#linuxcnc | Logs for 2013-02-09

[02:07:51] <DJ9DJ> moin
[03:15:00] <Loetmichel> mornin'
[05:52:55] <L84Supper> http://postimage.org/image/y2l75mw49/ http://s9.postimage.org/nu5ti9onj/A10.jpg another a10 board done as CoM module + IO board
[05:54:05] <L84Supper> the a10 is ~$6 and is a single cortex a9 1GHz with Mali GPU and tons of IO
[05:55:27] <L84Supper> still have to get around to benchmarking the IRQ latency jitter
[06:12:12] <mrsun> http://sphotos-e.ak.fbcdn.net/hphotos-ak-ash3/537978_10151410647318648_1783182546_n.jpg hmm stuff become larger when you lay it out compared to when its just standing in pieces against a wall :P
[07:19:30] <jthornton> what are you building?
[07:35:57] <JesusAlos> can do g1 with zero feed rate
[07:35:59] <JesusAlos> hi
[07:36:12] <JesusAlos> is the problem when open gcode file
[07:37:29] <archivist> can you rephrase to make a better question
[07:38:45] <JesusAlos> sorry my english
[07:39:41] <JesusAlos> I attempt a gcode file with linuxcnc, but appears a problem when try to run
[07:39:54] <JesusAlos> "can do g1 with zero feed rate "
[07:40:29] <jthornton> pretty much says it all
[07:40:46] * archivist imagines cannot
[07:40:52] <archivist> not can
[07:40:52] <JesusAlos> http://imagebin.org/246052
[07:40:54] <jthornton> http://linuxcnc.org/docs/html/gcode/gcode.html#sec:G1-Linear-Motion
[07:46:10] <r00t4rd3d> see his version
[07:49:33] <jthornton> and?
[07:52:33] <r00t4rd3d> have you compiled linuxcnc from source?
[07:54:06] <jthornton> me?
[07:54:17] <r00t4rd3d> sure
[07:54:45] <jthornton> yes all the time
[07:55:14] <r00t4rd3d> whats the flag to set for simulator?
[07:55:52] <jthornton> to make a pure simulator with no real time?
[07:56:26] <jthornton> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Installing_LinuxCNC#Building_LinuxCNC_simulator
[07:59:15] <r00t4rd3d> would building the simulator vs the real deal effect feed rates?
[07:59:55] <jthornton> no, if I understand your question
[08:01:09] <r00t4rd3d> would building the simulator vs the real deal effect feed rate errors**
[08:01:50] <jthornton> following errors?
[08:02:03] <jthornton> I don't know what you mean by feed rate errors
[08:03:00] <r00t4rd3d> i just find it weird is he using a pre alpha build, who knows what os and kernel he is using, does he have rtai ? Does he have an actual machine hooked up or is he just simulating ?
[08:03:40] <jthornton> that has no effect on trying to do a G1 with a zero feed rate
[08:09:05] <r00t4rd3d> oh i just figured if he was on a newer version of linux without a rtai kernel and using pre alpha without a simulator build that would lead to issues.
[08:10:31] <jthornton> doing a G1 with F0 is an error plain and simple
[08:25:54] <jthornton> I hate chasing an error down the wrong street
[08:33:34] <archivist> I wonder about the thought process that gets people stuck with an obvious error like that
[08:37:48] <archivist> Im having to type mach .... http://www.collection.archivist.info/searchv13.php?searchstr=mach
[08:44:09] <Tom_itx> even though stupid it should be a valid code
[08:45:20] <Tom_itx> not saying anyone should ever use it, but it shouldn't error
[08:46:56] <archivist> his code lacked setting F before using G1, it should error if no(0) speed set imo, impossible is....impossible and an error
[08:48:37] <archivist> it is also better that you get an error than silence and a stationary machine from that line
[08:48:37] <Tom_itx> k, if it were set after then i agree
[08:48:59] <Tom_itx> it would keep you wondering for a while
[09:01:36] <jthornton> all feed moves error with no feed rate
[09:02:37] <awallin> did anyone make progress with the olinuxino board? this one: https://www.olimex.com/Products/OLinuXino/
[09:05:26] <archivist> it only has 64MB of RAM memory
[09:07:57] <awallin> the A13 model has 512Mb... but there are so many ARM boards out there that it might be difficult to get xenomai and linuxcnc to run on a particular one...
[09:09:30] <rizo> i would like to write a hal component. Can i use the 16bit signed integer, or somehow truncate the 32 bit?
