#linuxcnc | Logs for 2012-04-25

Back
[00:06:07] <ReadError> hey hai hay
[00:19:17] -!- cylly2 [[email protected]] has joined #linuxcnc
[00:21:59] -!- Loetmichel has quit [Ping timeout: 265 seconds]
[00:23:35] -!- andypugh has quit [Quit: andypugh]
[00:25:17] <alex4nder> ReadError: sup
[00:30:22] -!- factor has quit [Remote host closed the connection]
[00:31:44] <Jymmm> Anybody want a free shipping container?
[00:32:09] <Tom_itx> bury it in your yard for dooms day
[00:33:18] -!- factor [factor!~factor@r74-195-218-67.msk1cmtc02.mskgok.ok.dh.suddenlink.net] has joined #linuxcnc
[00:34:10] <Jymmm> why?
[00:34:12] <ReadError> alex4nder: u gonan be here for a sec?
[00:34:22] <ReadError> going to try to get this z axis goin
[00:34:27] <ReadError> only thing that needs fixin
[00:37:56] -!- stevegt_1 has quit [Ping timeout: 244 seconds]
[00:39:21] <JT-Shop> Tom_itx: http://s47.photobucket.com/albums/f163/johnplctech/Plasma%20Cutter/
[00:43:18] <Tom_itx> did you put pc fans behind yours?
[00:43:20] <Tom_itx> 80mm
[00:44:54] -!- toast2 [[email protected]] has joined #linuxcnc
[00:45:02] -!- toastydeath has quit [Ping timeout: 245 seconds]
[00:49:59] -!- toast2 has quit [Ping timeout: 272 seconds]
[00:51:23] -!- toastydeath [[email protected]] has joined #linuxcnc
[01:04:51] <Jymmm> SWPadnos: ping
[01:05:36] -!- phantoxeD has quit []
[01:08:59] <alex4nder> ReadError: yah
[01:13:13] -!- rob_h has quit [Ping timeout: 260 seconds]
[01:17:35] -!- phantoxeD [[email protected]] has joined #linuxcnc
[01:23:52] <ReadError> alex4nder, having a bad time trying to get this z axis set
[01:24:01] <ReadError> it goes in the write direction, but code gives errors
[01:24:04] <ReadError> saying out of range
[01:24:11] <ReadError> i did -6 and 0
[01:24:22] <ReadError> home at -4 (2" above bottom)
[01:24:45] <ReadError> then, i set the touchoff value to 2
[01:24:52] <ReadError> since its 2" above the piece
[01:25:01] <ReadError> and it goes up really high
[01:25:16] <ReadError> or should it be -4 for touchoff?
[01:25:26] <ReadError> since its from the Z, which exist above the machine?
[01:26:17] <Tom_itx> sounds like it, 2 would be greater than 0
[01:26:39] <ReadError> but, all these values are positive
[01:27:12] <ReadError> the emc says linear move on line 589 exceeds limit
[01:27:27] <ReadError> its these dang negative numbers i think
[01:28:03] <Tom_itx> are you using fixture offsets?
[01:28:08] <Tom_itx> g54 etc
[01:28:12] <ReadError> oh no
[01:28:15] <ReadError> i left that default
[01:28:21] <ReadError> which should i be using?
[01:28:25] <jdhnc> what is on line 589
[01:28:32] <Tom_itx> g54 is the default
[01:28:47] <ReadError> its the default emc demo jdhnc
[01:28:52] <ReadError> ide paste but no internet on it yet
[01:29:02] <jdhnc> move z to about halfway then touch off
[01:29:28] <jdhnc> if it is trying to move to a save Z height (positive) and you don't touch off, anything positive will be out of range.
[01:29:41] <jdhnc> the 2.4 emc demo or the 2.5 linuxcnc demo?
[01:29:54] <ReadError> 2.4
[01:31:23] <ReadError> when i home the z
[01:31:26] <ReadError> even though its -4
[01:31:31] <ReadError> it sets it to 0.5
[01:31:37] <ReadError> not sure how its getting that number
[01:31:51] <ReadError> since im bound between -6 and 0
[01:32:33] <ReadError> i dont have any tools set
[01:32:41] <jdhnc> when you touch off, it sets zero to wherever you are (unless you enter a non-zero value)
[01:32:42] <ReadError> so im not sure where it would be getting these extra numbers from
[01:32:51] <ReadError> i did home though
[01:32:57] <ReadError> and thats what set the +0.5
[01:33:02] <ReadError> home should be -4
[01:33:03] <jdhnc> line 589 on my demo is just a scaled y/z
[01:33:12] <jdhnc> do you have a home offset?
[01:33:27] <ReadError> yea -4
[01:33:42] <ReadError> which is 2" above my table
[01:33:52] <ReadError> assuming -6 is the table..
[01:35:11] <ReadError> now im getting, linear move on line 22 would exceed joint 2s limit
[01:35:17] <ReadError> since its seeing it as a positive number
[01:35:34] <jdhnc> did you move x halfway and then touch off?
[01:35:38] <ReadError> and 0 is the "largest" number
[01:35:49] <ReadError> yea my "home" is in the middle of x, y
[01:35:53] <ReadError> then z is 2" above
[01:35:58] <jdhnc> that's not home
[01:36:12] <ReadError> thats what i put in the config ;/
[01:36:18] <jdhnc> or it is home, but that's not what I meant.
[01:36:32] <jdhnc> jog z halfway up, hit <end><enter>
[01:38:54] <ReadError> ok that put it at 0
[01:39:08] <jdhnc> now run it
[01:39:18] <ReadError> error on 22 and 589 ;/
[01:40:00] <alex4nder> hmm
[01:40:11] <jdhnc> 22 is joint 0 or 1?
[01:42:39] <alex4nder> ReadError: what I'm using on my Taig is I have Z as 0 to 5.5 .. and home at 5.5"
[01:43:06] <alex4nder> when I home my machine I bring Z up right before the gib screw hits the bearing block
[01:43:14] <alex4nder> and I home Z
[01:44:08] <alex4nder> you should be able to go all the way up to that point, and 5.5" down
[01:44:58] <jdhnc> I think my top Z is +(max travel)
[01:45:53] <Tom_itx> it should be
[01:46:10] <jdhnc> anybody know anything about ER collets?
[01:46:30] <jdhnc> like, are these ER or some random thing: http://www.cdcotools.com/index.php (on sale, top right)
[01:46:43] <ReadError> oh
[01:46:50] <ReadError> everyone was saying that z should be negative
[01:46:53] <ReadError> -6 0
[01:47:11] <ReadError> well thats what i got from joes config
[01:47:18] <ReadError> this positive axis seems to be better
[01:48:20] <Tom_itx> jdhnc, http://www.maritool.com/Collets-ER-Collets/c21_56/index.html
[01:49:00] <alex4nder> ReadError: it's all relative
[01:49:58] <Tom_itx> alex4nder is that G90 or G91?
[01:50:06] <alex4nder> haha
[01:53:44] <jdhnc> tim: nice, but severe price difference
[01:54:42] <Tom_itx> i posted it more for a visual aid
[01:55:37] <ReadError> awwwww yea
[01:55:41] <ReadError> we got it goin ;)
[01:55:45] <ReadError> screw that negative stuff
[01:56:03] <ReadError> how much fail will happen if i try to draw a png?
[01:56:52] <Tom_itx> jdhnc i would call them
[01:57:08] <Tom_itx> the mt2 bothers me a bit
[01:57:19] <Tom_itx> the drawbar helps
[01:57:39] <jdhnc> I have an r8 mill
[01:57:50] <Tom_itx> ok
[01:57:59] <Tom_itx> doesn't use bp collets?
[01:58:40] <jdhnc> I have a set of R8 collets, but to change them, you need a lot of Z room
[01:59:03] <Tom_itx> well yeah
[01:59:46] <jdhnc> I have an MT3 mill also (never used) that I could just get the mt3 arbor for
[02:01:47] <Tom_itx> if you get a set would it matter that they're ER?
[02:02:07] <Tom_itx> they look like it but..
[02:02:20] <jdhnc> no, only in that I would like to be able to get a matching collet holder for my other mill to share the collets
[02:02:44] <Tom_itx> i'd just call em
[02:03:58] <jdhnc> I sent them an email. I've impulse bought enough stuff this week.
[02:04:10] <ReadError> http://www.youtube.com/watch?v=pquiohEcamw&feature=plcp&context=C498f0f3VDvjVQa1PpcFPbUL-GhX6iZGUSGOozqxynljokN1Lh7uw%3D
[02:04:15] <ReadError> how we lookin? ;o
[02:04:25] <ReadError> 1st automated draw
[02:04:29] <ReadError> w/o holding the marker
[02:04:31] <jdhnc> woohoo!
[02:04:36] <ReadError> ;oooooo
[02:04:40] <ReadError> i need a stiffer tip
[02:04:45] <ReadError> im doing another one right now ;)
[02:04:47] <jdhnc> I hear that a lot
[02:04:58] <Jymmm> ReadError: viagra?
[02:05:01] <jdhnc> you need to upgrade emc
[02:05:01] <ReadError> cardboard isnt too flat either
[02:07:09] <alex4nder> ReadError: congrats
[02:08:06] <Tom_itx> ReadError wow! 2 in one night?
[02:10:22] -!- skunkworks__ [skunkworks__!~chatzilla@str-bb-cable-south-3-102.dsl.airstreamcomm.net] has joined #linuxcnc
[02:10:27] <ReadError> i want to do "wtf jackie chan"
[02:10:35] <ReadError> not sure how well it would convert to gcode
[02:13:52] <jdhnc> ReadError: http://pastebin.com/06YLy1u8
[02:14:52] <ReadError> awwwww yea
[02:14:56] <ReadError> let me get my flash drive
[02:15:51] <Tom_itx> jdhnc what do you use for txt?
[02:16:30] <jdhnc> that was DeskEngrave
[02:17:13] <jdhnc> but usually cad + cut2d or just simple stuff directly in cut2d
[02:24:45] -!- dimas_ has quit [Ping timeout: 265 seconds]
[02:26:16] -!- sumpfralle1 has quit [Ping timeout: 256 seconds]
[02:27:32] <ReadError> these 1-2-3 blocks are awesome
[02:27:52] <jdhnc> they're as easy as 1-2-3!
[02:28:47] <toastydeath> wait till you see 2-4-6 blocks!
