Back
[00:02:22] <anonimasu> hm
[00:02:34] <anonimasu> i just ran into the limitations of solidworks.
[00:02:41] <anonimasu> limitations/lack of features
[00:02:45] <anonimasu> how stupid.
[00:03:15] <alex_joni> hmm.. I reached the end of the internet one day
[00:03:37] <anonimasu> hehe
[00:03:38] <anonimasu> :9
[00:03:44] <anonimasu> http://www.io23.net/an0n/spjall.png
[00:03:57] <anonimasu> I were going to draw a seal there.. a groove..
[00:04:32] <anonimasu> hm that wouldnt work either..
[00:04:36] <anonimasu> very odd.
[00:04:47] <alex_joni> heading for bed now :)
[00:05:01] <anonimasu> alex_joni: how was the last boss?
[00:05:03] <anonimasu> g'nite
[00:05:08] <alex_joni> last boss?
[00:05:24] <anonimasu> yeah, at the end of all good videogames there's a boss..
[00:05:32] <anonimasu> that you have to defeat to clear the game..
[00:05:34] <alex_joni> oh, pretty simple
[00:05:39] <anonimasu> hm, that's a shame
[00:05:46] <alex_joni> left-left-kick-down-down-kick
[00:05:53] <alex_joni> fatality.. and he was out
[00:05:56] <alex_joni> :D
[00:05:58] <anonimasu> ^_^
[00:39:06] <CIA-12> 03jmkasunich 07HEAD * 10documents/lyx/emc2/hal_tools.lyx: added a stub for halshow documentation
[00:57:29] <davidf> hey.
[00:57:42] <jmkasunich> hey yourself ;-)
[00:57:56] <davidf> :)
[00:58:39] <davidf> hey, my mill is working perfectly at first brush, 70+ in/min. on all axes. But...
[00:59:05] <jmkasunich> there's always a "but"
[00:59:24] <davidf> :( At seeminly random times, x, y, or z mess up.
[00:59:42] <jmkasunich> define "mess up"
[00:59:49] <davidf> I tried lower acceleration till it is almost rubbery.
[01:00:27] <davidf> Thought that fixed it, but then the z axis tried to go through the table. Luckily I only had a pen in the spindle.
[01:00:46] <jmkasunich> define "mess up"
[01:00:56] <cradek> define "mess up"
[01:01:02] <davidf> But, I just realized while typing that maybe that happened due to too slow decelleration.
[01:01:24] <cradek> low accel can't make you overshoot
[01:01:39] <davidf> It is actually the opposite problem I'd been having, because it went too far rather than locking up momentarilly.
[01:02:03] <davidf> Yes it would seem that way. But,
[01:02:37] <jmkasunich> I still have no idea what you mean when you say "mess up"
[01:02:50] <jmkasunich> do you mean it looses steps?
[01:02:54] <jmkasunich> loses
[01:02:56] <davidf> Yes.
[01:03:20] <davidf> But after lowering the acc., it went too far on Z.
[01:03:23] <cradek> a stall on an 'up' will look like an overshoot on a subsequent 'down'
[01:03:28] <jmkasunich> so you go from 0 to 1, then back to 0, and its not where it started?
[01:03:33] <davidf> (About a quarter inch.
[01:03:48] <davidf> Ahah. Yes.
[01:04:02] <cradek> it's easy to get stalls going up in Z because of the weight
[01:04:10] <davidf> Oh, you're so smart. :)
[01:04:24] <cradek> I like to set my steppers so that I can push down considerably on the spindle and still jog up at full speed
[01:04:27] <jmkasunich> (depends on the type of mill - a quill isn't so heavy, but if you're lifting an entire head, thats harder
[01:04:32] <cradek> I'd rather have them reliable than fast
[01:04:46] <davidf> Especially Z.
[01:04:46] <cradek> it may be that you just don't have them set conservatively enough
[01:04:49] <cradek> right
[01:05:06] <davidf> But, I am able to jog it at 73 in/min.
[01:05:12] <cradek> I like to be unable to stop a Z lift with a fingers' pressure
[01:05:20] <davidf> And I was only using 30 in/min at the time.
[01:05:20] <jmkasunich> with no lost steps?
[01:05:39] <davidf> Well, That's a good question on Z isnt it?
[01:05:41] <davidf> :)
[01:05:42] <cradek> when you're jogging Z up at 73 in/min, push down on it, see how much power headroom you have
[01:05:57] <cradek> if you can easily stop it, it's NOT enough in my opinion
[01:06:12] <cradek> reducing speed only 10% makes a huge difference
[01:06:40] <jmkasunich> set Z a known height above the table, jog up (use a 1" incremental jog, for instance)
[01:06:53] <jmkasunich> measure with a ruler to see if it went 1"
[01:07:01] <davidf> I can put at least 10 lbs force on it without noticing anything, at 73 in/min.
[01:07:07] <jmkasunich> then jog down 1" and see if it went back to where it started
[01:07:15] <davidf> ok.
[01:07:23] <cradek> hmm ok
[01:08:00] <cradek> then repeat jmk's test 100 times :-)
[01:08:17] <cradek> if you get a stall going up an inch, you will sure notice it right away
[01:09:49] <jmkasunich> one inch might not be enough
[01:09:53] <jmkasunich> * jmkasunich does math
[01:10:00] <jmkasunich> 1.22 inches/sec max speed
[01:10:05] <jmkasunich> 12 inches/sec^2 accel
[01:10:06] <davidf> right. But I'll use a dial guage. Hang on.
[01:10:12] <cradek> yes you definitely want to get up to cruise
[01:10:13] <jmkasunich> so about 0.1 seconds to reach top speed
[01:10:16] <davidf> gauge
[01:10:21] <davidf> boy.
[01:10:25] <jmkasunich> and 0.1 seconds to stop again
[01:10:35] <anonimasu> that seems fast
[01:10:56] <cradek> I bet it's fine if he's got relatively little mass
[01:11:07] <cradek> I run mine at 20 ips2 for cutting PCBs
[01:11:21] <jmkasunich> avg speed during accel/decel = 1.22 / 2 = 0.6 inches/sec
[01:11:22] <cradek> it's very light, turns on a dime
[01:11:30] <jmkasunich> so 0.06" travel during accel and decel
[01:11:45] <cradek> ok so an inch is plenty
[01:11:46] <anonimasu> hm, so what's the speed limit for the planner now?
[01:11:48] <anonimasu> *curious*
[01:11:53] <jmkasunich> so 1" is fine... you get to full speed in 0.06"
[01:11:55] <cradek> what's a speed limit?
[01:12:00] <anonimasu> is there even one?
[01:12:09] <cradek> what do you mean? top velocity? no
[01:12:09] <anonimasu> the queue starvation issues that were before
[01:12:19] <cradek> oh different segments per second?
[01:12:21] <anonimasu> yeah
[01:12:49] <cradek> up to one every traj cycle
[01:13:09] <anonimasu> hm, with the adequate machine can you do hsm with it?
[01:13:10] <cradek> so 100 I think?
[01:13:20] <anonimasu> just speculating
[01:13:26] <cradek> depends how you define HSM and what your gcode is
[01:13:27] <anonimasu> I've got no idea about how many segments that means..
[01:13:27] <jmkasunich> depends on how long the segments are
[01:13:40] <anonimasu> hm short for contouring say a 3d thing..
[01:13:41] <jmkasunich> since we don't have multi-segment blends, I'
[01:13:49] <jmkasunich> I'd hesitate to call it HSM
[01:13:53] <anonimasu> hm ok
[01:13:55] <cradek> yeah
[01:14:00] <cradek> really depends on the gcode
[01:14:06] <anonimasu> well, I guess we will be there someday
[01:14:08] <cradek> jmkasunich: we sort of do have multi segment blends now
[01:14:29] <cradek> jmkasunich: we throw out points (join segments) when they make no difference considering the tolerance
[01:14:29] <anonimasu> I
[01:14:41] <anonimasu> I've got too old iron to machine that fast :9
[01:14:48] <cradek> depends what you call blending :-)
[01:14:52] <anonimasu> lol
[01:14:53] <cradek> * cradek waves his hands
[01:16:18] <davidf> Boy, guess I stimulated quite a conversation. Haven't read it yet... The Z is right on the money after 5 or 6 trials at 73 in/min.
[01:16:38] <davidf> +/- 0.0001
[01:16:38] <jmkasunich> nasty test case: a 1" move, followed by 100 moves of 0.0001 each, all parallel, followed by a 90 degree turn and a 1" move
[01:16:53] <jmkasunich> ok, so you're not losing steps during jog
[01:17:04] <cradek> jmkasunich: it'll blend all the parallels together
[01:17:29] <cradek> jmkasunich: in nonrealtime, then send just a couple segments to the planner
[01:17:29] <davidf> Don't think so. It isn't really acting like that so much as random.
[01:17:30] <jmkasunich> davidf: are you back to your original 12 inches/sec^2 accel limit? or using something else?
[01:17:48] <anonimasu> Hm, does this happen with hot/cold motors?
[01:18:09] <cradek> davidf: can you try it with the spindle on in case you've got an electrical noise problem?
[01:18:16] <davidf> Max acceleration 2.0
[01:18:17] <cradek> (not recommended with a dial indicator in the collet)
[01:18:36] <jmkasunich> first answer a question: was the spindle on when you noticed the problem before?
[01:18:52] <davidf> I was running with the spindle off when it happemed. Good idea, though.
[01:19:07] <cradek> darn
[01:19:19] <jmkasunich> first things first, that 2.0 accel is not going to help troubleshoot this
[01:19:30] <davidf> Happens either way, at random times. Sometimes not at all.
[01:19:47] <cradek> these kind of problems suck
[01:19:48] <davidf> Even at slow speeds.
[01:19:56] <jmkasunich> your original numbers: TRAJ accel, 12, AXIS accel, 10, STEPGEN accel 11
[01:19:58] <davidf> yes. Bigtime.
[01:20:14] <jmkasunich> personally I'd set TRAJ same as AXIS
[01:20:21] <jmkasunich> but EMC should use the lower of the two
[01:20:37] <cradek> no, traj is tooltip accel, depends on the move
[01:20:37] <jmkasunich> STEPGEN _must_ be higher than AXIS (and it is)
[01:20:50] <cradek> making traj sqrt(3)*axis makes perfect sense
[01:21:07] <jmkasunich> sqrt(2) maybe
[01:21:18] <cradek> 0,0,0 -> 1,1,1 = sqrt(3)
[01:21:24] <jmkasunich> ;-P
[01:21:26] <cradek> * cradek pokes jmkasunich
[01:21:46] <davidf> traj acc 12.0 axis acc 2.0 Stepgen acc 2.1
[01:22:13] <jmkasunich> and was it misbehaving with those settings?
[01:22:57] <davidf> It only does it occasionally. Seems like it would be a lot more frequent and predictable if it was a parameter issue.
[01:23:00] <cradek> when you were pushing down on Z while jogging up, were you making jogs long enough to let it get up to cruise speed? you can tell by the sound mostly
[01:23:16] <davidf> Yes.
[01:23:22] <davidf> And yes.
[01:23:31] <cradek> ok
[01:23:50] <jmkasunich> I'm suffering from lack of information
[01:23:54] <cradek> I still doubt your hardware, but we ran the latency test successfully for some time
[01:24:22] <cradek> the only overrun was when you pushed the special Break key, right?
[01:24:24] <davidf> I read something about some timer or nanosecond thing, where better performance on trajectories could be had by cutting it in half?
[01:24:26] <jmkasunich> how "occasional"? can you make it happen in a reasonable amount of time (a couple minutes) or is it every few hours?
[01:24:49] <jmkasunich> don't change _anything_ yet!
[01:24:56] <davidf> Yes re overrun.
[01:25:11] <cradek> sorry we're bombarding you
[01:25:20] <davidf> once in 2 - 5 minutes or so.
[01:25:28] <davidf> duration maybe 1 second.
[01:25:34] <jmkasunich> duration?
[01:25:36] <jmkasunich> what is duration?
[01:25:46] <jmkasunich> lost steps last forever
[01:26:08] <davidf> stall duration. Then it resumes, but offset by several thousand steps. Like maybe 1/4 inch.
[01:26:43] <davidf> like, I'll get a 1 inch vertical like instead of a 45.
[01:26:48] <cradek> a stall typically lasts until the end of the "move"
[01:26:53] <davidf> then resume the 45 deg line.
[01:27:08] <jmkasunich> what do you have to do to make the problem happen?
[01:27:19] <jmkasunich> jogs? mdi? some part program?
[01:27:49] <davidf> OK, that's entirely possible. Actually it was cutting or drawing) 1.5 inch tall script letters made of short lines.
[01:28:11] <jmkasunich> and only one axis stopped?
[01:28:40] <jmkasunich> (45 becoming vertical implies that one axis stopped and the other(s) kept going)(
[01:28:51] <davidf> It does seem to happen more at the bottom of the letter e, where it is almost horizontal.
[01:29:02] <davidf> Yes. You got that right jmkasunich
[01:29:12] <davidf> That's what is happening.
[01:29:22] <jmkasunich> do you have any backlash setting? (your pastebin ini has it at zero, but is it still zero?)
[01:29:31] <cradek> at the bottom of the e does x or y stall?
[01:29:42] <davidf> But it happens on all three axes, but so far only on one at a time at random spots.
[01:29:43] <cradek> since x is the one moving fast, y is slow
[01:31:38] <jmkasunich> hmm, you are really pushing the pulse rate... if the pastebin ini file is still what you have, you're running 50uS base period, which means an absolute maximum of 10,000 steps/second
[01:31:44] <cradek> ooh good question (backlash)
[01:32:00] <jmkasunich> 1.22 inches/sec at 8000 counts/inch = 9,760
[01:32:06] <davidf> Sorry, I had a phone call. Lets see, what did I miss?
[01:32:12] <cradek> jmkasunich: ouch
[01:32:49] <davidf> usually x. That's why I originally thought it might be a low freq. resonance.
[01:32:54] <cradek> about any machine can do 20-25 usec base period
[01:32:54] <jmkasunich> we really need to get the basic forumlas into a wiki page
[01:33:50] <davidf> So can you put that in English?
[01:33:52] <davidf> :)
[01:33:55] <cradek> heh
[01:34:14] <cradek> I suggest you change BASE_PERIOD in your ini from 50000 to 25000
[01:34:15] <jmkasunich> first, answer the backlash question? did you change it, or is it still zero?
