#emc | Logs for 2007-12-24

[00:05:31] <skunkworks> :) just research and make up your own mind
[00:29:28] <The_B> The_B is now known as The_Ball
[00:41:26] <JymmmEMC> I think each had it's merits, but... cool hand luke
[00:59:23] <gezr> yeah, just sad to see that it finally came to that
[01:08:11] <robin_sz> so, do we see les watts on here these days?
[01:12:47] <gezr> hey, robin, thought you might be sleeping
[01:12:52] <gezr> and yes, thats the one
[01:13:10] <robin_sz> gezr dude
[01:13:24] <robin_sz> long time no hear/type/talk whatever
[01:13:29] <robin_sz> are you well?
[01:13:43] <gezr> well, I went to look for ngc.vim and it was still on your site :)
[01:13:55] <robin_sz> heheh
[01:14:11] <robin_sz> liek I ever manage to tidy up?
[01:14:43] <gezr> I went to work for a used machiner dealer doing repair work on machines, to fix them up, then to travel to customers facilites to work on their equipment, warnty and contract, till I got tired of being under paid
[01:15:03] <robin_sz> oh, righty
[01:15:04] <gezr> so I decided i had all I could stand, getting ready to go back to college this spring
[01:15:05] <robin_sz> and now?
[01:15:13] <robin_sz> good call
[01:15:21] <robin_sz> if you can live on no money :0
[01:15:27] <gezr> so maybe ill accually get to help out on emc code for once
[01:15:38] <gezr> yeah ill be okay
[01:17:29] <gezr> I see that emc has made leaps and bounds lately too, thats good
[01:17:41] <gezr> accually has look ahead now too?
[01:18:01] <gezr> I think cradek and alex tweaked all that out but I cant remember
[01:21:52] <gezr> ?
[01:55:56] <skunkworks> cradek rewrote the trajectory planner..
[01:56:56] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?TrajectoryControl
[01:57:49] <skunkworks> also one of cradeks projects.. http://youtube.com/watch?v=7WCrqqoZkPg
[01:58:15] <skunkworks> http://youtube.com/watch?v=0jQy-zc5P6w
[02:02:27] <gezr> i had seen the first hadn't seen the second very nice
[02:13:21] <gezr> is the i guess b axix on that machine stable?
[02:40:40] <dmess> i feel your pain...
[02:46:37] <gezr> robin you gone again?
[02:50:43] <dmess> where is home BATMAN...
[02:54:23] <jmkasunich> dmess: hows the water level?
[02:54:42] <dmess> sweet... what ways that about 15 off horizontal??
[02:55:25] <dmess> still rising... not breeched.. thk GOD
[02:56:21] <dmess> about 9 inches NOW... lookes to maybe receding.....
[02:57:24] <a-l-p-h-a> dmess, sup?
[02:57:42] <a-l-p-h-a> how'd you get flooded?
[02:58:58] <dmess> creek in the back yard...
[02:59:06] <a-l-p-h-a> heh.
[02:59:14] <a-l-p-h-a> creek's are great... until they flood your house.
[03:32:05] <cradek> http://timeguy.com/cradek-files/emc/PICT0099.JPG
[03:32:19] <cradek> first part cut on my 5-axis desktop mill
[03:32:32] <cradek> the chamfering was done by tipping the end mill to 45 degrees
[03:38:23] <jmkasunich> the hole itself was straight up?
[03:38:52] <jmkasunich> I want to see the version where the hole is drilled at a 30 degree angle
[03:40:02] <cradek> heck I'd like to see that too
[03:40:06] <cradek> write the gcode and I'll run it!
[03:40:26] <jmkasunich> damn, look at the time, gotta go!
[03:40:31] <jmkasunich> ;-)
[03:40:57] <jmkasunich> actually, if the face that is being chamfered is perpendicular to the hole, is it that nasty?
[03:41:14] <jmkasunich> (I realise that an angled hole in a flat part requires an ellipse to chamfer it
[03:41:31] <jmkasunich> clamp the board in the vise vertically
[03:41:55] <jmkasunich> mill a flat surface tilted 30 degrees (tilt B, command U and V moves)
[03:42:21] <jmkasunich> oh, to make the hole you'd need UV arcs - those don't exist do they?
[03:42:38] <cradek> no, you'd have to approximate it
[03:42:57] <jmkasunich> lots of G1s you mean?
[03:43:13] <cradek> yes
[03:43:26] <jmkasunich> hmm, that takes the fun out of it
[03:44:15] <jmkasunich> for your test part, was the hole at least off-center on the rotary table?
[03:44:47] <jmkasunich> so during the chamfer, B and Z were constant, X, Y, and C were moving
[03:45:02] <cradek> yes both holes were made by one program
[03:45:16] <jmkasunich> chamfering an angled hole would move all five joints at once
[03:45:30] <cradek> yes that would be cool. I'll have to do it.
