#emc-devel | Logs for 2006-03-23

Back
[00:34:53] <rayh> rayh is now known as rayh_away
[02:04:53] <cradek> so what happens if you're cutting a thread and you hit pause?
[02:05:06] <cradek> I guess that's a bit rhetorical, because I know what happens - it ruins the work
[02:05:21] <cradek> I ask because currently if you hit pause, it doesn't pause until after the thread is done
[02:05:32] <cradek> I wonder if that's a feature.
[02:10:56] <cradek> ok, nobody has said no in five minutes, so I declare it officially a feature
[02:11:00] <rayh_away> That is one good way to handle the problem
[02:11:43] <cradek> which? not pausing, or feature declaring?
[02:11:59] <rayh_away> not pausing during the cut.
[02:12:24] <rayh_away> I was talking to a lathe guy this morning
[02:12:26] <cradek> I think so too; it seems like you would NEVER want to pause then.
[02:12:54] <rayh_away> the fanuc will jump out of the thread in x if pause.
[02:13:11] <cradek> wow
[02:13:22] <cradek> can it resume then?
[02:13:24] <rayh_away> but I don't know how you would continue after that
[02:13:34] <rayh_away> I'll ask tomorrow morning.
[02:13:42] <cradek> you'd have to back up a bit, or maybe just start that cut over
[02:13:45] <rayh_away> we declare it a feature for now.
[02:13:50] <cradek> haha agreed
[02:14:11] <cradek> I'm dying to try this - the plots look good, but I can hardly believe it until I see it cut
[02:14:43] <rayh_away> Fantastic. I've got a couple of fanuc post threads coming.
[02:15:09] <rayh_away> Right. I'll order that encoder tomorrow.
[02:15:23] <cradek> I think the groundwork for more complex canned cycles is done; you can do a series of cuts while synced now
[02:15:46] <rayh_away> great.
[02:16:02] <cradek> I've got an encoder coming for my mill too, but I'm not sure how well that will work out compared to a real lathe.
[02:16:28] <cradek> if I had spindle reversal, I bet I could pull off rigid tapping.
[02:16:47] <rayh_away> I've got a set of pmdx stepper drives and a power supply so I should be able to put the lathe together pretty quick.
[02:17:09] <cradek> do you have the sherline motors and mounts?
[02:17:10] <rayh_away> That will be something to test out on the Mazak.
[02:17:24] <rayh_away> Yes. It's a full cnc ready lathe.
[02:17:30] <cradek> anything bigger than me scares me...
[02:17:54] <cradek> I'd rather break some little 4-40 taps first!
[02:17:55] <rayh_away> You'll get used to it pretty quick when you get next to the Mazak.
[02:18:33] <cradek> it already has reversal?
[02:18:37] <rayh_away> euuu those break really easy. If you could run one of those in a 1/2 inch deep hole that would be an awesome demo.
[02:18:52] <cradek> what diameter of hole? hahaha
[02:18:58] <rayh_away> Yes the mazak has full speed control with tach and encoder feedback.
[02:19:19] <rayh_away> I figured you'd cheat it.
[02:19:25] <rayh_away> hole size that is.
[02:19:49] <cradek> how big a tap do you need before you can tap (say aluminum) without periodic reversals? I always reverse constantly when doing it by hand - they get stuck so easily otherwise
[02:20:13] <cradek> maybe #10? or 1/4?
[02:20:41] <cradek> (without cheating - real 60% threads)
[02:20:44] <rayh_away> I tap 6-32 with a battery drill with no reversal except at the bottom.
[02:20:55] <rayh_away> alum is harder than mild steel
[02:21:01] <cradek> in the recommended hole size?
[02:21:05] <rayh_away> need some good tapping gel.
[02:21:11] <rayh_away> Right.
[02:21:24] <cradek> yeah maybe my lube is not very good. I use whatever's handy.
[02:21:55] <cradek> alumicut is ok, but I've been known to use motor oil
[02:22:13] <rayh_away> I'm seeing a difference between hal variables and ini variables.