[09:12:32] * JT-Shop is happy to find the fire still burining in the shop stove
[09:17:50] <r00t4rd3d> http://forums.bit-tech.net/showthread.php?t=220958
[09:20:20] <r00t4rd3d> im gonna have to make one of those
[09:26:27] <r00t4rd3d> i guess that guys most used tool was a razor knife :/
[09:30:56] <archivist> r00t4rd3d, for thin wood see http://www.youtube.com/watch?v=v3Ad6tBdLbM
[09:31:30] <Tom_itx> hobby plywood
[09:32:41] <Tom_itx> archivist, :D
[09:43:00] <ReadError> https://dl.dropbox.com/u/25091878/Photo%20Feb%2009%2C%2012%2023%2043%20AM.jpg
[09:43:06] <ReadError> my big ass wood table
[09:46:07] <archivist> I must say dropbox is slow....had to get 1.3 gb off there yesterday
[09:59:47] <JT-Shop> I see the treadmill is in the std position that most are in
[10:05:25] <ReadError> lol
[10:08:14] <tjtr33> archivist, amazing 9u thick, and the edge was set by hand, knocking it into the shoe by the style of plane
[10:56:17] <ReadError> hey guys
[10:56:25] <ReadError> i got my tslot router bits
[10:56:35] <ReadError> if i need to make a channel thats straight
[10:56:42] <ReadError> whats the easiest way to do this?
[11:01:10] <r00t4rd3d> clamp a guide to the board
[11:01:15] <Tom_itx> yeah
[11:01:20] <r00t4rd3d> the your router will rest against
[11:01:37] <ReadError> on both sides?
[11:01:46] <Tom_itx> no
[11:02:54] <r00t4rd3d> http://www.theamericannight.com/voxproject/shelfcut1.jpg
[11:03:27] <r00t4rd3d> like that, then the side of your router will run along the lenght of the clamped on board
[11:03:48] <Tom_itx> cut a center groove first then the bottom T
[11:03:57] <Tom_itx> to relieve some material
[11:04:31] <r00t4rd3d> he got a tslot bit
[11:04:45] <Tom_itx> i noticed he said that
[11:04:53] <Tom_itx> cut a center groove first then the bottom T
[11:04:57] <r00t4rd3d> it cuts it all in one
[11:05:00] <Tom_itx> to relieve some material
[11:06:34] <r00t4rd3d> they are mad sharp, i dont think he will have a problem with mdf
[11:08:47] <tjtr33> if the board is meant to be a table , it might be attached to guides underneath.
[11:08:48] <tjtr33> if so, use them as the 'straight' reference for the cut, because the tslot would be straight AND alligned.
[11:15:10] <tjtr33> mdf tslot? what happens when the tnut is tightened? it seems it'd shear thu mdf.
[11:15:45] <ReadError> i was thinking about using ply
[11:16:26] <ReadError> r00t4rd3d: what do you use for tnuts ?
[11:16:33] <ReadError> those square nuts they sell?
[11:17:46] <tjtr33> build up a surface from strips of alum on top of ply/mdf. 1 layer makes the wide slot for nut, 2nd layer make narrow slots for bolts. then you're tightening against metal.
[11:18:34] <tjtr33> or buy an extrusion tslotted table (Isel etc)
[11:19:01] <ReadError> well i figure if i use wood and tslot it, and its the spoil board
[11:19:08] <ReadError> i can toss it out when it gets ratty
[11:19:54] <tjtr33> why tslots, why not embed threaded inserts as you need, toss the board, save the inserts for another layout.
[11:19:57] <Tom_itx> i use a grid of tapped holes instead
[11:20:51] <tjtr33> cheapo 'tnuts' (the wood worker kind a metal thread insert for wood) from the backside.
[11:21:11] <ReadError> its a bit over 48x48"
[11:21:21] <ReadError> putting all those inserts in would be a nightmare
[11:21:27] <ReadError> are they pretty cheap?
[11:21:47] <tjtr33> http://www.efc-intl.com/encyclopedia/nuts-self-retaining-t-type-capped-round-base-w3-wood-nibs.htm buy 'em by the bag at lowes/meodepot/menards
[11:21:48] <Tom_itx> yes
[11:21:51] <Tom_itx> http://tom-itx.dyndns.org:81/~webpage/boards/USBTiny_Mkii/Boxes/milling1.jpg
[11:22:22] <L84Supper> http://www.wired.com/dangerroom/2013/02/printed-magazine why all the hubub about 3d printing gun parts when machine tools have been around for decades?
[11:23:12] <Tom_itx> http://www.homedepot.com/buy/crown-bolt-10-32-x-5-16-in-zinc-plated-tee-nut-4-pieces--18881.html#.URaC2vJP_rQ
[11:23:20] <tjtr33> because 3d printer is cut & paste and a mill requires and apprenticeship of a few years
[11:25:43] <tjtr33> i used some wood tnuts yesterday because i needed a 5mm coupler nut, and couldnt find them. the tnut thread len was > 2 dia so i used that
[11:25:56] <L84Supper> you have to be able to import an STL and slice it
[11:26:41] <L84Supper> I suppose with a few weeks of training most people could do it :)
[11:26:53] <tjtr33> L84Supper, i just meant to say, 3d print was fast and no discipline, versus learnign a trade before getting a gun
[11:27:01] <L84Supper> I understand
[11:27:11] <tjtr33> L84Supper, are you planning on testing linux 3x on that A10?
[11:27:36] <L84Supper> tjtr33: I still have to get a cubieboard
[11:27:42] <L84Supper> been busy
[11:27:56] <tjtr33> what were the pix about?
