#emc-devel | Logs for 2007-07-07

[00:08:00] <jepler> hi lerneaen_hydra
[00:08:04] <jepler> up in the middle of the night?
[00:08:06] <jepler> mutt
[00:08:06] <jepler> oops
[00:08:18] <lerneaen_hydra> yeah pretty much :)
[00:11:14] <lerneaen_hydra> so what's happening?
[00:12:12] <jepler> not much
[00:12:19] <jepler> I'm playing with a hardware step generator, but not sure I can get the bugs out
[00:12:27] <jepler> (for the pluto board)
[00:12:30] <lerneaen_hydra> pluto p?
[00:12:34] <lerneaen_hydra> cool
[00:12:41] <lerneaen_hydra> what's sent to the pluto?
[00:12:56] <lerneaen_hydra> desired speed/dir for an axis?
[00:13:23] <jepler> yeah basically
[00:13:37] <jepler> the hal driver would take position, not velocity, as input
[00:15:09] <lerneaen_hydra> oh, so it's position/axis?
[00:15:17] <lerneaen_hydra> and over a certain time?
[00:15:46] <jepler> the communication between the HAL driver and the card is as you said at first: speed and direction
[00:16:00] <jepler> the HAL driver does the conversion from position command to velocity command
[00:16:53] <lerneaen_hydra> ah ok
[00:33:03] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/pluto-step-following-error-2.png
[00:34:12] <lerneaen_hydra> cool :)
[00:35:55] <jepler> http://emergent.unpy.net/index.cgi-files/sandbox/ferror-close-up.png
[00:36:40] <lerneaen_hydra> not bad :)
[00:36:44] <jepler> I wonder if this is due to a bug, or the quantization of the velocity request, or something else .. it sure has a distinctive pattern
[00:36:49] <lerneaen_hydra> hmm, now it's much too late here
[00:37:02] <lerneaen_hydra> wouldn't quantisation do that?
[00:37:04] <lerneaen_hydra> 'night
[00:37:29] <jepler> scale is 10000 steps per unit
[00:37:43] <lerneaen_hydra> oh
[00:37:58] <jepler> = 1/5 div on that second screenshot .. so maybe
[00:38:17] <jepler> I mean, the jitter in the error, not the error itself
[00:38:31] <lerneaen_hydra> yeah
[00:39:26] <jepler> the error is proportional to the speed, and for constant inputs it seems to settle on the nearest count and not jitter
[00:39:29] <jepler> so maybe it's right
[00:39:35] <jepler> I just need a stepper to test it on :-P
[00:39:54] <jepler> anyway .. 'night
[00:40:21] <lerneaen_hydra> 'night :)
[00:50:12] <jepler> sigh, I can't find bit sizes that do everything I want: 4 axes worth of stepgens to fit on the pluto, a high maximum step rate, but minimum nonzero step rate less than 1 step per ms
[01:32:41] <jmkasunich> hi guys
[01:34:13] <cradek> hi
[01:34:27] <jmkasunich> saw the discussion about jogwheel direction
[01:34:47] <jmkasunich> there was no rhyme or reason to it, just what happened
[01:34:54] <jepler> hi jmkasunich
[01:35:02] <jmkasunich> the scale values could be made negative to reverse one or more axes
[01:35:16] <jmkasunich> hi jeff
[01:35:16] <cradek> oh it's perfectly sane considering the sign of the tool movement
[01:35:45] <cradek> I just don't like it because Y and Z feel backward compared to a manual machine
[01:36:14] <cradek> I'd probably set mine with some negatives, but it would probably surprise anyone else using it
[01:46:51] <cradek> jmkasunich: sure looks like there's another ppmc bug doesn't it
[01:47:10] <jmkasunich> the latest email implies that its noise or crosstalk
[01:47:29] <cradek> a spike of noise exactly every .020 inches of movement?
[01:47:34] <cradek> seems unlikely to me
[01:47:59] <jmkasunich> well, Jon grounded his shield and it magically went away for him
[01:48:49] <cradek> I thought at first that every 400 counts could be a 1 bit error - but wrong 400
[01:49:11] <jepler> you mean, 400 is not octal in this case?
[01:49:18] <jepler> or do you mean something else?
[01:49:36] <cradek> right, I was thinking 0x400 but that's wrong
[01:51:34] <cradek> so what does anyone think about merging nineaxis?
[01:51:55] <jmkasunich> as in "before or after 2.2?"
