#emc-devel | Logs for 2011-10-05

Back
[00:03:49] -!- theorbtwo has quit [Ping timeout: 240 seconds]
[00:03:59] <jepler> andypugh: it looks that way, yes.
[00:04:04] theorb is now known as theorbtwo
[00:04:34] <andypugh> Ok. I only tried it by accident (bad memory)
[00:05:07] <JT-Shop> does stepgen only allow up to 8 motors or is the man page wrong?
[00:05:19] <jepler> /sys/module/kvm_intel/sections$ grep -w ^kvm_intel /proc/modules ; cat .text
[00:05:19] <jepler> kvm_intel 49912 0 - Live 0xffffffffa0848000
[00:05:19] <jepler> 0xffffffffa0848000
[00:07:11] -!- robh__ has quit [Ping timeout: 248 seconds]
[00:07:43] <andypugh> JT-Shop: Sadly true, it seems.
[00:08:05] <andypugh> Which seems short-sighted with 9 axes
[00:08:24] <JT-Shop> is this incorrect with ctrl_type loadrt stepgen step_type=type0[,type1...] [ctrl_type=type0[,type1...]]
[00:08:38] <JT-Shop> should be ctrl_type=p or v?
[00:08:49] <JT-Shop> http://www.linuxcnc.org/docview/html/man/man9/stepgen.9.html
[00:09:03] <JT-Shop> or am I reading this wrong?
[00:09:12] -!- Tom_itx has quit [Ping timeout: 256 seconds]
[00:10:34] <jepler> #define MAX_CHAN 8
[00:10:41] <jepler> the docs are right but it might not be hard to increase the limit
[00:11:49] <andypugh> Seems like it should be at least 9, possibly 18 for 9-way gantries :-)
[00:11:51] <JT-Shop> that would make sense to have MAX_CHAN 9 with 9 axis control
[00:13:01] -!- Loetmichel has quit [Ping timeout: 240 seconds]
[00:13:28] <JT-Shop> at one time was EMC only 8 axis?
[00:13:47] <skunkworks> I think it was 6 for a long time
[00:13:54] <skunkworks> so 8 would give you 2 extra
[00:14:09] <andypugh> I think it is just programmer habit to think in terms of bytes of flags.
[00:29:49] -!- JT-Shop has quit [Read error: Connection reset by peer]
[00:33:51] -!- jthornton has quit [Ping timeout: 255 seconds]
[00:40:48] -!- jthornton [[email protected]] has joined #emc-devel
[00:40:54] -!- elmo40 has quit [Ping timeout: 260 seconds]
[00:41:15] -!- JT-Shop [[email protected]] has joined #emc-devel
[00:43:09] -!- Valen has quit [Quit: Leaving.]
[00:44:12] -!- i_tarzan_ has quit [Ping timeout: 255 seconds]
[00:50:42] -!- skunkworks has quit [Remote host closed the connection]
[00:57:42] -!- i_tarzan has quit [Ping timeout: 255 seconds]
[01:03:06] -!- jthornton has quit [Read error: Connection timed out]
[01:04:11] -!- jthornton [[email protected]] has joined #emc-devel
[01:04:45] -!- mhaberler [[email protected]] has joined #emc-devel
[01:10:24] -!- jthornton has quit [Ping timeout: 256 seconds]
[01:11:06] -!- jthornton [[email protected]] has joined #emc-devel
[01:18:46] -!- andypugh has quit [Quit: andypugh]
[01:19:57] -!- jthornton has quit [Ping timeout: 260 seconds]
[01:21:27] -!- jthornton [[email protected]] has joined #emc-devel
[01:32:02] -!- tom3p has quit [Quit: Ex-Chat]
[01:32:12] -!- mhaberler has quit [Quit: mhaberler]
[01:46:14] -!- stormlight has quit [Quit: stormlight]
[01:52:22] -!- mhaberler [[email protected]] has joined #emc-devel
[02:01:31] -!- stormlight has quit [Client Quit]
[02:13:21] -!- jthornton has quit [Read error: Connection timed out]
[02:21:45] -!- jthornton [[email protected]] has joined #emc-devel
[02:33:07] -!- ries has quit [Quit: ries]
[02:41:42] -!- jthornton has quit [Read error: Connection timed out]
[02:43:09] -!- jthornton [[email protected]] has joined #emc-devel
[02:50:33] -!- jthornton has quit [Ping timeout: 255 seconds]
[02:51:10] -!- jthornton [[email protected]] has joined #emc-devel
[02:59:55] -!- jthornton has quit [Ping timeout: 258 seconds]
[03:01:01] -!- jthornton [[email protected]] has joined #emc-devel
[03:14:04] -!- mhaberler has quit [Quit: mhaberler]
[03:16:12] -!- A2Sheds has quit [Ping timeout: 256 seconds]
[03:18:16] -!- FinboySlick has quit [Quit: Leaving.]