[02:22:35] <cradek> what do you mean
[02:22:38] <rayh_away> like axis scale is 4000 in ini and 4000.0 in a hal return.
[02:22:58] <cradek> I guess that's not too surprising
[02:23:31] <rayh_away> I've a notion to simply write all of the hal read ini variables with the hal value.
[02:23:48] <cradek> I doubt that will hurt anything
[02:24:00] <rayh_away> Okay. You twisted my arm.
[02:24:36] <rayh_away> I'm nearly finished with emccalib.tcl but won't complete until morning now.
[02:24:58] <cradek> I haven't had a look yet - can it be run now?
[02:25:32] <rayh_away> Sure. You will have to change the permissions of emc2/tcl/bin/emccalib.tcl to a+x
[02:25:52] <rayh_away> Turns out all of those files were sources rather than executed.
[02:26:53] <rayh_away> The latest sf version will change values and let you revert to the last value but not save.
[02:26:54] <cradek> I don't get a window...?
[02:27:16] <rayh_away> odd.
[02:27:30] <cradek> chris@buster2:~/emc2.head$ chmod +x tcl/bin/emccalib.tcl
[02:27:30] <cradek> chris@buster2:~/emc2.head$ tcl/bin/emccalib.tcl
[02:27:35] <cradek> (with emc running)
[02:28:02] <rayh_away> I've only run it from the tkemc menu. That might be the difference.
[02:28:09] <rayh_away> Let me look
[02:28:25] <cradek> might need some kind of environment
[02:28:54] <rayh_away> Right.
[02:29:47] <rayh_away> "exec $TCLBIN/emccalib.tcl -- -ini $EMC_INIFILE &"
[02:29:55] <rayh_away> is what I use from inside tkemc
[02:29:57] <cradek> ok got it (added support in axis)
[02:30:09] <rayh_away> fantastic.
[02:30:19] <cradek> I don't have many options with my stepper setup
[02:30:29] <rayh_away> you should see three.
[02:30:45] <cradek> yes
[02:30:56] <cradek> it's a little too bad we can't change maxaccel/maxvel
[02:31:19] <rayh_away> I've tested with m5i20, univstep, univpwm, and the stepper versions.
[02:31:19] <cradek> allowing changing HAL_MAXVEL is misleading because it won't work when changed
[02:31:35] <cradek> same with SCALE, I guess
[02:31:43] <rayh_away> If there were a way to get at those emc global variables I believe we could.
[02:32:02] <cradek> PID is different, it is implemented in hal, so you can tune online
[02:32:45] <rayh_away> I read all of the setp lines in the hal files that read from the ini.
[02:33:01] <cradek> ah, I see
[02:33:17] <rayh_away> So changing them and re issuing the setp should make HAL run by the new value.
[02:33:38] <cradek> since we have bwidget, you could use a real tab widget instead of these pseudo-tabs
[02:33:43] <rayh_away> I've not tried that.
[02:34:03] <rayh_away> I thought about that but this was easier.
[02:34:13] <cradek> I think it would be more evident how it works, for the user who walks up to it for the first time
[02:34:17] <rayh_away> It was code I had elsewhere.
[02:34:31] <rayh_away> I'll look at that in the morning.
[02:34:39] <cradek> I'm sure that's something we could do in the "second pass"
[02:34:57] <rayh_away> Yep. Nothing is sacred.
[02:35:07] <cradek> more important to get it working than make it pretty at first
[02:35:39] <rayh_away> I'll have a servo powered sherline mill hooked up tomorrow.
[02:35:51] <rayh_away> That will be my test.
[02:35:57] <cradek> neat, how do you do that?
[02:36:30] <rayh_away> m5i20 with their 7i?? four axis 100 watt interface card.
[02:36:44] <cradek> that's one of the new pci cards?
[02:37:04] <rayh_away> right. $300 for the two cards.
[02:37:17] <cradek> other than the two cards what is needed?