[01:52:01] <cradek> yeah
[01:52:26] <cradek> I wonder if we've broken all the guis
[01:55:12] <cradek> tkemc has the very surprising behavior of showing X Y Z C A B
[01:55:26] <jepler> that's interesting
[01:55:40] <cradek> isn't it
[01:57:59] <cradek> mini shows X Y Z A B C correctly
[01:58:30] <cradek> xemc shows X Y Z like it always does
[01:59:21] <cradek> so they need a little massaging, but they're not totally broken
[02:00:55] <cradek> I definitely like the new error checking for undefined axes. it's a big improvement.
[02:02:17] <jepler> yeah -- that would be nice in trunk whether nineaxis is merged or not
[02:02:50] <cradek> lathe seems to work
[02:03:03] <cradek> even when I take out the complete [AXIS_1] section
[02:10:43] <jepler> jmkasunich: I think the pluto step generator is working, though I haven't been able to really scope the step waveform yet
[02:11:25] <jepler> the waveform looks OK in simulation though
[02:11:35] <jepler> bbl
[15:38:39] <jepler> pluto-step works well enough to run emc .. but still no motors to turn with it
[15:47:36] <cradek> 4 axis step plus dio?
[15:48:59] <jepler> yes
[15:49:24] <jepler> 15 din 13 dout 4 stepgen
[15:49:31] <jepler> no pwm, no watchdog, no encoder
[15:49:53] <cradek> nice
[15:50:07] <cradek> that's a lot of IO
[15:50:32] <jepler> er, 16 din, 14 dout
[15:50:38] <jepler> yes it is plenty for stuff like limit switches
[18:37:36] <Skullworks-PGAB> jepler: what do you need to test pluto-step? functional iron with motors or would just steppers/driver on a bench work?
[18:40:16] <jepler> Skullworks-PGAB: I'll probably bully chris into letting me try it on his mill, but one stepper motor and driver would be nice too
[18:41:54] <Skullworks-PGAB> cradek still using a 297/298 half step driver?
[18:42:17] <jepler> yeah
[18:44:36] <Skullworks-PGAB> I have a hobbyCNC 4 axis unipolar driver and 200 oz/in motors that I shelved - was planning a router table then came to my senses as I have no place to store the thing.
[18:45:43] <Skullworks-PGAB> allows 1/2 to 1/16 step
[18:46:27] <jepler> I was just reading the page about that driver
[18:46:30] <Skullworks-PGAB> if it would help for testing I could ship it - I have no plans for it.
[18:46:39] <jepler> sounds like a decent board
[18:47:30] <Skullworks-PGAB> the fixed chopper makes noise - and its unipolar - but otherwise its a great entry unit.
[18:47:28] <jepler> aside from being unipolar
[18:49:44] <jepler> * jepler has a little debate with himself
[18:50:13] <jepler> how much are you looking for?
[18:50:36] <Skullworks-PGAB> donation
[18:50:54] <Skullworks-PGAB> further the cause
[18:51:20] <jepler> at that price I hardly see how I could turn it down
[18:51:33] <jepler> except it probably means I have to deliver on pluto-stepper
[18:52:15] <Skullworks-PGAB> when the code is ready - its ready... and I wouldn't be using it anyway...
[18:53:15] <Skullworks-PGAB> besides - there are other applications you might be able to use it for.
[18:54:09] <jepler> will you send them USPS or some other service?
[18:54:30] <Skullworks-PGAB> my friends and I joked about using it to make a penn ploter to mount on the wall over a white board to draw stuff during presentations...
[18:54:47] <Skullworks-PGAB> yeah likely USPS
[18:54:47] <cradek> jepler: maybe you'll finally have something cnc
[18:55:13] <jepler> cradek: yeah I only have to build the rest of the machine in that case
[18:55:22] <cradek> right
[18:55:28] <jepler> Skullworks-PGAB: then send it to: Jeff Epler / 1666 Burr Street / Lincoln NE 68502
[18:56:11] <jepler> Skullworks-PGAB: you're sure you don't want any payment?
[18:56:31] <cradek> * cradek thinks jepler should pay shipping
[18:56:33] <Skullworks-PGAB> nope - you guys work hard enough as is
[18:56:36] <cradek> :-)
[18:56:41] <jepler> wow it's like christmas
[18:57:27] <cradek> 200 oz-in is real motors for a small machine
[18:57:32] <cradek> that's slick
[18:58:06] <Skullworks-PGAB> I have the address saved - I'll get everything boxed up and see if I can blackmail my cousin to get it to the Post Office
[18:58:58] <Skullworks-PGAB> During the week I work 5a to 5p and its a 45 minute drive so I can never get in the door while there open.