[01:34:21] <davidf> Lets see, You think I'm pushing for too fast a pulse rate?
[01:34:33] <davidf> Still 0.
[01:34:37] <jmkasunich> good
[01:34:42] <davidf> And it very nearly is.
[01:34:48] <jmkasunich> so backlash is not an issue, lets move on
[01:34:56] <davidf> Specs are 0.005 & I believe it.
[01:35:12] <jmkasunich> pulse rate: 73 ipm is marginal at best, for that particular BASE_PERIOD
[01:35:16] <jmkasunich> how fast is your PC?
[01:35:34] <davidf> So if the pulse rate is too high, and a process grabs the cpu?
[01:35:39] <davidf> Or can it?
[01:35:53] <jmkasunich> it has nothing to do with processes
[01:36:05] <davidf> Don't forget, this happens at 20 & 30 in/sec too.
[01:36:25] <davidf> 733 mhz UP
[01:36:30] <jmkasunich> inches/min I assume you mean, which is 0.33 and 0.50 inches/sec
[01:36:40] <jmkasunich> so that kind of rules out the base period thing
[01:36:45] <davidf> yes. sorry I keep doing that.
[01:37:15] <jmkasunich> you could still change BASE_PERIOD to 25000 or 30000, it will give you a slightly better step pulse signal
[01:37:37] <jmkasunich> but if the problem is happening at 20 inches/min, that probably not the cause
[01:37:49] <cradek> I want to do a test first if you'll humor me
[01:38:01] <davidf> I guess it could still be wiring problems. I think I should secure all the connections better (especially the ones to the motors from the drives. )
[01:38:28] <davidf> Sure. But first let me tell you,
[01:39:32] <davidf> re the pulse shape, When I first hooked up these drives, I found I got really bad performance till I switched to a pure sin 1/2 cycle instead of a sq wave step.
[01:39:56] <davidf> Does that ring any bells?
[01:39:59] <cradek> not sure what that is
[01:40:01] <jmkasunich> now I really have no idea what you are talking about
[01:40:06] <jmkasunich> what kind of drives are they?
[01:40:09] <cradek> full stepping (one winding at a time) works very badly
[01:40:21] <davidf> The waveform sent to the motors from the drives.
[01:40:44] <davidf> instead of a square step pulse, it is a sin wave
[01:40:55] <cradek> brb
[01:41:11] <jmkasunich> that sounds like you are describing the difference between fullstep and microstepping
[01:41:17] <jmkasunich> but I've never heard it put that way
[01:41:19] <jmkasunich> what kind of drives are they?
[01:41:25] <davidf> $1100.00 Parker Zeta4's
[01:41:49] <davidf> No, think of the waveform of a step pulse train, ok?
[01:41:52] <jmkasunich> gad zooks!
[01:41:58] <davidf> yeah.
[01:42:12] <davidf> But I got them used for $200.00 each.
[01:42:33] <davidf> But they were in service and almost brand new.
[01:42:54] <jmkasunich> zeta4-240?
[01:42:57] <davidf> OK, re the waveform.
[01:43:35] <davidf> No, the 120 Volt type.
[01:43:45] <davidf> Same thing, at 120 V.
[01:43:59] <davidf> Really dont get it?
[01:44:10] <jmkasunich> I'm googling for a manual
[01:44:18] <davidf> Square wave you get right?
[01:45:06] <jmkasunich> I probably understand the drive just fine, but I don't understand the terminology you are using... this is one of those things that needs a picture, not typing at each other
[01:45:18] <davidf> This looks just like that but picture a series of half ovals in a row. Thats it.
[01:45:48] <davidf> OK. Have you ever seen a sin wave on a scope?
[01:46:13] <jmkasunich> yeah
[01:46:29] <davidf> Just picture the top (positive) half.
[01:46:47] <jmkasunich> yeah
[01:46:48] <davidf> That's what my motors get.
[01:47:38] <davidf> Instead of instant rise from 0 to the max voltage of each step pulse, it is a gentle rise.
[01:48:00] <anonimasu> hm
[01:48:16] <davidf> But I don't think that's the problem. I really think this is acting like a wiring thing.
[01:48:43] <jmkasunich> microstepping uses sin/cosine currents
[01:48:54] <jmkasunich> I certainly hope thats what the drive is putting out
[01:49:17] <davidf> When I got the drives, they didn't come with a special plug, so I had to rig some connectors. They don't fit too well. So they might be getting loose, and causing missed steps.
[01:49:29] <davidf> Ahkam's Razor. (sp?)
[01:49:37] <davidf> Ahkim's
[01:49:37] <jmkasunich> Occams ;-)
[01:49:45] <davidf> yeah?
[01:49:53] <davidf> Boy was I off! :)
[01:50:21] <jmkasunich> http://en.wikipedia.org/wiki/Occams_razor
[01:50:25] <davidf> BTW who was that gut Dess Cat Tez?
[01:50:41] <davidf> Dess Cart Teez?
[01:51:46] <davidf> I think before I bother you guys any further I should make sure it's not just sloppy wires.
[01:51:56] <jmkasunich> which connector don't you have? the signal one (labeled indexer) or the power one labeled motor?
[01:52:13] <davidf> Power to the motor.
[01:52:46] <davidf> I have one, but oh, darn. That cant be it, because that axis does it too... Boo Hooo...
[01:53:11] <jmkasunich> bad motor connections will do a lot worse than missed steps
[01:53:15] <jmkasunich> they can damage drives
[01:53:26] <davidf> But still, I have wires all over the bench, proto board, really make shift at present.
[01:53:33] <jmkasunich> you have the manual for the drives, right?
[01:53:38] <davidf> Yes.
[01:53:52] <davidf> Got it on line you there?
[01:53:56] <jmkasunich> and you've done the installation and configuration by the book, including things like the automatic quick test?
[01:54:07] <jmkasunich> I found the manual
[01:54:12] <davidf> Yes, all of it.
[01:54:48] <davidf> Anyway, all 3 drives are doing this, so they probably are not all bad.
[01:54:54] <jmkasunich> right
[01:55:14] <jmkasunich> lets talk about dip switches
[01:55:24] <davidf> OK.
[01:55:26] <jmkasunich> page 14 of the manual
[01:55:48] <jmkasunich> switcehs 10 and 11, you have them both ON? (I assume that what you were talking about with the sine/square stuff)
[01:55:54] <davidf> I dont have it on this box. Cant look at it with you.
[01:55:57] <jmkasunich> switches
[01:56:09] <davidf> yes.
[01:56:13] <jmkasunich> http://www.parkermotion.com/manuals/ZETA/ZETA_Rev_A_Entire.pdf
[01:56:21] <davidf> ok thanks.
[01:56:57] <jmkasunich> how many steps/rev do you have it set for?
[01:57:02] <davidf> 400
[01:57:14] <davidf> half step.
[01:57:25] <jmkasunich> oh...
[01:57:30] <davidf> 200 steps/rev motors.
[01:57:31] <jmkasunich> thats almost as bad as fullstep
[01:57:35] <davidf> yes.
[01:57:45] <jmkasunich> so you have 20 pitch screws?
[01:57:55] <jmkasunich> 400 * 20 = 8000
[01:57:57] <davidf> But, my next choice was too fast as I recall. Let me look.
[01:58:10] <davidf> Yes that's what I remember.
[01:58:16] <davidf> yes.
[01:58:29] <davidf> 20 TPI
[01:58:33] <jmkasunich> next choice is 1000 steps/rev, or 20000 steps/inch
[01:58:41] <davidf> yes...
[01:58:42] <jmkasunich> that is gonna make for a slow machine
[01:58:52] <davidf> right.
[01:59:30] <davidf> And It seemed like there was a problem with the software or computer not fast enough.
[01:59:56] <jmkasunich> with your computer, you can probably lower BASE_PERIOD to 25000
[02:00:17] <jmkasunich> that would let you generate an absolute maximum of 20,000 steps/second
[02:00:28] <davidf> Right because my top speed on this box is like 50000 / sec or something like that.
[02:00:59] <jmkasunich> BASE_PERIOD = 50000 means top speed is 10,000 steps/second
[02:01:09] <jmkasunich> large base period = slow step rate
[02:01:23] <jmkasunich> slow PC's need a large base period, fast ones can use a smaller base period
[02:10:40] <davidf> jmkasunich, still there?
[02:10:51] <jmkasunich> yeah
[02:11:01] <jmkasunich> going to bed soon, but here for a few more minutes
[02:12:34] <davidf> I don't think we're going to find a problem with any settings. This developed only recently. Hadn't had any probles prior. Don't think it is the particular shapes I'm cutting either, since Z does it too, & never did before.
[02:12:50] <jmkasunich> before?
[02:12:53] <davidf> I think it's hardware. Probably wiring.
[02:12:57] <jmkasunich> I thought you just installed EMC?
[02:13:08] <davidf> No, About a month ago.
[02:13:18] <davidf> Cut lots of things with np
[02:13:24] <jmkasunich> and its been running fine until now?
[02:13:32] <davidf> Two days ago it started.
[02:13:38] <davidf> Yes.
[02:13:54] <davidf> Sorry, said that before, guess you didn't see it.
[02:13:59] <jmkasunich> ok, I was confused, because we were helping you install stuff only recently. I thought this was a new installation
[02:14:23] <davidf> No, I was actually just making changes, no installing.
[02:14:33] <jmkasunich> like the home switch?
[02:14:40] <davidf> You helped install quite awhile ago.
[02:14:46] <davidf> Yes.
[02:15:06] <jmkasunich> is it possible the home switch wiring is conducting noise?
[02:15:26] <davidf> The main thing was getting the home switches handled in .hal and config files.
[02:15:33] <jmkasunich> ok
[02:15:53] <davidf> I checked it on the scope. Didn't see anything.
[02:16:10] <davidf> Ran with them disconnected too. No help.
[02:16:17] <jmkasunich> ok
[02:16:34] <jmkasunich> you didn't tell us any of that ;-/
[02:17:01] <davidf> No, I think the next step is to get this rat's nest under control. There is a lot of room for electrical connecion probs.
[02:17:08] <jmkasunich> I'm more accustomed to helping people who have no scope, and very little experience, and usually with a freshly built machine
[02:17:17] <davidf> Didn't ask. :)
[02:17:31] <davidf> right.
[02:17:35] <jmkasunich> so I was in "help configure the machine" mode, instead of "it used to work, what went wrong" mode
[02:17:36] <davidf> I understand.
[02:18:24] <davidf> I seem pretty duh probably, 'cause I am about some of this, but I have pretty good trouble shooting skills when it comes to the EE part.
[02:18:39] <jmkasunich> ok
[02:18:48] <davidf> BSEE, BSCoE.
[02:18:50] <jmkasunich> IRC doesn't lend itself to quick communication
[02:18:54] <jmkasunich> BSEE here too
[02:18:59] <davidf> But ancient. :)
[02:19:06] <davidf> I'm 53.
[02:19:11] <davidf> ouch.
[02:19:16] <jmkasunich> thats only semi-ancient
[02:19:18] <jmkasunich> I'm 44
[02:19:32] <davidf> That's only almost young. :)
[02:19:54] <davidf> Your prostate will start acting up soon. Just wait.
[02:20:06] <jmkasunich> thanks.... just what I wanted to hear
[02:20:06] <davidf> :)
[02:21:41] <davidf> Look, I really think it is going to be a wire prob or something stupid like utility power blips. That's actually a real possibility. Heat wave here, lots of AC's loading the grid, etc.
[02:21:51] <anonimasu> hm
[02:22:00] <jmkasunich> ok
[02:22:18] <davidf> AC went off for 1.5 minutes 2 days ago. Could be power switching glitchews, who knows?
[02:22:34] <davidf> glitchews. Thats cute.
[02:23:23] <davidf> OK, let me try to get wired right, and then see how it goes. I'll also try it late night when power is more stable.
[02:23:53] <davidf> You guys have worked hard enough for what I'm paying you.
[02:25:13] <davidf> Thanks for the input, and all the help getting my files right. At least that part is good now, and I think I actually understand most of the important stuff now., far better than before, & for that I'm really grateful for your input.
[02:25:25] <jmkasunich> good luck
[02:25:29] <jmkasunich> let us know what happens
[02:25:31] <davidf> Thanks.
[02:26:11] <davidf> See you later. I'm optimistic. Just hope I do'nt find out I have $800.00 worth of bad drives.
[02:26:21] <jmkasunich> unlikely
[02:26:27] <davidf> G'nite.
[02:26:32] <jmkasunich> goodnight
[02:26:40] <davidf> Nite all.
[02:28:39] <anonimasu> night
[02:51:46] <A-L-P-H-A> shit someone died hang gliding in my town... dmess goes around there too.
[03:16:19] <skunkworks> 2.0.2 - nice
[03:16:36] <cradek> no, 2.0.3 now
[03:17:07] <skunkworks> wow - get lost for a month and look what happens :)
[03:19:16] <cradek> cradek has changed the topic to: Welcome! EMC (Enhanced Machine Controller) is a linux-based CNC control. | EMC 2.0.3 is out! | Home: www.linuxcnc.org | Wiki: wiki.linuxcnc.org | Usage map: www.frappr.com/emc2/
[03:19:59] <skunkworks> how is the lathe coming?
[03:20:05] <skunkworks> cradek:
[03:20:08] <cradek> it works great
[03:20:21] <cradek> very happy with it
[03:20:32] <skunkworks> nice work. Have you threaded with it?
[03:20:42] <cradek> but my quick change tool post is cheap, it needs some work to be any good
[03:21:09] <cradek> yes it threads great at ~ 1000rpm
[03:21:14] <skunkworks> wow
[03:21:25] <cradek> 1000 rpm, 20 tpi = 50 ipm on Z
[03:21:39] <skunkworks> wow - any issues?
[03:21:45] <cradek> nope, works perfectly
[03:21:50] <skunkworks> how is the tracking at that speed>"
[03:22:00] <skunkworks> nice. video)
[03:22:02] <skunkworks> ?
[03:22:32] <cradek> perfect as far as I can tell, very nice shaving peeled off one side of the thread, all in one piece
[03:23:00] <skunkworks> wow. neat
[03:23:25] <cradek> the nist lathe showed just a little variation per revolution in the shaving
[03:23:35] <cradek> but that must be because its encoder is a little wobbly
[03:23:48] <cradek> mine is nicely concentric
[03:58:31] <Jymmm> cradek: when you mounted a tool on the tbale of your mill as a pseudo lathe, how was the result?
