#linuxcnc-devel | Logs for 2014-08-11

[08:07:21] <jepler> I'm surprised that googling hasn't shown me a level translator designed for SPI (3 TX + 1 RX). maybe the bidirectional translators are good enough that having directional translators doesn't make sense?
[08:08:50] <jepler> anyway, if I want to continue with that project I think I (A) need the level translator as close as possible to the odroid, not 6" away on unshielded wires and (B) need to start by interfacing to an SPI slave I know works .. so that the only variable is the SPI functionality of the odroid
[08:09:17] <jepler> (A) would entail fabbing a board, so this all remains on hold at the moment
[08:23:57] <cradek> jepler: max is together and working.
[08:27:25] <malcom2073> So I'm having a bit of a stupid moment, trying to figure out where the source code to the linuxcncrsh utility is, is it in the main repo?
[08:28:14] <cradek> yes
[08:28:48] <cradek> ~/emc/src$ find . -name '*rsh*'
[08:28:48] <cradek> ./emc/usr_intf/emcrsh.cc
[08:28:59] <malcom2073> ahh ok thanks
[08:29:00] <cradek> perhaps the rename tricked you!
[08:29:10] <malcom2073> Nah, github search tricked me
[08:29:48] <cradek> websites are just a fad
[08:29:51] <malcom2073> Indeed.
[09:26:24] <cradek> http://linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/28209-261#49680
[09:35:33] <skunkworks> yeck
[09:36:08] <jepler> cradek: the "save to file" button is there, clicking it changes the timestamp on the inifile and creates an ini.bak file, but it doesn't seem to change anything
[09:36:23] <jepler> tested with a binary.hybrid.iso that has 0.pre5 because I was too lazy to update it
[09:36:34] <cradek> oh really. that's not what he describes
[09:36:37] <cradek> so we have n bugs
[09:36:51] <jepler> I tested with tklinuxcnc/servo_sim
[09:37:02] <cradek> jepler: fwiw I have a more modern one in ~/src
[09:37:03] <jepler> testing with axis it said I had no tunables
[09:37:26] <jepler> .. which is probably accurate
[09:37:30] <cradek> right
[09:38:03] <jepler> who knows what it means when the user says "function is gone":
[09:38:11] <jepler> it could mean the button is gone, or that it just does nothing
[09:38:28] <cradek> oh
[09:38:36] <cradek> I read it as the button is gone
[09:38:44] <cradek> but you're right, it could mean anything
[09:40:45] <jepler> reverting just emccalib.tcl to its contents at ref 145c609f4950ef94a711aacafcfe8552f46fabab fixes the problem I see
[09:42:38] <cradek> are you getting the "Not incompatible with" message?
[09:42:44] <jepler> cradek: no
[09:42:51] <cradek> (I think that should be "Not compatible")
[09:42:56] <jepler> oops, yeah
[09:43:07] <jepler> I do have the problem at 4a8a50154a8357e6f064a1386dd298c7e6be4800
[09:43:49] <cradek> - puts [list $j $tmpstring]
[09:44:20] <cradek> oh puts, forget it
[09:44:24] <cradek> my tcl is not
[09:45:05] <cradek> it must be the last blob?
[09:45:27] <jepler> seems like
[09:46:48] <jepler> so given tmpstr equal to "X = Y", the old code gave tmpvar "X", and the new code gave it "X "
[09:47:05] <jepler> for an input like "X=Y", the old code gave "X=Y" and the new code gave "X"
[09:47:14] <jepler> set tmpvar [lindex [split $tmpstr "="] 0]
[09:47:18] <jepler> + puts [list $tmpstr $tmpvar [lindex $tmpstr 0]]
[09:47:26] <jepler> this change seems to fix it for the standard (whitespace-y) inifile
[09:48:31] <cradek> I don't understand why lindex would do that, but I believe you
[09:49:14] <cradek> and also I don't understand why a puts would fix it
[09:49:28] <jepler> errr
[09:49:31] <jepler> I pasted the wrong line of course
[09:49:44] <jepler> + set tmpvar [string trim $tmpvar]
[09:53:40] <cradek> % format "'%s'" [lindex "a b" 0]
[09:53:40] <cradek> 'a'
[09:53:40] <cradek> % format "'%s'" [lindex "a " 0]
[09:53:40] <cradek> 'a'
[09:53:59] <cradek> I still don't understand why the thing happens that you say happens
[09:56:13] <cradek> I also don't understand how tclsh can not have readline in 2014
[09:57:05] <cradek> % format "'%s'" [lindex [split "a = b" "="] 0]
[09:57:05] <cradek> 'a '
[09:57:10] <cradek> ok what the hell
[09:58:10] <cradek> oh split returns '{a }'
[09:58:17] <cradek> that's pretty amazing
[11:32:57] <jepler> dgarr: if you see the scrollback, can you look at https://sourceforge.net/p/emc/bugs/389/ and my proposed patch?