[02:29:00] <jdhnc> they are surpsingly pricy
[02:29:11] <toastydeath> we always made that kind of thing on our own
[02:29:23] <toastydeath> everyone in the shop had a pair of bigass parallels
[02:29:55] <toastydeath> 12" long, 1/5
[02:30:12] <toastydeath> er, 1/2" wide and just under whatever the standard kurt vice jaw is tall
[02:30:20] <toastydeath> i think it gave 1/8th of jaw
[02:31:15] <ReadError> i got 4 blocks for 20$ i think
[02:31:26] <jdhnc> 123's are usually really cheap
[02:33:15] <ReadError> http://s3.amazonaws.com/ragefaces/3fa9aa4e19889110c01bae7601d70e25.png
[02:33:21] <ReadError> can you do that jdhnc
[02:33:29] <ReadError> like no larger than 3-4"
[02:33:44] <ReadError> getting some crazy stuff with the import in EMC
[02:33:52] <jdhnc> Read: how crazy?
[02:34:10] <ReadError> like a bunch of lines
[02:34:25] <ReadError> trying again with some different settings now
[02:34:52] <ReadError> on the bright side, 1 more day of work!
[02:35:04] <jdhnc> using a pen?
[02:35:06] <jdhnc> what's the tip size?
[02:35:13] <ReadError> unfortunately i have to wake up in about 5.5 hours
[02:35:16] <ReadError> its a sharpie
[02:35:23] <jdhnc> 1/16"?
[02:35:27] <ReadError> sure
[02:35:31] <ReadError> that sounds about riight
[02:37:44] -!- demacus_ [[email protected]] has joined #linuxcnc
[02:40:54] -!- dimas_ [[email protected]] has joined #linuxcnc
[02:41:24] -!- demacus has quit [Ping timeout: 245 seconds]
[02:44:20] <jdhnc> that was a bust
[02:44:31] <jdhnc> way too painful to play with for long.
[02:54:06] -!- skunkworks__ has quit [Remote host closed the connection]
[02:58:50] -!- ve7it has quit [Remote host closed the connection]
[03:00:17] <ReadError> do you have a copy ?
[03:00:21] <ReadError> or is it payware?
[03:00:36] <jdhnc> of what?
[03:01:12] <ReadError> deskengraver
[03:01:24] <jdhnc> google it, it's free (windows)
[03:01:32] <jdhnc> or try fengrave
[03:02:53] <ReadError> is there something simple that converts monochrome images to toolpath?
[03:03:36] <jdhnc> image2gcode
[03:03:52] <ReadError> http://www.youtube.com/watch?v=9AoC4s9HPo4
[03:03:55] <ReadError> ;o
[03:04:36] <jdhnc> crank up the feed rate
[03:05:07] <jdhnc> nice clamp for the pen.
[03:05:55] <jdhnc> http://www.youtube.com/watch?v=ELfONtG_Zkk&feature=related
[03:06:51] <ReadError> that alex4nder's
[03:13:04] -!- jthornton has quit [Read error: Connection reset by peer]
[03:13:04] -!- JT-Shop has quit [Read error: Connection reset by peer]
[03:13:33] -!- JT-Shop [[email protected]] has joined #linuxcnc
[03:13:38] -!- jthornton [[email protected]] has joined #linuxcnc
[03:27:58] -!- factor has quit [Ping timeout: 245 seconds]
[03:31:34] -!- factor [factor!~factor@r74-195-216-151.msk1cmtc02.mskgok.ok.dh.suddenlink.net] has joined #linuxcnc
[03:32:43] -!- jsr__ has quit [Quit: Leaving]
[03:34:24] -!- ktchk [[email protected]] has joined #linuxcnc
[03:55:29] -!- phantoxeD has quit [Read error: Connection reset by peer]
[04:10:57] -!- phantoxeD [[email protected]] has joined #linuxcnc
[04:16:05] -!- toastydeath has quit [Read error: Connection reset by peer]
[04:16:18] -!- toastydeath [[email protected]] has joined #linuxcnc
[04:17:09] -!- the-jub [[email protected]] has joined #linuxcnc
[04:17:22] <archivist> JT-Shop, I use http://www.arceurotrade.co.uk/Catalogue/Stepper-Motors
[04:24:06] <alex4nder> yoh
[04:26:46] <alex4nder> ReadError: hey, it's not a requirement, but I think you need some more gap in your couplers
[04:27:38] <alex4nder> it might be worth checking.
[04:48:29] <Thetawaves> http://learnyousomeerlang.com/static/img/un-bound.png
[04:55:36] -!- joe9 [[email protected]] has joined #linuxcnc
[05:03:03] -!- Fox_Muldr has quit [Ping timeout: 265 seconds]
[05:04:57] -!- Fox_Muldr [[email protected]] has joined #linuxcnc
[05:31:01] -!- mhaberler [[email protected]] has joined #linuxcnc
[05:40:04] -!- mk0 [[email protected]] has joined #linuxcnc
[06:00:09] -!- psha[work] [psha[work][email protected]] has joined #linuxcnc
[06:01:07] -!- phantoxe [[email protected]] has joined #linuxcnc
[06:01:13] e-ndy|afk is now known as e-ndy
[06:01:37] -!- phantoxe has quit []
[06:12:28] -!- factor has quit [Ping timeout: 260 seconds]
[06:14:05] -!- factor [factor!~factor@r74-195-216-151.msk1cmtc02.mskgok.ok.dh.suddenlink.net] has joined #linuxcnc
[06:38:10] -!- mhaberler has quit [Quit: mhaberler]
[06:43:57] -!- tiago [[email protected]] has joined #linuxcnc
[07:06:26] Jymmm is now known as MrAsshole
[07:06:32] MrAsshole is now known as Jymmm
[07:07:35] -!- stevegt_1 [[email protected]] has joined #linuxcnc
[07:15:19] -!- DJ9DJ [DJ9DJ!~Deejay@unaffiliated/dj9dj] has joined #linuxcnc
[07:15:29] <DJ9DJ> moin
[07:21:17] -!- joe9 has quit [Quit: leaving]
[07:24:25] cylly2 is now known as Loetmichel
[07:24:40] <Loetmichel> mornin'
[07:27:04] -!- maximilian_h [[email protected]] has joined #linuxcnc
[07:27:05] -!- maximilian_h [[email protected]] has parted #linuxcnc
[07:37:05] <alex4nder> hey
[08:08:37] -!- rob_h [[email protected]] has joined #linuxcnc
[08:41:00] -!- Valen [[email protected]] has joined #linuxcnc
[09:00:50] -!- asdfasd [[email protected]] has joined #linuxcnc
[09:13:06] -!- mhaberler [[email protected]] has joined #linuxcnc
[09:15:08] -!- n2diy has quit [Ping timeout: 240 seconds]
[09:22:10] -!- fliebel [[email protected]] has joined #linuxcnc
[09:50:08] -!- the-jub has quit [Remote host closed the connection]
[09:50:32] -!- the-jub [[email protected]] has joined #linuxcnc
[09:51:27] -!- the-jub has quit [Remote host closed the connection]
[10:00:40] -!- the-jub [[email protected]] has joined #linuxcnc
[10:44:58] -!- mhaberler has quit [Quit: mhaberler]
[11:03:02] -!- PCW has quit [Ping timeout: 246 seconds]
[11:03:51] -!- the-jub has quit [Remote host closed the connection]
[11:13:43] -!- the-jub [[email protected]] has joined #linuxcnc
[11:22:56] -!- PCW [[email protected]] has joined #linuxcnc
[11:27:07] -!- psha[work] has quit [Quit: Lost terminal]
[11:50:37] <Thetawaves> what was the name of that guy who did polyester granite spindle?
[11:59:21] -!- mhaberler [[email protected]] has joined #linuxcnc
[12:03:20] <A0Sheds> Thetawaves, from Australia? Had the howto with pics of his steps?
[12:08:33] <Thetawaves> yeah
[12:16:38] -!- mhaberler has quit [Read error: Connection reset by peer]
[12:17:17] -!- mhaberler [[email protected]] has joined #linuxcnc
[12:18:26] <A0Sheds> Thetawaves, http://www.cnczone.com/forums/general_metal_working_machines/118358-phenolic_basalt_head_our_hm45.html
[12:22:08] <Jymmm> Hmmmm, errr, ummm.... http://www.youtube.com/watch?v=ST3A4hQH8Ao&feature=relmfu and as a bonus, visit the website in the comments
[12:37:06] <Valen> whats all this about my mill?
[12:40:41] * archivist introduces Thetawaves to Valen
[12:41:38] -!- factor has quit [Ping timeout: 240 seconds]
[12:42:58] -!- WalterN has quit [Quit: Leaving]
[12:46:05] -!- factor [factor!~factor@r74-195-216-151.msk1cmtc02.mskgok.ok.dh.suddenlink.net] has joined #linuxcnc
[12:46:24] -!- WalterN [[email protected]] has joined #linuxcnc
[12:58:06] * Valen goes to bed ;->
[12:58:31] <Valen> but next time your on Thetawaves thats my mill lol
[13:00:08] <Jymmm> Does anyone play with RC Li-Ion batteries/chargers?
[13:00:26] <Valen> a fair bit why
[13:00:48] <Jymmm> Valen: I came across this, http://www.dealextreme.com/p/parallel-balance-charging-adaptor-board-67683
[13:01:05] <Valen> mmmk
[13:01:13] <Jymmm> Is the multi-connecotr plugs to charge and sense the voltage of each cell?
[13:01:34] <Jymmm> I'm ot understnad ing the zillions of connectors on the charger itself
[13:01:39] <Jymmm> s/ot/not/
[13:01:52] <Valen> that paticular one looks like you plug a bunch of packs into it, it then hooks them all into parallel then you join it to the charger by the wires on the back
[13:01:59] <Valen> thats not a charger
[13:02:04] <Valen> thats just an adaptor
[13:02:16] <Valen> most "modern" battery packs and chargers have something like that
[13:02:26] <Valen> but not for multiple packs
[13:02:40] <Jymmm> I say a balanced charger with a LCD giving each cells voltage
[13:02:44] <Jymmm> s/say/saw/
[13:03:03] <Valen> yep, generally you come out of that into an adaptor board
[13:03:14] -!- factor has quit [Ping timeout: 265 seconds]
[13:03:21] <Jymmm> but I didn't see those 4-8pin connectors, just the heavy 2pin
[13:03:23] <Valen> the adaptor board connects one battery main power wire and balance plug to a format the charger is happy with
[13:03:31] <Valen> url for charger?
[13:04:12] <Jymmm> looking
[13:04:48] <Valen> http://www.hobbyking.com/hobbyking/store/__15168__HobbyKing_ECO8_150W_7A_8S_Bal_Dis_Cyc_Charger_w_acc_USA_Warehouse_.html say is how they generally come
[13:04:55] <Valen> note the adaptor board
[13:05:26] <Jymmm> ah, so it's BOTH charger and adapter board
[13:05:41] <Valen> the one you sent me is like a bunch of adaptor boards
[13:05:53] <Valen> it can connect say 5 packs in parallell
[13:05:58] <Jymmm> ah
[13:06:01] <Valen> so you can charge and balance them at the same time
[13:06:07] <Jymmm> are these typically faster chargers?