[03:58:51] <cradek> fair
[03:59:08] <Jymmm> biggest drawback?
[03:59:17] <cradek> you can only cut a part as long as the distance from the tool to the table
[03:59:23] <cradek> anything like a tailstock is impossible
[03:59:33] <Jymmm> ah
[03:59:45] <Jymmm> but it did work cleanly?
[04:00:36] <cradek> yes I guess, for what it could do
[04:01:23] <Jymmm> ok, just curious
[04:05:30] <Jymmm> I don't have a lathe, but I wanted to make some small pulleys out of plastic. I didn't know if it was worth even trying, or just mold em, or somethign else.
[04:06:48] <cradek> what diameter?
[04:07:03] <Jymmm> max being around 3"
[04:07:18] <Jymmm> min around .5"
[04:07:21] <cradek> might be possible with care
[04:08:03] <Jymmm> I jsut can't find any is my problem
[04:08:12] <cradek> pulleys are a challenge even on a lathe because they must be perfectly concentric, so you have to either bore the hole and cut the outside in one mounting (usually impossible) or bore the hole and then turn the rest between centers
[04:08:22] <Jymmm> Well, unless I hit toy-r-us and rip some toy apart at least
[04:10:16] <Jymmm> Hmmm, I do have the right side of my grinder I could play with.... 2200 rpm iirc, just not sure about the rest.
[04:11:15] <Jymmm> eh, nm. I try to find some somewhere.
[04:11:30] <Jymmm> thanks
[04:11:33] <cradek> probably easier to buy them in this case unless they're really special
[04:11:51] <Jymmm> nah, I could probably use the cheapest
[04:13:08] <Jymmm> oh gawd, I can't remember the word atm....
[04:13:39] <Jymmm> What is the thing you put a buffing pad on, then put it into a motor?
[04:14:14] <cradek> arbor?
[04:14:22] <Jymmm> close
[04:14:39] <Jymmm> another word
[04:16:08] <Jymmm> ah mandrel
[04:16:29] <Jymmm> I need to make a "mandrel extender"
[04:16:47] <cradek> right mandrel, I couldn't come up with it either
[04:17:00] <Jymmm> just one of them words that slip your tounge =)
[04:17:57] <Jymmm> I'm just not sure how to hold/lock that mandrel into it. Some tiny stuff
[04:18:03] <cradek> do you use a tapered mandrel for buffing? they're really nice for changing wheels easily
[04:18:17] <cradek> my buffer has that
[04:18:42] <Jymmm> Actually, I bought a 6" bench grinder, and tossed on cotton buffying wheels instead
[04:19:19] <cradek> that works
[04:19:34] <Jymmm> Yeha, you stack multiple buffing pads together
[04:20:09] <cradek> but I bet mounting them is a pain because they don't self-center
[04:20:26] <cradek> they center themselves on the taper as soon as you turn it on
[04:20:51] <cradek> I'm off, goodnight
[04:20:56] <Jymmm> g'night
[07:46:39] <Lerneaen_Hydra> 'lo
[07:47:02] <Jymmm> lo
[07:47:39] <Jymmm> and, I'm outta here.... G'Night Lerneaen_Hydra =)
[07:47:46] <Lerneaen_Hydra> 'night
[11:33:15] <Lerneaen_Hydra_C> Lerneaen_Hydra_C is now known as Lerneaen_Hydra2
[11:33:19] <Lerneaen_Hydra2> 'lo
[11:33:26] <Lerneaen_Hydra2> * Lerneaen_Hydra2 pokes alex_joni
[11:38:30] <Lerneaen_Hydra2> * Lerneaen_Hydra2 pokes harder
[11:45:27] <Lerneaen_Hydra2> * Lerneaen_Hydra2 brings forth the cattle prod
[11:45:37] <Lerneaen_Hydra2> * Lerneaen_Hydra2 stokes it and gets it warm
[11:46:10] <Lerneaen_Hydra2> * Lerneaen_Hydra2 laughs demonicly
[11:49:45] <Lerneaen_Hydra2> * Lerneaen_Hydra2 takes out the glowing prod from the fire
[11:49:56] <Lerneaen_Hydra2> * Lerneaen_Hydra2 waves it around for good measure
[11:52:45] <A-L-P-H-A> think maybe he's just ignoring you to teach you something
[11:52:49] <Lerneaen_Hydra2> * Lerneaen_Hydra2 pokes alex_joni with the cattle prod
[11:52:51] <A-L-P-H-A> that something, maybe anything.
[11:52:54] <Lerneaen_Hydra2> A-L-P-H-A: you think so?
[11:53:02] <A-L-P-H-A> alex_joni is coy like that.
[11:53:02] <Lerneaen_Hydra2> hmm...
[11:53:07] <Lerneaen_Hydra2> I see
[11:53:17] <Lerneaen_Hydra2> * Lerneaen_Hydra2 puts away the prod, for now
[11:53:27] <A-L-P-H-A> he hates you now, that's all.
[11:54:04] <Lerneaen_Hydra2> that would be impractical from my point of view
[11:54:19] <Lerneaen_Hydra2> not to mention rather unfortunate
[11:54:24] <Lerneaen_Hydra2> hmm...
[11:54:47] <Lerneaen_Hydra2> * Lerneaen_Hydra2 brings forth a bottle of short-term amnesia
[11:55:08] <Lerneaen_Hydra2> think that will do any good?
[12:32:37] <anonimasu> yay
[12:32:42] <anonimasu> I just spun a axis on the mill
[12:40:26] <Lerneaen_Hydra> you mean mill-set-up-as-a-lathe-esque?
[12:41:02] <anonimasu> what?
[12:41:05] <anonimasu> I moved X
[12:41:06] <anonimasu> ;)
[12:41:28] <Lerneaen_Hydra> huh?
[12:41:34] <Lerneaen_Hydra> oh!
[12:41:36] <Lerneaen_Hydra> right
[12:41:39] <anonimasu> "mill-set-up-as-a-lathe-esque"?
[12:41:45] <anonimasu> not with computer control
[12:41:49] <anonimasu> I spun it from the power supply
[12:42:04] <Lerneaen_Hydra> cradek used his mill as a lathe once
[12:42:15] <Lerneaen_Hydra> material in spindle, tool attached to table
[12:42:17] <anonimasu> yeah
[12:42:20] <anonimasu> I know :)
[12:42:27] <anonimasu> I've milled with my lathe way back too :)
[12:42:35] <Lerneaen_Hydra> oh, diagnostic stuff is what you're doing?
[12:42:41] <Lerneaen_Hydra> testing various things?
[12:42:44] <anonimasu> nope, just built a axis finished
[12:42:51] <anonimasu> the current draw was a bit high
[12:42:51] <Lerneaen_Hydra> ok
[12:42:57] <anonimasu> about 10amps..
[12:43:07] <Lerneaen_Hydra> which voltage?
[12:43:10] <anonimasu> 24
[12:43:13] <Lerneaen_Hydra> that sounds like a lot
[12:43:20] <Lerneaen_Hydra> 240 watts
[12:43:25] <Lerneaen_Hydra> heh
[12:43:41] <Lerneaen_Hydra> thats nearly the power of my main spindle
[12:44:01] <Lerneaen_Hydra> it's 300-something
[12:44:03] <anonimasu> might have my preload/gibs adjusted tight
[12:44:07] <anonimasu> heh
[12:44:17] <anonimasu> I have/will have about 4.5kw
[12:44:18] <Lerneaen_Hydra> ballscrew?
[12:44:21] <anonimasu> yeah
[12:44:29] <Lerneaen_Hydra> what size mill?
[12:44:35] <Lerneaen_Hydra> this is semi-large
[12:44:37] <anonimasu> toolroom size..
[12:44:37] <Lerneaen_Hydra> I take it
[12:44:41] <anonimasu> ;)
[12:44:48] <anonimasu> not too small not too large
[12:45:02] <anonimasu> about 700kg I think
[12:45:15] <Lerneaen_Hydra> oh, not that bad
[12:46:27] <anonimasu> trying to find motor specs
[12:46:41] <Lerneaen_Hydra> servo motor?
[12:46:45] <anonimasu> ah peak current is 22a
[12:47:00] <anonimasu> and continous is 3.9
[12:47:22] <anonimasu> but that's at 72v
[12:47:41] <anonimasu> 280w..
[12:47:47] <anonimasu> continous..
[12:48:03] <anonimasu> 1584w peak
[12:48:04] <anonimasu> ^_^
[12:48:10] <Lerneaen_Hydra> oh dear
[12:48:21] <Lerneaen_Hydra> this motor must have some good cooling
[12:48:26] <Lerneaen_Hydra> heatsink+fan
[12:49:20] <anonimasu> nope
[12:49:26] <anonimasu> that's the motor as is from the manufacturer..
[12:49:40] <anonimasu> remember im at 24v
[12:49:47] <anonimasu> the current draw should drop as I raise the voltage
[12:50:10] <Lerneaen_Hydra> what type of motor is this? O.o
[12:50:14] <anonimasu> a galilmc servo
[12:50:26] <anonimasu> http://www.galilmc.com/catalog/catmotors.pdf
[12:50:49] <anonimasu> I just remembered you are a swede too :)
[12:50:51] <Lerneaen_Hydra> current draw lessens at higher voltage? I take it you're not running the motor directly
[12:50:58] <anonimasu> I am right now
[12:51:06] <Lerneaen_Hydra> borkborkbork
[12:51:17] <Lerneaen_Hydra> then how can the draw lessen with a higher voltage
[12:51:22] <Lerneaen_Hydra> ohm's law and all
[12:51:31] <anonimasu> 6
[12:51:33] <Lerneaen_Hydra> unless you get those coils *really* hot
[12:53:28] <Lerneaen_Hydra> I still fail to see how continous current draw lessens as the voltage rises
[12:53:49] <Lerneaen_Hydra> those servos look nice though
[12:53:57] <Lerneaen_Hydra> how much did they cost?
[12:53:58] <anonimasu> I wonder if I should hook up a meter to see if it's true
[12:54:06] <anonimasu> like 4500sek
[12:54:25] <Lerneaen_Hydra> hmm, ok
[12:54:27] <anonimasu> but I'm ordering them from the us next time..
[12:54:34] <Lerneaen_Hydra> where'd you get them from?
[12:54:39] <anonimasu> aaratron.se
[12:54:44] <anonimasu> I got a great price on one
[12:54:52] <anonimasu> around 3000sek
[12:55:07] <Lerneaen_Hydra> that sounds real nice
[12:55:08] <anonimasu> but the guy I talked to was on vacation when I ordered the other ones
[12:55:10] <Lerneaen_Hydra> some small defect?
[12:55:11] <Lerneaen_Hydra> ok
[12:55:13] <anonimasu> no
[12:55:18] <anonimasu> good deal that's all :9
[12:55:36] <Lerneaen_Hydra> haha
[12:55:37] <Lerneaen_Hydra> ok
[12:55:39] <Lerneaen_Hydra> ^^
[12:55:44] <anonimasu> ^_^
[12:55:50] <anonimasu> I'm going to oil it and see again later
[12:56:03] <anonimasu> Lerneaen_Hydra: reasonably the watt should stay the same
[12:56:51] <anonimasu> I cant remember how much I drew before
[12:57:06] <A-L-P-H-A> shit. did harborfreight fix their system, so I can only get one price now for items?
[12:57:15] <Lerneaen_Hydra> assuming it's just a standard DC motor without electronics then the power output (watts) will increase when you raise the voltage
[12:57:19] <Lerneaen_Hydra> ohms law and all
[12:57:21] <A-L-P-H-A> you used to be able to figit with their VGA numbers, to get a differnt price.
[12:57:27] <A-L-P-H-A> any yanks around?
[13:01:59] <Lerneaen_Hydra> r=v/i r=72/3.9, r=18 ohm (sounds suspiciously high), if V = 200, then i=200/18, i=11, and P=11*200=2.2kW
[13:04:19] <anonimasu> yeah
[13:04:28] <anonimasu> you are right
[13:04:30] <anonimasu> fsck.
[13:05:37] <anonimasu> need to check out how much I actually did preload everything
[13:05:52] <Lerneaen_Hydra> hmm
[13:06:04] <Lerneaen_Hydra> if you know the current then you van get an approximate voltage
[13:06:14] <Lerneaen_Hydra> since you know the resistance from the datasheet
[13:06:31] <Lerneaen_Hydra> contact resistance and stuff will only make it approximate though
[13:06:35] <anonimasu> yeah
[13:06:49] <anonimasu> I'm thinking it's just preload
[13:07:07] <anonimasu> the current when I pushed the table went to 12A
[13:07:22] <anonimasu> and I moved instead
[13:07:46] <Lerneaen_Hydra> haha
[13:08:06] <anonimasu> very fine pitched ballscrews and direct drive :)
[13:08:18] <anonimasu> so, that's why im thinking preload
[13:08:25] <Lerneaen_Hydra> just out of curiosity, what do you think a smaller servo motor (DC+encoder) would cost
[13:08:36] <anonimasu> no idea
[13:08:40] <anonimasu> depends on your size
[13:08:57] <Lerneaen_Hydra> this would be a smaller motor, somewhere in the 50W size
[13:09:21] <Lerneaen_Hydra> 1000-1500sek?
[13:09:24] <anonimasu> www.transmotec.se
[13:09:54] <anonimasu> but, gettning them to sell 1 of stuff might be hard
[13:10:32] <Lerneaen_Hydra> mainly system integrators?
[13:10:36] <anonimasu> yeah
[13:10:48] <anonimasu> you can call them up to talk with them and ask about it
[13:10:51] <anonimasu> it's probably the most accurate
[13:11:21] <Lerneaen_Hydra> ok
[13:11:27] <anonimasu> they shouldnt nescessarily be that expensive
[13:11:27] <Lerneaen_Hydra> bought from them before?
