#emc-devel | Logs for 2010-07-15

[02:44:05] <cradek> JT-Hardinge: http://linuxcnc.org/docs/2.4/html/gcode_main.html#sec:G76:-Threading-Canned
[02:44:28] <cradek> JT-Hardinge: I notice in here that your spell checker has changed degression into digression *again* :-)
[02:44:36] <cradek> degression is the right word
[07:01:09] <ries_> ries_ is now known as ries
[10:54:05] <CIA-2> EMC: 03jthornton 07v2.4_branch * r74eaa32b7aaa 10/docs/src/gcode/main.lyx: fix spelling error again!
[16:08:58] <cradek> if TOOL_INSIDE_ARC(side, turn) {
[16:08:58] <cradek> // tool is inside the arc
[16:09:06] <cradek> wow, what genius wrote that comment?
[16:09:56] <cradek> (it was from before I made a complicated test into a define)
[16:33:51] <skunkworks_> I was going through some code I had wrote a while ago... I had a comment 'it works - don't know why'
[16:34:42] <SWPadnos> the ones I hate are ones like "this seems to work better"
[16:38:51] <skunkworks_> my programmer here comments unworking/previous code out and keeps it. He is a code horder. ;)
[16:39:39] <SWPadnos> I do that since I have never actually set up an SCM server here (shameful, I know)
[16:39:51] <skunkworks_> *hoarder
[16:40:15] <skunkworks_> hard to do with vba
[16:40:18] <skunkworks_> :)
[16:40:36] <SWPadnos> icky
[16:40:36] <skunkworks_> someday we will have a real programming language
[16:41:22] <skunkworks_> seb gave a few of us a git 101 which was pretty cool.
[16:41:37] <SWPadnos> I wish I could have stayed
[16:42:11] <SWPadnos> part of my problem is that I'm too cheap to leave another computer on all the time - I can't help but think about the power usage :)
[16:43:11] <skunkworks_> heh
[17:25:08] <jepler> SWPadnos: dreamhost can be a perfectly good place to put git repositories; you don't need a new machine.
[17:25:23] <SWPadnos> that's true
[17:25:33] <jepler> SWPadnos: good fast ssh access for comitters and optional gitweb and (since 1.6.x?) good fast http access for read-only users
[17:25:46] <SWPadnos> though I'm not skilled enough in "rootless security administration" to do that for client code
[17:25:53] <jepler> as long as you give dreamhost shell accounts to all committers
[17:26:17] <SWPadnos> you're right though, it is something to think about
[17:28:49] <jepler> and if you don't want to go through setup/administration, github has private repositories starting at $7/mo
[17:31:42] <SWPadnos> I think that's about the cost of running a lightweight server at home :)
[17:34:36] <jepler> optimistically
[17:35:03] <cradek> you don't already have a unix machine that's on all the time?
[17:35:05] <SWPadnos> yeah, it would have to average ~70W for that cost
[17:35:07] <SWPadnos> nope
[17:35:13] <cradek> huh
[17:35:23] <SWPadnos> I have a Windows machine that's on all the time
[17:35:51] <cradek> maybe if you were smart enough you could run a git server ont hat?
[17:35:54] <cradek> on that?
[17:35:55] <SWPadnos> but, once I retire it (this one), soon-ish, I'll probably be able to leave the workstation off much of the time, so that would free up some power budget for a server
[17:36:05] <SWPadnos> no, I don't want to be that smart
[17:36:12] <cradek> yeah I wouldn't either.
[17:41:27] <SWPadnos> of course, I have the hardware for the server, and I have even upgraded it several times, even though I have never put it into actual service :)
[17:42:19] <cradek> silly, you don't upgrade servers, you use your lowly old machines for servers
[17:42:48] <SWPadnos> well, it kind of is a lowly old machine, I mean it has only one core (I think) and less than 4GB of RAM
[17:42:55] <cradek> haha
[17:43:20] <SWPadnos> and I don't think it has many giggle-hertzes either
[18:09:29] <jepler> it might well beat what serves git.linuxcnc.org.
