#emc-devel | Logs for 2008-05-30

[15:57:36] <SWPadnos> alex_joni, it looks like they may have done a perl change at DH - there are a lot more messages printed now when you run the logger
[15:58:11] <alex_joni> well.. if it still works ;)
[15:58:26] <SWPadnos> heh - it seems to be about the same
[16:05:11] <SWPadnos> did you mean pyvcp or vismach?
[16:26:06] <micges> bookmark
[17:04:54] <rayh> Hi cradek. I must have gotten it from trunk. I think that's all I've got here.
[17:06:27] <SWPadnos> hal_input is in the user_comps directory, rather than the user_comps/devices dir
[17:06:58] <SWPadnos> and of course doesn't have the word "joystick" anywhere in the file name :)
[17:08:31] <cradek> dangit, I see we missed that (one) file
[17:08:44] <cradek> jmkasunich: would you fix that license statement when you get a chance please
[17:09:21] <rayh> I believe that both of Roland's mazaks were fanuc CNCs.
[17:09:28] <SWPadnos> ok
[17:09:45] <SWPadnos> was I totally wrong about a CRT being there?
[17:10:00] <rayh> No there is a crt.
[17:10:08] <rayh> Seems like 9 inch.
[17:10:31] <SWPadnos> ok, phew! ;)
[17:11:13] <rayh> You're okay. I'm the one getting oldtimers.
[17:11:44] <SWPadnos> excelent!
[17:11:46] <SWPadnos> +l
[17:11:51] <SWPadnos> (see - it's contagious)
[17:12:21] <rayh> maybe so.
[17:14:12] <SWPadnos> the longer I wait the worse it gets :)
[17:15:30] <rayh> Nice post at cnc-zone. I don't understand any of it but...
[17:15:53] <rayh> Somebody said I could get on IRC and folk would help me.
[17:20:27] <SWPadnos> heh
[17:22:11] <alex_joni> that's just ad-scam
[17:22:45] <rayh> heh
[17:23:45] <SWPadnos> just for the heck of it, I checked the CNCzone ad prices. it looks like $200/month for one of those short banners
[17:24:07] <SWPadnos> there are special spots that are $400/mo, and a cheaper rotating ad thing that's only $50
[17:24:38] <SWPadnos> (thinking back to the Gecko thread on why EMC2 isn't "gaining groundswell")
[17:24:47] <SWPadnos> or however he put it
[17:24:50] <rayh> I find it interesting that folk assume that open source stuff ha as limited reconfigurability as what they've been using.
[17:25:19] <SWPadnos> there's also the assumption that you need to be able to write software to use it effectively
[17:25:23] <rayh> Who said EMC2 isn't "gaining groundswell?"
[17:25:26] <BigJohnT> most people figure if you pay more for something it HAS to be better
[17:25:34] <SWPadnos> even though many many people use Firefox and OpenOffice every day
[17:26:08] <rayh> ground swill -- otherwise known as dirty water.
[17:26:14] <SWPadnos> heh
[17:26:30] <SWPadnos> "We don't muddy our waters by paying for advertising"
[17:26:58] <rayh> There you go.
[17:27:26] <SWPadnos> hmmm - are there separate jog increments for each axis?
[17:27:35] <SWPadnos> (just crious - don't need that feature)
[17:27:38] <rayh> Advertising, we don't need no stinking advertising.
[17:27:39] <SWPadnos> curious
[17:27:42] <SWPadnos> heh
[17:28:03] <rayh> Mini changes increments between rotary and linear.
[17:28:15] <SWPadnos> ok, but not between linear axes
[17:28:20] <rayh> Or it did at one time.
[17:28:22] <rayh> No.
[17:28:44] <SWPadnos> I think the underlying NML messages may allow it (I think there's an increment passed with each JOG message)
[17:28:55] <rayh> That would seem to me to be really confusing.
[17:29:04] <SWPadnos> oh, it is possible with an MPG - there's a separate scale input for each joint
[17:29:04] <rayh> Yes.
[17:29:38] <SWPadnos> just thinking about the idea of 2 MPGs per joint, plus 1 each for FO and SO :)
[17:30:02] <SWPadnos> there's just no "continuous" with an MPG, which is fine
[17:30:36] <rayh> Okay! Okay! (Joe pesci) Question.
[17:30:43] <rayh> two actually
[17:30:46] <SWPadnos> Leo Getz
[17:31:24] <rayh> Yep. First is it possible to drive an axis, using any of our possible routes past the max vel set in the ini?
[17:32:00] <SWPadnos> I was just thinking about that. I think the only way may be backlash comp related
[17:32:11] <SWPadnos> but I'm not sure that's actually a way
[17:32:32] <rayh> Second, does g0 compute the actual tool tip vel at the hypotenuse of the angle between the axes involved.