[03:45:33] <cradek> later...
[03:45:47] <jmkasunich> I just realised that even UV arcs wouldn't be enough
[03:46:11] <jmkasunich> cause once you tilt the tool to do the chamfer, the hole is no longer parallel to the spindle
[03:46:36] <jmkasunich> that's a brain twister
[03:49:56] <jmkasunich> the 5 axis piece-de-resistance would be something like the top half of a dodecahedron sticking up out of the raw block, with a chamfered hole in each face
[03:50:02] <jmkasunich> all done with the same tool
[03:50:28] <dmess> or just a nice 3 axis wigglle......
[03:54:19] <dmess> top of a dodech is 3 axisable... bottom from the same set up is a stretch,... even for me... i cN MAKE THE CODE... BUT the setup is a little off
[03:56:33] <jmkasunich> with three axis, you'd need to use a ball end mill, and you'd wind up with grooves and such
[03:56:37] <dmess> JMK.... i've done that for DMG
[03:57:04] <jmkasunich> with five axis, you can make the tool perpendicular to the face and mill it with wide passes using a square end tool
[03:57:17] <dmess> yup
[03:57:42] <dmess> and heidenhien..makes it easy...
[03:58:35] <dmess> or siemens... 840d.. sweet
[04:01:13] <dmess> we intepolated a tooling ball... in 5 xis on an 840d.... under a tenth on the indidator.... machine was swinwiging....
[04:01:51] <dmess> numerouus witnesses
[04:02:17] <toastydeath> dmu?
[04:03:27] <dmess> was a DMG machine... deckel-maho-gildemieister...
[04:03:30] <cradek> jmkasunich: I understand lots of 5 axis code is done with short lines and inverse time feed. I think you hit on why
[04:04:28] <dmess> inverse time is a specislty of mine...
[04:05:19] <cradek> better still would be a tapped hole in each face... but that's for another day
[04:06:37] <dmess> 10 lines of code on a heidy or a siemens... honest.. NOT slaggin''''' im one of us
[04:07:15] <jmkasunich> cradek: that would be impressive
[04:09:48] <dmess> if i could show you fellas a shot of a good friends machine... that isnt an issue
[04:10:43] <dmess> pick an angle... mill a plane drill tap.. 3 minutes..max
[04:11:13] <cradek> dmess: emc can do it too - but not on my machine.
[04:12:02] <dmess> i KNOW emc could do it.. but WHO has the machine to try...
[04:12:22] <cradek> nobody yet
[04:13:20] <dmess> he's offered me that machine.. to play with and set to work...
[04:14:47] <dmess> its huge..... 5 meters x axis
[04:47:37] <Gamma-X> whats a good brand vfd? at a fair price.
[04:47:40] <Gamma-X> i need a 5hp.
[04:47:45] <Gamma-X> well i have a 3hp
[04:48:13] <cradek> I don't think you can find a single phase input 5hp vfd
[04:49:26] <a-l-p-h-a> dmess, I didn't remember what you said... how's it going?
[04:49:53] <Gamma-X> cradek i can live with a 3 hp
[04:51:16] <cradek> automation direct
[04:51:20] <jmkasunich> A 5HP drive with three phase input will probably work OK on single phase if you only load it to 3 HP
[04:52:09] <Gamma-X> what about a 3 hp 2 3 hp?
[04:52:13] <cradek> for the AD drives it doesn't say you can drive them that way once you get over the 3hp model (but I know squat about how they work)
[05:14:09] <gezr> cradek nice machine by the way
[05:14:19] <cradek> thanks
[05:16:49] <gezr> do you get good stability out of the rotating axis?
[05:17:15] <cradek> it's about as solid as the rest of the machine I think
[05:17:22] <gezr> very sweet
[05:17:23] <cradek> it has never been a good machine for cutting metal.
[05:17:55] <gezr> even aluminum?
[05:18:29] <cradek> yeah, I used it for metal only when I really needed to...
[05:18:38] <gezr> hmm
[05:18:51] <cradek> I guess it was OK for aluminum. had to be careful though.
[05:18:57] <gezr> Im sorry I havent been good to the project over the years
[05:19:19] <cradek> everyone does what he can, and no more
[05:19:42] <gezr> yeah i guess thats true
[05:20:44] <gezr> I dont even have a box running it lately
[05:21:26] <gezr> in fact last time I was playing really hard, paulc was still in here, and the tc stuff was just fixed
[05:22:05] <cradek> must have been a couple years then
[05:22:12] <cradek> so much has happened since then!
[05:22:18] <gezr> yeah, I seem to come and go every 2 or so years
[05:22:31] <gezr> I get frustrated and give up
[05:22:47] <gezr> I read the dev and norm emails still
[05:23:29] <gezr> hell, I remember when you first started comming in here:(
[05:23:34] <gezr> crap :)
[05:23:54] <jmkasunich> http://jmkasunich.com/pics/surplus_spindle.pdf
[05:24:04] <gezr> so you have or have not drilled at an angle like your video example shows?