[13:11:39] <anonimasu> no, just did some queries on some solenoids
[13:12:10] <Lerneaen_Hydra> ok
[13:12:25] <anonimasu> the dealer I buy plc's and stuff only seems to have larger stuff
[13:12:53] <anonimasu> yeah they start at 1,0kAV
[13:12:55] <anonimasu> kVA
[13:12:59] <Lerneaen_Hydra> not that bad, 50W seems to be in the 500-1000 sek range
[13:13:13] <Lerneaen_Hydra> with encoder and all
[13:13:23] <anonimasu> I'm gonna query them if they can fix me servos cheap ;)
[13:13:36] <Lerneaen_Hydra> haha
[13:13:48] <Lerneaen_Hydra> motor drivers were pricy though, that may be a DIY thing
[13:13:59] <anonimasu> hehe
[13:14:11] <anonimasu> I'd buy geckos any day unless you have lots of sparetime
[13:14:28] <Lerneaen_Hydra> That's what I've got ;)
[13:14:36] <Lerneaen_Hydra> and it's educational
[13:14:39] <anonimasu> ah
[13:14:40] <anonimasu> ok
[13:15:14] <anonimasu> http://www.lust-antriebstechnik.de/downloads/assets/lsh-t_catalogue_01-2006_en.pdf
[13:16:08] <anonimasu> I want that kind of servos :)
[13:16:18] <Lerneaen_Hydra> indeed
[13:16:22] <Lerneaen_Hydra> they wouldn't be bad
[13:18:41] <anonimasu> * anonimasu sighs
[13:18:49] <anonimasu> well, I guess I need to wait until tomorrow when I have my meter
[13:19:25] <anonimasu> I then have a look at what's sticky
[13:19:30] <anonimasu> sticky/sticks
[13:20:04] <anonimasu> *yawn*
[13:22:46] <anonimasu> Lerneaen_Hydra:
http://www.truetex.com/servomod.htm
[13:24:27] <Lerneaen_Hydra> oh
[13:24:35] <Lerneaen_Hydra> that looks relatively simple
[13:25:31] <anonimasu> yep
[14:01:03] <Lerneaen_Hydra> * Lerneaen_Hydra pokes alex_joni
[14:01:07] <Lerneaen_Hydra> * Lerneaen_Hydra yet again
[15:31:10] <alex_joni> Lerneaen_Hydra: what?
[15:37:49] <alex_joni> Lerneaen_Hydra: I'll be back in a couple of hours.. state your questions in here, I'll answer later
[15:37:53] <alex_joni> or drop me an email
[15:47:07] <skunkworks> Hi alex
[15:49:31] <Lerneaen_Hydra> alex_joni: hi
[15:49:34] <Lerneaen_Hydra> * Lerneaen_Hydra reads
[15:49:37] <Lerneaen_Hydra> gah!
[15:49:39] <Lerneaen_Hydra> anyway
[15:50:17] <Lerneaen_Hydra> I saw yesterday that you had added stuff to emc IRT FO wheels
[15:50:45] <Lerneaen_Hydra> so I was wondering if you had a hal file with the parameters that are needed
[15:51:21] <Lerneaen_Hydra> slow day today eh?
[15:52:32] <alex_joni> busy day today ;)
[15:52:39] <alex_joni> Lerneaen_Hydra: only 2 things needed
[15:52:47] <alex_joni> 1. halui.feed-override.scale
[15:53:06] <alex_joni> it's a pin, but meant as a parameter (I did set it to 0.1)
[15:53:28] <alex_joni> 2. halui.feed-override.counts (input from an counter: e.g. an encoder module output)
[15:53:53] <alex_joni> when the count increases by 1, FO will increase by 0.1
[15:54:09] <alex_joni> FO is from 0 to 1 (for 100%)
[15:54:22] <alex_joni> so 0 = 0, 0.1 = 10%, 0.2 = 20%, etc
[15:54:53] <alex_joni> Lerneaen_Hydra: got it?
[15:57:49] <Lerneaen_Hydra> alex_joni: here again
[15:57:56] <Lerneaen_Hydra> so there are only two rows?
[15:58:03] <alex_joni> two rows?
[15:58:15] <Lerneaen_Hydra> except for the stuff that's common between FO wheel and jogwheel
[15:58:20] <alex_joni> right
[15:58:44] <Lerneaen_Hydra> scale, would that be FO as how much to change per step?
[15:58:51] <alex_joni> right
[15:58:55] <Lerneaen_Hydra> so 1 step would give a change of 10% in your case?
[15:59:05] <alex_joni> if you set scale to be 0.01, then 1 step would change 1%
[15:59:12] <alex_joni> if you set scale to be 0.1, then 1 step would change 10%
[15:59:27] <alex_joni> if you set scale to be 1, then 1 step would change 100%
[15:59:32] <Lerneaen_Hydra> ok
[15:59:38] <Lerneaen_Hydra> so what does counts do?
[15:59:47] <Lerneaen_Hydra> or is that the actual input
[15:59:49] <alex_joni> there is where you feed the counting
[15:59:53] <alex_joni> right, actual input
[15:59:55] <Lerneaen_Hydra> oh, ok
[15:59:58] <Lerneaen_Hydra> from a pin or something?
[16:00:06] <alex_joni> say you have an jogwheel
[16:00:26] <alex_joni> while turning it, the ticks get counted
[16:00:34] <Lerneaen_Hydra> oh, ok
[16:00:42] <alex_joni> if you go one direction it increases, the other way it decreases
[16:00:51] <Lerneaen_Hydra> what is 0=
[16:00:55] <Lerneaen_Hydra> s/=/?
[16:01:01] <Lerneaen_Hydra> the startup value=
[16:01:04] <alex_joni> shouldn't matter
[16:01:09] <Lerneaen_Hydra> gah.. s/=/?
[16:01:23] <alex_joni> it works as a relative thing
[16:01:24] <Lerneaen_Hydra> ok. could there be an overflow error if you spin it a lot?
[16:01:27] <Lerneaen_Hydra> oh, ok
[16:01:30] <alex_joni> unlikely
[16:01:37] <alex_joni> it is clamped to the max_fo
[16:01:47] <alex_joni> so you can keep it spinning up, and it will stay at max_fo
[16:01:57] <Lerneaen_Hydra> I was thinking if you use it as a jogwheel
[16:02:09] <alex_joni> the next time you turn it lower, feed_override will decrease
[16:02:10] <Lerneaen_Hydra> or something else without a definite max & min
[16:02:16] <alex_joni> right
[16:02:25] <Lerneaen_Hydra> ok. so very intuitive behavior
[16:02:35] <alex_joni> I think the best way to use it is to use those cheap mechanical encoders
[16:02:45] <alex_joni> like a trimpot, but outputs quadrature
[16:02:49] <alex_joni> 4-5$ at most
[16:03:09] <alex_joni> they have about 10 ticks / rotation
[16:03:29] <Lerneaen_Hydra> the code basically says on_count, if current<1&>0, counter=counter+scale, else return?
[16:03:52] <Lerneaen_Hydra> yeah, my jogwheel & FO wheels have 16 positions per turn
[16:03:56] <alex_joni> new_fo = current_fo + (counts-old_counts)*scale
[16:03:59] <Lerneaen_Hydra> quite suitable for this
[16:04:18] <alex_joni> right
[16:04:34] <alex_joni> you could have the scale 0.625
[16:04:44] <alex_joni> that way a full turn goes from 0 to 100%
[16:04:48] <Lerneaen_Hydra> are these changes in head?
[16:04:51] <alex_joni> yes
[16:04:56] <alex_joni> and will be in 2.1.x
[16:05:07] <alex_joni> I also want to add up/down pins for feed_override
[16:05:09] <Lerneaen_Hydra> I'm probably going to have a scale of 0.05
[16:05:14] <Lerneaen_Hydra> for buttons?
[16:05:17] <alex_joni> right
[16:05:37] <Lerneaen_Hydra> oh, most commercial CNC's have quasi-logarithmic FO wheels
[16:05:43] <Lerneaen_Hydra> maybe something to look into
[16:05:52] <alex_joni> I might.. AXIS does that now
[16:05:54] <Lerneaen_Hydra> 1% steps from 1-10%, after that 5% steps
[16:05:57] <alex_joni> for jogging I think
[16:06:01] <Lerneaen_Hydra> or something similar
[16:06:07] <alex_joni> you can alter scale anyway you like
[16:06:10] <Lerneaen_Hydra> the version I have does that
[16:06:23] <Lerneaen_Hydra> so scale doesn't have to be linear?
[16:06:24] <alex_joni> because it's a pin, you can have it generated by a small math module
[16:06:28] <alex_joni> scale is a value
[16:06:29] <Lerneaen_Hydra> oh!
[16:06:32] <alex_joni> but it can change
[16:06:44] <Lerneaen_Hydra> so a variable is acceptable to have?
[16:06:50] <alex_joni> yeah
[16:06:54] <Lerneaen_Hydra> nice
[16:06:55] <alex_joni> but it needs to come from HAL
[16:07:10] <Lerneaen_Hydra> right. how much math can you do in hal?
[16:07:15] <alex_joni> so some HAL component that does some calculation on the current feedoverride value..
[16:07:25] <Lerneaen_Hydra> right
[16:07:29] <alex_joni> there are some basic functions already there (adding, ...)
[16:07:42] <Lerneaen_Hydra> a subs. table would also work
[16:07:51] <alex_joni> but you can always write a short component that does what you want
[16:07:54] <alex_joni> it's pretty easy
[16:08:00] <Lerneaen_Hydra> ok
[16:08:34] <Lerneaen_Hydra> I'll go and test those lines then
[16:08:48] <alex_joni> coo, I'm around if you need help
[16:08:59] <alex_joni> btw, there is another pin (halui.feed-override.value)
[16:09:06] <alex_joni> you can look at it with a halmeter
[16:09:13] <Lerneaen_Hydra> current FO?
[16:09:17] <alex_joni> it shows the current feed-override, as reported by emc
[16:09:21] <alex_joni> right
[16:13:50] <Lerneaen_Hydra2> hi again
[16:13:57] <Lerneaen_Hydra2> logger_aj: bookmark
[16:13:57] <Lerneaen_Hydra2> See
http://81.196.65.201/irc/irc.freenode.net:6667/emc/2006-08-05#T16-13-57
[16:19:32] <Lerneaen_Hydra2> the first pin (halui.feed-override.scale) how do I set it?
[16:19:40] <Lerneaen_Hydra2> just have: halui.feed-override.scale 0.05
[16:19:48] <Lerneaen_Hydra2> or something else?
[16:21:03] <Lerneaen_Hydra2> * Lerneaen_Hydra2 pokes alex_joni
[16:21:35] <jepler> Lerneaen_Hydra2: create a float signal, setf it, and linksp it to halui
[16:21:46] <Lerneaen_Hydra2> oh
[16:21:47] <Lerneaen_Hydra2> ok
[16:22:54] <alex_joni> sets
[16:23:47] <Lerneaen_Hydra2> hmm, what is the syntax for those rows?
[16:23:57] <alex_joni> newsig scale float
[16:24:02] <alex_joni> sets scale 0.05
[16:24:11] <alex_joni> linksp scale halui.feed-override.scale
[16:25:22] <Lerneaen_Hydra2> http://pastebin.ca/117009
[16:25:30] <Lerneaen_Hydra2> is a portion of the hal file
[16:25:42] <Lerneaen_Hydra2> does that look reasonable?
[16:25:54] <Lerneaen_Hydra2> other than some comments being old ;)
[16:26:37] <alex_joni> except line 69
[16:27:03] <alex_joni> that needs to be deleted
[16:27:11] <Lerneaen_Hydra2> right
[16:27:15] <alex_joni> and line 71. needs to read linkpp, not linksp
[16:27:39] <Lerneaen_Hydra2> ok
[16:27:42] <Lerneaen_Hydra2> anything else?
[16:27:56] <Lerneaen_Hydra2> * Lerneaen_Hydra2 tries to get emc to compile
[16:28:14] <alex_joni> nope
[16:29:55] <Lerneaen_Hydra2> hmm, compile failed
[16:30:11] <Lerneaen_Hydra2> make: Failed to remake makefile `Makefile'.
[16:30:11] <Lerneaen_Hydra2> make: *** No rule to make target `../configs/dallur-thc/emc.nml', needed by `configs'. Stop.
[16:33:36] <alex_joni> check permissions on that file
[16:34:20] <Lerneaen_Hydra2> that file doesn't exist O.o
[16:34:30] <Lerneaen_Hydra2> yet I did a cvs up
[16:35:05] <alex_joni> maybe jepler knows
[16:35:17] <alex_joni> try copying it yourself
[16:35:25] <alex_joni> from configs/common
[16:36:23] <Lerneaen_Hydra2> before I get too excited: the command "cvs up" is all that's needed, right?
[16:36:32] <Lerneaen_Hydra2> in the correct directory
[16:36:53] <alex_joni> try -dP
[16:37:00] <alex_joni> cvs up -dP
[16:37:04] <Lerneaen_Hydra2> ok
[16:38:35] <Lerneaen_Hydra2> now it got those files
[16:38:46] <Lerneaen_Hydra2> so cvs up doesn't always get everything?
[16:39:12] <alex_joni> except new dirs
[16:39:17] <Lerneaen_Hydra2> oh
[16:39:44] <Lerneaen_Hydra2> when compiling: hal/utils/halcmd.c:1964: warning: implicit declaration of function ‘nanosleep’
[16:40:35] <Lerneaen_Hydra2> the compile finished succesfully thoough
[16:41:57] <Lerneaen_Hydra2> have there been big changes to the config files?
[16:42:06] <Lerneaen_Hydra2> it didn't like my old one
[16:42:33] <wjweder> can anyone answer the question: in the EMC2 User's manual.pdf section 4.1.2 there is an example command "bin/hal_parport 0278" but the binary does not exist in emc2\bin
[16:43:10] <Lerneaen_Hydra2> line 315: axis-remote: command not found
[16:43:21] <Lerneaen_Hydra2> wjweder: AFAIK the command specifies which parport to use
[16:43:32] <Lerneaen_Hydra2> in this case the integrated one on the motherboard
[16:44:01] <Lerneaen_Hydra2> PCI parports have... varying adresses
[16:44:56] <wjweder> when i issue the command i get "command not found"
[16:46:33] <Lerneaen_Hydra2> I'm guessing here but I think the command has to go in a hal terminal and not a standard one
[16:47:04] <alex_joni> wjweder: are you using an installed version of emc2?
[16:47:24] <alex_joni> in that case all the commands in the manual starting with bin/ need to be issued without
[16:47:33] <wjweder> yes i am using ubuntu/emc2 package
[16:47:43] <alex_joni> ok, try hal_parport
[16:47:55] <alex_joni> instead of bin/hal_parport
[16:48:34] <wjweder> tried it, got same result, command not found
[16:49:10] <alex_joni> it might be missing .. I know it has been recently removed, wasn't sure that the 2.0.x doesn't have it
[16:49:24] <alex_joni> it's basicly the user-space version of the parport driver
[16:49:52] <wjweder> i did a dir in emc2/bin and it was not there
[16:50:12] <alex_joni> emc2/bin ? where did that come from?