[11:28:04] <tjtr33> oh nevermind :)
[11:28:14] <L84Supper> a new a10 board
[11:28:35] <L84Supper> there are a couple now with lots of GPIO
[11:28:53] <tjtr33> A10+ and the A30 i think
[11:29:10] <L84Supper> I have had an a10 net book for >6 months but it has no GPIO to play with
[11:31:44] <tjtr33> the whole 'life' of these new boards/processors is a worry. dunno what the devs will do ( beaglebone/arm choices are overwhelming )
[11:31:44] <Pitu> Hi guys.
[11:31:44] <tjtr33> hello
[11:31:48] <Pitu> I just installed LinuxCNC 2.5.1 on my laptop
[11:32:00] <L84Supper> tjtr33: it's difficult to find an ARM board with GPIO for <$60
[11:32:24] <ReadError> L84Supper: F4 discovery
[11:32:28] <ReadError> F3 discovery
[11:32:36] <L84Supper> tablets are no problem, but they don't route any of the GPIO signals
[11:32:42] <Pitu> installation went fine however I cannot run the latency test at all
[11:33:06] <Pitu> also main program doesn't start
[11:33:28] <L84Supper> cortex m3 and m4, the a10 is cortex a9 with a powerful GPU
[11:33:36] <ReadError> do you have a parallel port Pitu?
[11:33:44] <Pitu> yes I do
[11:33:48] <tjtr33> Pitu, are the apps in the Apllications menu?
[11:33:58] <micges> Pitu: try run linuxcnc from terminal and see errors
[11:33:59] <Pitu> it's a hp nw8000
[11:34:17] <Pitu> yes apps are available
[11:34:35] <Pitu> ok let me try
[11:34:52] <Pitu> there is also error log when I run from app menue
[11:34:59] <tjtr33> Pitu open a terminal and run 'latency-test'
[11:35:19] <tjtr33> the terminal will give you more information about the problem
[11:35:25] <Pitu> ok, gimme sec
[11:35:48] <pcw_home> L84Supper: cubie boards are available now
[11:36:29] <pcw_home> Working on low cost FPGA daughtercard
[11:36:54] <L84Supper> pcw_home: somebody posted a new a10 board on the arm-netbook ML, a10 on a COM with IO board
[11:37:30] <L84Supper> http://postimage.org/image/y2l75mw49/ http://s9.postimage.org/nu5ti9onj/A10.jpg
[11:37:40] <archivist> Pitu, most of us expect things will fail on a laptop due to power management
[11:37:45] <tjtr33> pcw_home, mesa plans on a fpga 'shield' for cubieboard?
[11:38:05] <pcw_home> Yes
[11:38:23] <Pitu> yes I read that but I wanted to give it a try
[11:38:54] <tjtr33> pcw_home, and an online school for fpga programming :)
[11:39:37] <pcw_home> We have pretty much all we need for FPGA motion in HM2
[11:39:53] <tjtr33> ah, bit files
[11:40:21] <pcw_home> well source (its all open source)
[11:41:04] <L84Supper> I haven't been playing with boards much lately, just mechanisms and polymers
[11:42:02] <pcw_home> Still have to play with A10 GPIO and decide best interface scheme.
[11:42:03] <pcw_home> Not much low level hardware data around
[11:42:10] <L84Supper> http://www.wandboard.org/ Wandboard started shipping a few days ago.
[11:42:34] <L84Supper> did you find a copy of the 400 or so page manual?
[11:42:37] <skunkworks> pcw_home: the local guy got the 5i25... He first tried stepconf wizard. :) I sent him the pncconf manual page and also the link to the forum with the xml and stuff.
[11:43:09] <skunkworks> If he can't get it - I think we are going to try to get together next weekend
[11:43:30] <pcw_home> Yes but the manual really has no serious data
[11:44:28] <tjtr33> "The Wandboard uses a so called EDM standard connector" arrrgh! overloaded acronyms!
[11:44:38] <Pitu_> pitu@NW8000:~/linuxcnc$ latency-test insmod: error inserting '/usr/realtime-2.6.32-122-rtai/modules/rtai_hal.ko': -1 Operation not permitted
[11:44:38] <pcw_home> Probably easiest to just use the hm2-stepper sample config and change the board name/config string
[11:45:06] <L84Supper> pcw_home: the kernel source unfortunately the only source of good info for it
[11:45:17] <Pitu_> i am not familiar with the terminal so hopefully I executed correctlz
[11:45:37] <pcw_home> I like the cubie as it includes the standard I/O on the main card but still brings out all the GPIO
[11:46:21] <tjtr33> Pitu, try same but boot from the LIVE cd ( what you got sounds like a 'permissions' or 'already there' problem )
[11:46:42] <L84Supper> http://linux-sunxi.org/A10
[11:47:03] <Pitu> ok, I will do - thanks
[11:47:26] <archivist> Pitu_ which ubuntu are you using, did you install rtai etc (realtime kernel)
[11:47:54] <tjtr33> Pitu,if permissions, then you may not have a good install, if 'already there' its just becasue you've loaded it once already and tried again ( and again :)
[11:49:58] <tjtr33> "<Pitu> I just installed LinuxCNC 2.5.1 on my laptop" oh he's gone
[11:49:59] <Pitu> i am still here
[11:49:59] <Pitu> i was logged in from the other laptop
[11:49:59] <Pitu> do copy the error message
[11:49:59] <tjtr33> did you install from a Linuxcnc cd?