[13:06:13] <Jymmm> fast
[13:06:24] <Valen> you can wind them up to whatever you want
[13:06:25] <jdhnc> I'd be wary of charging large Li battery packs with anything from DX in the middle.
[13:06:28] <Valen> generally 5A
[13:06:35] <Valen> jdhnc: there is that... lol
[13:06:59] <Jymmm> I just saw it for the first time
[13:07:49] <Jymmm> Valen: so you can slow charge them too?
[13:08:00] <Valen> .1a to 5-10a generally
[13:08:08] <Valen> depending on the total wattage
[13:08:35] <Jymmm> Valen: Ok, I'm thinking the old school N-Cad packs that doubled as stoves when being charged =)
[13:08:55] <Valen> lipo charge shouldnt get hot
[13:09:01] <Valen> otherwise they catch fire
[13:09:13] -!- mhaberler has quit [Ping timeout: 246 seconds]
[13:09:22] <Valen> the firefighting technique is to remove combustable materials from the surrounding area and watch it burn
[13:09:30] <Jymmm> Yeah, I dind't think so, and these charge LiFe too?
[13:10:14] <Valen> some will
[13:10:24] <Valen> keep in mind most are 12v powered
[13:10:28] <Valen> anyway gtg to bed
[13:10:40] <Jymmm> G'Night thanks
[13:11:00] <Valen> i got one recently with mains supply
[13:11:00] <Jymmm> Valen: I'm thinking solar here
[13:11:06] <Valen> is muchly a good idea
[13:11:12] <Valen> how do you mean?
[13:11:24] <Jymmm> Valen: I have a 15W solar panel
[13:11:43] <Valen> if you want to charge batteries with it these are not what you want
[13:12:09] <Valen> they need a fixed input voltage between 11 and (generally) 14 volts
[13:12:32] <Jymmm> I DO want LiPo/LiFe, just solar charge them
[13:12:32] <Valen> you want a mppt tracker + battery charging unit
[13:12:49] <Jymmm> mppt
[13:12:55] -!- mhaberler [[email protected]] has joined #linuxcnc
[13:12:58] <Valen> i know you want lithium, but these chargers probably won't work well with solar as an input
[13:13:02] <Valen> maximum power point
[13:13:22] -!- psha[work] [psha[work][email protected]] has joined #linuxcnc
[13:13:45] <Valen> anyway night
[13:13:47] <Jymmm> Hmmm http://www.timnolan.com/index.php?page=arduino-ppt-solar-charger
[13:13:50] <Jymmm> GNight
[13:13:51] -!- Valen has quit [Quit: Leaving.]
[13:14:18] <Jymmm> jdhnc: Ah come on... you can ALWAYS trust DX, HF, etc =)
[13:14:39] -!- skunkworks [[email protected]] has joined #linuxcnc
[13:14:55] e-ndy is now known as e-ndy|afk
[13:14:59] -!- mhaberler has quit [Read error: Connection reset by peer]
[13:16:05] -!- mhaberler [[email protected]] has joined #linuxcnc
[13:18:40] -!- the-jub has quit [Quit: ill be back soon! :-/ ....maybe..]
[13:19:16] -!- mhaberler_ [[email protected]] has joined #linuxcnc
[13:20:38] -!- mhaberler has quit [Ping timeout: 240 seconds]
[13:20:39] mhaberler_ is now known as mhaberler
[13:21:13] e-ndy|afk is now known as e-ndy
[13:34:25] -!- FinboySlick [[email protected]] has joined #linuxcnc
[13:38:55] e-ndy is now known as e-ndy|afk
[13:45:53] -!- Tom_itx has quit [Ping timeout: 245 seconds]
[13:46:29] -!- zlog has quit [Ping timeout: 248 seconds]
[13:48:12] -!- psha[work] has quit [Quit: Lost terminal]
[13:49:16] -!- zlog [[email protected]] has joined #linuxcnc
[13:49:24] -!- TekniQue has quit [Ping timeout: 252 seconds]
[13:49:57] e-ndy|afk is now known as e-ndy
[13:51:28] -!- TekniQue [[email protected]] has joined #linuxcnc
[13:51:28] -!- TekniQue has quit [Changing host]
[13:51:29] -!- TekniQue [TekniQue!baldur@unaffiliated/teknique] has joined #linuxcnc
[13:52:17] -!- mhaberler_ [[email protected]] has joined #linuxcnc
[13:52:17] -!- mhaberler has quit [Read error: Connection reset by peer]
[13:52:17] mhaberler_ is now known as mhaberler
[13:55:34] <Connor> Hey guys, is it possible to read a input single for RPM and spit the same clocking signal back out ?
[13:56:04] -!- mhaberler has quit [Client Quit]
[13:56:14] <Connor> I want to get RPM info from my speed controller, which already has built in display and output it to EMC.
[13:57:27] <cpresser> Connor: most VFD can be accessed via modbus. there you can read not only the RPM, but also the load and various other parameters
[13:57:51] <Connor> cpresser: not a VFD, it's a stock speed controller for the G0704 mill.
[13:58:54] * cpresser knows nothing about that specific speed controller.
[13:59:04] <Connor> I have access to the clk and gnd for the IR interrupter. When I hook a scope on it, it alters the signal going to the already installed RPM display and if gives more RPM than there really is..
[14:00:15] <Connor> So, I figured I could grab the signal, input it into EMC, then output a "duplicate" signal back to the existing external RPM indicator.
[14:00:45] <Connor> Hell, it may be easier to just install a 2nd speed indicator.
[14:03:00] <cpresser> so its a pulse counter? if yes, you can hook this up in emc with a single port-pin. use a 'encoder' hal-component for the counting
[14:04:48] <Connor> yea. it's a pulse pin.
[14:05:01] <Connor> but, then, I need to re-output the same pulse..
[14:05:30] <Connor> which may, or may not work...
[14:05:46] <cpresser> why would you want to re-output it ( by the way, its no problem 're-outputting' a signal)
[14:06:00] <Connor> to send back to the external display.
[14:06:47] <cpresser> just put the external display and emc2 on the same electrical network!?
[14:06:47] <Connor> I think splicing off the signal degrades it to much and the existing external display was miss counting because of it.
[14:06:51] <jdhnc> read the speed on the screen instead of the display
[14:07:12] <Connor> jdhnc: I want my cake and want to eat it too! :)
[14:07:20] <cpresser> if there is already an electrical problem you should fix that first^^
[14:07:21] -!- mhaberler [[email protected]] has joined #linuxcnc
[14:07:29] <jdhnc> did you try adding a resistor between the split signal and the emc input?
[14:08:01] <Connor> jdhnc: No. Like I said, I didn't even hook it up to EMC yet.. I was just checking the signal with my O-Scope.
[14:08:21] <Connor> and noticed that when the probes were on, it altered the signal.
[14:08:28] <cpresser> shild the cable.. move it so it gets lesser interference
[14:08:34] <jdhnc> high impedence scope probe?
[14:08:45] <cpresser> add a driver to the signal line
[14:09:01] <Connor> maybe. It's a cheap digital O-Scope from Parallax that hooks to the PC.
[14:09:11] <cpresser> make sure its a voltage and not a current driver
[14:09:13] <jdhnc> is it a single pulse per rev?
[14:09:34] <Connor> jdhnc: No, it's like 16 or more.
[14:09:52] <Connor> I haven't counted all the holes yet.
[14:11:53] <cpresser> i would suggest that you add a simple buffer/driver directly after the encoder. a 74LS06 or similar will do
[14:13:29] <jdhnc> I have a 1000 line encoder waiting for mine but no idea how to mount it
[14:19:12] -!- adb has quit [Ping timeout: 245 seconds]
[14:24:39] -!- joe9 [[email protected]] has joined #linuxcnc
[14:28:55] <Connor> jdhnc: 1000 line encoder? Dang, you planning to do spindle synchronization ?
[14:29:36] <joe9> ReadError: any luck with the mills. I was having some computer issues and hence was not on irc all the time.
[14:30:19] -!- factor [factor!~factor@r74-195-216-151.msk1cmtc02.mskgok.ok.dh.suddenlink.net] has joined #linuxcnc
[14:30:58] -!- hm2-buildmaster has quit [Quit: buildmaster reconfigured: bot disconnecting]
[14:30:59] -!- linuxcnc-build has quit [Quit: buildmaster reconfigured: bot disconnecting]
[14:33:21] -!- mozmck has quit [Quit: Leaving.]
[14:34:10] -!- adb [[email protected]] has joined #linuxcnc
[14:34:27] -!- mozmck [mozmck!~moses@client-74.117.92.175.dfwtx.partnershipbroadband.com] has joined #linuxcnc
[14:35:34] <skunkworks> jdhnc: you just do what ever it takes.. :) http://www.youtube.com/watch?v=5vvP4L_hr90
[14:38:20] <archivist> interesting related vid http://www.youtube.com/watch?v=8wt5EV2LMcc&NR=1&feature=endscreen
[14:40:44] -!- mhaberler has quit [Read error: Connection reset by peer]
[14:41:29] <jdhnc> Connor: was planning on it... and for feedback for the spindle speed control
[14:43:53] -!- viesturs [[email protected]] has joined #linuxcnc
[14:44:42] <viesturs> hello, folks!
[14:45:15] -!- syyl [[email protected]] has joined #linuxcnc
[14:46:14] <viesturs> I am trying to tune servos and I do not understand, why the following error for moving in one direction is huge - above 20 mm, while it is small for the opposite direction - less than 1 mm
[14:46:43] <cradek> that does sound strange. tell us about it.
[14:47:10] <viesturs> where should I start?
[14:47:28] <cradek> tell us about all the hardware?
[14:49:54] <viesturs> 3 pieces of Keling KL23BLS-115 servos (Nema 23) and USDigital E8 encoders with 512 cpr attached to 2 pieces of Mesa 7i39 LV, which are connected to 5i23. All that running on 2.5.0
[14:50:40] <archivist> uni directional friction, is this problem on all axes etc
[14:51:05] <viesturs> machine has ballscrews, I made my own design of fixed ballscrew and rotating nut.
[14:51:13] <cradek> so AC servos and dumb 3 phase drive
[14:51:53] <viesturs> archivist - I do not yet know, because that is gantry machine, I am starting with the single motor, I was hoping to do that first and then copy the settings to gantry motors, test and just slightly adjust
[14:52:15] <cradek> I don't know much about this brushless/bldc stuff
[14:52:36] <viesturs> cradek - yes, I do not have any smart drives to blame
[14:52:43] <cradek> what gives the phasing information?
[14:53:04] <cradek> it sounds like you have a fundamental problem with phasing/driving and you should not be worrying about pid yet
[14:53:57] <viesturs> In HalScope I see that when jogging in positive direction, the following error is accumulating. It is jerking, but the overall tendency is pretty straight line
[14:54:43] <viesturs> in negative direction the ferror increases to some 0,4-0,5 mm and stay there for whole move (except the point, where acceleration changes)
[14:55:07] <cradek> it sounds like you have a fundamental problem with phasing/driving and you should not be worrying about pid yet
[14:55:24] <cradek> what kind of feedback do you have to the bldc component?
