#emc-devel | Logs for 2010-11-07

[04:18:12] <logger_1> logger_1 is now known as 16SAANK32
[04:18:17] <16SAANK32> * 16SAANK32 is logging
[04:19:04] <SWPadnos> ?
[04:19:04] <SWPadnos> ?
[04:19:27] <SWPadnos> who changed the logger script?
[04:19:27] <SWPadnos> who changed the logger script?
[17:02:55] <qq-> psha, installed camview, what else we need to install for test (my camera is working too)
[17:09:10] <qq-> psha, btw, thanks for yesterday congrat ;)
[17:41:04] <skunkKandT> cradek: what do I put in the ini for locking rotory? I am not seeing it in the released.
[17:41:26] <skunkKandT> * not seeing it in the devel docs
[17:48:05] <cradek> [AXIS_4] LOCKING_INDEXER = 1
[17:50:26] <skunkKandT> Thanks!
[17:52:01] <skunkKandT> using the pid.N.feedback-deriv sure seems to quiet it down
[17:54:00] <skunkKandT> can someone refresh my memory on how to hook up the pid so there isn't a thunk when homing to index?
[17:55:11] <skunkKandT> I remember it being on the list.. (still looking)
[17:57:51] <skunkKandT> I think I have it..
[17:59:15] <jepler> hook pid's index-enable pin to the index signal. if that still doesn't fix it, check thread order (read hardware, motion, pid)
[17:59:39] <jepler> hmph :-/ I see pid index-enable is not documented
[18:00:59] <skunkKandT> yeh - that is why I was having a little problem remembering ;)
[18:02:10] <micges> I think there is problem in 2.4.5 sim_mm: when I'm issue m61 q3 Axis show no tool but emctop shows tool_in_spindle=3
[18:07:01] <jepler> I think the implementation of M61 is incomplete. two things in the stat buffer have to ultimately be updated, but only one is. http://pastebin.ca/1984632
[18:08:25] <micges> I think the same, just wanted confirmation of tihs
[18:09:22] <jepler> I agree it's a bug. assuming the fix is simple, I'd like to have it for a future 2.4 release.
[18:11:25] <micges> ok, I'll take care of it
[18:11:30] <jepler> thank you.
[18:13:14] <micges> second thing: if you have tool loaded, and tool length comp enabled, then you load another tool, TLO will not follow with present tool, you must issue g43 to update it
[18:13:22] <micges> is it intended?
[18:14:05] <cradek> yes it has always been that way
[18:15:52] <micges> ok
[18:16:11] <jepler> skunkKandT: how's this look for documentation of pid index-enable? http://pastebin.ca/1984640
[18:19:40] <skunkKandT> jepler: very good
[18:23:23] <skunkKandT> hmm - odd issue - I do have the index enable hooked to the pid - but when I do a home with ff1 set to non-zero - the servo osolates pretty bad as it tries to go home. (index only homing - 2 degree's home offset.
[18:23:45] <skunkKandT> threads seem to be in the right order in the hal file
[18:24:28] <skunkKandT> ok - never mind
[18:24:50] <jepler> you figured it out?
[18:25:21] <jepler> I was going to say, if you were using pid with feedback-deriv, but BEFORE my commit of last night 1187c3cadd63 then there's a new problem similar to the FF1 problem that the index-enable pin fixes
[18:25:46] <skunkKandT> no it seems to still be doing it
[18:25:53] <skunkKandT> I got the checkout this morning
[18:27:56] <skunkKandT> seems like the first time i home it.
[18:28:03] <skunkKandT> still playing
[18:29:52] <skunkKandT> and it doesn't seem to matter if the ff1 is non zero.