[03:20:36] -!- mhaberler [[email protected]] has joined #emc-devel
[03:29:06] -!- A2Sheds [A2Sheds!~ly@unaffiliated/l84supper] has joined #emc-devel
[03:32:59] -!- jthornton has quit [Read error: Connection timed out]
[03:35:04] -!- jthornton [[email protected]] has joined #emc-devel
[03:38:00] -!- stormlight has quit [Quit: stormlight]
[03:58:31] -!- stormlight has quit [Quit: stormlight]
[04:03:51] -!- theorbtwo has quit [Ping timeout: 252 seconds]
[04:24:17] -!- JT-Shop has quit [Max SendQ exceeded]
[04:26:50] -!- JT-Shop [[email protected]] has joined #emc-devel
[04:47:24] -!- psha[work] [psha[work][email protected]] has joined #emc-devel
[04:48:06] -!- i_tarzan_ has quit [Ping timeout: 255 seconds]
[04:55:37] -!- theos has quit [Ping timeout: 252 seconds]
[05:35:52] -!- vladimirek [[email protected]] has joined #emc-devel
[05:50:57] -!- vladimirek has quit [Remote host closed the connection]
[05:57:20] -!- ve7it has quit [Remote host closed the connection]
[06:09:51] -!- mhaberler has quit [Quit: mhaberler]
[07:14:46] -!- giz has quit [Quit: Page closed]
[07:19:58] -!- A2Sheds has quit [Quit: puff of smoke]
[07:23:10] -!- bootnecklad` has quit [Ping timeout: 244 seconds]
[07:26:10] -!- e-ndy [e-ndy!jkastner@nat/redhat/x-vpzskzvnoyawdksl] has joined #emc-devel
[07:31:52] -!- robh__ [[email protected]] has joined #emc-devel
[08:44:08] -!- jthornton has quit [Read error: Connection timed out]
[08:44:59] -!- jthornton [[email protected]] has joined #emc-devel
[08:55:07] -!- Tom_itx has quit []
[09:31:16] -!- mhaberler [[email protected]] has joined #emc-devel
[09:34:18] -!- i_tarzan has quit [Ping timeout: 255 seconds]
[10:30:44] -!- mhaberler has quit [Quit: mhaberler]
[11:00:55] -!- mhaberler [[email protected]] has joined #emc-devel
[11:03:00] -!- pjm has quit [Read error: Connection reset by peer]
[11:10:50] cpresser_ is now known as cpresser
[11:14:53] -!- mhaberler has quit [Quit: mhaberler]
[11:15:28] -!- steves_logging has quit []
[11:21:51] -!- pjm__ has quit [Ping timeout: 248 seconds]
[11:24:04] -!- mhaberler [[email protected]] has joined #emc-devel
[11:35:25] -!- mhaberler has quit [Read error: Connection reset by peer]
[11:36:18] -!- pingufan has quit [Remote host closed the connection]
[11:38:35] -!- SteveStallings [[email protected]] has joined #emc-devel
[11:38:58] SteveStallings is now known as steves_logging
[11:59:26] -!- mhaberler [[email protected]] has joined #emc-devel
[12:16:58] -!- mhaberler has quit [Quit: mhaberler]
[12:22:37] -!- theorbtwo has quit [Ping timeout: 276 seconds]
[12:25:01] -!- factor has quit [Quit: Leaving]
[12:47:49] -!- Valen has quit [Ping timeout: 240 seconds]
[12:52:27] -!- psha[work] has quit [Quit: leaving]
[13:09:51] -!- chester88 has quit [Ping timeout: 244 seconds]
[13:11:21] -!- ries has quit [Ping timeout: 276 seconds]
[13:23:30] -!- chester88 [[email protected]] has joined #emc-devel
[13:47:10] -!- skunkworks [[email protected]] has joined #emc-devel
[14:26:43] -!- Valen has quit [Quit: Leaving.]