[05:24:08] <jmkasunich> fenn: you around?
[05:24:48] <cradek> gezr: only air. I've only actually cut the one program shown in the photo
[05:24:55] <cradek> but the air cuts nicely
[05:25:01] <gezr> hahahah, it always does
[05:25:05] <jmkasunich> don't need coolant
[05:25:11] <jmkasunich> and you don't get chips in your shoes
[05:25:15] <gezr> jmksunich thats an interesting spindle design
[05:25:26] <cradek> yes it's easy to vacuum up the air chips
[05:25:31] <jmkasunich> gezr: actually, its a spindle that I got for $10 ;-)
[05:25:45] <jmkasunich> I need to figure out how to hold tooling and make a housing
[05:26:07] <cradek> jmkasunich: do you know what's wrong with them? just worn a bit?
[05:26:34] <jmkasunich> varies - some are tagged
[05:26:48] <gezr> what taper is that?
[05:26:53] <jmkasunich> a few have hairline cracks around the taper (there are coolant passages that I didn't draw)
[05:26:58] <jmkasunich> gezr: HSK
[05:27:27] <jmkasunich> some have sticky drawbars or other things that need tweaked
[05:27:47] <jmkasunich> as far as I can tell, this was the "needs rebuilt" bin of a shop that went out of business or something
[05:27:51] <gezr> this is full size stuff then eh?
[05:27:59] <jmkasunich> decent size, yes
[05:28:05] <gezr> ah fancy
[05:28:11] <jmkasunich> the taper is 24mm (bout 0.950") at the big end
[05:28:20] <jmkasunich> the spindle is close to a foot long overall
[05:28:25] <jmkasunich> bearing ODs are 68mm and 62mm
[05:29:14] <jmkasunich> the shaft at the top is 28mm - I can get QD bushings for that to mount a pulley
[05:29:14] <gezr> it looks like the drawing is missing the inner lock for the hsk though
[05:29:31] <jmkasunich> it is - I didn't bother drawing that stuff cause I'm not gonna use it
[05:29:44] <jmkasunich> are you familiar with tormach tooling system?
[05:29:47] <gezr> go with a standard draw bar then?
[05:30:06] <gezr> is it like the oh gosh, its been a while, its quick change though right?
[05:30:24] <gezr> with like a side screw that pulls everything together? balanced packages?
[05:30:38] <jmkasunich> http://www.tormach.com/Product_TTS.html
[05:30:59] <jmkasunich> 3/4" straight shank, goes in an R8 or Morse collet, and has a flange that bears on the spindle nose
[05:31:09] <jmkasunich> so when the collet draws it up you get flange contact for repeatable Z
[05:31:37] <jmkasunich> I'm hoping I can make a collet that fits in the taper and is drawn up the the existing drawbar (thats an M6 thread on the end)
[05:32:11] <gezr> that should be farly simple
[05:32:21] <jmkasunich> two issues
[05:32:33] <jmkasunich> the drawbar is too long - I'd have to shorten the tormach shanks
[05:33:02] <jmkasunich> and that reduced diameter section just above the taper will make for a very shallow and weirdly shaped (possibly weak) collet
[05:33:16] <jmkasunich> I'd like to bore that section out, but its hardened
[05:33:30] <jmkasunich> ditto for shortening the drawbar and re-threading it
[05:33:33] <gezr> case hardned I bet
[05:33:46] <gezr> the drawbar is going to be solid
[05:34:02] <jmkasunich> well, I bought two of them, and one has the tiny cracks, so I can give it a try
[05:34:04] <cradek> jmkasunich: could you get a grinder in there, like a dremel?
[05:34:17] <jmkasunich> yes, but it would take ages
[05:34:27] <jmkasunich> that is 17mm, I'd want to bring it out to 22mm or so
[05:34:29] <jmkasunich> and its 12mm long
[05:34:32] <cradek> uh cnc?
[05:34:45] <jmkasunich> uh, 0.0002 per pass?
[05:34:45] <cradek> :-)
[05:35:00] <cradek> ahhhh units clash!
[05:35:07] <gezr> it could be hard turned
[05:35:22] <cradek> if it's case hardened you could bore it once it gets easy
[05:35:53] <gezr> even hardened 41series steel is relatively easy to turn
[05:36:23] <jmkasunich> gezr: first I gotta get or make a steady rest
[05:36:35] <jmkasunich> the only part that fits inside my lathe spindle is the 28mm part at the back
[05:36:43] <gezr> what lathe is it ?