[16:50:30] <alex_joni> did you make that dir?
[16:51:11] <wjweder> sorry the dir command was a holdover from dos command prompt, i suppose
[16:51:52] <Mhel> hello.. I'm a hobbyists, linux newbie, also new to emc. I have successfully install emc in slackware. where can I get info on *.ini settings?
[16:52:04] <alex_joni> Mhel: in the wiki
[16:52:16] <wjweder> the ls command does not find Hal_parport either
[16:52:22] <alex_joni> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?Configuring_EMC2
[16:52:35] <Mhel> thanks I'll look
[16:53:21] <alex_joni> wjweder: the manual you are reading assumes run-in-place (where you get the source and put it in a directory, and compile it, then all the executables are in bin/ )
[16:53:41] <alex_joni> wjweder: for the installed version, you can find the installed executables in /usr/bin
[16:54:14] <Mhel> found it, thanks BTW the topic says emc 2.0.3 is out not in sourceforge yet tho.
[16:54:27] <alex_joni> Mhel: yeah, but you can get it..
[16:54:30] <alex_joni> hang on for a link
[16:54:41] <Mhel> thanks
[16:55:05] <alex_joni> Mhel:
http://cvs.linuxcnc.org/cgi-bin/cvsweb.cgi/emc2/emc2.tar.gz?only_with_tag=RELEASE_2_0_3;tarball=1
[16:55:21] <alex_joni> Mhel: maybe you can add your experience with slack to the wiki
[16:56:03] <alex_joni> Mhel: any different than this ?
http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?RtaiSteps
[16:56:10] <wjweder> i issued "ls hal*" command in usr/bin, the command is not there
[16:56:56] <Mhel> sure, I'm realy newbie just installed slack last monday just for EMC and been learning since, 'took me the whole week to install emc.
[16:57:20] <alex_joni> Mhel: sure sounds like a fast job :)
[16:57:26] <Lerneaen_Hydra2> Mhel: why slackware?
[16:57:26] <alex_joni> Mhel: and I'm not kidding
[16:58:01] <Mhel> I just happen to have it already, needless to say I'm on dial-up (a cheap person I am)
[16:58:15] <Lerneaen_Hydra2> Mhel: oh. ok.
[16:58:17] <alex_joni> Mhel: using rtai?
[16:58:22] <alex_joni> what kind of kernel?
[16:59:00] <Mhel> yes, I tried RTLinux ,but can't make it work, RTAi 2.4.31 adeos thingy :)
[16:59:14] <alex_joni> ok, that's a stable choice
[16:59:34] <alex_joni> not the bleeding edge, but nice ;)
[16:59:53] <alex_joni> Mhel: you're using emc2.. right?
[17:00:02] <Mhel> it's good enough to start with I guess
[17:00:06] <Mhel> yes EMC2
[17:00:47] <alex_joni> Mhel: one more thing.. can I ask you something?
[17:00:49] <Mhel> axis build has issue tho, it's build inside the EMC src and not the install dir
[17:01:01] <Mhel> sure
[17:01:11] <alex_joni> can you add a pin to www.frappr.com/emc2 ?
[17:01:21] <Mhel> ok.
[17:01:25] <alex_joni> coo :)
[17:01:34] <Lerneaen_Hydra2> alex_joni: it seems that my problem was something else, insmod: "error inserting '/usr/realtime-2.6.15-magma/modules/rtai_hal.ko': -1 File exists"
[17:01:53] <alex_joni> Lerneaen_Hydra2: sounds like a bad cleanup last time
[17:02:19] <alex_joni> sudo rmmod rtai_hal
[17:02:32] <alex_joni> but it will probably say it's in use
[17:02:44] <Lerneaen_Hydra2> yes
[17:02:53] <Lerneaen_Hydra2> /etc/init.d realtime stop?
[17:03:00] <Lerneaen_Hydra2> or something thereof?
[17:03:07] <alex_joni> Lerneaen_Hydra2: you're running run-in-place
[17:03:11] <alex_joni> so scripts/realtime stop
[17:03:17] <Lerneaen_Hydra2> oh, ok
[17:03:19] <alex_joni> but first rmmod motmod
[17:03:29] <alex_joni> and lsmod to see what else is there
[17:03:57] <Lerneaen_Hydra2> lots
[17:04:08] <alex_joni> lsmod | grep hal
[17:04:18] <Lerneaen_Hydra2> http://pastebin.ca/117059
[17:04:39] <alex_joni> rmmod hal_lib
[17:04:44] <alex_joni> then scripts/realtime stop
[17:05:35] <Lerneaen_Hydra2> which dir is that in?
[17:05:43] <Lerneaen_Hydra2> oh
[17:05:46] <Lerneaen_Hydra2> I'm blind
[17:06:03] <Lerneaen_Hydra2> after that start as usual?
[17:06:07] <alex_joni> right
[17:06:49] <Lerneaen_Hydra2> waiting for 'iocontrol' is taking lots of time
[17:06:58] <alex_joni> strange
[17:07:25] <Lerneaen_Hydra2> could head be borked?
[17:07:38] <alex_joni> unlikely
[17:07:40] <Lerneaen_Hydra2> or is this something else?
[17:07:53] <alex_joni> check emc2/VERSION
[17:07:57] <alex_joni> tell me what it says
[17:08:50] <Lerneaen_Hydra2> pre 2.1 cvs head
[17:08:57] <alex_joni> it seems ok.. odd
[17:11:02] <Lerneaen_Hydra2> anything more to test?
[17:11:11] <alex_joni> no.. does it still hang?
[17:11:19] <Mhel> alex_joni: thanks for the links, I'm still dual booting my modem doesn't work with slack yet. soon... gtg, am very eager to try my homebuilt frankenstein router.
[17:11:31] <alex_joni> Mhel: great to hear
[17:12:03] <Lerneaen_Hydra2> alex_joni: well, it doesn't hang with an error, it just sits there and waits
[17:12:16] <alex_joni> Lerneaen_Hydra2: try one of the default configs
[17:12:23] <alex_joni> I bet you used your old config
[17:12:27] <Lerneaen_Hydra2> yep
[17:12:31] <Lerneaen_Hydra2> oh!
[17:12:42] <Lerneaen_Hydra2> now there was an error after pressing ctrl c
[17:12:51] <Lerneaen_Hydra2> line 315: axis-remote: command not found
[17:14:09] <alex_joni> * alex_joni needs to run for a while
[17:14:15] <alex_joni> Lerneaen_Hydra2: try running a default config
[17:14:24] <alex_joni> configs have been changed a bit
[17:14:36] <alex_joni> especially the motmod stuff (BASE_PERIOD & co)
[17:14:44] <alex_joni> I'll be back later
[17:14:44] <Lerneaen_Hydra2> wtf! in my config there is now a line that says: #- tool table fileline 315: axis-remote: command not found
[17:14:48] <Lerneaen_Hydra2> ok
[17:14:57] <Lerneaen_Hydra2> so I copy over stuff from my config to a new one?
[17:15:15] <Lerneaen_Hydra2> only the ini I take it
[17:22:27] <Lerneaen_Hydra2> alex_joni: hmm, nope
[17:22:30] <Lerneaen_Hydra2> still that same thing
[17:22:53] <Lerneaen_Hydra2> and it also says the same error after pressing ctrl c
[17:23:01] <Lerneaen_Hydra2> exactly the same behavior
[17:23:08] <Lerneaen_Hydra2> when running the included scripts
[17:25:35] <jepler> hm, 'axis-remote: command not found' may be my fault
[17:26:09] <jepler> well, it means you haven't upgraded your AXIS CVS lately
[17:26:30] <jepler> * jepler just checked to be sure it doesn't get run when other GUIs are selected
[17:29:41] <Lerneaen_Hydra> hmm, I set the config to load axis
[17:30:00] <Lerneaen_Hydra> and I updated axis too, or at least I think I did
[17:36:38] <Lerneaen_Hydra> * Lerneaen_Hydra pokes jepler
[17:39:59] <jepler> setup.py in the CVS version of axis creates bin/axis-remote
[17:41:46] <Lerneaen_Hydra> and that file needs to be run?
[17:42:12] <jepler> when shutting down, the emc script new uses axis-remote to request that axis shut down if it hasn't yet
[17:42:20] <jepler> the emc script *now* uses
[17:42:27] <Lerneaen_Hydra> oh, ok
[17:42:42] <Lerneaen_Hydra> so that is not the reason emc doesn't start
[17:42:53] <jepler> $ bin/axis-remote --help
[17:42:54] <jepler> axis-remote: cause AXIS to open, reload its opened file, or exit
[17:42:54] <jepler> Usage: axis-remote [--ping|--reload|--quit|filename]
[17:42:54] <jepler> axis-remote [-p|-r|-q]
[17:43:06] <jepler> no, it's not -- it only runs axis-remote once it's shutting down
[17:43:14] <Lerneaen_Hydra> ok
[17:43:23] <Lerneaen_Hydra> is this a new feature?
[17:44:26] <jepler> axis-remote was added only recently
[17:45:21] <jepler> looks like it was added back on 7/31
[18:03:55] <Lerneaen_Hydra> oh
[18:04:14] <Lerneaen_Hydra> jepler: when should I execute the install script?
[18:04:25] <Lerneaen_Hydra> before/after compile?
[18:05:49] <cradek> hi all
[18:05:57] <Lerneaen_Hydra> 'lo cradek
[18:09:22] <jepler> Lerneaen_Hydra: if you're building axis manually, then you run the 'setup.py install' after you build emc2. but there's no reason to be building it manually, just put the 'axis' directory inside the 'emc2/src' directory and have it built every time.
[18:09:38] <Lerneaen_Hydra> thats strange
[18:09:44] <Lerneaen_Hydra> it is in emc2/src/axis
[18:09:58] <Lerneaen_Hydra> I have a cvs+compile script
[18:10:08] <CIA-12> 03jepler 07HEAD * 10emc2/src/hal/utils/halcmd.c: fix bug: 'all' always shows as a completion for unloadXXX
[18:12:03] <Lerneaen_Hydra> that script does:
[18:12:04] <Lerneaen_Hydra> #! /bin/sh
[18:12:06] <Lerneaen_Hydra> cd /home/cnc-lathe/emc2
[18:12:07] <Lerneaen_Hydra> cvs up -dP
[18:12:09] <Lerneaen_Hydra> cd src/axis
[18:12:10] <Lerneaen_Hydra> cvs up -dP
[18:12:12] <Lerneaen_Hydra> cd ..
[18:12:13] <Lerneaen_Hydra> ./configure --enable-run-in-place
[18:12:15] <Lerneaen_Hydra> make && sudo make setiud
[18:14:11] <jepler> does your setup.py refer to axis-remote?
[18:14:11] <jepler> $ grep axis-remote src/axis/setup.py
[18:14:11] <jepler> TERMINAL('axis-remote'): 'scripts/axis-remote.py',
[18:15:32] <Lerneaen_Hydra> TERMINAL('axis-remote'): 'scripts/axis-remote.py',
[18:16:10] <jepler> but you have no bin/axis-remote
[18:16:36] <Lerneaen_Hydra> apparently not...
[18:16:46] <jepler> does configure say this?
[18:16:46] <jepler> checking for AXIS source... will build axis from axis
[18:17:00] <Lerneaen_Hydra> uh.
[18:17:06] <Lerneaen_Hydra> I'll have to recompile
[18:17:09] <Lerneaen_Hydra> >.<
[18:17:12] <Lerneaen_Hydra> just a sec
[18:17:22] <Lerneaen_Hydra> did the script code look correct?
[18:17:37] <jepler> yes I didn't spot any errors
[18:19:39] <CIA-12> 03jepler 07HEAD * 10emc2/src/ (configure configure.in): avoid using 'find' when the realtime config script is in one of the obvious places (but still fall back to the 'find' if necessary)
[18:21:24] <Lerneaen_Hydra> yes it says it will build axis from source
[18:21:29] <cradek> cool, that was so silly
[18:22:37] <jepler> Lerneaen_Hydra: then I'm stumped as to why you don't get bin/axis-remote
[18:23:04] <jepler> cradek: do you have bin/axis-remote on your system with recent axis cvs?
[18:24:33] <Lerneaen_Hydra> everything works fine up till make, then I get some errors/warnings
[18:25:55] <cradek> no, let me up and build
[18:26:47] <Lerneaen_Hydra> those warnings being: directly after starting make: make: failed to remake makefile 'makefile'.
[18:26:49] <Lerneaen_Hydra> classic ladder warnings
[18:26:50] <Lerneaen_Hydra> halcmd warning
[18:27:06] <CIA-12> 03jepler 07HEAD * 10emc2/src/ (configure configure.in): detect common error: more than one axis source directory
[18:27:22] <cradek> jepler: works for me
[18:27:30] <jepler> if 'make' isn't finishing, then it's no use trying to run emc
[18:27:52] <Lerneaen_Hydra> jepler: make finishes, although with warnings
[18:27:56] <jepler> 13:24:01 <Lerneaen_Hydra> make && sudo make setiud
[18:28:00] <jepler> i notice there's a typo here
[18:28:03] <jepler> "setuid"
[18:28:19] <cradek> ;} { (exit 1); exit 1; }; }
[18:28:22] <cradek> wtf?
[18:28:28] <Lerneaen_Hydra> there's no typo in the script
[18:28:34] <Lerneaen_Hydra> I was just reading over vnc
[18:28:40] <Lerneaen_Hydra> and typing here
[18:28:41] <jepler> ok
[18:28:44] <cradek> oh sorry, that's generated
[18:29:54] <jepler> 'failed to remake makefile' is probably an error, even if 'make' continues
[18:30:05] <jepler> please pastebin the full output of 'make' and let's look at it
[18:31:06] <CIA-12> 03jepler 07HEAD * 10emc2/src/ (configure configure.in): apparently calling AC_OUTPUT multiple times is bad (it keeps recreating the 'output files' once for each call)
[18:34:19] <Lerneaen_Hydra> pastebin.ca/117162
[18:36:09] <jepler> huh, that's the very first line?
[18:36:21] <Lerneaen_Hydra> yes
[18:36:32] <Lerneaen_Hydra> you can see the bash line prior to that, right?