[14:41:04] -!- JT-Shop_ [[email protected]] has joined #emc-devel
[14:43:50] -!- JT-Shop has quit [Ping timeout: 260 seconds]
[14:44:00] JT-Shop_ is now known as JT-Shop
[14:57:08] -!- theos has quit [Excess Flood]
[14:57:52] -!- psha [[email protected]] has joined #emc-devel
[15:10:23] -!- sarariman_seb [[email protected]] has joined #emc-devel
[15:16:16] <sarariman_seb> good morning
[15:17:25] <cradek> it's a beautiful day in the neighborhood ...
[15:18:12] -!- Connor has quit [Read error: Connection reset by peer]
[15:23:00] <sarariman_seb> my morning music this morning: http://www.youtube.com/watch?v=acf7rqpICmQ
[15:25:45] <cradek> homemade music on a homemade instrument: http://www.youtube.com/watch?v=Y1MtdD7HrbY
[15:26:44] <psha> heh, if it's time for youtube links...
[15:26:50] <psha> http://www.youtube.com/watch?v=xmPwr7V0QHE
[15:27:25] <psha> TLC was pushed from the road
[15:27:38] <cradek> what's a TLC?
[15:27:45] <psha> land cruiser
[15:27:50] <psha> toyota
[15:27:57] <psha> one of the most popular cars for offroad
[15:27:58] <cradek> good job avoiding the light poles and trees
[15:28:25] <cradek> was it a tire blowout?
[15:28:51] <psha> no, another car crashed in tlc from back
[15:28:56] <psha> and pushed it from the road
[15:29:00] <sarariman_seb> sketchy
[15:29:11] <sarariman_seb> that fretles banjo sounds great
[15:29:22] <cradek> yeah, I've really got to get (or make) one.
[15:33:00] <cradek> http://www.youtube.com/watch?v=xJq04V2IoHk
[15:33:09] <cradek> you can make them out of about anything.
[15:34:04] <sarariman_seb> that looks like a fork on the bottom of the gas can
[15:54:15] <cradek> yep I'm sure that's what it is
[15:56:08] -!- nullie has quit [Quit: Ex-Chat]
[16:04:54] -!- cevad has quit [Quit: Leaving]
[16:05:37] <archivist> I thought the can would give a definite sound but does not seem to
[16:19:50] <skunkworks> I have my grandfathers banjo that he had modified with a guitar neck.
[16:20:10] <skunkworks> my wife wants me to learn so we can sing around the camp fire :)
[16:20:37] <cradek> I think every possible combination of guitar, banjo, and mandolin parts exists
[16:20:41] <archivist> hmm beer...camp fire...bliss
[16:25:04] <psha> \
[16:25:10] <psha> \
[16:25:33] <sarariman_seb> psha is doing "the wave" by himself
[16:25:41] <cradek> /
[16:25:54] <psha> heh, that's was psha jr.
[16:25:58] <psha> 3yrs old :)
[16:26:27] <sarariman_seb> hi jr! we're eagerly looking forward to your first commit to our repo! :-)
[16:26:53] <skunkworks> sarariman_seb: shouldn't your kids have git access?
[16:27:25] <sarariman_seb> heh
[16:27:40] <sarariman_seb> this winter i'm planning to teach python to my daughter (7 yrs)
[16:27:56] <sarariman_seb> we'll probably not use git right away tho� one thing at a time
[16:28:03] <skunkworks> heh
[16:28:14] <skunkworks> whatever happened to gwbasic?
[16:28:16] <skunkworks> ;)
[16:28:39] <sarariman_seb> it went the way of the floppy, and good riddance
[16:28:51] <sarariman_seb> i wonder if there's an IBM PC emulator for my phone...
[16:28:52] <archivist> buried 6 feet under
[16:29:43] <cradek> I *bought* floppies a couple weeks ago
[16:29:57] * sarariman_seb chokes on his coffee
[16:30:05] <cradek> ... after only being able to find one in the house when I needed one, and it (unsurprisingly) turned out to be bad
[16:30:17] <sarariman_seb> but� WHY
[16:30:25] <skunkworks> still need it for some computers to update the bios...