[14:55:38] <viesturs> hall sensors and encoders
[14:56:21] <cradek> what is your bldc component cfg?
[14:57:08] <viesturs> cfg=qh,qh,qh
[14:58:57] <cradek> it's not obvious to me from the bldc manpage that you can use both q and h
[14:59:40] <viesturs> I set that up last year with kind help from Andy Pugh
[15:00:20] <viesturs> the way I understand it is that it uses hall sensors to find the initial rotor position and then it uses encoder further on
[15:00:36] <cradek> wow there are 47 hall patterns
[15:00:48] <viesturs> theoretically I could use cfg = qhi, where i would be index of the encoder
[15:00:59] <viesturs> but I do not have index on encoders
[15:01:09] <viesturs> actually 48
[15:01:19] <cradek> yes
[15:01:22] <viesturs> a lot to test...
[15:01:30] <cradek> I don't think I can be any help about this
[15:01:36] <cradek> I don't have any relevant experience
[15:02:01] <viesturs> unfortunately Keling has not listed, what is hall patern of this particular or any other Nema 23 brushless motor
[15:02:04] <cradek> but your asymmetry makes me think you have a bldc problem not a pid problem
[15:03:38] <skunkworks> viesturs: unrelated: did you have the issue where you had a top speed limit that you could not get past? Did you figure out? (brushless motor and mesa stuff)
[15:04:12] <viesturs> I think that it is related, because it is the same machine
[15:04:30] <viesturs> no, I still have not the machine working
[15:05:43] <viesturs> I had all the wiring, motors etc removed and had them at home, got the 5i23 card back, tested as much as I could, everything was working fine, today I arrived, spent half day installing all the wiring back, now I just need to tune the servos and machine should be ready
[15:05:43] <skunkworks> hmm.. You may have to wait for andy to show up.. He is the expert on the bldc componant...
[15:06:03] <viesturs> yes, actually I was hoping to meet him here
[15:06:11] <cradek> I don't agree you have just a tuning problem
[15:06:57] <viesturs> chris, I do not reject possibility of being something wrong with bldc config
[15:07:53] <viesturs> I guess I will try to take a look at those hall paterns and try some other, I recall that I had 2 or 3 that seemed ok, chose one of them, but I might had made a wrong choice
[15:08:44] <skunkworks> I think there should only be 2 that work... (one would make the motor turn backwards...)
[15:09:38] <jdhnc> http://www.pmdcorp.com/downloads/app_notes/BrushlessSensorConfig.pdf
[15:09:42] <viesturs> could be, do not really remember, it was 6 or 7 months ago
[15:16:47] <viesturs> thanks for the link, I will have some reading to do now
[15:18:29] -!- mk0 has quit [Read error: Connection reset by peer]
[15:18:45] <joe9> which is the better/faster cpu for rtai/low latency: 1st cpu: http://codepad.org/5WxuJVWr , 2nd cpu: http://codepad.org/zDenDBd4
[15:18:58] <joe9> the celeron is faster, but everyone disses them.
[15:19:15] <joe9> not sure how that translates to rtai/latency performance though.
[15:19:26] <joe9> anyone running linuxcnc on a celeron?
[15:19:39] -!- WalterN has quit [Read error: Connection reset by peer]
[15:20:21] -!- WalterN [[email protected]] has joined #linuxcnc
[15:20:24] -!- DJ9DJ has quit [Remote host closed the connection]
[15:20:59] -!- DJ9DJ [DJ9DJ!~Deejay@unaffiliated/dj9dj] has joined #linuxcnc
[15:27:01] -!- H264 [[email protected]] has joined #linuxcnc
[15:30:32] -!- WalterN has quit [Ping timeout: 256 seconds]
[15:32:40] -!- ve7it [[email protected]] has joined #linuxcnc
[15:42:05] -!- Tom_itx [Tom_itx!~Tl@unaffiliated/toml/x-013812] has joined #linuxcnc
[15:52:36] e-ndy is now known as e-ndy|afk
[15:54:45] <pcw_home> unsymmetrical drive is almost guaranteed to be a commutation problem (BLDCs idea of the rotor position is wrong)
[15:55:47] <viesturs> ok, I guess I will have to retry all the bldc patterns
[16:00:15] <pcw_home> Are you using the latest LinuxCNC 2.5? there were some bugs in BLDC as well (fixed sometime around December last year IICRC)
[16:00:28] -!- tiago has quit [Remote host closed the connection]
[16:03:02] <viesturs> I have 2.5.0
[16:03:06] <viesturs> official release
[16:06:52] -!- psha [[email protected]] has joined #linuxcnc
[16:08:33] <pcw_home> That should be OK
[16:10:39] -!- syyl_ws [[email protected]] has joined #linuxcnc
[16:15:37] -!- erasmo [[email protected]] has joined #linuxcnc
[16:25:59] -!- mhaberler [[email protected]] has joined #linuxcnc
[16:30:47] -!- mhaberler has quit [Read error: Connection reset by peer]
[16:31:21] -!- mhaberler [[email protected]] has joined #linuxcnc
[16:32:56] -!- mhaberler has quit [Read error: Connection reset by peer]
[16:33:05] -!- mhaberler [[email protected]] has joined #linuxcnc
[16:33:22] -!- motioncontrol [motioncontrol!~io@host213-78-dynamic.3-79-r.retail.telecomitalia.it] has joined #linuxcnc
[16:35:40] <viesturs> another strange thing with bldc:
[16:35:42] <jdhnc> joe9: if latency is acceptable for your application then either shold suffice.
[16:35:47] <viesturs> I set the cfg=h
[16:36:04] <viesturs> disconnected bldc.0.rawcounts pin
[16:36:29] <viesturs> commented out poles and scale parameter lines
[16:36:46] <viesturs> added setp bldc.0.value 0,5
[16:37:06] <viesturs> and it moves just a little, when I press f2 and nothing else
[16:37:20] <viesturs> I tried also lower values, but nothing
[16:37:52] <viesturs> if I comment out the line to set bldc value and connect it to pid, I can jog the motor
[16:38:22] -!- ktchk [[email protected]] has parted #linuxcnc
[16:39:18] <viesturs> what coud I be missing?
[16:50:55] <jdhnc> both motors/drives/encoder give the same results?
[16:52:34] <viesturs> how do You mean that?
[16:53:07] <jdhnc> you have two of everything? do both sets do the same thing?
[16:54:45] -!- vladimirek [[email protected]] has joined #linuxcnc
[16:55:06] <viesturs> I have 3 servos. 2 of them are on gantry, so I do not touch them yet, unless somebody can advice me, how to tune them simultaneously totally from scratch. I would like to tune the one on gantry bridge at first, so that I have a starting for the gantry motors
[16:58:15] <pcw_home> what is your PWM frequency and deadtime?
[16:58:17] <pcw_home> (deadtime should be set to 0 and PWM frequency say 20 KHz)
[17:00:13] <viesturs> where do I set deadtime?
[17:00:23] <viesturs> pwm frequency is 30000
[17:00:48] <viesturs> if the deadtime means "deadband" PID parameter, then it is 0
[17:00:55] <pcw_home> man hostmot2
[17:04:31] -!- IchGuckLive [[email protected]] has joined #linuxcnc
[17:06:31] <viesturs> I have not specified the deadtime. I will put it to 0 and test, once I will finish going through all the hall patterns.
[17:09:09] <viesturs> Peter, before starting any tricks, I better shoud ask - what is recommended deadtime value for 7i39?
[17:19:01] <pcw_home> 0
[17:23:27] -!- mhaberler has quit [Quit: mhaberler]
[17:24:47] <pcw_home> with the default 5 uSec each edge deadtime (10 uSec total) you would have to get to 30% drive before you had _any_ motor voltage (at 30KHz = 33.3 usec period, 10 uSec is a long time to be off)
[17:24:49] <pcw_home> This may also account for asymmetry with a 7I39 as the duty cycle will go down to 0 but only up to 70% or so
[17:27:27] -!- mhaberler [[email protected]] has joined #linuxcnc
[17:27:34] <kb8wmc> IchGuckLive: good day sir
[17:27:41] <pcw_home> Deadtime is exclusively for bare MOSFET/IGBT/MODULE driving but is not needed/wanted for the 7I39 which only has PWM inputs and does its own deadtime management
[17:28:52] <IchGuckLive> kb8wmc: hi
[17:29:08] <kb8wmc> IchGuckLive: http://www.n2yo.com/?s=31789 a link for you to track one of the two Private orbiting space stations
[17:29:30] <IchGuckLive> i know it
[17:30:03] <kb8wmc> ok....I should have realized your were already aware of it
[17:31:11] <IchGuckLive> pcw_home: do you accept paypal for the sale of your cards ?
[17:31:20] <pcw_home> No
[17:31:58] <IchGuckLive> thats bad for me
[17:31:59] <pcw_home> We did at one time. our experience is that they are close to criminal
[17:32:30] <IchGuckLive> i know they charged 40USD fpr sale from China
[17:32:59] <viesturs> Peter, thank You for the explanation, I have set the deadtime to 0 for all 3 pwm gens. Now I will put the belt back on to try out. Going through bldc patterns did not give anything, because I tried to jog the motor after changing the pattern, but for many cases it did nothing. I expect that wrong sign at scale value or something similar is to blame. It would work, if motor would turn with setp bldc.0.value [number between 0 and 0,95 here]
[17:33:03] <IchGuckLive> but it wars orderd on friday and shiped on thuesday that is faster then from Berlin To munich
[17:34:07] <IchGuckLive> pcw_home: one of the CNC lathe broak down and the part needed only available in china
[17:35:40] <IchGuckLive> this turret parts ->http://www.cncmakers.com/uploads/allimg/100929/1_100929144950_1.jpg
[17:35:46] <pcw_home> we dont have much trouble shipping to/from China from here but Germany is a pain due to huge amounts of red tape
[17:36:32] <IchGuckLive> yes and the parts moved from the costum servise is then doubled in pay
[17:37:18] <IchGuckLive> so the cit on you is 199USD +40USD shipment +40USD pay +150USD VAT Tax
[17:37:34] <pcw_home> Thats a funny looking part (something dark inside)
[17:37:49] <pcw_home> Yow
[17:38:19] <archivist> rubber?
[17:38:38] <archivist> flex drive part?
