#emc-devel | Logs for 2007-08-13

[00:06:00] <a-l-p-h-a_> a-l-p-h-a_ is now known as a-l-p-h-a
[01:02:59] <jmkasunich__> jmkasunich__ is now known as jmkasunich
[13:55:41] <jepler> cradek: huh I remember a problem with toolchange and FPR but I thought I had fixed it
[13:55:43] <jepler> cradek: guess not
[13:56:14] <cradek> jepler: if you turn OFF fpr before the tool change, the bug happens. if you leave it on, the tool change works fine
[13:56:47] <cradek> I could find the problem in 5 minutes so I hacked my gcode and continued making parts
[13:57:00] <cradek> I can get you the program if I left the machine on - let me check
[14:00:14] <jepler> I think I wrote a program here that demonstrates the problem
[14:00:58] <jepler> http://pastebin.ca/655583
[14:02:50] <cradek> if you take away line 5 does it work?
[14:04:45] <jepler> yes
[14:05:01] <cradek> that's it then
[14:05:09] <cradek> I have a second symptom, let me isolate it
[14:07:32] <jepler> but if I remove the (G94) it stops on the next G1 (because the spindle's not spinning)
[14:07:36] <jepler> not an error, just zero feed rate
[14:08:14] <cradek> http://timeguy.com/cradek-files/emc/440.ngc
[14:08:56] <jepler> looks like the interpreter doesn't know that M6 toolchange stops the spindle turning
[14:09:15] <cradek> in 440.ngc the G0 on line 31 goes slow
[19:06:17] <jepler> emc latency test: http://pastebin.ca/655938
[19:07:14] <jepler> the intervals between thread invocations are systematically a little bit too close together (an averge of 4ns too soon); the worst interval on the fast thread was 29511ns compared to requested 24305ns
[19:09:40] <alex_joni> ns?
[19:10:12] <jepler> yeah those numbers should all be in nanoseconds
[19:10:29] <jepler> assuming that's what rtapi_get_time() returns
[19:11:40] <alex_joni> I think there was something bad about rtapi_get_time() :?
[19:12:39] <jepler> yeah it takes a long time
[19:12:59] <alex_joni> ah, ok..
[19:13:06] <alex_joni> * alex_joni vaguely remembered something
[19:15:18] <jepler> doesn't matter much in this case
[19:35:15] <jepler> http://emergent.unpy.net/files/sandbox/lat.png
[19:38:08] <jepler> I'm not sure I'm measuring quite the same as the rtai latency test but with a few tweaks I will be
[19:51:41] <skunkworks_> jepler: very cool
[19:53:09] <alex_joni> jepler: hope one of these days it'll say replace your <mumblesomething> because it's the one causing bad latencies :D
[20:00:03] <jepler> alex_joni: keep dreaming!
[20:00:41] <skunkworks_> a nice list of your hardware and what latency each is contributing. ;)
[20:03:42] <alex_joni> yeah ..
[20:04:03] <alex_joni> like.. oops a bigger latency.. what did you just do? oh, you plugged in a keychain.. please stop doing that!
[20:04:38] <alex_joni> oh, the users keeps insisting.. /etc/init.d/hotplud stop
[20:05:01] <alex_joni> hotplugd
[20:19:29] <alex_joni> good night all
[20:22:26] <skunkworks_> night alex
[20:22:50] <cradek> g'night alex
[23:06:06] <cradek> bridgeport update: it is safely on the driveway
[23:07:40] <cradek> now I have to make it shorter, then make a ramp to roll it into the garage, then make it taller again
[23:09:43] <skunkworks> Cool
[23:10:07] <skunkworks> we have moved a lot of heavy equipment on inch waterpipe.
[23:10:20] <cradek> yep that's the plan
[23:10:28] <cradek> but, it's too tall, that's the hard part
[23:10:39] <cradek> also there's a 2" step up
[23:10:46] <skunkworks> yeck
[23:10:53] <cradek> got some steel plate to make a ramp with
[23:11:01] <cradek> and a heavy duty comealong
[23:11:08] <skunkworks> nice - some sort of comealong?
[23:11:09] <skunkworks> ah
[23:11:20] <cradek> fortunately there's a tree in the right place to pull on! (through the back door of the garage)
[23:11:46] <skunkworks> very nice. We would drill holes in the concrete floor around the edged for pull points.
[23:11:56] <skunkworks> edges
[23:12:00] <cradek> that's a good idea too
[23:12:27] <skunkworks> if you don't mind the holes. :)
[23:13:02] <cradek> the correctly placed tree is a very nice surprise - I hadn't thought of it until just seeing it today
[23:13:25] <skunkworks> very nice.
[23:13:51] <cradek> if I protect it with something it will be fine - it's a huge tree
[23:13:53] <skunkworks> do you have some sort of cherry picker to take the top off?
[23:14:04] <cradek> I have a forklift until 8am
[23:14:42] <cradek> I really wish I had the manuals to see how the spindle housing comes apart, but I don't have them yet
[23:15:33] <skunkworks> that should work :) just be careful. I would think a car pulling strap - kinda like a seatbelt style (2inches wide) would work for the tree
[23:15:59] <cradek> yeah I only have a chain right now, but I can go get more stuff