[16:30:33] <cradek> (they were new! 20 for $.99 at goodwill)
[16:30:48] <archivist> ! what usb not good enough?
[16:31:10] <cradek> to get very old source code for a very old microcontroller off a very old computer
[16:31:23] <sarariman_seb> just scp it off over your wifi
[16:31:24] <sarariman_seb> oh wait
[16:31:27] <cradek> ha
[16:32:00] <archivist> clean the heads on the drive
[16:32:15] <cradek> that's what the first floppy does...
[16:32:31] <skunkworks> heh
[16:32:41] <archivist> old floppies can leave a residue rather than clean
[16:33:27] -!- e-ndy has quit [Quit: Ex-Chat]
[16:33:32] <sarariman_seb> i had a friend who went on a trip to nepal and wanted to ssh back home from internet cafes, he brough a floppy with three copies of putty, and was down to one good copy by the end of the trip but it did work for him
[16:33:45] <archivist> oxide comes off and sticks to the head like shit to a blanket
[16:40:19] <cradek> what a pleasant image that is
[16:42:50] <psha> sarariman_seb: heh, my brother instead of floppies brought strange infection from nepal :)
[16:43:37] <psha> and spend next several weeks in the hospital - doctors were looking at him and wait until his slightly higher temperature become normal
[16:44:27] -!- pingufan has quit [Remote host closed the connection]
[16:58:03] -!- ve7it [[email protected]] has joined #emc-devel
[16:58:52] <sarariman_seb> All archivists' talk of� oxide� makes me think of michelle bachmann (jackpot!): http://www.youtube.com/watch?v=LFB6LQ1-WKU
[17:03:45] -!- stormlight has quit [Quit: stormlight]
[17:11:25] -!- stormlight has quit [Ping timeout: 255 seconds]
[17:13:34] <skunkworks> sarariman_seb: you watched msnbc last night... ;)
[17:14:42] <sarariman_seb> i didnt, actually - was that on there?
[17:15:05] <skunkworks> yes - one of the shos
[17:15:07] <skunkworks> hows
[17:15:10] <skunkworks> shows
[17:15:10] <skunkworks> there
[17:15:28] <sarariman_seb> it was on boingboing recently
[17:17:55] <sarariman_seb> last night was tuesday, so hackspace night
[17:18:13] <sarariman_seb> helped a buddy cut a bunch of steel bar for a lathe he's building
[17:18:21] * sarariman_seb likes bandsaws
[17:20:04] <skunkworks> heh
[17:20:54] <skunkworks> I hole sawed through a 1-1/2" thick piece of aluminum last night.. (so I would not have to mill it all away)
[17:21:08] <skunkworks> 1-3/4 hole
[17:21:14] <sarariman_seb> wow
[17:21:30] <sarariman_seb> coolant?
[17:21:40] <skunkworks> a little bit of wd40
[17:22:18] <sarariman_seb> and now you have a chunk of round for the scrap pile, instead of a pile of chips for the trash can :-)
[17:22:19] <skunkworks> the trick that I learned from dad is you drill a hole all the way through the peice where the teeth cut so the shavings have a place to go.
[17:22:30] <skunkworks> otherwise it plugs up pretty quick
[17:22:37] <Jymmm> (denatured alcohol)
[17:22:40] <sarariman_seb> that sounds like a good idea
[17:23:06] <sarariman_seb> Jymmm, you use alcohol for cooling cuts in Al?
[17:23:36] <Jymmm> sarariman_seb: you can only use it ONCE ;)
[17:24:18] <Jymmm> ... in a lifetime that is
[17:25:21] <Jymmm> sarariman_seb: (just kidding)
[17:25:38] <sarariman_seb> i got too ambitious and took too big a cut once, Al cooled with wd40, and smoked up the garage
[17:25:55] <sarariman_seb> set off the fire alarm and woke my family up at like 1 in the morning :-(
[17:26:37] <skunkworks> ouch
[17:26:58] -!- IchGuckLive has quit [Ping timeout: 260 seconds]
[17:27:07] IchGuckLive_ is now known as IchGuckLive
[17:27:36] <Jymmm> sarariman_seb: Sounds perfect for denatured alcohol. Use the friction heat as a preheater, lets the alcohol burner much hotter once ignited
[17:28:29] <Jymmm> I've been looking at making some alcohol stoves lately =)
[17:28:38] <sarariman_seb> for camping maybe?