[11:49:59] <Pitu> yes
[11:49:59] <L84Supper> wandboard posted schematics as well
[11:50:08] <L84Supper> http://www.wandboard.org/index.php/downloads
[11:50:31] <pcw_home> L84Supper: yes Ive seen the sunix stuff map etc but this is no hardware manual...
[11:50:45] <pcw_home> s/map/register map/
[11:50:48] <tjtr33> Pitu, ok, insert the cd , reboot to the cd, and try the terminal ( and expect poor latency results because its a laptop) this is an educational effort, dont plan on using it for realtime control
[11:51:06] <L84Supper> pcw_home: there is a 500 or so page manual thats bit better than the 80 page floating around
[11:51:09] <Pitu> it's underway
[11:51:25] <Pitu> i hope i can start the program at all
[11:52:07] <Pitu> expecting my stepper motors and boards next week so at least i wanted to conduct a basic try run
[11:52:47] <Pitu> appreciate your support
[11:53:09] <tjtr33> Pitu, why not do all this testing on the computer you plan to use later? or ... do you plan to use the laptop?
[11:53:49] <Pitu> if the laptop would work I am all in
[11:53:58] <Pitu> if not i need to look for a PC
[11:54:18] <tjtr33> Pitu, the laptop will likely NOT work well
[11:54:28] <Pitu> too bad
[11:54:43] <Pitu> now the live system is up
[11:54:55] <Pitu> should i run from terminal again
[11:55:18] <tjtr33> Pitu, the ratio of laptops that work to laptops that do not is very very small
[11:55:18] <L84Supper> pcw_home: http://dl.linux-sunxi.org/A10/ http://dl.linux-sunxi.org/A10/A10%20User%20Manual%20-%20v1.20%20%282012-04-09%2c%20DECRYPTED%29.pdf
[11:55:46] <Pitu> ok, i am preparing myself to buy a pc
[11:55:49] <tjtr33> Pitu, yes, you can open a terminal and type latency-test <return>
[11:56:25] <Pitu> damn, same error
[11:58:16] <tjtr33> Pitu, dont beat up on yourself, the laptop is not suitable. work on other things till you find a useful old PC. test the same way :)
[11:58:55] <Pitu> ok, will do
[11:59:06] <Pitu> thank's anyway
[11:59:07] <tjtr33> Pitu, where are you?
[11:59:11] <Pitu> keep up the good work guys
[11:59:15] <Pitu> munich
[11:59:50] <tjtr33> Dunno where to look in Munich for used pc's but older P4's with 512 meg to 1g should be cheap
[12:00:38] <L84Supper> http://www.3ders.org/articles/20130209-microscale-printing-of-a-spaceship-on-worlds-fastest-3d-printer.html working with something similar only much faster
[12:00:39] <Pitu> i got some friends which will probably have one for me
[12:02:12] <tjtr33> http://munich.de.craigslist.de/
[12:05:11] <L84Supper> http://munich.de.craigslist.de/cto/3555377544.html heh, why is this in a Munich list?
[12:05:46] <tjtr33> florida retiree from DE?
[12:06:27] <pcw_home> L84Supper: thanks for manual!
[12:06:55] <L84Supper> yw!
[12:10:06] <L84Supper> submicron 3d printing lets you print your own inkjet printheads and micro-nozzles
[12:10:57] <L84Supper> this will be the real game changer since inkjet tech is so secret and difficult to RE
[12:11:09] <tjtr33> but will it bond microcircuits? that requires welding & metals
[12:11:19] <L84Supper> nope
[12:11:43] <L84Supper> I can print copper traces now with fluids that have the conductivity of bulk copper
[12:12:07] <tjtr33> cool, your technologyis not simple deposition like reprap
[12:12:43] <L84Supper> no, but it combines things like FDM/SLA/SLA into one unit
[12:12:54] <L84Supper> they all have their places
[12:13:03] <tjtr33> chicago? wicker park? iirc
[12:13:18] <L84Supper> mailing address anyway
[12:13:37] <tjtr33> ok, not getting personal. your work looked great
[12:13:56] <L84Supper> no problem, thanks
[12:14:53] <L84Supper> I've been back for 7 days and spent 6 in bed
[12:15:02] <tjtr33> i had some luck marrying an edm to a linuxcnc stepper, it cut well under hal control & with linuxcnc to pos the tool
[12:15:08] <L84Supper> nasty Chinese bug
[12:15:36] <tjtr33> oh, hope you get better ... dang chinese flu's
[12:15:48] <L84Supper> had a flu shot in the US
[12:16:01] <tjtr33> right you were in the north capitol
[12:16:05] <L84Supper> this one has kicked my butt
[12:16:20] <L84Supper> and the old south one, well pretty much all over
[12:17:03] <tjtr33> didja go to a doctor in China? get little pkts of chopped up pills? "red every meal, whit pouch before sleep"
[12:17:32] <L84Supper> started getting sick when I got to Korea
[12:17:53] <L84Supper> nap time, bbl
[12:18:01] <tjtr33> take care
[12:34:21] <gene77> Hello all;
[12:34:39] <JT-Shop> nap time gene77
[12:36:23] <gene77> Put in a HOME_ALL =1 in [TRAJ] and a HOME_SEQUENCE = 0 in z, 1 in x, in the .ini. Have a HOME ALL button that does do anything?