[05:36:48] <jmkasunich> so I'd have to support it by the front beearings
[05:36:58] <jmkasunich> my shoptask 3-in-1 :-(
[05:37:11] <gezr> I dont have anything that would fit that
[05:37:12] <jmkasunich> I also have access to a South Bend 13" at work
[05:37:24] <jmkasunich> but it doesn't have a steady either
[05:37:24] <gezr> thats a bit larger
[05:38:00] <gezr> I should kiss my last bosses ass a bit more, but that repair junk was not making me happy at all
[05:38:41] <jmkasunich> heh, I was just trying to do math for that cut and going "huh?", cause 22-17 = 3
[05:38:50] <jmkasunich> somehow the calculator got into octal mode
[05:39:09] <gezr> I have a calculator question for a ti89 if any of you have one
[05:39:37] <jmkasunich> ok, 0.098" to remove, if I grind it with a dremel at 0.0002 per pass, thats only 492 passes
[05:39:47] <jmkasunich> not counting wheel wear and dressing
[05:40:19] <jmkasunich> I'd still need the steady
[05:40:39] <gezr> whatabout a hydraulic hsk endmill holder?
[05:40:53] <jmkasunich> HSK tooling is WAY out of my league
[05:41:04] <jmkasunich> $300 for an endmill holder, $600 for a collet holder
[05:41:27] <jmkasunich> and that needs all the little bits to be good as well
[05:41:41] <gezr> oh wow
[05:41:48] <jmkasunich> this particular one was in the rebuild pile because one of the six internal collet fingers is bent
[05:42:00] <jmkasunich> I don't care if I'm gonna use some other (cheaper) tooling system
[05:42:25] <gezr> 492 passes is sounding almost fun about now
[05:42:59] <jmkasunich> I'm very tempted to build a steady that clamps onto the dovetail bed of my machine
[05:43:12] <jmkasunich> and make it fit the OD of the big bearings
[05:43:24] <jmkasunich> chuck up the tail end of the spindle, and support the front in the bearings
[05:43:28] <jmkasunich> use carbide and see what happens
[05:44:02] <gezr> like make an outer race block, then use the tail stock to hold the spindle while you insert shims and then clamp the whole thing down?
[05:44:05] <jmkasunich> with the steady, I could also use a grinder to touch up the taper and face, since it will be running in its own bearings - guaranteed concentric
[05:44:22] <jmkasunich> something like that
[05:44:35] <gezr> thats not a bad idea
[05:44:41] <jmkasunich> I was actually thinking of using a between centers boring bar to bore the bearing seat
[05:45:44] <jmkasunich> something to make sure the bearing seat is dead nuts centered
[05:49:36] <gezr> what would you make the bearing seat out of?
[05:49:52] <jmkasunich> for the steady? some big old chunk of steel
[05:50:03] <jmkasunich> actually, probably two chunks
[05:50:25] <jmkasunich> mill flat faces, dowel and screw them together, with a sheet of paper between, then bore
[05:50:40] <jmkasunich> that way I can open it up, insert spindle, and clamp down on the bearings
[05:51:37] <jmkasunich> for the actual spindle housing, I would turn and bore it out of a piece of 3.5" OD by 5/8" wall DOM (2.25" ID) about 6-1/2 inches long
[05:52:24] <jmkasunich> that wall is thick enough to allow for the 68mm and 62mm bearing bores, and a ring of screws around the end to bolt a retaining plate down over the tool end bearings
[05:52:53] <jmkasunich> OD would wind up something like 3-3/8
[05:53:13] <jmkasunich> I'm gonna draw up what I have in mind over the next week or so
[05:53:20] <jmkasunich> too many irons in the fire though
[06:05:04] <Jymmm> anyone running 2.6.20 by chance?
[06:39:59] <fenn> nice drawing..
[11:37:13] <BigJohnT> Dallur: Hello
[11:38:59] <Dallur> hey an bye
[13:07:00] <jepler> 91~
[13:07:02] <jepler> '
[13:07:20] <jepler> oops
[14:42:32] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[14:47:12] <Gamma-X> what uop all
[14:48:03] <jepler> I think everybody's on christmas break...
[14:49:53] <archivist> * archivist at work doin nuthin
[14:50:18] <jepler> same here, actually
[14:50:24] <cradek> yep
[14:51:14] <Gamma-X> lol
[14:51:51] <Gamma-X> anyone know if a 5hp vfd with single phase input to 3 phase output is a vailable? no luck finding it.
[14:57:57] <jepler> 5HP is about 16A 230V or 31A 110V if the vfd is 100% efficient (and it's unlikely to be) -- I think there's a reason you are only finding 3phase for 5HP motors.
[14:58:38] <cradek> you're not expecting 110 in are you?