[02:37:38] <cradek> (I'm looking forward to seeing this kind of hardware at fest)
[02:37:40] <rayh_away> a power supply <48 volts
[02:37:58] <rayh_away> and +5 for the encoders.
[02:38:25] <cradek> wow, that's an inexpensive setup then
[02:38:26] <rayh_away> I got some surplus pittman motors.
[02:38:36] <rayh_away> Roltek made me the motor mounts.
[02:38:49] <rayh_away> they had 1/4 inch shafts so that fit as is.
[02:39:10] <cradek> your sherline has acme screws?
[02:39:31] <rayh_away> No they are sherline's own threads.
[02:39:46] <cradek> oh really, I didn't know they did their own thing
[02:39:54] <rayh_away> 1/2 20 nominal for x and y and 3/8 in z
[02:40:09] <rayh_away> looks like allthread
[02:40:33] <rayh_away> oops 1/4 20 for x and y
[02:40:43] <cradek> 20 is a lot of threads per inch - what rapid speed do you think you can get?
[02:40:47] <rayh_away> They also make a metric thread version.
[02:40:54] <cradek> yeah mine is metric (1mm pitch)
[02:41:10] <rayh_away> I set up one for Abdul (vital) and he was getting 125 ipm.
[02:41:24] <cradek> wow
[02:41:34] <cradek> for 6" travel that's pretty darn fast
[02:41:37] <rayh_away> thread would last about a day if you ran it hard.
[02:42:03] <rayh_away> and the x would whip when it was at one end.
[02:42:13] <cradek> yikes
[02:42:28] <rayh_away> The stepper version from sherline runs bout 22 ipm with their little motors
[02:42:51] <rayh_away> That's fast enough for the newbee.
[02:42:56] <cradek> yeah, it seemed pretty slow, but fast enough
[02:43:13] <cradek> I get 30 on my hodgepodge mill and am happy enough
[02:43:31] <cradek> I can only cut pcbs at 20-22 so 30 is fast enough
[02:43:33] <rayh_away> I'm running 36 here with emc2 but that is about all the motors will torque.
[02:44:02] <rayh_away> The Smithy 1240 rapids at 180
[02:44:16] <cradek> their spindle is really slow - so you have to cut slow anyway
[02:44:21] <rayh_away> but that is rutex running 500 watt motors.
[02:44:41] <rayh_away> Yes it is but the speed is nice for rapid return to home.
[02:44:56] <cradek> yes or sloppy gcode (automatically generated) with lots of rapids here and there
[02:45:14] <rayh_away> They have talked about getting a single speed spindle to fit in the block.
[02:45:30] <cradek> 10k or so?
[02:45:47] <rayh_away> I seem to remember 6k or so
[02:46:03] <rayh_away> many of their customers are engraving with it.
[02:46:16] <rayh_away> motorcycle rims and such.
[02:46:27] <cradek> wonder how fast their current spindle goes
[02:46:38] <rayh_away> 3k i think.
[02:46:41] <cradek> ouch
[02:46:52] <cradek> that's way too slow for any small tools
[02:46:57] <rayh_away> right.
[02:47:16] <cradek> my pcb tools "require" 20krpm
[02:47:25] <cradek> I've used them at 10k but they work much better at 20k
[02:47:48] <cradek> I would think all engraving tools would be the same way
[02:48:02] <rayh_away> Certainly seems like it.
[02:48:25] <cradek> hey I should photograph my stepper driver board before I solder it
[02:48:32] <cradek> in fact I'm going to go do that.
[02:48:40] <cradek> back later!
[02:48:44] <rayh_away> You bet. Put it on the wiki.
[02:48:49] <rayh_away> see you.
[13:19:28] <alex_joni> morning ray
[13:22:09] <rayh> Hi Alex.
[13:22:40] <rayh> got a quick question about saving ini files after calibration.
[13:23:04] <rayh> We don't want to get into the diff stuff that emc did.
[13:23:15] <rayh> but should we save one backup?
[13:23:57] <rayh> name it xxx.ini.bak and it won't show on picker
[13:24:13] <rayh> name it xxx.back.ini and it will.