#emc | Logs for 2009-12-30

Back
[00:05:26] <jimbo> Picked up my Servo motor from the shop today. He said that it looked like it was skipping some of the contacts on the commutator. Hoping for the best.
[00:06:33] <Valen> as in it is or it was?
[00:07:53] <jimbo> It was. it is no longer.....
[00:08:53] <Valen> ahh thats good then
[00:10:52] <jimbo> Yes it is. Now onto soldering all the crimp ends. Install and test..... It is now snowing so I have lots of time to play.
[00:12:17] <Valen> funky
[00:20:25] <daniel> daniel is now known as Srpski
[01:04:34] <tom3p> i just saw a kindle running ubuntu ( it had TTS, i hadta look )
[01:04:55] <eric_unterhausen> that would be cool
[01:05:36] <eric_unterhausen> I'm much more reluctant to buy one after amazon unilaterally deleted a book off of their user's machines
[01:06:24] <tom3p> http://gadgets.boingboing.net/2009/09/02/photo-and-descriptio.html to expensive for me, i use an old ipaq 3650 with tts, tho i never put opie on it ( qt linux )
[01:07:05] <tom3p> ahah amazon has that drms disease
[01:07:39] <eric_unterhausen> yeah, but deleting without asking is going over the line
[01:08:20] <tom3p> yes, i hadnt heard of it, but thats trespass
[01:09:35] <jimbo> OK. When Tuning a PID for velocity mode I get a spike at the beginning of the movement of the joint and then it settles down. What parameter would effect that spike?
[01:10:00] <tom3p> acceleration
[01:10:19] <jimbo> Oh Yes the motor seems to be running fine now.... YEA
[01:10:31] <tom3p> and in physical realm stiction
[01:10:38] <tom3p> cool
[01:13:57] <jimbo> So lower acceleration would mean a lower spike if stiction is low..
[01:16:14] <Valen> you may want more P to get it started and then offset that with some D
[01:20:56] <Valen> if its stiction then acceleration probably wont stop the jump, the force needs to build up until it overcomes that
[01:21:07] <Valen> you then just need to minimise the overshoot
[01:27:33] <lilalinux> lilalinux is now known as lilalinux_away
[01:36:54] <jimbo> Valen: thank you I will look
[01:37:26] <Valen> 's only a wild ass guess so pinch of salt there ;->
[01:39:36] <jimbo> When you change acceleration the spike reduces a bit but it also affects deceleration.
[01:44:31] <PCW> You might try a little FF2 (accelleration Feed Forward), the velocity loop in the drive should be handling this but its worth a try
[01:49:49] <toastydeath> most box way machine tools overcome stiction by completely ignoring it
[01:50:48] <jimbo> Meaning? dovetail machines?
[01:51:07] <jimbo> Bridgeport
[01:51:10] <toastydeath> anything with a plain bearing, no rolling elements or hydrostatics
[01:51:21] <toastydeath> yeah, a bridgeport qualifies.
[01:51:53] <jimbo> So the choice of way lube is important?
[01:52:11] <Valen> yeah
[01:52:14] <toastydeath> not incredibly, bridgeports use a vactra 2 or equivalent oil
[01:52:47] <toastydeath> any waylube in that weight range will work
[01:53:08] <toastydeath> no matter what you use, you are going to have stiction on a plain way machine
[01:54:03] <jimbo> So with the Vactra 2 you are just trying to reduce it.
[01:54:08] <toastydeath> yep
[01:54:22] <Valen> way lube is probably more important than non way lube, though we are using hydraulic oil and dont seem to have an issue that I've noticed (yet)
[01:54:34] <toastydeath> waylube is very tacky so it tends to float the axis longer than other oils
[01:54:50] <PCW> You can probably determine if its stiction or a tuning problem by starting to accell from a slow creep instead of a stop
[01:55:16] <toastydeath> but anytime you actually stop an axis to go do something with the other axis, it'll fall back down and stick
[01:55:37] <toastydeath> all plain way machines do this, it cannot be avoided
[01:55:51] <PCW> Also if the error is tha same for accel and decel it mor liekely tuning than stiction
[01:57:23] <jimbo> The acceleration did help my tunning is much tighter. MNI_FERROR .001 FERROR .005 This may be good enough.....
[01:58:00] <Valen> thats pretty tight jimbo
[01:58:53] <jimbo> So I will leave well enough alone and move to the next one........
[01:59:00] <toastydeath> if you are using waylube, the axis will not sit down during most contouring work and you'll still get good accuracy there
[01:59:18] <toastydeath> so don't worry overly much if it's actually stiction at the beginning of movement
[02:01:16] <jimbo> Next on the list is I have a 110 oiler that I have been plugging in just to test how does one set this up so EMC will automatically control the application of the oil?