[12:36:39] <gene77> doesn't do anything
[12:37:37] <gene77> Did I miss something in the wiki?
[12:42:06] <gene77> Found it, damned typu's, I had 2 HOME_SEQUENCE=1's, no 0. Duhh
[12:44:44] <tjtr33> sometimes they count from 0 sometimes they count from 1... no rules, just 'its my ball' by the programmer
[12:45:05] <tjtr33> live with it or you cant play :)
[12:57:01] <gene77> So I've noticed :( hal file instances in particular, fun ain't it?
[12:58:07] <tjtr33> well really ots python vs c ( arrays start at 0 lists start at 1 ) or something like that, but i just sweeps up around here
[12:59:45] <Jymmm> JT-Shop: 7-8 sticks per lb
[13:05:56] <Jymmm> WOW, nitrogen exchange has gotten expensive
[13:06:24] <Jymmm> $25 exch for 40cf
[13:07:59] <gene77> Hot damn, I've fed K's of cf in T2 bottles for 10% of that
[13:08:14] <gene77> to transmission lines
[13:08:40] <Jymmm> how many CF in T2 ?
[13:09:13] <gene77> Is the a hot key to start an auto home, biggest bottle
[13:09:45] <gene77> 2 300 cf at least
[13:10:04] <gene77> Is there a hot key to start an auto home
[13:19:40] <skunkworks> gene77: cntl home
[13:19:41] <skunkworks> http://linuxcnc.org/docs/2.5/html/gui/axis.html#_keyboard_controls
[13:22:50] <cmorley1> connor: sorry I didn't answer you. There is a way to set values directly, the set-at button but see it is greyed out. I will fix that.
[13:23:39] <Icekiller> anyone seen ich.. something?
[13:23:47] <gene77> thanks am
[13:23:58] <gene77> thanks Sam .
[13:27:54] <gene77> But ctl-h didn't work, some magic needed?
[13:28:42] <Jymmm> restart emc?
[13:31:57] <gene77> Not in the last 5 minutes, had to change a chip...
[13:42:32] <skunkworks> gene77: home key
[13:42:34] <skunkworks> not h
[13:44:25] <jdh> home key, as in 'homing'?
[13:46:25] <pcw_home> skunkworks: if you neighbor with 5I25 is not using a 7I76, I don't know if pncconf will help. he may be better off with the hm2-stepper sample config
[13:53:06] <skunkworks> ok - I will let him know
[13:59:53] <mrsun> jthornton, a router =)
[14:07:10] <mrsun> or a plasma... if router plan falls to shit :P
[14:13:52] <skunkworks> pcw_home: would your g540 sample configs be even better to start with?
[14:15:25] <skunkworks> (http://freeby.mesanet.com/g540.zip)
[14:20:20] <skunkworks> It sounds like he has some sort of rutex breakout board and their step servos
[14:20:42] <skunkworks> So - I wonder if he is going to need a special pin file?
[14:22:05] <andypugh> Last night in the hotel in Lapland. 14 days really is too long to live in a Hotel.
[14:22:30] <andypugh> (Might be bearable in a hotel with CNC machines)
[14:22:32] <skunkworks> heh - I think the longest I have ever spent is a week.. (in mexico it wasn't long enough)
[14:23:25] <andypugh> True, 2 weeks in Cancu was nice. 2 weeks in Rovanemi doing the same tests in the same cars at -25 every day is less interesting.
[14:24:08] <andypugh> In Cancun my company was my girlfriend, here my company is work colleagues.
[14:24:25] <skunkworks> heh
[14:24:42] <skunkworks> right - my company was my wife and her girlfriend.
[14:25:10] <andypugh> On balance, the work colleagues make more sense.
[14:25:17] <skunkworks> heh
[14:25:45] <skunkworks> that is great - it isn't always true though...
[14:25:58] <andypugh> skunkworks: Do you have a pallet changer set up with LinuxCNC? There was a question on the forums.
[14:26:26] <skunkworks> yes - let me look
[14:26:43] <skunkworks> (it is run manually with pyvcp buttons
[14:26:45] <skunkworks> )
[14:31:14] <skunkworks> http://electronicsam.com/images/KandT/conversion/testing/Screenshot-axis.ngc%20-%20AXIS%202.5.0~pre%20on%20HM2-Servo.png
[14:32:37] <skunkworks> andypugh: do you have a link?