[18:30:35] <CIA-2> EMC: 03jepler 07v2.4_branch * r5237b3b14b7d 10/docs/man/man9/pid.9: pid: document index-enable pin
[18:30:44] <CIA-2> EMC: 03jepler 07master * r1b5432c084c1 10/src/hal/components/near.comp: near: fix for negative inputs (SF#3092250)
[18:30:45] <CIA-2> EMC: 03jepler 07master * rf4ae49cc6509 10/tests/near.0/ (README expected runstreamer test.hal): near: add a test for correctness
[18:30:46] <CIA-2> EMC: 03jepler 07master * rdebc639a0dab 10/src/hal/drivers/parport_common.h: pico/7i43: don't show a non-fatal message as error
[18:30:48] <CIA-2> EMC: 03jepler 07master * ra317d6196a60 10/src/emc/iotask/ioControl.cc: Fix current tool information after reloading table
[18:30:49] <CIA-2> EMC: 03jepler 07master * r00d512d30841 10/src/emc/usr_intf/pncconf/ (pncconf.glade pncconf.py): Add Touchy GUI choice. Fix spindle-at-speed code
[18:30:53] <CIA-2> EMC: 03jepler 07master * r65371b9d28e3 10/src/emc/usr_intf/pncconf/pncconf.py: use constants for descibing GUIs. fix Touchy signal names
[18:30:54] <CIA-2> EMC: 03jepler 07master * r7098f4f04396 10/src/emc/usr_intf/pncconf/pncconf.py: change touchy's quillup signalname
[18:30:55] <CIA-2> EMC: 03jepler 07master * r57ac6f7acdf0 10/src/emc/usr_intf/pncconf/pncconf.py: clean up generated HAL file text for jogging signals
[18:30:56] <CIA-2> EMC: 03jepler 07master * r16a80ca82e19 10/src/emc/usr_intf/pncconf/pncconf.py: fix error in hal file with default selected-jog-incr
[18:30:59] <jepler> the index pulse only causes a position jump the first time in a session..
[18:31:04] <CIA-2> EMC: 03jepler 07master * rb8fecffb8c1c 10/src/emc/usr_intf/pncconf/pncconf-help/ (help-extcontrols.txt help-gui.txt): Add info about TOUCHY
[18:31:05] <CIA-2> EMC: 03jepler 07master * r159df2fdba5f 10/ (12 files in 7 dirs): Merge branch 'v2.4_branch' of ../emc2.4
[18:31:06] <CIA-2> EMC: 03jepler 07master * r398675422123 10/docs/man/man9/pid.9: pid: index-enable is related to -deriv inputs too
[18:31:08] <CIA-2> EMC: 03jepler 07master * r5237b3b14b7d 10/docs/man/man9/pid.9: pid: document index-enable pin
[18:31:36] <skunkKandT> oh - should I git pull?
[18:31:56] <jepler> that push just now didn't change anything about pid
[18:31:59] <jepler> just the docs
[18:32:16] <skunkKandT> oh - didn't read that closely
[18:33:13] <jepler> scope pid.#.output and index-enable. trigger on falling edge of index-enable. if there's a spike in output at the same time as index-enable falls, then it's related to index. otherwise, it's not.
[18:33:31] <skunkKandT> ok
[18:33:41] <psha> qq-: camview-tiny
[18:34:02] <jepler> if you change your homing so that each repeated homing goes to the *next* index instead of the *same* index then you can do it more than once per start of emc..
[18:34:23] <skunkKandT> ah - ok
[18:37:55] <qq-> psha, done; camview-tiny /dev/video0 ; You must specify chain file, what's that 'chain file' ?
[18:42:53] <skunkKandT> jepler: this is re-home.. (yes my tune sucks)
[18:42:55] <skunkKandT> http://imagebin.ca/img/s77E9G.png
[18:43:06] <skunkKandT> this is finding the next index
[18:44:04] <skunkKandT> jepler: http://imagebin.ca/view/DwApulu.html
[18:44:17] <skunkKandT> or what happens during the first home
[18:49:05] <skunkKandT> I am pretty sure I have it hooked up right net b-index-enable hm2_[HOSTMOT2](BOARD).1.encoder.00.index-enable <=> axis.4.index-enable pid.3.index-enable
[18:49:50] <jepler> yuck http://imagebin.ca/view/DwApulu.html
[18:50:26] <skunkKandT> sorry.. did I break it again?
[18:50:28] <psha> qq-: create chain in ordinar camview and safe it, then give to tiny
[18:51:16] <jepler> skunkKandT: I don't know what's going on there
[18:52:56] <micges> skunkKandT: try look at pid.3.feedback signal while this weirdness
[18:53:21] <jepler> do you have either of feedback-deriv or command-deriv hooked up?
[18:53:51] <skunkKandT> net bvel pid.3.feedback-deriv hm2_5i20.1.encoder.00.velocity
[18:54:09] <skunkKandT> should I unhook and see if it still jumps?
[18:54:18] <jepler> sure, it's worth a try
[18:54:43] <skunkKandT> it will take a second because I will have to re-tune a bit
[18:55:43] <qq-> psha, dinner
[18:57:05] <psha> qq-: pa
[18:58:59] <skunkKandT> cradek: it does it with it unhooked.