[14:59:32] <cradek> I'd call automation direct and ask if the 5hp model can be run on single phase 240
[14:59:52] <cradek> their web site doesn't say it can, but jmk said it might work
[15:01:25] <jepler> the 5HP is specifically listed as 230v, 3ph only -- the 3HP one says 1/3ph
[15:02:02] <cradek> yep
[15:02:11] <jepler> at best you'd be operating it in a way that would void the warranty
[15:08:34] <BigJohnT> Gamma, why do you want a 5hp vfd for a 3hp motor
[15:26:40] <jmkasunich> Gamma-X: some drives may have single phase protection that will shut them down if you try to run them on single phase input, but I doubt they put that in small drives
[15:26:54] <jmkasunich> (for me, anything under 10-20HP is small)
[15:28:10] <jmkasunich> the main issue with singlephasing is that there is more current through the input diodes, and LOTs more ripple current through the filter capacitors - with single phase, when the line voltage goes to zero, the drive is running totally on stored cap energy - with three phase there is always incoming power from one phase or another
[15:28:36] <jmkasunich> the drives that are listed as 1/3 ph have larger caps and input diodes to compensate for taht
[15:28:37] <jmkasunich> that
[15:28:59] <jmkasunich> if you oversize the drive, that can also compensate
[15:29:22] <jmkasunich> (but yes, it will most likely void the warranty)
[15:36:05] <BigJohnT> jmkasunich: are you the one that did the xmas art g file?
[15:36:34] <cradek> if he has a 3hp motor and 1ph power, seems to me like he should use the 3hp drive with 1ph input rating
[15:36:44] <jmkasunich> BigJohnT: no
[15:36:57] <jmkasunich> cradek: I would agree
[15:37:19] <BigJohnT> ok can't remember who gave it to me
[15:37:29] <BigJohnT> cradek: I also agree
[15:38:40] <cradek> good we can move on then :-)
[16:13:26] <BigJohnT> anyone use a image to dxf converter?
[17:06:12] <skunkworks> the wind finally died down
[17:39:34] <dmess> here too.. but not before blowing down a bunch of trees
[17:45:57] <jmkasunich> did your wall hold?
[17:52:49] <lerneaen_hydra> cradek, nice image of the wood part with 5-axis
[17:52:56] <lerneaen_hydra> how fast can you run your mill?
[17:53:04] <lerneaen_hydra> got a video of it?
[17:53:18] <cradek> thanks
[17:53:18] <cradek> slow
[17:53:19] <cradek> no
[17:53:41] <lerneaen_hydra> the same speed as the previous video?
[17:53:53] <cradek> yes something like that
[17:54:00] <lerneaen_hydra> oh no..
[17:54:05] <lerneaen_hydra> that's really slow
[17:55:03] <cradek> it will be a bit better when I get joint constraints working. I won't have to be so conservative
[17:56:44] <lerneaen_hydra> joint contstraints, would that be limits that control each joints' maximum speed in contrast to the ini file which I assume only controls each axis' maximum speed?
[17:57:10] <cradek> yes
[17:57:37] <lerneaen_hydra> ah so as it is now you can get commanded joint speeds that are very high or even inf. for certain positions
[17:57:48] <cradek> yes
[17:58:05] <cradek> inf would mean a bad machine configuration where you try to move through a singularity of some kind
[17:58:23] <cradek> but yes the joint speed can add up, especially when moving several axes that each contribute
[17:59:39] <lerneaen_hydra> http://www.youtube.com/watch?v=BPGsMDM_UPU&mode=related&search= <-- I assume that setup can have a commanded roational axis speed of inf when passing through the center of rotation
[17:59:56] <lerneaen_hydra> I'm not sure I would call it a bad machine design though
[18:00:52] <cradek> yes R-theta is a good example of trouble
[18:01:40] <lerneaen_hydra> as long as the software is capable of controlling both axis and joint velocity so they're within limits I don't see any problem
[18:03:09] <lerneaen_hydra> btw, how did you limit the backlash in the index rotary table that you use as your C axis? (I take it that rotary A has its center of rotation along the X axis and so on)
[18:04:15] <cradek> the rotaries are B,C. they both have a bit of backlash.
[18:05:26] <lerneaen_hydra> have you reduced the backlash in them at all? all the rotary index tables I've seen have backlash in excess of 1/4 turn
[18:05:52] <lerneaen_hydra> * lerneaen_hydra is to lazy to figure out how many mRad that is
[18:22:04] <jmkasunich> a project for max5: http://www.youtube.com/watch?v=HsKMQG3brw4&feature=related
[18:41:24] <dmess> toshiba uses that syteme on their large VTL's... and the fanuc 21 it lights up like a xmas tree when you try to pass it thru center...
