they dont fling oil)
What material did you use for spindle bearings ?
301 brass? from mcmaster-carr
oh, no it's bronze
it was five years ago, dunno where the receipts are
I've used some si-bronze for a popet valve grinder spindle. Good stuff
why would you use babbit bearings for a small spindle btw?
Oh just brainstorming for something a beginner could use with built in precision with easily modified parts.
how about skate bearings?
or are you thinking bigger? (in that case i might suggest truck trailer bearings)
That may be more practical
car wheel bearings (taper roller)
Thats what I pictured
all depends on speed range
yes same thing
i bet trailer bearings could go up to 10krpm if lubricated properly
Probably I know seadoo tires can't be more than 12" dia
only trouble with them is the large diameter seals needed
why are seals a problem?
i just bought a $2 crankshaft oil seal for my honda which goes up to 7krpm
has a good oil supply in an engine
it was part of a kit, the original oil seal is still fine (after 150000miles == 60 hr at 10krpm)
* fenn wonders if those numbers are right
ok probably more like 400 hours at 10krpm
assuming the total number of turns is the same
but really it would blow up instantly if the oil ran dry
I think you may be off by an order of magnitude or two
400 hours would be 375miles/hour
or the equivalent of that anyway
top gear 4:1 back axle.....
and most normal driving is more like 2500 rpm
SWPadnos: yeah with no air resistance a honda engine at 10krpm in fifth gear is going about 400mph
air resistance doesn't matter (except for required engine power output, of course)
ok my gear ratios are a little off
what kinf og Honda? (car or motorcycle?)
CRX Si 1.6l
ok, so 2000 RPM ~= 60 MPH
(+/- 100% :) )
I'd like another CRX some day fun cars
150000 miles at 60mph is 2500 hours
yeah and 2500/(10000/1500) = 375 hours
so anyway what does that mean? what is a good MTBF for home-brew spindle seals?
and 3e8 total revs
Felt may do fine
labyrinth for lowest friction
i was thinking just let it leak and then catch the oil coming off a flinger plate
seals also keep dirt out of the spindle though
Old aircooled vw's had an oilslinger on the pulley crank and no seal at the pulley
they probably had a labyrinth as well
Screw vanes made on the pulley
oil flinger is probably a one bladed labyrinth
But a known issue people fix with a real seal when they add paper filtering to try to get any longetivity in dry areas
hi guys; whats the syntax to run stepconf using an existing set of configs for a seed?
click "the use a current config button"
assuming that config was generated with stepconf
humm, its not been modified other than correcting the scale signs, but the dir looks empty
so i make a new one.. I see I can change the sign of the tpi of the screw and reverse the tables motion.
Is the jog arrow the direction the table moves, or the bit in the work?
I used the bit in the work, but for the Y, the keyboard isn't used, only the jog buttons, so which way should the Y table move for a right arrow jog?
I don't think stepconf is quite ready for prime time. This time I made sure the machine had limit settings out of the way by several inches
but the file has MIN_LIMIT = -.01" for 2 of the axis's, and it won't come out of estop