[17:29:05] <Jymmm> for whatever, they're so tiny and super easy to store in a pack/mess kit
[17:29:46] <Jymmm> just get a tuna can, punch some holes it the side and you have a stove
[17:30:02] <sarariman_seb> i've seen them made from beer cans
[17:31:11] <Jymmm> Yeah, me too. But saw the tuna can one too and seems perfect. But I am going to try the soda can ones too. Saw a super easy one that seemd to work better than most.
[17:31:50] <Jymmm> the tuna can fits PERFECT under my mess kit cup
[17:34:00] nowox is now known as nowox_
[17:34:33] -!- stormlight has quit [Ping timeout: 252 seconds]
[17:34:33] stormlight_ is now known as stormlight
[17:35:00] nowox_ is now known as nowox
[17:37:23] <Jymmm> sarariman_seb: http://www.youtube.com/watch?v=apgjCfevGNg
[17:53:10] -!- psha has quit [Quit: Lost terminal]
[17:56:40] -!- psha [[email protected]] has joined #emc-devel
[17:58:40] -!- Calyp has quit [Ping timeout: 260 seconds]
[18:19:12] -!- piasdom has quit [Ping timeout: 252 seconds]
[18:38:20] -!- vladimirek [[email protected]] has joined #emc-devel
[18:55:07] -!- psha has quit [Quit: Lost terminal]
[18:55:09] -!- ewidance [[email protected]] has joined #emc-devel
[18:55:51] <ewidance> Hi all!
[18:58:31] -!- andypugh [andypugh!~andy2@cpc2-basl1-0-0-cust1037.basl.cable.virginmedia.com] has joined #emc-devel
[19:05:31] <ewidance> andypugh, may I ask a question? I'm currently reading (and trying) your work on tangentkins...
[19:05:57] <ewidance> However I got some problems on "You can then swap "trivkins" for "tangentkins" in your HAL and INI files. "
[19:06:54] <ewidance> When I replace trivkins for tangentlons in my machine hal file, ther's no more axis naming.. and no more moves...
[19:07:35] <andypugh> Odd
[19:08:05] <andypugh> If you start emc from the command line, do you get any error reports (just type "emc")
[19:08:49] <ewidance> Also, i don't see what to do in ini file... there's no reference to trivkins...
[19:08:52] <ewidance> I try..
[19:10:45] <ewidance> EMC2 - 2.4.6
[19:10:45] <ewidance> Machine configuration directory is '/home/jpcivade/emc2/configs/borsh9060'
[19:10:45] <ewidance> Machine configuration file is 'borsh9060.ini'
[19:10:45] <ewidance> Starting EMC2...
[19:10:46] <ewidance> INFO CLASSICLADDER- No ladder GUI requested-Realtime runs till HAL closes.
[19:10:57] <ewidance> That's all...
[19:11:58] <sarariman_seb> ewidance, the kins is specified in the hal file
[19:13:05] <ewidance> Yes, i've seen it. It's the first line. But andy said in his post : "You can then swap "trivkins" for "tangentkins" in your HAL and INI files. " (http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/catid,10/id,13083/lang,russian/ )
[19:13:29] -!- Guest184 has quit [Client Quit]
[19:15:48] <ewidance> I dont know many on kinematics, but i was thinking trivkins was doing a lot of thinks for XYZ. Since i only want A to be affected, there's no risk to replace trivkins with tangentkins?
[19:16:32] <andypugh> You do need to actually compile and install tangentkins, but it does only mess with A.
[19:16:57] <andypugh> trivkins actually does precisely nothing to any axis, that is what makes it trivial.
[19:22:40] <ewidance> tangentkins is compiled & installed. Thanks for pointing trivkins is doing notthing.
[19:24:55] <andypugh> Well, not _exactly_ nothing, it passes values from axis positions to joint positions and back again, but is just a pass-through.
[19:32:41] <ewidance> When i load tangentkins instead of trivkins in my machine hal file (the only one i modified...), in Axis, axis are not anymore named XYZA (instead, i've got 0/1/2/3);
[19:33:18] <ewidance> I verified with halmeter M151 & M152 change tangentkins.auto value. That's ok.
[19:33:46] <ewidance> In M152 mode, everything is ok, except axis namings.