[19:33:59] <gezr> just a few more days and I get to haul things back up into the attic
[19:48:41] <gezr> and all small rotary tables unless home made can have quite a bit of backlash, they typicaly have a quick production run and the accuracy of them is limited, taig, sheerline are models im talking about
[19:48:57] <gezr> even some large/better tables have some backlash
[19:50:22] <gezr> home gear cutting isnt easy to do either, especially a good worm/screw combo,
[19:50:43] <archivist> heh thats my interset
[19:50:49] <fenn> seems to me worm gears are the easiest to do
[19:50:49] <gezr> mine too
[19:51:20] <gezr> they are, but its getting the matchup correctly between the cut worm, and the pre machined wheel without using a burnishing method
[19:51:22] <fenn> cut the hob on the same piece as the worm and just flip it around once the teeth have been cut
[19:51:24] <archivist> Ive had problems with the error between worm and wheel
[19:52:02] <gezr> and normally we as home machinist use far too small a worm to have an effective combo
[19:52:37] <archivist> the best here is 72 to 1
[19:52:51] <fenn> 72 to 1 what?
[19:53:00] <gezr> worm to wheel ratio
[19:53:17] <archivist> worm to wheel in the dividing head
[19:53:35] <gezr> archivist what was your worms od?
[19:53:59] <fenn> i think i'd use a small "ineffective" worm and add a brake to supply the brute force
[19:53:59] <archivist> dunno its in a machine that works and is in spec
[19:54:29] <gezr> nice
[19:55:17] <fenn> any of you gearheads used a hypocycloidal gear set before?
[19:55:27] <archivist> gezr, I had some fun with an angle dekkor finding the best one here to make dividing plates with
[19:56:08] <archivist> fenn no
[19:56:38] <gezr> angle dekkor
[19:56:58] <gezr> is ?
[19:57:46] <archivist> optical erm tries to think of the other names..
[19:58:18] <fenn> traveling microscope?
[19:58:34] <archivist> mirror on the gear and a sight sending light to mirror
[19:58:46] <gezr> hmm
[19:59:06] <archivist> colimator
[19:59:08] <gezr> so were talking minute movements
[19:59:35] <archivist> found some enormous errors though
[20:00:01] <fenn> do you set it with a sine bar or angle plates or what?
[20:00:03] <archivist> nearly a degree on some worms
[20:01:06] <archivist> no site can see a line therefore sees angle head has moved
[20:03:01] <archivist> and as the error is bunched at worm rate it a sawtooth ish form high count gear can be very wrong
[20:03:22] <fenn> what do you use as a reference to get the "true" angle?
[20:04:26] <archivist> its relative, looking for peak to peak error
[20:04:43] <archivist> there is no true angle
[20:05:39] <archivist> we found the problem here when a few got sent back!
[20:06:03] <archivist> ordinary clock quality was not enough
[20:20:14] <jepler> I hear that for slow ADC it improves noise rejection to acquire data at 60Hz (US AC power frequency). If I'm not actually monitoring the AC power frequency (with a zero-crossing detector or whatever), would it still be good to sample at a synthesized frequency that "should be" around 60Hz instead?
[20:20:22] <jepler> http://axis.unpy.net/index.cgi/01198525592
[20:20:54] <jmkasunich> jepler: it depends on the nature of the A/D
[20:21:42] <jmkasunich> if its an integrating A/D (common in DMMs, precise but slow - not likely what you have on your board) then having the integration period a multiple of 1/line freq helps reject line noise
[20:21:51] <jmkasunich> but if its a sampling type, it makes little difference
[20:22:07] <jepler> OK
[20:22:09] <jepler> I won't worry about it
[20:23:14] <jepler> (that page shows the reason I'm asking; I don't know what specific kind of ADC is in there, but I am sure it's optimized for cost and size, not precision)
[20:23:41] <jmkasunich> I read that page but didn't follow thru to the board info
[20:24:07] <jepler> it's an AVR microcontroller with 6 channels of ADC built in (in the processor package)
[20:24:19] <lerneaen_hydra> jepler, which?
[20:24:24] <lerneaen_hydra> which microcontroller?
[20:24:25] <jepler> lerneaen_hydra: atmega168
[20:24:36] <lerneaen_hydra> nice, I use the same one myself :)
[20:24:44] <jepler> bbl
[20:24:57] <jmkasunich> "it takes about 100uS to read an analog input"
[20:25:01] <jmkasunich> = not an integrating converter
[20:27:53] <Jymmmmmmm> Merry Christmas!
[20:27:58] <lerneaen_hydra> according to the datasheet it takes 13-260┬ÁS, using some type of iterative comparator
[20:28:07] <lerneaen_hydra> merry christmas Jymmmmmmm
[20:28:23] <lerneaen_hydra> http://www.atmel.com/dyn/resources/prod_documents/doc2545.pdf page 245
[20:39:26] <dmess> merry Christmas.. to all and to all a better 2008
[20:41:07] <Jymmmmmmm> here here
[20:41:21] <Jymmmmmmm> I know I'm trying for that come Dec 26th for sure
[20:41:45] <Jymmmmmmm> (interview)
[20:41:59] <fenn> jepler: you can take multiple samples and average them, see http://www.atmel.com/dyn/resources/prod_documents/doc8003.pdf
[20:43:24] <fenn> it actually has one adc and an analog mux
[20:48:37] <BigJohnT> happy holidays to all
[20:49:18] <fenn> jepler: what does the bitwise or do on this line? uint16_t v = analogRead(i) | (i << 10);
[20:50:33] <fenn> ugh nevermind. i suppose i should read the documentation first :)
[21:05:57] <jmkasunich> you can't pause a program, give an MDI command, and then resume it, can you?