[17:38:39] <IchGuckLive> its the electrical connector for the 4way toolchanger
[17:39:00] <IchGuckLive> 110USD
[17:39:22] <IchGuckLive> oh no 55USD i ordert 2
[17:39:36] -!- Rogge [[email protected]] has joined #linuxcnc
[17:40:02] <pcw_home> so much for free trade
[17:40:34] <IchGuckLive> http://www.cncmakers.com/uploads/allimg/101009/1_101009150530_1.jpg thats the toolchanger
[17:41:20] <IchGuckLive> pcw_home: if the payment total reatches 150 USD it is 85% more
[17:42:14] <IchGuckLive> i think i will order via NY state buero
[17:42:39] <pcw_home> Of course you are in Europe so you may actually get some services for all your taxes...
[17:43:00] <viesturs> Peter, thank You so so so much! Ferror is symmetrical and does not exceed 4 mm!
[17:43:21] <viesturs> Since I have only P parameter specified, that is to be expected, right?
[17:43:46] <pcw_home> yes you will need FF1 and D and I
[17:43:47] <IchGuckLive> pcw_home: how mutch is the overall cost for the 5I25 kit with the 7I76 to NY
[17:45:37] -!- djdelorie has quit [Quit: Leaving]
[17:45:48] <pcw_home> If its cheapest shipping (USPS) maybe $215 or so
[17:46:09] <IchGuckLive> ok i will mail you if im in order
[17:46:16] -!- djdelorie [[email protected]] has joined #linuxcnc
[17:46:49] <djdelorie> can linuxcnc directly drive the PWMs for a bldc?
[17:47:05] <viesturs> I was thinking about tuning the velocity loop. Since I am doing servo tuning for the first time in my life, the dumb questions: do I specify it somewhere (like a PID parameter) or is it specified by the set of P, I or D used?
[17:48:51] -!- djdelorie has quit [Client Quit]
[17:49:10] <pcw_home> djdelorie yes
[17:49:24] <Jymmm> pcw_home: he just quit
[17:50:38] <pcw_home> well you can make a separate velocity loop but I dont think its worth the trouble
[17:51:01] <pcw_home> Think of the D term as being the velocity part of the loop
[17:51:27] <pcw_home> also make sure you use the fb-deriv input to the PID loop
[17:51:49] -!- djdelorie [[email protected]] has joined #linuxcnc
[17:51:58] -!- IchGuckLive has quit [Quit: ChatZilla 0.9.87 [Firefox 11.0/20120310193829]]
[17:53:32] <Jymmm> djdelorie: "pcw_home: djdelorie yes"
[17:54:01] <pcw_home> Andy's BLDC comp does it
[17:56:41] <viesturs> Peter, just to make sure I understand correctly: I should connect encoder.n.velocity to pid.n.feedback-deriv?
[17:56:54] * djdelorie wonders if a driver-only card would be useful, if that works...
[17:57:35] * djdelorie bounced IRC a few times to move it to my laptop. NEW VIDEO CARD TODAY!!! Hope it works as well as it's supposed to...
[18:03:21] <pcw_home> viesturs yes that will get you better velocity estimation then the PID comps default dP/dT
[18:03:51] <viesturs> Ok, thank You!
[18:03:54] <pcw_home> than the
[18:04:04] <viesturs> Will do it in a moment and then continue on PID tuning.
[18:04:35] <viesturs> Now the ferror is less than 2 mm with only P parameter specified
[18:05:05] <alex4nder> hey
[18:05:17] <djdelorie> ferror = following error?
[18:07:25] <Jymmm> yes
[18:07:48] <djdelorie> I had that in my controllers, had to add some feed-forward from the rpm feedback to the position offset
[18:10:19] <pcw_home> Since the 7I39 is a plain voltage mode drive you will definitely need some FF1 (for BEMF compensation)
[18:10:55] <pcw_home> and D
[18:11:15] <viesturs> do I understant that I should follow this sequence:
[18:11:18] <viesturs> first P
[18:11:20] <viesturs> then D
[18:11:22] <viesturs> then I
[18:11:24] <viesturs> then FF1
[18:11:41] <viesturs> add P until it starts oscillating
[18:11:46] <viesturs> then add D to calm it down
[18:11:57] <pcw_home> first P then D then P again then D again then FF1 then I
[18:12:13] * djdelorie thought it was "add P until it starts oscillating, then back it off a little until it doesn't"
[18:12:43] <viesturs> I thought " add P until oscillating, then add D to calm it down"
[18:12:48] <pcw_home> don't bother with I until its as close as you can get
[18:12:49] <pcw_home> it makes tuning FF1 hard/impossible
[18:13:16] <viesturs> ok, I will ask about the I, when I get there :)
[18:13:33] <pcw_home> yeah but you may be able to get more P as you increase D
[18:14:21] <djdelorie> D basically compensates for inertia?
[18:15:04] <pcw_home> D is first derivative so damping
[18:15:17] <pcw_home> first derivative of position
[18:15:45] -!- WorkingTaco [[email protected]] has joined #linuxcnc
[18:15:49] <djdelorie> I originally thought it was the "slow down as you approach the target" signal
[18:16:03] -!- WorkingTaco [[email protected]] has parted #linuxcnc
[18:16:33] <pcw_home> if you only have D you will have the equivalent of viscous damping
[18:16:38] -!- tom3p [[email protected]] has joined #linuxcnc
[18:17:16] <pcw_home> P by it self is like a spring (Hookes law)
[18:20:28] <alex4nder> anyone running an Atom D2800 yet and hae any latency numbers?
[18:20:28] <tom3p> interesting gimbal hinges might be useful for jig/fixtures http://whipmix.com/wp-content/uploads/via-product-catalog/product_docs/wide_vue_inst.pdf
[18:21:30] <alex4nder> tom3p: minorly creepy
[18:22:10] <pcw_home> so with only P position gain in a torque mode loop
[18:22:12] <pcw_home> the drive is the second derivative of the error
[18:22:14] <pcw_home> (the solution of this differential equation being a sine wave)
[18:22:15] <pcw_home> by adding D (first derivative) you add the exponential
[18:27:03] <Loetmichel> re @ home
[18:29:00] <pcw_home> bbl
[18:32:59] <Loetmichel> *sigh*... i am a litttle out of training... 4 hrs to mount a sattelite dish on the Company roof, throw 100m Coax cable around and install and config to digital sat recievers....
[18:33:23] <Loetmichel> there were times when i could do that in 1 hr straight.
[18:36:55] <viesturs> Peter, thanks for the explanation. Though I did not understand it completely... You said " will definitely need some FF1 (for BEMF compensation)". What is BEMF?
[18:38:57] <viesturs> another noob question - how should I move the motor after I adjust PID parameters? Just jogging forth/back/forth or is there something more effective?
[18:40:20] <jdhnc> Back ElectroMotive Force
[18:43:04] <viesturs> thanks! how do I see it in motion/ ferror graph? Those small spikes that currently are everywhere? I have got ferror below 0,3 mm with P parameter only. Does it mean I could set acceleration higher? Max velocity is limited by the bad balance in rotating nut assembly - it causes vibration...
[18:43:57] <viesturs> Most probably poor precision in manufacturing the parts. Should have checked it, before accepting. Yet another lesson learned with this macine...
[18:48:34] -!- tom3p has quit [Quit: Ex-Chat]
[18:51:05] <Jymmm> viesturs: what machine?
[18:58:56] <viesturs> Jymmm, how do You mean? This is my own build, I definitely will share some more details on mailing list, once I will get it working correctly.
[18:59:51] <Jymmm> viesturs: You said "you should have checked before accepting" as in look at it before singing for it from the shipper.
[19:01:11] -!- sumpfralle [[email protected]] has joined #linuxcnc
[19:03:31] <viesturs> No, that is a misunderstanding, I should have checked the parts for rotating ballscrew nut assembly, before accepting them from the guys that made them. Unfortunately I do not have a lathe (yet), so I have to outsource the process of actual metalworking and production of the parts. Currently I do designing part and also CNC control implementation. And after-sale user support as well.
[19:03:54] -!- acemi [acemi!~acemi@unaffiliated/acemi] has joined #linuxcnc
[19:04:02] <Jymmm> ah
[19:04:14] <PCW> FF1 is tuned so on a long slew, the lag or lead is nulled (too little FF1 = lag. too much = lead)
[19:10:37] -!- Rogge has quit [Quit: Never put off till tomorrow, what you can do the day after tomorrow]
[19:15:28] <ReadError> alex4nder
[19:15:31] <ReadError> you see the 2nd vid I did?
[19:15:58] <ReadError> joe9: delivered
[19:20:18] -!- m_kharboutly [m_kharboutly!6d0bbe60@gateway/web/freenode/ip.109.11.190.96] has joined #linuxcnc
[19:20:31] <skunkworks> viesturs: Make sure you are in exact stop mode (g61) and write a small gcode program that goes 1 inch back and forth. (or whatever is far enought that you get into cruise phase)
[19:21:10] <skunkworks> set halscope to trigger and keep playing with the pid and feed forward values
[19:21:41] -!- m_kharboutly has quit [Client Quit]
[19:24:41] <viesturs> yes, I am playing with HalScope! That is the only way I can imagine to actually trace the motion and all the parameters, like ferror etc.
[19:25:21] <viesturs> I am not homing the machine yet, running it in joint mode, so cannot currently run any test code. But I guess that with my current tuning I can let it home...
[19:28:03] -!- fatpandas has quit [Quit: leaving]
[19:28:17] <skunkworks> well - I have done it running a incremental jog..
[19:28:40] <viesturs> I am doing continuous jogging...
[19:28:56] <viesturs> Can anyone suggest, why are these lines so spikey?
[19:28:57] <viesturs> http://picpaste.com/pics/PID_spikes-97tp9Yva.1335382085.png
[19:30:07] -!- fatpandas [[email protected]] has joined #linuxcnc
[19:32:21] <alex4nder> ReadError: I missed it..
[19:34:06] * skunkworks isn't an expert at tuning...
[19:35:11] <archivist> looks a bit regular, is it at revolution rate
[19:35:28] <PCW> no nearly enough gain
[19:36:40] <PCW> and it does look periodic (once per turn?)
[19:36:47] <viesturs> which gain?
[19:37:21] <viesturs> yes i also noticed that it is periodical
[19:38:10] <PCW> Do you have any D? You will need more D to get more P
[19:38:14] <skunkworks> I did find out that vineger and salt works wonders for cleaning silver solder flux off of parts. Google for the win
[19:38:22] <viesturs> i was thinking that it might be the eccentricity in pulley on motor shaft
[19:38:44] <viesturs> I have
[19:38:48] <viesturs> P = 220
[19:38:52] <viesturs> I = 6
[19:38:57] <viesturs> D = 1,2
[19:39:12] <PCW> FF1?
[19:39:12] <viesturs> FF1 = 0,6
[19:39:29] <PCW> And you tuned FF1 before adding I?
[19:39:37] <viesturs> yes
[19:40:24] <PCW> And it oscillates with more P?