[18:59:03] <skunkKandT> heh
[18:59:13] <skunkKandT> jepler: it does it with it unhooked also
[18:59:29] <skunkKandT> second home is smooth.
[19:00:25] <psha> jepler: why ULAPI/RTAPI are not included in config.h and requred to be set in cmdline?
[19:03:20] <jepler> skunkKandT: if you want to finish excluding the recent pid changes as the cause of this behavior, you could check out the emc2.4 version of pid.c into your master
[19:03:30] <jepler> git checkout origin/v2.4:branch:src/hal/components/pid.c
[19:03:39] <jepler> err
[19:04:22] <psha> jepler: git diff | patch?
[19:04:27] <jepler> git checkout origin/v2.4_branchsrc/hal/components/pid.c
[19:04:28] <jepler> errrr
[19:04:31] <skunkKandT> heh
[19:04:32] <jepler> git checkout origin/v2.4_branch src/hal/components/pid.c
[19:04:36] <psha> :)
[19:05:01] <jepler> later to get back to the one from master: $ git checkout HEAD src/hal/components/pid.c
[19:05:32] <jepler> psha: RTAPI is defined for some files, ULAPI for others. some files are compiled once each way
[19:05:43] <jepler> psha: it indicates whether the current file is "realtime" or not
[19:05:45] <psha> jepler: thanks
[19:05:57] <jepler> hal_lib.c is an example of a file that is compiled once each way
[19:06:18] <psha> jepler: i'm still trying to compile 'one-for-all' version of video filter :)
[19:08:47] <skunkKandT> jepler: if I did it right - it does the same thing. what am I doing wrong I wonder?
[19:09:18] <jepler> skunkKandT: it doesn't ring a bell for me..
[19:10:01] <jepler> bbl
[19:12:03] <skunkKandT> ok - I will play with it.
[19:12:51] <skunkKandT> first fine http://imagebin.ca/img/5j6mrV6v.png
[19:12:53] <jepler> you did rebuild after the checkout, right?
[19:12:59] <skunkKandT> * first find
[19:13:43] <skunkKandT> second find http://imagebin.ca/img/G3B39N.png
[19:15:11] <skunkKandT> jepler: yes - but I am doing again with a make clea
[19:15:13] <skunkKandT> clean
[19:15:21] <skunkKandT> just to make sure.
[19:17:06] <jepler> if the feedback-deriv pin doesn't exist now, you did it right
[19:17:08] <jepler> bbl really this time
[19:21:20] <skunkKandT> still seems to do it
[19:21:23] <skunkKandT> odd
[19:58:04] <skunkKandT> here is with encoder position.. http://imagebin.ca/img/lfQ_AZ.png
[19:58:52] <skunkKandT> this is when it finds the new or initial indexl
[19:58:54] <skunkKandT> http://imagebin.ca/img/HTyUb7.png
[20:11:17] <qq-> psha, get Segmentation fault on camview , sry
[20:11:30] <psha> that's bad :)
[20:11:34] <psha> using cu-plugins?
[20:12:16] <qq-> psha, i dunno
[20:12:32] <psha> you'd just installed camview-tiny?
[20:12:42] <qq-> sure
[20:14:13] <psha> it's silently segfaulting or printing something before?
[20:14:15] <qq-> should i remove camview ?
[20:14:21] <psha> no
[20:14:33] <psha> what's your distro? squeeze?
[20:14:39] <qq-> segfault on save the chain
[20:14:46] <qq-> squeeze
[20:15:40] <psha> ah
[20:16:01] <psha> until saving it's working?
[20:16:30] <qq-> http://paste.debian.net/99243/
[20:16:45] <qq-> until saving works , see paste
[20:17:55] <qq-> but not show any videos
[20:18:13] <psha> replug camera
[20:20:59] <skunkKandT> jepler: could there be something with using the 7i48?
[20:21:11] <skunkKandT> the other axis still home normally to index
[20:21:19] <qq-> psha, nothing change
[20:22:48] <psha> still no camera?
[20:22:54] <qq-> i get for Format 640x480 UNRECOGNIZED
[20:23:30] <psha> and no other options in select?
[20:24:18] <qq-> 320x240 but UNRECOGNIZED too
[20:25:41] <skunkKandT> cradek - seems to lift and clamp correctly - now I just need to figure out the index
[20:25:58] <skunkKandT> I love th 'can't jog a locking axis...'