[21:06:49] <skunkworks> no
[21:07:05] <gezr> fanuc's you could
[21:07:27] <gezr> it all depended onthe machine though
[21:08:05] <gezr> you guys all runing the unbuto software I guess?
[21:08:15] <anonimasu> the livecd?
[21:08:23] <gezr> well, either live or installed
[21:09:14] <skunkworks> between that and jogging while the program is paused... (it gets asked a few times)
[21:09:33] <gezr> my emc box/dev box/gateway box, hasnt been inservice for half a year, and because of that and the fact I didnt put some important packages on it, I have to plop a new drive in, boot up, and do some magic file manupliation, was thinking I should just install unbuto on it
[21:14:37] <ALS> gezr: emc1?
[21:14:54] <gezr> newest
[21:15:07] <gezr> is what ide like to have
[21:15:21] <ALS> whats on the gatway
[21:15:47] <gezr> it was just the box i used for lots of stuff
[21:16:20] <gezr> my typical linux installs, are debian sid
[21:16:32] <ALS> ubuntu is cool emc 2.2 is great
[21:18:12] <dmess> pound it on
[21:18:25] <jmkasunich> * jmkasunich is liking this O-word looping stuff
[21:18:32] <skunkworks> :)
[21:19:33] <dmess> loves it
[21:20:26] <BigJohnT> jmkasunich: what are you looping?
[21:21:10] <jmkasunich> I needed to bore a timing belt pulley out from 1/4" to 1/2"
[21:21:14] <jmkasunich> about an inch deep
[21:21:30] <jmkasunich> wrote a loop to do all the roughing passes
[21:21:43] <BigJohnT> cool I do that with my anilam
[21:24:08] <ALS> sorry to bring up paul.c again ,but the BDI was great / IMO EMC 2 is leap's above and I can't figure out the beefs he has with EMC2
[21:25:07] <anonimasu> ALS: nobody else either :)
[21:25:37] <anonimasu> loops in g-code is the most useful thing ever
[21:25:42] <anonimasu> I rely on them heavily on the heidenhain
[21:26:29] <ALS> just curious he helped me with my bdi years back and have nothing but respect for him.
[21:32:47] <fenn> postscript function arguments are hard to wrap your head around :\
[21:33:16] <fenn> they get passed in on "the stack" which every subsequent statement modifies
[21:34:58] <fenn> maybe i should redo this in cairo
[22:00:34] <owhite> would Jon Elson happen to be on the channel?
[22:00:48] <owhite> from the name list, it doesnt look like it.
[22:03:30] <jmkasunich> he doesn't often come in here
[22:03:45] <jmkasunich> he was here a day or two ago, but before that it was months
[22:03:48] <owhite> Ah.
[22:04:12] <owhite> would you happen to know, if I want to run four axes on the USC, do I have to use P8? :-)
[22:04:33] <owhite> (P8 is a connector on the USC)
[22:11:45] <ALS> owhite: you check his web site for info?
[22:12:12] <owhite> yep, I did.
[22:12:45] <owhite> its got wiring diagrams for his old configuration -- there's .ini files and hal files for the four axes configuration, but not the wiring diagram.
[22:13:36] <jepler> owhite: http://pico-systems.com/univstep_pins.html says that axis 0, 1, 2, 3 are all on connector P2. Connector P8 is apparently for logic-level outputs; I think that this means they can only be updated at the slow rate of the emc "servo thread", and are not suitable for generating high speed step waveforms.
[22:15:47] <owhite> jepler. right. but what confused me is what that page reported for the encoders on p4.
[22:16:36] <owhite> it lists several unused pins, and I was wondering if they were genuinely unused, or I could use them for connecting to encoder for axis A.
[22:20:34] <jepler> owhite: The same document says that P3 has A, B, Z inputs for axis 0, 1, 2, 3
[22:21:28] <jepler> what I don't see is limit/home switch inputs for a 4th axis -- since that's all done by HAL configuration I think you're actually free to use the pins in any way you like
[22:22:17] <jepler> e.g., linksp Xminlim <= ppmc.0.din.01.in
[22:22:48] <jepler> but you have to understand, I'm just guessing from what I read on that page .. I have no experience with usc
[22:22:56] <owhite> got it.
[22:23:39] <owhite> I was thinking of digging into the hal component to see if it documents where the ppmc.0.dim.01.in corresponds to.
[22:24:05] <owhite> any suggestions where that might be in the code?