[19:40:25] -!- ve7it has quit [Remote host closed the connection]
[19:40:30] <ReadError> alex4nder: http://www.youtube.com/watch?v=9AoC4s9HPo4
[19:40:31] <viesturs> I had symmetrical ferror within 0,2 mm, added FF1 until symmetricity disappeared and it oscillated around zero line
[19:40:35] <ReadError> i got everything working now
[19:40:44] <ReadError> positive axis worked much better
[19:40:46] <viesturs> Adding I has not provided noticable improvement
[19:41:01] <viesturs> I will try adding more P...
[19:41:04] <viesturs> just a sec
[19:43:48] <PCW> When adding I add it until the system becomes unstable and then use about 1/2 that value
[19:44:53] <PCW> At least that's what works for me with voltage mode drives
[19:44:58] <viesturs> I changed P from 220 to 280 and also 340 - both were worse, some large ferror spikes up tp 0,36 mm appeared
[19:44:58] <viesturs> so I returned to 220
[19:47:15] -!- FinboySlick has quit [Remote host closed the connection]
[19:47:52] <skunkworks> I have done it where I add P until it become unstable. back it off a bit. then add d until it becomes unstable - back it off a bit. add more P ........repeate.....
[19:47:53] <viesturs> now reverted back to P=220 and there also are some large spikes
[19:48:44] <PCW> but it did not oscillate at the higher P values?
[19:49:02] <viesturs> should there be any spikes, if I start with all values at 0 and P also is very low, like 0,5?
[19:49:20] <PCW> I would adjust P up until it oscillates
[19:49:37] <viesturs> looking at graph it oscillates all the time. by the sound I also hear it is not perfectly smooth
[19:49:50] <viesturs> ok, I will try adding more P
[19:49:59] <PCW> sure its all spikes at that point sine the error is huge
[19:50:04] <viesturs> should I reset I and FF1 to 0?
[19:50:06] <PCW> since
[19:50:27] <PCW> no FF1 should be fine
[19:50:50] <PCW> i would not mess with I until its as good as it can be
[19:50:50] <viesturs> ok, I = 0 and more P...
[19:50:56] -!- vladimirek has quit [Ping timeout: 265 seconds]
[19:51:58] <PCW> does is oscillate when static? if not try skunkworks method
[19:52:50] <viesturs> no, it is calm, when static
[19:53:46] <PCW> The other way to tune P and D is by putting a step function in and adjust for maximum stable P gain and no (or small) overshoot on the step
[19:53:47] <viesturs> why would i add more P, when I am satisfied with absolute ferror values? I just want it to smooth out ferror curve and more importantly bldc.n.value curve
[19:54:12] <viesturs> I do not understand the last suggestion
[19:54:15] <PCW> more P gain should reduce your ferror
[19:54:38] <PCW> (if it no so much as to cause oscillation)
[19:54:45] <PCW> its not
[19:55:04] <djdelorie> does more P give you more speed/power too?
[19:55:49] <PCW> no just makes it stiffer (higher spring constant)
[19:56:02] <viesturs> I think that current ferror already is good - it is below 0,02 mm at 3000 mm/min
[19:56:26] <PCW> I make it stiffer as well but only for low frequencys
[19:57:10] <viesturs> the problem is that it is very spikey
[19:57:30] <PCW> what is spikey?
[19:58:13] -!- psha has quit [Quit: Lost terminal]
[19:58:14] <viesturs> I will post a screenshot of HalScope in a minute
[19:58:30] <PCW> the PID output will look like noise (bldc.n) but this is immaterial
[19:59:08] -!- syyl_ws has quit [Quit: Verlassend]
[20:00:21] <PCW> the ferror will get better with high P/D/I until you get to the point of instability
[20:00:53] -!- stevegt_1 has quit [Ping timeout: 248 seconds]
[20:01:11] <PCW> higher P/D/I i mean
[20:01:16] <djdelorie> then it explodes?
[20:01:31] <PCW> well it "rings"
[20:03:57] <PCW> if the gain is too low 'spikey' ferror may be just amplified mechanical drag
[20:04:40] -!- andypugh-iPhone [andypugh-iPhone!~andypughi@cpc2-basl1-0-0-cust639.basl.cable.virginmedia.com] has joined #linuxcnc
[20:05:28] <andypugh-iPhone> Hi Viesturs. Any luck?
[20:05:34] <viesturs> i think i can say that it rings for me, when in motion. it is totally quiet and silent, when not moving
[20:06:22] <viesturs> Hi, Andy! Actually yes, Peter's advice on setting pwmgen's deadtime to 0 changed the whole thing.
[20:06:44] <andypugh-iPhone> Glad to hear it.
[20:06:59] <viesturs> At least now I got the ferror within 0,03 s at 3000 mm/min.
[20:07:12] <viesturs> Will post a screenshot in a minute
[20:07:47] <andypugh-iPhone> I am migrating data between Macs, so using the IPhone. I guess I could use the CNC box.
[20:08:24] <viesturs> Here is the new screenshot:
[20:08:25] <viesturs> http://picpaste.com/ferror_spikes-jpeHU2qD.png
[20:08:29] -!- the-jub [[email protected]] has joined #linuxcnc
[20:08:54] <viesturs> Of course, if it has internet connection...
[20:08:56] <cradek> mm?
[20:09:57] <andypugh-iPhone> I started the transfer at 11am and it said 8 hours. I came back ar 8pm to it saying 40 hours. So I cancelled and connected with wire. Now it says 6 hours.
[20:10:08] <PCW> what happens if you double the P gain? (dont know if you've gone far enough until you've gone too far)
[20:11:06] <viesturs> the last screenshot is with P = 450, previously I had P = 220
[20:11:17] <viesturs> should I gradually increase even more?
[20:11:23] <PCW> well try 1000
[20:11:36] <viesturs> Ok
[20:11:58] <PCW> definitely looks like you have a periodic load variation
[20:12:17] <andypugh-iPhone> .03mm sound pretty good to me.
[20:12:53] <andypugh-iPhone> (at that speed)
[20:13:12] <viesturs> Had P=900, much louder vibration, almost all spikes in HalScope approach 0,1 mm, I had less than 0,03 before.
[20:13:14] <cradek> I agree the following is good but I also agree there's a mechanical issue
[20:13:41] <viesturs> Andy, yes, I also think that it is good. I just would like to smooth it out, it is very spikey
[20:14:04] <cradek> can you tell for sure if it's once per turn?
[20:14:13] -!- fliebel has quit [Remote host closed the connection]
[20:14:17] <andypugh-iPhone> Try a teeny amount of D? (possibly negative)
[20:14:28] <viesturs> It could be eccentricity in pulley on the motor shaft.
[20:15:27] <cradek> add the position to your plot so you can tell what position increment it is
[20:15:40] <viesturs> With 3000 mm/min = 50 mm/sec and with 2,5 mm travel per motor turn I have 20 revolutions per second, 0,05 s per each revolution.
[20:15:51] <cradek> see if the bumps stay the same distance apart at varying speeds
[20:16:10] <viesturs> is there difference, if I put pos-cmd or pos-fb?
[20:16:26] <cradek> you're only getting 5 bumps per second, nowhere near 20
[20:17:18] RagingComputer is now known as ragingcomputer
[20:17:41] ragingcomputer is now known as RagingComputer
[20:18:24] -!- FinboySlick [[email protected]] has joined #linuxcnc
[20:19:03] -!- sumpfralle1 [[email protected]] has joined #linuxcnc
[20:19:04] -!- sumpfralle has quit [Read error: Connection reset by peer]
[20:20:41] <viesturs> I changed bldc.0.value to axis.0.joint-pos-cmd
[20:21:03] <viesturs> should I post the screenshot?
[20:23:32] <viesturs> that will be difficult - LinuxCNC machine refuses to accept both my usb sticks, as it does not have internet access
[20:24:13] <andypugh-iPhone> How annoying.
[20:25:57] <PCW> what is the pully ratio? maybe the periodic mechanical stuff is on the leadscrew
[20:26:12] <viesturs> turns out there is I/O error on cf card
[20:26:21] <viesturs> it is 4:1
[20:27:05] <PCW> is it lumpy if turned by hand?
[20:27:17] <viesturs> the ballscrew is 10 mm pitch
[20:27:28] <viesturs> what is "lumpy"?
[20:27:37] <PCW> technical term
[20:27:58] <viesturs> I already thought so... :)
[20:28:03] <PCW> non constant torque
[20:28:18] <viesturs> I will try
[20:28:41] <PCW> related to "crunchy"
[20:29:24] <viesturs> yes, it is lumpy
[20:29:36] -!- motioncontrol has quit [Quit: Sto andando via]
[20:30:22] <andypugh-iPhone> Is there a gearbox.
[20:30:58] <viesturs> no, it is belt&pulleys only
[20:33:17] <viesturs> damn, looks like another cf card died
[20:33:30] <viesturs> now the LinuxCNC box won't start up
[20:33:47] <viesturs> now that is not good
[20:34:14] <andypugh-iPhone> Is the USB stick still in? It might be trying to boot from that.
[20:34:34] <viesturs> no, removed them
[20:35:34] <viesturs> there was black screen with some text, saying something like "starting additional modules"
[20:37:34] <viesturs> are cf cards sensitive to magnetic field?
[20:37:55] <viesturs> there is a transformer some 10 cm from it
[20:38:21] <andypugh-iPhone> Not as far as I know
[20:38:49] <viesturs> how do I convince the client to purchase new hdd?
[20:39:26] <andypugh-iPhone> Boot from CD. Run disk check utility.
[20:40:10] <andypugh-iPhone> I would re-seat the memory and cables first.
[20:40:23] <viesturs> there is no cd drive available and there is no chance for me to download the ubuntu livecd to put on usb drive
[20:40:37] <viesturs> I guess I will go to dormitory at first :)
[20:40:43] <viesturs> it is 23:40 for me
[20:41:06] <andypugh-iPhone> Working away from base is always tricky.
[20:41:29] <cradek> ubuntu live image on appropriate media is a basic troubleshooting/repair tool that you should have with you on service calls
[20:42:20] <cradek> it's so good at many things
[20:42:26] <andypugh-iPhone> Ideally a spare HDD too.
[20:43:23] <viesturs> Chris, this is yet another lesson learned with this machine...
[20:43:33] <cradek> :-)
[20:43:49] <viesturs> spare HDD is too expensive
[20:43:58] <viesturs> I could have spare cf card, but I bought this one last week, so it seems that they do not last long
[20:44:36] <viesturs> spare cf card would save me for another few days...
[20:44:42] <cradek> solid state media is shockingly unreliable
[20:45:48] <viesturs> I have had only good experience with the 3 SSDs I have used - one in the welding robot, second in a woodworking machine and the third in my laptop I am typing now
[20:46:12] <cradek> I have had only bad experiences with SSDs
[20:46:19] -!- acemi has quit [Quit: WeeChat 0.3.2]
[20:46:30] <cradek> I have one in my primary machine at home and it routinely logs weird errors
[20:46:41] <cradek> if I didn't have good backups I'd be worried about it.