[02:01:34] <toastydeath> does the oiler have a timer?
[02:01:43] <Valen> how do people stop the waylube from getting mixed with all the other oils and such floating around on a wet mill?
[02:01:50] <Valen> good bellows or some such?
[02:02:02] <toastydeath> no, we just let it fall in
[02:02:32] <toastydeath> and it gets vacuumed out when the coolant is changed
[02:02:44] <jimbo> No just a dumb motor in think. (Biju)
[02:02:45] <toastydeath> they make skimmers, if you don't change coolant often
[02:02:57] <toastydeath> bijur
[02:03:00] <toastydeath> v. good oilers
[02:03:08] <jimbo> Yes
[02:03:22] <toastydeath> but i can't answer, sorry
[02:03:48] <jimbo> OK thought I would ask......
[02:03:48] <Valen> i mean if the way is there and there is cutting fluid splashing all over it and thats carrying oil from the ballscrews or whatnot
[02:04:15] <toastydeath> Valen: yes, the coolant and waylube get mixed up
[02:04:25] <Valen> doesnt bother the ways?
[02:04:28] <toastydeath> the waylube winds up in a thin layer on top of the coolant in the tank, and is called tramp oil
[02:04:46] <toastydeath> most cncs have bellows or covers of some kind
[02:04:50] <toastydeath> mostly to keep chips out
[02:04:53] <Valen> yeah thats what I was wondering
[02:05:00] <Valen> might have to add some of those
[02:05:53] <toastydeath> they're way useful.
[02:07:04] <toastydeath> though sometimes machines kick them off spectacularly for no reason
[02:07:19] <toastydeath> they get jammed or otherwise lock up
[02:07:31] <toastydeath> then they go flying and bounce around inside the enclosure
[02:07:33] <eric_unterhausen> I wish I could get the wipers for mine
[02:08:28] <toastydeath> ownt
[02:08:49] <Valen> what sorts of depths, feeds and speeds would you guys run a 6mm (1/4?) magic coated solid carbide end mill at in stainless?
[02:09:05] <Valen> just want to validate my calculations with some real world experience
[02:09:12] <Valen> oh yeah spindle is limited to 1600 RPM
[02:09:27] <toastydeath> .15"
[02:09:36] <toastydeath> full slotting but slowly
[02:09:41] <toastydeath> only a couple ipm
[02:10:03] <toastydeath> you can usually cut about the depth of the cutter
[02:10:14] <toastydeath> without much issue, provided you can get the chips completely out (like we discussed before)
[02:10:22] <Valen> wow thats way deeper than I was thinking
[02:10:33] <toastydeath> obviously any tool overhang or long cutters will reduce your ability to do that
[02:10:38] <Valen> yeah
[02:10:43] <toastydeath> and as the cutter gets bigger the machine will start to limit things
[02:11:16] <Valen> http://www.youtube.com/watch?v=2KNit__LJE4&feature=player_embedded nice
[02:12:28] <Valen> I was under the impression that you needed to run the tools at about the cut per tooth rate or run the risk of rubbing?
[02:12:55] <toastydeath> the minimum chip load is very small for tiny tools
[02:13:05] <toastydeath> they're fairly sharp
[02:13:21] <Valen> I havent actually seen somebody spec a minimum chip load yet
[02:13:41] <toastydeath> .0005" per tooth or so
[02:16:19] <toastydeath> with small cutters you have to be more worried about clogging the gullet than you do breaking them from rubbing
[02:17:02] <Valen> shouldn't be a problem its all pretty open
[02:17:03] <toastydeath> clogging the gullet/recutting chips
[02:17:12] <toastydeath> well clogging the gullet is a function of feed per tooth
[02:17:15] <toastydeath> not really chip evac
[02:17:35] <toastydeath> if the gullet only has space for .002" of chip, and you try to stuff .003" in there
[02:17:38] <Valen> I havent heard that term before
[02:17:47] <toastydeath> cutter will break
[02:17:47] <Valen> what is the gullet?
[02:17:53] <toastydeath> the gullet is the space between teeth
[02:18:04] <toastydeath> the valley
[02:18:20] <Valen> the flute part?
[02:18:23] <toastydeath> yeah
[02:18:33] <toastydeath> it's another name for flute, sort of
[02:18:33] <Valen> ahh no that should be ok
[02:19:33] <Valen> oh the bit on the back side of the cutter before it joins the flute propper?
[02:19:50] <Valen> back side of the cutting edge?
[02:20:20] <toastydeath> it's the empty space
[02:20:44] <toastydeath> the form of the empty space is called the gullet.