[18:59:49] <jepler> I understand
[19:00:08] <cradek> they don't set their hours with the convenience of their customers in mind.
[19:01:31] <Skullworks-PGAB> its a .gov job - hardcopy spam delivery
[19:02:11] <jepler> cradek: yeah I think the flaw in your reasoning is that we are not the post office's customers
[19:02:37] <cradek> oh.
[19:04:00] <jepler> Skullworks-PGAB: let me know when it does get mailed, so that I can start getting impatient for its arrival
[19:04:34] <Skullworks-PGAB> :)
[19:08:04] <Skullworks-PGAB> hey - did either of you need anymore of those MAXON gearmotor servos like you used on your etch-a-sketch? - On impluse I bought like 10 and for the life of me I have no idea why. - I can throw some in the box also.
[19:08:43] <Skullworks-PGAB> * Skullworks-PGAB might actual have a clean spot on his desk....
[19:09:13] <jepler> I should probably say "no"
[19:09:27] <jepler> on the other hand, there are moments I regret that I didn't buy like 10 extra because clearly they have some good use
[19:10:20] <Skullworks-PGAB> I think I was reasoning that I could salvage the encoders
[19:10:33] <Skullworks-PGAB> which I did one one
[19:11:05] <Skullworks-PGAB> its quite a job - they used glue at one point - tamper proofing
[19:11:42] <jepler> yeah they do not just snap apart by any means
[19:11:58] <Skullworks-PGAB> but its doable - its just a small count and 3mm shafe size - makes it a pain to reuse
[19:13:03] <jepler> they're just 100 lines or something, right?
[19:13:42] <Skullworks-PGAB> 100 line - no index
[19:14:24] <Skullworks-PGAB> the disc was marked once I got it open
[19:14:45] <Skullworks-PGAB> watch screws...
[19:16:07] <Skullworks-PGAB> use a towel on the table so they don't bounce off the face of the earth and fall into the abyss
[19:30:20] <Skullworks-PGAB> bbl
[19:56:18] <jepler> Error: Project requires 614 logic cells, but the selected device can contain only 576 logic cells
[20:00:32] <jepler> 4 step + 1 PWM + 1 quadrature doesn't fit
[20:48:07] <alex_joni> hi guys
[20:51:07] <jepler> hi alex
[20:52:38] <alex_joni> seen you're starting to build a cnc
[20:53:10] <alex_joni> * alex_joni grins
[20:53:56] <jepler> something like that
[20:54:20] <jepler> I am pretty sure I'm what is derisively called a "motor spinner"
[20:58:11] <alex_joni> * alex_joni looks up "derisively"
[21:00:05] <alex_joni> oh :)
[21:02:43] <cradek> * cradek looks around, with a bit of panic, for his notes about how this used to be wired
[21:03:09] <jmkasunich> * jmkasunich is a motor spinner too
[21:03:41] <alex_joni> I did some CNC work today with emc2 :P
[21:03:44] <jepler> cradek: what's "this"?
[21:03:50] <alex_joni> lathe?
[21:04:20] <cradek> mill spindle motor
[21:04:24] <jmkasunich> cradek: I know it goes against your beliefs, but you really should document yoru machine ;-)
[21:04:28] <jmkasunich> your even
[21:04:41] <cradek> black green orange white red ...
[21:05:00] <jmkasunich> eenie meenie minie moe
[21:05:41] <cradek> jepler: found a scsi card you can have if you still need one
[21:14:36] <jepler> cradek: OK, I'll let you know
[22:46:48] <cradek> argh, I got it together, and it's totally dead.
[22:47:14] <cradek> jepler: do you have any leftover new thinkpad nipples?
[23:03:12] <jepler> cradek: I suspect I would never find them
[23:29:14] <cradek> that's ok, thanks
[23:29:40] <cradek> they're cheap on ebay
[23:46:59] <cradek> no, I don't think it should cost three times as much to ship three nipples...
[23:53:21] <cradek> http://xkcd.com/c243.html
[23:55:51] <jmkasunich> they need to be carefully packed - of course shipping more costs more!
[23:57:18] <cradek> um yeah
[23:57:51] <cradek> man I hate motors with brushes you can't insert after assembling the motor
[23:58:11] <jmkasunich> you need 12 hands and toothpicks to put it together
[23:58:38] <cradek> yep
[23:58:47] <jmkasunich> http://xkcd.com/c231.html
[23:58:51] <cradek> and, usually, x-ray vision
[23:59:04] <cradek> yep, I like that one too
[23:59:06] <cradek> what a great comic