[14:37:15] <andypugh> I am struggling to find it. For some reason the search function only returns two messages, neither of which contain the word "pallet"
[14:40:02] <andypugh> Ah, http://www.linuxcnc.org/index.php/forum/41-guis/17806-gscreen-a-gtk--glade--python-based-screen?start=492&lang=english#29775
[14:43:40] * JT-Shop had something to tell Andy but can't recall it now
[14:46:47] <andypugh> Do you recall if it was of significant importance?
[14:48:35] <JT-Shop> no, I'm a total blank...
[14:54:18] <JT-Shop> I have touchy with pyngcgui running on my BP :)
[15:00:43] <andypugh> Right, time to sleep. Back to normal tomorrow ;-)
[15:06:02] <JT-Shop> night
[15:13:31] <pcw_home> skunkworks: I think the G540 stuff is just the .pin, xml files for pncconf
[15:13:33] <pcw_home> If he has a special breakout it might need a custom bitfile
[15:14:33] <pcw_home> Though the g540x2 or prob_rfx2 _might_ do
[15:37:04] <skunkworks> pcw_home: when you say bit - that has to be sent to the 5i25 by utility?
[15:37:41] <pcw_home> MesaFlash
[15:38:31] <pcw_home> sudo mesaflash --device 5i25 --write foobar.bit
[15:40:26] <skunkworks> great
[15:40:46] <skunkworks> I am going to have him send me pinout info
[15:48:35] <DJ9DJ> gn8
[16:52:29] <tjtr33> hal edm with subtitles, just testing better oil bath and what it would do by itself http://videobin.org/+6cw/786.html
[17:22:05] <Jymmm> s/moonshine/oil/ =)
[17:22:15] <Jymmm> s/oil/moonshine/
[17:23:48] <L33TG33KG34R> someone here offered me some extrusion from canada, who was it?
[17:24:26] <L33TG33KG34R> I keep forgetting who it was. I just need some 5' long extrusions of 1545 for raising my CNC
[17:26:28] <spiderdijon> hi, I'm trying to adjust my axis acceleration and have got it some where I like it however it seems when the machine triggers a axis limit switch it still attempts to deccelerate instead of imeadiately stopping and it still crashes. This does not seem to happen with higher rates of acceleration. Is this behavior normal?
[17:28:57] <cradek> you would have to elaborate on the part of your findings where you report a difference between "higher" and not "higher" acceleration settings
[17:29:03] <cradek> I don't understand what you mean by "it still crashes"
[17:29:39] <cradek> but if you have your limit switches so near the end of travel that the machine cannot decelerate to a stop before running out of travel, you need to MOVE THEM inward. that is in no way a software problem.
[17:29:52] <ReadError> hey cradek, you are one of the linuxcnc devs right?
[17:31:37] <cradek> you can wire limit switches many ways, but in the most typical setup they will cause the servo amps to disable when they are hit. if you have some big mass moving, you can expect it to keep moving for a while in that unpowered state.
[17:32:10] <cradek> on a large or heavy machine, you may want to have mechanical OR electrical braking in that situation
[17:32:14] <spiderdijon> sorry yes that was a mistype, at higher MAX_ACCELERATION in the config ini file I find it stops before hitting the chassis, with lower values it crashes. I would have thought it would immeadiately stop generating step pulses when a limit switch was triggered instead of attempting to follow a acceleration (decceleration) profile.
[17:32:42] <cradek> what is your limit switch hooked to?
[17:33:04] <spiderdijon> parallel port, through some appropriate buffers
[17:33:13] <cradek> I mean in hal
[17:33:39] <spiderdijon> home and limit, is that what you mean?
[17:34:15] <cradek> perhaps you should pastebin your hal file
[17:34:23] <cradek> oops, brb
[17:34:25] <spiderdijon> ok just a sec
[17:34:26] <spiderdijon> thanks
[17:36:48] <spiderdijon> http://pastebin.com/TDAfK59f
[17:40:17] <tjtr33> and paste your .ini ( for soft limits and home method )
[17:41:41] <spiderdijon> http://pastebin.com/iqxbMCZq <-- ini
[17:43:08] <tjtr33> is err in all axis?
[17:43:36] <spiderdijon> yes, all make a clunk noise at lower acc
[17:44:29] <spiderdijon> definitely feels like it is following a profile on triggering a limit switch
[17:44:55] <tjtr33> and the distance form the switch change to the soft limit is about how far ( not 1um i hope :)
[17:47:17] <tjtr33> the soft limit has to be encountered before the physical switch. the distance between must be enuf to allow deceleration at maximum rate. (plus a but more for safety )
[17:47:28] <tjtr33> bit
[17:48:00] <spiderdijon> the soft limits do indeed stop it before the hard switch is triggered once it is homed
[17:48:16] <spiderdijon> its while it is homing it crashes at lower acc
[17:48:43] <tjtr33> crash meaning 'hits end' ?