[18:37:13] <cradek> Makefile: $(DEPS)
[18:37:27] <cradek> I guess that means it can't generate deps
[18:37:41] <cradek> but I'd expect an error before that line
[18:37:52] <CIA-12> 03jepler 07HEAD * 10emc2/src/Makefile: refer to the top directory in a better way when building axis
[18:37:57] <cradek> maybe make clean will fix it
[18:38:50] <robin_sz> try "make dinner" ;)
[18:39:30] <Lerneaen_Hydra> so make clean then a start from scratch?
[18:39:43] <cradek> can't hurt
[18:41:12] <alex_joni> cradek: I seem to recall that he got an error about emc.nml before that
[18:41:23] <alex_joni> I suggested to copy the file by hand
[18:41:33] <cradek> copy what file?
[18:41:42] <alex_joni> emc.nml#
[18:41:53] <jepler> 11:41:57 <Lerneaen_Hydra2> make: *** No rule to make target `../configs/dallur-thc/emc.nml', needed by `configs'. Stop.
[18:42:16] <alex_joni> jepler: that's the one
[18:43:57] <Lerneaen_Hydra> before that my script didn't have -dP in the cvs up command
[18:44:02] <Lerneaen_Hydra> so that may not be related
[18:44:13] <cradek> ah
[18:44:16] <cradek> that would explain it
[18:44:18] <alex_joni> right.. it probably didn't pull the dallur dir
[18:44:22] <alex_joni> as that's newer
[18:46:58] <Lerneaen_Hydra> there doesn't appear to be any errors any more
[18:47:15] <alex_joni> Lerneaen_Hydra: what config file are you using?
[18:47:24] <Lerneaen_Hydra> it's still compiling
[18:47:30] <alex_joni> can you make a new one based on one of the sample ones?
[18:47:34] <Lerneaen_Hydra> make didn't bomb out this time though
[18:47:39] <Lerneaen_Hydra> I've already done that
[18:47:46] <alex_joni> I suspect your ini still is in the old format (BASE_PERIOD = 0.00050)
[18:47:48] <Lerneaen_Hydra> all the .hal files are the "old" ones though
[18:48:00] <Lerneaen_Hydra> I don't think I've ever had one that old
[18:48:08] <Lerneaen_Hydra> all the ones I've had are in nS
[18:48:13] <alex_joni> OK
[18:51:05] <Lerneaen_Hydra> compile seems to be taking longer than usual
[18:51:17] <Lerneaen_Hydra> tons of CC <path>
[18:51:27] <Lerneaen_Hydra> and LD [M] <path>
[18:51:46] <alex_joni> Lerneaen_Hydra: that's normal
[18:51:54] <Lerneaen_Hydra> ok
[18:52:21] <Lerneaen_Hydra> uhoh
[18:52:37] <Lerneaen_Hydra> cradek had better stay for a while
[18:53:12] <Lerneaen_Hydra> still errors :(
[18:53:19] <Lerneaen_Hydra> though of a different variety
[18:53:27] <alex_joni> pastebin them
[18:53:36] <alex_joni> and cradek might still be around
[18:54:22] <cradek> wonder who tripped on chanserv's power cord
[18:54:36] <Lerneaen_Hydra> pastebin.ca/117194
[18:55:00] <jepler> you need to load the kinematics module
[18:55:14] <Lerneaen_Hydra> prior to the compile?
[18:55:20] <jepler> $ grep trivk src/emc2/configs/sim/core_sim.hal
[18:55:20] <jepler> loadrt trivkins
[18:55:23] <jepler> in your hal file
[18:55:29] <Lerneaen_Hydra> oh
[18:55:34] <alex_joni> cradek: +o ?
[18:55:35] <Lerneaen_Hydra> is kins in hal now? :D
[18:56:29] <jepler> kins is a separate kernel module, and it's loaded by halcmd loadrt
[18:56:49] <jepler> it's not exactly HAL, the axis and joint locations don't move through hal pins
[18:57:09] <Lerneaen_Hydra> well
[18:57:19] <Lerneaen_Hydra> uh, is it moved to somewhere after compile?
[18:59:14] <NickServ> This nickname is owned by someone else
[18:59:14] <NickServ> If this is your nickname, type /msg NickServ IDENTIFY <password>
[18:59:20] <Lerneaen_Hydra> wb
[19:00:05] <Lerneaen_Hydra> which hal files have been changed recently?
[19:08:09] <jepler> http://cvs.linuxcnc.org/cvs/emc2/configs/common/core_sim.hal?graph=1
[19:15:09] <Lerneaen_Hydra> hmm
[19:15:17] <Lerneaen_Hydra> looks like there's been lots of change
[19:15:29] <Lerneaen_Hydra> it didn't like my iocontrol.0.spindle-on
[19:25:40] <Lerneaen_Hydra> * Lerneaen_Hydra pokes jepler / cradek / alex_joni
[19:25:59] <cradek> spindle control was moved to motion
[19:26:07] <Lerneaen_Hydra> oh
[19:26:12] <cradek> I think it's motion.spindle-on now
[19:26:18] <cradek> 'bin/halcmd show pin motion' to be sure
[19:28:20] <Lerneaen_Hydra> motion.spindle-on
[19:28:28] <Lerneaen_Hydra> so the .0 was dropped
[19:28:36] <Lerneaen_Hydra> what if you have multiple spindles?
[19:29:48] <alex_joni> Lerneaen_Hydra: it's not supported by the g-code right now
[19:30:37] <Lerneaen_Hydra> ok
[19:43:32] <Lerneaen_Hydra> woo!
[19:43:35] <Lerneaen_Hydra> FO wheel works now
[19:43:40] <alex_joni> woo.. nice
[19:43:47] <Lerneaen_Hydra> much more RT feeling than a mouse too
[19:43:55] <Lerneaen_Hydra> not to mention easier to handle
[19:44:07] <alex_joni> * alex_joni is happy
[19:44:30] <Lerneaen_Hydra> I can't imagine not having jogwheels
[19:44:51] <Lerneaen_Hydra> heh, one thing that could be done (maybe) is having the mouse wheel as a jogwheel
[19:44:59] <Lerneaen_Hydra> in programs it would control feed
[19:45:14] <Lerneaen_Hydra> and when in manual control you can choose between X and Z
[19:45:27] <Lerneaen_Hydra> or whichever axes you have
[19:45:35] <Lerneaen_Hydra> resolution may be an issue though
[19:45:57] <alex_joni> I think the number of turns is an issue
[19:46:14] <alex_joni> but if it would be logarithmic it would be OK
[19:48:55] <Lerneaen_Hydra> yes
[19:49:04] <Lerneaen_Hydra> at 0.05 it's about 1.5 turns
[19:49:20] <Lerneaen_Hydra> and 0.05 is a bit much when going from 0% to 5%
[19:49:32] <Lerneaen_Hydra> though between 85% and 90% just about right
[19:51:40] <robin_sz> hmm ...
[19:51:57] <robin_sz> link X and Y to the mouse ... Z to the jog wheel ??
[19:52:08] <robin_sz> now htat would be kewl :)
[19:57:00] <jepler> that's done in one of the files linked from here:
http://axis.unpy.net/01154705314
[20:00:36] <robin_sz> * robin_sz looks
[20:00:55] <robin_sz> right
[20:01:17] <robin_sz> oh, while I'm on ... anyone know about cleaning stainless after welding?
[20:01:25] <robin_sz> normally ots done with a paste
[20:01:40] <robin_sz> phosphoric and hydrochloric mix usually
[20:02:10] <robin_sz> but ive seen some neat invertor powwered electrical things that do it with some sort of pad and a bit of DC current
[20:02:34] <robin_sz> I just didnt fancy the £1000+ price tag
[20:15:06] <alex_joni> robin_sz: should be cheaper
[20:15:10] <alex_joni> but paste is better
[20:15:24] <alex_joni> and you can get away with using lots of shielding gas
[20:15:30] <alex_joni> especially for root-backing
[20:20:05] <alex_joni> does www.bhphotovideo.com work for any of you?
[20:20:29] <Jymmm> bh == Big Hooters
[20:22:20] <Jymmm> alex_joni no, it doens't.
[20:23:05] <alex_joni> thx
[20:23:59] <Jymmm> cradek what mill do you have again?
[20:24:22] <alex_joni> max-nc ?
[20:24:26] <cradek> yes it started as a maxnc
[20:24:30] <alex_joni> it's on timeguy.com
[20:25:08] <Jymmm> I'm just trying to see the XY travel on sherline, taig, and maxnc
[20:25:30] <cradek> maxnc has a 4" wide table, I think sherline is 3"
[20:25:30] <alex_joni> sherline is very small
[20:28:42] <Jymmm> They came so close to actually having something here...
http://www.maxnc.com/page14.html
[20:30:53] <Jymmm> Ok, I think they are reaching here a bit...
http://www.maxnc.com/page11.html
[20:35:45] <Jymmm> cradek: I know you're interested in a vac table for making PCB's but why havnet' you?
[20:36:22] <cradek> recently worked out a design, but haven't tried it yet, too many projects
[20:36:43] <Jymmm> heh, I know that one.
[20:38:59] <CIA-12> 03jmkasunich 07HEAD * 10emc2/directory.map: new directory and map, beginning of moving docs to the main CVS module
[20:39:06] <cradek> yay
[20:39:46] <jmkasunich> I changed directory.map... I show docs/src/gcode, docs/src/gui, and docs/src/hal
[20:40:02] <jmkasunich> I'm creating /hal now, but the names of the others are subject to change
[20:40:06] <cradek> that sounds great
[20:40:18] <jmkasunich> maybe /gcode should be /rs274ngc?
[20:40:34] <cradek> no opinion
[20:40:46] <jmkasunich> others might have one, thats why I posted it here
[20:43:38] <jmkasunich> I just realized one nice benefit of putting images in the same dir as the corresponding lyx files... the paths in the lyx files aren't as messy
[20:43:44] <jmkasunich> no ../../images/foo
[20:43:54] <cradek> that'll be nice
[20:43:59] <cradek> that always seemed to get messed up
[20:44:40] <Jymmm> or put ALL images in say ./images/ =)
[20:44:51] <jmkasunich> thats the way it was before
[20:45:01] <Jymmm> ah, ok.
[20:45:10] <jmkasunich> sort of
[20:45:33] <Jymmm> Hey, whatever works and easy to maintain.
[20:45:34] <jmkasunich> originally, the lyx and the images were in two dirs at the same level
[20:45:47] <jmkasunich> so in the lyx you'd write ../images/foo to include an image
[20:46:06] <jmkasunich> then we moved the lyx files for emc2 down a layer, but the images stayed where they were
[20:46:12] <jmkasunich> so it was ../../images/foo
[20:46:27] <jmkasunich> now images will be in the same dir as the lyx, so it will be just "foo"
[20:46:49] <Jymmm> that's why I said ./iamges and not ../images
[20:46:58] <Jymmm> (less the typos =)
[20:47:13] <skunkworks> cradek: is it bad to mention I had a dream about you last night.... You used emc to control a large dish for seti.. very random. worked very well :)
[20:47:23] <cradek> haha
[20:47:39] <cradek> if I was wearing clothes, I don't mind
[20:47:52] <Jymmm> cradek: No, you were the dish!
[20:48:07] <Jymmm> cradek just trying to get a sun/moon tan
[20:48:17] <cradek> emc with kins could really control a telescope
[20:48:18] <skunkworks> ha ha. -- yes - you where fully clothed
[20:48:32] <cradek> ok good
[20:48:49] <Jymmm> would wearing a lion cloth be considered "fully clothed" ?
[20:49:01] <Jymmm> (sp)
[20:49:06] <cradek> Jymmm: the joke was over a bit ago
[20:50:28] <Jymmm> cradek Come on now, just because you're invisioning yourself in a loincloth doesn't mean it's over
[20:51:31] <skunkworks> I also had a epic dream about aliens invading. I think I drank a little too much last night. or it was the pizza
[20:51:49] <Jymmm> skunkworks what were you drinking? drain opener?
[20:51:53] <skunkworks> seemed to go on forever
[20:51:57] <cradek> too much pizza and beer can sure do that
[20:53:34] <skunkworks> I wonder if emc could be used run a telescope - with the objects mapped in 3d space.
[20:55:56] <Jymmm> but you could use a gps with it
[20:58:23] <Jymmm> there is a program already that tracks constellations, bet you could use it's output to guide the scope
[21:04:55] <jmkasunich> there is software out there (for dos/doze) to run telescopes
[21:05:08] <jmkasunich> it microsteps steppers using PWM thru a parport
[21:05:27] <jmkasunich> google "mel bartels"
[21:06:46] <cradek> wow, software microstepping would be cool
[21:07:55] <jmkasunich> I think he uses unipolar drive, 4 transistors per motor, two motors
[21:07:59] <jmkasunich> altitude and azimuth
[21:08:44] <jmkasunich> low supply voltage, he's not trying to get high rapid rates like we use for machining
[21:08:52] <cradek> sure
[21:09:25] <cradek> is this more for tracking than locating, or both?
[21:09:26] <jmkasunich> the program not only handles finding objects, it also tracks them as the earth rotates
[21:09:33] <Jymmm> http://www.bbastrodesigns.com/BBAstroDesigns.html#Computer_Operated_Telescopes
[21:10:01] <jmkasunich> thats the one
[21:10:19] <jmkasunich> its gotten fancy lately
[21:10:28] <Jymmm> released gpl , and source at the bottom
[21:10:33] <jmkasunich> (I was into telescopes for a while, but that was pre-2000)
[21:10:37] <cradek> nice
[21:10:55] <jmkasunich> I eventually figured out that I'd rather build things than look thru them, telescopes were just an interesting thing to build
[21:11:36] <jmkasunich> I don't see GPL, I see a EULA
[21:11:46] <Jymmm> VERY bottom of that page
[21:12:05] <jmkasunich> ok, the stepper stuff is GPL
[21:12:18] <jmkasunich> I guess the EULA is for the servo stuff
[21:13:45] <Jymmm> Chris Rowland has developed GPS software to connect scope.exe to GPS units via a serial port. See
http://groups.yahoo.com/group/scope-drive/files/Other_Software/gpsscope_v2.zip
[21:13:47] <robin_sz> jmkasunich, i have a nice 9.5" mirror here, just waiting for me to design a tube for it
[21:14:32] <robin_sz> must buy the elliptical flat for it one day soon
[21:14:45] <jmkasunich> I've never had a nice mirror
[21:14:52] <jmkasunich> not at all interested in grinding
[21:14:56] <jmkasunich> and too cheap to buy
[21:15:08] <robin_sz> this one is supposed to be lambda/10
[21:15:10] <jmkasunich> the one scope I built used a "replicated" 4" f/4 mirror
[21:16:01] <jmkasunich> cheap, claimed to be 1/4 wave, don't think it actually was
[21:16:10] <robin_sz> heh
[21:16:36] <robin_sz> mine needs re-silvering, but that can wait till I build the tube
[21:16:49] <robin_sz> or I figure out how to solver mirrors :)
[21:17:14] <alex_joni> robin_sz: I've seen a guy on the emc list that did his own mirrors
[21:17:17] <alex_joni> maybe you can ask
[21:17:25] <robin_sz> interesting ...