[14:33:19] <Tom_itx> jepler, take a look at the NXP translators
[14:36:28] <Tom_itx> http://www.nxp.com/documents/data_sheet/GTL2003.pdf
[14:36:36] <Tom_itx> that's the one i used on my avr programmers
[14:37:21] <Tom_itx> set GREF to your low side and SREF DREF to the high side
[14:37:51] <Tom_itx> The device allows
[14:37:51] <Tom_itx> bidirectional voltage translations between 0.8 V
[14:37:52] <Tom_itx> and 5.0 V without use of a direction pin.
[14:41:31] <Tom_itx> the GTL2005 is probably what you'd want, it's 4 bit http://www.nxp.com/documents/data_sheet/GTL2005.pdf
[14:42:32] <Tom_itx> $.49 from digikey quantity $1.26 singles
[14:43:55] <Tom_itx> 4.4ns
[14:45:22] <Tom_itx> most all the other translators i found at the time required a direction pin
[15:02:49] <dgarr> jepler: thanks for the fix, i can commit this patch if you want: http://www.panix.com/~dgarrett/stuff/0001-emccalib-fix-for-bug-389-not-writing-file.patch
[15:05:55] <dgarr> cut using git master on debian wheezy uspace this weekend: https://www.dropbox.com/sh/ke0x71704na5jpw/AAD_IsemlDU_Tf6fZpHbzcEGa
[15:08:29] <skunkworks> Coo!l!
[15:08:33] <cradek> your stuff is very cool
[15:09:39] <cradek> do I remember right that are you using ppmc with uspace?
[15:09:56] <dgarr> yes that was hal_ppmc, uspace
[15:10:01] <cradek> amazing
[15:10:14] <cradek> I should port pluto-servo
[15:10:48] <cradek> beweep is a very funny word
[15:11:43] <dgarr> i do get some rt errors at starting which i ignore and there is a traj planner nuisance message occasionally:
[15:11:45] <dgarr> "discriminant -0.000000 < 0 in velocity calculation"
[15:12:22] <cradek> in master I've seen a nuisance message at tool change too
[15:12:40] <cradek> we should figure out how to reproduce them and make bug reports. he's been really responsive.
[15:13:07] <skunkworks> I have not see the decriminant one for a long time
[15:13:49] <dgarr> i tried to reproduce the above message on a different machine but it didn't recur, this machine is XZC
[15:13:58] <skunkworks> cradek, what I call a tool change (and I don't know how we left it) it picks up the b axis and sets it back down.
[15:14:04] <skunkworks> when
[15:14:04] <cradek> hmm xzc might be important
[15:14:17] <cradek> I bet non-xyz configs are very lightly tested
[15:14:34] <cradek> skunkworks: oh good grief
[15:14:49] <skunkworks> that is in 2.6
[15:15:11] <skunkworks> scared the crap out of me the first time ;)
[15:16:08] <skunkworks> I need to switch over the the xyz postion that you fixed..
[15:17:29] <dgarr> this tar file has a complete simulator example setup and tp_xzc.ngc that sometimes shows the discriminant message if anyone wants to try it (38848 line ngc file)
[15:17:31] <dgarr> http://www.panix.com/~dgarrett/stuff/xzc.tgz
[15:39:59] <jepler> dgarr: if you tested it too, push away! thanks!
[15:40:07] <jepler> Tom_itx: thanks for the infos
[15:40:55] <jepler> dgarr: ooh are those penrose tiles?
[16:31:40] <KGB-linuxcnc> 03Dewey Garrett 05master 20f3f1a 06linuxcnc 10tcl/bin/emccalib.tcl emccalib: fix for bug 389 (not writing file) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=20f3f1a
[16:31:44] <dgarr> yes penrose tiles kites and darts
[16:32:52] <cradek> oops that should have been on 2.6. would you cherry pick it please?
[16:35:25] <KGB-linuxcnc> 03Dewey Garrett 052.6 615e834 06linuxcnc 10tcl/bin/emccalib.tcl emccalib: fix for bug 389 (not writing file) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=615e834
[16:35:34] <dgarr> hope i did that rigt
[16:39:37] <seb_kuzminsky> dgarr: that wooden box is super cool
[16:43:28] <dgarr> thanks, it is dyed black now
[16:50:04] <Tom_itx> what's it say on the inside?
[16:50:11] <Tom_itx> or did nobody notice that?
[16:50:50] <jepler> I saw the lettering but didn't take the time to read it.
[16:50:59] <Tom_itx> my head got dizzy
[16:53:45] <cradek> dgarr: looks right, thank you
[16:54:14] <dgarr> Sonnet 29
[17:03:04] <jepler> programmer, machinist, and poet
[17:11:20] <mozmck> looks like someone changed the forum email headers - messed up my email filters.