[17:48:49] <spiderdijon> sorry yes
[17:48:58] <spiderdijon> crashes into the chassis
[17:50:42] <tjtr33> was the limit switch also the home?... yes one switch for al
[17:50:58] <tjtr33> what is the distance from the switch change to the chassis?
[17:51:37] <spiderdijon> about 10mm from when in is triggered to the chassis
[17:53:06] <tjtr33> huge for me, ok, the homing method...
[17:53:23] <tjtr33> looking at your files for the homing type
[17:55:09] <tjtr33> got an index pulse or not?
[17:55:27] <spiderdijon> no, just dir and step
[17:56:41] <tjtr33> maybe you need USE_INDEX = No http://www.linuxcnc.org/docs/2.4/html/config_ini_homing.html see 0.1.2
[17:57:48] <tjtr33> mine is similar and i have HOME_USE_INDEX = NO ( and it works, bumps swx, backs off, creeps in, then rapids to my home posn, then next axis...
[17:58:16] <spiderdijon> hmm interesting, says default is no, do I have it set somewhere?
[18:00:15] <tjtr33> .ini in each [AXIS_whatever] stanza
[18:01:51] <spiderdijon> i'll put it in, thanks
[18:02:57] <tjtr33> mine pastebin.com/trw93kp7
[18:06:04] <tjtr33> lemme know if it works plz
[18:06:22] <spiderdijon> i'll try it now
[18:07:28] <tjtr33> the hi low accel thing is something else, this effort is just recognizing the matrix of params and not crashing.
[18:08:43] <spiderdijon> hmm seems to have had no noticable effect
[18:10:18] <tjtr33> move to ctr of travel, manually trgr the swx when re-homing, does it just keep moving towards end of travel?
[18:14:29] <spiderdijon> it does stop - the homing does work, im just wondering why changing the acceleration is effecting the behavior after the switches are triggered. Same homing velocity, lower acc crashes, higher does not.
[18:16:01] <cradek> well if the acceleration is lower, you have to move farther to come to a stop
[18:16:08] <tjtr33> just thinking out loud... if the homing method requires reversal, then it must decell to 0, the reverse, and IF the accel requires a huge distance, then i slide way past the switch looking back at the second baseman :)
[18:16:14] <cradek> I don't see that this has anything to do with homing at all
[18:16:19] <spiderdijon> yes put the acc is software controlled in this case right?
[18:16:55] <spiderdijon> the physical acc of the machine is fixed, I'm only changing the value in the ini
[18:17:25] <tjtr33> pyscial acc is not whats known to the control system, the specified acc is
[18:17:31] <cradek> the stepgen manpage says if enable is false, no steps are generated. I would check with halscope to determine whether this is really true
[18:17:49] <spiderdijon> good idea, i will look at that
[18:17:57] <cradek> if you abruptly stop the steps, you WILL lose machine position
[18:18:18] <spiderdijon> yes that makes sense
[18:18:22] <cradek> that means you will have to rehome, and since you don't have encoders and index, you may not get exactly the same origin
[18:18:33] <cradek> you must pick your poison
[18:18:51] <spiderdijon> ok that clears it up for me, thanks for your time
[18:18:58] <cradek> it would probably be better to adjust your switches so you have room for a controlled stop at whatever acceleration setting you end up using
[18:19:32] <cradek> it is arguable whether turning off stepgen.enable should abruptly stop stepping, or come to a controlled stop
[18:19:58] <cradek> you could use halscope to see which of those it is doing, and file a bug report if you think it's the wrong one, being aware the whole time that half of the people will disagree with you :-)
[18:19:59] <spiderdijon> it was the reasoning for following a profile on the limit trigger which was getting me, but ofc it needs to use it to keep track of position
[18:20:18] <spiderdijon> i'll modify the switches to support the lower acc
[18:20:27] <cradek> yes if you want to maintain position you have no choice but to decel in a controlled way
[18:20:34] <cradek> I bet that's the best answer
[18:20:48] <spiderdijon> yeah seems so, thanks for helping
[18:20:50] <cradek> once you're homed you will have soft limits in effect and you should never hit the switches anyway
[18:22:10] <cradek> (I can see an argument for stepgen stopping abruptly: if you ever hit the limit switch on a correctly configured machine, you have ALREADY lost position, losing it again is no sacrifice)
[18:22:30] <spiderdijon> yes that was my original thinking
[18:22:53] <cradek> I'm not sure how I feel but I can see it's pretty debatable either way
[18:23:12] <tjtr33> "<spiderdijon> about 10mm from when in is triggered to the chassis" the accel was low enuf to allow covering 10mm distance?
[18:23:48] <cradek> I agree, if so, that seems like an incorrectly-low accel setting
[18:24:02] <tjtr33> dang low, slower than me :)
[18:24:14] <spiderdijon> that does seem wrong actually...
[18:24:24] <spiderdijon> ah
[18:24:34] <cradek> unnecessarily low accel settings cause bad performance in all sorts of different ways. be sure you have it as high as possible.