[02:21:04] <toastydeath> there are other names for the actual parts of the physical cutter in that area, i forget what they are
[02:21:13] <Valen> there are so many names lol
[02:22:07] <toastydeath> most of them are unimportant
[02:24:15] <toastydeath> mostly because nobody else knows either so what's the point of knowing if nobody else has a clue what you're talking about
[02:25:19] <Valen> lol
[03:32:30] <tom3p> i was looking for some info on cutter geometry, this is is nice from aachen http://www.hssforum.com/Aachen/12_Lux.pdf
[03:33:18] <tom3p> doh! no THIS is nice http://www.hssforum.com/
[03:39:58] <eric_unterhausen> I wish I had learned all that stuff in my industrial engineering class
[04:52:13] <tom3p> there is no standard mfin sfin in emc, might be handy way for all mcodes/speed codes to tell control they are done
[04:52:33] <tom3p> afik
[05:15:08] <Dave911> logger_emc:bookmark
[05:15:08] <Dave911> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2009-12-30.txt
[05:32:36] <tom3p> argh! what do you get with man axis? info on Axis gui. how do you get info on axis.N.blah? man motion man! use the unix concept of capitalization significance
[05:42:41] <tom3p> i want to position a joint with emc2 using a stepgen of type 'p'.
[05:42:41] <tom3p> then i want to begin using a type 'v' stepgen to move it.
[05:42:41] <tom3p> assuming that can be done, i think axis.N.motor-pos-cmd will hold the position last accomplished by the 'p' stepgen.
[05:42:41] <tom3p> i'd like to know if axis.N.motor-pos-fb will track the actual position while in 'v' mode.
[05:42:41] <tom3p> Then, before returning to 'p' mode, i'd only need to insure motor-pos-fb equaled motor-pos-cmd.
[07:33:20] <cnc_valen> question, is there a way to get axis to display a "program time to go" IE when does it think thecurrent program will end?
[09:10:04] <Stinger> are there any other good machining/cnc irc channels around?
[09:13:30] <Stinger> I am looking at building a basic cnc router just to begin with, will ABEC 1 be good enough, or is it just going to be /really/ bad
[09:13:52] <eric_unterhausen> have you looked at cnczone?
[09:14:02] <eric_unterhausen> there is a lot of collected experience there
[09:14:05] <Stinger> not in regards to this, other things though
[09:28:33] <eric_unterhausen> one nice thing about cnczone is that you can see the full range of parts; from the very cheapest to the most expensive
[09:28:49] <eric_unterhausen> if you can use drawer slides, I think abec 1 bearings will probably work fine
[10:33:42] <lilalinux_away> lilalinux_away is now known as lilalinux
[10:48:14] <Valen> http://www.youtube.com/watch?v=b2-Kdud7eiA&feature=response_watch
[11:36:26] <MattyMatt> drawer slides are self adjusting :) the springiness of the outer race takes up any slack, and it's still stiff enough for woodwork afaics
[11:37:22] <MattyMatt> dunno how much woodwork I can do before they go sloppy tho
[11:39:08] <MattyMatt> Valen nice vid :)
[11:39:26] <Valen> not mine, but way cool though
[11:39:39] <MattyMatt> I like the commentary "It's like a lathe, isn't it?" that's quite perceptive for a girl
[11:39:52] <Valen> she seems actually interested
[11:40:00] <Valen> which is shocking in tits own right
[11:40:08] <MattyMatt> using the spindle as a toolholder is a cool idea
[11:40:35] <Valen> yeah
[11:40:41] <Valen> I wonder about the rotary table
[11:40:46] <Valen> what its resolution is like
[11:40:46] <MattyMatt> you just got freudslapped :) tits own right
[11:41:06] <Valen> ?
[11:42:16] <lilalinux> lilalinux is now known as lilalinux_away
[11:43:37] <lilalinux_away> lilalinux_away is now known as lilalinux
[11:44:19] <archivist_emc> I note that there is no room for a parting tool
[11:44:50] <Valen> just needs to mount a rotary tool changer on the head and she'll be right
[11:45:27] <archivist_emc> and upgrade to emc
[11:45:34] <Valen> well obviouslt
[11:45:38] <Valen> obviously
[11:46:05] <Valen> I think it might suck for doing "important" stuff on the rotary axis though, I cant see it getting much resolution or stability like that
[11:47:18] <archivist_emc> resolution should be as good as its encoder
[11:48:00] <Valen> yeah but its not going to hold it too well while its cutting I wouldn't imagine?