[21:18:11] <robin_sz> I have a vague idea ... its not so hard, but needs a good vacuum ... roughing pump followed by an oil diffusion pump and a cold trap
[21:18:51] <robin_sz> you probably do need to be a bit mad to try it though :)
[21:19:33] <alex_joni> robin_sz:
http://en.wikipedia.org/wiki/Amateur_telescope_making
[21:19:45] <robin_sz> alex_joni, the paste will work out probably more expensive than the electrical method
[21:19:48] <robin_sz> on the stainless
[21:20:26] <robin_sz> the health and safety and where do you wash the waste paste away to etc makes it very expensive
[21:20:47] <alex_joni> The mirror is aluminized by placing it in a vacuum chamber with electrically-heated nichrome coils that can sublime aluminum. In a vacuum, the hot aluminum atoms travel in straight lines. When they hit the surface of the mirror, they cool and stick. Some mirror coating shops evaporate a layer of quartz on the mirror, others expose it to pure oxygen or air in an oven so that it will form a tough, clear layer of aluminum oxide.
[21:21:13] <CIA-12> 03jmkasunich 07HEAD * 10emc2/directory.map: moved a doc file
[21:21:13] <CIA-12> 03jmkasunich 07HEAD * 10emc2/docs/src/common/GPLD_Copyright.lyx: moved a doc file
[21:21:55] <jmkasunich> some amateurs have set up their own aluminizing chambers
[21:22:37] <jmkasunich> Some have worked. All have resulted in the amateur becoming a bit of a vacuum expert, and spending far more time than they expected
[21:22:59] <Lerneaen_Hydra> jmkasunich: sounds like most hobbies ;)
[21:23:01] <jmkasunich> IOW, its like building your own stepper drive - you don't do it to save money ;-)
[21:23:33] <jmkasunich> Lerneaen_Hydra: true, but aluminizing a mirror more so than many
[21:24:18] <robin_sz> true enough
[21:24:29] <robin_sz> aluminizing is only like $40 anyway
[21:24:37] <jmkasunich> aluminizing needs a _lot_ of infrastructure
[21:24:46] <robin_sz> http://www.bbastrodesigns.com/software.html
[21:24:51] <jmkasunich> grinding/polishing a mirror on the other hand, can save you a crapload of money
[21:25:02] <robin_sz> yeah, but grinding is boring :)
[21:25:09] <jmkasunich> the infrastructure and materials are cheap, it just takes time
[21:25:18] <jmkasunich> right, I have no desire to make a mirror
[21:29:23] <robin_sz> a pc controlled scope .. now that might appeal
[21:30:12] <jmkasunich> I recently acquired a 60mm spotting scope for shooting, 18-36x... I might make a mount for that
[21:32:17] <skunkworks> wow - nice links - thanks
[21:33:02] <skunkworks> a coworker just got an 8 inch celestron - with go-to. smitt casagran or how ever you spell that.
[21:33:27] <jmkasunich> much bux there
[21:33:33] <skunkworks> got me wanting to make one - but I have to do everything big - thinking of 12 or so.
[21:33:43] <skunkworks> yes I would not buy one new.
[21:33:58] <skunkworks> really want to grind one. my dad has when he was a kid.
[21:34:05] <skunkworks> just need some time :)
[21:34:38] <skunkworks> another hobby :)
[21:36:45] <CIA-12> 03alex_joni 07HEAD * 10documents/lyx/emc2/ (Master_HAL.lyx hal_halui.lyx hal_motion.lyx): saved for lyx 1.3, seems the 1.4 can't be opened by older lyxes
[21:44:49] <Lerneaen_Hydra> 'night folks
[21:50:11] <CIA-12> 03alex_joni 07HEAD * 10emc2/docs/src/Master_User.lyx: brought Master_User from the documents/ module. Check there for older commit logs. Updated links to various planned subdirs
[21:55:04] <robin_sz> I did some research and its not dark enough where I am for anything much over 9" really
[21:55:10] <robin_sz> Id just see more light pollution
[21:55:53] <jmkasunich> where I live it isn't dark enough for a 4"
[21:56:50] <jmkasunich> from my back yard you can barely see the Pleidies (or however you spell it - the Seven Sisters)
[21:57:24] <alex_joni> in my home town you can barely see many stars.. (still can somehow count them :)
[22:07:57] <davidf> Hi all
[22:13:17] <alex_joni> hi davidf
[22:13:26] <gene> Hi all
[22:14:02] <gene> Question re thursdays HEAD, is backlash comp now working?
[22:14:47] <davidf> Hi alex_joni
[22:15:05] <davidf> I thought I'd post some results of troubleshooting my system. I learned a few significant things, but am still having the same old intermittent problem.
[22:16:22] <cradek> gene: I don't think backlash comp works well on steppers right now
[22:16:37] <cradek> davidf: sorry to hear that
[22:17:48] <davidf> boy, it's really wierd. I don't want to be a pain, so if you guys care to give it any more thought, please feel free, but if not, likewise.
[22:18:06] <alex_joni> davidf: what's the issue?
[22:18:07] <davidf> Anyway, here's what I have so far.
[22:18:46] <davidf> The stalling happens on all 3 axes. It IS random, nothing to do with Vel, acc, or trajectory, nor drawing.
[22:19:36] <davidf> (2) I put the scope on the input signal to the drive (Par Port pin, and also the inverter chip I buffer it with.)
[22:19:56] <davidf> The input signal is NORMAL even while the stall is occurring.
[22:20:26] <davidf> So that was a significant finding. We can forget about software settings, computer speed, etc.
[22:20:44] <davidf> It is somewhere between there and the motor.
[22:20:45] <alex_joni> ok, so it's probably a stepper/drive thing
[22:20:51] <davidf> Yes.
[22:20:52] <alex_joni> can you rule out mechanical?
[22:21:13] <davidf> I think so with 99 % confidence.
[22:21:23] <alex_joni> I would put the motors on the desk, and let them spin for a while
[22:21:31] <alex_joni> see if it still happens
[22:21:38] <davidf> I paused the program during a stall, and disconnected power to the drive.
[22:21:45] <alex_joni> but I suspect power loss
[22:22:01] <davidf> Then checked to see that the shaft / leadscrew was free. It was.
[22:22:03] <alex_joni> maybe your PSU is not strong enough? and when they all 3 drain power, it might happen
[22:22:14] <alex_joni> what PSU do you have?
[22:22:27] <alex_joni> can you put an RMS ampmeter on it?
[22:22:33] <davidf> Also it often happens that both X and Y will stall at the same time, so that pretty well rules out a bind.
[22:22:57] <alex_joni> sounds more like powerloss to me
[22:23:18] <cradek> yes that is an interesting twist
[22:23:26] <davidf> No, don't have one, but I checked the drive signal at the motor, on the scop.
[22:23:38] <davidf> scope.
[22:23:44] <alex_joni> you could put a shunt in series with the output of the PSU
[22:23:46] <alex_joni> and put the scope there
[22:24:14] <davidf> Saw 50 volt chopped square wave, as expected....
[22:24:24] <davidf> Hang with me a sec alex_joni ...
[22:24:34] <alex_joni> * alex_joni hangs
[22:24:40] <davidf> Good idea, let me tell you the rest on this first...
[22:25:16] <davidf> Anyway, During the stall, instead of seeing a series of chopped steps...
[22:26:34] <davidf> I saw a series of single 'chops' that is, the short spikes you see in the steps, but just one after another, separated by (i don't know how much, but not close together as normal.)
[22:26:45] <alex_joni> aha
[22:26:52] <alex_joni> what kind of PSU is this?
[22:26:57] <davidf> like oscillation in the driver
[22:27:15] <davidf> Which psu are you refering to?
[22:27:24] <davidf> You mean the drive unit?
[22:27:26] <alex_joni> the powersupply for the stepper drives
[22:27:27] <alex_joni> yes
[22:27:54] <davidf> Ok, I'm using Parker-Hanifin stepper drives...
[22:28:02] <davidf> Zeta4
[22:28:25] <davidf> They Use full 120 Volts for the PS, no Transformer.
[22:28:32] <alex_joni> ahh.. I see
[22:29:00] <alex_joni> darn.. not much you can do then
[22:29:04] <Jymmm> nasty
[22:29:13] <davidf> However, they are normally used with Parker Motors, that are very high inductance.
[22:30:14] <davidf> I tried setting the dip switches for the lowest Z range. It seemed it might have helped. Got the longest run so far without a stall.
[22:30:28] <davidf> BUT, still got it unfortunately.
[22:30:56] <davidf> And both X and Y stalled at the same time, then Y continued, and then X.
[22:31:14] <alex_joni> davidf: I know this might be farfetched.. do you have a signal generator?
[22:31:26] <davidf> Of sorts.
[22:31:44] <alex_joni> rectangular generator to drive one of the motors
[22:31:48] <alex_joni> then just leave it spinning
[22:31:53] <davidf> I have a 1, 1000, and 100K square on my Proto PS.
[22:32:00] <alex_joni> 1000 might be OK
[22:32:19] <alex_joni> probably you'll want to disconnect the motor from the coupling ;)
[22:32:45] <davidf> Yeah, I thought of that, but I does seem to be the drives or some outside power (AC supply) prob.
[22:32:57] <davidf> Ha.
[22:33:01] <alex_joni> do you happen to have an UPS handy?
[22:33:19] <davidf> You mean dont just let it bash & go eat dinner?
[22:33:27] <davidf> NO.
[22:33:32] <alex_joni> too bad ;)
[22:33:40] <davidf> Yeah.
[22:33:41] <alex_joni> that would have been worth a try
[22:33:50] <davidf> Ya know what I thought of?
[22:34:05] <alex_joni> might be the time to get one :D (they're not that expensive, a small one might be ~100$)
[22:34:23] <gene> cradek, I just tried it on two axis's, and it seems to be working rather nicely,
[22:34:37] <davidf> What if it is like, power glitches from the city power being overloaded by all the heat & Air conditioners etsc.
[22:34:47] <gene> Now if this $##@ micromill was predictable....
[22:34:57] <alex_joni> davidf: that's why I asked about the UPS
[22:35:31] <alex_joni> an UPS also does some very nice filtering
[22:35:35] <davidf> Also there is a power management program here, where the city can turn off Air Cond's that are equipped with a renmote breaker.
[22:35:59] <davidf> I wonder if they send some High-Hz signal over the power lines?
[22:36:11] <alex_joni> there are some phase-filters for AC mains
[22:36:19] <davidf> Right alex_joni I'm with you.
[22:36:25] <alex_joni> and you could use an 1:1 separation transformer
[22:36:38] <alex_joni> those usually cut-out high frequency
[22:36:39] <skunkworks> I had issues with parker drives when I was over powering the steppers.
[22:36:47] <davidf> Yeah, but a BF one.
[22:36:59] <alex_joni> davidf: what current do the motors need?
[22:37:11] <skunkworks> I didn't know what size the motors where but after I cut the current in half - no more issues.
[22:37:18] <davidf> Oh? Now you have my attention :)
[22:37:19] <skunkworks> (motors had no info on them_
[22:37:56] <davidf> I cut power from 2 to 1.25 amps, no help, but I could try further.
[22:38:04] <skunkworks> these where s6 drives. I had assumed the motor would take 6 amps (pretty big) but I assumed wronge
[22:38:11] <alex_joni> do they still spin OK?
[22:38:12] <robin_sz> * robin_sz boggles at all the talk of interference and motors ...
[22:38:33] <davidf> It does make sense. If the drives need higher Z, lowering the current might help reduce ringing etc.
[22:38:52] <alex_joni> davidf: cut power till they stop spinning OK, then increase to the next step
[22:39:15] <davidf> ok I'll try that.
[22:39:25] <davidf> robin_sz, whatcha mean?
[22:39:29] <robin_sz> guys ... wanna make some estimates of the -3db roll of point of a transformer and a 10,000uf capacitor? its going to be sub 5hz .. so mains born noise just isnt going to get through
[22:40:15] <alex_joni> if there is a transformer
[22:40:16] <davidf> No... but there is no Xformer in these drives.
[22:40:27] <robin_sz> oh, they are mains in?
[22:40:29] <robin_sz> coo.
[22:40:37] <davidf> They use 120 VAC direct, rectified.
[22:41:06] <alex_joni> robin_sz: btw, the new robots use 600V driver
[22:41:08] <alex_joni> robin_sz: btw, the new robots use 600V drives
[22:41:11] <alex_joni> and motors
[22:41:19] <alex_joni> because of the 400V rectified ;)
[22:41:22] <robin_sz> yes, but thats servos
[22:41:32] <alex_joni> yup, just pointing it out
[22:41:37] <alex_joni> cuts costs a LOT
[22:41:51] <alex_joni> smaller motors (less current at the same power)
[22:41:54] <alex_joni> smaller cables
[22:41:56] <alex_joni> etc
[22:41:57] <robin_sz> this I know ... modern servos are high voltage, low mass
[22:42:01] <robin_sz> * robin_sz nods
[22:42:14] <robin_sz> steppers are another matter entirely
[22:42:43] <davidf> The drives say 90 VAC minimum.
[22:43:01] <robin_sz> so probably 140V dc or so?
[22:43:27] <robin_sz> so that implies around 7V drop on the coils for running current
[22:43:29] <davidf> But I wonder if I could use DC from Series car Batteries at the equivalent rms?
[22:43:54] <robin_sz> those must be old motors ...
[22:44:19] <davidf> The current setting of 1.25 Amps shows a 50 V peak chopp.
[22:44:34] <davidf> Which... Mine?