[18:25:26] <spiderdijon> the far limit switch is a bit bent out of place from the crashing, (it's the one it homes to) so it is less distance, makes more sense now :D
[18:25:59] <spiderdijon> either way I'm sorted for what I need to change
[18:28:07] <tjtr33> why are you varying the accel? measure it and record it a bit conservatively in the ini. no guessing allowed :)
[21:56:33] <gene77> silly Q: I have the bearing adjustment nut quite tight on this chinese C7 screw, and my carriage is snugged up to needing about 100 pounds to move it. It is exhibiting about 0.006" of end play despite the tight preload. Bad bearings? Stiff carraige?
[21:59:02] <gene77> I put a dial on the far end too, its 0.0045" there, probably on both ends because the drive end was a finger dial & wasn't dead square.
[22:01:29] <jdh> slop in the screw/bearing, or in the nut/screw?
[22:03:50] <gene77> I am reading the end play of the screw in the bearings
[22:04:31] <L84Supper> http://www.kickstarter.com/projects/theflexiscaleco/3d-printed-kits-of-the-ffestiniog-englands-from-la
[22:04:33] <jdh> pre-made bearing block?
[22:04:53] <L84Supper> working models would be nicer, but it's a nice start
[22:05:02] <gene77> yes, came with the screw
[22:06:09] <gene77> I took 0.140" off the bottom of it to put it pretty close to the old ones position, and lapped it flat again
[22:06:15] <L84Supper> lots of those blocks come with just standard ball bearings, not thrust or angular
[22:06:22] <gene77> one both blocks
[22:06:29] <gene77> oneboth blocks
[22:06:35] <gene77> on both blocks
[22:07:00] <gene77> should I yell at the vendor?
[22:07:15] <jdh> captured on both ends?
[22:07:48] <L84Supper> I've gotten bearings from there that had *detents* in them
[22:07:48] <gene77> no, just the driven end
[22:08:37] <gene77> tail end floats in housing
[22:08:48] <jdh> I'd pull it out and check that the nut->bearing->block is acutally touching
[22:10:26] <gene77> That will be fun, but do-able tomorrow, The load nut seems to be coming up solidly
[22:10:54] <jdh> then where is the slop
[22:12:15] <gene77> still there whn the nut is tight
[22:13:05] <gene77> and I just backed the nut off about 1/16", no chagne?
[22:14:00] <jdh> magic!
[22:14:20] <jdh> tap it with a hammer.
[22:16:05] <gene77> solid, can't drive it back thru the bearings with a reasonable tap
[22:17:20] <jdh> sounds like maybe the bearing isn't seated on the screw?
[22:18:14] <jdh> guess if it is bound, you wouldn't know and it probably wouldn't matter.
[22:18:27] <gene77> Something like that, I got out a rod and moved it about 40 thou, but the play is the same
[22:18:44] <jdh> so the bearing is moving in the block?
[22:21:42] <gene77> Dunno, I just pulled the nut up tight again, and pulled the screw back to its original position again, but the play is still there, so either there is sleeve between the bearing hubs that tool long
[22:22:31] <gene77> or something fubar. It doesn't come tight like its crushing bearings, but come solid when its tight
[22:23:00] <gene77> Dunno, I just pulled the nut up tight again, and pulled the screw back to its original position again, but the play is still there, so either there is sleeve between the bearing hubs that too long
[22:23:39] <gene77> So I guess I'll take it apart tomorrow and see what I can learn.
[22:24:09] <jdh> that's half the fun.
[22:26:15] <gene77> Yup. Looks like I can take it apart without dismounting. The major ajdjustment when putting it back together is in the taper hub I made for ther drive gear, I have to adjust the draw bolts for minimum runout.
[22:27:25] <gene77> Thanks for the shoulder to cry on. ;-)
[22:28:15] <jdh> I had to add an extra washer to mine due to a slight mismeasure.
[22:48:11] <gene77> That looks like the cure here too, there is no spacer between the bearing inners
[22:54:41] <gene77> They are axial load balls, but I'm not impressed with the finish. Where can I get a 12 bore, 14mm OD washer, .15 mm thick?
[23:02:15] <jdh> mcmaster
[23:03:10] <jdh> they are only thrust bearings?
[23:06:29] <gene77> the races are at an angle, but not a large one, centers look the same on both sides, the outers are 28mm od, 23.75 on one face for id, and 25.7 on the opposite side
[23:09:47] <gene77> If I can shim the inners .15mm, that would about do it. I'll see if the local auto parts dweebs still know what shim stock is these days tomorrow. .15mm or 5 thou, I'm not that fussy. :)
[23:11:14] <gene77> Worst part is probably having to revive my edm lashup to cut it. :(
[23:11:17] <jdh> angular contact bearings
[23:11:55] <gene77> I guess thats what you'd call them.
[23:13:20] <gene77> Midnight, no beauty sleep tonight, and at 78, I could sure use some. :) Good night & thanks
[23:24:35] <r00t4rd3d> the before picture - http://i.imgur.com/2Phhw6h.jpg