[20:46:42] <rob_h> same here, very good results with SSDs here, 5years with one machine now still good, i think depends on makes also, OCZ here.. CF tho very bad times not worth it
[20:47:08] <cradek> mine's a kingston, which is as name-brand as I could find
[20:47:50] <rob_h> i did have to firmware update the first one i had so support trim on linux. to keep it good.. but others are newer now
[20:47:53] <viesturs> my 3 drives also are Kingstons, I do trust them and they have kept up with my expectations
[20:48:32] <viesturs> ok, I guess my day is over, now the hdd problem has to be solved tomorrow
[20:48:51] <andypugh-iPhone> I have been using a SATA DOM for about 2years doing lots of compiles, and it has been good.
[20:48:59] <viesturs> thank You, Chris and Peter, for You patience and help, I made a real progress today!
[20:49:09] <viesturs> see You all tomorrow or some other time!
[20:49:14] -!- viesturs [[email protected]] has parted #linuxcnc
[21:00:38] -!- the-jub has quit [Quit: ill be back soon! :-/ ....maybe..]
[21:02:06] <DJ9DJ> gn8
[21:02:30] -!- DJ9DJ has quit [Quit: bye]
[21:08:31] -!- andypugh-iPhone has quit [Quit: Colloquy for iPhone - http://colloquy.mobi]
[21:13:18] -!- stevegt_1 [[email protected]] has joined #linuxcnc
[21:21:08] -!- dimas_ has quit [Ping timeout: 240 seconds]
[21:22:58] -!- andypugh-iPhone [andypugh-iPhone!~andypughi@cpc2-basl1-0-0-cust639.basl.cable.virginmedia.com] has joined #linuxcnc
[21:24:04] -!- FinboySlick has quit [Quit: Leaving.]
[21:26:48] e-ndy|afk is now known as e-ndy
[21:31:57] H264 is now known as tiwake
[21:32:05] tiwake is now known as WalterN
[21:33:53] -!- dimas_ [[email protected]] has joined #linuxcnc
[21:34:45] e-ndy is now known as e-ndy|afk
[21:39:43] -!- Anon0063 [[email protected]] has joined #linuxcnc
[21:40:13] -!- Anon0063 has quit [Client Quit]
[21:40:27] -!- Anon0623 [[email protected]] has joined #linuxcnc
[21:41:03] <Anon0623> hello everyone
[21:41:38] <Anon0623> would need some help with linux to compile .c kinematics file
[21:42:32] <Anon0623> sudo comp -- install mykinematics.c returns me sudo:comp: command not found
[21:42:56] <cpresser> Anon0623: install emc2-dev
[21:44:22] <Anon0623> should run this in terminal?
[21:44:54] <Anon0623> install: missing destination file operand after `emc2-dev'
[21:45:54] -!- andypugh-iPhone has quit [Ping timeout: 272 seconds]
[21:46:34] <cpresser> no, you should not run this in a terminal :)
[21:46:53] <cpresser> use your prefered packet-manager-tool to install the package emc2-dev
[21:48:56] -!- mhaberler has quit [Ping timeout: 244 seconds]
[21:50:21] -!- andypugh [andypugh!~andypugh@cpc2-basl1-0-0-cust639.basl.cable.virginmedia.com] has joined #linuxcnc
[21:50:28] -!- andypugh has quit [Remote host closed the connection]
[21:50:46] -!- micges has quit [Quit: Leaving]
[21:51:31] -!- andypugh [andypugh!~andypugh@cpc2-basl1-0-0-cust639.basl.cable.virginmedia.com] has joined #linuxcnc
[21:52:15] <andypugh> Hmm, not as easy to join an IRC as one might like when there is no <hash> key on the keyboard.
[21:52:38] -!- mhaberler [[email protected]] has joined #linuxcnc
[21:53:29] <WalterN> ha
[21:53:32] <WalterN> here
[21:53:34] <WalterN> #
[21:53:38] <WalterN> copy-paste
[21:54:19] <andypugh> Thanks. That would actually have been genuinely useful except I googled and found it is right-alt + £
[21:54:28] <andypugh> #### :-)
[21:54:33] * cpresser actually has a text-file with various special chars in it on my mobilephone :)
[21:56:26] <andypugh> Yay! got the filebutton working.
[21:56:40] <andypugh> I am not cinvinced that I am going to like Python
[21:58:12] <Anon0623> hmm I have to admit I'm "noob" here to work with linux ... sry
[21:58:38] <Anon0623> so firs i need packet-manager-tool ?
[21:58:41] <Anon0623> first
[21:59:06] <andypugh> Anon0623: No, there should be one built in. Actually a few
[21:59:17] <Anon0623> synaptic
[21:59:18] <Anon0623> ?
[21:59:22] <andypugh> Yes
[21:59:24] <Anon0623> pl
[21:59:26] <Anon0623> ok
[21:59:40] <andypugh> That's under System->Administration
[22:00:10] <Anon0623> found it ...
[22:00:21] <andypugh> But for LinuxCNC-ish stuff the command-line is generally used. (sudo apt-get install <package>
[22:00:52] <andypugh> And for super-convenience, Aplications->Ubuntu Software Centre
[22:02:09] <Anon0623> ok I have both opened :)
[22:02:26] <andypugh> What are you trying to do?
[22:03:06] <Anon0623> comp and install kinematics
[22:03:58] <Anon0623> sudo comp --install mykinematics.c returns me sudo: comp: command not found
[22:04:52] <andypugh> Ah, sudo apt-get install linuxcnc-dev
[22:04:58] <Anon0623> its fresh install from LiveCD and I need to install my custom kinematics
[22:05:14] <JT-Shop> can comp compile a c file?
[22:05:21] <andypugh> possibly also sudo apt-get install buildessential
[22:05:46] <andypugh> JT-Shop: Yes, comp is quite a simple way to compile and install a kernel module written in C
[22:06:00] <JT-Shop> ok, thanks Andy
[22:06:48] <JT-Shop> hmmm http://linuxcnc.org/docview/html/hal/comp.html#_installing
[22:07:36] <Anon0623> so what I do with this?
[22:07:58] -!- dimas_ has quit [Ping timeout: 250 seconds]
[22:09:50] <JT-Shop> oh bother the sentence is run on and incomplete at the same time
[22:10:31] <andypugh> after you have installed buildessential and linuxcnc-dev then comp should work
[22:11:15] <JT-Shop> what does buildessential do?
[22:11:23] -!- FinboySlick [[email protected]] has joined #linuxcnc
[22:11:25] <Anon0623> runing this in terminal sudo apt-get linuxcnc-dev gives me E: Invalid operation linuxcnc-dev
[22:11:42] <andypugh> I think it makes sure you have all the compilers etc
[22:11:57] <JT-Shop> apt-get install linuxcnc-dev
[22:12:06] <JT-Shop> ah ok
[22:13:01] <JT-Shop> dang three errors in one sentence :/
[22:13:08] <Anon0623> E: Could not open lock file /var/lib/dpkg/lock - open (13: Permission denied)
[22:13:18] <Anon0623> E: Unable to lock the administration directory (/var/lib/dpkg/), are you root?
[22:13:23] -!- sumpfralle1 has quit [Ping timeout: 245 seconds]
[22:13:31] <Anon0623> I have admin rights
[22:13:55] <Anon0623> when I run only sudo, it asks me for password ...
[22:14:01] <JT-Shop> sudo apt-get install linuxcnc-dev
[22:14:28] <JT-Shop> I should not type right after getting up from a nap
[22:14:59] <Anon0623> E: Could not get lock /var/lib/dpkg/lock - open (11: Resource temporarily unavailable)
[22:15:09] <Anon0623> E: Unable to lock the administration directory (/var/lib/dpkg/), is another process using it?
[22:15:33] <JT-Shop> do you have the synaptic package manager open or another terminal open that is using apt-get
[22:16:09] <Anon0623> no, only terminal
[22:16:41] <Anon0623> aaaa ... happend something ....
[22:17:00] <Anon0623> I closed and started again terminal ... now it works ...
[22:17:20] <JT-Shop> ok... whew those errors are above my pay scale
[22:17:42] * JT-Shop wanders down to fix my mistakes in the manual before I forget
[22:17:56] <Anon0623> so now after loading this ... should sudo works also for kinematics?
[22:18:28] <Anon0623> or is there something more to do?
[22:20:13] -!- dimas_ [[email protected]] has joined #linuxcnc
[22:21:39] <Anon0623> it works! THANK YOU ALL !
[22:22:34] <andypugh> That seems like it was unnaturally easy
[22:23:40] <joe9> ReadError: how do you want to do the handover? where do you want me to come?
[22:24:07] <joe9> i am headed out tomorrow afternoon. so, would be good if I can pick it up before that.
[22:24:43] <joe9> ReadError: any thoughts, please?
[22:24:44] -!- erasmo has quit [Ping timeout: 245 seconds]
[22:24:45] <jthornton> now that the docs have the correct command and the sentence might make sense it might be easier
[22:25:23] <cradek> jthornton: "you're"
[22:25:40] -!- Anon0623 has quit [Quit: Bye Bye]
[22:25:55] <jthornton> crap, I shouldn't work right after my nap
[22:26:00] <cradek> sorry :-)
[22:26:29] <jthornton> np lol
[22:26:42] <jthornton> I'm not really grumpy after my nap
[22:28:57] * jthornton gets back to cannon construction
[22:29:16] <Tom_itx> yay
[22:35:13] <andypugh> How do I create a new instance of a type in Python?
[22:35:36] <andypugh> Do I need to find an object with a suitable creator, or can I just "dim" one?
[22:35:57] <andypugh> Specifically I want to create a gtk.FileFilter object.
[22:38:58] -!- Valen [[email protected]] has joined #linuxcnc
[22:39:02] -!- erasmo [[email protected]] has joined #linuxcnc
[22:39:38] -!- andypugh has quit [Remote host closed the connection]
[22:40:47] -!- erasmo has quit [Remote host closed the connection]
[22:41:21] -!- andypugh [andypugh!~andypugh@cpc2-basl1-0-0-cust639.basl.cable.virginmedia.com] has joined #linuxcnc
[22:43:37] <FinboySlick> andypugh: You want to create a new type that inherits gtk.FileFilter methods and properties?
[22:44:01] <andypugh> No
[22:44:10] <andypugh> I just want one.
[22:45:13] <FinboySlick> blah = gtk.FileFilter() if I remember it correctly.
[22:45:34] <andypugh> Yes, I found that just now in online docs
[22:45:38] <FinboySlick> Been a while since I played with it.
[22:45:59] <andypugh> But I get: NameError: global name 'gtk' is not defined
[22:46:13] <FinboySlick> you need to import first, most likely.