[11:50:03] <Valen> would need to be a pretty fancy encoder (or have 2) to read out at that speed I'd imagine
[11:52:26] <archivist_emc> I just watched the hard tapping vid, wrong type of tap for regular hard tapping
[13:01:17] <Valen> if I give you the money, would one of you buy this for me and drive it on over? http://www.governmentauctions.org/2009/04/grab-your-crew-and-get-ready-to-hit.asp
[13:02:22] <Valen> ;->
[13:02:25] <archivist_emc> that would be fun
[13:02:55] <archivist_emc> it would need larger fuel tanks
[13:03:05] <Valen> heh, be worth it, sell the house and live on it
[13:03:37] <Valen> you could island hop your way over, wouldn't be *too* bad
[13:04:05] <Valen> 1400 gallons should get you a fair way I'd imagine
[13:10:06] <archivist_emc> I imagine the mpg not being as good as a lorry, and they only do 8 ish
[13:10:17] <jthornton> 6V-53's are pretty stingy with fuel ie. it won't go fast
[13:11:04] <Valen> your right, this is more what I'm after I think http://au.apolloduck.com/feature.phtml?id=93811
[13:11:33] <Valen> just needs a car ramp, and garage with provision for a helipad ;->
[13:11:40] <Valen> (for a micro heli)
[13:14:46] <Valen> that is actually cheaper than a house towards the city here
[14:26:21] <MarkusBec> MarkusBec is now known as MarkusBec_away
[16:15:53] <lilalinux> lilalinux is now known as lilalinux_away
[16:18:00] <lilalinux_away> lilalinux_away is now known as lilalinux
[18:07:29] <lilalinux> lilalinux is now known as lilalinux_away
[19:35:55] <lilalinux_away> lilalinux_away is now known as lilalinux
[20:14:01] <skunkworks_> heh - the latest version of thunderbird has tabbed email :)
[20:33:38] <jimbo> It looks like my Bijur oiler outputs 1 pump per hour of run time. Would you enable it at the same time as the servo system? Or when the computer system is up?
[20:36:27] <cradek> when out of estop?
[20:36:37] <cradek> I like to be able to leave the computer on
[20:38:45] <jimbo> OK makes sense.....
[20:41:15] <jimbo> I am changing my oil to Vacuoline 1409 which from what I understand is the old formula for vectra 2. I will see if this helps the Y axis spike.
[20:50:05] <cradek> did you check all your metering nozzles?
[20:52:09] <jimbo> So I get to tear the table off again...... goody:)
[21:36:49] <MarkusBec_away> MarkusBec_away is now known as MarkusBec
[22:17:21] <MarkusBec> MarkusBec is now known as MarkusBec_away
[23:12:28] <GonMD> hey guys. im running into a slight problem. not sure if its with EMC, or my machine im running it on. twice in a row now, cutting a job, just a simple black and white photo into some MDF, using emc's built in code generator for images, it will stop 3/4 of the way through for no apparant reason. at first i thought the screensaver might be kicking in and messing with it, but i disabled it and tried a second time, and same thing. not exactly the same spot, but cl
[23:12:28] <GonMD> enought to 3/4 of the way through. thoughts ?
[23:14:35] <micges> can you pastebin.ca gcode?
[23:14:54] <micges> paste it at www.pastebin.ca
[23:14:55] <jimbo> On the mi520 card how fast are the hardware counters. Can they count around 600,000 per second.
[23:15:21] <micges> GonMD: and can you write what system you have and what emc version you're using?
[23:15:53] <GonMD> yeah, gimme a minute.
[23:18:19] <micges> brb
[23:21:49] <PCW> 5I20 encoder counters will count to roughly 4M counts/second with filters on and maybe 16M/sec with filters off
[23:21:50] <PCW> (HostMot2, with default filter rates)
[23:25:40] <GonMD> micges: http://gonmd.com/uploads/files/logo.ngc <gcode http://tiny.cc/FtWvS <hardware
[23:29:40] <GonMD> running EMC 2.3.4
[23:34:25] <frallzor> * frallzor is cuddling with an old gps
[23:51:05] <micges> GonMD: I don't see your problem on 2.3.4 or newest development snapshot
[23:51:25] <micges> what is your run procedure?
[23:51:52] <micges> something other that yust press start?
[23:53:16] <GonMD> after i touch off all axis, i hit run single line till the Z pulls up, then walk over and switch on the router, then un pause to let it run the rest of the program
[23:54:12] <micges> oh I see
[23:54:57] <micges> this is known problem (pause then start will just run 1000 lines)
[23:55:19] <GonMD> ah. very interesting.
[23:57:19] <micges> emc is buffering program to avoid errors, but in some cases it's just stops program
[23:58:44] <micges> (to be specific in that one case)
[23:59:21] <GonMD> understandable. is there a workaround that you know of, or am i gonna have to get really fast at hitting go and then hitting the power switch on my router :p