[22:44:46] <davidf> motors I mean.
[22:44:50] <robin_sz> if they are driven off 140V rails, yes
[22:45:16] <robin_sz> must be old, you couldnt drive modern motors with those sort of voltages
[22:45:19] <davidf> No, They are new. Like 3 volt DC or something like that.
[22:45:33] <robin_sz> oh,
[22:45:42] <davidf> The drives chop the PS to get the proper current.
[22:45:54] <robin_sz> this I know
[22:46:09] <davidf> As you are probably familiar with in general.
[22:46:12] <robin_sz> but normally you restrict the rail voltage to around 20 x dc drop, or they overheat
[22:46:16] <davidf> right. I know.
[22:46:36] <robin_sz> the rectified 120 AC is the weird part
[22:46:37] <davidf> Oh, no I didn't know that.
[22:46:47] <davidf> Why?
[22:46:54] <alex_joni> maybe it chops of the AC, then rectifies ?
[22:47:00] <robin_sz> dunno
[22:47:19] <alex_joni> that way I can imagine how they produce 50V chops
[22:47:25] <robin_sz> * robin_sz nods
[22:47:39] <davidf> Maybe. All I know is the wave form on the scope looks reasonable, and the motors don't get too hot with reasonable duty.
[22:47:48] <robin_sz> fair enough
[22:48:08] <robin_sz> actually, most motors are desinge dto run hot anyway
[22:48:16] <robin_sz> 100C is actually normal
[22:48:49] <davidf> right. These judging by feel are at around 150 F after 10 minutes continuous
[22:48:58] <robin_sz> sounds fine
[22:49:01] <davidf> so not bad.
[22:49:21] <alex_joni> robin_sz: I love the new motors on our robots
[22:49:22] <alex_joni> http://www.robcon.ro/new/images/upload/r320m.gif
[22:49:28] <alex_joni> look at axis 1 & 2
[22:49:37] <davidf> And, hot or not, this darn phenomenon keeps rearing it's ugly head.
[22:49:55] <robin_sz> nice
[22:49:57] <alex_joni> they are like a box, about 10x10x30cm
[22:49:58] <robin_sz> small huh
[22:50:01] <robin_sz> yeah
[22:50:04] <alex_joni> 3.2kW
[22:50:08] <alex_joni> * alex_joni grins
[22:50:09] <robin_sz> coo
[22:50:11] <davidf> But so seldom, you sometimes have to wait 10 minutes to see it, and it last ony 1/2 to 1 second.
[22:50:18] <alex_joni> resolver and brake inside
[22:51:09] <robin_sz> davidf, are you really really sure its not a pulses problem. I have seen this on wondows systems ... it misses about 1 second of pulses sand then sends them all at once, result all the axes stall
[22:51:22] <alex_joni> robin_sz: no, he scoped it
[22:51:31] <robin_sz> fairy nuff
[22:51:34] <alex_joni> I suggested to put a signal generator on it
[22:51:46] <robin_sz> I never did by my robot welder ...
[22:51:50] <davidf> Hey guys, what does anybody thing of some series resistance in the motor leads?
[22:52:00] <cradek> having pulses during the stall does not mean the stall was not caused by missing or misplaced pulses
[22:52:08] <robin_sz> indeed
[22:52:17] <robin_sz> davidf, no, dont do that
[22:52:28] <davidf> ok...
[22:52:35] <cradek> alex is right that using a signal generator would tell us for sure
[22:52:43] <alex_joni> unless it's really small, and you just want to measure current
[22:52:50] <alex_joni> 0.01 ohm or similar
[22:52:59] <robin_sz> cradek, by scoped it I am assumng he meant observed it carefully inthe moments leading up to a stall
[22:53:32] <robin_sz> alex_joni, thats a cloos right?
[22:53:38] <alex_joni> robin_sz: yes
[22:53:38] <cradek> sounds like it would be pretty hard to be sure of that
[22:53:40] <davidf> cradek, The pulse train was rock solid the same before during and after the stall as far as I could see, but the motor drive waveform was all wrong...
[22:53:51] <davidf> totally different during the stall.
[22:54:12] <alex_joni> is there any drive that has less stalls?
[22:54:18] <robin_sz> well its 3 seperate drives right ... so maybe it the mains browning out
[22:54:20] <alex_joni> Z for example?
[22:54:32] <cradek> so you were able to see nice even step pulses right at the moment of the start of the stall?
[22:54:33] <robin_sz> alex_joni, I never did buy my robot :(
[22:54:47] <alex_joni> robin_sz: I have a second hand one if you care..
[22:54:54] <robin_sz> cheap?
[22:54:56] <alex_joni> actually a customer has one
[22:54:59] <alex_joni> big :D
[22:55:02] <alex_joni> < 10k
[22:55:18] <alex_joni> 7m linear axis, 2 tables turn/tilt
[22:55:24] <davidf> Yes. I was carefully watching the waveform before and throughout the stall.
[22:55:29] <robin_sz> heh
[22:55:45] <robin_sz> thats a big linear axis ...
[22:55:51] <alex_joni> or was it longer?
[22:55:53] <cradek> ok that's great that you were watching at the right time, sounds hard to catch
[22:56:01] <alex_joni> it's a truck full of stuff
[22:56:06] <robin_sz> i bet
[22:56:16] <robin_sz> ex car plant?
[22:56:19] <cradek> can you beg/borrow/steal a UPS?
[22:56:24] <alex_joni> know komatsu?
[22:56:27] <davidf> But at the motor, (the drive output to one phase, the waveform was not even close to what it should have been.
[22:56:34] <robin_sz> alex_joni, yeah
[22:56:49] <davidf> I can try.
[22:56:53] <robin_sz> davidf, well, next time scope the mains
[22:57:04] <alex_joni> robin_sz: used to weld some parts for them
[22:57:17] <davidf> Did that too. yesterday. Saw nothing.
[22:57:20] <alex_joni> davidf: I would do it like this
[22:57:23] <cradek> do you have access to the rectified mains? You could trigger on that, AC-coupled
[22:57:40] <robin_sz> alex_joni, probably would prefer ABB for some reason .. seem more common over here
[22:57:45] <jmkasunich> are the drives getting warm?
[22:57:48] <alex_joni> 1. get one motor out of the mill, put it on the table
[22:57:49] <davidf> But my scope is anchient, so I would not have seen reall high freq. spikes.
[22:57:59] <davidf> sorry about the typos.
[22:58:04] <jmkasunich> maybe some kind of overtemp protection is shutting them down for a moment?
[22:58:08] <alex_joni> 2. let it run on the signal generator for a while (30 mins at least)
[22:58:11] <robin_sz> another thought .. are the drives opto-isolated input?
[22:58:17] <jmkasunich> robin_sz: yes
[22:58:24] <jmkasunich> (I was reading the manual yesterday)
[22:58:34] <robin_sz> ok, thats the other possibility ruled out then :)
[22:58:50] <davidf> Overtemp is out too. It latches off.
[22:58:52] <robin_sz> I was thinking reference volts bounce, hum etc
[22:59:12] <davidf> These are staying on, and buzzing the motors loudly during a stall.
[22:59:22] <robin_sz> indeed,
[22:59:32] <robin_sz> maybe just too damn close to the torque limits
[22:59:38] <jmkasunich> you said you scoped the drive output
[22:59:42] <davidf> Pulses are going out, just not right. The motor just vibrates.
[22:59:48] <jmkasunich> whats it look like again?
[22:59:54] <jmkasunich> 50V PWM?
[23:00:05] <davidf> yes.
[23:00:14] <jmkasunich> what PWM frequency?
[23:00:21] <davidf> not close to torque limit I think.
[23:00:30] <davidf> Hang onn re that...
[23:01:37] <CIA-12> 03alex_joni 07HEAD * 10emc2/docs/src/gui/ (17 files): gui related docs. previously in the documents/ module (check there for older changelogs)
[23:02:09] <davidf> Period is 0.8 * 50 uS.
[23:02:24] <davidf> punching the calc...
[23:02:31] <cradek> heh 40
[23:02:49] <jmkasunich> 40uS = 25KHz
[23:03:09] <davidf> right.
[23:03:13] <jmkasunich> roughly whats the duty cycle when its happy?
[23:03:32] <alex_joni> 25kHz sounds nice (you can't hear it at least ;)
[23:03:45] <davidf> 50%
[23:04:11] <davidf> You know, I do hear something though.
[23:04:27] <davidf> probably a harmonic.
[23:04:38] <jmkasunich> do you hear it only when its moving?
[23:04:57] <davidf> sounds like up around 15 to 20 kHz
[23:05:06] <davidf> no, at rest.
[23:05:58] <davidf> That's strange since theres nothing on the scope. Let me put it on auto trigger.
[23:07:34] <davidf> I can see about 5 V pp 60 Hz, and som very fast spikes.
[23:07:53] <jmkasunich> this is with the drive enabled, but no steps?
[23:07:59] <jmkasunich> the motor is "locked"?
[23:08:24] <jmkasunich> (locked = coils energized, hard to turn shaft)
[23:09:11] <davidf> yes.
[23:09:30] <davidf> There is some high freq. noise,
[23:09:32] <jmkasunich> you don't by any chance have a current probe do you?
[23:09:37] <jmkasunich> (long shot)(
[23:09:46] <davidf> goes from -10 to + 70 V
[23:09:58] <alex_joni> * alex_joni asked about a .01 ohm resistor to shunt it
[23:10:12] <davidf> got a current setting on my digital VOM...
[23:10:16] <jmkasunich> common mode noise makes the resistor really tough to use safely
[23:10:29] <jmkasunich> I was more interested in scoping the current
[23:10:44] <davidf> oh, nope.
[23:10:45] <jmkasunich> these drives (and all microstepping drives) use a closed current loop
[23:11:02] <davidf> mmm...
[23:11:04] <jmkasunich> if something fools the current feedback you can get all kinds of strange results
[23:11:08] <alex_joni> isn't a current probe basicly a shunt resistor?
[23:11:15] <jmkasunich> no, its magnetic
[23:11:21] <alex_joni> oh, those ..
[23:11:27] <alex_joni> hall?
[23:11:34] <robin_sz> sometimes
[23:11:45] <jmkasunich> current transformer, sometimes combined with hall for DC
[23:11:47] <alex_joni> not for DC I guess
[23:11:54] <alex_joni> err.. for
[23:11:54] <robin_sz> more usually they are a ferrite ring and some trickery
[23:12:09] <alex_joni> yeah, played with one once
[23:12:17] <alex_joni> for a plasma app
[23:12:23] <robin_sz> DC clampmeters are not, in my experience, very accurate
[23:12:33] <davidf> Anyway, the noise is constant.
[23:12:52] <jmkasunich> can you speed up the time axis?
[23:12:54] <robin_sz> they spend half the time telling you which way North is :)
[23:13:07] <jmkasunich> so you can see one or two noise pulses?
[23:13:15] <jmkasunich> are they at a constant frequency?
[23:13:26] <davidf> jmkasunich, my scope is very old. 10 Mhz.
[23:13:35] <jmkasunich> oh
[23:13:47] <jmkasunich> cd ..
[23:13:49] <jmkasunich> oops
[23:13:50] <davidf> Cant resole it enough to see anything but at low freq.
[23:14:03] <davidf> low freq sweep.
[23:15:07] <davidf> But the noise is constant, and the problem only intermittent so that would tend to rule it out as source, though not entirely.
[23:15:26] <davidf> Man, what a PITA.
[23:15:35] <alex_joni> davidf: try the signal generator
[23:15:43] <alex_joni> that completely rules out the input signal
[23:15:44] <davidf> OK.
[23:15:54] <alex_joni> then try an UPS if you can borrow one
[23:16:01] <alex_joni> that should rule out mains
[23:16:15] <davidf> Right.
[23:16:21] <alex_joni> after that ..I don't know :)
[23:16:52] <alex_joni> it would be interesting to know if this happens with no load on the motors
[23:17:08] <jmkasunich> how hard is it to decouple the motors from the screws?
[23:17:53] <davidf> Remember this is a micro mill, & I can turn the shafts easily by hand, & it happens at low, med & high speed.
[23:18:17] <davidf> I really dont think it's torque overload.
[23:18:39] <alex_joni> no, but even a little torque might matter
[23:19:04] <davidf> No hard. In fact, last time it happened, I ran the X Leadscrew out of the nut, so it's basically uncoupled now.
[23:19:06] <jmkasunich> we've already ruled out most of the common problems
[23:19:18] <jmkasunich> so now we're looking for uncommon ones
[23:19:24] <jmkasunich> the more things we can eliminate the better
[23:20:22] <davidf> OK... I'll get back in a few.
[23:22:34] <skunkworks> I spent a lot of time with the motors unhooked with a magic marker marking the shaft. back and forth - back and forth.
[23:23:01] <alex_joni> how many steps/ rot?
[23:23:40] <alex_joni> if it's 200, I tried this: make it go forward 200 steps, 199 back; repeat 200 times;)
[23:26:01] <davidf> Well, I found something that seems to help a lot. Everything seems better now.
[23:26:28] <alex_joni> current?
[23:26:31] <davidf> Beer.
[23:26:36] <alex_joni> lol
[23:26:46] <davidf> :)
[23:27:26] <davidf> So you want me to drive the Parker's step input with a 1 kHz signal, with the motor uncoupled?
[23:27:31] <alex_joni> right
[23:27:33] <jmkasunich> yeah
[23:27:41] <jmkasunich> just let it run and run
[23:27:47] <alex_joni> 30 mins ;)
[23:27:57] <davidf> ok. How about with just the leadscrew connected, but out of the nut?
[23:28:07] <jmkasunich> that works
[23:28:14] <davidf> That's easy.
[23:38:27] <CIA-12> 03alex_joni 07HEAD * 10emc2/docs/src/gcode/ (26 files): gcode related docs. previously in the documents/ module (check there for older changelogs)
[23:43:45] <alex_joni> heh.. cradek got me according to CIA stats
[23:43:53] <alex_joni> he has the most commits to emc ;)
[23:55:49] <CIA-12> 03jmkasunich 07HEAD * 10emc2/docs/src/hal/ (31 files): moved hal related lyx and images over, older commit history can be found in the documents module
[23:59:33] <davidf> OK, I have the motor /leadscrew running at the equivalent of 41 in/sec. Driving with a 555 chip at 2777 Hz. = 833 RPM. That's making it pretty noisey. I set it for a resonant freq.