[20:27:23] <psha> hm
[20:27:34] <psha> other apps like mplayer, cheese, etc are working fine?
[20:28:02] <qq-> no , only working is qv4l2
[20:28:17] <psha> ah, thats better :)
[20:28:23] <psha> what is your cam?
[20:28:46] <qq-> a photo device
[20:29:12] <qq-> not a (web)cam
[20:29:13] <psha> v4l2 compatible though?
[20:29:20] <qq-> yes
[20:29:50] <psha> so let's bother with your cam next and now just check that camview-tiny is embedding
[20:29:56] <psha> unplug it for a while
[20:31:35] <psha> is camview still segfaulting?
[20:31:39] <psha> you may use example input
[20:31:42] <psha> instead of cam
[20:32:35] <qq-> no segfault with example input
[20:32:48] <psha> nice
[20:33:29] <psha> so if you want i may help you with adding support for your cam later
[20:33:54] <psha> now save chain and add camview -c chain-file -w {XID} to axis INIfile
[20:33:57] <psha> is it working?
[20:34:49] <qq-> mom
[20:36:53] <qq-> searching axis INIfile/
[20:37:13] <psha> i meen ini file of emc
[20:37:19] <psha> embedded tab for axis
[20:37:47] <qq-> ah in configs/
[20:39:16] <jepler> skunkKandT: I don't know whether cradek is using index homing on his 7i48 yet..
[20:39:33] <jepler> I think he has index from a resolver-to-encoder board, so I imagine he would
[20:39:46] <SWPadnos> huh. I just "found" my old laptop - the one that actually has a parallel port :)
[20:40:22] <SWPadnos> I wonder if it's sufficiently old to not have too much ACPI crap
[20:42:21] <skunkKandT> jepler: is there anything else I could scope?
[20:48:47] <qq-> psha, done , saved in /mc2-2.4.5/configs/halui_pyvcp/halui.ini
[20:50:15] <qq-> put it as last line in that file
[20:53:41] <jepler> skunkKandT: oh I feel like at this point I'd just be guessing..
[20:53:50] <skunkKandT> heh
[20:54:36] <skunkKandT> if I set the homing offset to 0 I get the same result.
[20:54:47] <skunkKandT> Which I guess makes sense.
[20:56:51] <jepler> if you loadrt pid ... debug=1 then you can directly scope the pid error, errorI and errorD to see which of them is going nuts
[21:02:22] <cradek> yes I've homed to index on the 7i48 - that's where my rotary table was hooke dup
[21:02:51] <psha> qq-: two lines needed, as described in wiki in axis_embed_video
[21:03:13] <qq-> i see
[21:04:22] <psha> btw what is your cam name?
[21:05:13] <qq-> USB Digital Still Camera
[21:05:27] <qq-> driver sq905c
[21:05:46] <skunkKandT> jepler: cradek: http://imagebin.ca/img/AJTOAp.png
[21:05:55] <skunkKandT> wonder what I am doing wrong
[21:06:57] <cradek> pastebin the whole hal file?
[21:07:06] <psha> gspca?
[21:07:41] <psha> qq-: support may be easily added
[21:08:07] <skunkKandT> hold on - that last screen shot seems to be missing things
[21:08:43] <psha> qq-: with not very correct colors though :)
[21:09:33] <skunkKandT> cradek: http://pastebin.ca/1984759
[21:10:44] <skunkKandT> (sorry about the mess)
[21:11:11] <cradek> this is axis.4 right?
[21:11:18] <skunkKandT> http://imagebin.ca/img/giNFtSx.png
[21:11:20] <skunkKandT> tes
[21:11:21] <skunkKandT> yes
[21:11:29] <skunkKandT> B
[21:12:58] <qq-> psha, gspca_sq905c ; gspca_main
[21:13:23] <psha> it has specifig BGGR pixel format not included in camunits
[21:13:31] <psha> but it may be added
[21:13:34] <psha> if you need it :)
[21:14:44] <qq-> 'we' need it, because camview segfault on save for it
[21:15:06] <psha> :)
[21:17:06] <micges> skunkKandT: can you paste also ini file?
[21:17:26] <cradek> skunkKandT: I don't see anything wrong with the hal...
[21:17:52] <psha> qq-: qv4l2 is woking?
[21:18:02] <qq-> yes
[21:18:41] <cradek> do you have [AXIS_4] MAXOUTPUT set right?