[19:34:42] <ewidance> in M151 mode (tangent 'active'), Axis a gos to value '0.2' and... EMC2 fall in error ("Join 3 folowing error").
[19:36:29] <andypugh> Ah! yes, you get that with any non-trivial kinematics.
[19:36:57] <sarariman_seb> do you know why that is andy?
[19:37:10] <andypugh> You need to home in "joint mode" (0,1,2,3) then switch to "world" mode (XYZA)
[19:37:27] <sarariman_seb> yes...
[19:38:13] <ewidance> Sorry???
[19:38:16] <andypugh> It might be possible to make tangentkins an inverse-only kins (or some other combination) and not have it do that.
[19:39:03] <ewidance> What is join mode / world mode?
[19:40:53] <cradek> inverse and forward don't match, so this kins isn't going to work very well.
[19:41:22] <andypugh> ewidance: It makes more sense if you consider a robot arm. In "joint mode" jogging joint 0 would rotate the waist, jogging joint 1 might rotate the shoulder, and 2 the elbow, etc. Then in world mode jogging the X axis moves the tool in the X direction, moving all the joints the right amount to move the tool in a straight line in space.
[19:41:23] <cradek> you certainly don't get a following error with any nontriv kins.
[19:42:20] <andypugh> I know tangentkins is a bit of a hack. I have absolutely no intention of trying to release it.
[19:42:49] <cradek> this has been discussed a hundred times before, and this approach is doomed. you need to put the smarts in the interpreter (before motion interpolation) or directly in gcode.
[19:42:52] <andypugh> ewidance: You can switch to world mode using "$" once the joints are homed.
[19:43:52] <cradek> with this kins, any time you stop and then later go a different direction you're going to get a jump in A value, which will of course give you a following error.
[19:43:53] <andypugh> The kins approach only has one redeeming feature, you can install it without recompiling EMC2.
[19:44:09] -!- A2Sheds has quit [Remote host closed the connection]
[19:44:31] <andypugh> Yeah, you need to basically give A a 720 degree f-error limit.
[19:44:48] <cradek> um, yeah
[19:45:02] <cradek> I don't see how that's useful for any real world machine
[19:45:55] <ewidance> stil error.. even after going to world mode ($)... (i first quit, relauch, homed, switch to world , M151 and run...
[19:46:11] <cradek> what is M151?
[19:46:31] <ewidance> http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/id,13083/catid,10/limit,6/limitstart,6/lang,russian/
[19:46:40] <cradek> "In this moment i'm thinking to calculate the angle with a external python module and generate right g-code, ..."
[19:46:43] <andypugh> It's a user m-code to switch between trivkins and tangent
[19:46:56] <cradek> ^ this is an approach that will actually work
[19:46:58] <andypugh> A preprocessor is a better solution.
[19:47:10] <cradek> this kins approach will not work
[19:47:22] <ewidance> It's a command used to change the value of an hal ping, for activating / disactivating tangent mode...
[19:48:41] <andypugh> It's for a knife, which I am assuming can spin very quickly.
[19:49:13] <cradek> generating tangency for your gcode is not hard, and you can solve for both lines and arcs, and emc can do the motions
[19:49:22] <andypugh> If you do have any very sharp changes in direction, then you could switch back to normal, point A in the right direction, and switch back.
[19:50:49] <andypugh> cradek: True, and you don't even have to solve for intermediate points in the simple-A case. The original application was to attempt automatic kerf correction with a BC head, and that is not so simple.
[19:51:39] <cradek> you'll have to figure out what to do in corners, and code it explicitly in the gcode
[19:51:52] <cradek> lift and turn? make the corner into a small arc?
[19:51:57] <cradek> same for start/stop of cuts
[19:52:01] <ewidance> Andy : so what.
[19:52:06] <andypugh> ewidance: You can make the A f-error limit extremely big to get rid of the following error message. As the others are pointing out it does mask an underlying problem with the concept.
[19:52:11] <cradek> a kins approach cannot do any of that
[19:52:33] <ewidance> Do i have to abandon this approache for tangent knife?
[19:53:17] <cradek> well you don't have to do anything, but I believe the whole approach has fundamental problems
[19:54:15] <andypugh> I agree, but I think it can work for some applications.
[19:54:50] <andypugh> The correct solution is to put the calculations of A in tc, but that is much more intrusive.