[22:26:24] <owhite> bad day for shopping -- mesa electronics arent answering their phone :-)
[22:27:08] <jepler> the source code isn't too informative -- search down to "static void read_digins" in http://cvs.linuxcnc.org/cvs/emc2/src/hal/drivers/hal_ppmc.c?rev=1.52 but all you find is that they're numbered 0..15
[22:27:37] <jepler> but I would assume that matches up with the items labeled "bit 0 input" through "bit 15 input" in the pins.html page
[22:27:52] <owhite> jepler: hey thanks for digging that up.
[22:28:48] <owhite> looks good but I'll confirm with Jon by email.
[22:42:04] <ALS> Merry Chrismas All
[23:35:13] <Gamma-X> what would u guys do to clean a completely cover in oil wood chips, metal chips, and coolant machine lol.?
[23:35:48] <Jymmmmmmm> steam cleaner / pressure washer + degreaser
[23:36:59] <Jymmmmmmm> or... just toss on some solvent and gently wipe away and try not to get it where you want the grease to be
[23:37:44] <Gamma-X> ok
[23:37:53] <Gamma-X> what kind of solvent would u use?
[23:37:59] <Gamma-X> i started usin citrus cleaner
[23:38:09] <Gamma-X> but theres so many chips on it my hands are cut up lol
[23:38:40] <Jymmmmmmm> Got Gloves - dishwashing ones
[23:38:55] <Jymmmmmmm> use a chip brush, not your fingers =)
[23:39:51] <Jymmmmmmm> solvent could be gasoline, kerosen, even WD-40 if you must, just about anything to break up the grease
[23:39:59] <Gamma-X> i could prolly use a magnet to get the metal chips....
[23:40:22] <Gamma-X> spray solvent on it and suck it all up with a shop vac? sound acceptable?
[23:40:43] <Jymmmmmmm> solvent and electric motor dont usually go so well =)
[23:41:31] <Gamma-X> hmm
[23:41:36] <Gamma-X> citrus cleaner... alot of it lol
[23:41:49] <Jymmmmmmm> electric motor = sparks, solvent fumes ignite from sparks, Got marshmellows?
[23:42:18] <Jymmmmmmm> let the cirtrus clena soak for a while.
[23:42:43] <Gamma-X> lol ok
[23:42:55] <Jymmmmmmm> $0.99 chip/paint brush does pretty good
[23:43:00] <Gamma-X> what about the circutry in the control box?
[23:43:08] <Gamma-X> alcohol?
[23:43:12] <Jymmmmmmm> Um, dont' degrease it =)
[23:43:43] <Jymmmmmmm> Well, I use denatured alcohol on electronic circuitry
[23:43:50] <Gamma-X> just spray it on?
[23:43:53] <Jymmmmmmm> not rubbing/household alcohol
[23:43:59] <Gamma-X> i realyt dont wanna take all the boards out.
[23:44:12] <Jymmmmmmm> is it that bad that you need to?
[23:44:22] <Jymmmmmmm> is it that bad that you need to clean the pcb;s?
[23:44:47] <Gamma-X> ehhh
[23:44:49] <Gamma-X> dust and such
[23:45:02] <Gamma-X> some shavings in the lower area of the box
[23:45:06] <Jymmmmmmm> can of air, and if needed can of contact cleaner
[23:45:25] <Jymmmmmmm> og course power completely disconnected.
[23:46:01] <Gamma-X> yeah yeah. i havnt even check the power box on it to see how bad it is
[23:46:08] <Gamma-X> but hey i only payed 2 grand for a ycm-18! lol
[23:46:15] <Gamma-X> http://cgi.ebay.com/VERY-CLEAN-PROTOTRAK-SUPERMAX-YCM-40-CNC-VERTICAL-MILL_W0QQitemZ150196827733QQihZ005QQcategoryZ12584QQssPageNameZWDVWQQrdZ1QQcmdZViewItem
[23:46:28] <Jymmmmmmm> shop vac as much as you can first, then go in with solvents/liquids as needed.
[23:47:26] <a-l-p-h-a> used cnc machines are so cheap these days
[23:48:03] <BigJohnT> twas the night before Christmas and all though the house not a creature was stirring not even a mouse... When all of a sudden with a great blast of air out popped Gamma-X from behind his mill all covered in soot and old chips. He gave out a laugh and smiled at his clean machine...
[23:59:09] <Gamma-X> hahaha
[23:59:11] <Gamma-X> BigJohnT I LOVE U!
[23:59:15] <Gamma-X> Ur my home boy.
[23:59:20] <BigJohnT> you finally read it
[23:59:27] <Gamma-X> read what
[23:59:34] <Gamma-X> ohhh that lol i was eatin dinner
[23:59:37] <BigJohnT> the xmas story
[23:59:55] <Gamma-X> lol yeah
[23:59:59] <Gamma-X> i didnt take pics yet