[21:19:14] <skunkKandT> http://pastebin.ca/1984770
[21:19:38] <skunkKandT> it is set to 10 I think atm
[21:20:10] <cradek> I'm not seeing the problem...
[21:20:37] <qq-> camview show in Chain (10-11 fps) for USB Digital Still Camera (v4l2) [Streaming] , but in Display say 'No renderable units in chain' '
[21:20:56] <psha> with 'unknown' output format?
[21:21:43] <qq-> no 'Unrecognized'
[21:21:44] <psha> it's ok since opengl renderer don't know how to decompress BGGR format
[21:23:32] <psha> i'll write libv4l-convert based filter tomorrow
[21:23:47] <qq-> cool, thanks
[21:25:52] <qq-> libv4l2-convert
[21:31:17] <skunkKandT> I am not either..
[21:37:16] <qq-> psha, i get Current format: unknown (0x43353039) , maybe will help
[21:38:46] <psha> it's four_cc('S', '9', '0' ,'5')
[21:39:00] <psha> '9', '0', '5', 'c'
[21:39:06] <psha> missed a bit )
[21:39:10] <skunkKandT> cradek: jepler: would you believe that upping the HOME_LATCH_VEL from .1 to 1 fixed it?
[21:39:18] <psha> it's supported bu liv5lconvert
[21:39:20] <psha> v4l
[21:39:26] <cradek> skunkKandT: not really
[21:40:21] <skunkKandT> could it be a servo thread issue and how slow it was finding an index?
[21:40:41] <skunkKandT> it has worked 100% of the time now.
[21:40:49] <cradek> huh.
[21:40:53] <cradek> I don't get that.
[21:41:30] <qq-> psha, inputs: 0 = sq905c;
[21:41:55] <skunkKandT> it would have been .1 deg/sec
[21:44:22] <qq-> psha, mplayer use : Selected driver: v4l2
[21:46:19] <psha> qq-: thanks. i think i now have enought info to write plugin
[21:46:30] <psha> but first need to clean v4lconvert a bit :)
[21:47:26] <qq-> psha, thanks to you ;)
[21:55:25] <micges> skunkKandT: if so can you run halscope with axis.4.motor-pos-cmd ?
[21:55:45] <micges> but with HOME_LATCH_VEL = .1
[21:56:45] <skunkKandT> micges: it will be a bit
[22:34:08] <jepler> what I think I see from the screenshot is that after index-enable falls, the error (command-feedback) is very large
[22:34:12] <jepler> it should stay close to 0
[22:34:44] <jepler> what's your ferror? is it big enough that an error of 2 (degrees?) wouldn't trip it?
[22:35:23] <jepler> scope command and feedback and index-enable .. see if they really both step when they should
[22:36:57] <skunkKandT> following error is 10 at the moment
[22:39:40] <skunkKandT> I don't think it is going to happen today. running out of time and it seems to be working (have the chain hooked up and everything)
[22:40:10] <cradek> video! video!
[22:52:08] <skunkworks> cradek: uploading now
[22:52:17] <skunkworks> ;)
[22:52:30] <SWPadnos> so, I had a problem connecting to IRC while at airports. It seems SASL ident wasn't working through tmobile wifi
[22:52:34] <SWPadnos> or my phone
[22:52:40] <skunkworks> other than I think I might want to add a delay on the up switch - it seems to be working well
[22:52:55] <SWPadnos> did anyone else notice a change at freenode, or is it me?
[22:54:49] <skunkworks> I have no clue what you are talking about ;)
[22:55:38] <SWPadnos> me either, I just know I couldn't connect :)
[22:55:48] <SWPadnos> (and the error said all those other words I used)
[22:59:49] <skunkworks> heh
[23:21:30] <skunkworks> http://www.youtube.com/watch?v=vLE4lzPcEo8
[23:21:36] <skunkworks> first movement is homing
[23:28:48] <micges> neat
[23:41:38] <skunkworks> http://www.youtube.com/watch?v=gpz1tgTRyGc
[23:41:49] <skunkworks> (tach isn't being used)
[23:42:49] <cradek> yayy!!
[23:43:21] <skunkworks> cradek: thanks for all your work on locking rotories :)
[23:43:28] <skunkworks> bbl
[23:44:27] <cradek> skunkworks: you're welcome
[23:45:00] <cradek> skunkworks: looks like it starts rotating before it's fully up?