[20:07:31] -!- logger[psha] [logger[psha][email protected]] has joined #emc-devel
[20:07:55] <andypugh> For arcs, yes. A bit more tricky for splines, but then I don't think EMC2 actually does those.
[20:09:08] <skunkworks> I was think the gcode would be arcs or small line segments
[20:09:10] <skunkworks> or both
[20:10:45] <skunkworks> the pre-processor would go through the gcode and pre-postion the head when needed and calculate the a rotation for arcs and line segments. (I am sure it is a bit more complicated then that but it seems easy enough)
[20:10:54] <andypugh> I think it is, but there is a spline add-in (in master?) which converts to arcs internally, but appears as splines in the G-code (I think)
[20:11:13] <skunkworks> oh - that might get tricky probably
[20:11:39] -!- ve7it has quit [Remote host closed the connection]
[20:11:48] <skunkworks> the pre-processor would have to convert the splines to small line segments ;)
[20:15:21] <andypugh> Actually, splines seem to be in 2.4? It's certainly in 2.5 (I just tried typing G5.3 in MDI) :-)
[20:16:27] -!- ries has quit [Read error: Connection reset by peer]
[20:16:27] ries_ is now known as ries
[20:17:16] <andypugh> But no mention in the "Quick reference". Seems to be here though: http://linuxcnc.org/docs/html/gcode_main.html#r1_7
[20:17:35] <ewidance> ANdy ; tried for ferror :
[20:17:36] <ewidance> FERROR = 1000
[20:17:36] <ewidance> MIN_FERROR = 1000
[20:17:47] <ewidance> always error...
[20:19:19] -!- ries has quit [Read error: Connection reset by peer]
[20:19:32] <andypugh> Can you pastebin the HAL file?
[20:19:51] <andypugh> ( www.pastebin.com )
[20:21:12] -!- ries has quit [Read error: Connection reset by peer]
[20:21:13] ries_ is now known as ries
[20:22:32] <ewidance> little success... I tried with 360/360, and slowing down the speed... it WORKS!
[20:22:44] -!- ries has quit [Read error: Connection reset by peer]
[20:22:45] ries_ is now known as ries
[20:23:10] <andypugh> You might still find a problem on spirals.
[20:23:56] <andypugh> The simple solution (though not a very elegant one) is to short-circuit the feedback.
[20:24:53] <andypugh> Put the axis.3.feedback pin on the same line as the axis.3.motor-pos-cmd line, and comment out (#) the original feedbac line.
[20:26:09] -!- ries has quit [Read error: Connection reset by peer]
[20:26:10] ries_ is now known as ries
[20:27:54] <ewidance> I think my problem is more on acceleration on A axis. I'm using a real A axis (instead of doing a quick hack with a stand alone stepper). Acceleration time is setup relatively high. That might be the problem, and the reason why when i slowdown (speed corrector) that works.
[20:29:02] <ewidance> I'll do more investigations with a quick stepper, to see if limits are pushed away (not sure it's english : i'm french ;) ).
[20:30:01] <andypugh> The real problem has been pointed out by cradek and seb, that calculating your tangent on the basis of the current move is a bit too late. A pre-processor to modify the G-code is a better idea.
[20:31:04] -!- ries has quit [Read error: Connection reset by peer]
[20:34:26] <andypugh> But, as EMC2 typically doesn't move very far in 1ms, you get the new direction pretty quickly.
[20:35:28] -!- ries_ has quit [Ping timeout: 248 seconds]
[20:35:53] -!- ries has quit [Read error: Connection reset by peer]
[20:35:53] ries_ is now known as ries
[20:37:57] <ewidance> except if it's an important move from old to new position, and acceleration avoid stepper to follow quickly the new direction... Am I right?
[20:38:17] -!- skunkworks has quit []
[20:39:38] <andypugh> If you know it is a sharp corner, use your M151 and position the knife explicitly
[20:40:16] -!- ries has quit [Ping timeout: 248 seconds]
[20:40:30] <andypugh> You can set up EMC2 to run a filter file on the input G-code, so if you do have a preprocessor filter, it would be run automatically.
[20:40:31] -!- bootnecklad` has quit [Ping timeout: 276 seconds]
[20:41:11] <andypugh> That requires programming, but is quite a simple filter.
