#emc-devel | Logs for 2009-09-09

[06:31:39] <micges_work> hello
[07:51:12] <alex_joni> hi micges_work
[07:51:19] <alex_joni> sorry about last night, was a bit tied up ;)
[07:51:45] <micges_work> np
[08:11:11] <alex_joni> regarding the NML.. I think it can get overwritten
[08:14:17] <micges_work> do you know if trying to avoiding this can be done at emc_nml level or in libnml level?
[08:16:42] <micges_work> I mean trying to check buffer before write NML or such
[08:24:58] <CIA-8> EMC: 03micges 07joints_axes3 * r4f98c166d329 10/tcl/bin/emccalib.tcl: Make emccalib tool works with new inifile organisation
[08:35:33] <alex_joni> micges_work: I think emc_nml
[08:35:44] <alex_joni> but this means fixing in all clients
[08:37:26] <micges_work> mhm
[09:01:03] <micges_work> alex_joni: whee, I have machine with emc (vel=42000mm/min, acc=3000mm/s^2)
[09:04:43] <micges_work> alex_joni: I have question re. ja3: Axis is getting max velocity for jogging from [TRAJ]*_MAX_VEL, which may be much different that joging velocity for joints, how can we correct that?
[09:07:43] <alex_joni> hmm
[09:07:50] <alex_joni> jogging joints or world ? :P
[09:08:03] <alex_joni> AXIS should get the value for world jogging from [TRAJ]
[09:08:15] <alex_joni> and for joint jogging from [JOINT_*]MAX_VEL
[09:09:30] <micges_work> joint
[09:10:29] <micges_work> ok
[10:02:16] <micges_work> alex_joni: same situation for acc? (traj settings for coord mode, joint_n settings for free mode)
[10:04:33] <alex_joni> yeah, for now
[10:04:55] <alex_joni> eventually we need to get rid of [traj] settings and somehow compute the values from the [joint] vals
[10:05:05] <micges_work> ah ok
[11:38:05] <micges_work> alex_joni: you mean to get rid of ALL [TRAJ] settings?
[11:38:40] <micges_work> (all means all regards pos, vel, acc)
[11:44:58] <alex_joni> no, not all
[11:45:55] <alex_joni> maybe not even these
[11:46:14] <alex_joni> keep the values there, (not necessarely, so if not specified some computed values can be used instead)
[19:08:20] <micges> today after installing and running EMC on quite fast machine, I have few ideas to save time (which is expensive) and unnessesary errors
[19:10:37] <micges> on default servo-thread my machine (50000mm/min, 6500mm/s^2) I had strange following errors (up to 10mm)
[19:12:55] <micges> after rising servo-thread they dissapeared, so my idea is that motion will have some inputs about desired resolution at max speed and will show error if it can't be achieved (simmilar like steppers have)
[19:15:16] <micges> basically it could be usable to have in hal some info about interpolators
[19:26:53] <alex_joni> maybe deadband?
[19:27:51] <micges> alex_joni: ?
[19:28:40] <alex_joni> deadband is the distance a servo is supposed to stay without moving
[19:28:49] <alex_joni> usually the space between encoder counts
[19:29:10] <alex_joni> you could use that and max vel to compute something
[19:29:17] <micges> I know what deadband is ;)
[19:29:39] <micges> ok now I understand
[20:25:09] <cradek> micges: I think that is something you cannot calculate because it depends on the pid tuning
[20:25:34] <cradek> also I guess I don't understand what you mean by "strange" following errors
[20:25:58] <skunkworks> cradek: did you get your balls?
[20:26:06] <cradek> yes
[20:26:20] <cradek> I reballed it and that did not fix the problem - but then I found the real problem
[20:26:29] <cradek> (it needed reballing anyway, no problem)
[20:26:42] <skunkworks> ah - end bearings?
[20:27:01] <cradek> no, the motor-screw coupler was the real problem
[20:27:10] <cradek> it was not clamping straight
[20:27:21] <skunkworks> add
[20:27:24] <cradek> if I loosened the motor mounts, the motor would wave around in a little circle when moving
[20:27:38] <skunkworks> *odd
[20:27:58] <cradek> yeah it probably got mangled by the same people who badly reballed the screw
[20:29:01] <skunkworks> so - what is next on the agenda?
[20:29:21] <cradek> front panel and cleaning/painting/assembly
[20:29:27] <cradek> and trying to move it into place :-)
[20:29:39] <skunkworks> heh - through the double doors?
[20:30:03] <cradek> yep
[20:30:21] <cradek> it's too heavy for me to lift (by myself) with my pry bar - but I think with two people lifting it will be fine
[20:30:29] <cradek> a third to move the rollers
[20:31:15] <cradek> I *WILL* get my garage back
[20:31:20] <skunkworks> heh
[20:31:45] <skunkworks> mine is getting full of garage sale stuff.. (village wide rummage sale coming up)
[20:55:24] <skunkworks> there are worse things that could be in the garage :)
[22:04:58] <BigJohnT> BigJohnT is now known as BJT_Shop
[22:24:59] <BJT_basement> BJT_basement is now known as BJT_Shop