[22:46:24] <andypugh> I was wondering about that
[22:46:45] <FinboySlick> If you need just FileFilter() you can import just that from gtk.
[22:46:46] <andypugh> I think I prefer C
[22:46:50] <FinboySlick> Or you could import all of gtk.
[22:47:23] <andypugh> You know where you are with C in the Kernel. If you don't write it from scratch, you can't have it :-)
[22:47:40] <FinboySlick> C lets you get into more nitty gritty, but it doesn't take very long for python to let you do pretty funky things in just a few coherent lines (vs perl where the lines are incoherent ;)
[22:48:00] -!- factor has quit [Read error: Connection reset by peer]
[22:48:32] <FinboySlick> But yah, python keeps you hovering a few inches above the metal.
[22:49:27] <FinboySlick> BTW, if you import FileFilter() from gtk, you'll declare your class blah = FileFilter()
[22:49:27] <andypugh> It is getting easier now I have found "print dir(widget)
[22:49:41] <FinboySlick> If you import gtk, then blah = gtk.FileFilter()
[22:50:10] <FinboySlick> Messing with lists and sets and what not in python is a heck of a lot of fun.
[22:52:59] <andypugh> I liked Matlab, when I used it a lot
[22:53:50] <FinboySlick> Yeah. You'll probably have a great grasp of python within a day or two if you hammer at it. That's its strongest point if you ask me, it's very easy.
[22:53:51] <kanzure> FinboySlick: perl can be written coherently with patience (probably about the same as ruby)
[22:54:09] <jdhnc> kanzure: if that is true, why has it never happened?
[22:54:21] <FinboySlick> kanzure: I was mostly being humorous.
[22:54:40] <kanzure> FinboySlick: me too.. i meant no disrespect
[22:55:17] <FinboySlick> kanzure: Heh, non perceived... I can't really code that well.
[22:55:22] <kanzure> jdhnc: to be fair, i suppose all the perl i used to write looked like crap, then i started writing python
[22:55:24] <FinboySlick> :s/non/none/
[22:55:36] <kanzure> anyway, i didn't mean to start a language war
[22:55:47] <jdhnc> it's not a war, just an observation.
[22:56:03] <FinboySlick> Let's duke it out in the octogon!
[22:56:24] <kanzure> on a related note, are there ways to get >20 khz stepping with linuxcnc?
[22:56:31] <jdhnc> I think the more you know perl, the less legible it is.
[22:56:32] <Valen> mesa
[22:56:43] <andypugh> illegible: http://en.wikipedia.org/wiki/Brainfuck
[22:56:45] <kanzure> i played with mesa cards a looong time ago
[22:57:03] <kanzure> you mean the fpga controllers?
[22:57:30] <andypugh> But you have to like a language with a 200-byte compiler
[22:57:51] <FinboySlick> I haven't messed with my config a lot, but I think I got 60khz on my little atom board. But i never tested it.
[22:58:17] <kanzure> i thought there were limits due to rt-linux's minimum interrupt response timing?
[22:58:32] <FinboySlick> If you can manage a <9000 ns jitter.
[22:58:59] <FinboySlick> kanzure: Could well be, I just meant that it's what the dialog displays ;)
[23:00:27] -!- rob_h has quit [Ping timeout: 260 seconds]
[23:03:19] <Valen> mesa have a board for ~$100 or less now
[23:03:21] <Valen> just use it ;-P
[23:05:20] <kanzure> if it's an fpga, then what's the state of current open source fpga programmers?
[23:05:24] <kanzure> iirc they basically don't exist
[23:08:23] <cpresser> kanzure: http://opencores.org/
[23:10:39] -!- skunkworks__ [skunkworks__!~chatzilla@str-bb-cable-south-3-102.dsl.airstreamcomm.net] has joined #linuxcnc
[23:11:07] <andypugh> kanzure: No, there are no Free ones, but some free ones.
[23:11:43] <andypugh> Or, I might be wrong (looking at opencores)
[23:12:09] <andypugh> I think you need the zero-cost Xilinx compiler to create Mesa firmwares
[23:13:02] <Tom_itx> i wrote a how to for the bitfiles
[23:13:10] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/emc/xilinx/xilinx_install_index.php
[23:14:24] <andypugh> FinboySlick: Any idea how to access a different glade object from a callback? My callback is passed self and obj, but I want to set the value of a different object.
[23:14:55] <andypugh> Tom_itx: And it works, I have created working bitfiles, and a side-project is a web interface to make them on-the-fly
[23:15:28] <FinboySlick> andypugh: You'll probably need to pass that object as a param... Or add a reference to your glade object in obj. That essentially gives you a pointer to your glade object.
[23:15:34] <andypugh> (except I seem to have been distracted by Python)
[23:16:22] <kanzure> Tom_itx: looking..
[23:16:33] <andypugh> Or (to be unpythionic) set up a global?
[23:16:48] <kanzure> ah this is the xilinix software
[23:17:26] <FinboySlick> andypugh: I guess so. But that's a bit on the ugly side. Unless the glade object is central to your app. And even then, it's bad style.
[23:17:40] <FinboySlick> Heh, not that I could give style lessons when it comes to code.
[23:19:39] <andypugh> Any clues how to add a reference?
[23:19:58] <FinboySlick> obj.myglade = gladeobject
[23:20:07] <andypugh> I don't seem to get any choice about the parameters that are passed
[23:20:51] <FinboySlick> andypugh: Is this something you can pastebin?
[23:21:27] <andypugh> FinboySlick: So, in the __init__ I need to use builder.getobject to find both the button and the label, then add a reference to one to the other, so that the button callback can set the label?
[23:21:41] <andypugh> Seems, err, clunky?
[23:22:18] <FinboySlick> andypugh: You don't necessarily have to do it in the __init__... I'm just entirely sure what you're trying to do.
[23:23:59] <andypugh> Initially I want to load a bitmap, and set some labels in a GladeVCP dialog to show the height and width.
[23:24:42] <andypugh> So, a button to load the image (now working) and then I want to button callback to set the labels to show the size
[23:25:24] <andypugh> http://pastebin.com/ntMMEHKX
[23:28:24] <FinboySlick> Hmmm... And where are your labels defined?
[23:28:31] <FinboySlick> They're children of obj?
[23:28:56] <skunkworks__> the better question... andypugh : what are you writing?
[23:29:45] <andypugh> No, AFAIK obj is just the fileselectbutton widget, and the labels are on the same level.
[23:30:25] <andypugh> skunkworks__: It's an experiment prompted by a query on the forum, to see if I can read an image file into a HAL module
[23:31:51] <skunkworks__> interesting... Did you see this? http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Rastering_With_A_Laser
[23:32:12] <FinboySlick> OK. I sort of lack gtk/glade coding experience... You probably have a list of children that you can access like an array.
[23:32:12] <andypugh> Yes, and it seems horribly clunky and inelegant :-)
[23:32:43] <FinboySlick> So you could do self.fb["thelabel"] or something like that.
[23:32:46] <andypugh> hmm, obj.parent.child2?
[23:34:06] <andypugh> skunkworks__: My idea was a HAL module that creates a raster completely independent of any G-code, and modulates an output accordingly.
[23:34:11] <skunkworks__> andypugh: I agree
[23:34:20] <skunkworks__> andypugh: cool!
[23:34:41] <andypugh> No need to load EMC2 even, just a minimalist Glade UI and HAL
[23:35:17] <FinboySlick> andypugh: Was it any help?
[23:35:58] <andypugh> FinboySlick: The label is not a child of the button
[23:36:51] <FinboySlick> andypugh: I figured. Though you can probably do self.fb.parent.elemets["whateveristhelabel"].caption ="sometext"
[23:36:59] <skunkworks__> andypugh: gasp - no axis? I don't know about that... ;)
[23:39:21] <FinboySlick> Or replace self.fb by its parent and move up to the hierarchy when you add_filter.
[23:40:17] <FinboySlick> Things in python are usually organized in lists or catalogs or sets.
[23:41:36] <FinboySlick> Actually, that would be move down the hierarchy when calling add_filter
[23:41:43] <FinboySlick> But I'm sure you caught my drift.
[23:43:54] -!- factor [factor!~factor@r74-195-216-151.msk1cmtc02.mskgok.ok.dh.suddenlink.net] has joined #linuxcnc
[23:44:35] <FinboySlick> builder is a glade.Glade() object?
[23:44:45] <andypugh> I don't know.
[23:45:16] <andypugh> The thing is, I want to get to the labels from the LoadImage block, not the __Init__
[23:46:10] <FinboySlick> From a few google searches, it seems that glade coders make very liberal use of get_object()
[23:46:53] <andypugh> Which is fine, but can it look outside the passed object?
[23:46:54] <FinboySlick> So maybe there's no way to refer to child gui elements in lists.
[23:47:24] <FinboySlick> Oh, your label is not a child of builder ?
[23:47:55] <andypugh> FinboySlick: Yes, it is a child of builder (I think) but builder is not passed to LoadImage
[23:48:40] <FinboySlick> Oh, you can set a self.builder = builder
[23:49:01] <FinboySlick> Then you can use self.builder.get_object('whatever') in getimage
[23:49:08] <andypugh> Ah, here we come to a funamental question. Who or What is Self?
[23:49:54] <andypugh> Is it the class, or the function?
[23:50:01] <FinboySlick> It's the class.
[23:50:02] <andypugh> Or the whole UI?
[23:50:12] <FinboySlick> All methods on a class are passed self
[23:50:19] <FinboySlick> Which is the class itself.
[23:50:38] <andypugh> And you use self.variable to keep it local?
[23:50:39] <FinboySlick> That's why all your def whatever() statements have self as the first arg.
[23:51:00] <FinboySlick> Yes, that's how you set class properties from within the class.
[23:51:03] <andypugh> Or are variables local by default?
[23:51:24] <FinboySlick> If you just set a variable, it'll be local to the function.
[23:51:32] <FinboySlick> If you want it to be defined for the whole class, you use self
[23:51:42] <andypugh> (ie, I can remove all these self. things for local transitory variables?
[23:51:53] <andypugh> Right, good.
[23:52:17] -!- Valen has quit [Quit: Leaving.]
[23:53:11] -!- djdelorie has quit [Quit: Leaving]
[23:53:38] -!- djdelorie [[email protected]] has joined #linuxcnc
[23:55:52] <FinboySlick> Lemme know if it worked. All i say comes with a I haven't coded in a long time disclaimer.
[23:56:07] <andypugh> It seems close
[23:58:44] <andypugh> v
[23:58:46] <andypugh> http://pastebin.com/cnsWBgYH
[23:59:25] -!- H264 [[email protected]] has joined #linuxcnc
[23:59:42] <FinboySlick> v for victory or v for vendetta?
[23:59:51] -!- WalterN has quit [Ping timeout: 265 seconds]