[17:32:51] <SWPadnos> hmmm. good question
[17:33:16] <rayh> So any of the jog pins in hal honor that max vel?
[17:33:25] <skunkworks> how come when I post a message on the gecko list - no one post for days?
[17:33:32] <alex_joni> rayh: yes
[17:33:49] <alex_joni> the only thing that _might_ exceed ini settings si backlash
[17:33:54] <alex_joni> s/si/is/
[17:33:58] <rayh> Fantastic. That is what I wanted to hear.
[17:34:11] <alex_joni> (because the backlash move will be happening at max the ini settings)
[17:34:16] <rayh> So If I've got 6m backlash I can go pretty fast.
[17:34:28] <alex_joni> and if the axis was already going in that direction, the two will get added
[17:34:31] <rayh> ;)
[17:34:43] <alex_joni> although it's quite uncommon for a backlash move to happen if you were already going in that direction
[17:34:46] <rayh> okay. I understand that.
[17:34:51] <alex_joni> usually backlash kicks in at reversals
[17:34:59] <rayh> Sure.
[17:35:01] <alex_joni> so the joint speed should be low ..
[17:35:27] <rayh> Is accel honored by backlash comp?
[17:35:48] <rayh> Seem to remember something about that quite a while back.
[17:36:02] <alex_joni> yes. but the same limitations as for vel
[17:36:20] <alex_joni> (a while ago - I think around emc2_2.0.x - it wasn't)
[17:36:25] <rayh> Okay.
[17:36:52] <SWPadnos> regarding the G0 question, is the [TRAJ]MAX used as the commanded feed rate, or is the speed the vactor sum of the 3 joints max speeds?
[17:36:59] <SWPadnos> vector
[17:37:12] <rayh> Okay.
[17:37:33] <SWPadnos> (that was a typo correction, not a statement :) )
[17:37:51] <rayh> See BigJohnT was correct. You can get answers on IRC.
[17:38:06] <rayh> Oh. So the answer to the vector is still in the air.
[17:38:33] <rayh> I tried it with a g0 x5 y5 using axis and it said 72.
[17:41:58] <SWPadnos> it appears to honor the axis max vels, but not [TRAJ]MAX_LINEAR_VELOCITY
[17:42:09] <SWPadnos> which may not have been fully implemented, come to think of it
[17:43:24] <SWPadnos> ok, it does honor [TRAJ]MAX_VELOCITY
[17:43:57] <SWPadnos> so you cna have joint limits that would allow faster tooltip motion, but still limit traverse with the TRAJ setting
[17:43:59] <SWPadnos> can
[17:44:47] <BigJohnT> rayh: :)
[17:45:04] <rayh> Yea
[17:52:13] <alex_joni> bbl
[18:50:22] <rayh> SWPadnos, I confirmed that vector speed can be greater than axis max vel if traj max vel allows it.
[18:51:02] <rayh> The only problem I see with that is that the sliders in the gui then use traj and show a value faster than jogs can be.
[18:52:08] <SWPadnos> yes. it could be surprising to see a vel that's sqrt(3) * what you expect
[18:53:00] <cradek> that's why MAX_LINEAR_VELOCITY, MAX_ANGULAR_VELOCITY were made - to set the tops of the two sliders
[18:53:09] <cradek> it's a swirl, as jmkasunich would say
[18:53:23] <BigJohnT> rayh: during a G0 across X&Y I get 525 IPM I don't know what my max is set for but each axis is set for 415IPM...
[18:53:38] <SWPadnos> oh - those are for the GUI only, not for the TP?
[18:53:42] <cradek> yes
[18:54:28] <SWPadnos> maybe they should be walled MAX_*_JOG or something
[18:54:38] <rayh> BigJohnT, In theory on a 45 degree angle you should be able to see 586.
[18:55:07] <cradek> SWPadnos: they should be in [DISPLAY] (I think they are accepted there too)
[18:55:08] <BigJohnT> It might be that I'm going from memory and you know how that is
[18:55:20] <rayh> Don't I though.
[18:55:20] <SWPadnos> yep - I think I remember that :)
[18:57:01] <rayh> Just about finished with testing of the gOS $199 box from newegg.
[18:57:57] <SWPadnos> I didn't have good luck with that board, but I didn't do too much testing either
[18:58:16] <SWPadnos> actually it's the $60 MB+CPU thing that I have
[18:59:48] <skunkworks> rayh: how is it looking?
[19:00:04] <rayh> This mobo says PC2500G but ???
[19:00:17] <SWPadnos> what speed is that one?
[19:00:26] <rayh> Pretty good. Jitter is <20000
[19:00:29] <rayh> 1.5g
[19:00:59] <rayh> The google implementation of enlightenment sucks big time.