[20:41:19] <ewidance> You read in my mind.. I was just writing "But i'll have to write Gcode manually ... "
[20:41:55] <andypugh> If you are filtering the code, then doing _all_ the A moves that way makes sense.
[20:42:47] <andypugh> But, you might find that tangentkins works well enough.
[20:45:53] -!- giz has quit [Quit: Page closed]
[20:46:28] <ewidance> I'll try... Anyway, I wish t thank you very much for your assistance and explanations... I've learned a lot tonight about things i did not imagine in EMC2!
[20:47:57] <andypugh> No problem, and feel free to come back if you find you need a more robust solution.
[21:04:52] <ewidance> OK! (i'm at this time reading an interessting thread on emc-devel list, between cradek & Eric H Johnson... always the same people!)
[21:04:53] <andypugh> Talking about kinematics.... Does anyone have any more clue than me about the genserkins problem that Francesca has on the mailing list? He/She seems to be saying that the same issue occurs with the Puma560 sample config.
[21:05:20] <andypugh> Beter question, does anyone have less of a clue than me?
[21:09:35] -!- cjdavis has quit [Read error: Connection reset by peer]
[21:10:22] -!- FinboySlick has quit [Quit: Leaving.]
[21:17:56] <sarariman_seb> i just tried running the puma560 config in sim and it sort of worked
[21:18:03] <sarariman_seb> i got motion, but also warnings on stdout:
[21:18:18] <sarariman_seb> ERRkineInvers(joints: �), (iterations=100)
[21:18:33] <sarariman_seb> with the � actually being 6 floats
[21:19:24] <sarariman_seb> oh yeah, what andy said on the list
[21:21:56] <sarariman_seb> vismach is cool
[21:27:23] <sarariman_seb> francesca needs to post his broken configs or we can't help at all...
[21:30:19] -!- bootnecklad has quit [Ping timeout: 252 seconds]
[21:33:25] -!- cjdavis has quit [Read error: Connection reset by peer]
[21:46:50] <andypugh> I think they claim that the puma560 is broken the same way. Does everything work in sim?
[21:48:01] <andypugh> (and Francesca is a girls name)
[21:48:27] <sarariman_seb> whoops
[21:48:40] <sarariman_seb> puma560 ran in sim, but printed tons of errors
[21:49:08] <andypugh> I got no errors in a normal install. So that _might_ be the problem
[21:49:24] <sarariman_seb> "normal install" == realtime?
[21:49:29] <andypugh> Yes.
[21:49:48] <andypugh> A LiveCD install in a VM that I keep for debugging.
[21:49:57] <sarariman_seb> i'm booting my rtai vm now
[21:50:13] <andypugh> (So, I get a lot of realtime errors)
[21:57:51] -!- stormlight has quit [Quit: stormlight]
[22:02:56] <sarariman_seb> yeah, no warnings in realtime mode, but lots in sim
[22:03:15] <sarariman_seb> from kins
[22:03:16] <andypugh> Is that unexpected?
[22:03:28] <sarariman_seb> i'd expect kins to work equally well in sim as in realtime
[22:04:11] <andypugh> Hmm, my machine just crashed during a make: make[1]: Leaving directory `/usr/src/linux-headers-2.6.32-122-rtai'
[22:04:11] <andypugh> Killed
[22:04:11] <andypugh> make[1]: *** [_module_/home/andypugh/emc2-dev/src] Killed
[22:08:49] <sarariman_seb> oom?
[22:09:35] <andypugh> Went I went upstairs to it (which I need to do a lot) it had dropped out to command line, and seemed to be bust with fsck and lmsensors
[22:10:20] <sarariman_seb> ouch
[22:11:15] <andypugh> And now I need to go upstairs again, but this is probably a coding fault.
[22:11:53] <andypugh> (trying to pull a null-terminated string in from a sserial device, plenty of scope for problems)
[22:40:26] -!- nullie has quit [Quit: Ex-Chat]
[22:50:48] -!- isssy has quit [Quit: Visitor from www.linuxcnc.org]
[23:05:43] -!- sarariman_seb has quit [Quit: Leaving]
[23:07:58] -!- stormlight has quit [Quit: stormlight]
[23:26:23] -!- servos4ever has quit [Quit: ChatZilla 0.9.85 [SeaMonkey 2.0.11/20101206162726]]