[18:10:25] <SWPadnos> I think it is 64-bit, some socket 939 thing
[18:10:51] <SWPadnos> might even be an Opteron 146 or similar (but I think that motherboard ended up not working or something)
[18:11:53] <SWPadnos> oh hmm. I guess I was wrong, it's an Athlon 64 X2 3800+
[18:11:55] <SWPadnos> oh well
[18:12:04] <SWPadnos> 65W TDP though
[18:14:23] <SWPadnos> no wait, that's in my wifes machine
[18:14:30] <SWPadnos> oh well, it doesn't matter today
[19:54:40] <JT-Hardinge> is there any way to tell what EMC is waiting for when it stops for no reason that I can think of?
[19:55:36] <skunkworks_> Scotty 'computer... COMPUTER....'
[19:56:30] <cradek> JT-Hardinge: stops what where?
[19:56:49] <JT-Hardinge> after a tool change
[19:56:57] <JT-Hardinge> T3 M6
[19:56:59] <JT-Hardinge> G43
[19:57:09] <cradek> does it do any motion at all afterward?
[19:57:10] <JT-Hardinge> G1 X1.500 F150
[19:57:17] <JT-Hardinge> M3 M8
[19:57:19] <JT-Hardinge> ...
[19:57:23] <JT-Hardinge> no
[19:57:33] <cradek> you have G1 entry move before M3?
[19:57:40] <JT-Hardinge> the tool change completes
[19:57:42] <JT-Hardinge> yes
[19:57:42] <cradek> it could be waiting for spindle-at-speed if it's a G1
[19:58:19] <jepler> I wonder if using position-interpolated to pid gets most of the benefit of the high-precision velocity estimate
[19:58:24] <cradek> does your vfd give at-speed when stopped?
[19:58:29] <JT-Hardinge> before the M3?
[19:58:33] <JT-Hardinge> no
[19:58:35] <jepler> (since it includes the same information, and since hal_float_t is double it doesn't have precision problems as much)
[19:58:51] <JT-Hardinge> the at speed it internal in HAL
[19:58:54] <cradek> JT-Hardinge: I'm not sure without testing, but that's what I suspect
[19:59:22] <cradek> if you are M5 and the spindle is actually stopped you should probably have at-speed asserted because 0=0
[19:59:49] <cradek> try unlinking motion.spindle-at-speed and setp it true, and see if it goes
[19:59:58] <JT-Hardinge> at speed is false
[20:00:10] <JT-Hardinge> ok, I'll have to run it again give me 240 seconds
[20:01:13] <JT-Hardinge> I thought that it would only wait for the spindle to be at speed for the move after the M3?
[20:01:14] <cradek> or just change it to a G0
[20:01:40] <cradek> it doesn't make much sense to have G1 with the spindle stopped - I bet I never tested that situation
[20:02:20] <JT-Hardinge> If I use G0 then I get all the out of bounds errors :/
[20:02:21] <cradek> JT-Hardinge: I don't remember now exactly how it works - the tool change might confuse it because it stops the spindle as a side effect
[20:02:32] <cradek> oh
[20:02:35] <cradek> (AXIS,hide)
[20:02:36] <cradek> G0 ...
[20:02:38] <cradek> (AXIS,show)
[20:02:49] <cradek> grumble grumble
[20:03:12] <cradek> huh wonder why G0 vs G1 even affects that
[20:03:34] <cradek> I bet we should just take that check out.
[20:03:36] <JT-Hardinge> dunno but it does give a false warning with G0
[20:03:42] <cradek> interesting
[20:05:58] <JT-Hardinge> as soon as I unlinkp motion.spindle-at-speed it took off and completed the program
[20:06:13] <JT-Hardinge> I didn't even have to set it to true
[20:07:14] <cradek> yes it defaults to true, so you don't have to hook it up if you don't want
[20:08:44] <JT-Hardinge> ok, I'll leave it unhooked for now