[19:01:05] <SWPadnos> heh
[19:01:24] <rayh> Intended only for homes with a dhcp connection.
[19:01:37] <rayh> Cause it's hard to sudo and change to static.
[19:01:50] <rayh> Borked it for 8.04
[19:02:04] <rayh> That seems to be fine with the default Gnome manager.
[19:02:36] <rayh> I'm still twiddling with BIOS but should have numbers up by the end of the day.
[19:03:15] <SWPadnos> I think I wasn't able to get better than ~50k or so, even after stopping X (but only on the liveCD maybe - I don't remember how I tested now)
[19:03:44] <rayh> cradek, Same via v vesa driver issue that I had with 8.04 on the jetway.
[19:03:48] <SWPadnos> if you figure out what works, let me know - I have two boards :)
[19:03:57] <rayh> Okay.
[19:04:31] <rayh> Who makes that board?
[19:05:09] <SWPadnos> I think it's the same one you have - it was the gOS dev board from clubIT
[19:05:22] <rayh> Okay.
[19:05:24] <SWPadnos> 1.5GHz Via C7+CN700
[19:05:50] <SWPadnos> it appears to be made by Via
[19:06:54] <rayh> Same here.
[19:06:57] <SWPadnos> yep, there's a tag that says "ID-PCM7E PC2500 V1.0C"
[19:07:23] <SWPadnos> no G here, but I know it has integrated graphics, which is what that often means
[19:08:01] <rayh> Okay. I'm seeing 13200 about for jitter on both
[19:08:27] <SWPadnos> oh, it'sa PC2500E according to the manual
[19:09:09] <rayh> Right now I'm using the fail-safe BIOS defaults
[19:09:34] <SWPadnos> I was planning on bringing one or both to the workshop. we can experiment there :)
[19:09:53] <rayh> I'm thinking of it for the lab machines.
[19:10:31] <SWPadnos> ok. I'll be there all day Sunday (I think), so we can set up some stuff with mine if you need more PCs
[19:10:45] <rayh> Oh that would be great.
[19:11:26] <SWPadnos> maybe I should write down "power supplies, keyboards, monitors" too :)
[19:12:22] <BigJohnT> don't forget your mouse
[19:12:31] <rayh> The optimized BIOS defaults do a bit better <12000
[19:12:40] <SWPadnos> no way! they can use the terminal!
[19:22:15] <SWPadnos> heh - it would be funny to put those two google boards into these cases (which I have): http://www.rocketfishproducts.com/pc-93-3-rocketfish-aluminum-full-tower-case-black.aspx
[19:22:55] <SWPadnos> time to run some errands. BBL
[21:27:43] <cradek> jmkasunich: would you please fix the license statement in hal_joystick? seems we missed one, back when.
[21:27:47] <cradek> also, hi
[21:27:53] <cradek> I just didn't want to forget
[21:28:02] <jmkasunich> heh
[21:28:45] <jmkasunich> my weekend got an early start
[21:28:48] <jmkasunich> -;)
[21:28:51] <jmkasunich> oops
[21:28:55] <jmkasunich> ;-) even
[21:29:27] <jmkasunich> what is wrong with the license statement?
[21:29:34] <jmkasunich> * License: GPL Version 2.1
[21:29:40] <jmkasunich> is it supposed to be 2.0?
[21:29:53] <cradek> GPL2 and LGPL2.1 are valid choices
[21:29:58] <cradek> there is no GPL2.1
[21:30:20] <cradek> there is no GPL2.0 either
[21:32:03] <CIA-32> EMC: 03jmkasunich 07v2_2_branch * 10emc2/src/hal/user_comps/devices/hal_joystick.c: fix license
[21:33:03] <CIA-32> EMC: 03jmkasunich 07TRUNK * 10emc2/src/hal/user_comps/devices/hal_joystick.c: fix license
[21:33:07] <cradek> thanks
[21:33:11] <jmkasunich> np
[21:33:32] <jmkasunich> I discovered that my ups battery is good for 20ish minutes, maybe a bit more
[21:34:20] <jmkasunich> don't recall exactly when the power died, after 15-20 mins when the rumor mill said it wouldn't be back till 6pm I left work, got home just in time to hear it screaming for relief
[21:34:44] <cradek> 20 is a nice long time
[21:34:48] <jmkasunich> during an outage it goes "beep" <pause> "beep" <pause> etc - as it gets lower the beeps get closer to gether
[21:35:54] <jmkasunich> it was running _only_ the computer and kvm switch, till I got home and moved the mouse - I issued a shutdown command as the screaming got continuous, it was about 90% of the way to "power down" when the UPS cut out
[22:07:43] <jmkasunich> hah, I thought this was a HAL thing at first: http://imagebin.ca/view/Mo_2nUba.html